Mutual isomerization of uridine 2′- and 3′-alkylphosphates and cleavage to a 2′,3′-cyclic phosphate: The effect of the alkyl group on the hydronium- and hydroxide-ion-catalyzed reactions

Article abstract of DOI:10.1039/a704636g

Isopropyl, ethyl, 2-ethoxyethyl, 2-chloroethyl, 2,2-dichloroethyl and 2,2,2-trichloroethyl esters of uridine 3′-phosphate have been prepared. In aqueous acid the compounds undergo concurrent isomerization to 2′-alkylphosphates and cleavage to uridine 2′,3′-cyclic phosphate, but in aqueous alkali only cleavage to the cyclic phosphate takes place. Buffer-independent rate constants for these reactions have been determined. The hydroxide-ion-catalyzed reaction to the 2′,3′-cyclic monophosphate is exceptionally susceptible to the polar nature of the leaving group, the β_1g value being -1.28 ± 0.05. By contrast, the hydronium-ion-catalyzed isomerization and cleavage are both rather insensitive to the electron-withdrawing ability of the alkyl group, the β and β_1g values being -0.18 ± 0.02 and -0.12 ± 0.05, respectively. The transition state structures of the reactions are discussed on the basis of these structural effects.

Full text of DOI:10.1039/a704636g

View Article Online / Journal Homepage / Table of Contents for this issue
Mutual isomerization of uridine 2Ј- and 3Ј-alkylphosphates and
cleavage to a 2Ј,3Ј-cyclic phosphate: the effect of the alkyl group on
the hydronium- and hydroxide-ion-catalyzed reactions
Markus Kosonen,a Esmail Youseti-Salakdeh,b Roger Strömberg*,b,c and
Harri Lönnberg*,a
a Department of Chemistry, University of Turku, FIN-20014 Turku, Finland
b Department of Organic Chemistry, Arrhenius Laboratory, Stockholm University,
S-10691 Stockholm, Sweden
c Division of Organic and Bioorganic Chemistry, MBB, Scheele Laboratory,
Karolinska Institutet, S-17177 Stockholm, Sweden
Isopropyl, ethyl, 2-ethoxyethyl, 2-chloroethyl, 2,2-dichloroethyl and 2,2,2-trichloroethyl esters of uridine
3Ј-phosphate have been prepared. In aqueous acid the compounds undergo concurrent isomerization to
2Ј-alkylphosphates and cleavage to uridine 2Ј,3Ј-cyclic phosphate, but in aqueous alkali only cleavage to
the cyclic phosphate takes place. Buffer-independent rate constants for these reactions have been
determined. The hydroxide-ion-catalyzed reaction to the 2Ј,3Ј-cyclic monophosphate is exceptionally
susceptible to the polar nature of the leaving group, the â1g value being Ϫ1.28 0.05. By contrast,
the hydronium-ion-catalyzed isomerization and cleavage are both rather insensitive to the electron-
withdrawing ability of the alkyl group, the â and â1g values being Ϫ0.18 0.02 and Ϫ0.12 0.05,
respectively. The transition state structures of the reactions are discussed on the basis of these
structural effects.
HO
B
HO
B
Introduction
O
O
The mechanistic details of the hydrolysis of the phosphodiester
bonds of RNA have recently received considerable interest for
several reasons: to get a better insight into the action of ribo-
nucleases, above all RNaseA, to understand the action of
catalytic ribonucleic acids and to provide a rational basis for
the development of artificial nucleases. It is well known that
dinucleoside 3Ј,5Ј-monophosphates (1), the dimeric fragments
of RNA, undergo in aqueous solution two competing intra-
molecular transesterification reactions, viz. isomerization to
a 2Ј,5Ј-phosphodiester (2) and cleavage to a cyclic 2Ј,3Ј-
O
P
OH
OH
O
P
O
O
–O
–O
O
B
O
B
O
O
OH
OH
OH
OH
1
2
phosphodiester (3) with concomitant release of the 5Ј-linked
HO
B
1–8
O
nucleoside (Scheme 1).
