2′-C-branched ribonucleosides. 2. Synthesis of 2′-C-β-trifluoromethyl pyrimidine ribonucleosides

Article abstract of DOI:10.1021/ol0155687

(matrix presented) The first synthesis of 2′-C-β-trifluoromethyl pyrimidine ribonucleosides is described. 1,2,3,5-Tetra-O-benzoyl-2-C-β-trifluoromethyl-α-D-ribofuranose (3) is prepared from 1,3,5-tri-O-benzoyl-α-D-ribofuranose (1) in three steps and converted to 3,5-di-O-benzoyl-2-C-β-trifluoromethyl-α-D-1-ribofuranosyl bromide (5). The 1-bromo derivative (5) is found to be a powerful reaction intermediate for the synthesis of ribonucleosides. The reaction of silylated pyrimidines with (5) in the presence of HgO/HgBr₂ affords exclusively the β-anomers (6-8). Deprotection of (6-8) with ammonia in methanol yields the 2′-C-β -trifluoromethyl nucleosides (9-11).

Full text of DOI:10.1021/ol0155687

ORGANIC
LETTERS
2001
Vol. 3, No. 7
1025-1028
2′-C-Branched Ribonucleosides. 2.
Synthesis of 2′-C-â-Trifluoromethyl
Pyrimidine Ribonucleosides1
Nan-Sheng Li, Xiao-Qing Tang, and Joseph A. Piccirilli*
Howard Hughes Medical Institute, Department of Biochemistry and Molecular Biology
and Department of Chemistry, The UniVersity of Chicago, 5841 S. Maryland AVenue,
MC 1028, Chicago, Illinois 60637
jpicciri@midway.uchicago.edu
Received January 17, 2001
ABSTRACT
The first synthesis of 2′-C-â-trifluoromethyl pyrimidine ribonucleosides is described. 1,2,3,5-Tetra-O-benzoyl-2-C-â-trifluoromethyl-r-D-ribofuranose
(3) is prepared from 1,3,5-tri-O-benzoyl-r-D-ribofuranose (1) in three steps and converted to 3,5-di-O-benzoyl-2-C-â-trifluoromethyl-r-D-1-
ribofuranosyl bromide (5). The 1-bromo derivative (5) is found to be a powerful reaction intermediate for the synthesis of ribonucleosides. The
reaction of silylated pyrimidines with (5) in the presence of HgO/HgBr2 affords exclusively the â-anomers (6−8). Deprotection of (6−8) with
ammonia in methanol yields the 2′-C-â-trifluoromethyl nucleosides (9−11).
2′-C-Branched nucleosides exhibit anticancer2 and antiviral3
activities, as well as inhibitory activity against several
enzymes.4 For example, 2′-C-â-methyladenosine resists the
action of adenosine deaminase and inhibits the growth of
KB cells in culture.5 2′-C-â-Trifluoromethyl nucleosides may
be particularly interesting in this regard, as the trifluoro-
methyl group can enhance the therapeutic properties of
bioactive compounds.6 Additionally, 2′-â-trifluoromethyl
ribonucleosides may provide important tools for the analysis
of RNA structure and function. Specifically, a series of
ribonucleosides analogues in which the 2′-â-substituent on
the ribose ring is CH3, CH2F, CHF2, or CF3 might allow a
systematic variation of the pKa of the 2′-hydroxyl group over
a broad range and within a similar structural context. For
those enzymes and ribozymes that catalyze RNA strand
scission by activating the 2′-hydroxyl group for attack at the
adjacent 3′-phosphate diester (such as ribonuclease A or the
hammerhead and hairpin ribozymes), this series of nucleo-
sides could reveal a linear free relationship between the
catalytic rate and the pKa of the nucleophile, thereby
providing information about the degree of bond-making
between the phosphorus and the nucleophile in the transition
state. No existing series of nucleoside analogues possesses
these properties. So far, the 2′-â-CH3 derivative is the only
member of the series that has been synthesized.7 The
synthesis of 2′-C-â-trifluoromethyl ribonucleosides is par-
(1) For the initial publication in the series, see: Tang, X.-Q.; Liao, X.-
M.; Piccirilli, J. A. J. Org. Chem. 1999, 64, 747.
(2) (a) Yoshimura, Y.; Satoh, H.; Sakata, S.; Ashida, N.; Miyazaki, S.;
Matsuda, A. Nucleosides Nucleotides 1995, 14, 427 and references therein.
(b) Cory, A. H.; Samano, V.; Robins, M. J.; Cory, J. P. Biochem. Pharmacol.
