Hydrolysis of 2′,3′-O-methyleneadenos-5′-yl Bis(2′,5′-di-O-methylurid-3′-yl) Phosphate, A Sugar O-Alkylated Trinucleoside 3′,3′,5′-Monophosphate: Implications for the Mechanism of Large Ribozymes

Article abstract of DOI:10.1021/jo035094k

Hydrolytic reactions of 2′,3′-O-methyleneadenos-5′-yl bis(2′,5′-di-O-methylurid-3′-yl) phosphate (1), a sugar O-alkylated trinucleoside 3′,3′,5′-monophosphate, have been followed by RP HPLC over a wide pH range. Under neutral and mildly acidic conditions,

Full text of DOI:10.1021/jo035094k

Hyd r olysis of 2′,3′-O-m eth ylen ea d en os-5′-yl
Bis(2′,5′-d i-O-m eth ylu r id -3′-yl) P h osp h a te, a Su ga r O-Alk yla ted
Tr in u cleosid e 3′,3′,5′-Mon op h osp h a te: Im p lica tion s for th e
Mech a n ism of La r ge Ribozym es
Tuomas Lo¨nnberg* and Satu Mikkola
Department of Chemistry, University of Turku, Vatselankatu 2, FIN-20014 Turku, Finland
tuanlo@utu.fi
Received J uly 28, 2003
Hydrolytic reactions of 2′,3′-O-methyleneadenos-5′-yl bis(2′,5′-di-O-methylurid-3′-yl) phosphate (1),
a sugar O-alkylated trinucleoside 3′,3′,5′-monophosphate, have been followed by RP HPLC over a
wide pH range. Under neutral and mildly acidic conditions, the only reaction observed was a pH-
independent cleavage of the O-C5′ bond of the 5′-linked nucleoside. Under more alkaline conditions
nucleophilic attack by hydroxide ion starts to compete. The reaction is first order in [OH-] and
becomes predominant at pH 10. Each of the 3′-linked nucleosides is displaced 2.9 times as readily
as the 5′-linked one. To determine the âlg value for the hydroxide ion catalyzed hydrolysis of 1, two
diesters (2a ,b) having 2′,3′-O-methyleneadenosine (7) and 2′,5′-di-O-methyluridine (4) as leaving
groups were hydrolyzed under alkaline conditions. Since the âlg value for this reaction is known,
∆pKa between 4 and 7 could be calculated. The âlg for the hydrolysis of 1 was estimated to be -0.5
with use of this information. The mechanisms of the partial reactions and the role of leaving group
properties in ribozyme reactions of large ribozymes are discussed.
In tr od u ction
the corresponding data on the intermolecular transes-
terification mimicking the action of large ribozymes is
extremely scanty. In fact such a reaction has been
detected only in nonaqueous solutions.4 The present
study is aimed at partly filling this gap. Kinetics for the
hydrolysis of a trinucleoside 3′,3′,5′-monophosphate, 2′,3′-
O-methyleneadenos-5′-yl bis(2′,5′-di-O-methylurid-3′-yl)
phosphate (1), have been studied over a wide pH range.
An attack of hydroxide ion on the phosphorus atom of 1
results in a monoanionic pentacoordinated phosphorane
intermediate or phosphorane-like transition state. This
The naturally occurring catalytic ribonucleic acids,
ribozymes, fall according to their size and catalytic
mechanism into two categories.1 The small ribozymes
that include hammerhead, hairpin, hepatitis delta virus,
and varkud satellite ribozymes catalyze the cleavage of
a phosphodiester bond in the target chain by facilitating
the intramolecular attack of the neighboring 2′-hydroxy
function on the phosphorus atom.2 In other words, a cyclic
phosphorane intermediate (or transition state) preceding
the release of the 5′-linked nucleoside and concomitant
formation of a 2′,3′-cyclic phosphate is obtained intramo-
lecularly without covalent involvement of the ribozyme.
With large ribozymes, viz. group I and II introns and the
RNA subunit of RNase P, the 3′-linked nucleoside of the
scissile phosphodiester bond is displaced intermolecularly
by the 3′(or 2′-)hydroxy function of the entering nucleo-
side. Accordingly, the reaction proceeds via an acyclic
phosphorane intermediate (or transition state) having
two nucleosides bonded via a secondary and one via a
primary oxygen atom.
intermediate/transition state closely resembles the one
5-9
proposed
for the ribozyme reaction of group I introns
(Figure 1). The main difference is that one of the
nonbridged phosphorane oxygens is protonated in the
transition state of the nonenzymatic model reaction
(TS1) but coordinated to one or two magnesium ions in
the ribozyme reaction (TS2). In addition, the entering
and departing oxyanion, i.e. the two 3′-oxygen atoms
bonded to the phosphorus atom, in all likelihood interact
with a magnesium ion. Despite these differences, an
understanding of the factors that affect the relative rates
While numerous studies with small molecular models
have been carried out to clarify the underlying chemical
principles of the cleavage of RNA by an intramolecular
mechanism,3 similar to that utilized by small ribozymes,
(4) Roussev, C. D.; Ivanova, G. D.; Bratovanova, E. K.; Vassilev, N.
G.; Petkov, D. D. J . Am. Chem. Soc. 1999, 121, 11267-11272.
(5) Piccirilli, J . A.; Vyle, J . S.; Caruthers, M. H.; Cech, T. R. Nature
1993, 361, 85-88.
(6) Weinstein, L. B.; J ones, B. C.; Cosstick, R.; Cech, T. R. Nature
1997, 388, 805-808.
* Corresponding author.
(7) Yoshida, A.; Shan, S.; Herschlag, D.; Piccirilli, J . A. Chem. Biol.
2000, 7, 85-96.
(8) Shan, S.; Yoshida, A.; Sun, S.; Piccirilli, J . A.; Herschlag, D. Proc.
Natl. Acad. Soc. U.S.A. 1999, 96, 12299-12304.
(9) Yoshida, A.; Sun, S.; Piccirilli, J . A. Nat. Struct. Biol. 1999, 6,
318-321.
(1) Takagi, Y.; Warashina, M.; Stec, W. J .; Yoshinari, K.; Taira, K.
Nucleic Acids Res. 2001, 29, 1815-1834.
(2) Zhou, D.-M.; Taira, K. Chem. Rev. 1998, 98, 991-1026.
(3) Oivanen, M.; Kuusela, S.; Lo¨nnberg, H. Chem. Rev. 1998, 98,
961-990.