Studies with homopolymers have
shown that the situation remains similar for polyribo-
nucleotides.9,10 Under acidic conditions, both reactions proceed
in all likelihood via a common pentacoordinated intermediate
obtained by an attack of the 2Ј-hydroxy function on the
protonated monocationic phosphate group.4,5 This intermedi-
ate, when deprotonated to a neutral species, is sufficiently stable
to be able to pseudorotate.11 Accordingly, any of the bridging
phosphate oxygens, viz. O2Ј, O3Ј and O5Ј, may take an apical
position, and hence leave after protonation as an alcohol. The
breakdown of the intermediate is symmetrical, since form-
ation of the 2Ј,3Ј-cyclic monophosphate and mutual isomeriz-
ation of the 2Ј,5Ј- and 3Ј,5Ј-monophosphates occur at
approximately the same time.4,5 In contrast, formation of the
2Ј,3Ј-cyclic monophosphate is the only reaction detected under
alkaline conditions.4,5,12,13 The deprotonated 2Ј-oxyanion now
attacks at the monoanionic phosphate, and the resulting diani-
onic pentacoordinated species appears to be too unstable to be
able to pseudorotate (Scheme 2). Consistent with this kind of
in-line displacement, the Rp and Sp diastereomers of dinucleo-
side 3Ј,5Ј-phosphoromonothioates have been observed to
undergo hydroxide-ion-catalyzed cyclization to 2Ј,3Ј-cyclic
phosphoromonothioates by complete inversion of configur-
ation at phosphorus.14 The lifetime of the dianionic phos-
phorane is still being debated. Perreault and Anslyn15 have
recently reviewed the relevant literature and concluded that a
O
O
P
HO
O
B
O–
O
OH
OH
HO
B
O
O
O
P
O
O–
3
Scheme 1
dianionic phosphorane may possibly exist but it cannot be
kinetically distinguished from a transition state. In particular, it
cannot undergo protonation to a monoanionic phosphorane
which is known to pseudorotate. The ab initio calculations of
Taira and co-workers, however, suggest that dianionic phos-
J. Chem. Soc., Perkin Trans. 2, 1997
2661

View Article Online
HO
B
MMTrO
U
MMTrO
U
O
HO
B
O
O
O
i
ii
-H+
O
P
O–
O
P
OMthp
+HNEt3
O
P
OMthp
+HNEt3
O
P
OH
–O
O
–O
O
–O
H
–O
RO
B
B
O
O
O
O
O
O
5
6b–f
OH
OH
OH
OH
iii
MMTr = 4-methoxytrityl
Mthp = 4-methoxytetrahydropyran-4-yl
1
4b–f
Scheme 3 (i) bis(2-oxo-oxazolidin-3-yl)phosphoric chloride and ROH
in Py; (ii) I2–H2O–Py; (iii) HCl–H2O–THF
HO
B
O
HO
B
and the products were purified by RPHPLC. The syntheses
were not optimized, but they were considered to be adequate
for the purpose of obtaining pure material for kinetic investig-
ations. 5 was prepared from 3Ј,5Ј-O-tetraisopropyldisiloxane-
diyluridine18 via appropriate acetalization of the 2Ј-hydroxy
function,19 removal of the disiloxanediyl group,18 5Ј-O-mono-
methoxytritylation and 3Ј-phosphonylation.20,21 The isopropyl
ester of uridine 3Ј-phosphate (4a) was obtained by phosphotri-
ester approach.22 Accordingly, 5Ј-O-(4-methoxytrityl)-2Ј-O-
(tetrahydropyran-2-yl)uridine was converted to its 2-chloro-
phenyl isopropyl triester, hydrolyzed to 3Ј-isopropylphosphate
in aqueous alkali and deprotected with aqueous acid.
O
O
Oδ–
Oδ–
– Nuc-5'-O-
P
O
O
δ–O
Oδ–
P
–O
O
B
O
3
OH
OH
Scheme 2
phorane is sufficiently stable in aqueous solution to be regarded
as an intermediate.16,17
To learn more about the transition state structures of the
reactions indicated above, especially that proceeding via a
dianionic phosphorane, we now report the results of kinetic
studies carried out on a number of monoalkyl esters of uridine
3Ј-phosphate (4a–f). The pKa values of the esterified alcohols
Kinetic measurements
The concurrent intramolecular transesterifications of the alkyl
esters of uridine 3Ј-phosphate (4a–f) to uridine 2Ј,3Ј-cyclic
phosphate (reaction B in Scheme 4) and uridine 2Ј-alkyl-
O
O
NH
NH
A
HO
N
O
O
HO
N
O
k1
O
2′-isomer
k–1
O
P
OH
O
P
OH
–O
RO
O
–O
RO
k2
k2
O
B
4a R = CH(CH3)2
b
c
d
e
f
R = CH2CH3
R = CH2CH2OCH2CH3
R = CH2CH2Cl
R = CH2CHCl2
R = CH2CCl3
a
b
c
d
e
f
4
R = CH(CH3)2
R = CH2CH3
R = CH2CH2OCH2CH3
R = CH2CH2Cl
R = CH2CHCl2
R = CH2CCl3
2′,3′-cyclic phosphate
cover a range of 5 units (12–17). These studies are also relevant
because simple alkyl and aryl esters of nucleoside 3Ј-
phosphates, as well as their non-nucleosidic analogues, are
often used as mimics of dinucleoside 3Ј,5Ј-monophosphates.
Quantitative knowledge of the effects that this kind of struc-
tural simplification may exert on the rate of various transesteri-
fication reactions helps in assessing the general applicability
of the mechanistic conclusions drawn on the basis of the
behaviour of model compounds.
