Incorporation of a S-glycosidic linkage into a glyconucleoside changes the conformational preference of both furanose sugars

Article abstract of DOI:10.1016/j.carres.2006.11.007

A glyconucleoside containing a thioglycoside linkage, namely 1-(3-S-β-d-ribofuranosyl-2,3-dideoxy-3-thio-β-d-ribofuranosyl)-thymine, has been prepared through condensation of a suitably protected derivative of 3′-thiothymidine with an activated ribose sug

Full text of DOI:10.1016/j.carres.2006.11.007

Carbohydrate Research 342 (2007) 16–22
Incorporation of a S-glycosidic linkage into a
glyconucleoside changes the conformational preference
of both furanose sugars
Joanne Buckingham,a John A. Brazier,a Julie Fisherb,* and Richard Cossticka,*
aDepartment of Chemistry, University of Liverpool, Liverpool L69 7ZD, UK
bSchool of Chemistry, University of Leeds, Leeds LS2 9JT, UK
Received 8 August 2006; received in revised form 1 November 2006; accepted 7 November 2006
Available online 12 November 2006
Abstract—A glyconucleoside containing a thioglycoside linkage, namely 1-(3-S-b-D-ribofuranosyl-2,3-dideoxy-3-thio-b-D-ribo-
furanosyl)-thymine, has been prepared through condensation of a suitably protected derivative of 30-thiothymidine with an
activated ribose sugar. NMR has been used to study the conformation of the S-disaccharide and the unmodified O-disaccharide.
A full pseudorotational analysis showed that for the S-disaccharide, the ribose and deoxy ribose sugars have a preference for the
south and north pucker, respectively; which is the reverse of what is seen for the O-disaccharide.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Keywords: Glyconucleosides; Thioglycoside; Conformational analysis; NMR; Modified nucleoside
1. Introduction
a-D-ribofuranosyl-(100-20)-adenosine (3), which has an
a-glycosidic bond, and is a structural component in
polyADP-ribose.4 Glycosidic linkages to the 30-position
of the nucleoside have also been identified in several nat-
ural products including adenophostins5 A and B (4 and
5) and HF-76 (6). The adenophostins and their ana-
logues are potent agonists of myo-inisitol 1,4,5-triphos-
phate,7 whilst HF-7 is a neuroactive compound found
in the venom of the funnel-web spider.
Amongst the simplest of the glyconucleosides are the
30-O-b-D-ribofuranosyl-20-deoxynucleosides, which are
conveniently prepared by condensation of the partially
protected nucleoside with an activated ribofuranoside.
This procedure has been applied to the synthesis of
30-O-b-D-ribofuranosyl thymidine (7) and the 20-deoxy-
cytidine derivative.8 We have become interested in pre-
paring glyconucleosides in which the glycosidic oxygen
atom is replaced by sulfur. These S-glycosides are of
interest for two reasons: firstly, sulfur substitution can
enhance the stability of the glycosidic linkage to enzy-
matic hydrolysis9 and secondly, it can significantly alter
the conformation of ribofuranose sugars. The confor-
mational effect of replacing the 30-O atom by sulfur is
Glyconucleosides contain an additional carbohydrate
residue that is linked to one of the nucleoside hydroxyl
groups via an O-glycoside bond. These types of com-
pounds show considerable structural diversity depend-
ing on the nature of the sugar, its position of
attachment to the nucleoside and the configuration of
the O-glycosidic linkage.1 Of particular note are the 20-
O-ribofuranosyl nucleosides, which have been shown
to occur with either an a- or b-glycosidic bond. 20-O-
b-D-Ribofuranosyl-(100-20)-adenosine-500-O-phosphate (1)
and its analogous guanosine derivative (2) have a b-gly-
coside linkage, and were identified as modified nucleo-
sides at position 64 in initiator tRNAs of lower
eukaryotes.2 The 500-phosphororibofuranosyl moiety is
thought to be an important factor in determining
whether the tRNA participates in initiation or elonga-
tion during protein synthesis.3 Closely related is 20-O-
*
Corresponding authors. Tel.: +44 113 343 6577; fax: +44 113 343
6565 (J.F.); tel.: +44 151 794 3514; fax: +44 151 794 3588 (R.C.);
e-mail addresses: j.fisher@chem.leeds.ac.uk; rcosstic@liv.ac.uk
0008-6215/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2006.11.007

J. Buckingham et al. / Carbohydrate Research 342 (2007) 16–22
17
well established for 20-deoxynucleotides, where it has
been shown that the sugar to which the sulfur is
attached shifts to the C-30-endo/C-20-exo or north
conformation.
We now report the synthesis of a glyconucleoside con-
taining a S-glycoside linkage (8) and show through
NMR studies that in comparison to the O-glycoside
(7), the conformational preferences of both the ribose
and the deoxyribose sugars are significantly altered by
the sulfur substitution (Fig. 1).
tion of 50-O-tert-butyldiphenylsilyl (TBDPS) thymidine
(10, Scheme 1) with 1-O-acetyl-2,3,5-tri-O-benzoyl-b-D-
ribofuranose.
For the synthesis of the corresponding S-glycoside (8),
we adopted an analogous procedure based on 50-O-
TBDPS-30-deoxy-30-thiothymidine (14), as outlined in
Scheme 1. Thus, the anhydronucleoside (12) was pre-
pared in 85% yield from 50-O-TBDPS-thymidine (10)
by a Mitsunobu reaction. Opening of the anhydronucleo-
side was achieved by treatment with excess sodium thio-
benzoate in DMF at 110 ꢁC, to give the product
thioester (13) in 72%, following chromatography.