10.1021/jo035094k CCC: $27.50 © 2004 American Chemical Society
Published on Web 01/10/2004
802
J . Org. Chem. 2004, 69, 802-810

Hydrolysis of a 3′,3′,5′-Trinucleoside Monophosphate
yl bis(2′,5′-di-O-methylurid-3′-yl) phosphate (10) was
obtained by tetrazole-activated stepwise displacement of
the dimethylamino groups from tris(dimethylamino)-
phosphine with the appropriately protected nucleosides
(3, 4) prepared as described in the literature (Scheme
1).15-19 Removal of the N6-benzoyl protection with metha-
nolic ammonia, followed by RP-HPLC purification, then
gave the desired triester 1. Similar compounds have
previously been prepared with the H-phosphonate meth-
od.20
P r od u ct Distr ibu tion a n d Rea ction P a th w a ys.
The hydrolysis of 1 was followed over a wide pH range
(5-13) by analyzing the composition of the aliquots
withdrawn from the reaction mixture at appropriate time
intervals by HPLC. The products were identified by
spiking with authentic samples. The time-dependent
product distributions obtained under neutral and alkaline
conditions were somewhat different. In aqueous alkali,
the 2′,5′-di-O-methyluridylyl-3′,3′-(2′,5′-di-O-methyluri-
dine) [m2U(3′,3′)m2U, 5] and 2′,5′-di-O-methyluridylyl-
3′,5′-(2′,3′-O-methyleneadenosine) [m2U(3′,5′)m<A, 6] di-
esters were formed by the release of 2′,3′-O-methyl-
eneadenosine (m<A, 7) and 2′,5′-di-O-methyluridine
(m2U, 4), respectively (Figure 2). In neutral or slightly
acidic solutions, the only chromophoric product observed
to be accumulated was the uridylyl-3′,3′-uridine diester
[m2U(3′,3′)m2U] (Figure 3). In other words, no nucleosidic
product was accumulated, as could have been expected
and as was also detected at high pH. As discussed below
in more detail, m2U(3′,3′)m2U was in all likelihood formed
by phosphate elimination (route C), not by phosphoester
hydrolysis (route B), and the other product of the
elimination reaction, vinyl ether 8, was not accumulated,
owing to rapid subsequent decomposition to several
products. No nucleoside monophosphates were formed
either under neutral or alkaline conditions, indicating
that the diesters, m2U(3′,3′)m2U (5) and m2U(3′,5′)m<A
(6), were not hydrolyzed under the experimental condi-
tions.
F IGURE 1. Transition state structures for the model (TS1)
and the ribozyme reaction (TS2).
of the breakdown of TS1 by either P-O3′ or P-O5′
cleavage is of interest.
With group I introns, the intermediate is always
decomposed by departure of a 3′-linked nucleoside, lead-
ing to formation of either products or starting materials.10
Both processes have been shown to be subject to a similar
catalysis.11 Alternative explanations may a priori be
given for the overwhelmingly preferred cleavage of the
P-O3′ bonds: (i) the intermediate is too short-lived to
pseudorotate and the initial state geometry dictated by
the chain folding determines the course of the reaction,
(ii) the apicophilicity and leaving-group property of a
secondary alcohol is superior to that of the primary
alcohol, and (iii) hydrogen bonding of the neighboring
hydroxy group to the developing oxyanion greatly en-
hances the departure of a 3′-oxygen.4,7,12,13 Since with our
small molecule model (1) the nucleophile can attack from
any direction and the substrate is free of the constraints
present in ribozymes, the differences between rates of
cleavage of P-O3′ and P-O5′ bonds should reflect the
differences in the apicophilicities and leaving group
properties. The possible hydrogen bond stabilization of
the leaving 3′-oxyanion by the neighboring 2′-OH is ruled
out by alkylating all sugar hydroxy functions of the
molecule. Accordingly, the results of the present study
serve as necessary reference material for future inves-
tigations aimed at elucidating the role of leaving group
stabilization by intramolecular hydrogen bonding. In
addition, the âlg value for the hydrolysis via the acyclic
phosphorane intemediate or phosphorane-like transition
state has been estimated by comparing the relative rate
of the release of the 3′- and 5′-linked nucleosides, 4 and
7, from triester 1 to that of the release from diesters 2a ,b.
The latter compounds react by an intramolecular attack
of the 2′-hydroxy group on phosphorus, and the âlg value
for such a reaction is known.14
Studies in 18O-enriched water (95%) offered an expla-
nation for the different product compositions. HPLC-ESI-
MS analyses of the products revealed that no 18O was
incorporated into the m2U(3′,3′)m2U diester (5) in neutral
or slightly acidic solutions. Accordingly, the reaction in
all likelihood proceeded by cleavage of the O-C5′(m<A)
bond. By contrast, in 50 mmol L-1 aqueous sodium hy-
droxide, 78% of m2U(3′,3′)m2U had one 18O atom incor-
porated and all of the m2U(3′,5′)m<A diester (6) formed
contained one 18O atom. In other words, the trinucleoside
monophosphate 1 is decomposed either by nucleophilic
attack of hydroxide ion on the phosphorus atom or by
cleavage of the O-C5′(m<A) bond (Scheme 2).
Resu lts a n d Discu ssion
In the pH range 5-9 (T ) 50 °C), the cleavage of the
O-C5′ bond (route C) is the only reaction detected. As
P r ep a r a tion of th e Tr in u cleosid e 3′,3′,5′-Mon o-
p h osp h a te (1). N6-Benzoyl-2′,3′-O-methyleneadenos-5′-
(15) Hovinen, J . Helv. Chim. Acta 1997, 80, 851-855.
(16) Legorburu, U.; Reese, C. B.; Song, Q. Tetrahedron 1999, 55,
5635-5640.
(10) Cech, T. R.; Zaug, A. J .; Grabowski, P. J . Cell 1981, 27, 487-
496.
(11) Steitz, T. A.; Steitz, J . A. Proc. Natl. Acad. Sci. U.S.A. 1993,
90, 6498-6502.
(12) Tzokov, S. B.; Momtcheva, R. T.; Vassilev, N. G.; Kaneti, J .;
Petkov, D. D. J . Am. Chem. Soc. 1999, 121, 5103-5107.
(13) Bakalova, S.; Siebrand, W.; Ferna´ndez-Ramos, A.; Smedar-
china, Z.; Petkov, D. D. J . Phys. Chem. B 2002, 106, 1476-1480.
