Oligonucleotide analogues with integrated bases and backbones. Part 19: Gelation of organic solvents by self-complementary A*[s]U(*) dinucleosides

Article abstract of DOI:10.1002/hlca.200890236

The A*[s]U(*) dinucleosides 1 and 2 form thermoreversible gels in organic solvents. The basis of the gelation is the formation of linear aggregates by base pairing following desolvation of the nucleobases. This is evidenced by the absence of gel formation by the C(6)-deaminated analogue 3 of 1, the correlation of gelation with the anti-conformation, as preferred for 1, and the temperature-, concentration-, and time-dependent CD spectra. The gels were also characterized by the minimum gelation concentration, the gel - sol transition (melting) temperature, and rheological properties.

Full text of DOI:10.1002/hlca.200890236

Helvetica Chimica Acta – Vol. 91 (2008)
2181
Oligonucleotide Analogues with Integrated Bases and Backbones
Part 19
Gelation of Organic Solvents by Self-Complementary A*[s]U(*) Dinucleosides
by Nicolas Bogliotti, Anne Ritter, Se´verine Hebbe, and Andrea Vasella*
Laboratorium fꢀr Organische Chemie, ETH Zꢀrich, Wolfgang-Pauli Strasse 10, CH-8093 Zꢀrich
(e-mail: vasella@org.chem.ethz.ch)
The A*[s]U(*) dinucleosides 1 and 2 form thermoreversible gels in organic solvents. The basis of the
gelation is the formation of linear aggregates by base pairing following desolvation of the nucleobases.
This is evidenced by the absence of gel formation by the C(6)-deaminated analogue 3 of 1, the correlation
of gelation with the anti-conformation, as preferred for 1, and the temperature-, concentration-, and time-
dependent CD spectra. The gels were also characterized by the minimum gelation concentration, the
gel – sol transition (melting) temperature, and rheological properties.
Introduction. – Oligoribonucleotide analogues integrating backbone and bases
(ONIBs) possess linking elements between the nucleobases that replace the contiguous
backbone of the parent oligonucleotides. ONIBs form linear and/or cyclic duplexes [1],
as shown by the association of ethynylene- [2], ethenylene- [3], ethylene- [4], O- [5],
and S-linked [6] self-complementary U*[x]A(*)1) and A*[x]U(*) dinucleotide ana-
logues in CDCl3. Formation of cyclic duplexes (i.e., pairing) of self-complementary
dinucleotide analogues requires a specific conformation, and depends on the nature of
the linking unit, the sequence of the nucleobases, and their substitution at C(6)
(uridine) or C(8) (adenosine).
In the course of association studies of S-linked dinucleotide analogues one of us
discovered that the A*[s]U(*) dimers 1 and 2 [6] form thermoreversible gels in organic
solvents revealing a new type of association of these analogues that is observable on a
macroscopic scale [7].
Gels are viscoelastic materials usually consisting of a liquid and an immobilizing
agent (gelator). No definition takes all their properties into account [8 – 11], and,
according to a famous statement, gels are ꢁeasier to recognize than to defineꢂ. The liquid
component can be an organic solvent, as in an organogel [12 – 14], H2O, as in a hydrogel
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 hydroxymethylated adenosine and uridine
derivatives, resp. U(*) represents both unsubstituted and hydroxymethylated uridine derivative. The
replacement of the amino group at C(6) of adenosine by an H-atom is denoted by the letter ꢁ Hꢂ in
superscript (H); for example, HA represents C(6)-deamino adenosine derivative. The moiety x
linking C(8)ꢀCH2 of unit II and C(5’) of unit I is indicated in square brackets, i.e., [s] for a S-atom.
ꢃ 2008 Verlag Helvetica Chimica Acta AG, Zꢀrich

2182
Helvetica Chimica Acta – Vol. 91 (2008)
[15 – 17], or a liquid crystal [18][19]. The gel is characterized by a structure of
macroscopic dimensions that is permanent on the time scale of an analytical
experiment and by a stress limit below which it shows a solid-like rheological
behaviour [20]. The network formed by the gelator prevents the liquid from flowing
freely. A gel is either formed irreversibly by covalently cross-linking the gelator, as in
a chemical gel, or reversibly to form a physical gel by non-covalent interactions, such
as H-bonding, p-p stacking [21], Van der Waals forces, and solvophobic effects. Phys-
ical gels exhibit a reversible gel – sol transition at a characteristic melting temperature
(Tm).
We assumed that the gelation of organic solvents by the dinucleosides 1 and 2 is due
to the formation of a network by H-bonding, conceivably leading to base stacking and
perhaps assisted by it. To reach a satisfactory level of understanding of the gels, we
wished to determine the solvents that are gelled by 1 and 2, the physico-chemical
properties of the gels, the nature of the intermolecular interactions, and the structural
elements directly involved in the association. To see whether gelation of solvents by the
A*[s]U(*) dimers 1 and 2 requires the formation of H-bonds between the comple-
mentary nucleobases rather than between the CH2OH group(s), we also planned to
synthesise the 6-deamino analogue 3 of 1, and to compare its solubility and ability to
gelate solvents to those of 1 and 2.
Results and Discussion. – 1. Synthesis of the HA*[s]U Dimer 3. We synthesised this
dinucleotide analogue from the d-ribofuranosyl-9H-purine 9 and the known C(5’)-S-
acetate 10 [6] (Scheme). To obtain 9, we debenzoylated the C(8)-substituted adenosine
4 [6] to the amine 5 (94%). Its reductive deamination with 3-methylbutyl nitrite in dry
THF at 508 according to [22] provided 6 in yields of only 30 – 40% besides several by-
products, of which the inosine 7 was isolated in ca. 5% yield. It presumably results from

Helvetica Chimica Acta – Vol. 91 (2008)
2183
substitution of the intermediate diazonium salt by H2O generated upon diazotisation of
5. The yield of 6 was indeed increased to 59% by batchwise addition of a tenfold excess
of 3-methylbutyl nitrite to a 0.05m solution of 5 in dry THF and in the presence of
activated 4-ꢄ molecular sieves at 508, although 7 was still formed. It was isolated in a
yield of 12%, suggesting that the altered reaction conditions mostly suppressed its
transformation into follow-up products, in agreement with the decreased amount and
number of side-products. Detritylation of 6 with Cl2CHCOOH and Et3SiH afforded
the alcohol 8. The alcohol was transformed into the chloride 9 (96%) by mesylation,
completing the transformation of the intermediate methanesulfonate by adding 1 equiv.
of LiCl. A solution of 9 and the S-acetate 10 in MeOH was treated with MeONa,
resulting in S-deacetylation, nucleophilic substitution, and N-debenzoylation, and
leading to the HA*[s]U dimer 11 that was desilylated with (HF)3 · Et3N to give the
desired dinucleoside 3 in 95% yield.
Scheme
a) MeONa, MeOH, 08 to 258; 94%. b) 3-Methylbutyl nitrite, THF, mol. sieves (4 ꢄ), 508; 59% of 6, 12%
of 7. c) Et3SiH, Cl2CHCO2H, CH2Cl2; 68%. d) MsCl, EtN(i-Pr)2, CH2Cl2, 08 to 258, then LiCl; 96%. e)
MeONa, MeOH, 08 to 258; 98%. f) (HF)3 · Et3N, THF, 258; 95%.

