An efficient and diastereoselective synthesis of PSI-6130: A clinically efficacious inhibitor of HCV NS5B polymerase

Article abstract of DOI:10.1021/jo901345j

(Chemical Equation Presented) R7128 is the prodrug of 2′-deoxy- 2′-fluoro-2′-C-methylcytidine (PSI-6130), a potent and selective inhibitor of HCV NS5B polymerase. Currently, R7128 is in clinical trials for the treatment of HCV infection. To support clinical development efforts, we needed an efficient and scalable synthesis of PSI-6130. We describe an improved, diastereoselective synthetic route starting with protected D-glyceraldehyde. No chiral reagents or catalysts were used to produce the three new contiguous stereocenters. Introduction of fluorine at the C-2 tertiary carbon was accomplished in a highly regio- and stereoselective manner through nucleophilic substitution on a cyclic sulfate. Scale-limiting chromatographic purifications were eliminated through the use of crystalline intermediates.

Full text of DOI:10.1021/jo901345j

pubs.acs.org/joc
An Efficient and Diastereoselective Synthesis of PSI-6130: A Clinically
Efficacious Inhibitor of HCV NS5B Polymerase
Peiyuan Wang, Byoung-Kwon Chun, Suguna Rachakonda, Jinfa Du, Noshena Khan,
Junxing Shi,† Wojciech Stec,‡ Darryl Cleary, Bruce S. Ross,* and Michael J. Sofia
Pharmasset, Inc., 303A College Road E., Princeton, New Jersey 08540. †Present address: RFS Pharma, 186
Montreal Rd, Tucker, GA 30084. ‡Present address: Polish Academy of Sciences CMMS, Sienkiewicza 112,
90-363 Lodz, Poland.
bruce.ross@pharmasset.com
Received June 24, 2009
R7128 is the prodrug of 20-deoxy-20-fluoro-20-C-methylcytidine (PSI-6130), a potent and selective
inhibitor of HCV NS5B polymerase. Currently, R7128 is in clinical trials for the treatment of HCV
infection. To support clinical development efforts, we needed an efficient and scalable synthesis of
PSI-6130. We describe an improved, diastereoselective synthetic route starting with protected
D-glyceraldehyde. No chiral reagents or catalysts were used to produce the three new contiguous
stereocenters. Introduction of fluorine at the C-2 tertiary carbon was accomplished in a highly regio-
and stereoselective manner through nucleophilic substitution on a cyclic sulfate. Scale-limiting
chromatographic purifications were eliminated through the use of crystalline intermediates.
Introduction
Over the last several years, ribonucleoside analogues with
a 20-C-methyl substituent have been identified as potent and
selective inhibitors5 of NS5B polymerase of the HCV repli-
cation complex. Clark et al.6 disclosed the synthesis
and biological activity of a fluorinated cytidine analogue,
20-deoxy-20-fluoro-20-C-methylcytidine(PSI-6130,1,Figure1).
On the basis of its superior profile, PSI-6130 was selected for
clinical development. Currently its diisobutyryl prodrug
(R7128, 2) is in phase IIb clinical trials for Hepatitis C.
R7128 has demonstrated antiviral activity in genotype 1
nonresponders with 14 day monotherapy.7 In combination
with the standard of care over 28 days, treatment with R7128
led to undetectable viral levels in 88% of genotype 1 naive
patients and 90% undetectable viral levels in genotype 2 and
3 nonresponder patients.7 R7128 is the first direct acting
It is estimated that 130 million people1 worldwide are
infected with the Hepatitis C virus (HCV). Chronic infection
greatly increases the risk of hepatic carcinoma.2 In the
United States, there are estimated to be 3.9 million people
infected with HCV and 8 000-10 000 annual deaths attrib-
uted to HCV infection.3 The current standard of care,
pegylated interferon and ribavirin, produces viral response
rates in only 50% of the patients infected with the genotype 1
virus.4 Clearly, there is an unmet medical need for more
effective therapies.
*To whom correspondence should be addressed. Phone: 609-613-4117.
Fax: 609-613-4150.
(1) Alter, M. J. World J. Gastroenterol. 2007, 13 (17), 2436–41.
(2) (a) Tsukuma, H.; Hiyama, T.; Tanaka, S.; Nakao, M.; Yabuuchi, T.;
Kitamura, T.; Nakanishi, K.; Fujimoto, I.; Inoue, A.; Yamazaki, H.;
Kawashima, T. New Engl. J. Med. 1993, 328 (25), 1797–1801. (b) McHutchison,
J. G. Am. J. Manag. Care 2004, 10, S21–S29.
(3) Alter, M. Hepatology 1997, 26, 62S–65S.
