5-Exo versus 6-endo cyclization of nucleoside 2-sila-5-hexenyl radicals: Reaction of 6-(bromomethyl)dimethylsilyl 1′,2′-unsaturated uridines

Article abstract of DOI:10.1021/jo040260p

(Chemical Equation Presented) The mode of cyclization of 2-sila-5-hexen-1-yl radicals generated from 6-(bromomethyl)dimethylsilyl- 1′,2′-unsaturated uridines was investigated. In contrast to the case of the 2′-unsubstituted 6-silicon-tethered substrate (4

Full text of DOI:10.1021/jo040260p

5-Exo versus 6-Endo Cyclization of Nucleoside 2-Sila-5-hexenyl
Radicals: Reaction of 6-(Bromomethyl)dimethylsilyl
1′,2′-Unsaturated Uridines
Junko Ogamino,† Hideaki Mizunuma,† Hiroki Kumamoto,† Shingo Takeda,†
Kazuhiro Haraguchi,† Kazuo T. Nakamura,† Hiroshi Sugiyama,‡ and Hiromichi Tanaka*,†
School of Pharmaceutical Sciences, Showa University, 1-5-8 Hatanodai, Shinagawa-ku,
Tokyo 142-8555, Japan, and Department of Chemistry, Graduate School of Science,
Kyoto University, Kitashirakawa Oiwakecho Sakyo-ku, Kyoto, Japan
hirotnk@pharm.showa-u.ac.jp
Received September 13, 2004
The mode of cyclization of 2-sila-5-hexen-1-yl radicals generated from 6-(bromomethyl)dimethylsilyl-
1′,2′-unsaturated uridines was investigated. In contrast to the case of the 2′-unsubstituted 6-silicon-
tethered substrate (4), which undergoes exclusive 6-endo-cyclization, reactions of the 2′-substituted
(Me, CO2Me, OBz, and Cl) derivatives (14, 20, 22, and 24) uniformly proceeded in preferential or
exclusive 5-exo-mode. The Tamao oxidation of the resulting cyclized products was also carried out
to synthesize the corresponding 1′-C-hydroxymethyl derivatives.
Introduction
the expectation that, if its cyclization can be controlled
in favor of 5-exo-mode, the resulting product would act
as a precursor to 1′-C-hydroxymethylated nucleosides.
Despite the interest generated by the antibacterial and
antitumor activities found in the naturally occurring
antibiotic angustmycins C (3),7 no systematic study has
been executed so far on the structure-activity relation-
ships of this class of nucleosides.8
It is well appreciated that the 5-exo-ring closure of
5-hexenyl radicals is overwhelmingly preferred to the
alternative 6-endo-cyclization.1 5-Hexenyl radicals con-
taining various heteroatoms also follow the same trend2
with the exception of simple 2-sila and 3-sila counterparts
that have been reported to undergo preferential 6-endo-
cyclization.3 However, it has also been reported that both
6-endo and 5-exo pathways are operative in the radical
cyclization of allyl alcohols tethered with the (bromo-
methyl)dimethylsilyl group.4
As a part of our continuing studies on the chemistry
of 1′,2′-unsaturated uridine,5,6 1-(2-deoxy-D-erythro-pent-
1-enofuranosyl)uracil (1), we became interested in the
reaction of a nucleoside 2-sila-5-hexenyl radical (2) with
† Showa University.
Results and Discussion
‡ Kyoto University.
(1) Beckwith, A. L. J.; Schiesser, C. H. Tetrahedron 1985, 41, 3925.
(2) Beckwith, A. L. J. Tetrahedron 1981, 37, 3073. Curran, D. P.
Synthesis 1988, 417-439.
PreparationandRadicalReactionofthe6-(Bromo-
methyl)dimethylsilyl 1′,2′-Unsaturated Uridine. As
(3) Wilt, J. W. Tetrahedron 1985, 41, 3979.
(4) Nishiyama, H.; Kitajima, T.; Matsumoto, M.; Itoh, K. J. Org.
Chem. 1984, 49, 2298.
(7) Buchanan, J. G.; Wightman, R. H. In Topics in Antibiotic
Chemistry; Sammes, P. G., Ed.; Ellis Horwood Ltd.: Chichester, 1982;
Vol. 6, p 231.
(8) For recent reports concerning the synthesis of 1′-branched
nucleosides: (a) Kodama, T.; Shuto, S.; Ichikawa, S.; Matsuda, A. J.
Org. Chem. 2002, 67, 7706. (b) Haraguchi, K.; Kubota, Y.; Tanaka, H.
J. Org. Chem. 2004, 69, 1831.
(5) Itoh, Y.; Haraguchi, K.; Tanaka, H.; Gen, E.; Miyasaka, T. J.
Org. Chem. 1995, 60, 656.
(6) Kumamoto, H.; Shindoh, S.; Tanaka, H.; Itoh, Y.; Haraguchi,
K.; Gen, E.; Kittaka, A.; Miyasaka, T.; Kondo, M.; Nakamura, K. T.
Tetrahedron 2000, 56, 5363.
10.1021/jo040260p CCC: $30.25 © 2005 American Chemical Society
Published on Web 02/01/2005
1684
J. Org. Chem. 2005, 70, 1684-1690

5-Exo versus 6-Endo Cyclization of Nucleoside 2-Sila-5-hexenyl Radicals
SCHEME 1a
anionic migration of the stannyl group to yield the 2′-
stannyl derivative 8.6 This method enabled us to prepare
various types of 2′-substituted derivatives. Thus, iodina-
tion of 8 by using iodine gave 9, which was further
transformed to the 2′-methyl (10) and the 2′-carbomethoxy
(11) derivatives through halogen-lithium exchange reac-
tion. Similarly, by reacting 8 with NCS, the 2′-chloro
derivative (12) was obtained. The 2′-benzoyloxy deriva-
tive (13) was prepared from 3′,5′-bis-O-(tert-butyldi-
methylsilyl)-2′-ketouridine according to the published
procedure.11
a R ) TBDMS.
reported previously,6 the 6-position of 1 can be regio-
selectively lithiated with LDA. However, when BrCH2-
SiMe2Cl was reacted with the resulting 6-lithiated species
of 1, a large amount of the starting material was
recovered, presumably due to susceptibility of the reagent
to elimination. The use of LHMDS, which is less basic
than LDA, was found to be suitable for the present
purpose.9 Thus, treatment of a mixture of 1 and BrCH2-
SiMe2Cl (4 equiv) with LHMDS (6 equiv) at below -70
°C for 20 min in THF gave the 6-silicon-tethered deriva-
tive (4) in 87% yield.
