Efficient synthesis of [2′-18O]uridine and its incorporation into oligonucleotides: A new tool for mechanistic study of nucleotidyl transfer reactions by isotope effect analysis

Article abstract of DOI:10.1021/jo701727h

(Chemical Equation Presented) Lack of sufficient quantities of isotopically labeled materials has precluded the use of heavy atom isotope effects to investigate mechanisms of nucleotidyl transfer reactions in nucleic acids. Here we achieve regioselective

Full text of DOI:10.1021/jo701727h

Efficient Synthesis of [2′-18O]Uridine and Its
Incorporation into Oligonucleotides: A New Tool
for Mechanistic Study of Nucleotidyl Transfer
Reactions by Isotope Effect Analysis
these isotope effects on reactivity can provide essential informa-
tion for defining chemical mechanism. For nucleotidyl transfer
reactions, the nucleophile isotope effect on the attacking 2′-O
in particular offers the opportunity to distinguish unambiguously
between stepwise and concerted reaction mechanisms.3
Technical challenges severely limit the use of isotope effects
to study phosphoryl transferase enzymes that operate on nucleic
acids. One such limitation involves access to nucleic acid
substrates bearing the desired site-specific isotope enrichment.
Here we describe an efficient synthesis of [2′-18O]uridine, which
enables synthesis of RNA substrates for isotope effect analyses
on protein and RNA enzymes that catalyze nucleophilic attack
by the 2′-OH.
Few strategies exist for the synthesis of isotope-enriched
nucleosides, particularly sugar isotopomers.4 Previously, Pang
et al. obtained [3′-18O]uridine in low yield via reversible
hydration (HCl/H218O) of a 3′-ketouridine derivative followed
by reduction and separation from the predominant xylo epimer.5
Recently, Wnuk et al. reported access to [3′-18O]-1-(â-D-
arabinofuranosyl)uracil through a six-step synthesis from 2,3′-
cyclouridine involving a Fox thermal rearrangement as the key
step.6 However, this strategy generated arabinofuranosyl deriva-
tives and could not be used to prepare [2′-18O]uridine. Another
strategy reported in the literature to prepare 18O-labeled sugar
isotopomers makes use of SN2-substituted reaction of triflate
with 18O-labeled nucleophiles.7 Unfortunately, the 2′-â triflate
derivative of uridine failed to undergo the analogous reaction
to generate [2′-18O]uridine.
Qing Dai,† John K. Frederiksen,† Vernon E. Anderson,¶
Michael E. Harris,*,# and Joseph A. Piccirilli*,†,‡,§
Department of Biochemistry & Molecular Biology, Department
of Chemistry, and Howard Hughes Medical Institute, The
UniVersity of Chicago, 929 East 57th Street, MC 1028, Chicago,
Illinois 60637, and Department of Biochemistry and Center for
RNA Molecular Biology, School of Medicine, Case Western
ReserVe UniVersity, CleVeland, Ohio 44106
jpicciri@uchicago.edu; meh2@cwru.edu
ReceiVed August 8, 2007
Lack of sufficient quantities of isotopically labeled materials
has precluded the use of heavy atom isotope effects to
investigate mechanisms of nucleotidyl transfer reactions in
nucleic acids. Here we achieve regioselective opening of 2,2′-
cyclouridine with [18O2]benzoic acid/potassium hydride,
allowing an efficient “one-pot” synthesis of [2′-18O]uridine
in 88% yield. Conversion to the corresponding phosphora-
midite enables solid-phase synthesis of [2′-18O] RNA sub-
strates for isotope effect studies with nucleotidyl transferases
and hydrolases.
Commercially available 2,2′-cyclouridine (1) could provide
direct access to [2′-18O]uridine (2) if an 18O-enriched oxygen
nucleophile can be made to attack the ribose C-2′ regioselec-
tively. Initial attempts were directed toward hydrolysis of 1
under alkaline conditions (1 N NaOH, MeOH, rt, overnight).
Consistent with previous observations,6 the reaction gave only
1-(â-D-arabinofuranosyl)uracil, indicating that hydroxide attacks
exclusively at the nucleobase C-2 position. Therefore, to gain
access to 2 from 1, we sought to identify an oxygen nucleophile
with altered regioselectivity.