(4) Poynard, T.; Yuen, M.-F.; Ratziu, V.; Lai, C. L. Lancet 2003, 362,
2095–2100.
(6) Clark, J. L.; Hollecker, L.; Mason, J. C.; Stuyver, L. J.; Tharnish,
P. M.; Lostia, S.; McBrayer, T. R.; Schinazi, R. F.; Watanabe, K. A.; Otto,
M. J.; Furman, P. A.; Stec, W. J.; Patterson, S. E.; Pankiewicz, K. W. J. Med.
Chem. 2005, 48, 5504–5508.
(5) (a) Carroll, S. S.; et al. J. Biol. Chem. 2003, 278, 11979–11984.
(b) Pierra, C.; et al. J. Med. Chem. 2006, 49 (22), 6614–6620. (c) Gunic, E.;
Chow, S.; Rong, F.; Ramasamy, K. S.; Raney, D.; Li, J.; Huang, R.;
Hamatake, R.; Hong, Z.; Girardet, J. L. Bioorg. Med. Chem. Lett. 2007,
17 (9), 2456–2458.
(7) (a) Lalezari, J.; Gane, E.; Rodriguez-Torres, M.; DeJesus, E.; Nelson, D.;
Everson, G.; Jacobson, I.; Reddy, R.; Hill, G. Z.; Beard, A.; Symonds, W. T.;
Berrey, M. M.; McHutchison, J. G. 43rd Annual Meeting of EASL, Milan, Italy,
April 23-27, 2008. (b) Gane, E. et al. The 59th Annual Meeting of the American
Association for the Study of Liver Disease, November 2008, San Francisco, CA.
DOI: 10.1021/jo901345j
r
Published on Web 07/30/2009
J. Org. Chem. 2009, 74, 6819–6824 6819
2009 American Chemical Society

Wang et al.
JOCArticle
SCHEME 1. Retrosynthetic Analysis of PSI-6130
FIGURE 1. Structures of PSI-6130 and R7128.
antiviral to demonstrate potent activity in the clinic against
both genotype 2 and 3 virus.
Cubero, and Olea11 described the synthesis of ethyl 4,5-O-
isopropylidine-2-C-methyl-DL-arabinoate via osmium tetra-
oxide mediated dihydroxylation of a pentenoate precursor
that met our needs. In their earlier synthesis of the penteno-
ate,12 they used a Knoevenagel-Doebner condensation on
isopropylidene protected D-glyceraldehyde. The alkaline
conditions led to racemization of the stereocenter. A Wittig
reaction with an isolated ylide did not lead to any racemiza-
tion. They also observed high selectivity for the desired
E geometric isomer under their reaction conditions.
Proceeding forward with our synthesis of the ribonolac-
tone intermediate (Scheme 2), commercially available iso-
propylidine protected D-glyceraldehyde (6) can be made
economically on scale by symmetrical oxidative cleavage of
the diisopropylidine derivative of D-mannitol.13,14 Treating 6
with commercially available (carbethoxyethylidene)triphen-
ylmethylphosphorane gave the desired chiral pentenoate
ester (7) in 79% crude yield (ratio E/Z 97:3 by NMR). The
mixture was carried on to the next step without further
purification
Osmium tetraoxide can approach the double bond of the
pentenoate from either face and therefore create two dia-
stereomers for each of the geometric isomers. Cubero11
observed that treating purified geometric DL isomers with
catalytic osmium tetraoxide and potassium chlorate resulted
in roughly equal facial selectivity for the Z-isomer and 2:1 for
desired arabinoate from the E-isomer. In our early work, we
used Sharpless AD-mix-β catalyst15 to direct the attack to
the desired face and create the correct stereochemistry in
91% yield after chromatography. While this worked well, we
wanted to avoid using an expensive catalyst early in the
synthesis. Even without using a chiral catalyst, the reaction
on allylic alcohol derivatives should mostly occur on the face
opposite to the pre-existing hydroxyl or alkoxyl groups.16
After trying several oxidative conditions, we found that tert-
butyl hydroperoxide with catalytic osmium tetraoxide in an
tetraethylammonium acetate buffer17 worked exceptionally
Although structure 1 appears as a close analogue of the
natural ribonucleoside cytidine, it presented a synthetic
challenge to functionalize stereoselectively the C-20 position
to generate a quaternary center having 20-C-methyl and 20-
fluoro substituents. The discovery synthesis6 converted cyti-
dine to its 20-C-methyl arabinoside analogue in six steps
by using standard methods. Nucleophilic fluorination of the
C-20 tertiary alcohol with (diethylamino)sulfur trifluoride
(DAST) did result in inversion of configuration to give the
desired stereochemistry, but not surprisingly, the yield was
low due to the formation of elimination and hydrolysis
byproduct. Difficult chromatographic separation of the
mixture limited the scale as well. An alternative synthesis
of 1 from D-xylose proceeded via benzoyl protected 2-deoxy-
2-fluoro-2-C-methyl-ribonolactol, using again a low-yield-
ing DAST conversion of the tertiary alcohol, followed by
coupling with the cytosine base to give the benzoylated
cytidine precursor of 1.8 More recently, Mayes and Moussa
disclosed9 a synthesis of the sugar moiety by converting 3,4-
O-isopropylidene-2-C-methyl-D-arabinono-1,5-lactone to
its corresponding triflate, followed by fluoride displacement
with tris(dimethylamino)sulfur trimethylsilyl difluoride
(TASF) and deprotection to the six-membered-ring 1,
5-ribonolactone. This was subsequently converted to the
five-membered-ring 1,4-ribonolactone. From a process
chemistry standpoint, all of these routes have major limita-
tions. Most notably, the key fluorination step comes too late
in each of the synthetic schemes. This low-yielding step is
done on advanced and therefore more expensive intermedi-
ates and the fluorinating agent, DAST, is hazardous to use
on scale. Undesirable multiple chromatographic separations
are required in the sequences and overall yields are low.