(14) Kosonen, M.; Youseti-Salakdeh, E.; Stro¨mberg, R.; Lo¨nnberg,
H. J . Chem. Soc., Perkin Trans. 2 1997, 2661-2666.
(17) McGee, D. P. C.; Zhai, Y. Nucleosides Nucleotides 1996, 15,
1797-1803.
(18) Norman, D. G.; Reese, C. B.; Serafinowska, H. T. Synthesis
1985, 8, 751-754.
(19) Ti, G. S.; Gaffney, B. L.; J ones, R. A. J . Am. Chem. Soc. 1982,
104, 1316-1319.
(20) Garegg, P. J .; Lindh, I.; Stawinski, J . Bioactive Molecules; Vol.
3, Biophosphates Their Analogues; Elsevier Science: New York, 1987;
pp 89-92.
J . Org. Chem, Vol. 69, No. 3, 2004 803

Lo¨nnberg and Mikkola
SCHEME 1. P r ep a r a tion of th e Tr in u cleosid e 3′,3′,5′-m on op h osp h a te 1a
a
Reagents and conditions: (a) (1) TMSCl, Py, (2) BzCl, (3) H2O, NH3; (b) (PhO)2CdO, NaHCO3, DMA; (c) Mg(OMe)2, DMF; (d) P(NMe2)3,
tetrazole, MeCN; (e) 4, tetrazole, MeCN; (f) 3, tetrazole, MeCN; (g) I2, H2O, THF; (h) NH3, MeOH.
F IGURE 2. Time-dependent product distribution for the
hydrolysis of 1 in 0.5 mol L-1 aqueous NaOH at 50.0 °C: (9)
1, (b) 6, (O) 4, (2) 5, and (4) 7.
F IGURE 3. Time-dependent product distribution for the
hydrolysis of 1 at pH 4.8 and 50.0 °C: (9) 1 and (b) 5.
rahydrofuran, mimicking the probable elimination prod-
uct of 1, has been reported to hydrolyze rapidly under
neutral conditions (t1/2 ) 48 s, pH 6.87, T ) 25 °C, 0.00333
mol L-1 phosphate buffer, I ) 0.07 mol L-1), presumably
to 4-hydroxy-2-pentanone.23 The phenyl ester of thymi-
dine 3′,5′-cyclic phosphate has previously been shown to
undergo a pH-independent cleavage of the O-C5′ bond,
but in that case the 5′-linked nucleoside departs as an
mentioned above, only the m2U(3′,3′)m2U diester product
can be detected. One may tentatively assume that the
2′,3′-O-methyleneadenosyl moiety departs as a vinyl
ether type elimination product 8, and this subsequently
reacts to a variety of products, the concentration of each
being too low to be detected. Many organic phosphates,
including phosphotriesters, are known to undergo â-
elimination processes.21,22 In addition, 2-methylenetet-
(21) Cox, J . R.; Ramsey, O. B. Chem. Rev. 1964, 64, 317-352.
(22) Lapidot, A.; Samuel, D.; Silver, B. Chem. Ind. 1963, 468-471.
(23) Kankaanpera¨, A.; Taskinen, E.; Salomaa, P. Acta Chem. Scand.
1967, 21, 2487-2494.
804 J . Org. Chem., Vol. 69, No. 3, 2004

Hydrolysis of a 3′,3′,5′-Trinucleoside Monophosphate
SCHEME 2
alcohol, not as a vinyl ether.24 Evidently the carbocation
developing by cleavage of the C-O5′ bond may be
stabilized either by deprotonation to a vinyl ether or by
attack of hydroxide ion to give an alcohol. Which one of
these mechanisms prevail may depend on the detailed
structure of the starting material. Since the C-O5′ bond
cleavage was not of interest from the point of view of the
ribozyme action, the course of the elimination reaction
was not studied in more detail.
formed by a hydroxide ion catalyzed cleavage of the
C-O5′ bond, probably by an attack of hydroxide ion on
C5′(m<A) concerted with the bond rupture (route D).
p H-Ra te P r ofiles. Figure 4 shows the pH-rate
profile for the overall disappearance of 1. From pH 5 to
pH 9 the overall reaction is pH independent. Over this
relatively wide pH range the cleavage of the C-O bond
by route C predominates. On going to more alkaline
conditions (pH >10) the overall reaction becomes first
order in hydroxide ion concentration. This refers to the
cleavage of P-O3′ (route A) and P-O5′ (route B) bonds,
as well as the cleavage of the C-O5′ bond by route D.
The rate constants for the partial reactions were
calculated from the observed pseudo-first-order rate
constants as follows: since the m2U(3′,5′)m>A diester is
only formed by one route, namely by attack of hydroxide
ion, the rate constant for this reaction (route A) is ob-
tained directly from the observed rate constant (eq 1 ).
On passing pH 10, m2U(3′,5′)m<A (6) started to appear
as a product besides m2U(3′,3′)m2U (5) and both nucleo-
sidic products, m2U (4) and m<A (7), accumulated. On
the basis of 18O incorporation studies, m2U(3′,5′)m<A was
formed solely by a P-O bond cleavage (route B). m2U-
(3′,3′)m2U in turn exhibited only 78% incorporation of 18
O
(route B). Since under the experimental conditions (50
mmol L-1 of sodium hydroxide) the elimination reaction
(route C) accounts for only 5% of the overall formation
of m2U(3′,3′)m2U, the remaining 17% is in all likelihood
[m2U(3′,5′)m<A]
k1′ ) kobs
(1)
(24) Varila, J .; Hankama¨ki, T.; Oivanen, M.; Koole, L. H.; Lo¨nnberg,
H. J . Org. Chem. 1997, 62, 893-898.
[m2U(3′,5′)m<A] + [m2U(3′,3′)m2U]
J . Org. Chem, Vol. 69, No. 3, 2004 805

Lo¨nnberg and Mikkola
indicates that the 3′-linked m2U nucleoside (4) departs
the phosphorane intermediate 2.9 times as readily as the
5′-linked m<A nucleoside (7). Since all of the hydroxy
functions in triester 1 are methylated, this difference
cannot be attributed to a hydrogen bond stabilization of
the leaving 3′-oxyanion by the neighboring 2′-OH, but it
reflects the higher inherent leaving group ability of a 3′-
linked nucleoside. The difference is, however, smaller
than that observed with group I intron and gives, hence,
only a partial explanation for the course of the ribozyme
reaction.