2184
Helvetica Chimica Acta – Vol. 91 (2008)
2. Conformation of the Adenosine-Derived Monomers. The purine nucleosides 5, 6,
and 9 in CDCl3 adopt a syn-conformation, as revealed by d(HꢀC(2’)) of 5.81, 5.85, and
5.84 ppm, respectively (see Table 1). J(1’,2’)/J(3’,4’) of 0.74, 0.75, and 0.69, respectively,
evidence a predominant N conformation. J(4’,5’a) and J(4’,5’b) values for 5 and 6 (6.2 –
6.7 Hz) agree with a ca. 1:1 gt/tg mixture, typical for S-linked dinucleosides [6], while 9
shows a higher proportion of the gg-rotamer (gg/gt/tg 23 :36 :41), as revealed by
J(4’,5’a) and J(4’,5’b) values of 5.6 and 5.7 Hz, respectively2). The upfield shift for
HꢀC(2’) (5.62 ppm) and the downfield for HꢀC(1’) (6.37 ppm) of 8, as compared to 5,
6, and 9 (d(HꢀC(1’)) ¼ 6.19 – 6.32 ppm), reflect a ca. 4 :1 syn/anti mixture (see [6]).
Conceivably, an intramolecular H-bond between HOCH2ꢀC(8) (br. s at 4.70 ppm) and
OꢀC(5’) decreases the destabilization of the anti-conformer by this CH2OH group (cf.
[6]). The interpretation is in keeping with J(4’,5’a) and J(4’,5’b) values of 5.0 Hz,
evidencing a higher population of the gg-conformer of 8 (gg/gt/tg 36 :30 :34) than of 5,
6 (gt/tg ca. 1:1), and 9 (gg/gt/tg 23 :36 :41). The ratio of the ribofuranose conformers of
8 is not affected by the intramolecular H-bond, as revealed by J(1’,2’)/J(3’,4’) ¼ 0.71,
similar to the ratio observed for 5, 6, and 9.
Table 1. Selected 1H-NMR Chemical Shifts [ppm] and Coupling Constants [Hz] of the Adenosine-
Derived Monomers 5 – 9 in CDCl3
5
6
7
8
9
HꢀC(2)
HꢀC(6)
CHaꢀC(8)
CHbꢀC(8)
HꢀC(1’)
HꢀC(2’)
HꢀC(3’)
HꢀC(4’)
HaꢀC(5’)
HbꢀC(5’)
J(Ha,Hb)
J(1’,2’)
8.29
–
8.96
9.11
4.59
4.54
6.27
5.85
5.06
4.18
3.78
3.66
8.05
–
8.92
9.02
4.99
4.94
6.37
5.62
5.06
4.24
3.80
3.67
8.96
9.10
4.93
4.89
6.32
5.84
5.11
4.28
3.70
3.60
4.45
4.40
6.19
5.81
5.02
4.17
3.78
3.63
4.45
4.38
6.10
5.62
4.91
4.09
3.74
3.65
11.8
2.5
11.8
2.7
11.5
2.8
14.7
2.7
12.7
2.4
J(2’,3’)
6.5
6.6
6.7
6.6
6.4
J(3’,4’)
3.4
3.6
3.8
3.8
3.5
J(4’,5’a)
J(4’,5’b)
J(5’a,5’b)
6.7
6.4
11.8
6.2
6.2
10.7
6.1
6.2
10.7
5.0
5.0
11.0
5.6
5.7
10.9
The chemical shift for HꢀC(2’) of 7 (5.62 ppm) is lower than the one for 5 and 6,
while the S/N ratio of 7 is similar, as expressed by J(1’,2’)/J(3’,4’) ¼ 0.74; also J(4’,5’a)
and J(4’,5’b) values are similar (6.1 – 6.2, as compared to 6.2 – 6.7 Hz), revealing a ca.
1:1 gt/tg rotamer distribution. The upfield shift for HꢀC(2’) of 7 (Dd 0.19 – 0.23 ppm),
reflecting a different syn-conformation, or a higher proportion of the anti-conformer,
shows a subtle influence of the nature of the nucleobase.
2
)
See [6] for the calculation of the rotamer distribution.

Helvetica Chimica Acta – Vol. 91 (2008)
2185
3. Conformation and Association of the HA*[s]U Dimers in CHCl3. The deaminated
purine moiety in unit II of the HA*[s]U dimers 11 and 3 cannot form Watson – Crick- or
Hoogsteen-type H-bonds, while U · U base pairing [23] may lead to linear dimers, but
not to higher associates. The conformation of unit I is similar for the two HA*[s]U
dimers 11 and 3. The chemical shift for HꢀC(2’/I) of 11 and 3 (4.95 and 4.94 ppm, resp.;
Table 2) is typical for a ca. 1 :1 syn/anti orientation of the uracil moiety (see [6]). The
predominant N conformation of the furanose ring of unit I of 11 and 3 is characterized
by the same J(1’,2’)/J(3’,4’) ratio of 0.49. J(4’,5’a ¼ 6.3 Hz) and J(4’,5’b ¼ 6.4 Hz) for
unit I of 11, and the similar values of 6.0 and 6.3 Hz for 3 evidence a ca. 20 :40 :40 and a
ca. 24 :38 :38 gg/gt/tg equilibrium, respectively.
Table 2. Selected 1H-NMR Chemical Shifts [ppm] and Coupling Constants [Hz] of the HA*[s]U Dimers
11 and 3 in CDCl3
11
3
11
3
71 mm
26 mm
71 mm
26 mm
Uridine unit (I)
HꢀN(3/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)
Adenosine-derived unit (II)
9.71
5.59
4.95
4.71
4.27
2.95
2.92
2.1
6.6
4.3
6.3
6.4
9.58
5.58
4.94
4.67
4.26
3.00
2.94
2.1
6.6
4.3
6.0
6.3
HꢀC(6/II)
HꢀC(2/II)
CHaꢀC(8/II)
CHbꢀC(8/II)
HꢀC(1’/II)
HꢀC(2’/II)
HꢀC(3’/II)
HꢀC(4’/II)
HaꢀC(5’/II)
HbꢀC(5’/II)
J(Ha,Hb/II)
J(1’,2’/II)
9.06
8.91
4.18
4.07
6.30
5.94
5.11
4.26
3.62
3.53
9.10
8.90
4.07
4.07
6.14
5.30
5.11
4.52
3.96
3.79
)
5.0
5.9
1.6
2.0
a
14.6
2.1
6.3
3.2
6.2
6.1
10.7
–
14.0
14.0
J(2’,3’/II)
J(3’,4’/II)
J(4’,5’a/II)
J(4’,5’b/II)
J(5’a,5’b/II)
J(HO,5’a/II)
J(HO,5’b/II)
2.3
12.8
2.0
–
11.0
a) Not assigned.
The nature of the substituent at C(5’/II) influences the conformation of unit II. The
silyl ether 11 adopts a syn conformation, as revealed by d(HꢀC(2’/II)) ¼ 5.94 ppm. A
ca. 24 :40 :36 gg/gt/tg equilibrium is deduced from J(4’,5’a/II) ¼ 6.2 Hz and J(4’,5’b/
II) ¼ 6.1 Hz. A J(1’,2’/II)/J(3’,4’/II) ratio of 0.62 evidences a predominant N
conformation. HOꢀC(5’/II) of 3 forms a persistent intramolecular H-bond to N(3),
as evidenced by the dominant population of the gg-rotamer (J(4’,5’a/II) ¼ 2.0 Hz,
J(4’,5’b /II) ¼ 2.3 Hz), the preferred S conformation of the furanose (J(1’,2’)/J(3’/4’) ¼
3.1), the chemical shift of HOꢀC(5’/II) (5.61 ppm), the upfield shift of HꢀC(2’/II)
(5.30 ppm), and the typical J(5’,OH) values of 2.0 and 11.0 Hz (cf. [2][6][24]).

2186
Helvetica Chimica Acta – Vol. 91 (2008)
The self-association of 11 and 3 was investigated by analysing the concentration
dependence of the chemical shift for HꢀN(3/I). The HA*[s]U dimers 11 and 3
associate but weakly (Kass ¼ 23 and 20 mꢀ1, resp.) and form linear associates, as shown
by the weak curvature of the shift/concentration curves (SCC; Fig. 1), the absence of a
plateau at concentrations between 20 and 50 mm, and the numerical analysis according
to the method of Gutowsky and Saika [25][6] (Table 3). As expected, the association
constants for the HA*[s]U dimers 11 and 3 are slightly higher than those for 3’,5’-O-
diacetyl-2’-deoxyuridine (12.7 mꢀ1) [26] and 1-cyclohexyluracil (3.2 mꢀ1 [27] to 6.1mꢀ1
[28]), but an order of magnitude lower than those for self-complementary A*[s]U
dimers (221 – 225 mꢀ1, see [6]).
Fig. 1. Shift/concentration curves (SCCs) for the HA*[s]U dimers 11 and 3 in CDCl3 solution (including a
value of 7.70 ppm for a 0.0001-mm soln.)
Table 3. Association Constants Kass from the Concentration Dependence of d(HN(3)) in CDCl3 at 295 K
for the HA[s]U Dimers 11 and 3 (including a value of 7.70 ppm for a 0.0001-mm soln.), DG295 Values, and
Extrapolated Chemical Shifts of the Monoplexes and Duplexes
Dimer
Kass [mꢀ1
]
ꢀ DG295a) [kcal/mol]
dmonoplexb) [ppm]
dduplexc) [ppm]
11
3
23
20
1.8
1.8
7.74
7.74
10.88
11.17
a) Calculated from Kass b) Extrapolated for 0 mm. c) Extrapolated for infinite concentration.
.
The weak self-association of 11 and 3 in CHCl3 is also evidenced by the CD spectra,
the molar ellipticity ([q]) ranging from ca. ꢀ 5 · 103 to þ 104 deg · cm2 · dmolꢀ1, as
compared to [q] of the corresponding A*[s]U dimers, ranging from ca. ꢀ 6 · 104 to þ 2 ·