To produce enough material for preclinical and early
clinical development, we needed a new, efficient, and scalable
route that provided crystalline intermediates for purifica-
tion. Retrosynthetically, we desired a convergent synthesis
that would couple a fully elaborated sugar unit with the
requisite base (Scheme 1).
(11) Aparicio, F. J. L.; Cubero, I. I.; Olea, M. D. P Carbohydr. Res. 1984,
129, 99-109; 1983, 115, 250-253.
(12) Aparicio, F. J. L.; Cubero, I. I.; Olea, M. D. P. Carbohydr. Res. 1982,
103, 154–164.
Results and Discussion
The key intermediate needed was ribonolactone 3. We
envisioned10 introduction of the C-2 fluorine functionality
via stereoselective displacement of an activated hydroxyl
group on an acyclic intermediate (5) to form 4. Aparicio,
(13) Schmid, C. R.; Bryant, J. D.; Dowlatzedah, M.; Phillips, J. L.;
Prather, D. E.; Schantz, R. D.; Sear, N. L.; Vianco, C. S. J. Org. Chem.
1991, 56, 4056–4058.
(14) Schmid, C. R.; Bryant, J. D. Org. Synth. 1995, 72, 6-13.
(15) (a) Sharpless, K. B.; Amberg, W.; Bennani, Y. L.; Crispino, G. A.;
Hartung, J.; Kwong, H.-L.; Morikawa, K.; Wang, Z.-M.; Xu, D.; Zhang,
X.-L. J. Org. Chem. 1992, 57, 2768–2771. (b) Mokikawa, K.; Sharpless, K. B.
Tetrahedron Lett. 1993, 34 (35), 5575–5578.
(16) Christ, W. J.; Cha, J. K.; Kishi, Y. Tetrahedron Lett. 1983, 24 (37),
3947–3950.
(17) Katsuki, T.; Lee, A. W. M.; Ma, P.; Martin, V. S.; Masamune, K. B.;
Sharpless, K. B.; Tuddenham, D.; Walker, F. J. J. Org. Chem. 1982, 47, 1373.
(8) Clark, J. L.; Mason, J. C.; Hobbs, A. J.; Hollecker, L.; Schinazi, R. F.
J. Carbohydr. Chem. 2006, 25, 461–470.
(9) Mayes, B. A; Moussa, A. U.S. Pat. Appl. 2007/0203334.
(10) Wang, P.; Stec, W.; Clark, J.; Chun, B.-K.; Shi, J.; Du, J. Intl. Pat.
Appl. WO 2006/012440.
6820 J. Org. Chem. Vol. 74, No. 17, 2009

Wang et al.
JOCArticle
SCHEME 2. Synthesis of Protected Ribonolactone
well. The desired pure diol (8) was isolated in 87% yield after
chromatography. Due to the hazardous nature of osmium
tetraoxide, we also explored using permanganate salts for the
dihydroxylation. Typically, permanganate dihydroxylations
are performed in alkaline conditions to suppress overoxida-
tion. Due to the presence of an ester in our substrate, we
wanted to minimize exposure to strong base. We tried
conditions described by Hazra18 using tetradecyltrimethy-
lammonium permanganate (TDTAP), which included 0.1
equiv of potassium hydroxide. This produced 8 in an 8:1
mixture in 71% yield. The reaction also worked without the
addition of the base and gave a similar ratio and yield.
