Oligonucleotide analogues with integrated bases and backbones. Part 24: Synthesis, conformational analysis, and association of aminomethylene-linked self-complementary adenosine and uridine dinucleosides

Article abstract of DOI:10.1002/hlca.201000011

Inspection of Maruzen models and force-field calculations suggest that oligonucleotide analogues integrating backbone and bases (ONIBs) with an aminomethylene linker form similar cyclic duplexes as the analogous oxymethylene linked dinucleosides. The self

Full text of DOI:10.1002/hlca.201000011

668
Helvetica Chimica Acta – Vol. 93 (2010)
Oligonucleotide Analogues with Integrated Bases and Backbones
Part 24
Synthesis, Conformational Analysis, and Association of Aminomethylene-Linked Self-
Complementary Adenosine and Uridine Dinucleosides
by Katja Chiesa, Alyena Shvoryna, Bruno Bernet, and Andrea Vasella*
Laboratorium fꢀr Organische Chemie, ETH Zꢀrich, Wolfgang-Pauli Strasse 10, CH-8093 Zꢀrich
(e-mail: vasella@org.chem.ethz.ch)
Inspection of Maruzen models and force-field calculations suggest that oligonucleotide analogues
integrating backbone and bases (ONIBs) with an aminomethylene linker form similar cyclic duplexes as
the analogous oxymethylene linked dinucleosides. The self-complementary adenosine- and uridine-
derived aminomethylene-linked A*[n]U dinucleosides 15 – 17 were prepared by an aza-Wittig reaction of
the aldehyde 10 with an iminophosphorane derived from azide 6. The sequence-isomeric U*[n]A
dinucleosides 18 – 20 were similarly prepared from aldehyde 3 and azide 12. The N-ethylamine 5, the
acetamides 7 and 14, and the amine 13 were prepared as references for the conformational analysis of the
dinucleosides. In contradistinction to the results of calculations, the N-ethylamine 5 exists as
intramolecularly H-bonded hydroxyimino tautomer. The association in CDCl3 of these dinucleosides
1
was studied by H-NMR and CD spectroscopy. The A*[n]U dinucleosides 16 and 17 associate more
strongly than the sequence isomers 19 and 20; the cyclic duplexes of 16 form preferentially Watson –
Crick-type base pairs, while 17, 19, and 20 show both Watson – Crick- and Hoogsteen-type base pairing.
The cyclic duplexes of the aminomethylene-linked dinucleosides prefer a gg-orientation of the linker. No
evidence was found for an intramolecular H-bond of the aminomethylene group. The CD spectra of 16
and 17 show a strong, those of 19 a weak, and those of 20 almost no temperature dependence.
Introduction. – Oligonucleotide analogues integrating backbone and bases
(ONIBs) replace the backbone of nucleic acids by linking elements between adjacent
nucleobases. ONIBs form cyclic duplexes and/or linear associates, as shown by
analysing the association in CDCl3 of partially protected, self-complementary
ethynylene- [1], ethenylene- [2], ethylene- [3], oxymethylene- [4][5], and thio-
methylene-linked [6] dinucleosides of the type U*[x]A(*)1)) and A*[x]U(*). The
formation of cyclic duplexes (i.e., pairing) of partially protected, self-complementary
dinucleoside analogues in CDCl3 depends on the conformation of the linker, the
orientation of the nucleobase of unit I, the sequence of nucleobases, and the
conformation of the ribofuranose ring. Pairing may occur via Watson – Crick (WC),
1
)
Conventions for abbreviated notation: The substitution at C(6) of pyrimidines and C(8) of purines is
denoted by an asterisk (*); for example, A* and U* for thiomethylated adenosine and uridine
derivatives, respectively. The moiety x linking C(6)ꢀCH2 or C(8)ꢀCH2 of unit II and C(5’) of unit I
is indicated in square brackets, i.e., [c] for a C-, [o] for a O-, [n] for a N-, and [s] for a S-atom.
ꢁ 2010 Verlag Helvetica Chimica Acta AG, Zꢀrich

Helvetica Chimica Acta – Vol. 93 (2010)
669
reverse Watson – Crick (rWC), Hoogsteen (H), or reverse Hoogsteen (rH) H-bonding
[7].
A fully deprotected thiomethylene-linked adenosine and uridine-derived tetranu-
cleoside, and analogous dinucleosides proved insoluble in H2O. Considering that
biological applications of ONIBs will require a modicum of solubility in H2O, we turned
towards designing potentially H2O-soluble ONIBs. Aminomethylene-linked oligonu-
cleosides appeared promising. They may be sufficiently soluble in H2O, either as
amines, or as ammonium salts, or upon transformation into soluble derivatives by
substitution of the amino group of the linking element with an ionisable group.
Whereas ammonium groups are excellent H-bond donors, amino groups are weak H-
donors [8 – 10], but good H-bond acceptors, and, in 2-aminoethanols, NꢀH ··· O H-
bonds cannot successfully compete with OꢀH ··· N H-bonds [11 – 14]. The amino
groups should also allow to cross-link such oligonucleosides, thereby considerably
increasing the potential of generating novel structures.
We attempted at predicting the propensity of aminomethylene-linked ONIBs to
form duplexes, and present the result of modeling studies. We also describe the
synthesis of A*[n]U and U*[n]A dinucleosides devoid of a substituent at C(6/I) of U
and at C(8/I) of A, respectively. Such substituents favour a syn-conformation, as
required for the formation of cyclic duplexes. Thus, the envisaged dinucleosides should
exhibit a higher tendency to form linear associates than the C(6/I)- and C(8/I)-
substituted analogues, and present a more demanding test of the assumption that
aminomethylene-linked ONIBs form cyclic duplexes.
Results and Discussion. – Conformational Analysis and Molecular Modelling. The
pairing propensity of ethylene- [3], oxymethylene- [4][5], and thiomethylene-linked
[6] self-complementary U*[x]A(*)and A*[x]U(*) dinucleosides was analysed by
evaluating the relative energy of the conformers of the diastereoisomeric constitutional
isomers resulting from pairing. The analysis showed that oxymethylene- and thio-
methylene-linked dinucleosides undergo pairing, but that the duplexes adopt a
different conformation. While the cyclic duplexes of oxymethylene-linked U*[o]A(*)
dinucleosides adopt a gg-conformation of the linker [4][5], those of thiomethylene-
linked U*[s]A(*) and A*[s]U(*) dinucleosides adopt a gt-conformation [6]. The greater
similarity of N to O rather than to S in terms of electronegativity and bond length
suggests that aminomethylene-linked dinucleoside duplexes should adopt a gg-
conformation.
To analyse the formation of cyclic duplexes of A*[x]U and U*[x]A dinucleosides
(x ¼ O, S, and NH) more closely, we modelled Watson – Crick H-bonded cyclic duplexes
of 2’,3’-O-isopropylidenated and 5’/II-O-methylated derivatives using the programme
AMBER* implemented in Macromodel V. 6.0 [15]. We started with U*[o]A(*) and
A*[o]U(*) dinucleosides obtained by modifying the calculated cyclic duplex of a
U*[s]A*[s]U*[s]A* tetranucleoside [16], to avoid complications from H-bonds
involving the NꢀH group. We first set the tg-conformation for the MeOC(5’)H2 group
of unit II, and then constraints to maintain the Watson – Crick base pairing (1.7 ꢂ for
the H-bonds) and p-stacking (3.2 ꢂ distance between the base pairs), and to fix either
a gg- or a gt-orientation of the linking unit (torsion angle ꢀ 60 or þ 608). After
optimisation, all constraints were released, and the structures for all U*[x]A and

670
Helvetica Chimica Acta – Vol. 93 (2010)
A*[x]U dinucleosides (x ¼ O, S, and NH) were optimized; for x ¼ NH, the two
diastereoisomers with the same configuration at the invertomers of both CH2NHCH2
groups of the duplex were calculated. To obtain additional duplex structures, different
conformations of the linking units were similarly fixed at the start of the calculations,
and the constraints were released for the final optimization. Three characteristic
properties of the AMBER* force field were observed: 1) In contradistinction to
experimental observations, NH of the NHCH2 linker appears as a powerful H-bond
donor; if a H-bond acceptor was within reach, a NꢀH ··· X (X ¼ O or N) s- or p-type
H-bond was formed. 2) The base-stacking is maintained (distance of 3.1 – 3.3 ꢂ
between base pairs; concave base pairs were formed instead of flat base pairs with a
larger distance). 3) A high-syn-conformation is feasible at the cost of an energetic
penalty (c ꢁ 1208; this features a destabilizing steric interaction of HꢀC(2’) with the
nucleobase); experimentally, a weak H-bond of HꢀC(2’) to N(3) of A or to O¼C(2) of
U is expected to favour a high-syn conformation.
The calculations confirmed the preferred gt-conformation of the U*[s]A and
A*[s]U dinucleosides (DE ¼ 2.1 and 0.4 kcal/mol, resp.), and the gg-conformation of
the U*[o]A dinucleosides (DE ¼ 2.4 kcal/mol). They predict a preferred gg-conforma-
tion also of the A*[o]U sequence isomers (DE ¼ 0.4 kcal/mol), for which no
experimental data are available.
In the U*[n]A and A*[n]U series, the calculated energy values are strongly
influenced by the H-bonds of the NHCH2 group, meaning that the conformation of the
calculated duplexes is far more reliable than the energy values. The gg-rotamers of the
duplexes are favoured by an intra-residue H-bond to the nucleobase, whereas the gt-
rotamers can at best form a weak inter-residue H-bond to O(2’/II). The definition of
the torsion angles h, q, i, and k characterizing the conformation of the linker, and some
structural data of the gg-rotamers of the U*[n]A and A*[n]U duplexes are given in
Table 1. The most favoured duplexes UA-1 and AU-1 (Fig. 1) prefer a syn-orientation
of the nucleobase and a gꢀgþtgꢀ orientation of the angles h, q, i, and k. The less favoured
duplexes adopt a tt orientation of q and i (gꢀttgꢀ for UA-2, UA-3, and AU-4 ; gꢀttgþ for
AU-2, gꢀttt for AU-3). This conformation is similar to that of the most stable conformer
of Et2NH [17]. The conformers UA-2 and AU-2 may be less disfavoured than
calculated, since the calculated conformers possess a high-syn-orientation of the
nucleobase of unit I. A gauche k angle is expected to be favoured [6], and the energy of
0.71 kcal/mol required for a gauche q angle of UA-1 and AU-1 is easily gained by base
pairing. Thus, one is led to expect that NHCH2-linked dinucleosides pair more strongly
than the corresponding OCH2-linked dinucleosides, since the conformation of the
duplex resulting from base pairing and base stacking allows the formation of weak
intramolecular H-bonds of the NHCH2 linker.
Synthesis of the A*[n]U and U*[n]A Dinucleosides. We planned to synthesize these
dinucleosides by condensing a mononucleoside-derived aldehyde with an iminophos-
phorane, and reducing the resulting imine. The iminophosphoranes should be obtained
by treating 5’-azido-5’-deoxy nucleosides with Me3P, while the aldehydes were formed,
if not isolated, in the course of the synthesis of hydroxymethylated nucleosides [18]. We
thus required the 5’-azido-5’-deoxy nucleosides 6 and 12, and the aldehydes 3 and 10
(Schemes 1 and 2). In addition, the acetamides 7 and 14, the N-ethylamine 5, and the
amine 13 were of interest as references for the conformational analysis of the