Scheme 4
phosphates (reaction A in Scheme 4) were followed in 0.5 mol
dmϪ3 aqueous hydrogen chloride at 298.2 K by determining the
composition of the aliquots withdrawn at appropriate intervals
by RPHPLC. Uridine 2Ј,3Ј-cyclic phosphate is not actually
accumulated under these conditions, but it is hydrolyzed to a
1:2 mixture of uridine 2Ј- and 3Ј-phosphates, and subsequent
dephosphorylation to uridine is several orders of magnitude
slower than their formation.23–27 Accordingly, in addition to the
intermediate formation of uridine 2Ј-alkylphosphate, accumu-
lation of uridine 2Ј/3Ј-phosphates as final products was
observed in each case. Table 1 records the pseudo-first-order
rate constants obtained for reactions A and B. The dependence
of the logarithmic values of these rate constants on the pKa
value of the esterified alcohol is shown in Fig. 1. The values of
the reaction constants are Ϫ0.12 0.05 and Ϫ0.18 0.02 for
the cleavage (reaction B) and isomerization (reaction A),
respectively.
Results
Preparation of the model compounds
Uridine 3Ј-isopropylphosphate (4a) apart, the alkyl esters of
uridine 3Ј-phosphate (4b–f) were obtained by coupling 5Ј-O-
(4-methoxytrityl)-2Ј-(4-methoxytetrahydropyran-4-yl)uridine
3Ј-H-phosphonate (5) with appropriate alcohols, and oxidizing
the resulting H-phosphonate diester in situ with iodine in aque-
ous pyridine (Scheme 3). The sugar moiety protecting groups
were removed by acid-catalyzed hydrolysis in aqueous THF,
2662
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
Table 1 First-order rate constants for the interconversion of uridine
2Ј- and 3Ј-alkylphosphates (k1 ϩ kϪ1) and their transesterification to
uridine 2Ј,3Ј-cyclic monophosphate (k2) in aqueous hydrogen chloride
(0.5 mol dmϪ3) at 298.2 Ka
Table 2 Second-order rate constants for the transesterification of
uridine 3Ј-alkylphosphates to uridine 2Ј,3Ј-cyclic monophosphate in
aqueous alkali at 298.2 K. The ionic strength was adjusted to 1.0 mol
dmϪ3 with sodium chloride.
Compound
R
k2/10Ϫ5 sϪ1
(k1 ϩ kϪ1)/10Ϫ5 sϪ1
Compound
R
k2/10Ϫ5 dm3 molϪ1 sϪ1a
4a
4b
4c
4d
4e
4f
CH(CH3)2
CH2CH3
CH2CH2OEt
CH2CH2Cl
CH2CHCl2
CH2CCl3
0.636 0.009
2.00 0.05
2.93 0.06
3.12 0.02
4.05 0.19
2.61 0.10
0.47 0.03
0.72 0.02
1.02 0.01
1.05 0.08
2.23 0.05
3.87 0.21
4a
4b
4c
4d
4e
4f
CH(CH3)2
CH2CH3
CH2CH2OEt
CH2CH2Cl
CH2CHCl2
CH2CCl3
0.053 0.001b
4.6 0.8c
41.1 0.5c
459 6c
20 300 200d
114 000 3000e
a k1/kϪ1 ranged from 0.75 to 0.88 at pH 4.2 and 363.2 K.
a The pseudo-first-order rate constant divided by the concentration of
hydroxide ion. b The pseudo-first-order rate constant was obtained in
1.00 mol dmϪ3 aqueous sodium hydroxide. c The pseudo-first-order rate
constant was obtained in 0.10 mol dmϪ3 aqueous sodium hydroxide.
d The pseudo-first-order rate constant was obtained in 0.010 mol dmϪ3
aqueous sodium hydroxide. e The pseudo-first-order rate constants were
determined in aqueous ammonia–ammonium chloride buffers (buffer
ratio 1:1). The values obtained at the total concentrations of 0.1, 0.2,
0.3 and 0.4 mol dmϪ3 were 7.89, 8.66, 9.29 and 10.4 × 10Ϫ5 sϪ1, respect-
ively. The value extrapolated to buffer concentration zero was used to
calculate the second-order rate constant.
Fig. 1 Logarithmic first-order rate constants for the hydronium-ion-
catalyzed transesterification of uridine 3Ј-alkylphosphates to uridine
2Ј-alkylphosphates (᭺) and uridine 2Ј,3Ј-cyclic phosphate (᭹) plotted
against the pKa values of the esterified alcohols.35 The data refer to
[HCl] = 0.5 mol dmϪ3 at T = 298.2 K.
No isomerization to 2Ј-alkylphosphates could be observed in
aqueous sodium hydroxide. Uridine 3Ј-(2,2,2-trichloroethyl)-
phosphate (4f) hydrolyzed so readily that the rate constant
could not be determined in aqueous sodium hydroxide. With
this compound, the measurements were carried out at a lower
pH in an ammonia–ammonium chloride buffer (pH 9.54). The
rate constant slightly increased with the increasing buffer con-
centration, in spite of the fact that the ionic strength was kept
constant. Whether this modest rate-acceleration results from
medium effects or from buffer catalysis cannot be confidently
concluded. The buffer-independent second-order rate constant
was calculated from the first-order rate constant obtained by
extrapolation to zero buffer concentration. Table 2 summarizes
the kinetic data under alkaline conditions. In Fig. 2 the
logarithmic values of the second-order rate constants are
plotted against the pKa values of the alcohols corresponding
to the displaced alkoxide ions. The β1g value obtained in this
manner is Ϫ1.28 0.05.