Potassiumpermanganate with 18-crown-6ether19 indichloro-
methane gave a 6:1 mixture in 94% yield. We found the
most practical conditions to be potassium permanganate in
acetone with or without 10% triethylamine to give a 12:1
ratio. After an aqueous workup, the solid residue was
recrystallized to give the pure D-isomer diol 8 in 67% yield
from the crude alkene mixture without any chromatography.
With the diol ester 8 in hand, our first approach10 was to
selectively benzoylate the secondary C-3 hydroxyl (71%) and
then use a DAST fluorination of the C-2 tertiary alcohol
(68%). The acyclic ester could then be cyclized to form the
desired ribonolactone (58%). Alternatively, the acyclic ester
could be first cyclized with acid to form the arabinolactone,
then selectively dibenzoylated on the C-3,5 (55% two-step
yield) and finally treated with DAST (51%) to afford the
same ribonolactone. Although these routes were an improve-
ment over the D-xylose route, we observed that the yields of
the DAST mediated fluorinations fell as the scale increased
and there were still multiple chromatographic separations
required. Consequently, we searched for alternative scalable
fluorination methods that could be used with 8. Gao and
Sharpless20 reported efficiently opening cyclic sulfates of 1,2
diols with nucleophiles including fluoride. Interestingly,
opening cyclic sulfates from diols derived from R,β-unsatu-
rated esters resulted in a high selectivity for the R-position.
Furthermore, Avenoza21 et al. observed that a cyclic sulfate
substituted with an ester and methyl on the R-carbon and
unsubstituted on the β-carbon was opened by azide at the
more hindered R-position in a 4.5:1 ratio over the β-position.
These observations led us to try this approach on 8.22
Although the cyclic sulfate 9 could be made directly with
sulfuryl chloride or 1,10-sulfonyldiimidazole, we found high-
er and more reproducible yields using the two-step method of
first making the cyclic sulfite with thionyl chloride and then
further oxidizing to the cyclic sulfate with catalytic ruthe-
nium trichloride and sodium periodate20 or with catalytic
TEMPO and sodium hypochlorite. After an aqueous work-
up, 9 could be used without further purification. We ex-
plored a number of nucleophilic fluorinating reagents and
conditions. In the tetraalkylammonium fluoride series, tetra-
butylammonium fluoride led to some hydrolysis of the ester.
Tetramethylammonium fluoride worked well, but solubility
limited the reaction concentration. Tetraethylammonium
fluoride hydrate in acetonitrile or dioxane worked best in
the series. No fluorinated C-3 regioisomer could be observed
by 1H NMR of the crude product 10. The reaction mixture
was used directly in the next step.
Hydrolysis of a sulfate ester in the presence of an ester and
acetonide was reported by Kim and Sharpless23 using cata-
lytic concentrated sulfuric acid and 0.5 to 1.0 equiv of water
in THF. In our case, the reaction did not proceed to
completion. We could perform a global deprotection and
lactonization by adding Dowex 50Wx2-200 resin to give
directly crude lactone 3, but the resulting product mixture
required chromatographic separation. We found that hydro-
lysis of the sulfate could be accomplished while maintaining
the acetonide by treating the crude reaction mixture with
concentrated hydrochloric acid in 2,2-dimethoxypropane.
After an aqueous wash that removed most of the impurities,
the resulting acyclic product 11 was dissolved in ethanol and
treated with concentrated hydrochloric acid. Upon concen-
tration, the lactone product 3 presented as a white solid in
67% yield from the diol 8. This was benzoylated in pyridine
and then diluted with water to give 12 as a solid in 70% yield.
(18) Hazra.; et al. J. Chem. Soc., Perkin Trans. I 1994, 1667.
(19) Tabusa, F.; Suzuki, K.; Mukaiyama, T. Chem. Lett. 1983, 173.
(20) Gao, Y.; Sharpless, K. B. J. Am. Chem. Soc. 1988, 110, 7538–7839.
(21) Avenoza, A.; Busto, J. H.; Corzana, F.; Garcia, J. I.; Peregrina, J. M.
J. Org. Chem. 2003, 68, 4506–4513.
(22) Chun, K.; Wang, P. Intl. Pat. Appl. WO 2006/031725.
(23) Kim, B. M.; Sharpless, K. B. Tetrahedron Lett. 1989, 30 (6), 655–658.
J. Org. Chem. Vol. 74, No. 17, 2009 6821

Wang et al.
JOCArticle
SCHEME 3. Synthesis of PSI-6130
There was some yield loss due to hydrolysis during the
isolation. Structure 12 could also be isolated chromatogra-
phically in 91% yield.
production route24 with the most notable improvement atz
the glycosylation step (59%) and an overall yield over 20%.