The 2′-methyl-6-silicon-tethered derivative (14) was
prepared from 10 in 98% yield based on the aforemen-
tioned procedure using LHMDS, and its radical reaction
was carried out in refluxing benzene in a manner similar
to that for 4. As expected, the 5-exo-cyclized product 15
(41%) was formed predominantly over the 6-endo product
16 (8%). Additional products formed in this reaction were
When the radical reaction of 4 was carried out in
refluxing benzene by adding a mixture of AIBN (0.2
equiv) and Bu3SnH (2.0 equiv) via a motor-driven syringe
over 1 h at 80 °C (final concentration of the substrate:
0.01 M), only the 6-endo-cyclized products 5 (58%) and 6
(32%) were obtained (Scheme 1). Even at a lower tem-
perature of -50 °C (Et3B/Bu3SnH in toluene), no 5-exo-
cyclized product was formed, although a slight increase
was observed in the stereoselectivity of the R-face attack
(5/6 ) 3.1/1, total yield 55%). The regiochemistry of 5 and
6 was apparent from the presence of an anomeric proton
resonance (5, δ 6.25, J1′,2′ ) 3.6 Hz; 6, δ 5.97, J1′,2′ ) 4.0
Hz) that showed a JCH correlation to 13C-resonance of
3
the glycosidic bond-rearranged product 17 (29%), the 2′-
exomethylene derivative 18 (1%, stereochemistry not
known), and the reduced product 19 (4%). When carried
out at room temperature (Et3B/Bu3SnH/benzene), the
amount of 17 (36%) increased at the expense of 15 (26%).
The depicted structures of 15-17 came from the following
1H NMR observations. Compound 15 showed NOE cor-
relations between H-2′/one of the SiCH2 (11%) and 2′-
Me/H-3′ (4%). For 16, these were observed between H-3′/
H-1′ (13%), H-4′/one of the 6-SiMe2 (6%), 2′-Me/H-1′ (4%),
and 2′-Me/H-3′ (3%). The regiochemistry of 17 was
apparent from the presence of a coupling between H-1′
(double doublet, J ) 2.0 and 6.4 Hz) and 6-SiCH2, while
its stereochemistry was confirmed by the presence of
NOE enhancement between H-1′/2′-Me (5%) and H-1′/
H-4′ (5%).
the C6 in the HMBC (Heteronuclear Multiple Bond
Connectivity) spectrum. The depicted stereochemistry of
5 was unambiguously determined based on an NOE
experiment: 7.8% NOE correlation between H-1′ and
H-3′.
Although the preferred 6-endo-cyclization of a simple
2-sila-hexenyl radical has been rationalized in terms of
a longer C-Si bond,10 we reasoned that, in the present
case, the observed exclusive formation of 5 and 6 arises
at least in part from stabilization of the anomeric radical
A by the neighboring furanose ring oxygen. If this is the
case, it should be possible to alter the cyclization bias in
favor of 5-exo by introducing a substituent that stabilizes
a 2′-carbon radical.
Cyclization of 2′-Substituted 6-Silicon-Tethered
1′,2′-Unsaturated Uridines. Our recent study showed
that the 6-(tributyl)stannylated 1′,2′-unsaturated uridine
(7), prepared by LDA lithiation of 1 and subsequent
reaction with Bu3SnCl, undergoes an intramolecular
The above results for the reaction of 14, except for the
formation of minor products (16, 18, and 19), can be
rationalized as shown in Scheme 2. The high preference
for the C1′-attack of the R-silyl carbon-radial derived from
14 could be due to the formation of an incipient tertiary
C2′-radical B, although steric hindrance of the 2′-methyl
(9) Use of LHMDS in the C6-silylation of uridine derivatives has
recently been reported: Yoshimura, Y.; Kumamoto, H.; Baba A.;
Takeda, S.; Tanaka, H. Org. Lett. 2004, 6, 1793-1795.
(10) Chatgilialoglu, C.; Woynar, H.; Ingold, K. U.; Davies, A. J.
Chem. Soc., Perkin Trans. 2 1983, 555.
(11) Kodama, T.; Shuto, S.; Nomura, M.; Matsuda, A. Tetrahedron
Lett. 2000, 41, 3643.
J. Org. Chem, Vol. 70, No. 5, 2005 1685

Ogamino et al.
SCHEME 2a
SCHEME 3a
a R ) TBDMS.
data are in good agreement with the observed stereo-
chemical outcome of the cyclizations of 14 as well as 4.
To avoid the reaction pathway leading to glycosidic
bond-rearrangement, B f C f D (Scheme 2), it is
necessary to introduce a substituent that stabilizes the
intermediary 2′-carbon radical more efficiently.13 Since
there have been ample precedents that R,â-unsaturated
esters undergo radical addition reaction preferentially at
the â-position,14 a carbomethoxy group was selected as a
2′-substituent.
a R ) TBDMS.
TABLE 1. Minimized Energies (kcal/mol) of Conformers
E-H by Calculation
Compound 20 was prepared in 87% yield from 11 and
subjected to radical cyclization (Scheme 3). When exam-
ined in refluxing benzene, 20 gave the expected 5-exo-
product 21 in 53% yield together with the recovered 20
(13%), the reduced product (2%), and an unknown
product. There was no trace of the 6-endo or rearrange-
ment product. The best result was seen in the reaction
carried out at room temperature (Et3B/Bu3SnH/benzene)
to give 21 in 93% along with the reduced product (5%).