Ueda has summarized the reactions of 2,2′-cyclouridine with
various nucleophiles.8 Unfortunately, none of the reported
examples (Table 1) can be used to synthesize [2′-18O]uridine
(2). In general, reactions of 2,2′-cyclouridine with nucleophiles
have two possible regiochemical outcomes: (1) attack at the
A wide range of enzymes, including both protein and RNA
catalysts, promotes cleavage of RNA phosphodiester bonds by
attack of the adjacent 2′-OH and displacement of the 5′-O
leaving group. In solution, this reaction is catalyzed by both
acid and base, and both stepwise and concerted reaction
mechanisms are observed.1 Despite significant effort, little direct
experimental evidence exists that determines which specific
mechanism is followed for enzyme-catalyzed RNA cleavage.
The effect of isotopic substitution on chemical reaction kinetics
and equilibria provides an especially powerful way to investigate
enzymatic transition states and their active site interactions.
Isotopic substitution represents the smallest possible chemical
perturbation of a catalytic system but can nonetheless have
important influences on chemical reactivity.2 Interpretation of
(2) (a) Northrop, D. B. Methods 2001, 24, 117. (b) Hengge, A. C. FEBS
Lett. 2001, 501, 99.
(3) Anderson, V. E.; Ruszczycky, M. W.; Harris, M. E. Chem. ReV. 2006,
106, 3236.
(4) (a) Follman, H.; Hogencamp, H. P. C. J. Am. Chem. Soc. 1970, 92,
671. (b) Solsten, R. T.; McCloskey, J. A.; Schram, K. H. Nucleosides
Nucleotides 1982, 1, 57. (c) Schubert, E. M.; Schram, K. H. J. Labelled
Compd. Radiopharm. 1982, 19, 929. (d) Schwartz, H. M.; MacCoss, M.;
Danyluk, S. S. J. Am. Chem. Soc. 1983, 105, 5901. (e) Schwartz, H. M.;
MacCoss, M.; Danyluk, S. S. Magn. Reson. Chem. 1985, 23, 885.
(5) Pang, H.; Schram, K. H.; Smith, D. L.; Gupta, S. P.; Townsend, L.
B.; McCloskey, J. A. J. Org. Chem. 1982, 47, 3923.
(6) Wnuk, S. F.; Chowdhury, S. M.; Garcia, P. I., Jr.; Robins, M. J. J.
Org. Chem. 2002, 67, 1816.
(7) (a) Jiang, C.; Suhadolnik, R. J.; Baker, D. C. Nucleosides Nucleotides
1988, 7, 271. (b) D’Souza, F. W.; Lowary, T. L. J. Org. Chem. 1998, 63,
3166.
(8) Ueda, T. Synthesis and Reaction of Pyrimidine Nucleosides. In
Chemistry of Nucleosides and Nucleotides, 1st ed.; Townsend L. B., Ed.;
Plenum Press: New York and London, 1988; pp 1-112.
† Department of Biochemistry & Molecular Biology, The University of
Chicago.
‡ Department of Chemistry, The University of Chicago.
§ Howard Hughes Medical Institute, The University of Chicago.
¶ Department of Biochemistry, Case Western Reserve University.
# Center for RNA Molecular Biology, Case Western Reserve University.
(1) Oivanen, M.; Kuusela, S.; Lonnberg, H. Chem. ReV. 1998, 98, 961.
10.1021/jo701727h CCC: $40.75 © 2008 American Chemical Society
Published on Web 12/04/2007
J. Org. Chem. 2008, 73, 309-311
309

TABLE 1. Reaction of 2,2′-Cyclouridine with Various Nucleophiles
T
(°C)
time
(h)
yielda
(%)
entry
nucleophile
NaOH
Mg(OMe)2
B(OMe)3
solvent
product
16
211
MeOH/H2O
MeOH
MeOH+CH(OMe)3
MeOH
rt
65
150
37
150
20
16
5
42
48-168
<1
120
12
arabino
ribo
ribo
arabino
ribo
arabino
ribo
69
92
86
16
65
70
93
312
413
NH3
514
NaN3+BzOH
H2S+Et3N
EtSH+reagent 1b
t-BuSH+reagent 1b
AcSH
PhSe-SePh+NaBH4
HF
HCl
LiBr
HMPA
615
DMF
DMF
DMF
716
60
816
100
110
78
16
6
1
ribo
ribo
ribo
94
65
90
917
dioxane
EtOH
dioxane
dioxane
dioxane
acetone
1018
1119
1220
1321
1422
100-110
75-80
60
18
18
6
ribo
ribo
ribo
40-50
89
98
NaI+TsOH‚H2O
50
2.5
ribo
98
a Yields are for pure ribo and arabino product. b Reagent 1: N,N,N′,N′-tetramethylguanidine.
favors ribose attack.14 We used this apparent empirical trend to
guide our experiments. To favor nucleophilic attack at ribose,
we sought to replace the hydrogen atoms of H2O with
substituents that have larger øP values.