Fig. 2 Logarithimic second-order rate constants for the hydroxide-
ion-catalyzed transesterification of uridine 3Ј-alkylphosphates to urid-
ine 2Ј,3Ј-cyclic phosphate plotted against the pKa values of the esteri-
fied alcohols.35 The data refer to T = 298.2 K and I = 1.0 mol dmϪ3
.
of the alkoxy group being displaced (OR in Scheme 4); the β1g
value is Ϫ1.28 0.05. The electronegativity of the alkyl group
probably does not markedly affect the pre-equilibrium depro-
tonation of the 2Ј-hydroxy function, since the inductive effect
of the alkoxy group on the 2Ј-oxygen is transmitted via five
σ-bonds. The highly negative β1g suggests that the alkoxy ion
character of the leaving-group is well developed in the transi-
tion state. Comparable highly negative β1g values have also been
reported for other related intramolecular displacement reac-
tions of phosphodiesters: Ϫ0.9 for the cleavage of substituted
guanosine 3Ј-benzylphosphates,29 Ϫ1.13 for the cyclization of
the dianion of alkyl α,α-bis(trifluoromethyl)-2-hydroxybenzyl
phosphates,30 and Ϫ1.26 for the cyclization of the dianion of
aryl 2-carboxyphenyl phosphate.31 Actually, the β1g value of
Ϫ1.28 is sufficiently negative to be taken as an indication of a
clearly stepwise mechanism, i.e. pre-equilibrium formation of a
dianionic phosphorane followed by its unimolecular heterolysis
to an alkoxide ion and 2Ј,3Ј-cyclic phosphate. The fact that no
hydroxide-ion-catalyzed isomerization has ever been detected,
however, argues against this alternative. Evidently the dianionic
phosphorane is too short-lived to pseudorotate as a dianion or
via a kinetically invisible intermediary protonation to the
monoanionic phosphorane. Note that protonation to monoan-
ionic phosphorane is thermodynamically favoured under mildly
alkaline conditions (pKa is estimated to range from 13 to 1515),
and it hence should be a diffusion controlled process. Perhaps
the most conclusive evidence against the hydroxide-ion-
catalyzed isomerization has been obtained with the phos-
Discussion
As discussed in the Introduction, the hydroxide-ion-catalyzed
transesterification of ribonucleoside 3Ј-phosphodiesters to
ribonucleoside 2Ј,3Ј-cyclic monophosphates is generally
assumed to proceed by a simple ‘in line’ mechanism: the attack-
ing 2Ј-oxyanion and the departing 5Ј-oxyanion occupy the
apical positions within a phosphorane-type structure (cf.
Scheme 2 for NpN, 1). The β1g value for the cleavage of the
phenyl esters of uridine 3Ј-phosphate has been observed to be
Ϫ0.54, suggesting that both the P᎐O2Ј bond formation and
P᎐O5Ј bond cleavage are only weakly advanced in the transition
state.28 In other words, the reaction appears to be a concerted
process proceeding via a pentacoordinated transition state. The
situation is, however, different with the alkyl esters. As can be
seen from Fig. 2, this reaction is quite susceptible to the basicity
J. Chem. Soc., Perkin Trans. 2, 1997
2663

View Article Online
HO
O
O
OH
+
HO
+
HO
P
OH
P
OH
OR
OR
+H+
-H+
-H+
+
HO
+
OH
a
O
O
O
e
a
e
O
O
O
P
P
+H+
P
P
HO
OH
HO
OH
HO
OH
HO
OH
OR
OR
OR
OR
-H+
+H+
+H+
-ROH
-H+
O
O
O
O
P
P
+
OH
HO
OH
HO
+HOR
a
Scheme 5
phonate analog of GpU, having the O5Ј of the 5Ј-linked
nucleoside replaced with a methylene group.32 Since the P᎐C
bond cannot be cleaved, isomerization is the only reaction
taking place under alkaline conditions. The isomerization,
however, remains strictly pH-independent even at pH 10. At
higher pH the uracil base starts to degrade. Under these
conditions dinucleoside 3Ј,5Ј-monophosphates undergo rapid
hydroxide-ion-catalyzed cleavage, indicating that the 2Ј-oxy-
anion attacks the phosphate diester monoanion. It is probable
that the 2Ј-oxyanion also attacks the phosphonate ester
monoanion, but this attack does not lead to a detectable reac-
tion. The dianionic intermediate simply reforms the starting
material. Protonation to a monoanion would undoubtedly
lead to isomerization, since the pH-independent phosphate
migration via a monoanionic phosphorane is even faster with
the phosphonate analog than with an unmodified dinucleo-
side 3Ј,5Ј-monophosphate. Accordingly, the cleavage of ribo-
nucleoside 3Ј-alkylphosphates appears to proceed via a very late
transition state but without formation of a fully developed
dianionic phosphorane intermediate that could be stabilized by
protonation. The mechanism most likely lies on the borderline
between a concerted and a stepwise mechanism.
on the basis of the data available, but evidently they cancel each
other almost entirely. A similar observation has been made in
the hydronium-ion-catalyzed cyclization of the neutral ionic
form of alkyl α,α-bis(trifluoromethyl)-2-hydroxybenzyl phos-
phates: the β1g value is only Ϫ0.15.30
The isomerization (reaction A) also is rather insensitive to
the electronegativity of the alkyl group (β = Ϫ0.18 0.02).