Reactions of the 2′-benzoyloxy (22) and the 2′-chloro (24)
substrates, carried out under similar conditions, also
resulted in exclusive 5-exo-cyclization to give 23 (75%)
and 25 (66%), respectively. These results suggest that
polar effects15 rather than the steric hindrance of the 2′-
substituent are an important determinant for directing
this reaction in favor of 5-exo-cyclization.
The reaction of 24 deserves a comment. Although 25
was isolated after 1.5 h, it was also possible to obtain
the dechlorinated product 26 (69%) simply by further
continuing the reaction for 4 days. This compensates for
the aforementioned failure of 4 to give the corresponding
5-exo-cyclized product.
Oxidative Cleavage of the 5-Exo-Cyclized Prod-
ucts. Silicon-tethered molecules are widely used for
intramolecular reactions because the tether can be
readily removed after the reaction and it offers the scope
for facile transformation into other functionalities.16 The
most widely used transformation is the oxidative cleavage
of the C-Si bond, the Tamao oxidation,17 that allows the
silicon atom to be converted to a hydroxyl group. There
have been several examples of intramolecular radical
compd
conformer energy (kcal/mol)
difference (kcal/mol)
4
E
F
-239.718
G
-250.769
-237.793
H
1.925
4.430
14
-246.339
group cannot be ruled out. Also, the observed concurrent
formation of 17 may suggest that B is not sufficiently
stable to react exclusively with Bu3SnH, and thus, it
eliminated uracil-1-yl radical C which cyclizes in a 6-endo
manner to yield a stabilized anomeric radical D.
One may question why cyclization of 14 occurred
exclusively from the R-face of the furanoid glycal portion
leading to 15 and 16, in comparison with that of 4 where
the â-face attack took place to an appreciable extent. As
we have no clear explanation for these results, semi-
empirical molecular orbital calculations12 of the repre-
sentative two conformers of 4 (E and F) and 14 (G and
H) were carried out. Geometries of 4 and 14 were
generated from crystallographic coordinates of 1 (see
Supporting Information) and optimization was performed
with MM2 using CAChe program.The results are sum-
marized in Table 1.
Irrespective of the presence or absence of a substituent
at the 2′-position, the “down-conformer” (E or G) was
found to be more stable than the “up-conformer” (F or
H). Also, energy difference between the up and down
conformers is significantly larger in the case of 14 (ca.
4.4 kcal/mol) than the case of 4 (ca. 1.9 kcal/mol). These
(13) For radical stabilization energies of substituents: Brocks, J.
J.; Beckhaus, H.-D.; Beckwith, A. L. J.; Ru¨chardt, C. J. Org. Chem.
1998, 63, 1935.
(14) For recent examples: (a) Della, E. W.; Kostakis, C.; Smith, P.
A. Org. Lett. 1999, 1, 363. (b) Kim, K.; Okamoto, S.; Sato, F. Org. Lett.
2001, 3, 67.
(15) Giese, B. Angew. Chem., Int. Ed. Engl. 1983, 22, 753.
(16) Bols, M.; Skrydstrup, T. Chem. Rev. 1995, 95, 1253.
(17) For a review: Tamao, K. In Advances in Silicon Chemistry;
Larson, G. L., Ed.; JAI Press: London, 1996; Vol. 3, p 1.
(12) Semiempirical calculations (PM3) were performed using Spar-
tan (version 3.1) program on a SGI O2 computer.
1686 J. Org. Chem., Vol. 70, No. 5, 2005

5-Exo versus 6-Endo Cyclization of Nucleoside 2-Sila-5-hexenyl Radicals
SCHEME 4
reaction of silicon-tethered nucleosides for the synthesis
of their C-branched sugar derivatives.18 In these in-
stances, the silicon-tethers used are uniformly attached
to a hydroxyl group of the sugar moiety, therefore the
heteroatom needed for the oxidative cleavage of C-Si
bond is already attached to the silicon atom. In contrast
to this, the present cyclization products have a silicon
atom connected to four carbon-substituents.19
During our studies on the synthesis of 6-substituted
uridines,20 we experienced that the trimethylsilyl group
at the 6-position can be removed simply by reacting with
NH3/MeOH (rt, 1 h), presumably with simultaneous
formation of methoxytrimethylsilane. This fact encour-
aged us to examine the oxidative cleavage of the present
cyclized products under these frequently employed basic
conditions.
In contrast to these successful examples, the 2′-
unsubstituted 5-exo-cyclized product 26 gave a consider-
able amount of uracil under the above oxidation condi-
tions. This may be attributable to the lack of an electron-
withdrawing substituent at the 2′-position of 26, thus
rendering the formation of sugar-cation comparatively
facile. Compound 35 was, therefore, prepared in 77%
yield by the tin-radical mediated dechlorination of 34.
Conclusion
The mode of cyclization (5-exo versus 6-endo) of radicals
derived from 6-(bromomethyl)dimethylsilyl-1′,2′-unsatu-
rated uridines was studied by changing the substituent
at the 2′-position. Compound 4 having no 2′-substituent
underwent exclusive 6-endo-cyclization consistent with
the reported precedent.10) Although introduction of the
2′-methyl group to 4 changed the cyclization bias in favor
of the 5-exo-mode, the intermediate 2′-carbon radical is
not sufficiently stable. As a result, a considerable extent
of glycosidic bond-rearrangement took place in this
particular reaction.
Cyclization reactions of other 2′-substituted (CO2Me,
OBz, and Cl) derivatives occurred exclusively in 5-exo-
mode.
Oxidative cleavage of these cyclized products was also
carried out to disclose a new entry to 1′-hydroxymethyl-
branched nucleosides. In the oxidation of 5-exo-cyclized
products, employing CF3CH2OH as reaction solvent was
critical for an efficient transformation.
When the 6-endo-cyclized product 5 was heated in
refluxing MeOH containing KHCO3 (5 equiv), KF (5
equiv), and H2O2 (4 equiv), the reaction proceeded very
sluggishly and required 3 days for completion.
