Preparation and cleavage reactions of 3′-thiouridylyl-(3′→5′)-uridine

Article abstract of DOI:10.1039/b002155p

3′-Thiouridylyl-(3′→5′)-uridine [(Us)pU] 3 is prepared by coupling together the disulfide 14 and the 5′-H-phosphonate 18, and then removing the protecting groups. (Us)pU 3 readily undergoes cleavage in 0.05 mol dm^-3 sodium glycinate buffer (pH 10.06) at 50 °C to give, in the first instance, uridine 4 and 3′-thiouridine 2′,3′-cyclic phosphorothioate 21; in glacial acetic acid at 30 °C, it rapidly undergoes cleavage in essentially the same way. The behaviour of (Us)pU 3 is compared with that of uridylyl-(3′→5′)-uridine (UPU) 1a under the same basic and acidic reaction conditions. (Us)pU 3 and 3′-thiouridine 2′,3′-cyclic phosphorothioate 21 are both substrates for ribonuclease A; (Us)pU 3 is a substrate for Crotalus adamanteus snake venom phosphodiesterase but not for calf spleen phosphodiesterase.

Full text of DOI:10.1039/b002155p

View Online / Journal Homepage / Table of Contents for this issue
Preparation and cleavage reactions of 3Ј-thiouridylyl-(3Ј→5Ј)-
uridine
Xiaohai Liu and Colin B. Reese*
1
Department of Chemistry, King’s College London, Strand, London, UK WC2R 2LS.
Fax: ϩ44 (0)20 7848 1771; E-mail: colin.reese@kcl.ac.uk
Received (in Cambridge, UK) 16th March 2000, Accepted 22nd May 2000
Published on the Web 26th June 2000
3Ј-Thiouridylyl-(3Ј→5Ј)-uridine [(Us)pU] 3 is prepared by coupling together the disulfide 14 and the 5Ј-H-
phosphonate 18, and then removing the protecting groups. (Us)pU 3 readily undergoes cleavage in 0.05 mol dmϪ3
sodium glycinate buffer (pH 10.06) at 50 ЊC to give, in the first instance, uridine 4 and 3Ј-thiouridine 2Ј,3Ј-cyclic
phosphorothioate 21; in glacial acetic acid at 30 ЊC, it rapidly undergoes cleavage in essentially the same way.
The behaviour of (Us)pU 3 is compared with that of uridylyl-(3Ј→5Ј)-uridine (UpU) 1a under the same basic
and acidic reaction conditions. (Us)pU 3 and 3Ј-thiouridine 2Ј,3Ј-cyclic phosphorothioate 21 are both substrates
for ribonuclease A; (Us)pU 3 is a substrate for Crotalus adamanteus snake venom phosphodiesterase but not for
calf spleen phosphodiesterase.
characteristic property of an oligoribonucleotide in that the
Introduction
2Ј-thiol function does not appear to interact with the inter-
nucleotide phosphodiester linkage either under basic or acidic
conditions. Furthermore, 2Ј-thiouridylyl-(3Ј→5Ј)-uridine 1b
is not a substrate for ribonuclease A. We have also studied
the effect of replacing the 5Ј-bridging oxygen of UpU 1a by a
sulfur atom. The properties of the resulting analogue, uridylyl-
(3Ј→5Ј)-(5Ј-thiouridine) [Up(sU)]5 2 are somewhat similar
to those of UpU 1a itself except that Up(sU) is very much
(by a factor of ≈4 orders of magnitude) more susceptible to
hydrolysis under basic conditions than is UpU. These observ-
ations have been confirmed by other workers.6,7 In order to
complete this particular group of studies, we have examined
the effect of replacing the 3Ј-bridging oxygen of UpU 1a by a
sulfur atom to give 3Ј-thiouridylyl-(3Ј→5Ј)-uridine [(Us)pU] 3.
We now describe the detailed results of this last study, which
has already been reported in a preliminary form.8
The concept of antisense chemotherapy1 has stimulated studies
in the synthesis of oligodeoxyribo- and oligoribo-nucleotide
analogues, and especially of analogues in which the sugar resi-
dues or internucleotide linkages are modified. Research work
concerned with the elucidation of the mechanism of ribozyme
action2 has also led to the synthesis of oligonucleotide
analogues. Our own studies3 in this area have to a large extent
been motivated by a long-standing interest in the fundamental
chemistry of ribonucleic acids (RNAs). As part of these studies,
we have examined the effect of replacing the 2Ј-hydroxy func-
tion of uridylyl-(3Ј→5Ј)-uridine (UpU) 1a by a thiol function
(as in 1b). The results of this particular study proved to be
relatively uninteresting in that 2Ј-thiouridylyl-(3Ј→5Ј)-uridine4
1b does not display the equivalent of what is arguably the most
Results and discussion
3Ј-Thiouridylyl-(3Ј→5Ј)-uridine 3 was prepared from a nucleo-
side building block 14 (Scheme 1) and the triethylammonium
salt of 3-N-benzoyl-2Ј,3Ј-di-O-benzoyluridine 5Ј-H-phosphon-
ate 18 (Scheme 2). In the preparation of compound 14, uridine
4 was first converted (Scheme 1a) via its 5Ј-O-trityl-2Ј,3Ј-di-O-
mesyl derivative 5 into the lyxo-epoxide9 6 in high overall yield.
As previously reported,9 the lyxo-epoxide reacted with the con-
jugate base of 4-methoxytoluene-α-thiol 7 to give a mixture of
the isomeric sulfides 8 and 9 in the respective proportions of
≈2:1. The products were readily separated by chromatography
on silica gel. The configuration at C-2Ј of the major isomer 8
was inverted by Mitsunobu esterification10 with 4-nitrobenzoic
acid, followed by deacylation with alcoholic methylamine.
When the product was heated in 80% acetic acid solution, under
reflux, 3Ј-S-(4-methoxybenzyl)-3Ј-thiouridine 10 was obtained
in ≈16% overall yield for the seven steps starting from uridine.
The regio- and stereo-chemistry of compound 10 is based on 1H
and 13C NMR spectroscopic data. It is clear from the 1H COSY
spectrum of compound 10 that its secondary hydroxy func-
tion (at δ 5.92) is attached to H-2Ј (δ 4.01) and that the
(4-methoxybenzyl)sulfanyl group is attached to H-3Ј (δ 3.19).
The interactions between H-6 (δ 8.02) and H-2Ј and H-3Ј in the
NOESY spectrum of 3Ј-S-(4-methoxybenzyl)-3Ј-thiouridine 10
DOI: 10.1039/b002155p
J. Chem. Soc., Perkin Trans. 1, 2000, 2227–2236
This journal is © The Royal Society of Chemistry 2000
2227

View Online
and its conversion, via its 3Ј-O-mesyl derivative 16, into 3Ј-S-
(4-methoxybenzyl)-3Ј-thiouridine 10 proceeded in only 15%
overall yield for the three steps.
The triethylammonium salt of 3-N-benzoyl-2Ј,3Ј-di-O-
benzoyluridine 5Ј-H-phosphonate 18 was prepared (Scheme 2)
Scheme 2 Reagents and conditions: i, Ph3CCl, C5H5N, 100 ЊC; ii, BzCl,
C5H5N, room temp.; iii, AcOH–water (4:1 v/v), reflux; iv, a, reagent
prepared from PCl3, Et3N, 1H-1,2,4-triazole and THF, Ϫ35 ЊC; b, Et3N,
water, THF, Ϫ35 ЊC to room temp.
from uridine in four steps. Uridine was first converted via its
5Ј-O-trityl derivative into 3-N-benzoyl-2Ј,3Ј-di-O-benzoyl-
uridine14 17, which was isolated in ≈48% overall yield for the
three steps. Treatment of the latter compound 17 with a reagent
prepared15 from phosphorus trichloride, 1H-1,2,4-triazole and
triethylamine in THF followed by hydrolysis with aqueous
triethylamine gave the desired triethylammonium salt of the
corresponding 5Ј-H-phosphonate 18 in ≈90% overall yield. This
material, which was isolated as a colourless solid, was found by
reversed-phase HPLC to be ≈99% pure.
The nucleoside building block 14 and 3-N-benzoyl-2Ј,3Ј-di-
O-benzoyluridine 5Ј-H-phosphonate 18 were coupled together
by a modification of a procedure first reported by Cosstick and
his co-workers16 in the synthesis of a corresponding dinucleo-
side phosphorothioate in the deoxy-series. Thus the disulfide 14
and the H-phosphonate 18 were treated (Scheme 3) with chloro-
trimethylsilane and triethylamine in dry dichloromethane
solution at room temperature. The benzoyl protecting groups
were then removed by treatment with methanolic ammonia at
room temperature and the Fpmp protecting groups were also
removed at room temperature by hydrolysis under mild acidic
conditions.17 The products were finally fractionated by anion-
exchange chromatography to give the triethylammonium salt
of 3Ј-thiouridylyl-(3Ј→5Ј)-uridine [(Us)pU] 3. This material,
which was isolated as a colourless solid, was characterised on
the basis of 1H and 31P NMR spectroscopic data and its chem-
ically and enzymically promoted cleavage reactions (see below);
its homogeneity was established both by 31P NMR spectroscopy
(Fig. 1a) and reversed-phase HPLC (Fig. 1b). The chemical
shift of the phosphorus resonance (δP [D2O] 18.7) in the 31P
NMR spectrum of (Us)pU 3 (Fig. 1b) is fully consistent5,18 with
that of a phosphorothioate diester with a bridging sulfur atom.