The observed difference between the cleavage rates of
the P-O3′ and P-O5′ bonds allows one to estimate the
âlg value for the reaction via an acyclic trialkylated
phosphorane intermediate and, hence, to estimate the
extent of P-O bond cleavage in the transition state. For
this purpose, the pKa values of the departing m2U and
m<A alcohols are required. Since the âlg value for the
cleavage of ribonucleoside 3′-alkyl phosphates is known
to be -1.28 ( 0.05,14 this piece of information may be
exploited to determine the ∆pKa value for m2U and m<A.
The justification of this approach receives support from
the fact that the correlation line based on the pKa values
of alcohols that are either secondary or primary and
sterically very different, such as ethanol and 2,2,2-
trichloroethanol, exhibits practically no scattering. Ac-
cordingly, m2U and m<A were inserted as leaving groups
in guanosine 3′-diesters 2a and 2b (Scheme 3) and the
rate constants for the alkaline cleavage of these com-
pounds were determined under the same conditions used
previously for the determination of the âlg value for
uridine 3′-alkyl phosphates, i.e. at 25.0 °C in 0.1 mol L-1
of aqueous sodium hydroxide the ionic strength of which
was adjusted to 1.0 mol L-1 with sodium chloride. The
P-O3′(m2U) bond (2b) was cleaved 19.2 times as rapidly
as the P-O5′(m<A) bond (2a ), the observed first-order
rate constants obtained with 2a and 2b being 9.92 × 10-5
and 1.90 × 10-4 s-1, respectively. Accordingly, the âlg
value for the hydrolysis of triester 1 can be obtained by
(eq 5).
F IGURE 4. pH-rate profile for the decomposition of 1 at 50
°C.
There are, however, two similar P-O3′ bonds present
and, hence, k1′ must be divided by 2 to obtain the actual
rate constant k1 for the bond cleavage. Routes B and D
(Scheme 2) lead to identical products. 18O incorporation
studies reveal that in 50 mmol L-1 of aqueous sodium
hydroxide the attack on phosphorus is more rapid by a
factor of 4.6. Accordingly, under these particular condi-
tions the overall rate constant for the formation of m2U-
(3′,3′)m2U and m<A may be broken down to the rate
constants for reactions B and D (eqs 2 and 3). Finally,
since the elimination is the only reaction detected over a
wide pH range, its rate constant is simply the observed
rate constant under those conditions (eq 4). The following
partial rate constants were obtained: k1/[OH-] ) 2.3 ×
10-2 ( 1 × 10-4 L mol-1 s-1, k2/[OH-] ) 8.0 × 10-3 ( 2 ×
10-4 L mol-1 s-1, k3/[OH-] ) 1.8 × 10-3 ( 2 × 10-4
L
mol-1 s-1, k4 ) 5.6 × 10-6 ( 1 × 10-8 s-1
.
[m<A]
k2 ) 0.82k
(2)
obs[m2U(3′,5′)m<A] + [m2U(3′,3′)m2U]
log(k1/k2, triester)
[m<A]
âlg(triester) ) âlg(diesters)
)
k3 ) 0.18k
(3)
(4)
obs[m2U(3′,5′)m<A] + [m2U(3′,3′)m2U]
log(k1/k2, diesters)
-0.46 (5)
k4 ) kobs
(pH 5-9)
The value of approximately -0.5, although susceptible
to marked uncertainty, represents a relatively early
transition state, where the leaving group is still largely
bound to the phosphorus atom. For comparison, the âlg
value for the cleavage of ribonucleoside 3′-aryl phos-
phates is of the same magnitude and this reaction is
generally assumed to be a one-barrier process where both
the formation of the P-O2′ bond and the cleavage of the
P-OAr bond are only moderately advanced.27,28 The
cleavage of ribonucleoside 3′-phosphotriesters and 3′-
phosphodiesters, suggested to be cleaved via a very late
Reactions A and B are first order in hydroxide ion
concentration over the entire pH range where they can
be detected, while reaction C is pH independent at pH
<10, i.e. under conditions where it can be clearly de-
tected. Rate constants k1 and k2 are in reasonably good
agreement with that measured for methyl thymid-3′-yl
thymid-5′-yl phosphate.25 On the other hand, the hydrox-
ide ion catalyzed cleavage of trimethyl phosphate is about
2 orders of magnitude slower.26 The difference is ex-
pected, since both 3′- and 5′-linked nucleoside oxyanions
are much better leaving groups than methoxide ion.
Comparison of the rate constants obtained for reactions
A and B in 50 mmol L-1 of aqueous sodium hydroxide
transition state, in turn, both exhibit markedly negative
29
âlg values, -1.38 ( 0.18
and -1.28 ( 0.05,14 respec-
(27) Davis, A. M.; Hall, A. D.; Williams, A. J . Am. Chem. Soc. 1988,
110, 5105-5108.
(25) Conrad, J .; Mu¨ller, N.; Eisenbrand, G. Chem.-Biol. Interact.
1986, 60, 57-65.
(26) Barnard, P. W. C.; Bunton, C. A.; Llewellyn, D. R.; Vernon, C.
A.; Welch, V. A. J . Chem. Soc. 1961, 2670-2676.
(28) Ba-Saif, S. A.; Davis, A. M.; Williams, A. J . Org. Chem. 1989,
54, 5483-5486.
(29) Kosonen, M.; Seppa¨nen, R.; Wichmann, O.; Lo¨nnberg, H. J .
Chem. Soc., Perkin Trans. 2 1999, 2433-2439.
806 J . Org. Chem., Vol. 69, No. 3, 2004

Hydrolysis of a 3′,3′,5′-Trinucleoside Monophosphate
SCHEME 3. P r ep a r a tion of Diester s 2a a n d 2ba
a
Reagents and conditions: (a) 3, tetrazole, MeCN; (b) (1) I2, H2O, THF, 2,6-lutidine, (2) HOAc, H2O; (c) NH3, MeOH; (d) Et3N‚3HF,
THF; (e) 4, tetrazole, MeCN; (f) (1) I2, H2O, THF, 2,6-lutidine, (2) HOAc, H2O; (g) NH3, MeOH; (h) TBAF, THF.