1994, 47, 365.
(3) Awano, H.; Shuto, S.; Baba, M.; Kira, T.; Shigeta, S.; Matsuda, A.
Bioorg. Med. Chem. Lett. 1994, 4, 367.
(4) (a) Baker, C. H.; Banzon, J.; Bollinger, J. M.; Stubbe, J.; Samano,
V.; Robin, M. J.; Lippert, B.; Jarvi, E.; Resvick, R. J. Med. Chem. 1991,
34, 1879. (b) Ong, S. P.; McFarlan, S. C.; Hogenkamp, H. P. C. Biochemistry
1993, 32, 11397. (c) van der Donk, W. A.; Yu, G.; Silva, D. J.; Stubbe, J.;
McCarthy, J. R.; Jarvi, E. T.; Matthews, D. P.; Resvick, R. J.; Wagner, E.
Biochemistry 1996, 35, 8381.
(5) (a) Walton, E.; Jenkins, S. R.; Nutt, R. F.; Zimmerman, M.; Holly,
F. W. J. Am. Chem. Soc. 1966, 88, 4524. (b) Walton, E.; Jenkins, S. R.;
Nutt, R. F.; Holly, F. W.; Nemes, M. J. Med. Chem. 1969, 12, 306.
(6) (a) Welch, J. T. SelectiVe Fluorination in Organic and Bioorganic
Chemistry; American Chemical Society: Washington, DC, 1991; pp 3-15.
(b) Welch, J. T.; Eswarakrishnan, S. Fluorine in Bioorganic Chemistry; J.
Wiley and Sons: New York, 1991. (c) Hudlicky, M.; Pavlath, A. E.
Chemistry of Organic Fluorine Compounds II; American Chemical
Society: Washington, DC, 1995.
(7) Harry-O’kuru, R. E.; Smith, J. M.; Wolfe, M. S. J. Org. Chem. 1997,
62, 1754.
10.1021/ol0155687 CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/08/2001

ticularly challenging because of the electron-withdrawing and
sterically demanding character of the 2′-CF3 group.8,9 Here
we describe the first synthesis of 2′-C-â-trifluoromethyl
ribonucleosides.
Perbenzoylated 2′-C-â-trifluoromethyl-R-D-ribofuranose 3
was synthesized as shown in Scheme 1. Oxidation of 1,3,5-
The Hilbert-Johnson glycosylation reaction of pyrimdines
with peracylated ribose usually proceeds efficiently and
stereoselectively at room temperature to yield the â-anomer.13
However, the glycosylation of bis(trimethylsilyl)uracil with
3 in the presence of either trifluoromethylsilyl trifluo-
romethanesulfonate or SnCl4 at room temperature failed to
give any 2′-C-trifluoromethyluridine. Even after these reac-
tions were refluxed for 1 week in acetonitrile (83 °C), less
than 5% of the nucleoside was obtained, though the â-anomer
4b was formed exclusively. Heating the reaction mixture in
a sealed heavy-wall pressure tube at 120 °C for 3 days in
the presence of trifluoromethylsilyl trifluoromethanesulfonate
improved the yield of nucleoside 4 to 33% and allowed
greater recovery of starting material 3 (57%), but both
anomers were formed (4a/4b ) 4:96).14 At 140 °C for 3
days, the yield of 4 increased to 56% yield, but both the
stereoselectivity (4a/4b ) 14:86) and the recovery of starting
material (13%) decreased. At 180 °C, the reaction went to
completion more rapidly (<24 h); however, both the yield
(21%) and the stereoselectivity decreased significantly (4a/
4b ) 40:60). The stereochemical assignment of 4a and 4b
as the R and â anomers, respectively, was determined by
the 19F-1H NOE experiments. For 4a, we recorded strong
NOE signals for 1′-H (δ 6.87) and 3′-H (δ 6.40) when the
19F nuclei were irradiated. For 4b, we observed strong NOE
signals for 3′-H (δ 6.23) and 6-H (δ 7.50) upon irradiation
of the 19F nuclei.
Scheme 1a
a (a) Dess-Martin periodinane, CH2Cl2, rt, 24 h; (b) (i) CF3SiMe3,
TBAF(5%), THF, rt, 18 h; (ii) TBAF, rt, 10 min; (iii) BzCl, DMAP,
Et3N, CH2Cl2, rt, 5 h; (c) bis(trimethylsilyl)uracil, TMSOTf,
CH3CN, 120 °C, 3 days, 4a/4b ) 4:96.