Hydrolysis of the 50-O-TBDPS protected thioester (13)
was performed using sodium hydroxide in aqueous eth-
anol, as previously described10 for the corresponding
dimethoxytrityl-protected nucleoside and gave 50-O-
TBDPS-30-deoxy-30-thiothymidine (14) in 78% yield.
Whilst thionucleosides are susceptible to oxidation to
the symmetrical disulfides, 14 was clearly established
as the thiol from the 1H NMR spectrum, which revealed
a single compound with a characteristic doublet
attributable to the mercapto proton (1.53 ppm). Conden-
sation of the thionucleoside (14) with 1-O-acetyl-2,3,5-
tri-O-benzoyl-b-D-ribofuranose using tin tetrachloride
in dichloromethane gave 15 in 67% yield, comparable
to the 74% yield obtained for the corresponding
O-disaccharide (11). Deprotection was accomplished in
two steps, initial treatment with tetrabutylammonium
fluoride removed the TBDPS group (60% yield, follow-
ing chromatography) and subsequent reaction with
methanolic ammonia gave the fully deprotected thiogly-
coside (8) as an analytically pure amorphous solid after
extraction with dichloromethane. The beta configuration
2. Results and discussion
2.1. Synthesis of 1-(3-S-b-D-ribofuranosyl-2,3-dideoxy-3-
thio-b-D-ribofuranosyl)-thymine (8)
In what served as a model study for the synthesis of the
S-glycoside (8), O-b-D-ribofuranosyl-(100-30)-thymidine
(7) was prepared using the method previously described
by Mikhailov.8 In the key step, the fully protected O-
glycoside (11) was prepared in 74% yield by condensa-
HO
B
HO
O
Ade
O
2-O3PO
HO
O
HO
O
O
O
OHHO
2-
OH
HO
OPO3
1, B = Adenin-9-yl
2, B = Guanin-9-yl
3
HO
O
Ade
-O3SO
Gua
RO
HO
O
Thy
Thy
O
O
2-
OPO3
O
HO
RO
-
BzO
HO
OSO3
O
Thy
X
O
2-O3PO
2-O3PO
X
O
O
(ii), (iii)
(i)
O
O
OH
H3C
OH
10
OBz
BzO
OH
OH
HO
OR
AcO
11, X = O
15, X = S
4, Adenophostin A, R = H
5, Adenophostin B, R = OAc
7, X = O
8, X = S
6
(i)
(iv)
O
N
2-O3PO
HO
5'
Thy
1'
CH3
Thy
O
O
N
RO
RO
Thy
4'
3'
Thy
O
O
(vi)
(v)
2'
RO
O
HO
HO
O
5'R
X
O
O
O
4'R
SH
14
SBz
1'R
3'R
2'R
13
12
R = tert-butyldiphenylylsilyl
OH
HO
OH
HO
7, X = O
8, X = S
9
Scheme 1. Reagents and conditions: (i) 1-O-acetyl-2,3,5-tri-O-benzoyl-
b-D-ribofuranose, SnCl4, CH2Cl2, 0 ꢁC; (ii) TBAF, THF; (iii) NH3
saturated MeOH; (iv) DIAD, PPh3, ethyl acetate; (v) sodium
thiobenzoate, DMF, 110 ꢁC; (vi) NaOH (0.4 M) H2O, EtOH.
Figure 1. Glyconucleoside structures. Subscript R indicates a ribose
sugar.

18
J. Buckingham et al. / Carbohydrate Research 342 (2007) 16–22
of the ribose glycosidic bond of the S-disaccharide (8)
was established through NOE measurements (see
below and Supplementary data).
of these simulations are shown in Table 2. From the
error analysis, all singular values are large and compara-
ble in magnitude to within a factor of 100, indicating
well determined data. From the variance–covariance
matrix, the estimated standard deviations of all para-
meters were at most of the order of 10ꢀ3. This procedure
was repeated for spectra recorded at 21, 26 and 31 ꢁC.
The conformation of five-membered rings can be de-
scribed by the concept of pseudorotation.15 Using this
approach the phase angle, P, and puckering amplitude,
U, are sufficient to uniquely define the ring shape.15,16
Furanose rings of nucleotides generally exist in an
equilibrium between two extreme conformers (Fig. 2)
notated north (N) and south (S); these labels encompass
the conformations C-20-exo/C-30-endo and C-20-endo/
C-30-exo, respectively.16 Consequently, five quantities
are actually required for a complete description: the
pseudorotational values for the north and south con-
formers (PN,UN,PS,US) and the mole fraction of one
of the conformers (XN or XS). These may be calculated
2.2. Conformational studies
The O-disaccharide (7) has previously been prepared8
and its 1H NMR spectrum recorded at 200 MHz, in
D2O solution, has been analyzed. Use of a 500 MHz
instrument in this current study has now resulted in
the previously reported overlapped-multiplets being
resolved and a full assignment of the protons is now
1
described (Table 1). The 500 MHz H NMR spectrum
of S-disaccharide (8) in D2O solution was assigned using
DQF-COSY and TOCSY spectra. The relative assign-
ments of the 20/200 protons for 7 were made as described
previously.11 For 8, the 20/200 assignments were con-
firmed by analysis of 1D NOE data; H-6-20 (2%), and
comparison with literature data12 for a related com-
pound. Nuclear Overhauser data (Supplementary data)
recorded for 8 also permitted confirmation of the pres-
ence of the b-anomer for each sugar ring. In the case
of the ribose sugar of 8, the characteristic NOEs13 estab-
lished that H-40 and H-10 are on the same face as each
other and no NOE connection was observed between
H-10 and H-30. Care is needed when interpreting the
absence of an NOE; however, as the stereochemistry
at C-40 is established at the outset, these data taken to-
gether are consistent with the glycosidic bond being beta.