2′,5′-Di-O-m eth ylu r id in e (4). 4 was prepared from 5′-O-
methyl-2,2′-anhydro-1-â-D-arabinofuranosyluracil (9) as previ-
ously described for 5′-O-(4,4′-dimethoxytrityl)-2′-O-methyl-
tively. We therefore presume the reaction to be rather of
Sn2(P) type than to proceed via a well-developed phos-
phorane intermediate.
1
uridine.13 Yield 2.4 g (89%). H NMR (δH) (400 MHz, DMSO-
In summary, the results of the present study suggest
that a nucleoside esterified through its 3′-hydroxy departs
upon nucleophilic attack of hydroxide ion on the phos-
phorus atom of a phosphotirester only 3 times as fast as
its 5′-esterified counterpart, as far as stabilization of the
leaving group by intramolecular hydrogen bonding is
prevented. Accordingly, additional factors, such as hy-
drogen-bonding stabilization of the leaving group and
noncovalent interactions steering the nucleophilic attack
to take place opposite the 3′-linked nucleoside, must be
considered in future studies.
d6) 7.71 (d, 1H, J ) 8.1 Hz), 5.82 (d, 1H, J ) 4.3 Hz), 5.67 (d,
1H, J ) 8.1 Hz), 5.26 (d, 1H, J ) 6.0 Hz), 4.06 (q, 1H, J ) 5.6
Hz), 3.91 (dd, 1H, J 1 ) 10.9 Hz, J 2 ) 3.0 Hz), 3.48 (dd, 1H, J 1
) 11.1 Hz, J 2 ) 4.1 Hz), 3.34 (s, 3H), 3.31 (s, 3H). HRMS (FAB)
M+ calcd 273.1087, obsd 273.1089.
N6-Ben zoyl-2′,3′-O-m eth ylen ea d en osin e (3). 3 was pre-
pared from 2′,3′-O-methyleneadenosine14 as previously de-
scribed for N6-benzoyl-2′-deoxyadenosine.15 Yield 3.6 g (83%).
1H NMR (δ) (400 MHz, DMSO-d6) 11.3 (s, 1H), 8.83 (s, 1H),
8.74 (s, 1H), 8.11 (d, 2H, J ) 7.5 Hz), 7.70 (t, 1H, J ) 7.5 Hz),
7.61 (t, 2H, J ) 7.5 Hz), 6.34 (d, 1H, J ) 3.0 Hz), 5.45 (dd, 1H,
J 1 ) 2.8 Hz, J 2 ) 6.4 Hz), 5.25 (s, 1H), 5.21 (s, 1H), 5.19 (t,
1H, J ) 5.6 Hz), 5.01 (dd, 1H, J 1 ) 3.4 Hz, J 2 ) 6.4 Hz), 4.28
(dd, 1H, J 1 ) 4.72 Hz, J 2 ) 8.36 Hz), 3.64 (m, 2H). HRMS
(FAB) M+ calcd 384.1308, obsd 384.1307.
N6-Ben zoyl-2′,3′-O-m eth ylen ea d en os-5′-yl Bis(2′,5′-d i-
O-m eth ylu r id -3′-yl) P h osp h a te (10). 2′,5′-Di-O-methyluri-
dine (4, 0.583 mmol, 0.1587 g) was dissolved in anhydrous
MeCN (2.1 mL). Tris(dimethylamino)phosphine (0.644 mmol,
0.117 mL) and a solution of tetrazole (0.571 mmol, 40 mg) in
anhydrous MeCN (2 mL) was added and the reaction mixture
was stirred for 21 h at 50 °C. A solution of 2′,5′-di-O-
methyluridine (4, 0.551 mmol, 0.15 g) in anhydrous MeCN (2
mL) was then added, followed by a solution of tetrazole (0.571
mmol, 40 mg) in anhydrous MeCN (1.5 mL). After the mixture
was stirred for 45 h at 60 °C, a solution of N6-benzoyl-2′,3′-
O-methyleneadenosine (3, 0.548 mmol, 0.21 g) in anhydrous
Exp er im en ta l Section
Ma ter ia ls. Nucleosides and protected nucleoside phos-
phoramidites were commercial products that were used as
received after checking the purity by HPLC. The 18O atom
fraction of H218O was 95%. The buffer constituents were of
reagent grade.
5′-O-Me t h yl-2,2′-a n h yd r o-1-â-D-a r a b in ofu r a n osylu -
r a cil (9). 9 was prepared from 5′-O-methyluridine11 by a
method previously described for 2,2′-anhydro-1-â-D-arabino-
1
furanosyluracil.12 Yield 4.3 g (86%). H NMR (δH) (400 MHz,
DMSO-d6) 7.82 (d, 1H, J ) 7.5 Hz), 6.30 (d, 1H, J ) 5.8 Hz),
5.83 (d, 1H, J ) 7.5 Hz), 5.18 (d, 1H, J ) 5.8 Hz), 4.32 (s, 1H),
4.19 (m, 1H), 3.30-3.18 (m, 2H), 3.06 (s, 3H). ESI+-MS m/z
241.1 [M + H]+.