Helvetica Chimica Acta – Vol. 91 (2008)
2187
104 deg · cm2 · dmolꢀ1 [6]3). The CD spectra of 1 mm solutions of 11 and 3 in CHCl3 in
the temperature range of 0 – 508 (Fig. 2) show a weak variation of the ellipticity, in
agreement with poor stacking, as expected for weakly associating species. Two positive
Cotton effects around 270 and 245 nm are observed for 11, whereas 3 shows a negative
effect around 280 nm.
Fig. 2. Temperature-dependent CD spectra (solid lines, in 108 steps from 08 to 508) and UV spectra
(dashed lines) of 1-mm solutions of a) 11 and b) 3 in CHCl3
4. Characterization of the Gels. 4.1. Solvents. We first evaluated the gelation of a
selection of solvents by the A*[s]U(*) dimers 1 and 2, and by the HA*[s]U dimer 3. The
selection is based on Chastretteꢂs classification of solvents into nine categories that
agrees well with chemical intuition [29]. To overcome its limitations, we modified it by
introducing a class of aliphatic apolar solvents, including solvents from an ꢁunexpectedꢂ
class [29] and solvents that were not considered in the original classification, according
to structural analogy. This resulted in the 33 solvents in Table 4.
The solubility of 1, 2, and 3 in the selected solvents, and the properties of the
resulting gels were evaluated at a gelator concentration of 1% (w/v), considering the
availability of the dinucleosides and the efficiency of other nucleobase-containing
organogelators [8][30 – 33] (Table 4; see Exper. Part for details).
Inspection of Table 4 shows that 3, the deaminated analogue of 1, does not form a
gel with any of the solvents tested, while the A*[s]U dimer 1 forms partial gels4) in
acetophenone, CH2Cl2, acetone, and CHCl3, and clear or turbid gels in many of the
aprotic dipolar and H-bonding solvents. Turbid gels presumably contain aggregates of a
size comparable to the wavelength of visible light, whereas clear gels contain smaller
aggregates [34]. The A*[s]U* dimer 2 is soluble in a larger number of solvents than 1.
While 1 forms gels in most of the alcohols tested, 2 is soluble in all of them. It forms
partial gels in benzene, toluene, and CH2Cl2, and turbid gels in MeCN and AcOEt. In
most solvents, 3 is about as soluble as 2. No significant increase of the viscosity of the
solutions of 3 was observed that could indicate the formation of associates.
3
)
)
For the sake of comparison, the ellipticity (q [mdeg]) values reported in [6] were converted to [q]
[deg · cm2 · dmolꢀ1].
See Exper. Part for a definition and evaluation of partial gels.
4

2188
Helvetica Chimica Acta – Vol. 91 (2008)
Table 4. Solubility of the Dinucleosides 1, 2, and 3 in Selected Solvents, and Properties of the Gelsa)
Class
Solvent
1
2
3
Aliphatic, apolarb)
Pentaneb)
Hexanec)
Ie)
I
I
Ie)
I
I
Ie)
I
I
Cyclohexanec)
c
CCl4
)
I
I
I
Aromatic, apolar
Benzene
Toluene
I
I
PG
PG
S
Ig)
Aromatic, relatively polar
Electron-pair donor
Acetophenone
PGd)
S
S
Et2O
Ie)
I
I
Ie)
I
I
Ie)
I
I
(i-Pr)2O
t-BuOMeb)
1,4-Dioxane
S
S
S
Aprotic, dipolar
CH2Cl2
PG
PG
S
S
S
TG
TG
S
S
S
S
S
S
Acetone
ClCH2CH2Cl
Butan-2-one
MeCN
PGf)
CG
CG
CG
TG
AcOEt
Aprotic, highly dipolar
DMF
DMSO
S
S
S
S
S
S
Aprotic, highly dipolar, and highly polarizable
H-Bonding
Sulfolane
S
S
S
2,2,2-Trifluoroethanol
MeOH
EtOH
PrOH
BuOH
Pentanol
Decanolb)
i-PrOH
t-BuOH
H2O
S
S
S
S
S
S
S
S
S
S
S
I
S
S
S
S
S
S
S
S
S
I
TG
TG
TG
TG
TG
TG
TG
I
H-Bonding, strongly associated
Miscellaneous
CHCl3
1,2-Dimethoxyethanec)
THFc)
PG
Se)
Se)
S
Se)
Se)
S
Se)
Se)
a) [gelator] ¼ 1% (w/v), I: insoluble, S: soluble, PG: partial gel, CG: clear gel, TG: turbid gel. b) Missing
f
in Chastretteꢂs original classification [29]. c) Reclassified solvent. d) After 5 d. e) At 258. ) After 36 h.
g) Soluble at 708.
The results in Table 4 show the crucial role of the C(6)ꢀNH2 group of the adenosine
moiety for the formation of gels, and evidence that H-bonding between the
complementary nucleobases is the (or a) main driving force leading to a network of

Helvetica Chimica Acta – Vol. 91 (2008)
2189
the gelator, rather than just the interaction of the CH2OH groups. These results and the
tendency of the dinucleosides to form cyclic duplexes, as discussed in the preceding
paper [6], suggest that gelation is the result of a subtle balance between the formation
of cyclic and linear associates, considering that 2 shows a weaker tendency to form gels
than 1, but a stronger one to form cyclic duplexes [6]. Thus, 1 forms linear associates
more readily than 2, in agreement with the higher population of the anti-conformation
of unit I in 1. anti-Conformers indeed lead exclusively to linear associates, while a syn-
conformation is required for the formation of cyclic duplexes [2][3][6].
4.2. Gel – Sol Transition Temperature and Minimum Gelation Concentration. The
gel – sol transition (melting) temperature (Tm) of the gels formed by 1 and 2 depends
strongly on the gelled solvent (Table 5 and Exper. Part). The gels of 1 and
ClCH2CH2Cl, butan-2-one, and MeCN exhibit Tm values of 628, 588, and 498
respectively, while a broad transition between ca. 62 and 758 was observed for the
gel with AcOEt. Tm Values of the gels of 2, and MeCN or AcOEt are 548 and 568,
respectively. Tm Values of the gels of 1 in linear alcohols increase with the length of the
alkyl chain of the solvent, ranging from 508 in EtOH to 848 in decanol, whereas Tm of
the branched i-PrOH and t-BuOH are 548 and 488, respectively.
Table 5. Gel – Sol Transition Temperature (Tm) of 1% (w/v) Gels of 1 and 2
Solvent
1
2
Aprotic dipolar solvents
ClCH2CH2Cl
Butan-2-one
MeCN
AcOEt
EtOH
628
588
498
62 – 758
508
–
–
548
568
–
H-Bonding solvents
PrOH
528
–
BuOH
538
–
Pentanol
Decanol
i-PrOH
658
848
548
–
–
–
t-BuOH
488
–
The minimum gelation concentration (MGC), i.e., the minimum amount of 1 and 2
required to gel a given volume of solvent, varies significantly with the solvent (Table 6
and Exper. Part). The lowest MGC for the gels of 1 in an aprotic dipolar solvent was
observed for ClCH2CH2Cl (0.4%), while gelation of butan-2-one, MeCN, and AcOEt
requires 0.5, 0.6, and 0.8% of 1, respectively. The MGC of 2 is 0.6% in MeCN and 0.9%
in AcOEt, comparable to the MGC of 1. Gelation of i-PrOH and t-BuOH required 0.4
and 0.6% of 1, respectively. A regular decrease of the MGC of 1 from 0.8 to 0.3% was
observed for linear alcohols from EtOH to decanol. The gelation efficiency of 1 and 2
in terms of MGC is similar to that of other nucleobase-containing gelators that gel
alkanes at a concentration of ca. 0.3% [31][33], but lower than the one of trehalose
esters which gel AcOEt [35], or of benzylidene derivatives of pyranosides that gel
cyclohexane [36] at concentrations as low as 0.04%.
The variation of Tm and MGC for the gels of 1 in linear alcohols shows a
strengthening of the self-association of the gelator with increasing length of the alkyl

2190
Helvetica Chimica Acta – Vol. 91 (2008)
Table 6. Minimum Gelation Concentration (MGC) of the Gels of 1 and 2
1 (w/v)
Solvent
2 (w/v)
Aprotic dipolar solvents
ClCH2CH2Cl
Butan-2-one
MeCN
AcOEt
EtOH
PrOH
BuOH
Pentanol
Decanol
i-PrOH
0.4%
0.5%
0.6%
0.8%
0.8%
0.7%
0.6%
0.5%
0.3%
0.4%
0.6%
–
–
0.6%
0.9%
H-Bonding solvents
–
–
–
–
–
–
–
t-BuOH
chain of the solvent, the gel of 1 and decanol exhibiting the lowest MGC and the
highest Tm. Conversely, the gel of 1 and EtOH exhibits the highest MGC and the lowest
Tm. In a series of homologous linear alcohols, the elongation of the alkyl chain affects
mainly the polarity of the bulk solvent. Zhu and Dordick reported a correlation
between the gelation number (related to MGC) of trehalose diester gels and the
polarity parameter ET(30) [37], the most comprehensive empirical solvent polarity
scale known so far [38][39]. Similarly, the MGC for the gels of 1 and linear alcohols
increases with the increasing polarity parameter ENT of the solvent, while Tm decreases
(Fig. 3). This is in agreement with the observation that H-bonding is the main force
responsible for the association of the gelator, as highly polar solvents weaken H-bonds
to a greater extent than less polar ones. The trend followed by gels in i-PrOH and t-
BuOH (not shown) is different, evidencing that the constitution of the alcohols affects
the association of the gelator, independently of the polarity of the solvent.
Fig. 3. Effect of the polarity parameter ETN of the solvents on a) MGC and b) Tm of the gels of 1. ETN Values
are taken from [39].
4.3. Rheological Characterization. The rheological properties of some of the gels of
1 and 2 were evaluated by amplitude and temperature sweeps [40].