SCHEME 1. The Two Modes of Nucleophilic Attack to
2,2′-Cyclouridine
Although in dimethylformamide (DMF) Mg(OMe)2 favors
ribose attack to form 2′-O-methyluridine,11 we cannot access
uridine readily from 2′-O-methyluridine because the methyl
group is difficult to remove. We reasoned that sodium acetate
(NaOAc) might favor ribose attack as an acetyl group withdraws
electrons more strongly than a methyl group. Hydrolytic removal
of the acetyl group from the resulting product would then allow
facile access to uridine. The commercial availability of [18O2]-
NaOAc makes this strategy especially attractive. We heated
NaOAc with 1 in DMF at 140 °C for 24 h. TLC showed that
uridine formed along with the regioisomer 1-(â-D-arabinofura-
nosyl)uracil in a ratio of 2:3 ribo/arabino. Using potassium
acetate (KOAc) rather than NaOAc improved the ribo/arabino
ratio to 3:2. Despite this partial success, the poor solubility of
NaOAc and KOAc in DMF and the significant amount of
byproduct from nucleobase attack (which in larger scale
reactions proved difficult to separate from the desired uridine
by column chromatography) led us to explore other alternatives.
To enhance solubility in DMF, we tested potassium benzoate
(KOBz) as the nucleophile, hoping that the greater electron-
withdrawing power23 of the benzoyl moiety (relative to acetyl
moiety) might improve regioselectivity. We heated a mixture
of 1 (1.0 equiv) and KOBz (1.0 equiv) in DMF at 140 °C and
monitored the reaction by TLC. After 1 week, TLC revealed
that little starting material remained, and that two major products
and one minor product had formed. The major products
corresponded to uridine and 2′-O-benzoyluridine. The minor
product co-migrated with 1-(â-D-arabinofuranosyl)uracil and
presumably forms via benzoate attack on the nucleobase. After
treating the crude reaction mixture with NaOMe/MeOH to
nucleobase C-2 (nucleobase attack) to give C-2-substituted 1-(â-
D-arabinofuranosyl)uracil derivatives, or (2) attack at the pentose
C-2′ (ribose attack) to give 2′-substituted uridine derivatives
(Scheme 1).9 Inspection of the literature data reveals an apparent
regioselectivity trend: for the same nucleophilic atom, the
nucleophile associated with substituents having large electrone-
gativity values (øP)10 tends to favor ribose attack. To illustrate,
for sulfur nucleophiles, t-BuSH, EtSH, or AcSH (øP of C is
2.55) favor ribose attack,16,17 while H2S (øP of H is 2.20) favors
nucleobase attack.15 For oxygen nucleophiles, NaOMe (øP of
Na is 0.93) favors nucleobase attack, while Mg(OMe)2 (øP of
Mg is 1.31) and B(OMe)3 (øP of B is 2.03) favor ribose
attack.11,12 With nitrogen nucleophiles, NH3 (øP of H is 2.20)
favors nucleobase attack,13 whereas NaN3 (øP of N is 3.04)
(9) Sebesta, D. P.; O’ Rourke, S. S.; Martinez, R. L.; Pieken, W. A.;
McGee, D. P. C. Tetrahedron 1996, 52, 14385.
(10) Shriver, D. F.; Atkins, P. W.; Salvador, P. Inorganic Chemistry,
3rd ed.; W. H. Freeman: New York, 2006.
(11) Roy, S. K.; Tang, J.-Y. Org. Process Res. DeV. 2000, 4, 170.
(12) Ross, B. S.; Springer, R. H.; Tortorici, Z.; Dimock, S. Nucleosides
Nucleotides 1997, 16, 1641.
(13) (a) Delia, T. J.; Beranek, J. J. Carbohydr., Nucleosides Nucleotides
1977, 4, 349. (b) Ozaki, H.; Nakajima, K.; Tatsui, K.; Izumi, C.; Kuwahara,
M.; Sawai, H. Bioorg. Med. Chem. Lett. 2003, 13, 2441.
(14) (a) Li, H.; Jiang, Z.; Wang, X.; Zheng, C. Synth. Commun. 2006,
36, 1933. (b)Vasil’eva, S. V.; Abramova, T. V. Sil’nikov, V. N. Russ. J.
Bioorg. Chem. 2004, 30, 264.
(18) Ozaki, H.; Nakajima, K.; Tatsui, K.; Izumi, C.; Kuwahara, M.;
Sawai, H. Bioorg. Med. Chem. Lett. 2003, 13, 2441.
(19) Codington, J. F.; Doerr, I.; Van Praag, D.; Bendich, A.; Fox, J. J.