To determine accurate coupling constants necessary
1
from vicinal H–1H coupling constants, involving sugar
ring protons, using a generalized form of the Karplus
equation17 embedded in the computer program PSEU-
ROT.18 In using the latest version of this program, the
methodology previously adopted was followed here.12
A summary of the results of the pseudorotational anal-
ysis of 7 and 8 is provided in Table 3.
An initial interpretation of the H-10 to H-20/H-200
coupling constant data11 for the deoxyribose ring sug-
gested that in the O-disaccharide (7), the sugar pucker
would be south and that in S-disaccharide (8) would
be north (Table 2). These preliminary conclusions were
confirmed following the pseudorotation analysis. Thus,
in the deoxysugar of the O-disaccharide (7), the south (or
1
for conformational analysis (see below), H NMR spec-
tra recorded for 7 and 8 were computer simulated fol-
lowing our previous methodology.12 The chemical
shifts and coupling constants from each experimental
spectrum were entered into a spectral simulation pro-
gram, gNMR14 and new spectra calculated. To extract
the coupling constants, peaks in the calculated spectrum
were assigned to observed transitions in the correspond-
ing experimental spectrum, and then an automated iter-
ative approach to improving agreement between the
calculated and the observed spectrum was adopted.
Each sugar system was taken separately and the results
Table 2. Vicinal coupling constants (Hz) for sugar ring protons in each
of the disaccharides
Coupling constants at 21 ꢁC
10-20
S-Disaccharide
10-200
20-30
200-30
30-40
Deoxy
Ribo
3.48
5.76
7.62
—
8.51
5.08
9.13
4.71
8.86
3.63
O-Disaccharide
Deoxy
Ribo
6.99
1.23
6.58
—
6.96
4.55
3.42
—
3.75
6.98
Table 1. 1H NMR chemical shifts (ppm) for 7 and 8 in D2O solution at
21 ꢁC from gNMR simulations
Proton
Chemical shifts of sugar protons
S-Disaccharide (8) O-Disaccharide (7)
5'
3'
5'
Z
2'
X
X
Deoxy
Ribo
Deoxy
Ribo
O
O
4'
1'
100
2
6.164
2.635
2.577
3.587
4.009
3.954
3.842
5.125
4.055
—
4.169
4.006
3.746
3.671
6.232
2.359
2.532
4.426
4.106
3.799
3.740
5.091
4.067
—
4.195
4.006
3.806
3.640
1'
4'
3'
2'
200
30
400
5
Z
North
South
Figure 2. Representations of the two extreme sugar conformations:
500
X = N, S or O, Z = O or S.

J. Buckingham et al. / Carbohydrate Research 342 (2007) 16–22
19
Table 3. Pseudorotational parameters for the north and south
conformers of the sugar rings in each of the disaccharides
more equal population of the conformers with the south
pucker being slightly favoured (57–58% S).
S-Disaccharide
O-Disaccharide
Deoxy Ribo
In relation to our studies on the O-disaccharide (7),
comparison can be drawn to previous work on its phos-
phorylated derivative (9). Compound 9 was found to
have ribose and deoxyribose sugars with pseudorotation
parameters22 (mole fraction south 74% and 11% and PS
160ꢁ, and 162ꢁ, for the deoxyribose and ribose rings,
respectively) very similar to those calculated here for 7
(see Table 3). Slight differences in the pseudorotational
parameters between 9 and 7 are likely to arise from
the presence of the bulky 50-phosphate group in 9 and
the slight temperature difference between the two sets
of measurements; 19 ꢁC for 9,22 whilst data for 7 were
collected over three different temperatures, which is gen-
erally accepted to give more reliable solutions to the
pseudorotation analysis.18 Alternatively, as coupling
constants used for the analysis of 9 were taken directly
from spectra with the assumption of first order behav-
iour,22 this may be another cause of the difference.