In order to characterise the main nucleotide cleavage
products (see below) of (Us)pU 3, the preparation of 3Ј-thio-
uridine 3Ј-phosphorothioate 20 and 3Ј-thiouridine 2Ј,3Ј-cyclic
phosphorothioate 21 was undertaken (Scheme 4). 3Ј-S-(2,4-
Dinitrophenylsulfanyl)-3Ј-thiouridine 12 (Scheme 1a) was
allowed to react with tris(trimethylsilyl) phosphite.19 Following
hydrolytic work-up and fractionation of the products by anion-
exchange chromatography, triethylammonium 3Ј-thiouridine
3Ј-phosphorothioate 20 was obtained and isolated as a colour-
Scheme 1 Reagents and conditions: i, Ph3CCl, C5H5N, 100 ЊC; ii,
CH3SO2Cl, C5H5N; iii, aq. NaOH, 1,4-dioxane, 100 ЊC; iv, 7, NaH,
DMA, 100 ЊC; v, 4-nitrobenzoic acid, Ph3P, EtO2CN᎐NCO2Et, THF,
᎐
room temp.; vi, MeNH2, EtOH, room temp.; vii, AcOH–H2O (4:1, v/v),
reflux; viii, 11, CF3CO2H, CH2Cl2, 0 ЊC; ix, 13, CF3CO2H, CH2Cl2,
room temp.
provide confirmatory evidence for the configuration both at
C-2Ј and C-3Ј. Further firm evidence in support of the constitu-
tion of the latter compound 10 comes from its alternative
preparation (see below) from 1-(2Ј,5Ј-di-O-trityl-β--xylo-
furanosyl)uracil11 15. 3Ј-S-(4-Methoxybenzyl)-3Ј-thiouridine
10 reacted rapidly9 with 2,4-dinitrobenzenesulfenyl chloride 11
in the presence of trifluoroacetic acid (TFA) in dichloro-
methane solution to give 3Ј-S-(2,4-dinitrobenzenesulfanyl)-3Ј-
thiouridine 12. When the latter compound 12 was treated with
1-(2-fluorophenyl)-4-methoxy-1,2,3,6-tetrahydropyridine12,13
13 also in the presence of TFA in dichloromethane solution,
the required nucleoside building block 14 was obtained in ≈70%
overall yield for the two steps starting from compound 10. The
alternative preparation (Scheme 1b) of the latter compound
10 also involved seven steps, starting from uridine. However,
1-(2Ј,5Ј-di-O-trityl-β--xylofuranosyl)uracil11 15, which may
be prepared from uridine in four steps, is relatively inaccessible
2228
J. Chem. Soc., Perkin Trans. 1, 2000, 2227–2236

View Online
Fig.
1
(a) 31P NMR spectrum (145.8 MHz; D2O) of triethyl-
ammonium 3Ј-thiouridylyl-(3Ј→5Ј)-uridine [(Us)pU] 3; (b) reversed-
phase HPLC profile (programme 2) of (Us)pU 3.
Fig. 2 Reversed-phase HPLC profiles (programme 2) of (a) triethyl-
ammonium 3Ј-thiouridine 3Ј-phosphorothioate 20 and (b) 3Ј-thio-
uridine 2Ј,3Ј-cyclic phosphorothioate 21.
Scheme 4 Reagents and conditions: i, a, Me3SiCl, Et3N, CH2Cl2, room
temp.; b, (Me3SiO)3P; ii, ClCO2Et, Bu3N, water, room temp.
hydrolytic cleavage to give 3Ј-thiouridine 3Ј-phosphorothioate
20. In a separate experiment, carried out in 0.05 mol dmϪ3
sodium glycinate buffer (pH 9.87) at 30 ЊC, the cyclic phos-
phorothioate 21 was cleanly converted into the 3Ј-phosphoro-
thioate 20. This reaction also displayed pseudo-first-order
kinetics with t1/2 ca. 164 min. The product composition for the
hydrolysis of (Us)pU 3 after 45 min under the above reaction
conditions (i.e. pH 10.06, 50 ЊC; Scheme 5, i) is illustrated in
Fig. 3a. On the basis of their retention times, the four major
components were identified as uridine 4 (tR 3.37 min, 37% of
total absorbance), 3Ј-thiouridine 2Ј,3Ј-cyclic phosphorothioate
21 (tR 4.83 min, 14% of total absorbance), 3Ј-thiouridine 3Ј-
phosphorothioate 20 (tR 6.27 min, 22.9% of total absorbance)
and unchanged (Us)pU 3 (tR 7.76 min, 25.3% of total absorb-
ance). The action of 0.05 mol dmϪ3 sodium glycinate buffer (pH
10.06) on (Us)pU 3 at 50 ЊC was also monitored by 31P NMR
spectroscopy. The reaction was carried out on a larger scale
and, after 30 min, the products were neutralised and worked up.
The 31P NMR spectrum (D2O) of the products is illustrated in
Fig. 4a: the resonance signals at δ 37.2, 18.7 and 16.3 ppm may
be assigned to 3Ј-thiouridine 2Ј,3Ј-cyclic phosphorothioate
21, (Us)pU 3 and 3Ј-thiouridine 3Ј-phosphorothioate 20,
respectively. This larger scale reaction (see Experimental
section) appeared to proceed more slowly (t1/2 > 30 min) than
the reaction described above. The point of greatest interest to
emerge from these results is that 3Ј-thiouridylyl-(3Ј→5Ј)-uridine
Scheme 3 Reagents and conditions: i, Me3SiCl, Et3N, CH2Cl2, room
temp.; ii, NH3, MeOH, room temp.; iii, AcOH–water (2:98 v/v),
room temp.
1
less solid. This material was characterised on the basis of H
and 31P (δP[D2O] 16.4) NMR data and its homogeneity was
confirmed by HPLC analysis (Fig. 2a). When an aqueous
solution of 3Ј-thiouridine 3Ј-phosphorothioate 20 was treated
with ethyl chloroformate and tributylamine, according to
Michelson’s procedure,20 3Ј-thiouridine 2Ј,3Ј-cyclic phosphoro-
thioate 21 was obtained. This material, which was isolated as a
colourless solid following anion-exchange and silanised silica
gel chromatography, was again characterised on the basis of 1H
and 31P (δP[D2O] 37.0) NMR data. The relative homogeneity
of the cyclic phosphorothioate 21 can be seen from its reversed-
phase HPLC profile (Fig. 2b).
It was important to investigate the base- and acid-catalysed
hydrolysis of (Us)pU 3 and to determine whether or not it was a
substrate for common phosphodiesterases. (Us)pU 3 was found
to undergo hydrolysis rapidly in 0.05 mol dmϪ3 sodium
glycinate buffer (pH 10.06) at 50 ЊC. Pseudo-first-order kinetics
were observed with a half-life (t1/2) of ca. 25 min. The primary
hydrolysis products (Scheme 5) were found to be uridine 4 and
3Ј-thiouridine 2Ј,3Ј-cyclic phosphorothioate 21. Under these
conditions, the cyclic phosphorothioate 21 underwent further
J. Chem. Soc., Perkin Trans. 1, 2000, 2227–2236
2229

View Online
Scheme 5 Reagents and conditions: i, 0.05 mol dmϪ3 sodium glycinate buffer (pH 10.06), 50 ЊC; ii, glacial acetic acid, 30 ЊC.
[(Us)pU] 3 undergoes base-catalysed hydrolysis at pH ≈ 10
much more rapidly than does uridylyl-(3Ј→5Ј)-uridine (UpU)
1a. In 0.05 mol dmϪ3 sodium glycinate buffer (pH 10.06) at
50 ЊC, the half-life of hydrolysis of UpU 1a was found to be ca.
80–90 h. Thus, under these conditions, (Us)pU 3 undergoes
base-catalysed hydrolysis at a rate approximately 200 times
faster than does unmodified UpU 1a. A similar result was sub-
sequently reported by other workers21,22 in connection with a
study relating to 3Ј-thioinosylyl-(3Ј→5Ј)-uridine 27 (see below).