J. Am. Chem. Soc. 1961, 83, 5030.
(20) Vanheusden, V.; Munier-Lehmann, H.; Pochet, S.; Herdewijn, P.;
Van Calenbergh, S. Bioorg. Med. Chem. Lett. 2002, 12, 2695.
(21) Aoyama, Y.; Sekine, T.; Iwamoto, Y.; Kawashima, E.; Ishido, Y.
Nucleosides Nucleotides 1996, 15, 733.
(15) Sekiya Takao Ukita, T. Chem. Pharm. Bull. 1967, 15, 1498.
(16) Divakar, K. J.; Reese, C. B. J. Chem. Soc., Perkin Trans. 1 1982,
1625.
(22) Sukeda, M.; Shuto, S.; Sugimoto, I.; Ichikawa, S.; Matsuda, A. J.
Org. Chem. 2000, 65, 8988.
(17) Imazawa, M.; Ueda, T.; Ukita, T. Chem. Pharm. Bull. 1975, 23,
604.
(23) Benzoic acid (BzOH, pKa ) 4.2) is a stronger acid than acetic acid
(pKa ) 4.8).
310 J. Org. Chem., Vol. 73, No. 1, 2008

SCHEME 2. Synthesis of [2′-18O]Uridine
type 5′-UG and 5′-UGGGUCGGC gave [M - H]- peaks at
m/z 588 and 2881, respectively.
In conclusion, regiochemistry of nucleophilic reactions with
2,2′-cyclouridine appears to be influenced by the chemical
context of the nucleophilic atom and the presence of acid. Using
these hypotheses as a guide, we developed reaction conditions
in which an oxygen nucleophile (potassium benzoate in the
presence of benzoic acid) favors ribose attack over nucleobase
attack by 20-fold. This regioselectivity allows for efficient
synthesis of [2′-18O]uridine and its phosphoramidite from 2,2′-
cyclouridine and [18O2]benzoic acid. This approach may also
be applied for converting the 2,3′-cyclouridine to [3′-18O]uridine.
We can now construct RNA substrates containing [2′-18O]-
uridine isotopologues, thereby enabling isotope effect analyses
of protein and RNA enzymes that catalyze 2′-O-transphospho-
rylation reactions. Through known transformations of uridine,28
we may also access [2′-18O] isotope-enriched cytidine, adenos-
ine, and guanosine.
convert 2′-O-benzoyluridine to uridine, the 1H NMR spectrum
indicated that uridine and 1-(â-D-arabinofuranosyl)uracil formed
in a 5:1 ratio.
Using 2 equiv of KOBz instead of 1 equiv had little effect
on the reaction rate or the ratio of the ribo/arabino products.
However, we found that the presence of benzoic acid accelerated
the reaction and improved regioselectivity further.24 A mixture
of BzOH (1.0 equiv) and KOBz (1.0 equiv) at 140 °C in DMF
consumes 1 completely within 48 h to give the products in a
ratio of 20:1 ribo/arabino.25 The increased reaction rate and
regioselectivity upon addition of benzoic acid may reflect acid
catalysis, in which protonation of the imino nitrogen alters the
regiochemical outcome of the reaction.
We then exploited our observations to synthesize the target
isotopomer, 2 (Scheme 2). Acidic hydrolysis of benzonitrile in
H218O gave [18O2]benzoic acid containing 80% isotope enrich-
ment (MS).6 To avoid loss of 18O, we prepared a mixture of
1:1 [18O2]benzoic acid and potassium [18O2]benzoate by treating
[18O2]benzoic acid (2.0 equiv) with potassium hydride (1.0
equiv) in DMF. After addition of 1 (1.0 equiv), the mixture
was heated to 140 °C. After 4 days, TLC showed that 1 was
almost completely consumed. After removal of DMF under high
vacuum, we treated the residue with NaOMe/MeOH to convert
2′-[18O2]benzoyluridine to [2′-18O]uridine. After column chro-
matography, we obtained 2 in 88% yield (80% isotope enrich-
ment as determined by MS). As expected, the new compound
gave essentially the same 1H and 13C NMR spectra as an
authentic sample of uridine. In addition, we observed an 18O-
induced 13C NMR shift26 of 1.54 Hz upfield for 2′-C of [2′-
18O]uridine, further confirming the position of 18O.7a
Experimental Section
[2′-18O]Uridine (2): To a pressure tube (35 mL) under argon
were added anhydrous DMF (20 mL), KH (35%, 457 mg, 4 mmol),
and [18O2]benzoic acid (1.01 g, 8 mmol, 2 equiv). After the reaction
was stirred for 10 min at rt, 2,2′-cyclouridine (904 mg, 4 mmol, 1
equiv) was added. The mixture was heated to 140 °C for 4 days.