J . Org. Chem, Vol. 69, No. 3, 2004 807

Lo¨nnberg and Mikkola
SCHEME 4. P r ep a r a tion of th e Refer en ce Diester s 5 a n d 6a
a
Reagents and conditions: (a) NCCH2CH2OP(NiPr2)2, tetrazole, MeCN; (b) 4, tetrazole, MeCN; (c) I2, H2O, THF, 2,6-lutidine; (d) NH3,
MeOH; (e) 3, tetrazole, MeCN; (f) I2, H2O, THF, 2,6-lutidine; (g) NH3, MeOH
MeCN (2 mL) was added. The reaction mixture was stirred
for an additional 23 h at 60 °C, after which iodine (1.74 mmol,
0.45 g) in a mixture of THF (12 mL) and water (6 mL) was
added. After the solution was stirred for 19 h at room
temperature, 50 mL of CH2Cl2 was added. The resulting
mixture was washed with saturated aqueous NaHSO3 (3 ×
10 mL) and the aqueous phase was extracted with CH2Cl2 (2
× 50 mL). The organic phase was dried with Na2SO4 and
evaporated to dryness. No attempt was made to optimize the
synthesis. The product was purified by HPLC on a LiChro-
spher RP-18 column (10 × 250 mm2, 10 µm) eluting with a
mixture of 0.05 mol L-1 of aqueous NH4OAc and MeCN (76:
24, v/v). The main purpose was to obtain a highly pure sample
for the kinetic measurements and, hence, no attention was
paid to the quantitativeness of the isolation. Yield 11.5 mg
purified by HPLC on a LiChrospher RP-18 column (10 × 250
mm2, 10 µm) eluting with a mixture of water and MeCN (75:
25, v/v). Yield 9.85 mg (99%). 31P NMR (δP) (202 MHz, DMSO-
1
d6) 0.76. H NMR (δH) (500 MHz, DMSO-d6) 8.34 (s, 1H), 8.17
(s, 1H), 7.67 (d, 1H, J ) 8.2 Hz), 7.66 (d, 1H, J ) 8.1 Hz), 7.37
(s, 2H), 6.23 (d, 1H, J ) 2.8 Hz), 5.90 (d, 1H, J ) 6.3 Hz), 5.87
(d, 1H, J ) 6.3 Hz), 5.71 (d, 1H, J ) 8.1 Hz), 5.71 (d, 1H, J )
8.1 Hz), 5.36 (dd, 1H, J 1 ) 2.9 Hz, J 2 ) 6.6 Hz), 5.18 (s, 1H),
5.18 (s, 1H), 5.06 (dd, 1H, J 1 ) 2.9 Hz, J 2 ) 6.6 Hz), 4.92 (m,
1H), 4.86 (m, 1H), 4.38 (m, 2H), 4.22 (m, 2H), 4.13 (m, 1H),
4.07 (m, 2H), 3.54 (dd, 2H, J 1 ) 3.2 Hz, J 2 ) 11.1 Hz), 3.48
(dd, 2H, J 1 ) 3.75 Hz, J 2 ) 11.0 Hz), 3.45 (dd, 1H, J 1 ) 3.6
Hz, J 2 ) 11.0 Hz), 3.39 (dd, 1H, J 1 ) 3.6 Hz, J 2 ) 11.0 Hz),
3.35 (s, 3H), 3.32 (s, 3H), 3.31 (s, 3H), 3.26 (s, 3H). HRMS
(FAB) M- calcd 866.2358, obsd 866.2362.
2′,5′-Di-O-m et h ylu r id ylyl-(3′,3′)-(2′,5′-d i-O-m et h ylu r i-
d in e) (5). 2′,5′-Di-O-methyluridine (4, 1.14 mmol, 0.3111 g)
and 2-cyanoethyl N,N,N′,N′-tetraisopropylphosphorodiamidite
(1.43 mmol, 0.455 mL) were dissolved in a solution of tetrazole
(1.11 mmol) in MeCN (3.46 mL). After being stirred for 40 min
at room temperature, the reaction mixture was evaporated to
dryness and a conventional CH2Cl2/aq NaHCO3 workup was
carried out. The organic phase was evaporated to dryness,
coevaporated from anhydrous MeCN, and dried in a vacuum
desiccator overnight. A solution of 2′,5′-di-O-methyluridine (4,
0.646 mmol, 0.1756 g) and tetrazole (0.608 mmol) in acetoni-
trile (2.350 mL) were added and the reaction mixture was
stirred for 3.5 h at room temperature. Iodine (1.37 mmol, 0.354
g) in a mixture of water (4 mL), THF (8 mL), and 2,6-lutidine
(2 mL) was added and the reaction mixture was stirred for an
additional 2.5 h at room temperature. CH2Cl2 (50 mL) was
1
(2%). 31P NMR (δH) (202 MHz, DMSO-d6) 0.75. H NMR (δH)
(500 MHz, DMSO-d6) 11.43 (d, 1H, J ) 2.2 Hz), 11.42 (d, 1H,
J ) 2.2 Hz), 8.76 (s, 1H), 8.68 (s, 1H), 8.03, (d, 2H, J ) 7.2
Hz), 7.7-7.4 (m, 6H), 6.36 (d, 1H, J ) 2.9 Hz), 5.89 (d, 1H, J
) 6.3 Hz), 5.85 (d, 1H, J ) 6.4 Hz), 5.71 (d, 1H, J ) 8.1 Hz),
5.71 (d, 1H, J ) 8.1 Hz), 5.42 (dd, 1H, J 1 ) 2.8 Hz, J 2 ) 6.4
Hz), 5.20 (s, 1H), 5.19 (s, 1H), 5.09 (dd, J 1 ) 4 Hz, J 2 ) 6.4
Hz), 4.92 (m, 1H), 4.86 (m, 1H), 4.44 (m, 1H), 4.38 (m, 1H),
4.26 (m, 1H), 4.22 (dd, 1H, J 1 ) 6.6 Hz, J 2 ) 3.2 Hz), 4.13 (dd,
1H, J 1 ) 6.7 Hz, J 2 ) 3.5 Hz), 4.08 (dd, 2H, J 1 ) 5.8 Hz, J 2
)
11.8 Hz), 3.58-3.43 (m, 4H), 3.34 (s, 3H), 3.31 (s, 3H), 3.30 (s,
3H), 3.24 (s, 3H). ESI--MS m/z 970.7 [M - H]-.
2′,3′-O-Meth ylen ea d en os-5′-yl Bis(2′,5′-d i-O-m eth ylu -
r id -3′-yl) P h osp h a te (1). Compound 10 was dissolved in
saturated methanolic ammonia. After being stirred for 5.5 h
the solution was evaporated to dryness and the product was
808 J . Org. Chem., Vol. 69, No. 3, 2004

Hydrolysis of a 3′,3′,5′-Trinucleoside Monophosphate
added and the resulting mixture was washed with saturated
aqueous NaHSO3 (2 × 20 mL). The aqueous phase was
extracted with dichloromethane (2 × 50 mL). The combined
organic phases were dried with Na2SO4 and evaporated to
dryness. The product was purified on a silica gel column with
a mixture of CH2Cl2 and MeOH as an eluent (90:10, v/v). The
purified product was then dissolved in saturated methanolic
ammonia. After being stirred for 2 h at room temperature the
reaction mixture was evaporated to dryness. Yield 0.11 g
(34%). 31P NMR (δP) (202 MHz, DMSO-d6) -0.02. 1H NMR (δH)
(400 MHz, DMSO-d6) 7.72 (d, 2H, J ) 8.1 Hz), 5.83 (d, 2H, J
) 5.8 Hz), 5.69 (d, 2H, J ) 8.1 Hz), 4.47 (m, 2H), 4.20 (m, 2H),
3.84 (m, 2H), 3.60-3.48 (m, 4H), 3.33 (s, 6H), 3.32 (s, 6H).