The reason for this enhanced rate of hydrolysis is not obvious
as the cleavage of (Us)pU 3 to give the cyclic phosphorothioate
21, like the corresponding cleavage of UpU 1a, involves a
simple ester-exchange reaction without direct participation of
the sulfur atom. An explanation or partial explanation for the
greater lability of (Us)pU 3 is that the presence of the relatively
large 3Ј-sulfur atom leads to a lower-energy transition state in
Fig. 3 (a) Reversed-phase HPLC profile (programme 2) of products
obtained by treating 3Ј-thiouridylyl-(3Ј→5Ј)-uridine [(Us)pU] 3 with
0.5 mol dmϪ3 sodium glycinate buffer (pH 10.06) at 50 ЊC for 45 min;
(b) reversed-phase HPLC profile (programme 2) of products obtained
by treating (Us)pU 3 with glacial acetic acid at 30 ЊC for 10 min; (c)
reversed-phase HPLC profile (programme 3) of products obtained by
treating uridylyl-(3Ј→5Ј)-uridine [UpU] 1a with glacial acetic acid at
30 ЊC for 37 min. See appropriate parts of the text for the identification
of the component peaks.
the cyclisation reaction and to a less strained product 21.
3Ј-Thiouridylyl-(3Ј→5Ј)-uridine 3 was found to decompose
rapidly in glacial acetic solution at 30 ЊC. Precise kinetic data
were not obtained but, under these conditions (Scheme 5, ii), t1/2
for the decomposition of (Us)pU 3 was ca. 4 min. The product
composition after 10 min is illustrated in Fig. 3b. On the basis
of their retention times, the major components were identified
as uridine 4 (tR 3.30 min, 38.3% of total absorbance), 3Ј-
thiouridine 2Ј,3Ј-cyclic phosphorothioate 21 (tR 4.71 min,
36.2% of total absorbance), 3Ј-thiouridine 3Ј-phosphorothioate
20 (tR 6.19 min, 4.6% of total absorbance) and unchanged
(Us)pU 3 (tR 7.67 min, 18.6% of total absorbance). Thus the
cleavage products obtained in glacial acetic acid and indeed the
course of the cleavage reactions themselves (Scheme 5, ii) are
qualitatively similar to those observed in pH 10.06 sodium
glycinate buffer (Scheme 5, i; see also Fig. 3a). There are, how-
ever, two significant differences. First, the proportions of cyclic
phosphorothioate 21 and 3Ј-phosphorothioate 20 obtained are
different and secondly, a fifth, albeit minor product (tR 7.51
min, 2.3% of total absorbance; Scheme 3b) is obtained in the
glacial acetic acid reaction. As the latter reagent is essentially
free from water, it is hardly surprising that relatively little
hydrolysis of the cyclic phosphorothioate 21 occurs and there-
fore that the proportion of 3Ј-phosphorothioate 20 obtained
is small. Although it has not been fully characterised, the fifth
and least abundant product (Fig. 3b; tR 7.51 min) may be 3Ј-
thiouridylyl-(2Ј→5Ј)-uridine 22 i.e. the product of phosphoryl
migration. There is some evidence in favour of this conclusion.
First, this product appears, like the isomeric 2Ј-thiouridylyl-
(3Ј→5Ј)-uridine4 1b, to be stable under acidic conditions.
Secondly, NMR spectroscopic evidence has been obtained in
support of the fifth product being a phosphate rather than a
phosphorothioate ester. A solution of (Us)pU 3 in glacial acetic
acid was maintained at 30 ЊC for 25 min. The 31P NMR spec-
trum (in D2O) of the products is illustrated in Fig. 4b: the
resonance signals at δ 37.2 (≈81%), 18.7 (≈11%) and 16.8 (≈5%)
may, like the corresponding signals in Fig. 4a, be assigned to
3Ј-thiouridine 2Ј,3Ј-cyclic phosphorothioate 21, unchanged
(Us)pU 3 and 3Ј-thiouridine 3Ј-phosphorothioate 20, respec-
tively. The fourth resonance signal at δ Ϫ0.44 (≈3%) may pos-
sibly be assigned to 3Ј-thiouridylyl-(2Ј→5Ј)-uridine 22. The
phosphorus signal in the 31P NMR spectrum (D2O) of the iso-
meric 2Ј-thiouridylyl-(3Ј→5Ј)-uridine 1b resonates4 at δ Ϫ0.3.
However, the signal at δ Ϫ0.44 (Fig. 4b) could alternatively
be assigned to the phosphorus resonance of 3Ј-thiouridine 2Ј-
phosphate, which would be obtained if acetic acid also pro-
moted the hydrolytic cleavage of the P–S bond of 3Ј-thiouridine
2Ј,3Ј-cyclic phosphorothioate 21.
It has been known for many years23 that, under acidic condi-
tions (e.g. pH 1.0, 25 ЊC), UpU 1a both isomerises to uridylyl-
(2Ј→5Ј)-uridine 23 and undergoes hydrolytic cleavage. We now
report that UpU 1a is more stable (t1/2 ca. 60 min) in glacial
acetic acid solution at 30 ЊC (Scheme 6) than is (Us)pU 3. The
product composition after 37 min, which includes uridine 4 (tR
3.29 min, 11.8% of total absorbance), uridine 2Ј,3Ј-cyclic phos-
phate 24 (tR 3.53 min, 10.0% of total absorbance), uridine
2Ј(3Ј)-phosphates 25/26 (tR 4.99, 5.17 min, 1.5% of total
absorbance), uridylyl-(2Ј→5Ј)-uridine 23 (tR 7.71 min, 12.8% of
total absorbance) and unchanged UpU 1a (tR 9.27 min, 63.9%
of total absorbance), is illustrated in Fig. 3c. Compared with
(Us)pU 3, UpU 1a undergoes acetic acid-promoted cleavage to
give uridine 4 and uridine 2Ј,3Ј-cyclic phosphate 24 compar-
atively slowly, but phosphoryl migration to give the isomeric
dinucleoside phosphate 23 occurs much more readily. Indeed,
it is by no means firmly established from the data described
2230
J. Chem. Soc., Perkin Trans. 1, 2000, 2227–2236

View Online
Scheme 6 Reagents and conditions: i, glacial acetic acid, 30 ЊC.
comparative rates of cleavage (to give uridine and 3Ј-thio-
inosine 2Ј,3Ј-cyclic phosphorothioate 28 instead of uridine
2Ј,3Ј-cyclic phosphate 24) and isomerisation [to give 3Ј-thio-
inosylyl-(2Ј→5Ј)-uridine 29 instead of uridylyl-(2Ј→5Ј)-uridine
23]. Indeed these authors claimed22, that, at pH 2, the isomeris-
ation of (Is)pU 27 to give the (2Ј→5Ј)-dinucleoside phosphate
29 proceeded more rapidly than the cleavage reaction. It is per-
haps worth noting that the reactions between 0.1 mol dmϪ3
hydrochloric acid (pH 1.0)17,23 and UpU 1a and glacial acetic
acid and UpU are similar except that in the aqueous medium
uridine 2Ј,3Ј-cyclic phosphate 24 does not accumulate but is
converted into a mixture of uridine 2Ј- and 3Ј-phosphates 25 and
26. Like Elzagheid et al.,22 we shall not attempt to offer an
explanation for the differences between the results obtained
from the action of glacial acetic acid on (Us)pU8 3 and
those obtained from the action of hydrochloric acid on (Is)pU
27. However, this apparent discrepancy has encouraged us to
report our own results in greater detail and to illustrate them
with what we believe to be firm HPLC and 31P NMR spectro-
scopic evidence. We also considered it important to describe in
some detail the independent synthesis and characterisation of
the main cleavage products (i.e. 3Ј-thiouridine 2Ј,3Ј-cyclic
phosphorothioate 21 and 3Ј-thiouridine 3Ј-phosphorothioate
20) that were obtained both by the action of pH 10.06 sodium
glycinate buffer and glacial acetic acid on (Us)pU 3.
Fig. 4 31P NMR spectra (145.8 MHz; D2O) of products obtained by
treating 3Ј-thiouridylyl-(3Ј→5Ј)-uridine [(Us)pU] 3 (a) with 0.05 mol
dmϪ3 sodium glycinate buffer (pH 10.06) at 50 ЊC for 30 min and
(b) with glacial acetic acid at 30 ЊC for 25 min.
above (Figs. 3b and 4b) that (Us)pU 3 displays any tendency
whatsoever to isomerise to 3Ј-thiouridylyl-(2Ј→5Ј)-uridine 22 in
glacial acetic acid solution. Furthermore, if isomerisation does
occur (i.e. if the minor component with tR = 7.51 min in Fig. 3b
is 3Ј-thiouridylyl-(2Ј→5Ј)-uridine 22), the isomerisation reac-
tion is very much slower than the cleavage reaction that leads
initially to 3Ј-thiouridine 2Ј,3Ј-cyclic phosphorothioate 21 and
uridine 4.