The reaction mixture was concentrated to dryness under vacuum.
The residue was dissolved in methanol (20 mL), and sodium
methoxide in methanol (30%, 1.52 mL, 2 equiv) was added. The
mixture was stirred overnight at rt, and acetic acid (690 µL, 3 equiv)
was added. The mixture was stirred at rt for an additional 10 min
and concentrated to dryness under vacuum. The residue was purified
by silica gel chromatography, eluting with 10% methanol in ethyl
acetate, to give 2 (859 mg, 88%) as a white solid: 1H NMR (500
MHz, DMSO-d6) δ 11.30 (br, 1H), 7.88 (d, J ) 8.2 Hz, 1H), 5.78
(m, 1H), 5.65 (d, J ) 8.2 Hz, 1H), 5.36 (br, 1H), 5.07 (br, 2H),
4.01 (m, 1H), 3.95 (m, 1H), 3.83 (m, 1H), 3.60 (m, 1H), 3.55 (m,
1H); 13C NMR (125.8 Hz, DMSO-d6) δ 163.5, 151.1, 141.1, 102.1,
88.0, 85.2, 73.9, 70.2, 61.2; HRMS calcd for C9H12N2O518O, [MH+]
247.0816, found 247.0818.
Acknowledgment. Q.D. is a Research Assistant Professor
at the University of Chicago, and J.A.P. is an Investigator of
the Howard Hughes Medical Institute. We thank the following
members of the Piccirilli laboratory for critical comments on
the manuscript: C. Lea, R. Fong, T. Novak, R. Sengupta, K.
Sundaram, J. Ye, Dr. N.-S. Li, Dr. J. Lu, Y. Koldobskaya, S.
Koo, and J. Min. We also thank Dr. N.-S. Li for oligo synthesis.
We transformed [2′-18O]uridine to the corresponding phos-
phoramidite 3 using standard methods (see Supporting Informa-
tion).27 We used 3 to incorporate [2′-18O]uridine into two
oligonucleotides, [2′-18O]UG and [2′-18O]UGGGUCGGC, ac-
cording to standard RNA synthesis protocols. After deprotection
and purification by reversed-phase HPLC, MALDI-TOF MS
confirmed the masses of the oligonucleotides, giving the
expected [M - H]- peaks at m/z 590 and 2883, while the wild-
Supporting Information Available: A scheme and experimen-
1
tal details for synthesis of phosphoramidite 3; H and 13C NMR
spectra for compounds 2 and 3. This material is available free of
charge via the Internet at http://pubs.acs.org.
(24) This reactivity trend should be viewed with caution, as it is based
on a limited number of independent examples rather than systematic
investigation of factors such as solvent, pH, temperature, bases, and the
“chemical context” of the nucleophilic atom.
JO701727H
(25) In the presence of 2 equiv of BzOH (without KOBz), the reaction
produced only a trace amount of product as indicated by TLC, presumably
due to the weak nucleophilicity of benzoic acid.
(26) Risley, J. M.; Van Etten, R. L. J. Am. Chem. Soc. 1979, 101, 252.
(27) (a) Hakimelahi, G. H.; Proba, Z. A.; Ogilvie, K. K. Tetrahedron
Lett. 1981, 22, 4775. (b) Moyroud, E.; Biala, E.; Strazewski, P. Tetrahedron
2000, 56, 1475. (c) Milecki, J.; Zamaratski, E.; Maltseva, T. V.; Foldesi,
A.; Adamiak, R. W.; Chattopadhyaya, J. Tetrahedron 1999, 55, 6603.
(28) For conversion of uridine to cytidine, see: (a) Saladino, R.; Crestini,
C.; Bernini, R.; Frachey, G.; Mincione, E.; J. Chem. Soc., Perkin Trans. 1
1994, 21, 3053. For conversion of uridine to adenosine, see: (b) Trelles, J.
A.; Valino, A. L.; Runza, V.; Lewkowicz, E. S.; Iribarren, A. M., Biotechnol.
Lett. 2005, 27, 759. (c) Trelles, J. A.; Fernandez, M.; Lewkowicz, E. S.;
Iribarren, A. M.; Sinisterra, J. V.; Tetrahedron Lett. 2003, 44, 2605. For
conversion of uridine to guanosine, see: (d) Mikami, Y.; Matsumoto, S.;
Hayashi, Y.; Sato, T. Jpn. Kokai Tokkyo Koho, JP 2001269192 A, 2001.
J. Org. Chem, Vol. 73, No. 1, 2008 311