Although the desired product was isolated in 81% yield
as its tri-O-acetyl derivative (27), we reexamined this
reaction. It was found that the addition of 18-crown-6 (5
equiv) to the reaction medium considerably shortened the
reaction time. Thus, refluxing for 22 h followed by
acetylation gave 27 in 96% yield (Scheme 4). Similarly,
28 was isolated in 66% yield from 6.
The oxidative cleavage of the 5-exo-cyclized products
(21, 23, 25, and 26)21 was next examined. Upon treatment
under the above-mentioned reaction conditions, the 2′-
carbomethoxy derivative 21 was not converted to the
expected 1′-hydroxymethyl derivative 29 but instead to
a product assumed to be 30. This product was unstable,
giving a more polar compound (31?) during silica gel
column chromatography. This result suggested that a
more electron-withdrawing alkoxy substituent was nec-
essary to increase the Lewis acidity of the silicon atom,
rendering the formation of penta- and hexacoordinated
silicon intermediates feasible.17) When the reaction of 21
was carried out in CF3CH2OH at rt for 15 h, and it was
followed by acetylation, the desired 32 was isolated in
87% yield. Similarly, 33 was obtained in 91% yield from
23. In the case of the 2′-chloro derivative 25, the corre-
sponding 1′-hydroxymethyl compound 34 was isolated in
93% yield without acetylation.
Experimental Section
6-(Bromomethyl)dimethylsilyl-1-[3,5-bis-O-(tert-butyl-
dimethylsilyl)-2-deoxy-D-erythro-pent-1-enofuranosyl]uracil
(4). To a THF (60 mL) solution of 1 (2.0 g, 4.4 mmol) and
BrCH2SiMe2Cl (2.4 mL, 17.6 mmol) was added LHMDS (1.54
M THF solution, 17.1 mL, 26.4 mmol) at below -70 °C under
positive pressure of dry Ar. After stirring for 20 min, the
reaction mixture was diluted with saturated aqueous NH4Cl.
Extraction with EtOAc followed by column chromatography
(hexane/EtOAc ) 4/1) gave 4 (2.33 g, 87%) as a foam: UV
(MeOH) λmax 267 nm (ꢀ 10100), λmin 237 nm (ꢀ 4000); 1H NMR
(CDCl3) δ 0.08, 0.09, 0.10, and 0.12 (12H, each as s), 0.49 and
0.50 (6H, each as s), 0.90 and 0.91 (18H, each as s), 2.69 and
2.81 (2H, each as d, J ) 13.0 Hz), 3.61 (1H, dd, J ) 10.4 and
8.4 Hz), 3.81 (1H, dd, J ) 10.4 and 6.0 Hz), 4.42 (1H, ddd, J
) 8.4, 6.0, and 2.0 Hz), 4.98 (1H, dd, J ) 2.8 and 2.0 Hz), 5.20
(18) (a) Xi, Z.; Agback, P.; Plavec, J.; Sandstrom, A.; Chatto-
padhyaya, J. Tetrahedron 1992, 48, 349. (b) Ueno, Y.; Nagasawa, Y.;
Sugimoto, I.; Kojima, N.; Kanazaki, M.; Shuto, S.; Matsuda, A. J. Org.
Chem. 1998, 63, 1660. (c) Kodama, T.; Shuto, S.; Nomura, M.; Matsuda,
A. Chem. Eur. J. 2001, 7, 2332.
(19) A phenyldimethylsilyl group attached carbon has been con-
verted to the hydroxyl group in two steps, first with HBF4 and then
with m-CPBA/Et3N: Fleming, I.; Henning, R.; Plaut, H. J. Chem. Soc.,
Chem. Commun. 1984, 29.
(20) Tanaka, H.; Hayakawa, H.; Miyasaka, T. In Nucleosides and
Nucleotides as Antitumor and Antiviral Agents; Chu, C. K., Baker, D.
C., Eds.; Plenum Press: New York, 1993; p 23.
(21) Due to a small amount of the 2′-methyl-5-exo-cyclized product
(15), the oxidative cleavage of this compounds was not examined.
(1H, d, J ) 2.8 Hz), 5.94 (1H, d, J ) 2.0 Hz), 8.34 (1H, br); 13
C
NMR (CDCl3) δ -5.5, -5.3, -4.7, -4.6, -3.1 and -3.0, 15.0,
17.9, 18.2, 25.7, 61.6, 75.3, 89.0, 101.4, 112.2, 149.6, 149.7,
156.9, 161.8; FAB-MS m/z 605 and 607 (M+ + H). Anal. Calcd
for C24H45BrN2O5Si3‚H2O: C, 46.21; H, 7.59; N, 4.49. Found:
C, 46.46; H, 7.68; N, 4.74.
J. Org. Chem, Vol. 70, No. 5, 2005 1687

Ogamino et al.
Radical Cyclization of 4: Formation of the 6-Endo-
Cyclized Products 5 and 6. To a refluxing benzene (16.5
mL) solution of 4 (200 mg, 0.33 mmol) was added a mixture of
AIBN (10.8 mg, 0.066 mmol) and Bu3SnH (0.18 mL, 0.66
mmol) in benzene (16.5 mL) over 1 h via a motor-driven
syringe under positive pressure of dry Ar. The reaction mixture
was further refluxed for 20 min and then evaporated. Column
chromatography (hexane/EtOAc ) 4/1) of the residue gave a
mixture of 5 and 6 (157 mg, 90%, 5/6 ) 1.8/1.0 calculated by
integrating H-1′). Separation of these products was carried out
by HPLC (hexane/EtOAc ) 3/2) to give analytically pure 5 (tR
) 9.6 min, solid) and 6 (tR ) 11.5 min, foam).