Tentatively one may assume that an electron-withdrawing alkyl
group facilitates the nucleophilic attack on phosphorus, and
hence increases the equilibrium concentration of the phos-
phorane intermediate. The increasing electronegativity of the
alkyl group also seems to favour isomerization over cleavage,
although the effect is small. Since both reactions proceed via
a common intermediate, this effect has to be attributed to
pseudorotation/breakdown of this intermediate. In other
words, the increasing electronegativity of the alkyl group
favours the cleavage of the P᎐O3Ј bond over cleavage of the
P᎐OR bond. The origin of this influence is difficult to define.
The rate-limiting step of both reactions is assumed to be a uni-
molecular rupture of the P᎐OHϩ bond. The increasing elec-
tronegativity of the alkyl group retards the pre-equilibrium pro-
tonation of the P᎐OR bond and accelerates the departure of
the protonated alkoxy ligand. In contrast, the protonation and
subsequent cleavage of the P᎐O3Ј bond may be expected to be
rather insensitive to electron-withdrawal by the alkyl group.
Perhaps the effect on the basicity of the alkoxy oxygen is so
dominant that the P᎐OR bond cleavage becomes retarded
compared to that of P᎐O2Ј.
The studies of Dalby et al.30 on cyclization of methyl α,α-
bis(trifluoromethyl)-2-hydroxybenzyl phosphate by the dis-
placement of methoxide ion with the neighbouring phenoxide
ion suggest that the departure of the methoxide ion is suscep-
tible to general acid catalysis by ammonium ions. The modest
rate-accelerating effect of NH3–NH4ϩ buffers on the cleavage of
uridine 3Ј-(2,2,2-trichloroethyl)phosphate may tentatively be
accounted for in a similar way.
Experimental
The hydronium-ion-catalyzed cleavage and isomerization
(Scheme 5) proceed via a pseudorotating pentacoordinated
intermediate obtained by the attack of 2Ј-OH on the mono-
cationic phosphodiester bond. According to ab initio calcul-
ations, the pseudorotation barrier of a neutral phosphorane is
sufficiently low to allow rapid equilibration between the forms
having either O2Ј or O3Ј apical. The leaving group is now an
alcohol, not an alkoxide ion.4,5 The rate of cleavage (reaction B)
is, as expected, rather insensitive to the polar nature of the
departing alcohol (β1g = Ϫ0.12 0.05). Electron-withdrawal by
a polar substituent, for example, weakens the P᎐OR bond by
increasing its polarity and facilitates the attack of the 2Ј-OH on
phosphorus, but simultaneously the pre-equilibrium proton-
ation of the starting material and the protonation of the leaving
oxygen are retarded. These partial effects cannot be quantified
General
Pyridine, acetonitrile, ethanol, 2-chloroethanol, 2,2-dichloro-
ethanol and 2,2,2-trichloroethanol were dried by storage over
molecular sieves. 2-Ethoxyethanol and phosphorus trichloride
were distilled prior to use. Triethylamine and THF were dis-
tilled over calcium hydride and lithium aluminium hydride,
respectively. Uridine, 4-methoxy-5,6-dihydro-2H-pyran, tetra-
butylammonium fluoride, 4-methoxytrityl chloride, bis(2-oxo-
oxazolidin-3-yl)phosphonic chloride, 1,3-dichloro-1,1,3,3-
1
tetraisopropyldisiloxane and iodine were used as received. H
NMR spectra were recorded at 270 MHz, and all signals were
assigned by H–1H homocoupling experiments to verify that
1
phosphoric acid was esterified to the 3Ј-hydroxy function of
uridine.
2664
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
Table 3 1H and 31P NMR chemical shifts for uridine 3Ј-alkylphosphates (4a–f) in 2H2O at 25 ЊCa
δ
Compound
H1Ј
H2Ј
H3Ј
H4Ј
H5Ј,H5Љ
H5
H6
OR
P
4a
4b
4c
4d
4e
4f
5.79
5.88
5.84
5.88
5.90
5.96
4.28
4.38
4.34
4.38
4.46
4.49
4.39
4.58
4.49
4.52
4.62
4.68
4.15
4.21
4.18
4.29
4.32
4.37
3.74; 3.65
3.72; 3.80
3.70; 3.77
3.74; 3.84
3.82; 3.90
3.85; 3.94
5.73
5.85
5.74
5.82
5.88
5.91
7.71
7.88
7.74
7.81
7.86
7.89
1.10; 4.28
3.88
1.16; 3.47; 3.59; 3.93 Ϫ0.96
3.79; 4.08
4.25; 6.08
4.56
ϩ3.00
Ϫ0.85
Ϫ1.04
Ϫ0.50
Ϫ1.36
a The 1H and 31P chemical shifts are given in ppm from external TMS and phosphoric acid, respectively.