In the light of the results of the above experiments relating to
the action of glacial acetic acid on (Us)pU 3, our attention was
drawn to a very recent report by Elzagheid et al.22 concerning
the action of hydrochloric acid at 90 ЊC and over a range
of pHs on 3Ј-thioinosylyl-(3Ј→5Ј)-uridine [(Is)pU] 27. These
authors reported22 that at pH 2 and below the behaviour of
(Is)pU 27 corresponds to that of UpU 1a with regard to the
Finally, the action of ribonuclease A, Crotalus adamanteus
snake venom phosphodiesterase and calf spleen phospho-
diesterase on (Us)pU 3 was examined. Both (Us)pU 3 and 3Ј-
thiouridine 2Ј,3Ј-cyclic phosphorothioate 21 proved to be very
good substrates for ribonuclease A. At 30 ЊC (Us)pU 3 under-
went over 80% digestion in the presence of ribonuclease A in
J. Chem. Soc., Perkin Trans. 1, 2000, 2227–2236
2231

View Online
pH 7.5 Tris hydrochloric buffer solution in 26 min to give a
mixture of uridine 4, 3Ј-thiouridine 2Ј,3Ј-cyclic phosphorothio-
ate 21 and 3Ј-thiouridine 3Ј-phosphorothioate 20. The relative
proportions of the cyclic phosphorothioate 21 and the 3Ј-
phosphorothioate 20 were ≈1:2. After 146 min, the substrate
had been virtually completely digested to uridine 4 and 3Ј-
thiouridine 3Ј-phosphorothioate 20. In a separate experiment,
which was carried out under similar conditions, 3Ј-thiouridine
2Ј,3Ј-cyclic phosphorothioate 21 underwent ≈30% ribonuclease
A-promoted digestion to 3Ј-thiouridine 3Ј-phosphorothioate 20
in 5 min and complete digestion in 155 min. Perhaps unsurpris-
ingly, (Us)pU 3 proved (see Experimental section), like UpU
1a, to be a substrate for Crotalus adamanteus snake venom
phosphodiesterase but, unlike UpU, not to be a substrate for
calf spleen phosphodiesterase.
organic layers were combined, dried (MgSO4), and evaporated
under reduced pressure. The residue was dissolved in 1,4-
dioxane (1400 cm3). Aq. sodium hydroxide (2.5 mol dmϪ3; 140
cm3, 0.35 mol) was then added and the reaction mixture was
heated under reflux. After 3 h, the products were cooled to
room temperature; they were then neutralised (to pH 7) by the
slow addition of glacial acetic acid and concentrated to dryness
under reduced pressure. A solution of the residue in chloroform
(500 cm3) was washed first with saturated aq. sodium hydrogen
carbonate (2 × 250 cm3) and then with brine (250 cm3). The
dried (MgSO4) organic layer was evaporated under reduced
pressure. The residue was fractionated by short-column chrom-
atography on silica gel: the appropriate fractions, which were
eluted with chloroform–methanol (99:1 v/v), were evaporated
under reduced pressure to give 1-(5Ј-O-trityl-2Ј,3Ј-anhydro-β-
-lyxofuranosyl)uracil9 6, as a colourless foam (23.66 g); Rf
0.60 (system A); δH [(CD3)2SO] 3.21 (2 H, m), 4.11 (2 H, m),
4.27 (1 H, m), 5.63 (1 H, d, J 8.1), 6.12 (1 H, s), 7.25–7.45 (15 H,
m), 7.52 (1 H, d, J 8.1), 11.48 (1 H, br s); δC [(CD3)2SO] 55.5,
55.7, 62.4, 75.9, 80.9, 86.3, 102.0, 127.2, 128.0, 128.2, 140.9,
143.4, 150.4, 163.0.
Experimental
1H and 13C NMR spectra were measured at 360.1 and 90.6
MHz, respectively, with a Bruker AM 360 spectrometer;
tetramethylsilane was used as an internal standard. J-Values are
given in Hz. 31P NMR spectra were measured at 145.8 MHz
with the same spectrometer; 85% orthophosphoric acid was
used as an external standard. UV spectra were measured with
a Perkin-Elmer model 552 spectrophotometer. Merck silica
gel 60 F254 pre-coated plates (Art 5715 and 5642) which,
unless otherwise stated, were developed in solvent system A
[chloroform–methanol (9:1 v/v)], were used for TLC. Liquid
chromatography (HPLC) was carried out on a Jones Apex
Octyl 10µ column, which was eluted with 0.1 mol dmϪ3 triethyl-
ammonium acetate buffer–acetonitrile mixtures with a flow rate
of 1.5 cm3 minϪ1: programme 1 involved a linear gradient over a
period of 10 min starting with buffer–acetonitrile (70:30 v/v)
and ending with buffer–acetonitrile (30:70 v/v); programme
2 involved a linear gradient over 10 min starting with buffer–
acetonitrile (97:3 v/v) and ending with buffer–acetonitrile
(80:20 v/v), followed by a linear gradient over 5 min ending
with buffer–acetonitrile (70:30 v/v); programme 3 involved a
linear gradient over 10 min starting with buffer–acetonitrile
(97:3 v/v) and ending with buffer–acetonitrile (93:7 v/v); pro-
gramme 4 involved a linear gradient over a period of 10 min
starting with buffer–acetonitrile (100:0 v/v) and ending with
buffer–acetonitrile (80:20 v/v). Merck Kieselgel H (Art 7729)
was used for short-column chromatography; Merck silan-
ised silica gel (Art 7719) was used for reversed-phase
column chromatography. DEAE Sephadex A-25 was used for
anion-exchange chromatography. Acetonitrile, tetrahydrofuran
(THF) and triethylamine were dried by heating with calcium
hydride, under reflux, for 3–5 h and were then distilled under
atmospheric pressure; N,N-dimethylacetamide (DMA) was
dried by heating over calcium hydride and was then distilled
under reduced pressure; dichloromethane was dried by heating,
under reflux, over phosphorus pentaoxide and was then dis-
tilled. All solvents were stored over molecular sieves (no.
4 A). Phosphorolytic enzymes were purchased from the Sigma
Chemical Company.
4-Methoxytoluene-α-thiol 7 (10.2 cm3, 73.2 mmol) was added
dropwise to a stirred suspension of sodium hydride (60% dis-
persion in mineral oil; 2.36 g, 59.0 mmol) in DMA (35 cm3) at
0 ЊC (ice–water-bath). The resulting solution was stirred at
room temperature for a further period of 30 min. 1-(5Ј-O-
Trityl-2Ј,3Ј-anhydro-β--lyxofuranosyl)uracil 6 (13.84 g) was
added in four portions over a period of 15 min and the stirred
reactants were then heated at 100 ЊC. After 1 h, the products
were concentrated under reduced pressure. The residue was dis-
solved in chloroform (500 cm3) and the resulting solution was
washed first with brine (4 × 300 cm3) and then with water (300
cm3). The dried (MgSO4) organic layer was evaporated under
reduced pressure. TLC (system A) revealed a major (Rf 0.51)
and a minor (Rf 0.60) product. This mixture was fractionated
by short-column chromatography on silica gel. The minor
component was eluted with petroleum spirit (distillation range
40–60 ЊC)–ethyl acetate (4:1 v/v). Elution of the column with
chloroform–methanol (99:1 v/v) and evaporation of the
appropriate fractions gave 1-[3Ј-S-(4-methoxybenzyl)-3Ј-thio-
5Ј-O-trityl-β--arabinofuranosyl]uracil9 8 as a colourless foam
(9.36 g); Rf 0.51 (system A); δH [(CD3)2SO] 3.20 (3 H, m), 3.64
(3 H, s), 3.75 (1 H, d, J 13.1), 3.83 (1 H, d, J 13.4), 3.86 (1 H, m),
4.35 (1 H, m), 5.21 (1 H, d, J 8.2), 6.04 (2 H, m), 6.81 (2 H, d,
J 8.7), 7.19 (2 H, d, J 8.7), 7.31 (15 H, m), 7.68 (1 H, d, J 8.1),
11.34 (1 H, br s); δC [(CD3)2SO] 33.7, 47.2, 55.0, 62.3, 76.4, 79.8,
84.0, 86.3, 100.2, 113.8, 127.2, 128.0, 128.3, 129.6, 130.0, 141.7,
143.2, 150.5, 158.2, 163.1.
The above material 8 (4.75 g), 4-nitrobenzoic acid (2.55 g,
15.3 mmol) and triphenylphosphine (4.41 g, 16.8 mmol) were
dissolved in dry acetonitrile (20 cm3) and the resulting solution
was evaporated under reduced pressure. The residue was re-
evaporated from its solution in dry THF (20 cm3) and then re-
dissolved in dry THF (30 cm3). The stirred solution was cooled
to 0 ЊC (ice–water-bath) and diethyl azodicarboxylate (2.64
cm3, 16.8 mmol) was added dropwise over a period of 25 min.