Reduction of ribonolactone 12 to the ribonolactol 13
(Scheme 3) required a hindered reducing reagent to minimize
over-reduction to the acyclic 1,4-diol. Titration with lithium
tri-tert-butoxyaluminum hydride maximized production of
13 in a 2:1 β/R anomeric ratio. Clark reported coupling the
1-O-benzoyl derivative of 13 with silylated N4-benzoylcyto-
sine using stannic chloride catalyst in acetonitrile to give an
even mixture of anomers. After screening a number of
coupling conditions and reagents, improved results were
obtained by the coupling the 1-O-acetate of ribonolactol
(14) and changing the solvent and temperature to chloro-
benzene at 65 °C to give a 4:1 ratio of the β/R anomers. A
possible explanation for the increased β selectivity in the
nonpolar chlorobenzene solvent is enhanced coordination of
the tin catalyst with the C-2 fluoro and C-1 chloro on the
R-face and thus favoring attack from the β-face. The β/R
anomeric ratio of 14 had no influence on the anomeric ratio
of the subsequent nucleosides under the Lewis acid mediated
coupling conditions. For the process run, the crude lactol 13
was converted to 14 and used directly for coupling to form
15. After filtering off the tin salts, the pure β-anomer crystal-
lized out of a methanolic solution of the reaction mixture in
29% yield from protected lactone 12. Deprotection of 12 to
the target nucleoside could be readily accomplished in
methanolic ammonia to give the product PSI-6130 (1) as a
pure solid in 88% yield. The resulting product was identical
in all respects with the discovery synthesis product.
Experimental Section
(2S,3R)-Ethyl 3-((R)-2,2-dimethyl-1,3-dioxolan-4-yl)-2,3-di-
hydroxy-2-methylpropanoate (8). (Carbethoxyethylidene)triph-
enylmethylphosphorane (66.2 g, 182.6 mmol) was dissolved in
anhydrous dichloromethane (250 mL) and the solution was
cooled to -40 °C with stirring under a nitrogen atmosphere.
(R)-(þ)-2,2-Dimethyl-1,3-dioxolane-4-carboxaldehyde (6, 25.0 g,
192.3 mmol) was dissolved in anhydrous dichloromethane (100 mL)
and added to the reaction solution over 20 min at -40 °C. The
stirred reaction mixture was allowed to warm to ambient tem-
perature for 17 h. The mixture was concentrated under reduced
pressure to dryness and the residue was suspended in tert-butyl
methyl ether (200 mL). The suspension was filtered to remove
triphenylphosphine oxide and the filtrate was concentrated under
reduced pressure to afford 25.8 g of crude alkene (7) in 97:3 E/Z
ratio by H NMR consistent with literature values.12 The crude
alkene 7 (15.0 g, 70.1 mmol) was dissolved in acetone (700 mL) at
0-5 °C. Potassium permanganate (13.24 g, 83.8 mmol) was added
in one portion. After being stirred at 0-5 °C for 5 h, the reaction
was quenched by the addition of saturated aq sodium sulfite
(150 mL). After 30 min, a colorless suspension formed. The solid
was removed by filtration andwashed withethyl acetate (200 mL).
The filtrate was extracted with ethyl acetate (3 ꢀ 75 mL). The
combined extracts were dried over sodium sulfate, filtered, and
concentrated under reduced pressure to give a white solid reside as
a 12:1 mixture of diol isomers by NMR. The residue was dissolved
in warm ethyl acetate (30 mL). Hexanes (120 mL) was added
slowly to form a precipitate. The solid was collected by filtration,
washed with hexanes (2 ꢀ 50 mL), and dried under vacuum
(0.2 mmHg, ambient temperature, 24 h) to give 11.2 g of white
crystalline solid 8 (11.2 g, 53% from the phosphorane) as a single
isomer by NMR, mp 75.0-75.5 °C. Optical rotation [R]25D þ37.8
(c 0.50, EtOH). 1H NMR (DMSO-d6) δ 1.18 (t, 3H, J=7.2 Hz,
CH2CH3), 1.23 (s, 3H, CH3), 1.24 (s, 3H, CH3), 1.28 (s, 3H, 2-
CH3), 3.66 (dd, 1H, J=7.4Hz, 3-H), 3.80-3.89 (m, 2H, 5-H), 4.05
(dt, 2H, J=7.2 and 2 Hz, CH2CH3), 4.11 (q, 1H, J=6.4 Hz, 4-H),
4.90 (s, 1H, 2-OH), 5.09 (d, 1H, J=7.4 Hz, 3-OH). 13C NMR
(DMSO-d6) δ 14.73, 22.74, 26.23, 27.11, 61.00, 66.51, 75.44, 75.48,
77.35, 108.52, 175.33. IR (neat, cm-1) 3271.45, 2984.72, 1724.80,
1456.12, 1370.25, 1292.13, 1254.14, 1207.58, 1141.85, 1043.18,
954.12, 895.63, 844.93. HR-MS calcd for C11H20NaO6 (M þ Na)
271.1160, obsd 271.1152.