HRMS (FAB) M- calcd 605.1496, obsd 605.1496.
dissolved in saturated methanolic ammonia. After being
stirred for 7.5 h at room temperature the reaction mixture was
evaporated to dryness. The residue was dissolved in anhydrous
THF (1 mL) and triethylamine trihydrofluoride (0.27 mL) was
added. The reaction mixture was stirred for 45 min at 45 °C,
after which 10 mL of 0.1 mol L-1 of aqueous NaOAc (10 mL)
was added. The mixture was evaporated to dryness and a CH2-
Cl2/aq NaOAc workup was performed. The aqueous phase was
concentrated under reduced pressure and the product was
purified by HPLC on a LiChrospher RP-18 column (10 × 250
mm2, 10 µm) eluting with a mixture of 0.06 mol L-1 of aqueous
NaOAc and MeCN (91:9, v/v). Finally the buffer salts were
removed on the same column by eluting with a mixture of
water and MeCN. Yield 67.7 mg (21%). 31P NMR (δP) (162
MHz, DMSO-d6) 0.64. 1H NMR (δH) (400 MHz, DMSO-d6) 8.40
(s, 1H), 8.16 (s, 1H), 7.89 (s, 1H), 7.32 (s, 2H), 6.53 (s, 2H),
6.14 (d, 1H, J ) 5.6 Hz), 5.72 (d, 1H, J ) 6.8 Hz), 5.28 (dd,
1H, J 1 ) 6.0 Hz, J 2 ) 3.4 Hz), 5.17 (s, 1H), 5.12 (s, 1H), 4.96
(dd, 1H, J 1 ) 8.1 Hz, J 2 ) 4.1 Hz), 4.34 (m, 2H), 4.28 (m, 1H),
3.86 (m, 1H), 3.44-3.41 (m, 4H). HRMS (FAB) M- calcd
623.1364, obsd 623.1376.
2′,5′-Di-O-m et h ylu r id ylyl-(3′,5′)-(2′,3′-O-m et h ylen ea d -
en osin e) (6). 2′,5′-Di-O-methyluridine (4, 0.763 mmol, 0.2076
g) was dried by coevaporating from anhydrous pyridine and
anhydrous MeCN. To the residue were added a solution of
tetrazole (0.916 mmol) in MeCN (2.035 mL) and 2-cyanoethyl-
N,N,N′,N′-tetraisopropylphosphorodiamidite (0.916 mmol, 0.291
mL). After being stirred for 60 min at room temperature, the
reaction mixture was evaporated to dryness. A conventional
CH2Cl2/aq NaHCO3 workup was performed and the combined
aqueous phases were evaporated to dryness. To the residue
was added N6-benzoyl-2′,3′-O-methyleneadenosine (3, 0.787
mmol, 0.3015 g), and the resulting mixture was dried by
coevaporating from anhydrous pyridine and anhydrous MeCN.
The residue was dissolved in a solution of tetrazole (0.916
mmol) in MeCN (3.035 mL). The reaction mixture was stirred
for 67 h at room temperature, after which iodine (1.43 mmol,
0.37 g) in a mixture of THF (8 mL), water (4 mL), and 2,6-
lutidine (2 mL) was added. After being stirred for an additional
7 h at room temperature, the reaction mixture was concen-
trated under reduced pressure and a CH2Cl2/aq NaHSO3
workup was carried out. The combined organic phases were
evaporated to dryness and the product was purified on a silica
gel column with a mixture of CH2Cl2 and MeOH as an eluent
(93:7, v/v). The purified product was then dissolved in satu-
rated methanolic ammonia. After being stirred for 7 h at room
temperature the reaction mixture was evaporated to dryness.
The product was purified by HPLC on a LiChrospher RP-18
column (10 × 250 mm2, 10 µm) eluting with a mixture of 0.06
mol L-1 of aqueous NaOAc and MeCN (91.5:8.5, v/v). Finally
the buffer salts were removed on the same column by eluting
with a mixture of water and MeCN. Yield 55.5 mg (11%). 31P
NMR (δP) (162 MHz, DMSO-d6) -0.36. 1H NMR (δH) (400 MHz,
DMSO-d6) 11.36 (s, 1H), 8.44 (s, 1H), 8.14 (s, 1H), 7.68 (d, 1H,
J ) 8.4 Hz), 7.31 (s, 2H), 6.12 (d, 1H, J ) 3.0 Hz), 5.81 (d, 1H,
J ) 5.6 Hz), 5.68 (dd, 1H, J 1 ) 8.1 Hz, J 2 ) 2.1 Hz), 5.30 (dd,
1H, J 1 ) 6.4 Hz, J 2 ) 3.2 Hz), 5.16 (s, 1H), 5.11 (s, 1H), 4.95
(dd, 1H, J 1 ) 6.4 Hz, J 2 ) 3.2 Hz), 4.41 (m, 1H), 4.28 (m, 1H),
4.08 (m, 1H), 3.77 (m, 3H), 3.45-3.36 (m, 2H), 3.27 (s, 3H),
3.23 (s, 3H). HRMS (FAB) M- calcd 612.1455, obsd 612.1472.