The stirred reactants were then allowed to warm up to room
temperature. After 17 h, the products were evaporated under
reduced pressure and the residue was fractionated by short-
column chromatography on silica gel. The appropriate frac-
tions, which were eluted with petroleum spirit (40–60 ЊC)–ethyl
acetate (70:30 to 60:40 v/v), were combined, and evaporated
under reduced pressure. The residue was dissolved in alcoholic
methylamine (≈5 mol dmϪ3; 10 cm3) at room temperature. After
10 min, the products were evaporated under reduced pressure
and the residue was fractionated by short-column chromato-
graphy on silica gel. The appropriate fractions, which were
eluted with chloroform–methanol (99:1 v/v), were combined,
and evaporated under reduced pressure. A solution of the resi-
due in acetic acid–water (4:1 v/v; 20 cm3) was heated, under
3Ј-S-(4-Methoxybenzyl)-3Ј-thiouridine 10
(a) A solution of uridine 4 (14.5 g, 59.4 mmol) and chloro-
triphenylmethane (18.2 g, 65.3 mmol) in dry pyridine (200 cm3)
was heated at 100 ЊC for 2 h. The stirred products were cooled
to 0 ЊC (ice–water-bath) and methanesulfonyl chloride (13.8
cm3, 0.178 mol) was added. The reactants were maintained at
4 ЊC for 15 h and water (20 cm3) was then added. After 15
min, the products were evaporated under reduced pressure.
The residue was dissolved in chloroform (500 cm3) and the
resulting solution was washed with saturated aq. sodium
hydrogen carbonate (3 × 300 cm3). The combined aqueous
layers were back-extracted with chloroform (300 cm3). The
2232
J. Chem. Soc., Perkin Trans. 1, 2000, 2227–2236

View Online
reflux, for 30 min. The products were evaporated under reduced
pressure and then fractionated by short-column chromato-
graphy on silica gel; the appropriate fractions, which were
eluted with chloroform–methanol (97:3 v/v) were evaporated
under reduced pressure to give the title compound 10 (1.06 g,
15.8% overall yield for the 7 steps starting from uridine) (Found,
in material crystallised from absolute ethanol: C, 53.5; H, 5.3;
N, 7.3. C17H20N2O6S requires C, 53.67; H, 5.30; N, 7.36%), mp
144–146 ЊC; δH [(CD3)2SO] 3.19 (1 H, dd, J 4.9 and 9.3), 3.63
(1 H, m), 3.72 (3 H, s), 3.73–3.83 (3 H, m), 4.01 (2 H, m), 5.27
(1 H, t, J 4.7), 5.58 (1 H, d, J 8.1), 5.65 (1 H, d, J 1.7), 5.92 (1 H,
d, J 5.2), 6.84 (2 H, d, J 8.6), 7.24 (2 H, d, J 8.6), 8.02 (1 H, d,
J 8.1), 11.29 (1 H, br s); δC [(CD3)2SO] 34.4, 45.3, 55.0, 59.4,
75.5, 84.5, 90.1, 100.9, 113.0, 130.1, 140.2, 150.4, 158.2, 163.2.
(b) A solution of 1-(2Ј,5Ј-di-O-trityl-β--xylofuranosyl)-
uracil)11 15 (3.32 g, 4.5 mmol) and methanesulfonyl chloride
(1.76 cm3, 22.7 mmol) in dry pyridine (18 cm3) was stirred at
room temperature. After 24 h, the products were cooled to 0 ЊC
(ice–water-bath) and saturated aq. sodium hydrogen carbonate
(10 cm3) was added, followed by dichloromethane (300 cm3).
The resulting mixture was extracted, first with saturated aq.
sodium hydrogen carbonate (2 × 200 cm3) and then with water
(200 cm3). The dried (MgSO4) organic layer was evaporated
under reduced pressure and the residue was co-evaporated with
dry toluene (2 × 15 cm3) to give the putative 3Ј-O-mesyl
derivative 16 as a foam.
The above material 12 (0.224 g) and 1-(2-fluorophenyl)-4-
methoxy-1,2,3,6-tetrahydropyridine13 13 (0.71 g, 3.4 mmol)
were dissolved in dry acetonitrile (6 cm3) and the resulting solu-
tion was evaporated under reduced pressure. After this process
had been repeated, the residue was dissolved in dichloro-
methane (9 cm3), and TFA (0.29 cm3, 3.8 mmol) was added.
The reaction solution was stirred at room temperature. After 17
h, triethylamine (1.05 cm3, 7.5 mmol) was added and the prod-
ucts were evaporated under reduced pressure. The residue was
dissolved in chloroform (50 cm3) and the resulting solution was
washed with saturated aq. sodium hydrogen carbonate (2 × 25
cm3). The combined aqueous layers were back-extracted with
chloroform (25 cm3). The combined organic layers were dried
(MgSO4), and evaporated under reduced pressure. The residue
was fractionated by short-column chromatography on silica gel:
the appropriate fractions, which were eluted with chloroform–
methanol (95.5:0.5 to 99:1 v/v), were combined and evapor-
ated under reduced pressure to give the title compound 14
as a yellow foam [0.32 g; ≈70% overall yield for the two steps
based on 3Ј-S-(4-methoxybenzyl)-3Ј-thiouridine 10 as starting
material]; δH [(CD3)2SO] 1.73 (4 H, m), 1.86 (1 H, m), 2.00 (3 H,
m), 2.75–2.95 (6 H, m), 3.05 (3 H, s), 3.15 (5 H, m), 3.51 (1 H,
dd, J 4.1 and 11.0), 3.62 (1 H, m), 3.93 (1 H, m), 4.45 (1 H, m),
4.98 (1 H, t, J 6.0), 5.77 (1 H, d, J 8.1), 6.07 (1 H, d, J 5.3), 6.8–
7.2 (8 H, m), 7.89 (1 H, d, J 8.1), 8.48 (1 H, d, J 9.0), 8.54 (1 H,
dd, J 2.4 and 9.0), 8.82 (1 H, d, J 2.3), 11.49 (1 H, br s);
δC [(CD3)2SO] 32.5, 32.6, 32.8, 33.5, 47.3, 47.4, 47.5, 47.7, 48.0,
52.5, 60.2, 72.9, 79.2, 82.3, 87.2, 98.2, 100.5, 102.5, 140.8, 150.7,
162.9 and signals assignable to the resonances of the 18 aryl
carbon atoms.
4-Methoxytoluene-α-thiol 7 (2.09 cm3, 15 mmol) was added
dropwise to a solution of sodium hydride (60% dispersion in
mineral oil; 0.546 g, 13.65 mmol) in DMA (20 cm3) at 0 ЊC (ice–
water-bath). The cooling bath was then removed. The solution
obtained was stirred at room temperature for 15 min and was
then added to the above 3Ј-O-mesyl derivative 16. The reactants
were heated at 110 ЊC. After 2 h, the products were cooled and
dichloromethane (200 cm3) was added. The resulting solution
was washed with saturated aq. sodium hydrogen carbonate
(2 × 200 cm3). The combined aqueous layers were back-
extracted with dichloromethane (100 cm3). The organic layers
were combined, dried (MgSO4), and evaporated under reduced
pressure. The residue was heated with acetic acid–water (4:1
v/v; 30 cm3), under reflux. After 30 min, the products were
concentrated under reduced pressure and re-evaporated with
cyclohexane (2 × 15 cm3). The residue was fractionated by
short-column chromatography on silica gel: the appropriate
fractions, which were eluted with chloroform–methanol (99:1
to 98:2 v/v), were combined, and evaporated under reduced
pressure to give the title compound 10 (0.265 g, 15.3% overall
yield for the 3 steps), identical (1H, 13C NMR) to the material
obtained in (a) above.
Triethylammonium salt of 3-N-benzoyl-2Ј,3Ј-di-O-benzoyl-
uridine 5Ј-H-phosphonate 18
Uridine 4 (5.00 g, 20.5 mmol), chlorotriphenylmethane (6.29 g,
22.6 mmol) and pyridine (60 cm3) were stirred together in an
atmosphere of argon at 100 ЊC. After 1.5 h, the products were
cooled to 0 ЊC (ice–water-bath) and benzoyl chloride (21.43
cm3, 0.184 mol) was added. The reactants were then allowed to
warm up to room temperature. After 14 h, the stirred products
were cooled to 0 ЊC and water (20 cm3) was added slowly.