Helvetica Chimica Acta – Vol. 91 (2008)
2191
Amplitude Sweeps. The variation of the storage modulus (G’), related to the elastic
behaviour of the sample, and of the loss modulus (G’’), related to the viscous behaviour,
were determined for gels formed by 1% (w/v) of 1 in ClCH2CH2Cl, MeCN, AcOEt,
EtOH, pentanol, and decanol as a function of increasing and decreasing strain at a
constant temperature and frequency (Fig. 4; see Exper. Part).
For all gels of 1 studied, the elastic behaviour dominated over the viscous one at low
amplitudes (G’ > G’’), revealing the rigidity of the gel, i.e., a solid-like behaviour.
Moreover, each one of the functions G’(g) and G’’(g) [41] showed a constant plateau
value below the limit of the linear viscoelastic (LVE) range [41], showing that the
structure of the gel is stable under low-deformation conditions. At amplitudes higher
than the limit of the LVE range (gL), the structure of the gel started to change
irreversibly [41]; for 1, the gL values ranged between 0.3% in MeCN, and 0.6% in EtOH
and decanol. As the amplitude increased further, the viscous behaviour started to
dominate over the elastic one (G’’ > G’), and the sample exhibited a liquid character.
Depending on the solvent, the gel – sol transitions occurred at strain values ranging
from 3 to 9%. The reversibility of the gelation was evidenced by a sweep with
decreasing amplitudes. As the strain exerted on a liquid mixture of gelator 1 and solvent
(G’’ > G’) decreased, the elastic behaviour increased more rapidly than the viscous one,
and the gel character of the sample (G’ > G’’) was fully recovered at strain values
ranging from 0.5 to 3%.
Similar oscillatory tests of gels formed by 1% (w/v) of 2 in MeCN and AcOEt
(Fig. 5) also evidenced the domination of the elastic over the viscous behaviour (G’ >
G’’). In contradistinction to the gels formed from 1, neither the function G’(g) nor
G’’(g) showed a constant value upon increasing the strain from 0.1 to 100%, evidencing
a low gL value (<0.1%). The gel of 2 and MeCN exhibited a gel – sol transition at a
strain of ca. 2%, while the sol – gel transition of the gel formed with AcOEt required a
significantly higher strain (ca. 30%). Reversible gelation was observed upon decreasing
the amplitude, reconstitution of the gel occurring at a strain of ca. 0.3% for 2 in MeCN
and ca. 10% for 2 in AcOEt.
The significantly lower gL values observed for gels derived from 2 as compared
to those derived from 1 indicate that the former are less resistant to strain than
the latter i.e., that the strength of the inter-gelator association is weaker for 2 than for
1, in keeping with the interpretation of the gel-forming tendency of these dinucleo-
sides.
Temperature Sweeps. We only investigated the temperature effect on the dynamic
moduli G’ and G’’ for the gels of 1 (Fig. 6; see Exper. Part), considering the low gL
values for the gels of 2. The domination of the storage modulus of the gels of 1 and
ClCH2CH2Cl, MeCN, EtOH, pentanol, and decanol over the loss modulus (G’ > G’’, gel
character) was observed for a rather wide range of temperatures, up to the melting
temperature (Tm) where the viscous behaviour became predominant (G’’ > G’). The
agreement between the gel – sol transition temperatures determined by this technique
and by visual inspection was satisfactory (Table 5; 408 vs. 498 for MeCN, 538 vs. 508 for
EtOH, 658 vs. 658 for pentanol, 878 vs. 848 for decanol), with the exception for the gel
with ClCH2CH2Cl (788 vs. 628). The difference observed for the gel with ClCH2CH2Cl
illustrates the difficulty of visual discrimination between a gel and a viscous solution. In
addition, no crossing of the curves related to the temperature dependence of G’ and G’’

2192
Helvetica Chimica Acta – Vol. 91 (2008)
Fig. 4. Amplitude sweep – variation of the storage (G’) and loss (G’’) moduli with increasing and
decreasing strains for the 1% (w/v) gels of 1 in a) ClCH2CH2Cl, b) MeCN, c) AcOEt, d) EtOH,
e) pentanol, and f) decanol at 258. The dotted lines represent the limit of the linear viscoelastic range
(gL).
could be observed for the gel of 1 and AcOEt, in parallel to a broad gel – sol transition
temperature.

Helvetica Chimica Acta – Vol. 91 (2008)
2193
Fig. 5. Amplitude sweep – variation of the storage (G’) and loss (G’’) moduli with increasing and
decreasing strains for the 1% (w/v) gels of 2 in a) MeCN and b) AcOEt at 258
4.4. Circular Dichroism. Temperature-dependent CD spectra of the gels formed by
1% (w/v) of 1 and ClCH2CH2Cl, MeCN, AcOEt, EtOH, PrOH, pentanol, and decanol,
and of the gels of 2 in MeCN and AcOEt were measured to characterize their
supramolecular structure [42]. Due to the relatively high concentration (ca. 16 mm)
required for gelation as compared to the range of concentrations usually employed for
CD studies, we modified the acquisition parameters (see Exper. Part). In contra-
distinction to the CD spectra of 1-mm solutions of A*[s]U(*) dinucleosides in CHCl3 [6]
which show extrema between ca. 260 and 280 nm, and no bands above 300 nm, the gels
of 1 and 2 (ca. 16 mm) show intense CD signals between ca. 275 and 400 nm, and a weak
signal to noise ratio below ca. 275 nm, due to the strong absorbance of the samples.
Shape and intensity of the bands depend on the gelator and the solvent (cf. Figs. 7 and 8).
At 08, the CD spectrum of the gel of 1 and MeCN shows a broad minimum around
315 nm, a maximum at ca. 300 nm, and a minimum at ca. 280 nm (Fig. 7). Increasing
the temperature to 308 lowered the intensity of the minimum at ca. 315 nm, while the
intensity of the extrema at ca. 300 and at ca. 280 nm increased. Further increasing the
temperature to 408 resulted in a decrease of the intensity of the minimum around
315 nm, a slight increase of the intensity of the maximum at ca. 300 nm, and a decrease
of the intensity of the minimum at ca. 280 nm. At 508, above Tm of the gel, the CD bands
could hardly be observed, evidencing a significantly lower degree of organisation.
The CD spectra of the gels of 1 and AcOEt, EtOH, PrOH (not shown), pentanol,
and decanol exhibit a similar shape with a broad minimum around 310 nm (315 nm for
decanol) and a maximum at ca. 290 nm (295 nm for decanol), but differ strongly from
the CD spectra of 1 in MeCN. The temperature-dependent CD spectra of these gels
exhibit isodichroic points at ca. 300 nm, and showed a regular decrease of intensity of
the bands with increasing temperature. This suggests an equilibrium between two
species, conceivably a weakly and a highly organised aggregate corresponding to weak
and strong CD signals, respectively.
The CD spectra of the gel of 1 and ClCH2CH2Cl were not reproducible, as the shape
of the CD curve for a given sample depended on the orientation of the cell relative to
the axis of the light beam, evidencing birefringency of the gel [42].