The glycosylation reaction with 3 required unusually high
temperatures possibly because the CF3 group hinders the
approach of nucleophile9 and destabilizes the intermediate
carbocation at C-1, both by electron withdrawal directly
through the σ framework and by weakening the ability of
the C-2 benzoyl group to donate electron density to C-1.
We explored therefore the possibility of converting the 1-O-
benzoyl derivative 3 to the more reactive 1-bromo derivative.
2,3,5-Tri-O-acyl-D-1-ribofuranosyl halides have been widely
investigated as glycosylating agents for the synthesis of
ribonucleosides and generally react to give â-anomers
exclusively.15 Giese et al. prepared 2,3,5-tri-O-benzoyl-D-1-
ribofuranosyl bromide by treating 1,2,3,5-tetra-O-benzoyl-
R-D-ribofuranose in CH2Cl2 with 33% HBr in acetic acid (0
°C to room temperature).16 In contrast, 3 was completely
unreactive under these conditions, even after the mixture was
stirred at room temperature for 9 days. However, heating 3
in a solution of 30% HBr in acetic acid at 80-85 °C for 5
h installed a bromine atom at C-1 but removed the 2-O-
benzoyl group to give 3,5-di-O-benzoyl-2-C-â-trifluoro-
tri-O-benzoyl-R-D-ribofuranose 1 with Dess-Martin perio-
dinane gave pure 1,3,5-tri-O-benzoyl-R-D-2-ketoribofuranose
2 via a modified procedure in 97% yield.10 Nucleophilic
trifluoromethylation of 2 with Ruppert’s reagent (CF3SiMe3)
catalyzed by TBAF (5%) in THF11 followed by desilylation
with stoichiometric TBAF and treatment with benzoyl
chloride/DMAP/Et3N afforded 1,2,3,5-tetra-O-benzoyl-2-C-
â-trifluoromethyl-R-D-ribofuranose 3 in 73% overall yield.
19F-1H NOE experiments indicated that the trifluoromethyl
nucleophile added stereoselectively to the â-face of the
sugar.12 When the 19F nuclei of the trifluoromethyl group
were irradiated, we observed strong NOE signals for 1-H (δ
7.42) and 3-H (δ 6.17), suggesting that 1-H, 3-H, and the
CF3 group are on the same side of the ribofuranosyl ring.
(8) It has been reported that the reactions at C-1 of carbohydrates via
cationic intermediates are much more difficult to achieve when a trifluo-
romethyl group is attached to C-2. Logothetis, T. A.; Eilitz, U.; Hiller, W.;
Burger, K. Tetrahedron 1998, 54, 14023 and references therein.
(9) Seebach argued on the basis of van der Waal hemispheres that a
trifluoromethyl group is between two and three times larger than a methyl
group. See: Seebach, D. Angew Chem., Int. Ed. Engl. 1990, 29, 1320. Other
assessments suggest that the steric influence of the trifluoromethyl group
is closer to that of an isopropyl group or even a tert-butyl group. See: (a)
Smart, B. E.; Middleton, W. J. J. Am. Chem. Soc. 1987, 109, 4982. (b)
Mosher, H. S.; Stevenot, J. E.; Kimble, D. O. J. Am. Chem. Soc. 1956, 78,
4374. (c) Ramachandran, P. V.; Teodorovic, A. V.; Brown, H. C.
Tetrahedron 1993, 49, 1725.
(13) (a) Niedballa, U.; Vorbruggen, H. J. Org. Chem. 1974, 39, 3654,
3664, 3668, 3672. (b) Niedballa, U.; Vorbruggen, H. J. Org. Chem. 1976,
41, 2084. (c) Vorbruggen, H.; Hofle, G. Chem. Ber. 1981, 114, 1256.
(14) The anomers were separated with difficulty by silica gel chroma-
tography as described in Supporting Information.
(15) (a) Wittenburg, E. Chem. Ber. 1968, 101, 1095. (b) Winkley, M.
W.; Robins, R. K. J. Org. Chem. 1968, 33, 2822. (c) Winkley, M. W.;
Robins, R. K. J. Org. Chem. 1969, 34, 431. (d) Currie, B. L.; Robins, R.
K.; Robins, M. J. J. Heterocycl. Chem. 1970, 7, 323. (e) Ozaki, S.;
Watanabe, Y.; Hoshiko, T.; Fujisawa, H.; Uemura, A.; Ohrai, K. Tetrahe-
dron Lett. 1984, 25, 5061. (f). Kissman, H. M.; Pidacks, C.; Baker, B. R.