6-(Bromomethyl)dimethylsilyl-1-[3,5-bis-O-(tert-butyl-
dimethylsilyl)-2-deoxy-2-C-methyl-D-erythro-pent-1-eno-
furanosyl]uracil (14). This compound was obtained as a foam
in 98% yield from 10 (100 mg, 0.213 mmol) by the procedure
described for the preparation of 4: UV (MeOH) λmax 265 nm
1
(ꢀ 10100), λmin 236 nm (ꢀ 4400); H NMR (CDCl3) δ 0.08 and
0.13 (12H, each as s), 0.46 and 0.48 (6H, each as s), 0.91 (18H,
s), 1.62 (3H, s), 2.69 and 2.77 (2H, each as d, J ) 13.2 Hz),
3.58 (1H, dd, J ) 10.5 an 8.0 Hz), 3.77 (1H, dd, J ) 10.5 and
6.2 Hz), 4.28 (1H, ddd, J ) 8.0, 6.2, and 1.6 Hz), 4.72 (1H, d,
J ) 1.6 Hz), 5.96 (1H, d, J ) 2.4 Hz), 8.26 (1H, br); 13C NMR
(CDCl3) δ -5.4, -5.3, -4.8, -4.4, -3.3, 3.2, 9.5, 14.7, 17.9,
18.5, 25.7, 26.0, 62.0, 78.4, 86.7, 109.3, 112.4, 143.1, 149.5,
157.5, 162.0; FAB-MS m/z 619 and 621 (M+ + H). Anal. Calcd
for C25H47BrN2O5Si3: C, 48.45; H, 7.64; N, 4.52. Found: C,
48.84; H, 7.80; N, 4.50.
Radical Cyclization of 14: Formation of 15-19. The
reaction was carried out by using 14 (75 mg, 0.122 mmol) by
the procedure described for the case of 4. HPLC separation
(hexane/EtOAc ) 2/1) of the reaction mixture gave 15 (tR 12.0
min, 27 mg, 41%, foam), 16 (tR 10.9 min, 4.9 mg, 8%, foam), a
mixture of 17 and 19 [tR 9.8 min, 21.8 mg, 17 (29%) and 19
(4%): the yields were calculated from 1H NMR integration]
and 18 (tR 14.0 min, 0.7 mg, 1%, foam).
Physical data of 5: mp 67-75 °C; UV (MeOH) λmax 271 nm
(ꢀ 10400), λmin 235 nm (ꢀ 1700); 1H NMR (CDCl3) δ 0.08, 0.10,
and 0.11 (12H, each as s), 0.30 and 0.41 (6H, each as s), 0.68
(1H, dd, J ) 9.2 and 15.6 Hz), 0.92 and 0.93 (18H, each as s),
1.26 (1H, dd, J ) 15.6 and 5.6 Hz), 2.52-2.66 (1H, m), 3.62-
3.68 (2H, m), 3.83 (1H, dd, J ) 10.6 and 2.0 Hz), 4.72 (1H, dd,
J ) 7.4 and 5.8 Hz), 5.83 (1H, d, J ) 2.0 Hz), 6.25 (1H, d, J )
3.6 Hz), 8.10 (1H, br); 13C NMR (CDCl3) δ -5.6, -5.4, -5.0,
-4.8, -2.3, -1.5, 2.0, 18.0, 18.3, 25.7, 25.9, 40.3, 62.1, 71.0,
81.4, 81.9, 109.3, 151.4, 158.3, 161.6; FAB-MS m/z 527 (M+
+
H). Anal. Calcd for C24H46N2O5Si3‚1/4H2O: C, 54.25; H, 8.82;
N, 5.27. Found: C, 54.42; H, 9.11; N, 5.28.
Physical data of 15: UV (MeOH) λmax 252 nm (ꢀ 14600), λmin
225 nm (ꢀ 5600); 1H NMR (CDCl3) δ 0.02, 0.03, and 0.08 (12H,
each as s), 0.40 and 0.42 (6H, each as s), 0.88 and 0.89 (18H,
each as s), 1.05 (3H, d, J ) 7.2 Hz), 1.55 (1H, d, J ) 15.6 Hz),
2.47 (1H, dd, J ) 8.2 and 7.2 Hz), 3.80-3.91 (3H, m), 4.42
(1H, dd, J ) 8.2 and 7.0 Hz), 5.72 (1H, d, J ) 2.4 Hz), 8.16
(1H, br); 13C NMR (CDCl3) δ -5.2, -5.0, -4.2, -3.9, -3.3, -2.7,
11.0, 17.7, 18.5, 25.7, 26.0, 29.3, 55.7, 64.4, 77.5, 86.9, 105.7,
106.4, 150.2, 161.5, 163.8; FAB-MS m/z 541 (M+ + H). Anal.
Calcd for C25H48N2O5Si3: C, 55.51; H, 8.94; N, 5.18. Found:
C, 55.82; H, 9.15; N, 5.17.
Physical data of 6: UV (MeOH) λmax 270 nm (ꢀ 9800), λmin
234 nm (ꢀ 1500); 1H NMR (CDCl3) δ 0.06, 0.07, and 0.08 (12H,
each as s), 0.30 and 0.35 (6H, each as s), 0.69 (1H, dd, J )
14.8 and 5.0 Hz), 0.89 (18H, s), 1.19 (1H, dd, J ) 14.8 and
12.8 Hz), 2.18-2.23 (1H, m), 3.79-3.84 (3H, m), 4.04 (1H, d,
J ) 2.0 Hz), 5.77 (1H, d, J ) 2.4 Hz), 5.97 (1H, d, J ) 4.0 Hz),
8.16 (1H, br); 13C NMR (CDCl3) δ -5.5, -5.3, -4.7, -4.6, -2.6,
-2.5, 7.9, 17.9, 18.5, 25.7, 26.0, 44.0, 62.6, 80.5, 82.5, 85.4,
108.9, 151.8, 157.6, 161.5; FAB-MS m/z 527 (M+ + H). Anal.
Calcd for C24H46N2O5Si3: C, 54.71; H, 8.80; N, 5.32. Found:
C, 54.44; H, 8.93; N, 5.26.