Helvetica Chimica Acta – Vol. 93 (2010)
671
Table 1. Some Structural Parameters of the Calculated gg-Configured U*[n]A and A*[n]U Duplexes Possessing
Watson – Crick Base Pairing
U*[n]A Duplexes
UA-1
UA-2
5.4
UA-3
9.5
Erel [kcal/mol]
0
c/1 (/OꢀC(1’/I)ꢀN(9/I)ꢀC(5/I)) [8]
h (/OꢀC(4’/I)ꢀC(5’/I)ꢀN) [8]
q (/C(4’/I)ꢀC(5’/I)ꢀNꢀCH2) [8]
i (/C(5’/I)ꢀNꢀCH2ꢀC(6/II)) [8]
k (/NꢀCH2ꢀC(6/II)ꢀN(1/II)) [8]
c/2 (/OꢀC(1’/II)ꢀN(1/II)ꢀ
C(2(/II)) [8]
66
ꢀ 47
104
ꢀ 170
ꢀ 68
55
119
ꢀ 54
156
ꢀ 161
ꢀ 75
57
88
ꢀ 52
172
154
ꢀ 88
63
NHCH2 H-Bond
NꢀH ··· N(3/I)
1.86 (168)
12 and 17
small
NꢀH ··· N(9/I) NꢀH ··· N(3/I)
l [ꢂ] (/NꢀH ··· X [8])
2.00 (154)
43 and 46
small
1.93 (148)
small
Propeller twist [8]
Buckle twist [8]
Roll angle [8]
14 and 15
small
small
22 and 24
A*[n]U Duplexes
AU-1
AU-2
AU-3
AU-4
11.1
77
ꢀ 41
171
Erel [kcal/mol]
0
62
2.1
125
ꢀ 58
161
156
74
11.0
68
c/1 (/OꢀC(1’/I)ꢀN(1/I)ꢀC(2/I)) [8]
h (/OꢀC(4’/I)ꢀC(5’/I)ꢀN) [8]
q (/C(4’/I)ꢀC(5’/I)ꢀNꢀCH2) [8]
i (/C(5’/I)ꢀNꢀCH2ꢀC(8/II)) [8]
k (/NꢀCH2ꢀC(8/II)ꢀN(9/II)) [8]
c/2 (/OꢀC(1’/II)ꢀN(9/II)ꢀ
C(6(/II)) [8]
ꢀ 48
ꢀ 37
ꢀ 153
ꢀ 166
ꢀ 158
39
101
ꢀ 173
ꢀ 67
58
144
ꢀ 100
44
44
NHCH2 H-Bond
NꢀH ··· O¼C(2/I) NꢀH ··· N(1/I) NꢀH ··· O¼C(2/I) NꢀH ··· O¼C(2/I)
l [ꢂ] (/NꢀH ··· X [8])
1.73 (171)
small
1.93 (156)
24 and 26
small
1.75 (175)
small
1.73 (168)
twice 30
small
Propeller twist [8]
Buckle twist [8]
Roll angle [8]
small
small
10 and 11
small
25 and 26
30 and 31
dinucleosides. We synthesized these compounds from the known 2’,3’-O-isopropylidene-
uridine (1) [19] and -adenosine 8 [20][21].
Protecting HOꢀC(5’) of 1 by silylation with thexyl(dimethyl)silyl chloride (TDS)
according to [22] gave 2 that was treated with excess LDA and DMF to yield, upon
hydrolysis [18], 98% of aldehyde 3 (Scheme 1). The azide 6 was prepared in a yield of
80% from 1 by tosylation to 4, followed by reaction with NaN3 [23][24]. To obtain the
acetamide 7, we hydrogenated 6 in the presence of Pd/C and Ac2O [24][25], while
treatment of the p-toluenesulfonate 4 with EtNH2 in DMF gave 85% of the crystalline
ethylamine 5 (m.p. 191 – 1938)2). The structure of 5 is evidenced by DQF-COSY,
HSQC, and HMBC spectra (see below and Exper. Part).
2
)
Sayed Ahmed claimed to have obtained the imido tautomer of 5 by substitution of 2’,3’-O-
isopropylidene-5’-(thiophenyl)uridine with EtNH2 in boiling MeOH [26]. Although the melting
point of 1968 is similar to that of 5, the reported 1H-NMR data are incompatible with both the
hydoxyimino and the imido structure of 5.

672
Helvetica Chimica Acta – Vol. 93 (2010)
Fig. 1. Front and top view of the Amber*-calculated duplexes UA-1 and AU-1
The adenine mononucleosides were prepared by an analogous reaction sequence as
the uridine analogues. Silylation of 8 gave the silyl ether 9 (Scheme 2). Deprotonation
of 9 with excess LDA, followed by treatment with DMF and hydrolysis, yielded 97% of
aldehyde 10. The azide 12 was obtained in 77% overall yield via the known p-
toluenesulfonate 11 [27 – 29]. Pd-Catalyzed hydrogenation of 12 in MeOH gave 80% of
the known amine 13 [28][29]. In the presence of Ac2O, the catalytic hydrogenation
afforded the acetamide 14 (76%).
With the required aldehydes and azides in hand, we proceeded to prepare the
dinucleosides. Staudinger reaction of azide 6 with Me3P in THF [30], followed by
treatment of the crude iminophosphorane 6A with aldehyde 10, generated an imine
[30][31] that was reduced in situ with NaBH3CN [32] to yield 70% of the A*[n]U
dinucleoside 15 (Scheme 3). Debenzoylation of 15 with MeONa in MeOH gave the
silyl ether 16 (90%) that was desilylated by treatment with (HF)3 · Et3N in THF [33] to
yield 88% of the alcohol 17. Desilylation with Bu4NF in THF led in an incomplete
reaction to several side-products.
The sequence-isomeric U*[n]A dinucleoside 18 was similarly synthesized by the
reaction of azide 12 with Me3P to the iminophosphorane 12A and its condensation with
aldehyde 3 (Scheme 4). The resulting imine was directly reduced to the protected
U*[n]A dinucleoside 18 (84%). Debenzoylation of 18 gave the silyl ether 19 (87%)
that was desilylated to the alcohol 20 (82%).
Conformation of the Uridine and the Adenosine Monomers. Three main factors, the
orientation of the nucleobase (syn or anti), the pucker of the furanose ring (northern or

Helvetica Chimica Acta – Vol. 93 (2010)
673
Scheme 1
a) TDSCl, 1H-imidazole, CH2Cl2; > 99%. b) LDA (¼ lithium diisopropylamide), DMF, THF, ꢀ 768;
98%. c) TsCl, pyridine; 92%. d) 70% EtNH2 in H2O, DMF, 808; 85%. e) NaN3, DMF, 808; 80%. f) H2,
Pd/C, Ac2O, MeOH; 72%.
southern), and the orientation of the ROCH2 moiety at C(4’) (gg, gt, or tg) influence
the conformation of mononucleosides and mononucleotides [34]. These factors can
easily be analyzed by 1H-NMR spectroscopy. The syn/anti-orientation of the
nucleobase may be deduced from the NOE ratio of HꢀC(6) of U or HꢀC(8) of A
with HꢀC(1’) and with HꢀC(2’)/HꢀC(3’) [35] rather than from the chemical shift of
HꢀC(2’)3). A northern (N) ring conformation is evidenced by J(1’/2’)/J(3’/4’) ꢂ 1 and a
southern (S) conformation by J(1’,2’)/J(3’,4’) ꢃ 1 [34]. The conformation of the
RXCH2 (X ¼ O, S, or NH) group at C(4’) is deduced from J(4’,5’a) and J(4’,5’b). A gg-
conformer is evidenced by a sum of J(4’,5’a) and J(4’,5’b) that is smaller than 3 Hz, and
a mixture of gt- and tg-conformers is evidenced by a sum of J(4’,5’a) and J(4’,5’b) in
excess of 12 Hz. Similar J(4’,5’a) and J(4’,5’b) values indicate a similar population of
3
)
Typical chemical shifts for 2’,3’-O-isopropylidenated derivatives in CDCl3 solution: anti-configured
U: 4.70 – 4.80 ppm, syn-configured U: 5.10 – 5.20 ppm, anti-configured A: 5.20 – 5.30 ppm, syn-
configured A: 5.70 – 5.80 ppm, and syn-configured A with an intramolecular H-bond of HOꢀC(5’)
to N(3): 5.20 – 5.30 ppm [6][1].

674
Helvetica Chimica Acta – Vol. 93 (2010)
Scheme 2
a) TDSCl, 1H-imidazole, CH2Cl2; 90%. b) LDA, DMF, THF, ꢀ 768; 97%. c) TsCl, pyridine; 86%. d)
NaN3, DMF, 808; 90%. e) H2, Pd/C, MeOH; 80%. f) 12, H2, Pd/C, Ac2O, MeOH; 76%.
the gt- and tg-conformations. The gg/gt/tg rotamer distribution may be calculated4), but
this requires an unambiguous assignment of the HꢀC(5’) signals to Hpro-RꢀC(5’) and
Hpro-SꢀC(5’). Usually, Hpro-SꢀC(5’) of ribosides and nucleosides resonates at lower field
[34 – 39], but the relative chemical shift of Hpro-RꢀC(5’) and Hpro-SꢀC(5’) may be
interchanged by the effect of the surrounding groups, e.g., by AcOꢀC(3’) [40], by a
syn-oriented nucleobase [6], or by the tertiary structure of RNA [41].
As expected, the uridine-6-carbaldehyde 3 adopts a syn-conformation. This is
evidenced by d(HꢀC(2’)) ¼ 5.09 ppm, and leads to a preferred gt- and tg-orientation of
the silyloxymethyl group (J(4’,5’a) þ J(4’,5’b) ¼ 10.8 Hz; Table 3 in the Exper. Part).
For the 6-unsubstituted uridines, a larger contribution of the syn-conformer of azide 6
than of the p-toluenesulfonate 4 is evidenced by a stronger downfield shift of HꢀC(2’)
4
)
See [6] for the calculation of the rotamer distribution of 5’-O and 5’-S nucleosides. The parameters
for 5’-amino-5’-deoxy nucleosides were calculated with the force-field programme MM3* [15]
leading to the following equations:
J(4’,5’pro-S) ¼ 3.6 Pgg þ 1.7 Pgt þ 10.9 Ptg and
J(4’,5’pro-R) ¼ 1.9 Pgg þ 10.4 Pgt þ 4.2 Ptg
We are not aware of any assignment of the HꢀC(5’) signals of 5’-nitrogenated nucleosides to
Hpro-RꢀC(5’) and Hpro-SꢀC(5’). Here, we assume that Hpro-SꢀC(5’) of the 5’-amino and 5’-azido
nucleosides resonates at lower field.

Helvetica Chimica Acta – Vol. 93 (2010)
675
Scheme 3
a) Me3P, THF. b) 1. THF; 2. NaBH3CN, AcOH, MeOH; 70% from 6. c) MeONa, MeOH; 90%. d)
(HF)3 · Et3N, THF; 88%.
(4.99 vs. 4.91 ppm). This is also reflected by a stronger preference of 6 for a gt/tg
equilibrium (J(4’,5’a) þ J(4’,5’b) ¼ 10.2 for 6 and 9.0 Hz for 4). The amide 7 adopts
completely the syn-conformation, as evidenced by a NOESY cross peak between
HꢀC(6) and HꢀC(1’), but not between HꢀC(6) and HꢀC(2’), and by the downfield
shift of HꢀC(2’) resonating at 5.15 ppm. Thus, 7 prefers the syn-conformation more
strongly than the corresponding acetate and thioacetate (d(HꢀC(2’)) ¼ 5.00 ppm)
which form a ca. 3 :1 syn/anti equilibrium [6]5). This is due to a partially persistent
intramolecular H-bond of HNꢀC(5’) to O¼C(2), as evidenced by a clear preference
for the gg-rotamer (J(4’,5’pro-S) ¼ 4.5 Hz, J(4’,5’pro-R) ¼ 3.8 Hz, gg/gt/tg ¼ 65 :18 :17),
while there is hardly a downfield shift of the AcNH signal (6.48 vs. 6.45 ppm for methyl
5-deoxy-5-acetamido-2,3-O-isopropylidene-b-d-ribofuranoside [43]). The four com-
pounds 3, 4, 6, and 7 adopt a (N)-conformation.
5
)
The stronger preference for the syn-conformation of 5’-XAc (X ¼ O, S, or NH) and 5’-N3 derivatives
of 6-unsubstituted U and of 8-unsubstituted A mononucleosides than of the corresponding 5’-OSiR3
or 5’-OTr analogues may be due to a stronger preference of the gt/tg-conformers which was already
observed in the b-d-ribofuranoside series ((J(4’,5’a) þ J(4’,5’b) ¼ 14.0 – 14.7 Hz for 5-O-acetyl- and
5-azido-5-deoxyribosides; (J(4’,5’a) þ J(4’,5’b) ¼ 8.5 – 10.0 Hz for 5-O-silyl- and 5-O-tritylribosides
[42]).