Conclusion
In summary, we developed an efficient, diastereoselective,
and scalable synthesis of the nucleoside HCV NS5B inhibi-
tor, PSI-6130. We started from a protected three-carbon
synthon with a single chiral center and created three addi-
tional contiguous chiral centers without the use of chiral
reagents or catalysts. For the key C-20 fluorinated quatern-
ary center, we used a nucleophilic fluoride reagent to open a
cyclic sulfate in a highly regio- and stereoselective manner.
There were three solid intermediates that could be isolated
after multiple steps: diol 8 (53%), protected lactone 12
(47%), and protected product 15 (29%). No chromatogra-
phy was used throughout the process to produce pure final
product 1 in 6.4% overall yield from (carbethoxyethylid-
ene)triphenylmethylphosphorane. Further process develop-
ment of this route led to the multikilogram clinical material
((2R,3R,4R)-3-(Benzoyloxy)-4-fluoro-4-methyl-5-oxotetrahy-
drofuran-2-yl)methyl Benzoate (12) [Common Name: 3,5-Di-O-
benzoyl-2-deoxy-2-fluoro-2-C-methyl-D-ribono-γ-lactone]. To a
stirred solution of diol 8 (7.2 g, 29.0 mmol) in anhydrous
dichloromethane (80mL) and triethylamine (12.1 mL, 87.1 mmol)
at 0 °C was slowly added thionyl chloride (3.2 mL,
43.8 mmol). After 15 min, the reaction was diluted with
dichloromethane (100 mL) and washed with cold water (2ꢀ
50 mL) and brine (50 mL). The organic layer was concentrated
under reduced pressure to roughly one-third the volume and
(24) Axt, S.; Chun, B.-K.; Jin, Q.; Rachakonda, S.; Ross, B.; Sarma, K.;
Vitale, J.; Zhu, J. Intl. Pat Appl. WO 2008/045419.
6822 J. Org. Chem. Vol. 74, No. 17, 2009

Wang et al.
JOCArticle
then diluted with acetonitrile (100 mL). The solution was
cooled to 0 °C and then TEMPO catalyst (44 mg) was added.
A sodium hypochlorite solution (10-13% available chlorine,
79 mL) was added and the reaction was stirred vigorously at
0 °C for 20 min and then at ambient temperature for 1 h. The
upper organic layer was separated and dried over sodium
sulfate, filtered, concentrated under reduced pressure, coeva-
porated with dichloromethane (2 ꢀ 50 mL), and dried
(0.2 mmHg, ambient temperature, 17 h) to 8.2 g of crude
cyclic sulfate (9). In turn, this was dissolved in anhydrous
dioxane (100 mL). Tetraethylammonium fluoride hydrate
(6.5 g, 37 mmol) was added and the mixture was heated
to 100 °C for 1 h and then cooled to ambient temperature.
2,2-Dimethoxypropane (100 mL), followed by conc. aq
hydrochloric acid (6 mL) were added and the mixture was
stirred for 3 h at ambient temperature. The reaction was
diluted with ethyl acetate (100 mL) and then washed with
cold saturated aq sodium bicarbonate (2ꢀ50 mL) and brine
(2ꢀ50 mL). The combined aqueous layer was back-extracted
with ethyl acetate (50 mL). The combined organic layer was
dried over sodium sulfate, filtered, and concentrated under
reduced pressure to give the crude acyclic 11 as a semisolid
(5.1 g). This was dissolved in reagent ethanol (50 mL) and
conc. aq hydrochloric acid (1 mL). The resulting solution was
stirred at ambient temperature for 15 h and then concentrated
under reduced pressure followed by coevaporations with
toluene (5 ꢀ 15 mL) to give the unprotected lactone 3 as a
white solid (3.2 g). Compound 3 was combined with other lots
made in a similar way for a total of 15.0 g (91.5 mmol). This
was dissolved in anhydrous pyridine (150 mL). Benzoyl
chloride (42 mL, 362.16 mmol) was added slowly over 5 min
at 0-5 °C. The resulting reaction mixture was stirred at
ambient temperature for 25 min, then water was added
(50 mL) and the mixture was stirred for 5 min to form a
suspension. The precipitated product was collected by filtra-
tion. The filter cake was suspended in cold water (200 mL),
and the solid was collected. This was repeated three times