After a further period of 10 min, solid sodium hydrogen
carbonate (20 g) was added very carefully in portions. After 15
min, the products were concentrated under reduced pressure
and the residue was partitioned between chloroform (400 cm3)
and saturated aq. sodium hydrogen carbonate (300 cm3). The
solid precipitate was removed by filtration and the layers were
separated. The organic layer was back-extracted with chloro-
form (2 × 150 cm3). The combined organic layers were dried
(MgSO4), and evaporated under reduced pressure. After the
residue had been coevaporated with toluene under reduced
pressure, it was dissolved in acetic acid–water (4:1 v/v; 75 cm3)
and the solution was heated under reflux. After 1.5 h, the
products were concentrated under reduced pressure and the
residue was coevaporated with cyclohexane (2 × 50 cm3). The
residual material was fractionated by short-column chromato-
graphy on silica gel: the appropriate fractions, which were
eluted with chloroform–methanol (99:1 v/v), were combined,
and evaporated under reduced pressure. Crystallisation of the
residual glass from absolute ethanol gave 3-N-benzoyl-2Ј,3Ј-di-
O-benzoyluridine 17 as colourless needles (5.51 g, 48.3%), mp
191–193 ЊC (lit.14 191–193 ЊC); δH [(CD3)2SO] 3.83 (2 H, m),
4.53 (1 H, m), 5.63 (1 H, m), 5.79 (2 H, m), 6.12 (1 H, d, J 8.2),
6.30 (2 H, d, J 5.3), 7.40 (2 H, m), 7.45–7.7 (6 H, m), 7.77 (3 H,
m), 7.79 (4 H, m), 8.31 (1 H, d, J 8.2).
3Ј-S-(2,4-Dinitrophenylsulfanyl)-2Ј,5Ј-bis-O-[1-(2-fluorophenyl)-
4-methoxypiperidin-4-yl]-3Ј-thiouridine 14
TFA (0.115 cm3, 1.5 mmol) was added to a stirred solution
of 3Ј-S-(4-methoxybenzyl)-3Ј-thiouridine 10 (0.190 g, 0.5
mmol) and 2,4-dinitrobenzenesulfenyl chloride 11 (0.234 g, 1.0
mmol) in dry dichloromethane (8 cm3) at 0 ЊC (ice–water-bath).
After 15 min, the products were evaporated under reduced pres-
sure. A solution of the residue in chloroform–methanol (95:5
v/v; 10 cm3) was pre-adsorbed on silica gel (3 g). The products
were then fractionated by short-column chromatography on
silica gel: elution of the column with chloroform–methanol
(97:3 v/v) and concentration of the appropriate fractions gave
3Ј-S-(2,4-dinitrophenylsulfanyl)-3Ј-thiouridine 12 as a yellow
solid (0.217 g); δH [(CD3)2SO] 3.56 (1 H, dd, J 5.3 and 9.0), 3.62
(1 H, m), 3.82 (1 H, m), 4.22 (1 H, m), 4.43 (1 H, m), 5.28 (1 H,
m), 5.54 (1 H, d, J 8.1), 5.72 (1 H, d, J 2.0), 6.63 (1 H, d, J 5.4),
7.93 (1 H, d, J 8.2), 8.56 (1 H, d, J 9.0), 8.62 (1 H, dd, J 2.4 and
9.0), 8.90 (1 H, d, J 2.3), 11.34 (1 H, br s); δC [(CD3)2SO] 52.6,
59.7, 75.2, 84.2, 90.4, 101.2, 121.4, 128.2, 129.2, 140.3, 144.7,
145.0, 145.3, 150.4, 163.2.
Dry triethylamine (5.42 cm3, 39.0 mmol) and phosphorus
trichloride (1.05 cm3, 12.0 mmol) were added to a stirred solu-
tion of 1H-1,2,4-triazole (2.48 g, 36.0 mmol; recrystallised from
dry acetonitrile) in dry THF (72 cm3) at Ϫ35 ЊC (methanol–
solid CO2-bath). After 15 min, a solution of 3-N-benzoyl-2Ј,3Ј-
J. Chem. Soc., Perkin Trans. 1, 2000, 2227–2236
2233

View Online
di-O-benzoyluridine 17 (1.67 g, 3.00 mmol) in dry THF (60
cm3) was added. The stirred reactants were maintained at
Ϫ35 ЊC. After a further period of 30 min, triethylamine–water
(1:1 v/v; 20 cm3) was added. The products were allowed to
warm up to room temperature and were then concentrated
under reduced pressure. The residue was partitioned between
chloroform (200 cm3) and 0.5 mol dmϪ3 triethylammonium
hydrogen carbonate (pH 7.5; 2 × 100 cm3). The organic layer
was dried (MgSO4), and evaporated under reduced pressure.
The residue was fractionated by short-column chromatography
on silica gel: the appropriate fractions, which were eluted
with chloroform–methanol (95:5 v/v), were combined, and
evaporated under reduced pressure. A solution of the residue
in chloroform (15 cm3) was added dropwise to stirred petrol-
eum spirit (30–40 ЊC; 400 cm3) to give triethylammonium
3-N-benzoyl-2Ј,3Ј-di-O-benzoyluridine 5Ј-H-phosphonate 18
as a colourless solid (1.95 g, ≈90%, based on 3-N-benzoyl-
2Ј,3Ј-di-O-benzoyluridine 17); tR 8.02 min (programme 1);
δH [(CD3)2SO–D2O] 1.15 (6 H, t, J 7.3), 3.03 (4 H, q, J 7.3),
4.11 (2 H, m), 4.63 (1 H, m), 5.77 (2 H, m), 5.90 (0.5 H, s), 6.14
(1 H, d, J 8.2), 6.28 (1 H, d, J 5.5), 7.36 (2 H, t, J 7.8), 7.4–7.7
(6.5 H, m), 7.73 (3 H, m), 7.94 (4 H, m), 8.34 (1 H, d, J 8.3);
δP 2.6 (d, JP, H 602.8).
[(Us)pU] 3 as a colourless solid (387 A260 units); tR 7.70 min
(programme 2) (Fig. 1b); δH (D2O) 1.24 (9 H, m), 3.16 (6 H, m),
3.35 (1 H, dt, J 4.7 and 11.9), 3.90 (1 H, dd, J 3.1 and 13.6), 4.04
(1 H, d, J 13.8), 4.11 (2 H, m), 4.24 (3 H, m), 4.35 (1 H, dd, J 4.9
and 11.7), 4.42 (1 H, d, J 4.6), 5.68 (1 H, d, J 7.7), 5.70 (1 H, d,
J 8.1), 5.72 (1 H, d, J 8.3), 5.82 (1 H, d, J 2.1), 7.98 (1 H,
d, J 8.1), 8.02 (1 H, d, J 8.2); δP (D2O) 18.7 (Fig. 1a).
Triethylammonium salt of 3Ј-thiouridine 3Ј-phosphorothioate 20
3Ј-S-(2,4-Dinitrophenylsulfanyl)-3Ј-thiouridine 12 (0.046 g,
≈0.1 mmol), prepared as above by the action of 2,4-dinitro-
benzenesulfenyl chloride and TFA on 3Ј-S-(4-methoxybenzyl)-
3Ј-thiouridine 10, was dissolved in dry acetonitrile (5 cm3) and
the resulting solution was evaporated under reduced pressure.
Triethylamine (0.076 cm3, 0.55 mmol) and chlorotrimethyl-
silane (0.064 cm3, 0.5 mmol) were added to a stirred suspension
of the residue in dichloromethane (4 cm3) at room temperature.
After 15 min, tris(trimethylsilyl) phosphite (0.05 cm3, 0.15
mmol) was added. After a further period of 30 min, the
products were evaporated under reduced pressure. The residue
was dissolved in 0.2 mol dmϪ3 aq. triethylammonium hydrogen
carbonate (pH 7.5, 10 cm3) and the resulting solution was
extracted with dichloromethane (10 cm3). The aqueous layer
was concentrated to small volume under reduced pressure
and was then applied to a column (17 cm × 2 cm diameter) of
DEAE Sephadex A-25, which was eluted with a linear gradient
(0.0 to 0.50 mol dmϪ3 over 600 cm3) of triethylammonium
hydrogen carbonate buffer (pH 7.5). The appropriate fractions,
which were eluted with an average buffer concentration of 0.34
mol dmϪ3, were combined and lyophilised to give the triethyl-
ammonium salt of 3Ј-thiouridine 3Ј-phosphorothioate 20 as a
colourless solid [507 A260 units; tR 6.24 min (programme 2)]
(Fig. 2a); δH (D2O) includes the following signals: 3.55 (1 H, dt,
J 4.8 and 11.2), 3.99 (1 H, dd, J 3.2 and 13.5), 4.05 (1 H, dd,
J 2.1 and 13.4), 4.13 (1 H, m), 4.49 (1 H, d, J 4.8), 5.86 (1 H, s),
5.88 (1 H, d, J 8.2), 8.08 (1 H, d, J 8.1); δP (D2O) 16.4 (d, JP, H
11.2).