676
Helvetica Chimica Acta – Vol. 93 (2010)
Scheme 4
a) Me3P, THF. b) 1. THF; 2. NaBH3CN, AcOH, MeOH; 84% from 12. c) MeONa, MeOH; 87%. d) HF ·
pyridine, THF; 82%.
The tautomeric hydroxyimino structure of the ethylamine 5 in CDCl3 is evidenced
by the absence of an imido NH signal at low field and by a broad t (J ꢁ 4.0 Hz) at
6.15 ppm assigned to the OH group. However, the EtNH signal is not visible (fast H/D
exchange?). The ethylamine 5 adopts a southern and a gg-conformation (J(4’,5’a) þ
J(4’,5’b) ¼ 3.2 Hz; Table 3 in the Exper. Part). These structural aspects are character-
istic of an intramolecular H-bond, and an OH ··· N H-bond is indeed further evidenced
by the downfield shift of the OH signal and its coupling with the MeCH2 group.
Surprising at first view is the absence of a coupling between OH and H2C(5’), an
observation that is confirmed by the absence of a cross-peak in the DQF-COSY
spectrum between the OH and HꢀC(5’) signals. AM1 Calculations (programme
Spartan 04 for Macintosh [44]) corroborate the hypothesis of an intramolecular H-
bond (OH ··· N H-bond: 1.63 ꢂ), but result in a preferred flat (N)-conformation
(Fig. 2). The HꢀC(5’)ꢀN ··· HO torsion angles of 109 and ꢀ 1338 suggest a mean
absolute value of 1208 for both torsion angles of a flexible molecule, in agreement with
the absence of a coupling between OH and H2C(5’). The upfield shift for HꢀC(2’)
resonating at 4.86 ppm may be due to a different structure of the nucleobase, or to the
intramolecular H-bond (similarly as observed in the adenosine series; see [1] and refs.
cit. therein). The intramolecular OH ··· N H-bond of 5 persists completely in

Helvetica Chimica Acta – Vol. 93 (2010)
677
(D6)DMSO, with OH resonating as t at 7.15 ppm, and presumably also in CD3OD, as
suggested by the gg-conformation in both solvents.
Fig. 2. AM1-Calculated structures of 5 and of its imido tautomer 5a
However, according to AM1 calculations, the intramolecular H-bonded imido
tautomer 5a (Fig. 2) is more stable than 5 by 20.7 kcal/mol. Although this surprising
result agrees well with similar calculations of the tautomers of 1-methyluracil [45], it is
incompatible with our observations.
The CD spectrum of a 1-mm solution of 4 in CHCl3 at 208 shows a positive Cotton
effect with a maximum at 260 nm, suggesting an anti-conformer [46] (Fig. 3). A weak
negative Cotton effect of 6 (minimum at 290 nm) evidences a substantial amount of the
syn-conformer, and confirms the conclusion drawn from the NMR data. Finally, 5 and 7
show the typical negative Cotton effects (minimum at 270 nm) for cis-configured
uridines, again in agreement with the results of NMR analysis.
The C(8)-formylated adenosine 10 adopts a syn-conformation, as evidenced by the
downfield shift of HꢀC(2’) (5.62 ppm; Table 5 in the Exper. Part), a gt/tg equilibrium
(J(4’/5’a) þ J(4’/5’b) ¼ 12.9 Hz), and an (N)-conformation. Also the sulfonate 11, the
azide 12, and the amine 13 adopt (N)-conformations. A moderate upfield shift of their
HꢀC(2’) signal as compared to the one of 10 (Dd ¼ 0.15 to 0.34 ppm) suggests syn/anti-
equilibria for these 8-unsubstituted adenosines. This interpretation is supported by a
reduced preference for the gt/tg relative to the gg rotamer population (gg/gt/tg ratio of
ca. 2 :2 :1, as suggested by (J(4’/5’a) þ J(4’/5’b) of 10.2 – 10.4 Hz). The acetamide 14
shows all characteristics of a persistent intramolecular NꢀH ··· N(3) H-bond, i.e., the
downfield shift of AcNH (7.53 – 7.58 ppm, as compared to 6.48 ppm for 7), the upfield
shift for HꢀC(2’) (5.28 ppm), small J(4’,5’a) and J(4’,5’b) values (3.6 and 2.7 Hz), a
large J(5’a,OH) (8.1 Hz), a small J(5’b,OH) value (ca. 3.0 Hz), and a (S) ring
conformation (J(1’/2’)/J(3’/4’) ¼ 1.6; see [1][6][47][48] for the analogous OꢀH ··· N(3)
H-bond). This intramolecular H-bond is responsible for the slow exchange of AcNH
with D2O; no exchange was observed within one hour at room temperature, when

678
Helvetica Chimica Acta – Vol. 93 (2010)
Fig. 3. CD Spectra of 1-mm solutions at 208 of the A*[n]U and U*[n]A dinucleosides 4 – 7 in CDCl3
BzNH was completely exchanged. In the NOESY spectrum of 14, the intramolecular
H-bond is corroborated by a cross-peak between Me and HꢀC(2), a strong cross-peak
between NH and HꢀC(5’b), and a weak cross-peak between NH and HꢀC(5’a). The
syn-conformation is revealed by a strong cross-peak between HꢀC(8) and HꢀC(1’)
and a weak cross-peak between HꢀC(8) and HꢀC(2’). The presence of an intra-
molecular H-bond in 14, but not in 13, reflects the good H-donating properties of the
NH group of amides.
Association and Conformation of the A*[n]U and U*[n]A Dinucleosides in CDCl3.
The self-association of the A*[n]U dinucleosides 16 and 17, and of the U*[n]A
sequence isomers 19 and 20 was investigated by analysing the concentration depen-
dence of the chemical shift of HꢀN(3/I) of 16 and 17, and HꢀN(3/II) of 19 and 20
(shift/concentration curves, SCCs), and the determination of conformational aspects
from 1H-NMR spectra, similarly as it was described for thiomethylene-linked
analogues [6]. Base stacking was investigated by CD spectroscopy.
The SCCs of the A*[n]U dinucleosides 16 and 17 show the characteristics denoting
the formation of cyclic duplexes, i.e., a strong bending at concentrations below 15 mm
and formation of a plateau at concentrations above 15 mm (Fig. 4). The SCC of the

Helvetica Chimica Acta – Vol. 93 (2010)
679
U*[n]A dinucleoside 20 and especially that of 19 show a distinctly weaker bending at
low concentrations and a weak, but steady increase at concentrations above 20 mm,
revealing equilibria between the monoplex, linear associates, and cyclic duplexes. The
downfield shift of the plateau of 16 (13.02 ppm at 30 mm) indicates Watson – Crick-type
base pairing und the upfield shift of the plateau of 17 (12.40 ppm) Hoogsteen-type base
pairing (cf. [1][6][49]). Unfortunately, a broad HꢀN(3/I) signal in the ROESY
spectrum of 17 prevents the detection of cross peaks with HꢀC(2/II) and/or CH2ꢀC(8/
II). The signal of HꢀC(6/I) shows strong cross-peaks with the signals of HꢀC(1’/I) and
HꢀC(2’/I), and weak cross-peaks with the signals of HꢀC(3’/I) and H2C(5’/I),
evidencing a syn/anti-equilibrium. Thus, the upfield shift of HꢀN(3/I) is partially due
to minor amounts of linear associates possessing an anti-oriented uracil moiety.
Although substantial amounts of linear associates of 20 and especially of 19 lead to an
upfield shift for HꢀN(3/I), the chemical shift at 30 mm (12.56 and 12.27 ppm, resp.)
suggests mixtures of Watson – Crick- and Hoogsteen-type base-paired cyclic duplexes,
and a 1:1 mixture is evidenced for 19 by cross-peaks of similar intensity in the NOESY
spectrum between HꢀN(3/II) and both HꢀC(2/I) and HꢀC(8/I).
Fig. 4. Shift/concentration curves (SCCs) of the A*[n]U and U*[n]A dinucleosides 16, 17, 19, and 20 in
CDCl3 solution (including a value of 7.70 ppm for a 0.001-mm solution).
A syn-conformation of the uridine moiety of the A*[n]U dinucleosides 16 and 17 is
evidenced by d(HꢀC(2’)) of 5.18 – 5.19 ppm (Table 7 in the Exper. Part), suggesting
only a minor contribution of the anti-conformer of 17 detected by NOESY cross peaks.
The gg-orientation of the linker is more strongly preferred by 16 than by 17 as indicated
by smaller J(4’/5’a) and J(4’/5’b) values (3.1/4.4 vs. 3.5/5.4 Hz). Calculations suggest a
gg/gt/tg equilibrium of 70 :29 :1 for 16 and of 50 :40 :10 for 17. J(1’/2’)/J(3’/4’) Ratios of
1.3 and 0.8 evidence that 16 prefers a (S)- and 17 a (N)-conformation. The formation of
mainly linear associates of the benzamide 15 is suggested by the preference for the anti-
conformer (d(HꢀC(2’)) ¼ 4.94 ppm), and by a 2 :2 :1 gg/gt/tg ratio. HꢀN(5’/I) is
increasingly deshielded from 17 (1.76 ppm) via 15 (2.13 ppm) to 16 (2.73 ppm), and
may indicate an increasing H-donating NH group. A deeper analysis would require

680
Helvetica Chimica Acta – Vol. 93 (2010)
knowledge about the H-bond acceptor (O or N; p- or s-orbitals as H-acceptors). It is,
however, clear that association prevents tautomerisation to a hydroxyimino species and
formation of an OꢀH ··· N H-bond, as observed for the monomer 5.
HꢀC(2’/I) of the U*[n]A dinucleosides 18 – 20 resonates at 5.51 – 5.52 ppm
(Table 9 in the Exper. Part), at the same position as the monomeric AcOC(5’) and
AcSC(5’) analogues [6]. This may be interpreted as suggesting a ca. 85 :15 syn/anti-
equilibrium for the adenosine unit. In the NOESY spectrum of 19, cross-peaks of
similar intensity between HꢀC(8/I), and both HꢀC(1’/I) and HꢀC(2’/I) suggest even
a 1:1 syn/anti-equilibrium. Similar J(1’/2’/I) and J(3’/4’/I) values for 18 – 20 show a ca.
1:1 (N)/(S) equilibrium, and similar J(4’/5’a/I) and J(4’/5’b/I) values of 3.1 – 4.2 Hz
denote a dominant gg-rotamer (85% for 18 and 19; 70% for 20). With HNꢀC(5’/I) of
18 – 20 resonating at 1.9 – 2.2 ppm there is no spectroscopic evidence for an intra-
molecular H-bond.
Unit II of 15 – 20 shows the same conformational properties as the corresponding
mononucleosides, i.e., a syn-orientation of the nucleobase, the gt/tg-equilibrium of the
TDSOCH2 group of 15, 16, 18, and 19, a completely persistent intramolecular OꢀH ···
N(3) H-bond of 17, and a partially persistent OꢀH ··· O¼C(2) H-bond of 20.
These NMR investigations led to the following conclusions. The cyclic duplexes of
16 possess Watson – Crick-type base pairing, whereas the cyclic duplexes of 17, 19, and
20 possess both Watson – Crick- and Hoogsteen-type base pairing. The cyclic duplexes
of the U*[n]A and A*[n]U dinucleosides prefer a gg-orientation of the linker. A minor
contribution of cyclic duplexes possessing a gt-oriented linker is suggested by a
preference of the gt- over the tg-conformation of the dinucleosides (gt/tg 29 :1 for 16,
4 :1 for 17, 14 :1 for 19, and 9 :1 for 20) that is more pronounced than for the closely
related mononucleosides (gt/tg 1:1 for 6 and 2 :1 for 13). Thus, the conformational
preferences of the cyclic duplexes of the NHCH2-linked dinucleosides are indeed more
similar to those of the OCH2-linked analogues than to those of the SCH2-linked
analogues. No evidence was found for an intramolecular H-bond of the CH2NH group.
The SCCs of 16, 17, 19, and 20 (Fig. 4) were analysed numerically by the method
proposed by Gutowsky and Saika [50], including a value of 7.70 ppm for a 0.001 mm
solution, corresponding to the chemical shift of the monoplex, as deduced from
d(HꢀN(3) of monomeric uridine derivatives [6]. Inclusion of this value significantly
reduced the variance of Kass. The thermodynamic parameters (Table 2) were
1
determined by vanꢀt Hoff analysis of the H-NMR spectra recorded of ca. 2 – 5-mm
solutions in CDCl3 in intervals of 108 and in the temperature range of 7 to 508.
However, the U*[n]A dinucleosides 19 and 20 (Kass ¼ 872 and 1291mꢀ1, resp.;
Table 2) associate distinctly more weakly than the corresponding U*[o]A dinucleosides
(Kass ¼ 18400 and 12300 mꢀ1 [6]), but about as strongly as the corresponding U*[s]A
dinucleosides (Kass ¼ 198 and 1529 mꢀ1 [6]). The A*[n]U dinucleosides 16 and 17
(Kass ¼ 3454 and 2429 mꢀ1, resp.) associate 11- and 15-times more strongly than the
A*[s]U analogues (Kass ¼ 225 and 221mꢀ1; no data are available for A*[o]U
analogues). The ꢀ DH values for 19 (10.8 kcal/mol) and 20 (9.5 kcal/mol) are typical
for equilibria of the monoplex, cyclic duplexes, and substantial amounts of linear
associates [3]. The ꢀ DH values for 16 (16.6 kcal/mol) and 17 (16.8 kcal/mol) evidence
equilibria between monoplex and cyclic duplexes possessing Watson – Crick-type base
pairing. These values are larger than expected and may comprise a (small) contribution