J. Am. Chem. Soc. 1955, 77, 18. (g) Kessman, H. M.; Child, R. G.; Weiss,
M. J. J. Am. Chem. Soc. 1957, 79, 1185.
(10) Cook, G. P.; Greenberg, M. M. J. Org. Chem. 1994, 59, 4704.
(11) Prakash, G. K. S.; Krishnamurti, R.; Olah, G. A. J. Am. Chem. Soc.
1989, 111, 393.
(12) We synthesized 2′-C-R-trifluoromethyl-â-D-uridine by the nucleo-
philic addition of trimethyltrifluoromethylsilane to 3′,5′-O-(di-tert-butylsi-
1
lyldiyl)-2′-keto-â-D-uridine. Its H, 13C, and 19F spectra are different from
those of the corresponding 2′-C-â-trifluoromethyl-â-D-uridine 10 described
(16) Koch, A.; Lamberth, C.; Wetterich, F.; Giese, B. J. Org. Chem.
1993, 58, 1083.
in this paper.
1026
Org. Lett., Vol. 3, No. 7, 2001

Table 1. Glycosylation of 4-N-Benzoyl Cytosine with 5 under Various Conditions
entry
1
reagents
temp (°C)
80
time (h)
8
yield (%) of 6
(i) 5, DMAP, Et3N, PhCOCl, CH2Cl2;
0
(ii) persilylated base (2.0 equiv), HgO (1.3 equiv), HgBr2 (0.8 equiv), benzene
5, persilylated base (2.7 equiv), HgO (1.3 equiv), HgBr2 (0.8 equiv), toluene
5, persilylated base (1.9 equiv), HgO (1.3 equiv), HgBr2 (0.8 equiv), benzene
5, base (2.0 equiv), HgO (1.3 equiv), HgBr2 (0.8 equiv), toluene
5, persilylated base (2.2 equiv), toluene
5, persilylated base (3.0 equiv), HgO (1.5 equiv), benzene
5, persilylated base (2.7 equiv), Hg(OAc)2 (0.9 equiv), toluene
(i) 5, HgBr2 (1.25 equiv), benzene, reflux, 1 h;
2
3
4
5
6
7
8a
110
25
110
110
80
10
62
16
52
14
10
19
48
0
0
0
24
32
48
110
80
(ii) persilylated base (2.4 equiv), benzene
a The product contained a 2′-O-trimethylsilyl group, which could be removed by TMAF to give 6 in quantitative yield.
methyl-R-D-1-ribofuranosyl bromide 5 in 77% yield. The
stereochemistry of 5 was assigned by 19F-1H NOE experi-
ments. We observed strong NOE signals for 1-H (δ 6.94)
and 3-H (δ 5.68) when the 19F nuclei were irradiated.
Because the 2-O-benzoyl group directs nucleophiles to the
opposite face of the sugar during glycosylation reactions,
we reintroduced it. Treatment of 5 with benzoyl chloride in
the presence of DMAP and triethylamine gave quantitatively
2,3,5-tri-O-benzoyl-2-C-â-trifluoromethyl-R-D-1-ribofurano-
syl bromide. Unfortunately, 2,3,5-tri-O-benzoyl-2-C-â-tri-
fluoromethyl-R-D-1-ribofuranosyl bromide was ineffective
as a glycosylating agent; reaction in the presence of HgO/
HgBr215a with nucleobases such as persilylated 4-N-benzoyl
cytosine gave only a trace of product after 8 h at reflux in
benzene (Table 1, entry 1). To our surprise, however, the
reaction of 5, which lacks the 2-benzoyl group, with
persilylated 4-N-benzoyl cytosine gave exclusively the cor-
responding â-nucleoside derivative, 3′,5′-di-O-benzoyl-4-N-
benzoyl-2′-C-â-trifluoromethyl-â-D-cytidine 6 in 48% yield
(Table 1, entry 2). Although 3,5-di-O-benzoyl-D-1-ribofura-
nosyl chloride and bromide have been prepared previously,17
to the best of our knowledge these glycosyl halides have
not been used for the synthesis of nucleosides.18 Conse-
quently, we studied the conversion of 5 to 6 under a variety
of reaction conditions (Table 1).
tetramethylammonium fluoride to give 6 in quantitative yield
(Table 1, entry 8). These results suggest that transformation
of 5 to 6 may occur via an SN2 reaction with the assistance
of mercury(II) (Scheme 2).