Deoxy
Ribo
Puckers
UN
US
PN
31.2ꢁ
34.5ꢁ
42.8ꢁ
40.1ꢁ
38.4ꢁ
39.5ꢁ
35.0ꢁ
28.1ꢁ
9.5ꢁ
36.8ꢁ
35.0ꢁ
ꢀ16.3ꢁ
162.6ꢁ
PS
RMS error/Hz
162.6ꢁ
0.540
149.2ꢁ
0.053
143.8ꢁ
0.160
0.120
% South
21 ꢁC
26 ꢁC
3
2
3
58
57
57
70
72
70
5
3
4
31 ꢁC
C-20-endo/C-30-exo) conformer is dominant (70% mole
fraction), which is typical of deoxyribonucleosides.16
The preference for this conformation arises from a com-
bination of electronic and steric factors covered by the
anomeric and gauche effects;19 the anomeric ‘contribu-
tion’ or generalized anomeric effect is now considered
as a special case of a general preference for gauche con-
formations.20 The anomeric effect of the nucleobase (or
any other substituent with electron lone pairs attached
to the C-10 (anomeric) carbon) is thought to originate
from the tendency of one of the lone pairs on O-40 to
adopt an antiperiplanar orientation relative to the glyco-
sidic bond. The gauche effect is the stabilization of the
gauche versus the trans orientation in the X–C–C–Y
fragment, where X and Y are electronegative groups,
O or S in this instance. Thus, both factors are deter-
mined to some extent by the electronegativity of substit-
uents attached to the sugar ring.20,21 If the anomeric
effect were to dominate then the sugar puckers would
be north. In a normal deoxyribose ring, that is, with a
30-O, the gauche effect dominates and the sugar pucker
is south. When the 30-O is substituted for a sulfur atom
(with much lower electronegativity), the gauche effect is
dramatically reduced and the anomeric effect wins. This
is clearly seen for the deoxyribose ring in the S-disaccha-
ride (8) where the N conformer is seen almost exclusively
(97–98% mole fraction). This result is consistent with
our previous studies on 30-S-modified dinucleotides,
which revealed the same conformational preference for
this sugar.12
2.3. Conclusion
In summary, the first example of a glyconucleoside
containing a S-glycosidic linkage has been prepared. The
S-glycosidic linkage was introduced by condensation
of a 30-thiothymidine derivative with 1-O-acetyl-2,3,5-
tri-O-benzoyl-b-D-ribofuranose and the yield (67%) for
this condensation reaction was slightly lower than that
used to generate the O-glycoside (74% yield). The beta
configuration of the S-glycosidic linkage was established
by NOE experiments. Conformational studies, including
a full pseudorotational analysis, showed that for the S-
disaccharide the ribose and deoxy ribose sugars have a
preference for the south and north pucker, respectively.
In the south, arrangement, the C-20 atom is above the
plane formed by the bonds between C-10, O-40, C-40
and C-30 below this plane. In the north arrangement,
C-30 is above the plane described by C-40, O-40, C-10
and C-20 below. These observations for the S-disaccha-
ride (8) are the reverse of what is seen for the O-
disaccharide (7). The switch in the conformational
preferences of the sugars, which is induced by the sulfur
substitution, results through changing the interplay of
the different anomeric and gauche factors. The altered
conformational features of the S-disaccharide may
prove useful in exploring the relationship between sugar
pucker and biological activity in glyconucleosides.
In ribose rings, additional gauche interactions arise
and the balance is such that the north pucker is generally
preferred and this is seen for the ribose ring of the O-
disaccharide (7) where this conformer dominates to the
extent of about 95% mole fraction. However, this bal-
ance is disturbed when the electronegativity of ribose
ring substituents is changed.21 Thus, when the O-10
atom is substituted by a sulfur atom, as in the case of
the ribose sugar of S-disaccharide (8), the anomeric
effect is reduced, as is the gauche interaction between
O-20 and the C-10 sulfur substituent and the result is a
3. Experimental
3.1. General methods
1-O-Acetyl-2,3,5-tri-O-benzoyl-b-D-ribofuranose
was
purchased from Chemgenes, Wilmington, MA, USA.

20
J. Buckingham et al. / Carbohydrate Research 342 (2007) 16–22
Anhydrous N,N-dimethylformamide (Sure-sealTM) was
purchased from Aldrich. Dichloromethane was distilled
from calcium hydride. Analytical thin layer chromato-
graphy was performed on Merck TLC aluminium sheets
coated with silica gel 60 F254. The developed chromato-
graphs were visualized with a UV lamp (254 nm).
Carbohydrates were developed by treatment with
p-anisaldehyde stain [p-anisaldehyde (6 cm3) was mixed
with sulfuric acid (8.0 cm3), acetic acid (2.4 cm3) and
ethanol (218 cm3)], turning blue upon charring with a
heat gun. Flash chromatography was performed on
Merck silica (particle size 40–63 lM, supplied by
BDH), prepared as a homogeneous slurry in the column
eluent and applied to the column over a base layer of
sand. The material to be purified was then introduced
on to the column in a minimal volume of the eluent
and ran with the appropriate eluents. All NMR spectra
were recorded on a Bruker 400 MHz, a Bruker 500 MHz
or a Varian Unity Inova 500 MHz NMR spectrometer.
For 1H and 13C NMR spectra ran in deuterated dimethyl-
sulfoxide ((CD)3SO), all chemical shifts are reported
in parts per million relative to (CD)3SO (1H ppm:
2.50, 13C ppm: 39.52 0.06). For the disaccharide
derivatives, resonances referring to the ribose sugars
are denoted with a subscript ‘R’. All mass spectra
were recorded on a Micromass LCT mass spectro-
meter using electrospray ionization (EI) and direct
infusion syringe pump sampling. All samples were
diluted with MeOH. For the preparation of sodium
thiobenzoate, thiobenzoic acid (10.11 g, 73.2 mmol)
was added dropwise to a stirred solution of 3 M sodium
hydroxide (24.13 cm3, 72.4 mmol) over a 30 min period.
The precipitate formed was removed by filtration and
the filtrate concentrated in vacuo to yield sodium
thiobenzoate as a pale yellow solid (10.99 g, 68.4 mmol,
94%).