Helvetica Chimica Acta – Vol. 93 (2010)
681
Table 2. Association Constants Kass, and Extrapolated Chemical Shift of the Monoplexes (c ¼ 0 mm) and
Duplexes (c ¼1 ) as Calculated from the Concentration Dependence of d(HN(3)) in CDCl3 at 295 K for
the A*[n]U Dinucleosides 16 and 17, and the U*[n]A Dinucleosides 19 and 20 (including a value of
7.70 ppm for a 0.001-mm solution), and Determination of the Thermodynamic Parameters by vanꢄt Hoff
Analysis of the Temperature Dependence of d(HN(3)) for ca. 2 – 5-mm Solutions in CDCl3 at 7 – 508
a
b
c
Dinucleoside
Kass [mꢀ1
]
dmonoplex
)
dduplex
[ppm]
)
ꢀ DG295
)
ꢀ DH
ꢀ DS
[ppm]
[kcal/mol]
[kcal/mol]
[cal/mol · K]
A*[n]U Series
16
17
3454 ꢄ 323
2429 ꢄ 409
7.66 ꢄ 0.08
7.67 ꢄ 0.14
13.43 ꢄ 0.06
12.80 ꢄ 0.08
4.8
4.6
16.6
16.8
38.7
41.3
U*[n]A Series
19
20
872 ꢄ 183
1291 ꢄ 71
7.77 ꢄ 0.22
7.69 ꢄ 0.06
12.97 ꢄ 0.11
13.11 ꢄ 0.04
3.9
4.2
10.8
9.5
23.4
20.3
a) Extrapolated for c ¼ 0. b) Extrapolated for c ¼1 . c) Calculated from Kass
.
of a weak intramolecular H-bond of CH2NH. The similarity of the ꢀ DH values for 16
and 17 is surprising, as a smaller ꢀ DH value is expected for 17, considering that it is a
mixture of Watson – Crick- and Hoogsteen-type base paired cyclic duplexes.
The CD spectra of the dinucleosides 16, 17, 19, and 20 were recorded of 1 mm
solutions in CHCl3 in the temperature range from 08 to 508 (Fig. 5). They show a
significant dependence of the molar ellipticity ([q] [deg · cm2 · dmolꢀ1]) on the
temperature for the A*[n]U dinucleosides 16 and 17, a weaker one for the sequence
isomer 19 and almost no temperature dependence for 20. These values denote a strong
base stacking in the cyclic duplexes of 16 and 17. The low values of the ellipticity, and
the poor temperature dependence of the CD spectra of 19 and 20 may be characteristic
for duplexes of U*[n]A dinucleosides; this may be relevant for the pairing of NHCH2-
linked ONIBs in H2O.
We expect that protonation of CH2NH moiety of the U*[n]A and A*[n]U
dinucleosides will lead to strong intramolecular H-bond of the ammonium group with
N(3/I) of the U*[n]A and possibly also with O¼C(2/I) of the A*[n]U dinucleosides.
Such H-bonds lead to a gg-conformation and to a syn-orientation of unit I, and they
prevent the formation of parallel base pairs in a cyclic duplex. Therefore, protonation is
expected to lead to linear associates only. The positive Cotton effect of 16 · TFA and
19 · TFA at ca. 260 nm agrees indeed with a syn-oriented nucleobase of unit I (Fig. 5).
We thank Syngenta AG, Basel, and F. Hoffmann-La Roche AG, Basel, for generous financial support.
Experimental Part
General. Solvents were distilled: CH2Cl2, MeOH, (i-Pr)2NH, EtN(i-Pr)2, and pyridine over CaH2;
THF over Na/benzophenone. DMF was dried over 4-ꢂ mol. sieves. Reactions were run under N2. Qual.
TLC: precoated silica-gel glass plates (Merck silica gel 60 F 254); detection by spraying with ꢃmostainꢄ
(400 ml of 10% H2SO4 soln., 20 g of (NH4)6Mo7O24 · 6 H2O, 0.4 g of Ce(SO4)2) and heating. Flash
chromatography (FC): Merck silica gel 60 (0.063 – 0.200 mm). Optical rotation: 1-dm cell at 258 and
589 nm; concentration c in g/100 ml. Temp.-dependent CD (108 steps from 08 to 508): 1-mm soln. in
CHCl3 in a 1-mm Suprasil cell. FT-IR Spectra: solid state (ATR), absorption in cmꢀ1. UV Spectra: 10ꢀ5
m

682
Helvetica Chimica Acta – Vol. 93 (2010)
Fig. 5. Temperature-dependent CD spectra (solid lines, in 108 steps from 08 to 508) and UV spectra
(dashed lines) of 1-mm solutions of the A*[n]U and U*[n]A dinucleosides 16, 17, 19, and 20 in CDCl3
soln. in CHCl3 at 208 in a 1-cm or 1-mm Suprasil cell. 1H- and 13C-NMR spectra: at 300, 400, or 500 MHz,
and 75, 100 or 125 MHz, resp. MS: high-resolution matrix-assisted laser desorption/ionization time-of-
flight mass spectrometry (HR-MALDI-TOF) with 3-hydroxypicolinic acid (3-HPA) matrix.

Helvetica Chimica Acta – Vol. 93 (2010)
683
General Procedure for NMR Studies. NMR Experiments were performed at 295 K on a Varian
Gemini 300 spectrometer (300 MHz) in CDCl3 (passed through basic aluminium oxide and dried over 4-
ꢂ mol. sieves prior to use). Experiments started at the highest concentration, with stepwise replacement
of 0.3 ml of the 0.8 ml soln. with 0.3 ml pure CDCl3. The data were analysed by non-linear least-squares
fitting using MATLAB (trust-region algorithm); the parameters were Kass, d(HꢀN(3/I or II); c ¼ 0 mm),
and d(HꢀN(3/I or II); c !1 ). The thermodynamic parameters were determined by vanꢀt Hoff analysis.
The uracil d(HꢀN(3/I or II)) was monitored at 7, 15, 22, 30, 40, and 508 and at a fixed concentration,
typically between 2 and 5 mm.
5’-O-[Dimethyl(1,1,2-trimethylpropyl)silyl]-6-formyl-2’,3’-O-isopropylideneuridine (3). A soln. of
(i-Pr)2NH (11.03 ml, 84.11 mmol) in dry THF (80 ml) was treated dropwise, at ꢀ 768 under N2, with 1.6m
BuLi in hexane (52.57 ml, 84.11 mmol), stirred for 15 min at ꢀ 768 and for 15 min at 08. The mixture was
cooled to ꢀ 768, treated dropwise with a soln. of 2 [6] (7.17 g, 16.82 mmol) in THF (80 ml) and stirred for
1.5 h, treated dropwise with dry DMF (32.52 ml, 0.42 mol), and stirred for an additional h at ꢀ 768. The
mixture was allowed to reach 248, treated with sat. NH4Cl soln. (100 ml), and neutralized with AcOH to
pH 7. After extraction with AcOEt (100 ml), the org. phase was washed with H2O (5 ꢅ 100 ml), dried
(MgSO4), and evaporated, to afford 3 (7.51 g, 98%). Yellow foam. Rf (AcOEt) 0.76. [a]2D5 ¼ þ14.5 (c ¼
1.0, CHCl3). IR (ATR): 3204w (br.), 3059w, 2957w, 2868w, 1687s, 1462m, 1373m, 1251m, 1211m, 1157w,
1
1079s, 967w, 874m, 827s. H-NMR (300 MHz, CDCl3): see Table 3; additionally, 9.69 (s, CH¼O); 8.81
(br. s, NH); 1.61 (sept., J ¼ 6.9, Me2CH); 1.57, 1.35 (2s, Me2C); 0.86 (d, J ¼ 6.9, Me2CH); 0.84 (s, Me2CSi);
0.09 (s, Me2Si). 13C-NMR (75 MHz, CDCl3): see Table 4; additionally, 184.62 (d, CH¼O); 114.47 (s,
Me2C); 33.94 (d, Me2CH); 27.09, 25.25 (2q, Me2C); 25.25 (s, Me2CSi); 20.22, 20.17 (2q, Me2CH); 18.37 (q,
Me2CSi); ꢀ 3.47 (q, Me2Si). HR-MALDI-MS: 477.2021 (100, [M þ Na]þ, C21H34N2NaO7Siþ; calc.
477.2027). Anal. calc. for C21H34N2O7Si (454.59): C 55.48, H 7.54, N 6.16; found: C 55.20, H 7.38, N 6.06.
Table 3. Selected 1H-NMR Chemical Shifts [ppm] and Coupling Constants [Hz] of the Uridine Monomers
3 – 7 in CDCl3
3
4
5a)
6
7
HꢀC(5)
HꢀC(6)
HꢀC(1’)
HꢀC(2’)
HꢀC(3’)
HꢀC(4’)
HaꢀC(5’)
HbꢀC(5’)
J(5,6)
6.22
–
5.70
7.23
5.66
4.91
4.78
4.34
4.28
4.24
8.1
5.76
5.76
5.76
7.20
5.33
5.15
4.78
4.24
3.74
3.47
8.0
7.35
5.26
4.86
4.97
4.42
4.02
3.91
8.0
7.29
5.66
4.99
4.81
4.24
3.62
3.62
8.1
6.39
5.09
4.81
4.14
3.81
3.75
–
J(1’,2’)
2.4
2.1
4.0
2.1
2.7
J(2’,3’)
6.6
6.6
6.0
6.6
6.6
J(3’,4’)
4.2
3.6
2.0
4.5
4.3
J(4’,5’a)
J(4’,5’b)
J(5’a,5’b)
4.5
6.3
11.1
3.6
5.4
10.8
2.0
1.2
11.6
5.1
4.5
3.8
14.4
5.1
b
)
a) Assignments based on a DQF-COSY and a HSQC spectrum. b) Not assigned.
2’,3’-O-Isopropylidene-5’-O-(4-toluenesulfonyl)uridine (4). Prepared according to [23]. M.p. 149 –
1508 ([51]: 1508). [a]2D5 ¼ þ15.0 (c ¼ 0.09, DMF) ([52]: [a]2D0 ¼ þ39.3 (c ¼ 0.09, DMF)). 1H-NMR
(300 MHz, CDCl3): see Table 3; additionally, 8.40 (br. s, NH); 7.77 (d, J ¼ 8.4, 2 arom. H); 7.34 (d, J ¼ 8.4,
2 arom. H); 2.45 (s, Me); 1.55, 1.33 (2s, Me2C). 13C-NMR (75 MHz, CDCl3): see Table 4; additionally,
145.16 (s); 132.34 (s); 129.79 (2d); 127.85 (2d); 114.59 (s, Me2C); 27.10, 25.27 (2q, Me2C); 21.77 (q, Me).