Physical data of 16: UV (MeOH) λmax 272 nm (ꢀ 10000), λmin
236 nm (ꢀ 1800); 1H NMR (CDCl3) δ 0.08, 0.10, 0.11, and 0.12
(12H, each as s), 0.26 (1H, d, J ) 16.0 Hz), 0.30 and 0.39 (6H,
each as s), 0.93 and 0.94 (18H, each as s), 1.08 (3H, s), 1.51
(1H, d, J ) 16.0 Hz), 3.55 (1H, ddd, J ) 7.2, 2.4, and 2.0 Hz),
3.67 (1H, dd, J ) 11.6 and 2.4 Hz), 3.82 (1H, dd, J ) 11.6 and
2.0 Hz), 4.27 (1H, d, J ) 7.2 Hz), 5.84 (1H, d, J ) 2.0 Hz),
5.94 (1H, s), 7.99 (1H, br); 13C NMR (CDCl3) δ -5.6, -5.4, -4.5,
-4.3, -1.3, 0.5, 10.3, 18.0, 18.2, 23.7, 25.7, 25.8, 44.8, 62.1,
76.5, 81.3, 86.1, 109.3, 151.6, 158.3, 161.6; FAB-MS m/z 541
(M+ + H). Anal. Calcd for C25H48N2O5Si3: C, 55.51; H, 8.94;
N, 5.18. Found: C, 55.39; H, 9.19; N, 4.97.
Preparation of Compounds 7-10 and 12. See ref 6.
1-[3,5-Bis-O-(tert-butyldimethylsilyl)-2-C-carbomethoxy-
D-erythro-pent-1-enofuranosyl]uracil (11). Under positive
pressure of dry Ar, BuLi (1.6 M in hexane, 1.47 mL, 2.35 mmol)
was added dropwise to a mixture of 9 (136 mg, 0.234 mmol)
and ClCO2Me (181 µL, 2.34 mmol) in THF (5 mL). After
stirring for 20 min, the reaction was quenched by adding
saturated aqueous NH4Cl. Extraction with EtOAc followed by
column chromatography (hexane/EtOAc ) 6/1) gave 11 (73.3
mg, 61%) as a solid: mp 150-153 °C; UV (MeOH) λmax 245
nm (ꢀ 17100), λmin 224 nm (ꢀ 14800); 1H NMR (CDCl3) δ 0.09,
012, and 0.15 (12H, each as s), 0.89 (18H, s), 3.68 (1H, dd, J
) 11.0 and 6.8 Hz), 3.71 (3H, s), 3.83 (1H, dd, J ) 11.0 and
5.2 Hz), 4.51 (1H, ddd, J ) 6.8, 5.2, and 2.0 Hz), 5.18 (1H, d,
J ) 2.0 Hz), 5.79 (1H, dd, J ) 8.2 and 2.0 Hz), 7.23 (1H, d, J
) 8.2 Hz), 8.33 (1H, br); 13C NMR (CDCl3) δ -5.5, -5.4, -4.8,
-4.7, 18.0, 18.3, 25.7, 25.8, 51.3, 61.7, 74.3, 90.3, 102.5, 104.3,
142.5, 147.6, 156.6, 162.5, 163.1; FAB-MS m/z 513 (M+ + H).
Anal. Calcd for C23H40N2O7Si2: C, 53.88; H, 7.86; N, 5.46.
Found: C, 53.83; H, 7.90; N, 5.43.
1H NMR and 13C NMR data of 17: 1H NMR (CDCl3) δ -0.01,
0.11, 0.20, 0.27, and 0.40 (18H, each as s), 0.85 and 0.90 (18H,
each as s), 1.02 (1H, dd, J ) 2.0 and 15.6 Hz), 1.44 (1H, dd, J
) 15.6 and 6.4 Hz), 1.49 (3H, s), 3.27 (1H, dd, J ) 10.2 and
7.0 Hz), 3.49 (1H, dd, J ) 10.2 and 5.6 Hz), 3.79 (1H, ddd, J
) 7.0, 5.6, and 1.2 Hz), 4.36 (1H, dd, J ) 6.4 and 2.0 Hz), 4.71
(1H, d, J ) 1.2 Hz), 5.75 (1H, d, J ) 2.4 Hz), 8.18 (1H, br); 13
C
NMR (CDCl3) δ -5.6, -5.5, -5.4, -5.3, 0.0, 0.2, 8.7, 16.7, 17.9,
18.5, 25.8, 26.0, 63.3, 71.3, 77.4, 80.5, 87.6, 109.2, 151.4, 161.0,
161.4.
1-[2-O-Benzoyl-3,5-bis-O-(tert-butyldimethylsilyl)-D-
erythro-pent-1-enofuranosyl]uracil (13). This compound
was prepared as a foam in 71% yield from 3′,5′-bis-O-(tert-
butyldimethylsilyl)-2′-ketouridine by the published procedure:
Physical data of 18: UV (MeOH) λmax 263 nm (ꢀ 10300), λmin
232 nm (ꢀ 3500); 1H NMR (CDCl3) δ -0.01, 0.01, 0.14, and
0.18 (12H, each as s), 0.45 and 0.46 (6H, each as s), 0.86 and
0.94 (18H, each as s), 1.37 and 1.62 (2H, each as d, J ) 15.4
Hz), 3.58 (1H, ddd, J ) 8.2, 6.0, and 2.4 Hz), 3.76 (1H, dd, J
) 11.6 and 6.0 Hz), 3.89 (1H, dd, J ) 11.6 and 2.4 Hz), 4.97
(1H, ddd, J ) 8.2, 2.4, and 2.2 Hz), 5.09 (1H, d, J ) 2.2 Hz),
5.23 (1H, d, J ) 2.4 Hz), 5.76 (1H, d, J ) 2.4 Hz), 7.75 (1H,
br); 13C NMR (CDCl3) δ -5.4, -5.3, -4.5, -4.2, -3.5, -2.5,
17.9, 18.3, 25.8, 26.0, 27.4, 62.9, 72.1, 83.4, 100.6, 105.8, 107.8,
149.0, 155.4, 160.6, 163.5; FAB-MS m/z 539 (M+ + H). Anal.
Calcd for C25H46N2O5Si3: C, 55.72; H, 8.60; N, 5.20. Found:
C, 55.69; H, 8.55; N, 4.96.