Gu a n osyl-(3′,5′)-(2′,3′-O-m eth ylen ea d en osin e) (2a ). N6-
Benzoyl-2′,3′-O-methyleneadenosine (3, 0.561 mmol, 0.2147 g)
and N2-acetyl-5′-O-(4,4′-dimethoxytrityl)-2′-O-(triisopropylsi-
lyloxymethyl)guanosine 3′-(2-cyanoethyl-N,N-diisopropylphos-
phoramidite) (11, 0.5 mmol, 0.5 g) were dissolved in a solution
of tetrazole (0.63 mmol) in MeCN (1.4 mL). The reaction
mixture was stirred for 1 h at room temperature, after which
iodine (1.82 mmol, 0.47 g) in a mixture of THF (8 mL), water
(4 mL), and 2,6-lutidine (2 mL) was added. After being stirred
for an additional 1 h, the reaction mixture was concentrated
under reduced pressure and a CH2Cl2/aq NaHSO3 workup was
performed. The organic phase was evaporated to dryness and
the residue was dissolved in a mixture of acetic acid and water
(25 mL, 80:20, v/v). After being stirred for 50 min at room
temperature the reaction mixture was evaporated to dryness,
followed by a conventional CH2Cl2/aq NaHCO3 workup. The
organic phase was evaporated to dryness and the product was
purified on a silica gel column eluting with a mixture of CH2-
Cl2 and MeOH (98:2, v/v). The purified product was then
Gu a n osyl-(3′,3′)-(2′,5′-d i-O-m eth ylu r id in e) (2b). 2′,5′-Di-
O-methyluridine (4, 0.551 mmol, 0.15 g) and N2-acetyl-5′-O-
(4,4′-dimethoxytrityl)-2′-O-(triisopropylsilyloxymethyl)gua-
nosine 3′-(2-cyanoethyl-N,N-diisopropylphosphoramidite) (11,
0.5 mmol, 0.5 g) were dissolved in a solution of tetrazole (0.63
mmol) in MeCN (1.4 mL). The reaction mixture was stirred
at room temperature for 30 min, after which iodine (1.67 mmol,
0.43 g) in a mixture of water (4 mL), tetrahydrofuran (8 mL),
and 2,6-lutidine (2 mL) was added. After being stirred for an
additional 3 h at room temperature, the reaction mixture was
concentrated under reduced pressure. A CH2Cl2/aq NaHSO3
workup was carried out, the organic phase was evaporated to
dryness, and the residue was dissolved in a mixture of acetic
acid and water (25 mL, 80:20, v/v). After being stirred for 3 h
at room temperature, the reaction mixture was evaporated to
dryness, followed by a conventional CH2Cl2/aq NaHCO3 work-
up. The organic phase was evaporated to dryness and the
product was purified on a silica gel column eluting with a
mixture of CH2Cl2 and MeOH, the MeOH content of which was
increased stepwise from 5% to 10%. The purified product was
then dissolved in saturated aqueous ammonia. After being
stirred for 4 h at room temperature the reaction mixture was
evaporated to dryness. The residue was dissolved in a solution
of tetrabutylammonium fluoride (1.62 mmol, 0.422 g) in
anhydrous MeCN (1.6 mL), and the solution was stirred for 3
h
at room temperature. The reaction mixture was then
evaporated to dryness and a CH2Cl2/aq NaOAc workup was
carried out. The aqueous phase was concentrated under
reduced pressure and the product was purified by HPLC on a
LiChrospher RP-18 column (10 × 250 mm2, 10 µm) eluting
with a mixture of 0.06 mol L-1 of aqueous NaOAc and MeCN
(88:12, v/v). The purified product was passed through a Na+-
form Dowex 50-W (100-200 mesh) cation exchange column.
Finally, the buffer salts were removed by HPLC on the same
column as previously eluting with a mixture of water and
MeCN (88:12, v/v). Yield 64.7 mg (20%). 31P NMR (δP) (162
MHz, DMSO-d6) 0.64. 1H NMR (δH) (400 MHz, DMSO-d6) 7.87
(s, 1H), 7.70 (d, 1H, J ) 8.1 Hz), 7.35 (d, 1H, J ) 7.3 Hz), 5.68
(d, 1H, J ) 8.1 Hz), 4.57-4.51 (m, 2H), 4.40 (m, 1H), 4.18 (dd,
1H, J 1 ) 6.2 Hz, J 2 ) 3.0 Hz), 3.94 (m, 1H), 3.87 (t, 1H, J )
5.6 Hz), 3.60-3.49 (m, 4H), 3.34 (s, 3H), 3.30 (s, 3H). HRMS
(FAB) M- calcd 616.1404, obsd 616.1403.
Kin etic Mea su r em en ts. The reactions were carried out
in sealed tubes immersed in a thermostated water bath, the
temperature of which was adjusted to 50 °C within (0.1 °C.
The hydronium ion concentration of the reaction solutions was
adjusted with sodium hydroxide and acetate, triethanolamine,
glycine, and triethylamine buffers. The pH values of the
buffers were calculated from the literature data of the pKa
values of the buffer acids under experimental conditions. The
ionic strength of the solutions was adjusted to 1.0 mol L-1 with
J . Org. Chem, Vol. 69, No. 3, 2004 809

Lo¨nnberg and Mikkola
NaNO3, except with experiments on diesters 2a ,b, in which
case NaCl was used. The effect of the buffer concentration on
the reaction rate was determined at pH 8 by carrying out two
runs at different buffer concentrations. No buffer catalysis was
observed. Otherwise, a buffer concentration of 0.1 mol L-1 was
used.
The initial substrate concentration in the kinetic runs was
ca. 10-4 mol L-1. The composition of the samples withdrawn
at appropriate intervals was analyzed by HPLC on a Supelcosil
LC-18-S column (4.6 × 150 mm2, 5 µm) or a Hypersil-Keystone
Aquasil C18 column (4 × 150 mm2, 5 µm), using 0.05 mol L-1
of aqueous NH4OAc and MeCN as an eluent. The amount of
MeCN was linearly increased from 8 to 30%. The observed
retention times (tR, min) for the hydrolytic products of 1
(the flow rate was 1.5 mL min-1) were as follows: 14.3 (1),
12.2 (5), 11.5 (6), 5.4 (7), and 4.6 (4). The products were
identified by spiking with authentic reference samples (Scheme
4), and the characterization was further verified by LC/MS
analysis.
and 2.375 mol L-1 of aqueous sodium hydroxide (2 µL). The
vial was sealed and immersed in a water bath at 50 °C for 2
h, after which the solution was neutralized by adding 1 mol
L-1 of aqueous nitric acid (4.76 µL). For reference purposes,
an analogous experiment was carried out by using a corre-
sponding sodium hydroxide solution prepared in 16O water.
Composition of the product mixture was analyzed by HPLC/
MS.
Ack n ow led gm en t. We thank Prof. Harri Lo¨nnberg
for fruitful discussions.
Note Ad d ed a fter ASAP P ostin g. Due to a produc-
tion error, the wrong figures were published in the
version posted ASAP J anuary 10, 2004; the corrected
version was posted J anuary 13, 2004.
Su p p or tin g In for m a tion Ava ila ble: 13C NMR spectra of
compounds 9, 4, 3, 10, 1, 5, 6, 2a , and 2b. This material is
available free of charge via the Internet at http://pubs.acs.org.
Ma ss Sp ectr om etr ic An a lysis of th e P r od u cts of Hy-
d r olysis in 18O-En r ich ed Wa ter . Compound 1 (ca. 35 µg)
was dissolved in a mixture of H218O (95 atom % 18O, 98 µL)
J O035094K
810 J . Org. Chem., Vol. 69, No. 3, 2004