684
Helvetica Chimica Acta – Vol. 93 (2010)
Table 4. Selected 13C-NMR Chemical Shifts [ppm] of the Uridine Monomers 3 – 7 in CDCl3
3
4
5a)
6
7
C(2)
C(4)
C(5)
C(6)
C(1’)
C(2’)
C(3’)
C(4’)
C(5’)
149.79
161.76
111.54
147.29
90.86
84.15
81.04
149.81
163.11
102.64
142.15
94.78
84.30
80.79
85.01
69.32
152.84
169.50
105.92
143.79
97.72
83.25
81.38
149.77
162.96
102.81
142.21
94.65
84.27
81.42
150.36
163.44
102.70
143.73
96.89
83.40
81.02
87.94
63.06
85.93
61.42
85.71
52.35
85.32
41.08
a) Assignments based on a HSQC spectrum.
5’-Deoxy-5’-(ethylamino)-2’,3’-O-isopropylideneuridine (5). A soln. of 4 [23] (40 mg, 0.09 mmol) in
DMF (3 ml) was treated with 70% aq. EtNH2 (13 ml, 0.18 mmol) and stirred at 808 for 2 h. The soln. was
cooled to 238 and evaporated. FC (AcOEt/MeOH 50 :1 ! 20 :1) gave 5 (60 mg, 85%). Colourless foam.
Rf (AcOEt/MeOH 10 :1) 0.53. M.p. 191 – 1938 ([26]: 1968, but with inconsistent 1H-NMR data). [a]2D5
¼
ꢀ35.7 (c ¼ 0.4, MeOH). UV (MeOH): 227 (13300), 262 (8500). IR (ATR): 3298w (br.), 3072w, 2985w,
2936w, 2861w, 1688w, 1643s, 1612s, 1540s, 1499s, 1456m, 1374m, 1349m, 1333w, 1267w, 1248m, 1213m,
1202m, 1154m 1116m, 1094s, 1080s, 1035m, 1011m, 968w, 947w, 910m, 884w, 852m, 823m, 787w, 760w,
729s, 683w, 644w, 612w. 1H-NMR (400 MHz, CDCl3; assignments based on a DQF-COSY and a HSQC
spectrum): 6.15 (br. t, J ꢁ 4.0, HOꢀC(2)); 3.47 – 3.33 (m, MeCH2N); 1.60, 1.38 (2s, Me2C); 1.17 (t, J ¼ 7.2,
MeCH2N); signal of HꢀN(5’) not visible. 1H-NMR (300 MHz, (D6)DMSO): 7.53 (d, J ¼ 7.5, HꢀC(6));
7.15 (t, J ¼ 5.4, HOꢀC(2)); 5.63 (d, J ¼ 4.2, HꢀC(1’)); 5.56 (d, J ¼ 7.8, HꢀC(5)); 4.92 (dd, J ¼ 6.3, 4.2,
HꢀC(2’)); 4.82 (dd, J ¼ 6.3, 3.0, HꢀC(3’)); 4.13 (q, J ¼ 3.0, HꢀC(4’)); 3.64 (br. s, 2 HꢀC(5’)); 3.22 – 3.18
(m, MeCH2N); 1.53, 1.30 (2s, Me2C); 1.09 (t, J ¼ 7.2, MeCH2N). 1H-NMR (400 MHz, CD3OD;
assignments based on a DQF-COSY, a HSQC, and a HMBC spectrum): 7.70 (d, J ¼ 7.6, HꢀC(6)); 5.84
(d, J ¼ 7.6, HꢀC(5)); 5.59 (d, J ¼ 3.2, HꢀC(1’)); 5.00 – 4.96 (m, HꢀC(2’), HꢀC(3’)); 4.38 (br. q, J ꢁ 2.0,
HꢀC(4’)); 3.89 (dd, J ¼ 11.8, 2.4, HaꢀC(5’)); 3.85 (dd, J ¼ 11.8, 2.0, HbꢀC(5’)); 3.47 (q, J ¼ 7.2,
MeCH2N); 1.65, 1.42 (2s, Me2C); 1.25 (t, J ¼ 7.2, MeCH2N). 13C-NMR (100 MHz, CDCl3; assignments
based on a HSQC spectrum): 114.87 (s, Me2C); 36.75 (t, MeCH2N); 27.40, 25.33 (2q, Me2C); 14.50 (q,
MeCH2N). 13C-NMR (100 MHz, CD3OD; assignments based on a HSQC and a HMBC spectrum):
174.30 (s, C(4)); 154.54 (s, C(2)); 142.87 (d, C(6)); 116.09 (s, Me2C); 105.57 (d, C(5)); 98.45 (d, C(1’));
86.65 (d, C(4’)); 83.21 (d, C(2’)); 81.78 (d, C(3’)); 61.88 (t, C(5’)); 37.31 (t, MeCH2N); 27.32, 25.32 (2q,
Me2C); 14.46 (q, MeCH2N). HR-MALDI-MS: 334.1370 (38, [M þ Na]þ, C14H21N3NaO5þ ; calc.
334.1373), 312.1554 (100, [M þ H]þ, C14H22N3O5þ ; calc. 312.1554).
5’-Azido-5’-deoxy-2’,3’-O-isopropylideneuridine (6). Prepared according to [23][24]. [a]2D5 ¼ þ149.1
(c ¼ 0.02, DMF) ([52]: [a]2D0 ¼ þ68.5 (c ¼ 0.02, DMF)). 1H-NMR (300 MHz, CDCl3): see Table 3;
additionally, 9.21 (br. s, NH); 1.57, 1.35 (2s, Me2C). 13C-NMR (75 MHz, CDCl3): see Table 4; additionally,
114.74 (s, Me2C); 27.18, 25.32 (2q, Me2C).
5’-Acetamido-5’-deoxy-2’,3’-O-isopropylideneuridine (7). A suspension of 6 [23][24] (300 mg,
0.97 mmol), 10% Pd/C (40 mg), and Ac2O (0.18 ml, 1.94 mmol) in MeOH (5 ml) was stirred under 1 atm
of H2 at 248 for 20 h, and filtered through Celite. Evaporation and FC (CH2Cl2/MeOH 100 :1 ! 50 :1)
gave 7 (226 mg, 72%). Colourless foam. Rf (CH2Cl2/MeOH 10 :1) 0.62. [a]2D5 ¼ ꢀ42.1 (c ¼ 0.24, CHCl3).
UV (CHCl3): 257 (14300). IR (ATR): 3301w, 3059w, 2988w, 2927w, 1667s, 1547m, 1454m, 1426m, 1376s,
1264m, 1211m, 1157m, 1066s, 970w, 907w, 880w, 856m, 809m. 1H-NMR (500 MHz, CDCl3): see Table 3;
additionally, 8.54 (br. s, HꢀN(3)); 6.68 (br. dd, J ¼ 6.8, 3.8, AcNH); 2.01 (s, AcN); 1.55, 1.34 (2s, Me2C).
13C-NMR (75 MHz, CDCl3): see Table 4; additionally, 170.78 (s, MeC¼O); 114.67 (s, Me2C); 27.25, 25.32
(2q, Me2C); 23.08 (q, MeC¼O). HR-MALDI-MS: 348.1165 (83, [M þ Na]þ, C14H19N3NaO6þ ; calc.
348.1166), 326.1343 (100, [M þ H]þ, C14H20N3O6þ ; calc. 326.1347).

Helvetica Chimica Acta – Vol. 93 (2010)
685
N6-Benzoyl-5’-O-[dimethyl(1,1,2-trimethylpropyl)silyl]-8-formyl-2’,3’-O-isopropylideneadenosine
(10). A soln. of (i-Pr)2NH (16.35 ml, 0.12 mol) in dry THF (100 ml) was treated dropwise at ꢀ 788 under
N2, with 1.6m BuLi in hexane (77.95 ml, 0.12 mol), and stirred for 15 min at ꢀ 788 and for 15 min at 08.
The mixture was cooled to ꢀ 788, treated dropwise with a soln. of 9 [6] (13.80 g, 24.94 mmol) in THF
(100 ml), stirred for 1 h, treated with dry DMF (48.28 ml, 0.62 mol), and stirred for an additional h at
ꢀ 788. The mixture was allowed to reach 258, treated with sat. NH4Cl soln. (100 ml), and neutralized with
AcOH to pH 7. After extraction with AcOEt (100 ml), the org. phase was washed with H2O (4 ꢅ 100 ml),
dried (MgSO4), and evaporated to afford 10 (14.00 g, 97%). Yellow/red foam. Rf (AcOEt) 0.83. [a]D25
¼
ꢀ16.7 (c ¼ 1.0, CHCl3). IR (ATR): 3252w (br.), 2957w, 2868w, 1703m, 1604m, 1581m, 1509m, 1488m,
1464m, 1373m, 1248s, 1211m, 1157w, 1072s, 929w, 828s. 1H-NMR (300 MHz, CDCl3): see Table 5;
additionally, 10.04 (s, CH¼O); 9.16 (s, NH); 8.04 – 7.99 (m, 2 arom. H); 7.64 – 7.50 (m, 3 arom. H); 1.62,
1.38 (2s, Me2C); 1.56 (sept., J ¼ 6.9, Me2CH); 0.82 (d, J ¼ 6.9, Me2CH); 0.79, 0.78 (2s, Me2CSi); 0.00,
ꢀ 0.01 (2s, Me2Si). 13C-NMR (75 MHz, CDCl3): see Table 6; additionally, 183.24 (d, CH¼O); 164.21 (s,
NC¼O); 133.06 (s and d); 128.90 (2d); 127.79 (2d); 114.46 (s, Me2C); 34.11 (d, Me2CH); 27.30, 25.56 (2q,
Me2C); 25.31 (s, Me2CSi); 20.34 (q, Me2CSi); 18.53 (q, Me2CH); ꢀ 3.31 (q, Me2Si). HR-MALDI-MS:
582.2732 (100, [M þ H]þ, C29H40N5O6Siþ; calc. 582.2742). Anal. calc. for C29H39N5O6Si (581.74): C 59.88,
H 6.76, N 12.04; found: C 59.95, H 6.83, N 11.91.
Table 5. Selected 1H-NMR Chemical Shifts [ppm] and Coupling Constants [Hz] of the Adenosine
Monomers 10 – 14 in CDCl3
10
8.87
11
8.73
12
8.84
13
8.78
14a)
HꢀC(2)
HꢀC(8)
HꢀC(1’)
HꢀC(2’)
HꢀC(3’)
HꢀC(4’)
HaꢀC(5’)
HbꢀC(5’)
J(1’,2’)
8.70
8.04
5.89
5.28
4.84
4.42
3.95
3.32
4.2
–
8.08
6.15
5.38
5.06
4.52
4.29
4.22
2.4
8.16
6.19
5.47
5.07
4.42
3.61
3.61
2.4
8.16
6.10
5.46
5.01
4.28
3.04
2.95
3.0
6.99
5.62
5.12
4.27
3.77
3.67
2.7
J(2’,3’)
6.6
6.3
6.3
6.3
6.0
J(3’,4’)
3.9
3.3
3.6
3.3
2.7
J(4’,5’a)
J(4’,5’b)
J(5’a,5’b)
6.3
6.6
10.5
4.5
5.7
10.5
5.4
4.5
5.7
13.2
3.6
2.7
14.4
5.4
b
)
a) J(5’a,OH) ¼ 8.1, J(5’b,OH) ꢁ 3.0 Hz. b) Not assigned.
N6-Benzoyl-2’,3’-O-isopropylidene-5’-O-[(4-methylphenyl)sulfonyl]adenosine (11). Prepared ac-
1
cording to [28]. [a]2D5 ¼ ꢀ17.8 (c ¼ 1.0, CHCl3). H-NMR (300 MHz, CDCl3): see Table 5; additionally,
8.95 (br. s, NH); 8.04 (d, J ¼ 7.2, 2 arom. H); 7.67 – 7.52 (m, 5 arom. H); 7.21 (d, J ¼ 8.4, 2 arom. H); 2.40 (s,
Me); 1.61, 1.39 (2s, Me2C). 13C-NMR (75 MHz, CDCl3): see Table 6; additionally, 164.02 (s, NC¼O);
145.08 (s); 133.38 (s); 132.75 (d); 131.97 (s); 129.64 (2d); 128.77 (2d); 127.78 (2d); 127.70 (2d); 114.69 (s,
Me2C); 27.13, 25.34 (2q, Me2C); 21.72 (q, Me).
5’-Azido-N6-benzoyl-5’-deoxy-2’,3’-O-isopropylideneadenosine (12) [27][29]. A soln. of 11 (1.00 g,
1.66 mmol) in dry DMF (10 ml) was treated with NaN3 (2.16 g, 33.24 mmol) and stirred at 808 for 5 h. The
soln. was cooled to 248, diluted with AcOEt (50 ml), washed with H2O (5 ꢅ 50 ml), dried (MgSO4), and
evaporated to give 12 (656 mg, 90%). Colourless foam. Rf (AcOEt) 0.66. 1H-NMR (300 MHz, CDCl3):
see Table 5; additionally, 8.95 (br. s, NH); 8.05 – 7.99 (m, 2 arom. H); 7.65 – 7.50 (m, 3 arom. H); 1.64, 1.41