18c
UV (MeOH) λmax 231 nm (ꢀ 17700), λmin 212 nm (ꢀ 12500),
λshoulder 270 nm (ꢀ 9500); 1H NMR (CDCl3) δ 0.00, 0.07, and
0.10 (12H, each as s), 0.84 and 0.93 (18H, each as s), 3.81 (1H,
dd, J ) 11.0 and 6.0 Hz), 3.89 (1H, dd, J ) 11.0 and 5.6 Hz),
4.56 (1H, ddd, J ) 6.0, 5.6, and 3.0 Hz), 5.27 (1H, d, J ) 3.0
Hz), 5.77 (1H, dd, J ) 8.0 and 2.4 Hz), 7.40 (1H, d, J ) 8.0
Hz), 7.45-7.49 (2H, m), 7.58-7.63 (1H, m), 8.04-8.07 (2H,
m), 8.22 (1H, br); 13C NMR (CDCl3) δ -5.4, -5.3, -4.7, -4.6,
17.9, 18.4, 25.6, 25.9, 62.4, 74.1, 87.3, 103.0, 123.4, 128.3, 128.6,
130.2, 133.9, 137.0, 141.6, 147.2, 162.3, 163.5; FAB-MS m/z
575 (M+ + H). Anal. Calcd for C28H42N2O7Si2: C, 58.51; H 7.37;
N, 4.87. Found: C, 58.57; H, 7.46; N, 4.78.
1H NMR and 13C NMR data of 19: 1H NMR (CDCl3) δ 0.08,
0.11, and 0.12 (12H, each as s), 0.32 (9H, s), 0.89 and 0.91
1688 J. Org. Chem., Vol. 70, No. 5, 2005

5-Exo versus 6-Endo Cyclization of Nucleoside 2-Sila-5-hexenyl Radicals
(18H, each as s), 1.60 (3H, s), 3.61 (1H, dd, J ) 10.6 and 8.4
Hz), 3.78 (1H, dd, J ) 10.6 and 6.0 Hz), 4.27 (1H, ddd, J )
8.2, 6.0, and 2.0 Hz), 4.72 (1H, s), 5.91 (1H, d, J ) 2.4 Hz),
8.13 (1H, br); 13C NMR (CDCl3) δ -5.4, -5.3, -4.3, -1.0, 9.6,
17.9, 18.5, 25.6, 26.0, 62.1, 78.4, 86.6, 108.8, 110.9, 143.4, 149.8,
161.3, 162.2.
and 0.90 (18H, each as s), 1.78 and 2.22 (2H, each as d, J )
15.4 Hz), 3.86-4.01 (3H, m), 4.92 (1H, dd, J ) 8.0 and 6.2
Hz), 5.46 (1H, d, J ) 6.2 Hz), 5.56 (1H, d, J ) 2.4 Hz), 7.40-
7.44 and 7.55-7.59 (3H, each as m), 7.65 (1H, br), 7.93-7.95
(2H, m); 13C NMR (CDCl3) δ -5.2, -5.1, -4.8, -4.2, -4.1, -3.4,
17.7, 18.5, 25.6, 26.0, 27.6, 64.1, 75.5, 82.6, 89.2, 102.3, 105.4,
128.4, 128.7, 129.6, 133.8, 149.6, 161.2, 163.0, 166.0; FAB-MS
m/z 647 (M+ + H). Anal. Calcd for C31H50N2O7Si3: C, 57.55;
H, 7.79; N, 4.33. Found: C, 57.49; H, 8.08; N, 4.29.
6-(Bromomethyl)dimethylsilyl-1-[3,5-bis-O-(tert-butyl-
dimethylsilyl)-2-C-carbomethoxy-2-deoxy-D-erythro-pent-1-
enofuranosyl]uracil (20). This compound was obtained as
a solid (mp 117-120 °C) in 87% yield from 11 (361 mg, 0.703
mmol) by the procedure described for the preparation of 4: UV
(MeOH) λmax 262 nm (ꢀ 13200), λmin 221 nm (ꢀ 9000); 1H NMR
(CDCl3) δ 0.09, 0.14, and 0.18 (12H, each as s), 0.45 and 0.46
(6H, each as s), 0.90 and 0.91 (18H, each as s), 2.66 (2H, s),
3.62 (1H, dd, J ) 10.5 and 8.5 Hz), 3.84 (1H, dd, J ) 10.5 and
6.0 Hz), 4.52 (1H, ddd, J ) 8.5, 6.0, and 1.4 Hz), 5.15 (1H, d,
J ) 1.4 Hz), 6.00 (1H, d, J ) 2.4 Hz), 8.13 (1H, br); 13C NMR
(CDCl3) δ -5.4, -4.9, -3.5, 14.4, 17.9, 18.4, 25.6, 25.9, 51.5,
60.9, 74.2, 90.3, 107.5, 112.4, 149.2, 156.0, 157.1, 161.6, 163.0;
FAB-MS m/z 663 and 665 (M+ + H). Anal. Calcd for C26H47-
BrN2O7Si3: C, 47.04; H, 7.14; N, 4.22. Found: C, 47.07; H,
7.39; N, 4.19.
Radical Cyclization of 20: Formation of 21. To a
mixture of 20 (50 mg, 0.076 mmol) and Et3B (1M THF solution,
30 µL, 0.0304 mmol) in benzene (6.7 mL) was added a benzene
(7.6 mL) solution of Bu3SnH (31 µL, 0.114 mmol) via a motor-
driven syringe over 1 h at rt, during which time the same
amount of Et3B was added three times every 15 min. After
addition of Bu3SnH, the reaction mixture was stirred for
further 0.5 h. Evaporation of the solvent followed by HPLC
separation (hexane/EtOAc ) 2/1) gave 21 (tR 12.4 min, 41 mg,
93%, solid) and the reduced product (tR 9.2 min, 2.4 mg, 5%,
foam). Physical data of 21: mp 194-195 °C; UV (MeOH) λmax
267 nm (ꢀ 10200), λmin 233 nm (ꢀ 1700); 1H NMR (CDCl3) δ
-0.03, 0.01, 0.11, and 0.14 (12H, each as s), 0.43 and 0.44 (6H,
each as s), 0.85 and 0.88 (18H, each as s), 1.69 and 1.77 (2H,
each as d, J ) 15.2 Hz), 3.39 (1H, d, J ) 7.2 Hz), 3.65 (3H, s),
3.71 (1H, dd, J ) 11.6 and 7.2 Hz), 3.79 (1H, dd, J ) 11.6 and
2.4 Hz), 3.83 (1H, dt, J ) 7.2 and 2.4 Hz), 4.92 (1H, t, J ) 7.2
Hz), 5.75 (1H, d, J ) 2.4 Hz), 7.91 (1H, br); 13C NMR (CDCl3)
δ -5.4, -5.3, -4.8, -4.5, -3.6, -2.6, 18.0, 18.4, 25.8, 25.9,
30.0, 51.5, 63.5, 64.5, 73.6, 86.2, 102.2, 106.1, 150.1, 160.7,
163.0, 169.5; FAB-MS m/z 585 (M+ + H), 623 (M+ + K). Anal.