Triethylammonium salt of 3Ј-thiouridylyl-(3Ј→5Ј)-uridine
[(Us)pU] 3
3Ј-S-[(2,4-Dinitrophenylsulfanyl]-2Ј,5Ј-bis-O-[1-(2-fluoro-
phenyl)-4-methoxypiperidin-4-yl]-3Ј-thiouridine 14 (0.087 g,
≈0.1 mmol) and the triethylammonium salt of 3-N-benzoyl-
2Ј,3Ј-di-O-benzoyluridine 5Ј-H-phosphonate 18 (0.144 g, ≈0.2
mmol) were dissolved in dry acetonitrile (10 cm3) and the result-
ing solution was evaporated under reduced pressure. The resi-
due was then dissolved in dry dichloromethane (4 cm3), and dry
triethylamine (0.167 cm3, 1.2 mmol) and chlorotrimethylsilane
(0.13 cm3, 1.0 mmol) were added. The reactants were then
stirred at room temperature. After 17 h, the products were
poured into 0.2 mol dmϪ3 aq. triethylammonium hydrogen
carbonate (pH 7.5; 100 cm3) and the resulting mixture was
extracted with chloroform (3 × 100 cm3). The combined
organic extracts were dried (MgSO4), and evaporated under
reduced pressure. The residue, compound 19, was dissolved in 2
mol dmϪ3 methanolic ammonia (5 cm3) and the resulting solu-
tion was stirred at room temperature. After 17 h, the products
were evaporated under reduced pressure. The residue was
fractionated on a column of Merck silanised silica gel: the
appropriate fractions, which were eluted with acetonitrile–
water (15:85 v/v), were combined, and evaporated under
reduced pressure. The residue was then rechromatographed on
a column (17 cm × 2 cm diameter) of DEAE Sephadex A-25,
which was eluted with a linear gradient (0.0 to 0.50 mol dmϪ3
over 600 cm3) of triethylammonium hydrogen carbonate buffer
(pH 7.5). The appropriate fractions, which were eluted with an
average buffer concentration of 0.34 mol dmϪ3, were combined,
and concentrated under reduced pressure. The residue was
coevaporated with ethanol (2 × 10 cm3) under reduced pressure
and then dissolved in chloroform (2 cm3). When this solution
was added, with stirring, to petroleum spirit (30–40 ЊC, 100
cm3), a colourless solid precipitate (0.070 g) was obtained. This
material (0.035 g) was dissolved in acetic acid–water (2:98 v/v;
5 cm3) and the solution was stirred at room temperature. After
17 h, the products were evaporated under reduced pressure. The
residue was coevaporated with cyclohexane (3 × 5 cm3); it was
then fractionated on a column (17 cm × 2 cm diameter) of
DEAE Sephadex A-25, which was eluted as above. The
appropriate fractions, which were eluted with an average buffer
concentration of 0.22 mol dmϪ3, were combined, and evapor-
ated under reduced pressure. The residue was coevaporated
under reduced pressure with ethanol (2 × 5 cm3) to give
the triethylammonium salt of 3Ј-thiouridylyl-(3Ј→5Ј)-uridine
Triethylammonium salt of 3Ј-thiouridine 2Ј,3Ј-cyclic
phosphorothioate 21
Triethylammonium 3Ј-thiouridine 3Ј-phosphorothioate 20 (210
A260 units), tributylamine (0.12 cm3, 0.5 mmol), ethyl chloro-
formate (0.019 cm3, 0.2 mmol) and water (0.5 cm3) were vigor-
ously stirred together at room temperature. After 40 min,
methanol (0.5 cm3) was added and the products were applied to
a column (17 cm × 2 cm diameter) of DEAE Sephadex A-25.
The column was eluted with a linear gradient (0.0 to 0.50 mol
dmϪ3 over 600 cm3) of triethylammonium hydrogen carbonate
buffer (pH 7.5). Appropriate fractions, which were eluted with
an average buffer concentration of 0.22 mol dmϪ3, were com-
bined and lyophilised. The residual solid was dissolved in water
(2 cm3) and the solution was applied to a column (30 × 2 cm
diameter) of Merck silanised silica gel. The product-containing
fractions, which were eluted with water, were combined and
lyophilised to give the triethylammonium salt of 3Ј-thiouridine
2Ј,3Ј-cyclic phosphorothioate 21 as a colourless solid (160 A260
units); tR 4.69 min (programme 2) (Fig. 2b); δH (D2O) includes
the following signals: 3.76 (1 H, dd, J 4.2 and 13.0), 3.94 (1 H,
dd, J 2.2 and 12.9), 4.1–4.25 (2 H, m), 4.90 (1 H, m), 5.84 (1 H,
d, J 8.1), 5.98 (1 H, s), 7.79 (1 H, d, J 8.1); δP (D2O) 37.0 (d, JP, H
13.5).
Hydrolysis of 3Ј-thiouridylyl-(3Ј→5Ј)-uridine [(Us)pU] 3 in 0.05
mol dmϪ3 sodium glycinate buffer (pH 10.06) at 50 ЊC
Substrate [(Us)pU] 3 (1 A260 unit) was dissolved in sodium
glycinate buffer (0.15 cm3) and the resulting solution was heated
at 50 ЊC. After appropriate intervals of time, aliquots (10 mm3)
2234
J. Chem. Soc., Perkin Trans. 1, 2000, 2227–2236

View Online
of reaction solution were removed, and analysed by HPLC
(programme 2). A straight line was obtained by plotting ln
[%(Us)pU remaining] against time. The half-life (t1/2) of
hydrolysis was found to be 25 min. A typical HPLC profile (Fig.
3a) obtained after 45 min revealed unchanged (Us)pU 3 (tR 7.76
min, 25.3% of total absorbance), uridine 4 (tR 3.37 min, 37.0%
of total absorbance), 3Ј-thiouridine 2Ј,3Ј-cyclic phosphoro-
thioate 21 (tR 4.83 min, 14.0% of total absorbance) and 3Ј-
thiouridine 3Ј-phosphorothioate 20 (tR 6.27 min, 22.9% of total
absorbance). In a separate experiment, the substrate [(Us)pU]
(30 A260 units) was dissolved in the same sodium glycinate buf-
fer (3.0 cm3) and the resulting solution was heated at 50 ЊC.
After 30 min, acetic acid was added until the pH dropped to
4.0. The products were then evaporated under reduced pressure.
The 31P NMR spectrum of the residue (in D2O) (Fig. 4a)
revealed the following resonance signals: δ 37.2 (s, ≈20%,
assigned to 3Ј-thiouridine 2Ј,3Ј-cyclic phosphorothioate 21),
18.7 [s, ≈70%, assigned to unchanged (Us)pU 3] and 16.3 (s,
≈10%, assigned to 3Ј-thiouridine 3Ј-phosphorothioate 20). The
half-life of hydrolysis of uridylyl-(3Ј→5Ј)-uridine (UpU) 1a
in the same sodium glycinate buffer solution was found to be
80–90 h at 50 ЊC.
Hydrolysis of 3Ј-thiouridine 2Ј,3Ј-cyclic phosphorothioate 21 in
0.05 mol dmϪ3 sodium glycinate buffer
Substrate 21 (1 A260 unit) was dissolved in 0.05 mol dmϪ3
sodium glycinate buffer (pH 9.87; 0.15 cm3) and the resulting
solution was maintained at 30 ЊC. After appropriate intervals of
time, aliquots (10 mm3) of reaction solution were removed and
analysed by HPLC (programme 2). A straight line was obtained
by plotting ln (% substrate 21 remaining) against time. The half-
life (t1/2) of hydrolysis was found to be 164 min. A typical HPLC
profile, obtained after 131 min, revealed unchanged substrate
21 (tR 4.82 min, 54% of total absorbance) and 3Ј-thiouridine
3Ј-phosphorothioate 20 (tR 6.27 min, 46% of total absorbance).
In a separate experiment, the substrate (10 A260 units) was dis-
solved in a mixture of 0.05 mol dmϪ3 sodium glycinate buffer
(pH 10.6; 0.30 cm3) and D2O (0.30 cm3 at room temperature.
The resulting solution was maintained at 21 ЊC in an NMR
tube. After ≈7.4 h, the 31P NMR spectrum revealed two reson-
ance signals: δ 37.3 (s, ≈40%, assigned to 3Ј-thiouridine 2Ј,3Ј-
cyclic phosphorothioate 21) and 16.0 (s, ≈60%, assigned to
3Ј-thiouridine 3Ј- phosphorothioate 20).