686
Helvetica Chimica Acta – Vol. 93 (2010)
Table 6. Selected 13C-NMR Chemical Shifts [ppm] of the Adenosine Monomers 10 – 14 in CDCl3
10
11
12
13
14
C(2)
C(4)
C(5)
C(6)
C(8)
C(1’)
C(2’)
C(3’)
C(4’)
C(5’)
155.37
151.87
122.83
151.99
145.33
89.79
83.39
81.88
87.37
62.89
152.43
149.56
123.36
150.74
145.08
91.04
84.10
81.38
84.59
68.91
153.09
150.06
123.84
151.37
142.38
90.94
84.29
82.03
85.70
52.49
152.58
149.62
123.63
151.18
142.14
90.81
83.71
81.81
87.74
43.86
152.11
150.20
124.29
150.77
142.72
92.10
82.49
81.38
83.94
41.04
(2s, Me2C). 13C-NMR (75 MHz, CDCl3): see Table 6; additionally, 164.86 (s, NC¼O); 133.77 (s); 133.07
(d); 129.09 (2d); 128.12 (2d); 115.22 (s, Me2C); 27.37, 25.55 (2q, Me2C).
5’-Amino-N6-benzoyl-5’-deoxy-2’,3’-O-isopropylideneadenosine (13). Prepared according to
[28][29]. [a]2D5 ¼ ꢀ48.5 (c ¼ 1.0, CHCl3). 1H-NMR (300 MHz, CDCl3): see Table 5; additionally, 8.01
(d, J ¼ 6.9, 2 arom. H); 7.59 (t, J ¼ 6.9, 1 arom. H); 7.50 (t, J ¼ 7.2, 2 arom. H); 1.62, 1.39 (2s, Me2C); signals
of BzNH and NH2 not visible. 13C-NMR (75 MHz, CDCl3): see Table 6; additionally, 164.00 (s, NC¼O);
133.47 (s); 132.69 (d); 128.73 (2d); 127.78 (2d); 114.66 (s, Me2C); 27.34, 25.46 (2q, Me2C). HR-MALDI-
MS: 433.1603 (91, [M þ Na]þ, C20H22N6NaOþ4 ; calc. 433.1595), 411.1783 (100, [M þ H]þ, C20H23N6Oþ4 ;
calc. 411.1775).
5’-Acetamido-5’-deoxy-N6-benzoyl-2’,3’-O-isopropylideneadenosine (14). A suspension of 12
(500 mg, 1.14 mmol), 10% Pd/C (40 mg), and Ac2O (0.16 ml, 1.71 mmol) in MeOH (5 ml) was stirred
under 1 atm of H2 at 248 for 20 h, and filtered through Celite. Evaporation and FC (cyclohexane/AcOEt
2 :1 ! AcOEt) gave 14 (394 mg, 76%). Colourless foam. Rf (AcOEt/MeOH 10 :1) 0.33. [a]2D5 ¼ ꢀ103.1
(c ¼ 0.9, CHCl3). UV (CHCl3): 240 (17300), 280 (20800). IR (ATR): 3277w (br.), 3074w (br.), 2987w,
2936w, 1659m, 1608s, 1579s, 1511s, 1486m, 1454s, 1373m, 1329m, 1246s, 1210s, 1156m, 1093s, 1073s, 1028m,
1002w, 969w, 933w, 893w, 866m, 796m, 756w, 707s, 643m. 1H-NMR (300 MHz, CDCl3): see Table 5;
additionally, 9.39 (br. s, exchange with D2O, BzNH); 7.98 (d, J ¼ 7.2, 2 arom. H); 7.58 – 7.53 (m, addn. of
D2O ! weak upfield shift, but no exchange, AcNH); 7.55 (t, J ¼ 7.5, 1 arom. H); 7.46 (t, J ¼ 7.5, 2 arom.
H); 2.06 (s, AcN); 1.57, 1.31 (2s, Me2C). 13C-NMR (75 MHz, CDCl3): see Table 6; additionally, 170.47 (s,
MeC¼O); 164.65 (s, NC¼O); 133.22 (s); 132.77 (d); 128.69 (2d); 127.86 (2d); 114.84 (s, Me2C); 27.49,
25.36 (2q, Me2C); 23.23 (q, MeC¼O). HR-MALDI-MS: 491.1443 (30, [M þ K]þ, C22H24KN6O5þ ; calc.
491.1440), 475.1706 (96, [M þ Na]þ, C22H24N6NaO5þ ; calc. 475.1700), 453.1886 (100, [M þ H]þ,
C22H25N6Oþ5 ; calc. 453.1881).
N6-Benzoyl-5’-O-[dimethyl(1,1,2-trimethylpropyl)silyl]-2’,3’-O-isopropylideneadenosine-8-methyl-
(81 ! 5’-N)-5’-amino-5’-deoxy-2’,3’-O-isopropylideneuridine (15). A soln. of 6 (200 mg, 0.65 mmol) in
THF (4 ml) was treated dropwise with 1m Me3P in THF (0.71 ml, 0.71 mmol) and stirred for 6 h at 238
(TLC: complete conversion of 6). The mixture was treated with a soln. of 10 (378 mg, 0.65 mmol) in
THF (3 ml), stirred for 48 h at 238, and evaporated. A soln. of the residue in AcOEt (15 ml) was washed
with H2O (2 ꢅ 20 ml), dried (MgSO4), and evaporated. A soln. of the residue in MeOH (6 ml) was
treated with NaBH3CN (163 mg, 2.60 mmol) and AcOH (0.15 ml, 2.60 mmol), stirred for 24 h at 238, and
evaporated. A soln. of the residue in AcOEt (15 ml) was washed with sat. NaHCO3 soln. and brine, dried
(MgSO4), and evaporated. FC (CH2Cl2/MeOH 50 :1) gave 15 (390 mg, 70%). Yellow solid. Rf (CH2Cl2/
MeOH 15 :1) 0.43. M.p. 124 – 1268. [a]2D5 ¼ ꢀ16.7 (c ¼ 0.46, CHCl3). UV (CHCl3): 260 (14800), 279
(14800). IR (ATR): 3204w, 2956w, 2931w, 2865w, 1690s, 1609m, 1582w, 1482w, 1455m, 1375m, 1249s,
1
1211m, 1157m, 1072s, 970w, 926w, 828s. H-NMR (300 MHz, CDCl3): see Table 7; additionally, 9.86 (s,
HꢀN(3/I)); 9.33 (s, HNꢀC(6/II)); 7.99 (d, J ¼ 6.9, 2 arom. H); 7.60 – 7.44 (m, 3 arom. H); 4.18 (s,
CH2ꢀC(8/II)); 2.13 (br. s, HNꢀC(5’/I)); 1.60, 1.53, 1.40, 1.29 (4s, 2 Me2C); 1.58 (sept., J ¼ 6.9, Me2CH);

Helvetica Chimica Acta – Vol. 93 (2010)
687
0.81 (d, J ¼ 6.9, Me2CH); 0.77, 0.76 (2s, Me2CSi); ꢀ 0.04, ꢀ 0.05 (2s, Me2Si). 13C-NMR (75 MHz, CDCl3):
see Table 8; additionally, 164.90 (s, PhC¼O); 133.54 (s); 132.58 (d); 128.62 (2d); 127.96 (2d); 114.55,
113.82 (2s, 2 Me2C); 34.10 (d, Me2CH); 27.27, 25.44 (2q, 2 Me2C); 25.26 (s, Me2CSi); 20.34 (q, Me2CSi);
18.54 (q, Me2CH); ꢀ 3.35 (q, Me2Si). HR-MALDI-MS: 871.3715 (55, [M þ Na]þ, C41H56N8NaO10Siþ;
calc. 871.3781), 849.3945 (100, [M þ H]þ, C41H57N8O10Siþ; calc. 849.3961). Anal. calc. for C41H56N8O10Si
(849.03): C 58.00, H 6.65, N 13.20; found: C 57.80, H 6.69, N 12.96.
Table 7. Selected 1H-NMR Chemical Shifts [ppm] and Coupling Constants [Hz] of the A*[n]U
Dinucleosides 15 – 17 in CDCl3
15
16a)
17a)
15
16
17
Uridine unit (I)
HꢀC(5/I)
5.62
7.25
5.59
4.94
4.82
4.20
3.04
2.97
5.71
7.19
5.30
5.19
5.06
4.36
3.13
2.99
5.72
7.20
5.34
5.18
5.00
4.29
3.01
2.97
HꢀC(6/I)
J(5,6/I)
8.1
2.4
6.6
4.5
4.2
8.0
3.7
6.4
2.8
3.1
8.1
2.7
6.6
3.5
3.5
5.4
HꢀC(1’/I)
HꢀC(2’/I)
HꢀC(3’/I)
HꢀC(4’/I)
HaꢀC(5’/I)
HbꢀC(5’/I)
J(1’,2’/I)
J(2’,3’/I)
J(3’,4’/I)
J(4’,5’a/I)
J(4’,5’b/I)
J(5’a,5’b/I)
6.0
12.6
4.4
12.3
12.3
Adenosine unit (II)
HꢀC(2/II)
8.74
4.20
4.20
6.39
5.82
5.15
4.27
3.68
3.56
8.33
4.20
3.99
6.42
5.90
5.14
4.29
3.72
3.58
8.30
4.19
4.00
6.12
5.30
5.16
4.51
3.97
3.79
CHaꢀC(8/II)
CHbꢀC(8/II)
HꢀC(1’/II)
b
J(Ha,Hb/II)
J(1’,2’/II)
J(2’,3’/II)
)
14.0
1.6
6.3
14.6
5.1
5.5
1.8
6.6
HꢀC(2’/II)
HꢀC(3’/II)
J(3’,4’/II)
3.3
2.8
1.0
HꢀC(4’/II)
J(4’,5’a/II)
J(4’,5’b/II)
J(5’a,5’b/II)
6.9
6.6
10.8
6.8
6.9
10.6
1.2
< 1.0
11.2
HaꢀC(5’/II)
HbꢀC(5’/II)
a) Assignments based on a DQF-COSY and a HSQC spectrum. b) Not assigned.
5’-O-[Dimethyl(1,1,2-trimethylpropyl)silyl]-2’,3’-O-isopropylideneadenosine-8-methyl-(81 ! 5’-N)-
5’-amino-5’-deoxy-2’,3’-O-isopropylideneuridine (16). A soln. of 15 (470 mg, 0.55 mmol) in MeOH (3 ml)
was treated with a soln. of MeONa (299 mg, 5.53 mmol) in MeOH (2 ml), stirred for 48 h at 248, and
evaporated. A soln. of the residue in AcOEt (20 ml) was washed with sat. NH4Cl soln. and brine, dried
(MgSO4), and evaporated. FC (CH2Cl2/MeOH 50 :1 ! 30 :1) gave 16 (370 mg, 90%). Yellow solid. Rf
(CH2Cl2/MeOH 10 :1) 0.47. M.p. 129 – 1308. [a]2D5 ¼ ꢀ161.0 (c ¼ 1.0, CHCl3). UV (CHCl3): 260. IR
(ATR): 3326w, 3192w, 2955w, 2865w, 1693s, 1635m, 1603m, 1574w, 1447m, 1374m, 1329w, 1251m, 1211m,
1157m, 1070s, 969w, 933w, 828s, 776m, 660w. 1H-NMR (400 MHz, CDCl3; assignments based on a DQF-
COSY and a HSQC spectrum): see Table 7; additionally, 12.72 (s, HꢀN(3/I)); 7.26 (br. s, H2NꢀC(6/II));
2.63 (br. s, HNꢀC(5’/I)); 1.59, 1.57, 1.39, 1.34 (4s, 2 Me2C); 1.53 (sept., J ¼ 6.9, Me2CH); 0.82 (d, J ¼ 6.9,
Me2CH); 0.78, 0.76 (2s, Me2CSi); ꢀ 0.04, ꢀ 0.07 (2s, Me2Si). 13C-NMR (100 MHz, CDCl3; assignments
based on a HSQC spectrum): see Table 8; additionally, 114.20, 113.44 (2s, 2 Me2C); 34.11 (d, Me2CH);
27.34, 27.21, 25.43, 25.24 (4q, 2 Me2C); 25.19 (s, Me2CSi); 20.29, 20.26 (2q, Me2CSi); 18.48, 18.44 (2q,
Me2CH); ꢀ 3.46 (q, Me2Si). HR-MALDI-MS: 767.3502 (19, [M þ Na]þ, C34H52N8NaO9Siþ; calc.
767.3519), 745.3686 (100, [M þ H]þ, C34H53N8O9Siþ; calc. 745.3699). Anal. calc. for C34H52N8O9Si
(744.92): C 54.82, H 7.04, N 15.04; found: C 55.09, H 7.00, N 14.90.
2’,3’-O-Isopropylideneadenosine-8-methyl-(81 ! 5’-N)-5’-amino-5’-deoxy-2’,3’-O-isopropylideneuri-
dine (17). In a polyethylene flask, a soln. of 16 (200 mg, 0.27 mmol) in THF (2 ml) was treated with