2194
Helvetica Chimica Acta – Vol. 91 (2008)
Fig. 6. Temperature sweep – variation of the storage (G’) and loss (G’’) moduli with increasing
temperature for the 1% (w/v) gels of 1 in a) ClCH2CH2Cl, b) MeCN, c) EtOH, d) pentanol, e) decanol,
and f) AcOEt at constant strain (0.3%) and frequency (1 Hz)
The CD spectrum of the gel of 2 and MeCN at 08 shows a positive Cotton effect,
with a broad maximum at ca. 300 nm and a minimum at ca. 285 nm. The intensity of the
bands decreased steadily with increasing temperature (Fig. 8). The CD spectrum of the

Helvetica Chimica Acta – Vol. 91 (2008)
2195
Fig. 7. Temperature-dependent CD spectra of the 1% (w/v) gels of 1 (16 mm) in a) MeCN, b) AcOEt,
c) EtOH, d) pentanol, and e) decanol
gel of 2 and AcOEt at 08 shows a negative Cotton effect with a strong sharp minimum
at ca. 295 nm, and a similar maximum at ca. 280 nm. Although no significant changes of
intensity could be observed upon heating to 408, increasing the temperature stepwise to
608, above the Tm of the gel (548), resulted in a decrease of the molar ellipticity. For

2196
Helvetica Chimica Acta – Vol. 91 (2008)
Fig. 8. Temperature-dependent CD spectra of the 1% (w/v) gels of 2 (15 mm) in a) MeCN and b) AcOEt
these two gels, we observed isodichroic points at ca. 290 nm, suggesting an equilibrium
between a highly and a weakly organised aggregate of 2, similarly as it was observed for
the gels of 1.
An intriguing feature of all CD spectra of the gels of 1 and 2 are strong bands with
long tails above 300 nm. CD Bands at such wavelengths can hardly be the consequence
of a red shift of the p – p* transition of the nucleobases. The maximal absorbance of the
gels, too intense to be determined precisely, was found around 260 nm, typical for the
p – p* transitions of uracil and adenine [43][44]. Cholesteric or chiral smectic C phases
that may be formed upon gelation could give rise to CD signals outside the absorption
region of the chromophore [42]. However, the formation of such liquid-crystalline
phases was excluded upon inspecting the samples under polarized light in a microscope.
An extremum at 280 nm in the CD spectra of polyriboadenylic acids was reported
by Bush and Scheraga [45] who suggested that it is due to a n – p* transition. That this
band is absent in the CD spectra of mono- or dinucleotides is ascribed to solvation of
the adenine ring and a blue shift of the n – p* band. Another example of a long-
wavelength CD band around 295 nm was reported for stacked 6-N-methylated
adenylyluridine [46] that exhibited a large CD effect as compared to the UV
absorption, corroborating the n – p* character of the transition. The CD bands
observed above 275 nm for the gels of 1 and 2 are thus likely to arise from a n – p*
transition, and their intensity is expected to increase with the degree of association (and
stacking) of the gelator.
Broad tail-shaped bands at longer wavelength were also observed for carotenoid
aggregates [47] and ascribed to non-packed head-to-tail assemblies. Conceivably, the
broad tail-shaped bands characterizing the CD spectra of the gels of 1 and 2 could also
originate from non-packed aggregates that would slowly form a more tightly organized
aggregate. The intensity of the broad tail-shaped bands is expected to decrease with
aging of the gel. To test these hypotheses, we studied the influence of the gelator
concentration and of aging the gel on the CD spectra (cf. Fig. 9 and 10).
The CD spectrum of a 0.3% solution of 1 in MeCN shows only a weak negative
band at ca. 290 nm (Fig. 9). Increasing the concentration of 1 to 0.6% resulted in
gelation of the solvent, and a broad negative CD band around 320 nm, a positive one

Helvetica Chimica Acta – Vol. 91 (2008)
2197
around 295 nm, and a negative one at ca. 280 nm. Further increasing the concentration
of 1 to 1% has no effect on the negative tail-shaped band at ca. 320 nm, while the
positive band was slightly red-shifted to ca. 300 nm, and the intensity of the negative
band around 280 nm increased. This evidences the relation between the degree of
association of the gelator and the apparition of CD signals above 300 nm. Conceivably,
at a low concentration of 1, the gelator– gelator interactions are weak, and the
nucleobases solvated. This would result in a blue shift of the n – p* transition band (vide
supra). As the concentration of 1 increases, gelation of the solvent occurs, and a
network of the gelator is formed. This requires desolvating the nucleobases, and results
in a red shift of the n – p* transition, occurring around 300 nm.
Fig. 9. CD Spectra of the 0.3% (w/v; solution), 0.6% (w/v; gel), and 1% (w/v; gel) mixtures of 1 and
MeCN at 208, recorded 1 h after gelation
The CD spectrum of the 1% (w/v) gel of 1 and MeCN recorded 10 min after gel
formation shows a broad negative, a positive, and a sharp negative band at ca. 320, 300,
and 280 nm, respectively (Fig. 10). As the gel aged, the intensity of the broad tail-
shaped negative band around 320 nm decreased. The intensity of the positive band at
ca. 300 nm decreased slightly, while the band at ca. 280 nm was not affected. After 27 h,
the CD spectrum remained unchanged. This suggests a slow rearrangement of the
network of the gelator, as revealed by the decrease of the intensity of the broad tail-
shaped band above 300 nm that may originate from tightly packed aggregates of the
gelator.
We thank the Swiss National Science Foundation and Syngenta AG, Basel, for generous support.
Experimental Part
General. THF was distilled from Na/benzophenone, CH2Cl2, MeOH, DMF, pyridine, EtN(i-Pr)2,
and (i-Pr)2NH from CaH2. Reactions were run under N2. Qual. TLC: precoated silica-gel plates (Merck
silica gel 60 F 254); detection by spraying ꢁanisaldehydeꢂ (EtOH (425 ml), AcOH (5 ml), H2SO4 (16 ml),
p-anisaldehyde (10 ml)). Flash chromatography (FC): silica gel Fluka 60 (0.04 – 0.063 mm). Optical
rotations: 1-dm cell at 258 and 589 nm. Temp.-dependent CD (108 steps from 08 to 508): 1-mm solns. in
CHCl3 in a 1-mm Suprasil cell. UV Spectra: 10ꢀ5 m solns. in CHCl3 at 208 in a 1-cm Suprasil cell. FT-IR:

2198
Helvetica Chimica Acta – Vol. 91 (2008)
Fig. 10. Time-dependent CD spectra of the 1% (w/v; 16 mm) gel of 1 and MeCN at 208
solid state (ATR). 1H- and 13C-NMR: at 300 or 400 MHz, and 75 or 100 MHz, resp. MS: matrix-assisted
laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF) with 0.05m indol-3-acrylic
acid (IAA) in THF, or with 0.05m a-cyano-4-hydroxycinnamic acid (CCA) in MeCN/EtOH/H2O, and
high-resolution (HR) MALDI-TOF with 0.05m 2,5-dihydrobenzoic acid (DHB) in THF.
NMR Studies. NMR Experiments were performed at 295 K and at 300 MHz in CDCl3 (passed
through basic aluminium oxide and dried over mol. sieves (4 ꢄ) prior to use). Experiments started at the
highest concentration, with stepwise replacement of 225 ml of the 700 ml soln. with 225 ml of CDCl3. The
data were analysed by non-linear least-squares fitting using MATLAB (trust-region algorithm); the
parameters were Kass, d(HꢀN(3), c ¼ 0 mm), and d(HꢀN(3), c ¼1 ).
5’-O-[Dimethyl(1,1,2-trimethylpropyl)silyl]-2’,3’-O-isopropylidene-8-{[(4-methoxyphenyl)(diphe-
nyl)methoxy]methyl}adenosine (5). A soln. of 4 [6] (4.117 g, 4.88 mmol) in dry MeOH (12 ml) was
cooled to 08, treated with a freshly prepared 1.22m soln. of MeONa in MeOH (12.0 ml, 14.65 mmol) and
stirred at 258 for 15 h. The mixture was cooled to 08, and treated with sat. aq. NH4Cl soln. (20 ml). After
evaporation of MeOH, the residue was extracted three times with AcOEt. The combined org. layers were
dried (Na2SO4) and evaporated. FC (cyclohexane/AcOEt 4 :1 ! 2 :1 ! 1:1) gave 5 (3.459 g, 94%).
White solid. Rf (cyclohexane/AcOEt 1:2) 0.24. M.p. 84 – 888. [a]2D5 ¼ ꢀ201.9 (c ¼ 0.075, CHCl3). UV
(CHCl3): 263 (40470). IR (ATR): 3464w, 3316w, 3157w, 2955w, 2867w, 1735w, 1638m, 1604m, 1578w,
1509m, 1490w, 1463w, 1446m, 1415w, 1373m, 1328m, 1298m, 1250s, 1213m, 1170m, 1154m, 1061s, 1034s,
1
976m, 932w, 900w, 869m, 830s. H-NMR (300 MHz, CDCl3): see Table 1; additionally, 7.55 – 7.22 (m, 12
arom. H); 6.88 – 6.83 (m, 2 arom. H); 6.04 (br. s, NH2); 3.78 (s, MeO); 1.57 (sept., J ¼ 6.9, Me2CH); 1.50,
1.38 (2s, Me2C); 0.85, 0.82 (2s, Me2CSi); 0.79 (d, J ¼ 6.9, Me2CH); 0.00, ꢀ 0.02 (2s, Me2Si). 13C-NMR
(75 MHz, CDCl3): see Table 7; additionally, 158.94 (s); 143.67, 143.62 (2s); 134.70 (s); 130.61 (2d); 128.56
(4d); 128.13 (4d); 127.32 (2d); 113.97 (s, Me2C); 113.44 (2d); 87.93 (s, Ph2C); 55.31 (q, MeO); 34.19 (d,
Me2CH); 27.37, 25.73 (2q, Me2C); 25.32 (s, Me2CSi); 20.39 (q, Me2CSi); 18.58, 18.55 (2q, Me2CH);
ꢀ 3.30, ꢀ 3.34 (2q, Me2Si). HR-MALDI-MS: 752.3841 ([M þ H]þ, C42H54N5O6Siþ; calc. 752.3838).
Reductive Deamination of 5. A mixture of 5 (2.241 g, 2.98 mmol) and 4-ꢄ mol. sieves (3 g) in dry
THF (60 ml) was heated to 508, treated hourly for 4 h with 3-methylbutyl nitrite (4 ꢁ 4 ml, 120 mmol),
and evaporated. FC (cyclohexane/AcOEt 2 :1 ! 0 :1) gave 6 (1.297 g, 59%) and 7 (274 mg, 12%). A
sample of 7 for analysis was recrystallized in MeCN/CHCl3 (ca. 100 :1).
Data of 5’-O-[Dimethyl(1,1,2-trimethylpropyl)silyl]-2’,3’-O-isopropylidene-8-{[(4-methoxyphenyl)-
(diphenyl)methoxy]methyl}-b-d-ribofuranosyl-9H-purine (6). White solid. Rf (cyclohexane/AcOEt 1:2)
0.53. M.p. 55 – 588. [a]2D5 ¼ ꢀ23.2 (c ¼ 0.07, CHCl3). UV (CHCl3): 267 (12990). IR (ATR): 3057w, 2956w,
2867w, 1727w, 1600m, 1584w, 1509m, 1491w, 1456w, 1448w, 1427w, 1372w, 1355w, 1300w, 1250s, 1213m,
1
1179m, 1154m, 1066s, 1034s, 974m, 916w, 901w, 870w, 830s. H-NMR (300 MHz, CDCl3): see Table 1;
additionally, 7.57 – 7.24 (m, 12 arom. H); 6.91 – 6.85 (m, 2 arom. H); 3.81 (s, MeO); 1.58 (sept., J ¼ 6.9,