Action of hydrolytic enzymes on 3Ј-thiouridylyl-(3Ј→5Ј)-uridine
[(Us)pU] 3
Action of glacial acetic acid on 3Ј-thiouridylyl-(3Ј→5Ј)-uridine
[(Us)pU] 3 at 30 ЊC
(a) Ribonuclease A. A solution of ribonuclease A (10 µg) in
0.1 mol dmϪ3 Tris hydrochloride buffer (pH 7.5; 10 mm3) was
added to a solution of (Us)pU 3 (1 A260 unit) in the same buffer
solution (0.1 cm3). The resulting solution was maintained at
30 ЊC. After 26 min, HPLC analysis (programme 2) revealed
substrate 3 (tR 7.6 min, 15.7% of total absorbance), 3Ј-thio-
uridine 3Ј-phosphorothioate 20 (tR 6.13 min, 28.3% of total
absorbance), 3Ј-thiouridine 2Ј,3Ј-cyclic phosphorothioate 21
(tR 4.60 min, 15.0% of total absorbance) and uridine 4 (tR 3.19
min, 41.0% of total absorbance). After 146 min, HPLC analysis
revealed substrate 3 (0.3% of total absorbance), 3Ј-thiouridine
3Ј-phosphorothioate 20 (50.4% of total absorbance) and urid-
ine (49.3% of total absorbance). Under the same reaction
conditions, 3Ј-thiouridine 2Ј,3Ј-cyclic phosphorothioate 21
underwent 30% digestion (to 3Ј-thiouridine 3Ј-phosphoro-
thioate 20) by ribonuclease A in 5 min.
(b) Crotalus adamanteus snake venom phosphodiesterase. A
stock solution of enzyme in 0.1 mol dmϪ3 Tris hydrochloride
buffer (pH 7.5; 10 mm3) was added to a solution of (Us)pU 3
(1 A260 unit) in 0.1 mol dmϪ3 Tris hydrochloride buffer (pH 7.5;
0.01 mol dmϪ3 with respect to magnesium chloride; 0.10 cm3).
The reaction solution was maintained at 30 ЊC. After 4 h,
HPLC analysis (programme 4) revealed that the substrate
accounted for only 9% of the total absorbance at 260 nm.
(c) Calf spleen phosphodiesterase. A stock solution of enzyme
in 0.1 mol dmϪ3 Tris hydrochloride buffer (pH 7.5; 10 mm3) was
added to a solution of (Us)pU 3 (1 A260 unit) in 0.1 mol dmϪ3
Tris hydrochloride buffer (pH 7.5; 0.002 mol dmϪ3 with respect
to EDTA and containing 0.05% Tween 80; 0.10 cm3). The solu-
tion was maintained at 30 ЊC. The substrate remained com-
pletely undigested after 24 h. Under the same conditions,
uridylyl-(3Ј→5Ј)-uridine (UpU) 1a was quantitatively digested
to uridine 3Ј-phosphate and uridine 4.
A freshly prepared solution of (Us)pU 3 (1 A260 unit) in glacial
acetic acid (0.15 cm3) was maintained at 30 ЊC. After appropri-
ate intervals of time, aliquots (10 mm3) of the solution were
removed, rapidly evaporated under reduced pressure (oil-
pump), re-dissolved in water (15 mm3) and analysed by HPLC
(programme 2). The half-life of decomposition of (Us)pU 3
was found to be ca. 4 min. A typical HPLC profile, after a
reaction time of 10 min (Fig. 3b), revealed uridine 4 (tR 3.3 min,
38.3% of total absorbance), 3Ј-thiouridine 2Ј,3Ј-cyclic phos-
phorothioate 21 (tR 4.71 min, 36.2% of total absorbance), 3Ј-
thiouridine 3Ј-phosphorothioate 20 (tR 6.19 min, 4.6% of
total absorbance), an unidentified product that might be 3Ј-
thiouridylyl-(2Ј→5Ј)-uridine 22 (tR 7.51 min, 2.3% of total
absorbance) and unchanged (Us)pU 3 (tR7.67 min, 18.6%
of total absorbance). In a separate experiment, the substrate
[(Us)pU] 3 (10 A260 units) was dissolved in glacial acetic acid
(0.50 cm3) and the resulting solution was maintained at 30 ЊC.
After 30 min, the products were evaporated under reduced
pressure (oil-pump). The 31P NMR spectrum (D2O) (Fig. 4b)
of the residue revealed the following resonance signals: δ 37.2 (s,
≈81%, assigned to 3Ј-thiouridine 2Ј,3Ј-cyclic phosphorothioate
21), 18.7 [s, ca. 11%, assigned to (Us)pU 3], 16.8 (s, ≈5%,
assigned to 3Ј-thiouridine 3Ј-phosphorothioate 20) and Ϫ0.44
[s, ≈3%, assigned tentatively to 3Ј-thiouridylyl-(2Ј→5Ј)-uridine
22].
Action of glacial acetic acid on uridylyl-(3Ј→5Ј)-uridine [UpU]
1a at 30 ЊC
The above experiment was repeated with UpU 1a (1 A260 unit).
Aliquots were removed and evaporated as above, and then
analysed by HPLC (programme 3). After 10 min, the absorb-
ance percentages relating to remaining substrate 1a (tR 9.27
min), uridylyl-(2Ј→5Ј)-uridine 23 (tR 7.71 min), uridine 2Ј,3Ј-
cyclic phosphate 24 (tR 3.53 min) and uridine 4 (tR 3.29 min)
were 87.8, 4.5, 3.7 and 4.0. After 37 min (Fig. 3c), the absorb-
ance percentages relating to the remaining substrate 1a,
uridylyl-(2Ј→5Ј)-uridine 23, uridine 2Ј(3Ј)-phosphates 25 and
26 (tR 4.99, 5.17), uridine 2Ј,3Ј-cyclic phosphate 24 and urid-
ine 4 were 63.9, 12.8, 1.5, 10.0 and 11.8. After 60 min, the
absorbance percentages relating to remaining substrate 1b,
uridylyl-(2Ј→5Ј)-uridine 23, uridine 2Ј(3Ј)-phosphates 25 and
26, uridine 2Ј,3Ј-cyclic phosphate 24 and uridine 4 were 49.8,
15.7, 4.6, 12.3 and 17.6.
Acknowledgements
One of us (X. L.) thanks the K. C. Wong Foundation for a
research scholarship, and the C.V.C.P. for an Overseas Research
Students Award.
References
1 Oligonucleotides: Antisense Inhibitors of Gene Expression, ed. J. S.
Cohen, Macmillan,London, 1989.
2 Nucleic Acids and Molecular Biology, Vol 10. Catalytic RNA,
ed. F. Eckstein and D. M. J. Lilley, Springer, Berlin, 1997.
J. Chem. Soc., Perkin Trans. 1, 2000, 2227–2236
2235

View Online
13 M. Faja, C. B. Reese, Q. Song and P.-Z. Zhang, J. Chem. Soc., Perkin
3 C. B. Reese, in Nucleic Acids and Molecular Biology, ed. F.
Eckstein and D. M. J. Lilley, Springer, Berlin, 1989, vol. 3,
pp. 164–181.
4 C. B. Reese, C. Simons and P.-Z. Zhang, J. Chem. Soc., Chem.
Commun., 1994, 1809.
5 X. Liu and C. B. Reese, Tetrahedron Lett., 1995, 36, 3413.
6 R. G. Kuimelis and L. W. McLaughlin, Nucleic Acids Res., 1995, 23,
4753.
7 J. B. Thomson, B. K. Patel, V. Jimberg, K. Eckart and F. Eckstein,
J. Org. Chem., 1996, 61, 6273.
Trans. 1, 1997, 191.
14 R. Lohrmann and H. G. Khorana, J. Am. Chem. Soc., 1964, 86,
4188.
15 X. Liu and C. B. Reese, J. Chem. Soc., Perkin Trans. 1, 1995, 1685.
16 J. S. Vyle, X. Li and R. Cosstick, Tetrahedron Lett., 1992, 33, 3017.
17 D. C. Capaldi and C. B. Reese, Nucleic Acids Res., 1994, 22, 2209.
18 R. Cosstick and J. S. Vyle, Nucleic Acids Res., 1990, 8, 829.
19 C. E. Müller and H. J. Roth, Tetrahedron Lett., 1990, 31, 501.
20 A. M. Michelson, J. Chem. Soc., 1959, 3655.
21 L. B. Weinstein, D. J. Earnshaw, R. Cosstick and T. R. Cech, J. Am.
Chem. Soc., 1996, 118, 10341.
8 X. Liu and C. B. Reese, Tetrahedron Lett., 1996, 37, 925.
9 R. Johnson, B. V. Joshi and C. B. Reese, J. Chem. Soc., Chem.
Commun., 1994, 133.
22 M. I. Elzagheid, M. Oivanen, K. D. Klika, B. C. N. M. Jones,
R. Cosstick and H. Lönnberg, Nucleosides, Nucleotides, 1999, 18,
2093.
10 O. Mitsunobu, Synthesis, 1981, 1.
11 N. C. Yung and J. J. Fox, J. Am. Chem. Soc., 1961, 83, 3060.
12 C. B. Reese and E. A. Thompson, J. Chem. Soc., Perkin Trans. 1,
1988, 2881.
23 B. E. Griffin, M. Jarman and C. B. Reese, Tetrahedron, 1968, 24,
639.
2236
J. Chem. Soc., Perkin Trans. 1, 2000, 2227–2236