688
Helvetica Chimica Acta – Vol. 93 (2010)
Table 8. Selected 13C-NMR Chemical Shifts [ppm] of the A*[n]U Dinucleosides 15 – 17 in CDCl3
15
16a)
17a)
15
Adenosine unit (II)
16
17
Uridine unit (I)
C(2/I)
C(4/I)
C(5/I)
C(6/I)
151.79
151.77
163.50
103.34
143.49
151.57
163.56
103.22
143.57
C(2/II)
C(4/II)
C(5/II)
C(6/II)
152.49
149.90
122.24
153.49
148.87
46.75
89.89
83.77
82.12
87.96
152.40
151.02
118.12
155.52
150.36
46.43
90.23
83.24
82.58
88.65
63.21
152.21
150.63
118.67
155.94
149.57
46.29
92.54
82.92
81.77
163.02
102.64
142.20
C(8/II)
CH2ꢀC(8/II)
C(1’/II)
C(2’/II)
C(3’/II)
C(4’/II)
C(5’/II)
C(1’/I)
C(2’/I)
C(3’/I)
C(4’/I)
C(5’/I)
93.97
83.08
81.43
85.66
50.57
98.25
82.11
81.41
85.28
51.47
98.08
82.13
81.77
86.93
51.24
85.85
63.35
62.89
a) Assignments based on a HSQC spectrum.
(HF)3 · Et3N (0.44 ml, 2.70 mmol) and stirred for 24 h at 248. The mixture was treated with 1m NaOH and
extracted with AcOEt. The org. phase was washed with brine, dried (MgSO4), and evaporated. FC
(CH2Cl2/MeOH 80 :1 ! 20 :1) gave 17 (143 mg, 88%). Colourless foam. Rf (CH2Cl2/MeOH 10 :1) 0.23.
[a]2D5 ¼ ꢀ233.6 (c ¼ 1.0, CHCl3). UV (CHCl3): 261 (21400). IR (ATR): 3323w, 3190w, 2986w, 2936w,
2857w, 2817w, 1690s, 1639s, 1605m, 1578w, 1452m, 1374s, 1331w, 1266m 1237m, 1212s, 1156m, 1072s, 968w,
945w, 851m, 799m, 763w, 713w, 620w. 1H-NMR (500 MHz, CDCl3; assignments based on a DQF-COSY
and a HSQC spectrum): see Table 7; additionally, 12.34 (s, HꢀN(3/I)); 7.47 (br. s, H2NꢀC(6/II)); 6.85 –
6.30 (br. s, OH); 1.76 (br. s, HNꢀC(5’/I)); 1.66, 1.55, 1.39, 1.32 (4s, 2 Me2C). 13C-NMR (125 MHz, CDCl3;
assignments based on a HSQC spectrum): see Table 8; additionally, 114.11, 113.96 (2s, 2 Me2C); 27.71,
27.22, 25.28, 25.23 (4q, 2 Me2C). HR-MALDI-MS: 641.2060 (20, [M þ K]þ, C26H34KN8O9þ ; calc.
641.2080), 625.2329 (52, [M þ Na]þ, C26H34N8NaO9þ ; calc. 625.2341), 603.2511 (100, [M þ H]þ,
C26H35N8Oþ9 ; calc. 603.2522).
5’-O-[Dimethyl(1,1,2-trimethylpropyl)silyl]-2’,3’-O-isopropylideneuridine-6-methyl-(61 ! 5’-N)-5’-
amino-N6-benzoyl-5’-deoxy-2’,3’-O-isopropylideneadenosine (18). A soln. of 12 (100 mg, 0.23 mmol) in
THF (4 ml) was treated dropwise with 1m Me3P in THF (0.25 ml, 0.25 mmol), and stirred for 6 h at 238
(TLC: complete conversion of 12). The mixture was treated with a soln. of 3 (104 mg, 0.23 mmol) in
THF (3 ml), stirred for 48 h at 238, and evaporated. A soln. of the residue in AcOEt (15 ml) was washed
with H2O (2 ꢅ 20 ml), dried (MgSO4), and evaporated. A soln. of the residue in MeOH (4 ml) was
treated with NaBH3CN (58 mg, 0.92 mmol) and AcOH (0.052 ml, 0.92 mmol), stirred for 24 h at 238, and
evaporated. A soln. of the residue in AcOEt (15 ml) was washed with sat. NaHCO3 soln. and brine, dried
(MgSO4), and evaporated. FC (CH2Cl2/MeOH 100 :1 ! 50 :1) gave 18 (165 mg, 84%). Yellow foam. Rf
(CH2Cl2/MeOH 20 :1) 0.24. [a]2D5 ¼ ꢀ13.4 (c ¼ 0.8, CHCl3). UV (CHCl3): 260 (13000), 275 (13700). IR
(ATR): 3248w, 3178w, 3090w, 3041w, 2958w, 2927w, 2866w, 1689s, 1608m, 1581w, 1513w, 1484w, 1455m,
1381m, 1331w, 1249m, 1211m, 1156m, 1129w, 1072s, 1029w, 987w, 909m, 873m, 829s, 795m, 777m, 729s,
1
708s, 687w, 660w, 645m, 619w. H-NMR (500 MHz, CDCl3; assignments based on a DQF-COSY and a
HSQC spectrum): see Table 9; additionally, 10.10 (s, HꢀN(3/II)); 9.58 (s, HNꢀC(6/I)); 8.10 – 8.00 (m, 2
arom. H); 7.60 – 7.40 (m, 3 arom. H); 2.18 (br. s, HNꢀC(5’/I)); 1.64, 1.53, 1.41, 1.31 (4s, 2 Me2C); 1.57
(sept., J ¼ 6.9, Me2CH); 0.82 (d, J ¼ 6.9, Me2CH); 0.80, 0.79 (2s, Me2CSi); 0.05, 0.04 (2s, Me2Si). 13C-NMR
(125 MHz, CDCl3; assignments based on a HSQC spectrum): see Table 10; additionally, 165.05 (s,
PhC¼O); 133.32 (s); 132.87 (d); 128.75 (2d); 128.29 (2d); 114.75, 113.42 (2s, 2 Me2C); 34.01 (d, Me2CH);
27.51, 27.36, 25.67, 25.51 (4q, 2 Me2C); 25.27 (s, Me2CSi); 20.35, 20.30 (2q, Me2CSi); 18.49, 18.45 (2q,
Me2CH); ꢀ 3.30 (q, Me2Si). HR-MALDI-MS: 887.3484 (22, [M þ K]þ, C41H56KN8O10Siþ; calc.
887.3520), 871.3765 (100, [M þ Na]þ, C41H56N8NaO10Siþ; calc. 871.3781).

Helvetica Chimica Acta – Vol. 93 (2010)
689
Table 9. Selected 1H-NMR Chemical Shifts [ppm] and Coupling Constants [Hz] of the U*[n]A
Dinucleosides 18 – 20 in CDCl3
18a)
19a)
20
18
19
20
Adenosine unit (I)
HꢀC(2/I)
8.77
8.14
6.03
5.51
5.25
4.42
3.08
2.91
8.30
7.95
6.04
5.51
5.31
4.41
3.06
2.88
8.28
7.95
6.01
5.52
5.22
4.42
2.99
2.90
HꢀC(8/I)
HꢀC(1’/I)
HꢀC(2’/I)
HꢀC(3’/I)
HꢀC(4’/I)
HaꢀC(5’/I)
HbꢀC(5’/I)
J(1’,2’/I)
J(2’,3’/I)
J(3’,4’/I)
J(4’,5’a/I)
J(4’,5’b/I)
J(5’a,5’b/I)
3.5
6.3
3.3
3.4
3.6
3.1
6.8
3.4
3.4
3.1
3.0
6.3
3.0
3.3
4.2
12.3
12.3
12.3
Uridine unit (II)
HꢀC(5/II)
5.75
3.66
3.66
6.07
5.27
4.76
4.12
3.74
3.70
5.59
3.63
3.54
6.17
5.40
4.87
4.15
3.78
3.75
5.63
3.63
3.54
5.95
5.32
5.07
4.17
3.86
3.80
CHaꢀC(8/II)
CHbꢀC(8/II)
HꢀC(1’/II)
HꢀC(2’/II)
HꢀC(3’/II)
HꢀC(4’/II)
HaꢀC(5’/II)
HbꢀC(5’/II)
b
J(Ha,Hb/II)
J(1’,2’/II)
J(2’,3’/II)
)
13.8
< 1.0
6.3
14.1
1.8
6.3
< 1.0
6.3
J(3’,4’/II)
4.3
4.4
4.5
J(4’,5’a/II)
J(4’,5’b/II)
J(5’a,5’b/II)
5.6
7.3
10.7
5.5
7.3
10.7
2.7
4.5
12.3
a) Assignments based on a DQF-COSY and a HSQC spectrum. b) Not assigned.
Table 10. Selected 13C-NMR Chemical Shifts [ppm] of the U*[n]A Dinucleosides 18 – 20 in CDCl3
18a)
19a)
20
18
19
20
Adenosine unit (I)
Uridine unit (II)
C(2/II)
C(4/II)
C(5/II)
C(6/II)
C(2/I)
C(4/I)
C(5/I)
C(6/I)
C(8/I)
C(1’/I)
C(2’/I)
C(3’/I)
C(4’/I)
C(5’/I)
152.43
150.25
124.24
153.80
142.63
91.17
82.99
81.57
84.99
50.30
153.58
149.39
119.96
155.97
139.78
90.84
83.18
81.73
85.44
50.31
153.03
148.87
119.57
155.73
139.90
91.09
83.14
80.57
85.74
50.14
151.06
162.79
102.94
151.63
50.75
91.35
151.40
163.82
103.27
153.06
51.21
151.95
163.45
103.38
152.80
50.85
91.31
CH2ꢀC(6/II)
C(1’/II)
91.47
C(2’/II)
C(3’/II)
C(4’/II)
C(5’/II)
84.40
82.28
89.53
64.01
84.61
82.50
89.55
64.15
84.05
81.83
88.06
62.67
a) Assignments based on a HSQC spectrum.
5’-O-[Dimethyl(1,1,2-trimethylpropyl)silyl]-2’,3’-O-isopropylideneuridine-6-methyl-(61 ! 5’-N)-5’-
amino-5’-deoxy-2’,3’-O-isopropylideneadenosine (19). A soln. of 18 (165 mg, 0.19 mmol) in MeOH
(2 ml) was treated with a soln. of MeONa (105 mg, 1.9 mmol) in MeOH (2 ml), stirred for 48 h at 238, and
evaporated. A soln. of the residue in AcOEt (15 ml) was washed with sat. NH4Cl soln. and brine, dried
(MgSO4), and evaporated. FC (CH2Cl2/MeOH 50 :1 ! 20 :1) gave 19 (123 mg, 87%). Colourless foam.
Rf (CH2Cl2/MeOH 10 :1) 0.48. [a]2D5 ¼ ꢀ3.4 (c ¼ 1.0, CHCl3). UV (CHCl3): 259 (22700). IR (ATR):
3329w, 3192w, 2956w, 2936w, 2865w, 1693s, 1638m, 1602m, 1576w, 1457m, 1374m, 1330w, 1292w, 1251m,