followed in the last run by washing a minimal amount of
methanol. The filter cake was dried (0.2 mmHg, ambient
temperature, 17 h) to give 12 as a white solid (24.0 g, 70%
from 3, 47% from 8), mp 137.2-137.8 °C. Optical rotation
[R]25D þ131 (c 0.50, CHCl3). 1H NMR (DMSO-d6) δ 1.68 (d,
3H, J=24.2 Hz, CH3), 4.62-4.74 (m, 2H, H-5, 50), 5.11-5.15
(m, 1H, H-4), 5.76 (dd, 1H, J=7.0, 18.4 Hz, H-3), 7.46 (m, 2H,
m-Ar), 7.55 (m, 2H, m-Ar), 7.62 (m, 1H, p-Ar), 7.70 (m, 1H,
p-Ar), 7.93 (m, 2H, o-Ar), 8.06 (m, 2H, p-Ar), 8.08 (m, 2H,
Ar). 13C NMR (DMSO-d6) δ 18.69 (d, J = 24.3 Hz), 63.90,
72.53 (d, J=7.0 Hz), 78.30, 92.38 (d, J=183.5 Hz), 128.95,
129.43, 129.59, 129.67, 130.00, 133.36, 134.32, 134.81, 165.47,
165.94, 170.24 (d, J = 21.4 Hz). IR (neat, cm-1) 1724.96,
1454.88, 1370.30, 1291.90, 1207.60, 1140.99, 1069.90,
1043.33, 953.48, 845.22. HR-MS calcd for C20H17O6FLi
1(M þ Li) 379.1169, obsd 379.1151.
reaction was diluted with ethyl acetate (400 mL) and water
(200 mL). The layers were separated and the aqueous layer was
extracted with ethyl acetate (2ꢀ100 mL). The combined organic
layer was washed with water (3ꢀ150 mL) and brine (150 mL),
dried over sodium sulfate, filtered, concentrated under reduced
pressure, and coevaporated with toluene (2ꢀ100 mL) to give the
crude acetate 14 as clear brown oil. The oil was poured onto
a plug of silica gel (50 g) in a sintered glass Buchner funnel and
eluted with 20% ethyl acetate in hexanes until all the acetate
was recovered. The solution of the acetate 14 was concentrated
under reduced pressure to a colorless, thick oil (32 g) in roughly
2:1 β/R anomers as determined by NMR in DMSO-d6. Silylated
base was prepared by heating to reflux a suspension of
N-benzoylcytosine (19.39 g, 90.14 mol) and ammonium sulfate
(300 mg) in hexamethyldisilazane (200 mL) for 6 h and concen-
trating the resulting solution by vacuum distillation under
aspiration, followed by drying under vacuum (0.2 mmHg)
for 2 h at ambient temperature. The oily residue was dissolved
in chlorobenzene (250 mL). To this solution was added a portion
(25.0 g) of the semipurified acetate 14 and neat tin(IV) chloride
(31 mL, 265 mmol). After stirring under a nitrogen atmosphere
for 2 h at ambient temperature, the reaction was heated to
60-70 °C for 19 h. The reaction mixture was cooled to 0 °C and
solid sodium bicarbonate (96 g, 1.14 mol) and ethyl acetate
(500 mL) were added. To the stirred solution was slowly added
water (20 mL) (caution: vigorous evolution of carbon dioxide).
After the mixture was stirred for 30 min at ambient temperature,
the suspension was filtered and the collected solid was washed
with ethyl acetate (200 mL). The filtrate was washed with water
and brine (2ꢀ250 mL each), dried over sodium sulfate, filtered,
and concentrated under reduced pressure to give a pale yellow-
brown solid. Methanol (250 mL) was added and the mixture
was heated under reflux for 30 min thencooled to ambient
temperature, and the resulting precipitate was collected
by filtration, washed with methanol (2 ꢀ 30 mL), and dried
(0.2 mmHg, 24 h, ambient temperature) to give 15 as a white
solid (8.0 g, 29% from lactone 12), mp 240-241 °C. 1H NMR
(CDCl3) δ 1.47 (d, 3H, J=22.3 Hz, CH3), 4.63 (dd, 1H, J=
2.8, 12.7 Hz, H-50), 4.72 (d, 1H, J=9.4 Hz, H-40), 4.87 (d, 1H, J=
12.7 Hz, H-500), 5.55 (br dd, 1H, 8.4, 20.9 Hz, H-30), 6.50 (br d,
1H, J=16.8 Hz, H-10), 7.41-7.55 (m, 7H, Ar and H-5), 7.61-
7.69 (m, 3H, Ar), 7.88 (d, 1H, J = 6.8 Hz, H-6), 8.06-8.10
(m, 5H, Ar), 8.65 (s, 1H, NH). 13C NMR (CDCl3) δ 17.19 (d, J=
25.8 Hz), 61.75 (s), 71.88 (s), 90.20 (br s), 96.98 (br s), 100.08
(d, J=197 Hz), 127.51, 128.34, 128.65, 128.83, 129.14, 128.37,
129.54, 130.12, 132.80, 133.38, 133.78, 134.02, 143.80
(br s), 154.30, 157.50, 162.51, 165.43, 165.95. IR (neat, cm
-1
)
1726.99, 1485.04, 1314.13, 1254.93, 1090.94, 1024.45, 708.53.