5.87 (1H, d, H1 , J 3.66 Hz), 7.35–7.65 (11H, m,
10 · CH(Ph, TBDPS), H6). 13C NMR (100 MHz;
0
DMSO-d6): d 13.4 (5-CH3), 19.1 (SiC(Me)3), 26.8
t
0
0
0
(CH3 Bu(TBDPS)), 39.8 ðC2 Þ, 63.2 ðC5 Þ, 77.3 ðC3 Þ,
0
85.1 ðC4 Þ, 116.4 (C5), 128.3 (C6), 130.3, 132.6, 132.7,
137.1 (CH(TBDPS), SiC(Ph)(TBDPS)), 153.7 (C2),
171.1 (C4). m/z (ES+): 463.2 [M+H]+, 485.2
[M+Na]+, 501.2 [M+K]+. Acc mass: [M+H]+
463.2040, C26H31N2O4Si requires 463.2053; [M+Na]+
485.1859,
C26H30N2O4SiNa
requires
485.1873;
[M+K]+ 501.1616, C26H30N2O4SiK requires 501.1612.
3.3. 30-Deoxy-30-thio-30-S-benzoyl-20-deoxy 50-O-(tert-
butyldiphenylsilyl)-thymidine (13)
Sodium thiobenzoate (0.443 g, 2.77 mmol) was added to
a stirred solution of compound 12 (0.512 g, 1.11 mmol)
dissolved in anhydrous dimethylformamide (6 cm3)
under N2 at 110 ꢁC. Additional portions of sodium
thiobenzoate (0.443 g, 2.77 mmol) were then added
every 30 min for the next 21/2 h. Following the addition
of the final portion of sodium thiobenzoate, the reaction
mixture was left to stir for a further 30 min after which
the solvent was removed under reduced pressure. The
deep red residue that remained was taken up in ethyl
acetate (100 cm3) and washed with satd aq NaHCO3
(2 · 100 cm3). The organic layer was then dried over
anhydrous sodium sulfate, filtered and concentrated in
vacuo to afford the impure product as a pale brown
foam. Purification by flash silica chromatography, elut-
ing with CH2Cl2 followed by a gradient of 0–1% MeOH
in CH2Cl2, yielded the title compound as a pale brown
amorphous solid (0.478 g, 0.795 mmol, 72%). 1H
NMR (400 MHz; DMSO-d6):
d
1.02 (9H, s,
tBu(TBDPS)), 1.59 (3H, s, 5-CH3), 2.46–2.72 (2H, m,
0
00
0
H2 , H2 ), 3.91 (1H, dd, H5 , J 3.97, 7.79 Hz), 3.98
00
0
(1H, dd, H5 , J 2.54, 6.99 Hz), 4.12–4.16 (1H, m, H4 ),
3.2. 50-O-(tert-Butyldiphenylsilyl)-2,30-anhydrothymidine
(12)
4.40 (1H, q, H3 , J 7.92 Hz), 6.23 (1H, dd, H1 , J 5.40,
6.27 Hz), 7.34–7.93 (16H, m, 10 · CH(Ph, TBDPS),
5 · CH(Bz), H6), 11.36 (1H, br s, N3H). 13C NMR
0
0
Diisopropylazodicarboxylate (1.23 cm3, 6.24 mmol)
dissolved in ethyl acetate (2 cm3) was added dropwise
to a solution of triphenyl phosphine (1.64 g, 6.24 mmol)
and 50-O-(tert-butyldiphenylsilyl)thymidine (2.00 g,
4.16 mmol) dissolved in ethyl acetate (11 cm3) and the
reaction mixture left to stir overnight at rt. The solvents
were subsequently removed under reduced pressure to
afford a crude yellow foam. Purification by flash silica
chromatography, eluting with 5:1 ethyl acetate–n-hex-
ane followed by 5:1:2 ethyl acetate–n-hexane–MeOH,
afforded the title compound as a white amorphous solid
(100 MHz; DMSO-d6):
d
11.8 (5-CH3 thy), 18.8
t
0
(SiC(Me)3), 26.6 (CH3 Bu(TBDPS)), 37.4 ðC2 Þ, 39.8
0
0
0
ðC3 Þ 63.3 ðC5 Þ, 83.7 ðC1 Þ, 109.7 (C5), 127.3, 128.2,
129.6, 130.3, 134.6, 135.3, 135.4, 136.0 ((CH(Ph,
TBDPS)), SiC(Ph)(TBDPS), CH(Bz), SCOC(Bz)),
136.3 (C6) 150.3 (C2), 163.6 (C4), 190.0 (SCOBz). m/z
(ES+): 623.2 [M+Na]+. Acc mass: [M+Na]+ 623.1995,
C33H36N2O5SSiNa requires 623.2012.
3.4. 30-Deoxy-30-thio-50-O-(tert-butyldiphenylsilyl)-
thymidine (14)
1
(1.64 g, 3.55 mmol, 85%). H NMR (400 MHz; DMSO-
t
d6): d 0.97 (9H, s, CH3 Bu(TBDPS)), 1.76 (3H, d, 5-CH3,
N2 was bubbled through a stirred solution of 0.4 M
sodium hydroxide in ethanol (52 cm3) for 30 min.