690
Helvetica Chimica Acta – Vol. 93 (2010)
1
1210m, 1156m, 1127w, 1072s, 991w, 969w, 921w, 862m, 828s, 798m, 777m, 728w, 647w, 623w. H-NMR
(500 MHz, CDCl3; assignments based on a DQF-COSY and a HSQC spectrum): see Table 9;
additionally, 11.95 (s, HꢀN(3/II)); 6.60 (br. s, H2NꢀC(6/I)); 2.04 (br. s, HNꢀC(5’/I)); 1.63, 1.55, 1.41,
1.38 (4s, 2 Me2C); 1.58 (sept., J ¼ 6.9, Me2CH); 0.82 (d, J ¼ 6.9, Me2CH); 0.80, 0.79 (2s, Me2CSi); 0.04, 0.03
(2s, Me2Si). 13C-NMR (125 MHz, CDCl3; assignments based on a HSQC spectrum): see Table 10;
additionally, 114.37, 113.35 (2s, 2 Me2C); 34.10 (d, Me2CH); 27.56, 27.46, 25.86, 25.71 (4q, 2 Me2C); 25.57
(s, Me2CSi); 20.36, 20.32 (2q, Me2CSi); 18.49, 18.45 (2q, Me2CH); ꢀ 3.27 (q, Me2Si). HR-MALDI-MS:
767.3457 (59, [M þ Na]þ, C34H52N8NaO9Siþ; calc. 767.3519), 745.3685 (100, [M þ H]þ, C34H53N8O9Siþ;
calc. 745.3699). Anal. calc. for C34H52N8O9Si · MeOH (776.39): C 54.11, H 7.26, N 14.42; found: C 54.29, H
7.02, N 14.51.
2’,3’-O-Isopropylideneuridine-6-methyl-(61 ! 5’-N)-5’-amino-5’-deoxy-2’,3’-O-isopropylideneadeno-
sine (20). In a polyethylene flask, a soln. of 19 (90 mg, 0.12 mmol) in THF (2 ml) was treated with HF ·
pyridine (0.11 ml, 1.20 mmol), and stirred for 1 h at 238. The mixture was treated with 1m NaOH and
extracted with AcOEt. The org. phase was washed with brine, dried (MgSO4), and evaporated. FC
(CH2Cl2/MeOH 70 :1 ! 40 :1) gave 20 (59 mg, 82%). Colourless foam. Rf (CH2Cl2/MeOH 10 :1) 0.38.
[a]2D5 ¼ þ2.2 (c ¼ 0.23, CHCl3). UV (CHCl3): 259. IR (ATR): 3468w, 3398w, 3323w, 3211w, 2975w, 2933w,
1704s, 1690s, 1649s, 1603m, 1579m, 1500w, 1467m, 1434m, 1422w, 1375m, 1328w, 1300w, 1244m, 1211s,
1156m, 1100s, 1067s, 1049s, 991m, 974m, 940w, 909w, 866m, 844m, 828m, 797m, 769m, 717w, 642w, 617w.
1H-NMR (300 MHz, CDCl3): see Table 9; additionally, 12.52 (s, HꢀN(3/II)); 6.78 (br. s, H2NꢀC(6/I));
2.30 – 1.80 (HOꢀC(5’/II), HNꢀC(5’/I)); 1.61, 1.55, 1.39, 1.36 (4s, 2 Me2C). 13C-NMR (75 MHz, CDCl3):
see Table 10; additionally, 114.16, 113.84 (2s, 2 Me2C); 27.49, 27.48, 25.76, 25.50 (4q, 2 Me2C). HR-
MALDI-MS: 641.2074 (26, [M þ K]þ, C26H34KN8O9þ ; calc. 641.2080), 625.2332 (61, [M þ Na]þ,
C26H34N8NaOþ9 ; calc. 625.2341), 603.2512 (100, [M þ H]þ, C26H35N8Oþ9 ; calc. 603.2522).
REFERENCES
[1] X. Zhang, B. Bernet, A. Vasella, Helv. Chim. Acta 2006, 89, 2861.
[2] X. Zhang, B. Bernet, A. Vasella, Helv. Chim. Acta 2007, 90, 864.
[3] X. Zhang, B. Bernet, A. Vasella, Helv. Chim. Acta 2007, 90, 891.
[4] A. J. Matthews, P. K. Bhardwaj, A. Vasella, Chem. Commun. 2003, 950.
[5] A. J. Matthews, P. K. Bhardwaj, A. Vasella, Helv. Chim. Acta 2004, 87, 2273.
[6] A. Ritter, D. Egli, B. Bernet, A. Vasella, Helv. Chim. Acta 2008, 91, 673.
[7] W. Saenger, ꢃPrinciples of Nucleic Acid Structureꢄ, Springer-Verlag, New York, 1984, p. 556.
[8] S. Millefiori, A. Raudino, F. Zuccarello, Z. Phys. Chem. (Wiesbaden) 1980, 123, 67.
´
[9] I. Vorobyov, M. C. Yappert, D. B. DuPre, J. Phys. Chem. A 2002, 106, 668.
[10] R. Maruyama, K. Ohno, J. Phys. Chem. A 2004, 108, 4211.
´
[11] A. Lꢅmmermann, I. Szatmari, F. Fꢀlçp, E. Kleinpeter, J. Phys. Chem. A 2009, 113, 6197.
´
[12] S. C. C. Nunes, M. E. S. Eusebio, A. J. L. Jesus, M. T. S. Rosado, J. S. Redinha, J. Mol. Struct.:
THEOCHEM 2008, 863, 73.
[13] T. D. Smith, J. B. Gerken, P. V. Jog, J. D. Roberts, Org. Lett. 2007, 9, 4555.
[14] Y. Liu, C. A. Rice, M. A. Suhm, Can. J. Chem. 2004, 82, 1006.
[15] F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp, M. Lipton, C. Caufield, G. Chang, T.
Hendrickson, W. C. Still, J. Comput. Chem. 1990, 11, 440.
[16] B. Bernet, Z. Johar, A. Ritter, B. Jaun, A. Vasella, Helv. Chim. Acta 2009, 92, 2596.
[17] H. Wolff, G. Gamer, Spectrochim. Acta, Part A 1972, 28, 2121; H. Takeuchi, T. Kojima, T. Egawa, S.
Konaka, J. Phys. Chem. 1992, 96, 4389.
[18] W. Czechtizky, A. Vasella, Helv. Chim. Acta 2001, 84, 594.
[19] P. A. Levene, R. S. Tipson, J. Biol. Chem. 1934, 106, 113.
´
[20] S. Chladek, J. Smrt, Collect. Czech. Chem. Commun. 1964, 29, 214.
´
[21] L. Poppe, W. E. Hull, J. Retey, Helv. Chim. Acta 1993, 76, 2367.
[22] H. Wetter, K. Oertle, Tetrahedron Lett. 1985, 26, 5515.

Helvetica Chimica Acta – Vol. 93 (2010)
691
[23] R. Wang, D. H. Steensma, Y. Takaoka, J. W. Yun, T. Kajimoto, C.-H. Wong, Bioorg. Med. Chem.
1997, 5, 661.
ˇ
ˇ
´
´
´
[24] V. Skaric, D. Katalenic, D. Skaric, I. Salaj, J. Chem. Soc., Perkin Trans. 1 1982, 2091.
[25] T. Wada, N. Minamimoto, Y. Inaki, Y. Inoue, Chem. Lett. 1998, 1025.
[26] A. F. Sayed Ahmed, Indian J. Chem., Sect. B 1994, 33, 69.
[27] A. Gaudemer, F. Nief, R. Pontikis, J. Zylber, Org. Magn. Reson. 1977, 10, 135.
[28] P. Ciuffreda, A. Loseto, E. Santaniello, Tetrahedron 2002, 58, 5767.
[29] P. Brown, C. M. Richardson, L. M. Mensah, P. J. Oꢄ Hanlon, N. F. Osborne, A. J. Pope, G. Walker,
Bioorg. Med. Chem. 1999, 7, 2473.
[30] A. Jayanthi, V. K. Gumaste, A. R. A. S. Deshmukh, Synlett 2004, 979.
´
´
[31] P. Molina, J. Alcantara, C. Lopez-Leonardo, Tetrahedron 1996, 52, 5833.
[32] R. C. Reynolds, N. Bansal, J. Rose, J. Friedrich, W. J. Suling, J. A. Maddry, Carbohydr. Res. 1999, 317,
164.
[33] M. C. Pirrung, S. W. Shuey, D. C. Lever, L. Fallon, Bioorg. Med. Chem. Lett. 1994, 4, 1345.
[34] D. B. Davies, Prog. Nucl. Magn. Reson. Spectrosc. 1978, 12, 135.
´
[35] H. Rosemeyer, G. Toth, B. Golankiewicz, Z. Kazimierczuk, W. Bourgeois, U. Kretschmer, H.-P.
Muth, F. Seela, J. Org. Chem. 1990, 55, 5784.
[36] D. B. Davies, A. Rabczenko, J. Chem. Soc., Perkin Trans. 2 1975, 1703.
[37] P. C. Kline, A. S. Serianni, Magn. Reson. Chem. 1988, 26, 120; P. C. Kline, A. S. Serianni, Magn.
Reson. Chem. 1990, 28, 324.
[38] R. J. Parry, J. Chem. Soc., Chem. Commun. 1978, 294.
[39] M. Remin, D. Shugar, Biochem. Biophys. Res. Commun. 1972, 48, 636.
[40] R. G. S. Ritchie, A. S. Perlin, Carbohydr. Res. 1977, 55, 121.
[41] J. P. Marino, H. Schwalbe, S. J. Glaser, C. Griesinger, J. Am. Chem. Soc. 1996, 118, 4388.
´
[42] M. W. Logue, B. H. Han, Carbohydr. Res. 1983, 121, 287; C. G. Francisco, R. Freire, C. C. Gonzalez,
´
´
E. I. Leon, C. Riesco-Fagundo, E. Suarez, J. Org. Chem. 2001, 66, 1861; M. Kiso, A. Hasegawa,
Carbohydr. Res. 1976, 52, 95.
´
[43] F. Sarabia-Garcꢆa, F. J. Lopez-Herrera, Tetrahedron 1996, 52, 4757.
[44] ꢃSpartan 04 for Macintoshꢄ, Wavefunction, Inc., Irvine, CA 92612 USA.
[45] A. R. Katritzky, M. Karelson, J. Am. Chem. Soc. 1991, 113, 1561; P. ꢇ. Civcir, THEOCHEM 2000,
532, 157.
[46] D. W. Miles, M. J. Robins, R. K. Robins, M. W. Winkley, H. Eyring, J. Am. Chem. Soc. 1969, 91, 824.
[47] P. K. Bhardwaj, A. Vasella, Helv. Chim. Acta 2002, 85, 699.
[48] H. Gunji, A. Vasella, Helv. Chim. Acta 2000, 83, 2975.
[49] A. Dunger, H.-H. Limbach, K. Weisz, J. Am. Chem. Soc. 2000, 122, 10109.
[50] H. S. Gutowsky, A. Saika, J. Chem. Phys. 1953, 21, 1688.
[51] J. Baddiley, G. A. Jamieson, J. Chem. Soc. 1955, 1085.
ˇ
ˇ
[52] A. Babic, S. Gobec, C. Gravier-Pelletier, Y. Le Merrer, S. Pecar, Tetrahedron 2008, 64, 9093.
Received January 7, 2010