Helvetica Chimica Acta – Vol. 91 (2008)
2199
Table 7. Selected 13C-NMR Chemical Shifts [ppm] of the Adenosine-Derived Monomers 5 – 9 in CDCl3
5
6
7
8
9
C(2)
C(4)
C(5)
C(6)
152.77
149.33
119.07
155.48
150.68
59.48
90.38
82.35
83.00
87.31
152.38
152.08
133.69
148.44
154.42
59.67
90.29
82.71
81.97
144.27
158.96a)
124.46
159.25a)
149.33b)
59.19
90.30
83.22
81.86
86.93
152.63
152.24
132.90
148.03
156.81
58.27
89.80
83.63
80.87
86.83
153.10
151.98a)
133.31
149.12
152.53a)
36.76
90.16
83.00
81.46
87.45
C(8)
CH2ꢀC(8)
C(1’)
C(2’)
C(3’)
C(4’)
C(5’)
87.02
63.05
63.22
63.17
62.69
62.73
a) Assignment may be interchanged. b) May be interchanged with 149.67 ppm (see Exper. Part).
Me2CH); 1.52, 1.40 (2s, Me2C); 0.85 (d, J ¼ 6.9, Me2CH); 0.81, 0.80 (2s, Me2CSi); 0.00 (s, Me2Si).
13C-NMR (75 MHz, CDCl3): see Table 7; additionally, 158.99 (s); 143.47, 143.42 (2s); 134.44 (s); 130.62
(2d); 128.49 (4d); 128.16 (4d); 127.37 (2d); 114.29 (s, Me2C); 113.48 (2d); 88.19 (s, Ph2C); 55.32 (q,
MeO); 34.16 (d, Me2CH); 27.37, 25.68 (2q, Me2C); 25.32 (s, Me2CSi); 20.36 (q, Me2CSi); 18.56, 18.54 (2q,
Me2CH); ꢀ 3.36, ꢀ 3.38 (2q, Me2Si). HR-MALDI-MS: 737.3736 ([M þ H]þ, C42H53N4O6Siþ; calc.
737.3729). Anal. calc. for C42H52N4O6Si (736.98): C 68.45, H 7.11, N 7.60; found: C 68.23, H 7.12, N 7.47.
Data of 5’-O-[Dimethyl(1,1,2-trimethylpropyl)silyl]-2’,3’-O-isopropylidene-8-{[(4-methoxyphenyl)-
(diphenyl)methoxy]methyl}inosine (7). White solid. Rf (cyclohexane/AcOEt 1:1) 0.29. M.p. 2068. [a]D25
¼
ꢀ18.5 (c ¼ 0.93, CHCl3). UV (CHCl3): 246 (20230), 255 (sh, 18040), 276 (sh, 8990). IR (ATR): 3098w,
3055w, 2956w, 2867w, 1707s, 1607w, 1589w, 1555w, 1528w, 1509m, 1491w, 1460w, 1446m, 1375w, 1345w,
1298w, 1251m, 1212m, 1176w, 1155w, 1086s (sh), 1061s, 1034s, 983w, 970w, 931w, 900w, 862m, 828s.
1H-NMR (300 MHz, CDCl3): see Table 1; additionally, 13.05 (br. s, NH); 7.53 – 7.21 (m, 12 arom. H);
6.88 – 6.83 (m, 2 arom. H); 3.80 (s, MeO); 1.57 (sept., J ¼ 6.9, Me2CH); 1.48, 1.36 (2s, Me2C); 0.83 (d, J ¼
6.9, Me2CH); 0.80, 0.79 (2s, Me2CSi); 0.00 (s, Me2Si). 13C-NMR (75 MHz, CDCl3): see Table 7;
additionally, 149.67 (s); 143.68, 143.58 (2s); 134.73 (s); 130.66 (2d); 128.63 (4d); 128.14 (4d); 127.30 (2d);
114.46 (s, Me2C); 113.50 (2d); 87.87 (s, Ph2C); 55.38 (q, MeO); 34.25 (d, Me2CH); 27.41, 25.72 (2q, Me2C);
25.41 (s, Me2CSi); 20.45 (q, Me2CSi); 18.62 (q, Me2CH); ꢀ 3.23, ꢀ 3.26 (2q, Me2Si). HR-MALDI-MS:
775.3501 ([M þ Na]þ, C42H52N4NaO7Siþ; calc. 775.3503). Anal. calc. for (C42H52N4O7Si) · 0.5 H2O
(761.98): C 66.20, H 7.01, N 7.35; found: C 66.11, H 6.89, N 7.37.
5’-O-[Dimethyl(1,1,2-trimethylpropyl)silyl]-8-(hydroxymethyl)-2’,3’-O-isopropylidene-b-d-ribofura-
nosyl-9H-purine (8). A soln. of 6 (1.275 g, 1.73 mmol) in dry CH2Cl2 (17 ml) was treated with Et3SiH
(2.21 ml, 13.84 mmol) and ClCH2COOH (1.71 ml, 20.76 mmol), and stirred for 1 h at 258. The mixture
was diluted with CH2Cl2 (10 ml) and H2O (10 ml), and neutralised with 1m aq. NaOH. The aq. layer was
extracted three times with AcOEt. The combined org. layers were dried (Na2SO4) and evaporated. FC
(cyclohexane/AcOEt 1:1 ! 1:2 ! 1:4) gave 8 (550 mg, 68%). A sample for analysis was recrystallized
from H2O/MeCN. White solid. Rf (cyclohexane/AcOEt 1:2) 0.25. M.p. 95.1 – 95.48. [a]2D5 ¼ ꢀ1.9 (c ¼
1.68, CHCl3). UV (CHCl3): 266 (14570). IR (ATR): 3197w (br.), 2985w, 2935w, 2876w, 1741w, 1606m,
1581w, 1520w, 1465w, 1434w, 1381w, 1373m, 1355m, 1327w, 1303m, 1252m, 1226w, 1206w, 1156w, 1125w,
1
1110m, 1075s, 1061s, 1054s, 1027m, 998w, 975w, 962w, 942w, 930m, 913w, 875m, 853m, 833s. H-NMR
(300 MHz, CDCl3): see Table 1; additionally, 4.70 (br. s, HOCH2ꢀC(8)); 1.61, 1.38 (2s, Me2C); 1.55 (sept.,
J ¼ 6.9, Me2CH); 0.82 (d, J ¼ 6.9, Me2CH); 0.78 (s, Me2CSi); 0.01 (s, Me2Si). 13C-NMR (75 MHz,
CDCl3): see Table 7; additionally, 114.88 (s, Me2C); 34.11 (d, Me2CH); 27.32, 25.49 (2q, Me2C); 25.49 (s,
Me2CSi); 20.38, 20.35 (2q, Me2CSi); 18.51 (q, Me2CH); ꢀ 3.41 (q, Me2Si). HR-MALDI-MS: 465.2527
([M þ H]þ, C22H37N4O5Siþ; calc. 465.2533). Anal. calc. for C22H36N4O5Si (464.63): C 56.87, H 7.81, N
12.06; found: C 57.15, H 7.76, N 12.04.