30-Deoxy-30-thio-30-S-benzoyl-50-O-(tert-butyldiphenylsi-
lyl)-thymidine (0.478 g, 0.80 mmol) was then added and
0
00
J 1.11 Hz), 2.46–2.61 (2H, m, H2 , H2 ), 3.61 (1H, dd,
0
00
H5 , J 7.48, 10.8 Hz), 3.84 (1H, dd, H5 , J 5.56,
0
0
10.64 Hz), 4.43–4.46 (1H, m, H4 ), 5.33 (1H, m, H3 ),

J. Buckingham et al. / Carbohydrate Research 342 (2007) 16–22
21
0
0
0
0
0
the reaction mixture stirred under N2 at rt for a further
2 h. A 1 M solution of potassium dihydrogen phosphate
(93 cm3, 93.70 mmol) was then added dropwise over a
15 min period. The cream precipitate that formed was
filtered, washed with distilled water (3 · 50 cm3) and
dried over phosphorus pentoxide to yield the title com-
ðC2 Þ, 40.8 ðC3 Þ, 63.2 ðC5 Þ, 64.3 ðC5 RÞ, 72.4 ðC3 RÞ,
0
0
0
0
75.4 ðC1 RÞ, 80.3 ðC4 RÞ, 83.6 ðC1 RÞ, 84.8 ðC4 Þ, 85.3
0
ðC2 RÞ, 109.9 (C5), 128.1, 128.2, 128.8, 129.0, 129.1,
129.1, 129.6, 129.7, 129.7, 130.2, 130.2, 133.1, 133.9,
135.3, 135.4 (CH(TBDPS), SiC(Ph)(TBDPS)/CH(Bz),
OCOC(Bz)), 135.9 (C6), 150.7 (C2), 164.0 (C4), 164.9.
165.1, 165.8 (COBz). m/z (ES+): 963.3[M+Na]+. Acc
mass: [M+Na]+ 963.2999, C52H52N2O11SSiNa requires
963.2959.
1
pound as a cream solid (0.308 g, 0.619 mmol, 78%). H
NMR (400 MHz; DMSO-d6):
d
1.01 (9H, s,
tBu(TBDPS)), 1.53 (1H, d, SH, J 5.70 Hz), 1.56
0
(3H, d, 5-CH3, J 0.80 Hz), 2.29–2.37 (1H, m, H2 ),
00
0
2.50–2.57 (1H, m, H2 ), 3.52 (1H, q, H3 , J 8.59 Hz),
3.6. 1-[2,3-Dideoxy-3-thio-3-S-(2,3,5-tri-O-benzoyl-b-D-
ribofuranosyl)-b-D-ribofuranosyl]thymine
0
00
0
0
3.80–3.98 (3H, m, H5 , H5 , H4 ), 6.14 (1H, dd, H1 ,
0
J 3.66, 5.73 Hz), 7.39–7.68 (11H, m, H6 , 10 · CH(Ph,
TBDPS)), 11.29 (1H, br s, N3H). 13C NMR
A 0.5 M solution of tetrabutyl ammonium fluoride tri-
hydrate in tetrahydrofuran (0.71 cm3, 0.7085 mmol)
was added to a solution of 15 (0.480 g, 0.510 mmol)
and the reaction mixture stirred for 2 h at rt. Co-evapo-
ration of the solvent with chloroform afforded the
impure product as a cream foam. Purification by flash
silica chromatography, eluting with CH2Cl2 followed
by a gradient of 0–2% MeOH in CH2Cl2, yielded the
title compound as a white amorphous solid (0.214 g,
(100 MHz; DMSO-d6): d 12.2 (5-CH3), 19.2 (SiC(Me)3),
t
0
0
27.0 (CH3 Bu(TBDPS)), 34.7 ðC3 Þ, 41.4 ðC2 Þ, 62.8
0
0
0
ðC5 Þ, 83.4 ðC1 Þ, 87.9 ðC4 Þ, 109.9 (C5), 128.2,
130.3, 135.4, 136.0 (CH(TBDPS), SiC(Ph)(TBDPS)),
136.2 (C6) 150.8 (C4), 164.0 (C2). m/z (ES+): 519.2
[M+Na]+. Acc mass: [M+Na]+ 519.1749, C26H32N2O4-
SSiNa requires 519.1750.
1
3.5. 1-[5-O-tert-Butyldiphenylsilyl-2,3-dideoxy-3-thio-3-
S-(2,3,5-tri-O-benzoyl-b-D-ribofuranosyl)-b-D-ribo-
furanosyl]-thymine (15)
0.304 mmol, 60%). H NMR (400 MHz; DMSO-d6): d
0
1.76 (3H, d, 5-CH3, J 0.95 Hz) 2.39–2.48 (2H, m, H2 ,
00
0
0
00
H2 ), 3.59–3.75 (3H, m, H3 , H5 , H5 ), 3.84 (1H, dt,
0
0
H4 , J 3.02, 8.10 Hz), 4.56 (1H, dd, H5 R, J 4.6,
0
A 1.0 M solution of tin tetrachloride in CH2Cl2
(1.07 cm3, 1.073 mmol) was added to a cold solution
(0 ꢁC) of 1-O-acetyl-2,3,5-tri-O-benzoyl-b-D-ribofura-
nose (0.489 g, 0.936 mmol) in dry CH2Cl2 (4 cm3) under
N2 and the resulting reaction mixture stirred at 0 ꢁC for
10 min. A solution of 30-deoxy-50-O-(tert-butyldiphenyl-
silyl)-30-thiothymidine (0.395 g, 0.795 mmol) in dry
CH2Cl2 (2 cm3) was then added and the reaction mix-
ture stirred under the same conditions for a further
1 h. A 10% aqueous solution of sodium bicarbonate
(10 cm3) was then added and the reaction mixture left
to stir for a further 20 min. The white suspension that
was observed to form was removed by filtration through
Celite and the filtrate was washed with water (30 cm3)
and reduced in vacuo to afford a crude cream foam.