5Ј-O-(4-Methoxytrityl)-2Ј-O-(4-methoxytetrahydropyran-4-yl)-
uridine 3Ј-H-phosphonate triethylammonium salt (5)
groups. With 6d, a more concentrated acid solution was used
([HCl] = 0.038 mol dmϪ3, THF–water 6:4). The progress of
deprotection was followed by TLC (propan-2-ol–conc. aq.
ammonia–water 8:1:1). After complete deprotection, the pH
was adjusted to 6 with aqueous sodium hydroxide, and the solu-
tions were washed with diethyl ether. The products (4b–f) were
purified by RPHPLC (Lichrosorb RP-18, 25 × 250 mm, 7 µm)
using a linear gradient from 20 to 100% of buffer B in buffer A
over30min[bufferA:25mmoldmϪ3 aqueoustriethylammonium
acetate (pH 6.5); buffer B: 25 mmol dmϪ3 triethylammonium
acetate in 50% aqueous acetonitrile]. The retention times were:
4b, 18 min; 4c, 17 min; 4d, 13 min; 4e, 15 min; 4f, 17 min. The
1H and 31P NMR chemical shifts are listed in Table 3, and the
RPHPLC retention times and electrospray mass spectroscopy
(ESMS) signals in Table 4.
Uridine was converted to 3Ј,5Ј-O-(1,1,3,3-tetraisopropyldi-
siloxane-1,3-diyl)uridine,18 and this to 2Ј-O-(4-methoxytetra-
hydropyran-4-yl)-3Ј,5Ј-O-(1,1,3,3-tetraisopropyldisiloxane-1,3-
diyl)uridine,19 as described previously. The silyl protection was
removed with tetrabutylammonium fluoride in acetonitrile,18
and the product was purified by silica gel chromatography,
using a stepwise gradient of methanol (1–10%, v/v) in chloro-
form containing 1% triethylamine. 2Ј-O-(4-Methoxytetra-
hydropyran-4-yl)uridine obtained (2.44 g, 7.21 mmol) was
treated with 4-methoxytrityl chloride (3.34 g, 10.8 mmol) in dry
pyridine (69 ml). Conventional sodium bicarbonate workup
and subsequent purification on silica gel, using a mixture of
chloroform and methanol (19:1) containing 1% triethylamine
as an eluent, gave 2Ј-O-(4-methoxytetrahydropyran-4-yl)-5Ј-O-
(4-methoxytrityl)uridine in 58% yield. This compound (0.88 g,
1.44 mmol) was dried by coevaporations from pyridine and
acetonitrile, and dissolved in acetonitrile (36 ml). The solution
was added dropwise (over a period of 1 h) to a mixture of
imidazole (1.05 g, 15.5 mmol), phosphorus trichloride (5.04 g,
3.65 mmol) and triethylamine (2.25 ml, 16.2 mmol) in
acetonitrile (1.8 ml) in an ice bath.20,21 After an additional 15
min mixing at room temperature, aqueous triethylammonium
bicarbonate (5 ml, 0.1 mol dmϪ3) was added, and the mixture
was evaporated to dryness. The residue was coevaporated from
a mixture of pyridine and triethylamine (4:1), and then from
toluene. The crude product was equilibrated between chloro-
form and aqueous triethylammonium bicarbonate (0.1 mol
dmϪ3). The aqueous phase was extracted twice with chloro-
form, and the combined organic phases were evaporated to
dryness. Silica gel chromatography using a stepwise gradient of
methanol (1–10%) in chloroform containing 1% triethylamine
gave 5 in 89% yield.
5Ј-O-(2-Methoxytrityl)-4Ј-O-(tetrahydropyran-2-yl)uridine33
was converted to its 3Ј-(2-chlorophenylisopropyl)phosphate by
the method of Marugg et al.22 2-(Chlorophenoxy)phosphoryl-
bis(1,2,4-triazole) was, however, used as a phosphorylating
agent instead of 2-chlorophenyl-O,O-bis(benzotriazol-1-yl)-
phosphate. The phosphotriester obtained (0.26 g, 0.31 mmol)
was treated with a 1:4 mixture (100 ml) of aqueous sodium
hydroxide (0.1 mol dmϪ3) and dioxane overnight at room tem-
perature. Aqueous hydrogen chloride (2 ml, 2 mol dmϪ3) was
added, and the mixture was kept overnight at room temper-
ature. The solution was neutralized with triethylamine and
evaporated to dryness. The residue was dissolved in water (20
ml), washed twice with an equal volume of dichloromethane,
and the aqueous layer was evaporated to dryness. The product
(4a) was purified by RPHPLC on a LiChrospher 100 RP-18
column using a 93:7 (v/v) mixture of aqueous ammonium
bicarbonate (0.05 mol dmϪ3) and methanol as an eluent. The
yield was 45%. The 1H and 31P NMR chemical shifts are listed
in Table 3, and the RPHPLC retention times and ESMS signals
in Table 4.