Scheme 2
Direct coupling of crude intermediate 5 (generated in situ
from perbenzoylated 3) with persilylated 4-N-benzoylcy-
tosine, uracil, or thymine in the presence of HgO/HgBr2 gave
exclusively the â-nucleoside derivatives 6, 7, and 8 in yields
of 42%, 47%, and 57%, respectively (Scheme 3). The
Scheme 3a
The reaction requires silylation of the nucleobase and the
presence of mercury(II), as no reaction occurs with the free
nucleobase (Table 1, entry 4) or in the absence of the
mercury(II) salt (Table 1, entry 5). Mild heating (refluxing
benzene or toluene) also was required (Table 1, entry 3).
Interestingly, after 1 h of reflux in benzene in the presence
of mercury(II) bromide, 5 remained unchanged in the reaction
mixture. However, subsequent addition of the persilylated
nucleobase gave in 48% yield the corresponding 2′-O-
trimethylsilylated product, which can be desilylated by
a 30% HBr in AcOH, 80-85 °C, 5 h; (b) persilylated nucleo-
bases, HgO/HgBr2, benzene, or toluene, reflux; (c) NH3, MeOH, 4
°C, 2 days.
(17) (a) Gorin, P. A. J. Can. J. Chem. 1962, 40, 276. (b) Ness, R. K.;
Fletcher, H. G., Jr. J. Am. Chem. Soc. 1956, 78, 4710.
(18) A few papers report the reaction of 3,5-di-O-acyl-D-1-ribofuranosyl
chloride with reactive pyridine derivatives but not with nucleobases. See:
(a) Konno, K.; Hayano, K.; Shirahama, H.; Saito, H.; Matsumoto, T.
Tetrahedron 1982, 38, 3281. (b). Freyne, E. J.; Esmans, E. L.; Lepoivre, J.
A.; Alderweireldt, F. C. Carbohydr. Res. 1980, 78, 235. (c) Alderweireldt,
F. C.; Kemps, L.; Esmans, E. L.; Dommisse, R. A.; Lepoivre, J. A.
Heterocycles 1984, 21, 795.
stereochemistries of 6-8 were confirmed by 19F-1H NOE
experiments. For 6, we observed strong NOE signals for 3′-H
Org. Lett., Vol. 3, No. 7, 2001
1027

(δ 5.85) and 6-H (δ 8.04) upon irradiation of the 19F nuclei.
For 7 and 8, we also observed strong NOE signals for 3′-H
(δ 5.96), 6-H (δ 7.55) of 7 and 3′-H (δ 5.98) and 6-H (δ
7.30) of 8 upon irradiation of the 19F nuclei.
Debenzoylation of 6, 7, and 8 with ammonia in methanol
at 4 °C for 2 days provided the free 2′-C-â-trifluorormethyl
ribonucleosides 9, 10, 11 in yields of 91-99%. CD spec-
troscopy also confirmed the anomeric configuration of each
nucleoside. The CD spectra of 9-11 show a positive cotton
effect from 260 to 280 nm as expected for a pyrimidine
nucleosides in the â-configuration;19 the R-anomers show a
negative cotton effect in the same range.
achieved the first synthetic entry into this potentially
important class of nucleosides, thereby allowing investigation
of their biological and biochemical properties. We plan to
convert these trifluoromethyl ribonucleosides to their phos-
phoramidite derivatives for site-specific incorporation into
oligonucleotides via solid-phase synthesis.
Acknowledgment. N.S.L is a Research Specialist, X.Q.T.
was a Research Associate, and J.A.P. is an Associate
Investigator of the Howard Hughes Medical Institute. We
thank Dr. A. Bates for help in obtaining NOE spectra and
Q. Dai, S. Das, M. Hamm, and J. Schwans for critical
comments on the manuscripts.
In summary, 2′-C-â-trifluoromethyl ribonucleosides are
challenging compounds to synthesize because of the cis
relationship between the heterocycle and the large, electron-
withdrawing CF3 group. By discovering that 5 effectively
and stereoselectively glycosylates pyrimidines, we have
Supporting Information Available: Full experimental
and analytical data for compounds 2, 3, 4a,b, and 5-11;
1
19F-1H NOE spectra for 3, 4a,b, and 5-8, and H and 13C
NMR spectra of 9-11. This material is available free of
charge via the Internet at http://pubs.acs.org.
(19) Miles, D. W.; Inskeep, W. H.; Robins, M. J.; Winkley, M. W.;
Robins, R. K.; Eyring, H. J. Am. Chem. Soc. 1970, 92, 3872.
OL0155687
1028
Org. Lett., Vol. 3, No. 7, 2001