2200
Helvetica Chimica Acta – Vol. 91 (2008)
8-(Chloromethyl)-5’-O-[dimethyl(1,1,2-trimethylpropyl)silyl]-2’,3’-O-isopropylidene-b-d-ribofurano-
syl-9H-purine (9). A soln. of 8 (541 mg, 1.16 mmol) in dry CH2Cl2 (8 ml) was cooled to 08, treated with
EtN(i-Pr)2 (404 ml, 2.32 mmol) and MsCl (99 ml, 1.28 mmol), stirred at 258 for 18 h, treated with LiCl
(49 mg, 1.16 mmol), and stirred for 8 h. The mixture was diluted with CH2Cl2 (8 ml), and washed three
times with H2O. The combined org. layers were dried (Na2SO4) and evaporated. FC (cyclohexane/
AcOEt 9 :1 ! 4 :1) gave 9 (537 mg, 96%). White solid. Rf (cyclohexane/AcOEt 1:1) 0.67. M.p. 81.0 –
82.08. [a]2D5 ¼ ꢀ23.1 (c ¼ 1.4, CHCl3). UV (CHCl3): 268 (10670). IR (ATR): 2993w, 2957m, 2875w,
1594m, 1581m, 1509w, 1466w, 1455w, 1425w, 1397m, 1373m, 1362m, 1307w, 1295w, 1251m, 1232w, 1212m,
1150m, 1122m, 1103m, 1073s, 1052s, 981m, 962w, 943m, 925m, 871m, 831s. 1H-NMR (300 MHz, CDCl3):
see Table 1; additionally, 1.62, 1.40 (2s, Me2C); 1.54 (sept., J ¼ 6.9, Me2CH); 0.81 (d, J ¼ 6.9, Me2CH);
0.76, 0.75 (2s, Me2CSi); ꢀ 0.05 (s, Me2Si). 13C-NMR (75 MHz, CDCl3): see Table 7; additionally, 114.56
(s, Me2C); 34.16 (d, Me2CH); 27.37, 25.55 (2q, Me2C); 25.39 (s, Me2CSi); 20.37, 20.34 (2q, Me2CSi); 18.54,
18.52 (2q, Me2CH); ꢀ 3.42 (q, Me2Si). HR-MALDI-MS: 483.2194 ([M þ H]þ, C22H36ClN4O4Siþ; calc.
483.2189).
5’-O-[Dimethyl(1,1,2-trimethylpropyl)silyl]-2’,3’-O-isopropylidene-b-d-ribofuranosyl-9H-purine-8-
methyl-(81 ! 5’-S)-2’,3’-O-isopropylidene-5’-thiouridine (11). A soln. of 9 (532 mg, 1.1 mmol) and 10
(377 mg, 1.1 mmol) in dry degassed MeOH (11 ml) was cooled to 08, treated with a freshly prepared
0.66m soln. of MeONa in MeOH (5.0 ml, 3.3 mmol), and stirred at 258 for 3 h. The mixture was cooled to
08, and treated with sat. aq. NH4Cl soln. (20 ml). After evaporation of MeOH, the residue was extracted
four times with AcOEt. The combined org. layers were dried (Na2SO4) and evaporated. FC (CH2Cl2/
MeOH 95 :5) gave 11 (802 mg, 98%). White solid. Rf (CH2Cl2/MeOH 95 :5) 0.32. M.p. 95 – 978. [a]D25
¼
ꢀ58.2 (c ¼ 0.78, CHCl3). UV (CHCl3): 266 (18630). IR (ATR): 3192w (br.), 3054w, 2957w, 2937w,
2868w, 1691s, 1631w, 1597w, 1510w, 1455m, 1423w, 1375m, 1360m, 1303w, 1252m, 1211m, 1156m, 1076s,
971m, 923w, 860m, 828s. 1H-NMR (300 MHz, CDCl3): see Table 2; additionally, 7.30 (d, J ¼ 8.1, HꢀC(6/
I)); 5.71 (dd, J ¼ 8.1, 1.8, HꢀC(5/I)); 1.61, 1.49, 1.40, 1.27 (4s, 2 Me2C); 1.52 (sept., J ¼ 6.9, Me2CH); 0.79
(d, J ¼ 6.9, Me2CH); 0.74, 0.73 (2s, Me2CSi); ꢀ 0.08 (s, Me2Si). 13C-NMR (75 MHz, CDCl3): see Table 8;
additionally, 114.79, 114.21 (2s, 2 Me2C); 34.16 (d, Me2CH); 27.33, 27.16, 25.53, 25.33 (4q, 2 Me2C); 25.33
(s, Me2CSi); 20.36, 20.35 (2q, Me2CSi); 18.53 (q, Me2CH); ꢀ 3.43 (q, Me2Si). HR-MALDI-MS: 747.3213
([M þ H]þ, C34H51N6O9SSiþ; calc. 747.3207).
Table 8. Selected 13C-NMR Chemical Shifts [ppm] of the HA*[s]U Dimers 11 and 3 in CDCl3
11
3
11
3
Uridine unit (I)
C(2/I)
C(4/I)
C(5/I)
C(6/I)
C(1’/I)
C(2’/I)
C(3’/I)
C(4’/I)
Adenosine-derived unit (II)
154.53a)
163.46
102.87
142.62
94.74
84.46
83.34
87.02
33.63
153.80a)
163.36
102.90
142.53
94.63
84.40
83.17
86.65
33.85
C(2/II)
C(4/II)
C(5/II)
C(6/II)
152.35
150.09a)
133.35
148.10
152.16a)
28.49
151.94
150.03a)
134.00
148.90
151.75a)
28.64
92.22
82.74
81.62
85.99
C(8/II)
CH2ꢀC(8/II)
C(1’/II)
C(2’/II)
C(3’/II)
C(4’/II)
C(5’/II)
90.09
82.83
C(5’/I)
81.89
87.75
62.85
63.30
a) Assignment may be interchanged.
2’,3’-O-Isopropylidene-b-d-ribofuranosyl-9H-purine-8-methyl-(81 ! 5’-S)-2’,3’-O-isopropylidene-
5’-thiouridine (3). In a polyethylene flask, a soln. of 11 (696 mg, 0.93 mmol) in dry THF (4.7 ml) at 258
was treated with a soln. of (HF)3 · Et3N (1.53 ml, 28.0 mmol) and stirred at 258 for 64 h. The mixture was