5Ј-O-(4-Methoxytrityl)-2Ј-O-(4-methoxytetrahydropyran-4-yl)-
uridine 3Ј-monoalkylphosphate triethylammonium salts (6b–f)
5 (0.20 g, 0.25 mmol) was dissolved in dry pyridine (10 ml) and
the appropriate alcohol (0.28 mmol) was added. To the stirred
reaction mixture, bis(2-oxo-oxazolidin-3-yl)phosphonic chlor-
ide (0.13 g, 0.50 mmol) was then added. After the reaction was
completed, iodine in a mixture of pyridine and water (4% w/v,
96:4 v/v, 10 ml) was added. The oxidation of H-phosphonate
diester to phosphate diester took place in 5 min. The reaction
mixture was diluted with dichloromethane and evaporated to
near dryness. The evaporation was repeated several times from
the same solvent. The residue was dissolved in dichloromethane
and washed with aqueous sodium thiosulfate (10%, w/v) and
then with saturated aqueous sodium bicarbonate. The crude
product was purified by silica gel chromatography using a step-
wise gradient of methanol (1–10%) in chloroform containing
1% triethylamine. The yields of 6b–f varied from 70 to 95%.
Kinetic measurements
The reactions were carried out in stoppered bottles immersed in
a water bath, the temperature of which was adjusted to 298.2 K
within 0.1 K. The initial concentration of the nucleotide
esters was approximately 2 × 10Ϫ4 mol dmϪ3. The progress of the
reactions was followed by withdrawing 100 µl aliquots from the
reaction mixture at appropriate intervals. The reactions were
stopped by adjusting the pH to 4.5, and the composition of the
aliquots was analyzed by RPHPLC on a Hypersil ODS column
(250 × 4 mm, 5 µm) column. Isocratic elution with a mixture of
acetic acid–sodium acetate buffer (0.045–0.015 mol dmϪ3, pH
4.2, [NH4Cl] = 0.1 mol dmϪ3) and acetonitrile was applied
throughout. The acetonitrile content varied from 3.5 to 10%
(v/v), depending on the hydrophobicity of the esterfied alcohol.
Uridine 2Ј- and 3Ј-phosphates formed as hydrolysis products
were identified by spiking with authentic samples. Isomeriz-
ation of nucleoside 3Ј-alkylphosphates to their 2Ј-counterparts
under acidic conditions has been established previously.34 Table
4 records the acetonitrile contents and the retention times for
the starting materials and their 2Ј-isomers.
Uridine 3Ј-monoalkylphosphate triethylammonium salts (4a–f)
The protected phosphodiesters (6b,c,e,f) were dissolved in a
mixture of THF and water (9:1, 10 ml) containing hydrogen
chloride (0.01 mol dmϪ3) to remove the acid labile protecting
The rate constants for the transesterification of 4a–f to urid-
J. Chem. Soc., Perkin Trans. 2, 1997
2665

View Article Online
Table 4 Retention times of uridine 3Ј-alkylphosphates (4a–f) and their 2Ј-isomers on a Hypersil ODS column,a and the ESMS signals of the
3Ј-alkylphosphates
tR/min
Alkyl group
3Ј-Isomer
2Ј-Isomer
Eluentb
m/zc
CH(CH3)2
CH2CH3
CH2CH2OEt
CH2CH2Cl
CH2CHCl2
CH2CCl3
6.9
7.4
21.0
14.5
23.8
16.5
4.1
5.1
11.1
7.5
10.2
7.5
5.0
3.5
3.5
3.5
5.0
365.4 (100%), 253.6 (20%)
351.2 (100%), 239.3 (20%)
395.3 (100%), 283.4 (19%)
387.2 (33%), 385.3 (100%)
421.1 (64%), 419.1 (100%), 347.3 (12%)
457.0 (28%), 455.0 (93%), 453.1 (100%)
10.0
a Column 4 × 250 mm, particle size 5 µm, flow rate 1 ml minϪ1
.
b Acetonitrile content (%, v/v) in acetic acid–sodium acetate buffer (0.045/0.015
mol dmϪ3), containing ammonium chloride (0.1 mol dmϪ3). c Signals exhibiting the relative abundance >10% of that of the [M᎐H]Ϫ indicated.
17 A. Yliniemelä, T. Uchimaru, K. Tanabe and K. Taira, J. Am. Chem.
Soc., 1993, 115, 3032.
18 W. J. Markiewicz, J. Chem. Res. (M), 1979, 181.
19 C. B. Reese, R. Saffhill and J. E. Sulston, Tetrahedron, 1970, 26,
1023.
20 P. J. Garegg, T. Regberg, J. Stawinski and R. Strömberg, Chem. Scr.,
1986, 26, 59.
21 J. Stawinski, R. Strömberg, M. Thelin and E. Westman, Nucleic
Acids Res., 1988, 16, 9285.