HR-MS calcd for C31H27FN3O7 (M þ H) 572.1838, found
572.1828.
4-Amino-1-((2R,3R,4R,5R)-3-fluoro-4-hydroxy-5-(hydroxy-
methyl)-3-methyltetrahydrofuran-2-yl)pyrimidin-2(1H)-one (1)
[Common Names: 20-Deoxy-20-fluoro-20-C-methylcytidine and
1-(2-Deoxy-2-fluoro-2-C-methyl-β-D-ribofuranosyl)cytosine].
Protected nucleoside 15 (16.7 g, 30.8 mmol) was suspended
in methanolic ammonia (7 M, 750 mL). The mixture was
stirred at atmospheric pressure at ambient temperature for
12 h, then concentrated under reduced pressure to a yellow
solid. Tetrahydrofuran (400 mL) was added and the suspen-
sion was heated under reflux with stirring for 30 min and
then allowed to cool to ambient temperature. The resulting
solid was collected by filtration, washed with methanol (2ꢀ
100 mL), and dried (0.2 mmHg, 24 h, ambient temperature)
to give 6.7 g of 1 (88% from 15) as an off-white powder, mp
(2R,3R,4R,5R)-5-(4-Benzamido-2-oxopyrimidin-1(2H)-yl)-
2-(benzoyloxymethyl)-4-fluoro-4-methyltetrahydrofuran-3-yl
Benzoate (15) [Common Names: 30,50-O-N4-Tribenzoyl-20-deoxy-
20-fluoro-20-C-methylcytidine and 1-(3,5-Di-O-benzoyl-2-deoxy-
2-fluoro-2-C-methyl-β-D-ribofuranosyl)-N4-benzoylcytosine].
Protected lactone 12 (23.0 g, 61.8 mmol) was dissolved in
anhydrous tetrahydrofuran (500 mL) and the solution was
cooled to -20 °C under a nitrogen atmosphere. Lithium tri-
tert-butylaluminum hydride (1.0 M in THF, 75 mL, 75 mmol)
was added in over 15 min with stirring while maintaining the
temperature at -20 °C. After 5 h, based on TLC, an additional
amount of the hydride (10 mL) was added. After 1.5 h more, the
reaction was complete to give lactol 13. To the reaction mixture
were added DMAP (7.5 g, 62 mmol) and acetic anhydride (58.1 g,
569 mmol) and the reaction was stirred at -20 °C for 2 h. The
218-219 °C. Optical rotation [R]25 þ128.6 (0.50, water).
D
TLC and spectral data matched the authentic standard.6 1H
ΝΜR (DMSO-d6) δ 1.15 (d, 3H, J=22.5 Hz, CH3), 3.62 (m,
1H, H-50), 3.78 (m, 3H, H-30, H-40, H-500), 5.20 (br s, 1H,
J. Org. Chem. Vol. 74, No. 17, 2009 6823

Wang et al.
JOCArticle
30-OH), 5.56 (d, 1H, J = 4.8 Hz, 50-OH), 5.72 (d, 1 H, J =
7.6 Hz, H-5), 6.08 (d, J=19.2 Hz, H-10), 7.26 (s, 2H, NH2),
7.86 (d, 1H, J =7.4 Hz., H-6). 13C NMR (DMSO-d6) δ 17.08
(d, J=25.8 Hz), 59.19 (s), 71.21 (d, J=17.6 Hz), 81.95 (s),
89.17 (d, J=40.5 Hz), 94.96 (s), 100.85 (s), 102.65 (s), 140.95
(s), 155.86 (s), 166.21 (s). IR (neat, cm-1) 3218.64, 1610.93,
1496.63, 1397.15, 1283.51, 1192.45, 1045.18, 1028.91,
842.98, 789.67, 706.70. LR-MS m/z 259. HPLC purity at
254 nm 99.9%.
Acknowledgment. The authors wish to thank Drs. Michael
J. Otto and Anthony J. Shuker for their helpful discussions
and organization of this work.
Supporting Information Available: General analytical meth-
ods and copies of 1H and 13C NMR spectra for compounds 8,
12, 15, and 1. This material is available free of charge via the
Internet at http://pubs.acs.org.
6824 J. Org. Chem. Vol. 74, No. 17, 2009