Purification by flash silica chromatography, eluting
with 2:1 n-hexane–ethyl acetate followed by 1:1 n-
hexane–ethyl acetate, afforded the title compound as a
12.0 Hz), 4.68 (1H, dd, H5 R, J 3.81, 12.0 Hz), 4.75
(1H, q, H4 R, J 4.29 Hz), 5.18 (1H, br t, 50-OH, J
4.93 Hz), 5.66–5.69 (1H, m, H1 R), 5.77–5.81 (2H, m,
0
0
0
0
0
H
2 R, H3 RÞ, 6.04 (1H, dd, H1 , J 4.29, 6.76 Hz), 7.44–
8.04 (16H, m, H6, 15 · CH(Bz)), 11.24 (1H, br s,
N3H). 13C NMR (100 MHz; DMSO-d6): d 12.6 (5-
0
0
0
0
CH3), 40.1 ðC2 Þ, 40.6 ðC3 Þ, 60.2 ðC5 Þ, 64.3 ðC5 RÞ,
0
0
0
0
72.5 ðC3 RÞ, 75.5 ðC1 RÞ, 80.1 ðC4 RÞ, 83.9 ðC1 Þ, 85.5
0
0
ðC2 RÞ, 85.8 ðC4 Þ, 109.3 (C5), 128.9, 129.0, 129.1,
129.2, 129.6, 129.7, 133.9, 134.2, 134.3 (CH(Bz),
OCOC(Bz)), 136.5 (C6), 150.7 (C2), 164.1 (C4), 164.9,
165.1, 165.8 (COBz). m/z (ES+): 725.2 [M+Na]+. Acc
mass: [M+Na]+ 725.1760, C36H34N2O11SNa requires
725.1781.
3.7. 1-(3-S-b-D-Ribofuranosyl-2,3-dideoxy-3-thio-b-D-
ribofuranosyl)-thymine (8)
1
white amorphous solid (0.501 g, 0.532 mmol, 67%). H
A solution of 1-[2,3-dideoxy-3-thio-3-S-(2,3,5-tri-O-
NMR (400 MHz; DMSO-d6):
d
0.95 (9H, s,
benzoyl-b-D-ribofuranosyl)-b-D-ribofuranosyl]-thymine
(0.206 g, 0.293 mmol) dissolved in 7 M methanolic
ammonia (3.52 cm3) was stirred at rt under N2 for
5 days. The solvents were then removed under reduced
pressure and the residue that remained was taken up
in distilled water (20 cm3) and washed with CH2Cl2
(2 · 20 cm3). The aqueous layer was subsequently
co-evaporated to dryness with acetone to yield the title
compound as a white amorphous solid (0.113 g,
0.289 mmol, 99%). 1H NMR (500 MHz; DMSO-d6):
tBu(TBDPS)), 1.49 (3H, d, 5-CH3, J 0.64 Hz), 2.42–
0
00
0
0
0
2.56 (2H, m, H2 , H2 ), 3.82–3.98 (4H, m, H3 , H4 , H5 ,
00
0
H5 ), 4.58 (1H, dd, H5 R, J 4.61, 12.16 Hz), 4.69 (1H,
00
0
dd, H5 R, J 3.66, 12.08 Hz), 4.78 (1H, q, H4 R, J
0
4.45 Hz), 5.70 (1H, t, H1 R, J 4.77 Hz), 5.76 (1H, d,
0 0
2 R, J 4.45 Hz), 5.83 (1H, t, H3 R, J 5.25 Hz), 6.13
H
0
(1H, dd, H1 J 4.77, 7.31 Hz), 7.32–8.05 (26H, m,
10 · CH(Ph, TBDPS), 15 · CH(Bz), H6), 11.31 (1H, br
s, N3H). 13C NMR (100 MHz; DMSO-d6): d 12.2 (5-
CH3), 19.2 (SiC(Me)3), 27.0 (CH3 tBu(TBDPS)), 39.0
0
00
d 1.76 (3H, s, 5-CH3), 2.32–2.45 (2H, m, H2 , H2 ),

22
J. Buckingham et al. / Carbohydrate Research 342 (2007) 16–22
0
0
00
3.38–3.55 (3H, m, H3 , H5 R, H5 R), 3.58–3.63 (1H, m,
References
0
00
0
0
0
H5 ), 3.71–3.80 (4H, m, H5 , H4 , H4 R, H2 R), 3.84–3.87
(1H, m, H3 R), 4.70 (1H, t, 50R-OH, J 5.77 Hz), 4.94
1. Efimtseva, E. V.; Mikhailov, S. N. Biochemistry (Mos-
cow) 2002, 67, 1374–1384.
2. Desgres, J.; Keith, G.; Kuo, K. C.; Gehrke, C. W. Nucl.
Acids Res. 1989, 17, 865–882.
0
(1H, d, 30 -OH, J 5.56 Hz), 4.96 (1H, d, H1 R, J
0
R
5.21 Hz), 5.18 (1H, t, 50-OH, J 5.34 Hz), 5.23 (1H, d,
0
0
2R-OH, J 5.98 Hz), 6.03 (1H, dd, H1 , J 7.05, 4.27 Hz),
3. Fourster, C.; Chakraburtty, K.; Sprinzl, M. Nucl. Acids
¨
7.76 (1H, s, H6), 11.24 (1H, br s, N3H). 13C NMR
Res. 1993, 21, 5679–5683.
4. Diefenbach, J.; Burkle, A. Cell. Mol. Life Sci. 2005, 62,
721–730.
5. Takahashi, M.; Kagasaki, T.; Hosoya, T.; Takahashi, S.
J. Antibiot. 1993, 46, 1643–1647; Takahashi, S.; Kino-
shita, T.; Takahashi, M. J. Antibiot. 1994, 47, 95–
100.