22 J. E. Marugg, M. Tromp, P. Jhurani, C. F. Hoyng, G. A. van der
Marel and J. H. van Boom, Tetrahedron, 1984, 40, 73.
23 H. I. Abrash, C.-C. Cheung and J. C. Davis, Biochemistry, 1967, 6,
1298.
24 M. R. Eftink and R. L. Biltonen, Biochemistry, 1983, 22, 5134.
25 M. Oivanen and H, Lönnberg, J. Org. Chem., 1989, 54, 2556.
26 S. N. Mikhailov, M. Oivanen, P. Oksman and H. Lönnberg, J. Org.
Chem., 1992, 57, 4122.
ine 2Ј,3Ј-cyclic monophosphate (k2; reaction B in Scheme 4)
and the interconversion of 4a–f and their 2Ј-isomers (k1 ϩ kϪ1
;
reaction A) were calculated by rate laws of irreversible and
reversible first-order reactions, as described previously in
detail.34 Calculation of the rate constant, k1 ϩ kϪ1, of reversible
first-order reactions is based on the equilibrium constant of the
reaction. Since reaction B (Scheme 4) is, under the markedly
acidic conditions employed in kinetic measurements (0.5 mol
dmϪ3 aq. HCl) somewhat faster than the interconversion of
nucleoside 2Ј- and 3-alkylphosphates (reaction A), these equi-
librium constants could not be determined in the same solution,
but were obtained at pH 4.2 and 363.2 K. However, our
previous results4,34 show that the equilibrium constant is not
markedly sensitive to the hydronium ion concentration.
27 M. Ora, M. Oivanen and H. Lönnberg, J. Org. Chem., 1997, 62,
3246.
28 A. M. Davis, A. D. Hall and A. Williams, J. Am. Chem. Soc., 1988,
110, 5105.
29 K. Satoh and Y. Inoue, Chem. Lett., 1972, 1097.
30 K. N. Dalby, A. J. Kirby and F. Hollfelder, J. Chem. Soc., Perkin
Trans. 2, 1993, 1269.
31 S. A. Khan, A. J. Kirby, M. Wakselman, D. P. Horning and
J. M. Lawlor, J. Chem. Soc. B, 1970, 1182.
References
1 E. Anslyn and R. Breslow, J. Am. Chem. Soc., 1989, 111, 4473.
2 R. Breslow and D.-L. Huang, J. Am. Chem. Soc., 1990, 112, 9621.
3 R. Breslow, E. Anslyn and D.-L. Huang, Tetrahedron, 1991, 47,
2365.
4 P. Järvinen, M. Oivanen and H. Lönnberg, J. Org. Chem., 1991, 56,
5369.
5 M. Oivanen and H. Lönnberg, Trends Org. Chem., 1991, 2, 183.
6 R. Breslow and R. Xu, J. Am. Chem. Soc., 1993, 115, 10 705.
7 R. Breslow and R. Xu, Proc. Natl. Acad. Sci. USA, 1993, 90, 1201.
8 R. Breslow, S. D. Dong, Y. Webb and R. Xu, J. Am. Chem. Soc.,
1996, 118, 6588.
9 R. Breslow and M. Labelle, J. Am. Chem. Soc., 1986, 108, 2655.
10 S. Kuusela and H. Lönnberg, J. Chem. Soc., Perkin Trans. 2, 1994,
2109.
11 T. Uchimaru, M. Uebayasi, T. Hirose, S. Tsuzuki, A. Yliniemelä,
K. Tanabe and K. Taira, J. Org. Chem., 1996, 61, 1599.
12 T. Koike and Y. Inoue, Chem. Lett., 1972, 569.
13 D. M. Brown, D. I. Magrath, A. H. Neilson and A. R. Todd, Nature
(London), 1956, 177, 1124.
32 M. Oivanen, S. Mikhailov, N. Padyukova and H. Lönnberg, J. Org.
Chem., 1993, 58, 1617.
33 J. H. van Boom and C. T. J. Wreesman, in Oligonucleotide Synthesis.
A Practical Approach, ed. M. J. Gait, IPL Press, Oxford, 1984,
pp. 153–182.
34 M. Oivanen, R. Schnell, W. Pfleiderer and H. Lönnberg, J. Org.
Chem., 1991, 56, 3623.
35 E. P. Serjeant and B. Dempsey, Ionization Constants of Organic
Acids in Aqueous Solution, IUPAC Chemical Data Series 23,
Pergamon Press, Oxford, 1979.
14 M. Oivanen, M. Ora, H. Almer, R. Strömberg and H. Lönnberg,
J. Org. Chem., 1995, 60, 5620.
15 D. M. Perreault and E. V. Anslyn, Angew. Chem., Int. Ed. Engl.,
1997, 36, 432.
16 K. Taira, T. Uchimaru, J. W. Storer, A. Yliniemelä, M. Uebayasi and
K. Tanabe, J. Org. Chem., 1993, 58, 3009.
Paper 7/04636G
Received 1st July 1997
Accepted 22nd August 1997
2666
J. Chem. Soc., Perkin Trans. 2, 1997