Helvetica Chimica Acta – Vol. 91 (2008)
2201
cooled to 08, treated with aq. 1m NaOH soln. until the pH reached ca. 10, and extracted four times with
AcOEt. The combined org. layers were dried (Na2SO4) and evaporated. FC (CH2Cl2/MeOH 97:3) gave 3
(533 mg, 95%). White solid. Rf (CH2Cl2/MeOH 97:3) 0.40. M.p. 126.0 – 129.58. [a]2D5 ¼ ꢀ47.2 (c ¼ 0.73,
CHCl3). UV (CHCl3): 265 (16760). IR (ATR): 3496w (sh.), 3219w (br.), 3058w, 2987w, 2937w, 2876w,
2823w, 1688s, 1631w, 1598m, 1510w, 1456m, 1425w, 1374m, 1362m, 1303w, 1266m, 1212m, 1155m, 1077s,
987m, 946w, 854m, 810m. 1H-NMR (300 MHz, CDCl3): see Table 2; additionally, 7.26 (d, J ¼ 8.1, HꢀC(6/
I)); 5.68 (dd, J ¼ 8.1, 1.7, HꢀC(5/I)); 5.61 (dd, J ¼ 11.0, 2.0, HOꢀC(5’/II)); 1.64, 1.47, 1.37, 1.24 (4s,
2 Me2C). 13C-NMR (75 MHz, CDCl3): see Table 8; additionally, 114.83, 114.45 (2s, 2 Me2C); 27.83, 27.12,
25.47, 25.26 (4q, 2 Me2C). HR-MALDI-MS: 627.1853 ([M þ Na]þ, C26H32N6NaO9Sþ; calc. 627.1849).
Evaluation of the Solubility and Gelation Properties. In a 5-ml vial closed with a stopper, a mixture of
5 mg of the dimers 1, 2, or 3, and 500 ml of the solvent was stirred at 758 (958 when decanol was used). A
compound was considered insoluble, when no complete dissolution occurred upon heating for ca. 10 min.
A compound was considered as soluble, when a soln. was obtained upon heating, and remained a soln.
after 36 h at 258. A mixture of a compound and a solvent was considered a gel when, after dissolution
upon heating, followed by cooling to 258, no flowing could be observed upon inversion of the vial (ꢁstable
to inversionꢂ criterion) [48]. A partial gel refers to a jelly-like soln. that shows a significant flow upon
inversion of the vial.
Determination of the Gel – Sol Transition Temperature (Tm). A vial containing a gel was placed in a
thermostated bath, and the temp. was slowly increased by steps of 18. Tm was considered to be reached
when the material flowed upon inversion of the vial.
Determination of the Minimum Gelation Concentration (MGC). MGC was determined by addition
of a defined amount of solvent to a vial containing a gel, so as to reduce the gelator concentration in steps
of 0.1% (w/v). The mixture was then heated above Tm, until a clear soln. was obtained, and the sample
was cooled to 258. The process was repeated until the material flowed upon inversion of the vial.
Rheological Characterisation of the Gels. All measurements were performed on a Anton Paar
Physica MCR 300 rheometer equipped with a cone-plate system CP 30-4 (49.95-mm diameter, 4.08
angle). The storage and loss moduli were not corrected, due to the inaccurate measurement of the
effective radius of the sample. The amplitude sweeps were performed at increasing and decreasing strains
(1 min pause between the experiments), at constant temp. (258) and frequency (1 Hz). The limiting value
of the linear viscoelastic range, gL, was determined manually with a tolerated deviation of 5%. The temp.
sweeps were performed in a chamber containing an atmosphere saturated with solvent vapours, at
increasing temp. (rate 5 – 108/min), and constant strain (0.3%) and frequency (1 Hz).
CD Spectra of the Gels. The gel was heated above its Tm, rapidly transferred as a soln. to a 1-mm
path-length cell, and kept at 258 until gelation occurred (10 min to 3 h). The CD measurements were
performed on a Jasco J-710 spectropolarimeter, at increasing temp., using the following settings: 10-nm
bandwidth, 0.5-s response, low sensivity, 0.1-nm data pitch, and 50 – 200-nm/min scanning speed.
REFERENCES
´
[1] S. Eppacher, N. Solladie, B. Bernet, A. Vasella, Helv. Chim. Acta 2000, 83, 1311.
[2] X. Zhang, B. Bernet, A. Vasella, Helv. Chim. Acta 2006, 89, 2861.
[3] X. Zhang, B. Bernet, A. Vasella, Helv. Chim. Acta 2007, 90, 864.
[4] X. Zhang, B. Bernet, A. Vasella, Helv. Chim. Acta 2007, 90, 891.
[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] A. Viger, Ph.D. Thesis No. 16060, ETH-Zꢀrich, 2005.
[8] K. Araki, I. Yoshikawa, Top. Curr. Chem. 2005, 256, 133.
[9] N. M. Sangeetha, U. Maitra, Chem. Soc. Rev. 2005, 34, 821.
[10] J. H. van Esch, B. L. Feringa, in ꢁEncyclopedia of Supramolecular Chemistryꢂ, Eds. J. L. Atwood,
J. W. Steed, Marcel Dekker, New York, Basel, 2004, p. 586 – 596.
[11] P. J. Flory, Faraday Discuss. Chem. Soc. 1974, 57, 7.
[12] O. Gronwald, E. Snip, S. Shinkai, Curr. Opin. Colloid Interface Sci. 2002, 7, 148.

2202
Helvetica Chimica Acta – Vol. 91 (2008)
[13] D. J. Abdallah, R. G. Weiss, Adv. Mater. 2000, 12, 1237.
[14] P. Terech, R. G. Weiss, Chem. Rev. 1997, 97, 3133.
[15] M. de Loos, B. L. Feringa, J. H. van Esch, Eur. J. Org. Chem. 2005, 3615.
[16] L. A. Estroff, A. D. Hamilton, Chem. Rev. 2004, 104, 1201.
[17] K. Y. Lee, D. J. Mooney, Chem. Rev. 2001, 101, 1869.
[18] T. Kato, Y. Hirai, S. Nakaso, M. Moriyama, Chem. Soc. Rev. 2007, 36, 1857.
[19] T. Kato, N. Mizoshita, M. Moriyama, T. Kitamura, Top. Curr. Chem. 2005, 256, 219.
[20] M. George, R. G. Weiss, Acc. Chem. Res. 2006, 39, 489.
[21] A. Ajayaghosh, V. K. Praveen, Acc. Chem. Res. 2007, 40, 644.
[22] V. Nair, S. D. Chamberlain, Synthesis 1984, 401.
[23] A. Ritter, B. Bernet, A. Vasella, Helv. Chim. Acta 2008, 91, 1675.
[24] H. Gunji, A. Vasella, Helv. Chim. Acta 2000, 83, 1331.
[25] H. S. Gutowsky, A. Saika, J. Chem. Phys. 1953, 21, 1688.
[26] A. Dunger, H. H. Limbach, K. Weisz, Chem. – Eur. J. 1998, 4, 621.
[27] H. Iwahashi, Y. Kyoguku, J. Am. Chem. Soc. 1977, 99, 7761.
[28] Y. Kyogoku, R. C. Lord, A. Rich, J. Am. Chem. Soc. 1967, 89, 496.
[29] M. Chastrette, M. Rajzmann, M. Chanon, K. F. Purcell, J. Am. Chem. Soc. 1985, 107, 1.
[30] T. Sato, M. Seko, R. Takasawa, I. Yoshikawa, K. Araki, J. Mater. Chem. 2001, 11, 3018.
[31] E. Snip, K. Koumoto, S. Shinkai, Tetrahedron 2002, 58, 8863.
[32] Y. J. Yun, S. M. Park, B. H. Kim, Chem. Commun. 2003, 254.
[33] I. Yoshikawa, S. Yanagi, Y. Yamaji, K. Araki, Tetrahedron 2007, 63, 7474.
[34] Y. Jeong, K. Hanabusa, H. Masunaga, I. Akiba, K. Miyoshi, S. Sakurai, K. Sakurai, Langmuir 2005,
21, 586.
[35] G. John, G. Zhu, J. Li, J. S. Dordick, Angew. Chem., Int. Ed. 2006, 45, 4772.
[36] R. Luboradzki, O. Gronwald, A. Ikeda, S. Shinkai, Chem. Lett. 2000, 1148.
[37] G. Y. Zhu, J. S. Dordick, Chem. Mater. 2006, 18, 5988.
[38] C. Reichardt, in ꢁSolvents and Solvent Effects in Organic Chemistryꢂ, Wiley-VCH, Weinheim, 2003,
p. 389 – 469.
[39] C. Reichardt, Chem. Rev. 1994, 94, 2319.
[40] R. Brummer, in ꢁRheology Essentials of Cosmetic and Food Emulsionsꢂ, Springer-Verlag, Berlin,
2006, p. 63 – 80.
[41] T. Mezger, in ꢁThe rheology handbook: for users of rotational and oscillatory rheometersꢂ, Ed. U.
Zorll, Vincentz Verlag, Hannover, 2002, p. 122 – 126.
[42] G. Gottarelli, G. Spada, E. Castiglioni, in ꢁMolecular Gels: Materials with Self-Assembled Fibrillar
Networksꢂ, Eds. R. G. Weiss, P. Terech, Springer-Verlag, Dordrecht, 2006, p. 431 – 446.
[43] P. R. Callis, Annu. Rev. Phys. Chem. 1983, 34, 329.
[44] P. S. Song, Ann. Rep. Prog. Chem., Sect. B: Org. Chem. 1977, 74, 18.
[45] C. A. Bush, H. A. Scheraga, Biopolymers 1969, 7, 395.
[46] C. S. M. Olsthoorn, C. A. G. Haasnoot, C. Altona, Eur. J. Biochem. 1980, 106, 85.
´
[47] M. Simonyi, Z. Bikadi, F. Zsila, J. Deli, Chirality 2003, 15, 680.
[48] S. R. Raghavan, B. H. Cipriano, in ꢁMolecular Gels: Materials with Self-Assembled Fibrillar
Networksꢂ, Eds. R. G. Weiss, P. Terech, Springer-Verlag, Dordrecht, 2006, p. 241 – 252.
Received July 10, 2008