6. McCormick, J.; Li, Y.; McCormick, K.; Duynstee, H. I.;
van Engen, A. K.; van der Marel, G. A.; Ganem, B.; van
Boom, J. H.; Meinwald, J. J. Am. Chem. Soc. 1999, 121,
5661–5665.
0
(125 MHz; DMSO-d6): d 12.7 (5-CH3), 38.9 ðC3 Þ, 39.5
0
0
0
0
0
ðC2 Þ 60.5 ðC5 Þ, 62.8 (C5 R), 71.6 (C3 R), 75.6 ðC2 RÞ,
0
0
0
0
84.2 (H1), 85.7 ðC4 Þ, 86.3 ðC4 RÞ, 87.7 ðC1 RÞ, 109.8
(C5), 136.9 (C6), 151.0 (C2), 164.7 (C4). m/z (ES+):
413.1 [M+Na]+. Acc mass: [M+Na]+ 413.0987,
C15H22N2O8S requires 413.0995.
3.8. Conformational analysis
7. Mochizuki, T.; Kondo, Y.; Abe, H.; Taylor, C. W.; Potter,
B. V. L.; Matsuda, A.; Shuto, S. Org. Lett. 2006, 8, 1455–
1458.
8. Mikhailov, S. N.; De Clercq, E.; Herdewijn, P. Nucleosides
Nucleotides 1996, 15, 1323–1334.
9. Wilson, J. C.; Kiefel, M. J.; Angus, D. I.; von Itzstein, M.
Org. Lett. 1999, 1, 443–446.
10. Sabbagh, G.; Fettes, K. J.; Gosain, R.; O’Neil, I. A.;
Cosstick, R. Nucl. Acids Res. 2004, 32, 495–501.
11. Rinkel, L. J.; Altona, C. J. Biomol. Struct. Dyn. 1987, 4,
621–649.
12. Beevers, A. P. G.; Witch, E. M.; Jones, B. C. N. M.;
Cosstick, R.; Arnold, J. R. P.; Fisher, J. Magn. Reson.
Chem. 1999, 37, 814–820.
13. Rosemeyer, H.; Toth, G.; Golankiewicz, B.; Kazimierc-
zuk, Z.; Bourgeois, W.; Kretschmer, U.; Muth, H.-P.;
Seela, F. J. Org. Chem. 1990, 55, 5784–5790; Wuthrich, K.
In NMR of Proteins and Nucleic Acids; J. Wiley & Sons:
New York, 1986.
14. gNMR v. 4.1.2. Adept Scientific, Oxford, 2001.
15. Altona, C.; Sundaralingham, M. J. Am. Chem. Soc. 1972,
94, 8205–8212.
16. Saenger, W. Principles of Nucleic Acid Structure; Springer:
New York, 1984.
1
1D H NMR spectra for 7 and 8 in D2O solution were
recorded on a Varian Unity Inova 500, at 21, 26 and
31 ꢁC, in 16k of complex data points over a spectral
width of 6000 Hz. A relaxation delay of 1 s was used
followed by 1.5 s presaturation time. Prior to Fourier
transformation a, ꢀ1 Hz line broadening and a 0.6
Gaussian factor were applied for resolution enhance-
ment. 1D NOE spectra were recorded using the ‘cycle-
noe’ routine and a total NOE buildup time of 7 s.
DQF-COSY and TOCSY spectra were recorded with
hypercomplex phase cycling, in 2k of data points and
256 pairs of t1 increments. An 80 ms spinlock period
was used in the TOCSY experiment. Prior to Fourier
transformation a Gaussian window function using
LB = ꢀ1 Hz and GB = 0.06 was applied and the fid
zero filled to 2k in F1. All spectral processing was per-
1
formed using vnmr. 1D H NMR spectra were simu-
lated using gNMR14 version 4.1.2 (Adept Scientific,
Oxford). Computation of the conformational analysis
was performed on a PC using the program PSEUROT
version 6.3.18
17. Hasnoot, C. A. G.; De Leeuw, F. A. M.; Altona, C. Org.
Magn. Reson. 1981, 15, 43–52.
18. De Leeuw, F. A. M.; Altona, C. J. Compt. Chem. 1983, 4,
438–448; van Wijk, J.; Haasnoot, C. A. G.; de Leeuw, F.
A. M.; Huckriede, B. D.; Westra Hoekzema, A.; Altona,
C.; Leiden Institute of Chemistry, Leiden University, The
Netherlands, 1999.
19. Plavec, J.; Tong, W.; Chattopadhyaya, J. J. Am. Chem.
Soc. 1993, 115, 9734–9746.
20. Juaristi, E.; Cuevas, G. Tetrahedron 1992, 48, 5019–
5087.
Acknowledgements
We would like to thank the BBSRC for support (Grant
B18146) and the EPSRC for studentships to J.B. We are
also grateful to Allan Mills (Liverpool) for obtaining
mass spectra.
21. Thibaudeau, C.; Plavec, J.; Garg, N.; Papchikhin, A.;
Chattopadhyaya, J. J. Am. Chem. Soc. 1994, 116, 4038–
4043.
Supplementary data
22. Efimtseva, E. V.; Shelkunova, A. A.; Mikhailov, S. N.;
Nauwelaerts, K.; Rozenski, J.; Lescrinier, E.; Herdewijn,
P. Nucleosides Nucleotides Nucl. Acids 2003, 22, 359–
371.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.carres.2006.11.007.