Calcd for C26H48N2O7Si3: C, 53.39; H, 8.27; N, 4.79. Found:
C, 53.39; H, 8.66; N, 4.73.
6-(Bromomethyl)dimethylsilyl-1-[3,5-bis-O-(tert-butyl-
dimethylsilyl)-2-chloro-2-deoxy-D-erythro-pent-1-enofuran-
osyl]uracil (24). This compound was obtained as a powder
(mp 127-129 °C) in 89% yield from 12 (330 mg, 0.675 mmol)
by the procedure described for the preparation of 4: UV
(MeOH) λmax 263 nm (ꢀ 10500), λmin 239 nm (ꢀ 6000); 1H NMR
(CDCl3) δ 0.09, 0.10, 0.15 and 0.17 (12H, each as s), 0.50 and
0.51 (6H, aech as s), 0.91 and 0.92 (18H, each as s), 2.73 (2H,
s), 3.66 (1H, dd, J ) 10.6 and 8.2 Hz), 3.84 (1H, dd, J ) 10.6
and 6.2 Hz), 4.40 (1H, ddd, J ) 8.2, 6.2, and 2.0 Hz), 4.85 (1H,
d, J ) 2.0 Hz), 5.98 (1H, d, J ) 2.4 Hz), 8.20 (1H, br); 13C
NMR (CDCl3) δ -5.4, -4.7, -4.5, -3.5, -3.4, 1.6, 14.5, 17.6,
18.4, 25.6, 25.9, 61.4, 76.7, 87.7, 108.8, 112.8, 144.9, 148.9,
156.8, 161.7; FAB-MS m/z 639 and 641 (M+ + H). Anal. Calcd
for C24H44BrClN2O5Si3: C, 45.02; H, 6.93; N, 4.38. Found: C,
45.33; H, 7.02; N, 4.38.
Radical Cyclization of 24: Formation of 25 and 26.
This reaction was carried out by using 24 (50 mg, 0.078 mmol)
in a similar manner as described for 20. After finishing
addition of Bu3SnH, the reaction mixture was stirred for 0.5
h at rt. HPLC separation (hexane/EtOAc ) 2/1) of the mixture
gave 25 (tR 11.6 min, 28.7 mg, 66%, foam), 26 (tR 14.2 min, 2.5
mg, 6%, foam), and the reduced product (tR 8.3 min, 3.7 mg,
8%, foam). When the reaction mixture was stirred for 4 days,
26 was isolated in 69% together with 25 (trace amount) and
the reduced product (7%).
Physical data of 25: UV (MeOH) λmax 262 nm (ꢀ 10100), λmin
229 nm (ꢀ 1900); 1H NMR (CDCl3) δ 0.05, 0.11, and 0.15 (12H,
each as s), 0.43 and 0.44 (6H, each as s), 0.88 and 0.90 (18H,
each as s), 1.70 and 1.78 (2H, each as d, J ) 15.6 Hz), 3.81
(1H, dd, J ) 11.2 and 1.6 Hz), 3.91 (1H, ddd, J ) 8.3, 8.1, and
1.6 Hz), 4.01 (1H, dd, J ) 11.2 and 8.3 Hz), 4.23 (1H, d, J )
8.0 Hz), 4.77 (1H, dd, J ) 8.1 and 8.0 Hz), 5.73 (1H, d, J ) 2.4
Hz), 8.47 (1H,br); 13C NMR (CDCl3) δ -5.2, -5.1, -4.6, -3.8,
-3.1, -3.0, 17.8, 18.5, 25.7, 26.0, 29.0, 64.0, 73.1, 78.5, 84.8,
103.4, 105.9, 149.7, 160.3, 163.4; FAB-MS m/z 561 (M+ + H).
Anal. Calcd for C24H45ClN2O5Si3: C, 51.35; H, 8.08; N, 4.99.
Found: C, 51.59; H, 8.31; N, 4.99.
Physical data of 26: UV (MeOH) λmax 258 nm (ꢀ 7400), λmin
232 nm (ꢀ 2400); 1H NMR (CDCl3) δ 0.00, 0.01, 0.08, and 0.09
(12H, each as s), 0.38 and 0.44 (6H, each as s), 0.87 and 0.90
(18H, each as s), 1.42 and 1.72 (2H, each as d, J ) 15.2 Hz),
2.05 (1H, dd, J ) 13.3 Hz), 3.51 (1H, dd, J ) 13.3 and 7.7 Hz),
3.67 (1H, dd, J ) 12.0 and 7.6 Hz), 3.77-3.82 (2H, m), 4.59
(1H, dd, J ) 7.7, and 6.0 Hz), 5.74 (1H, d, J ) 2.4 Hz), 8.05
(1H, br); 13C NMR (CDCl3) δ -5.4, -5.2, -4.8, -3.8, -3.5, -2.7,
17.9, 18.4, 25.8, 25.9, 29.2, 47.1, 63.4, 72.4, 87.1, 103.7, 105.9,
149.7, 161.2, 163.2; FAB-MS m/z 527 (M+ + H). Anal. Calcd
for C24H46N2O5Si3: C, 54.25; H, 8.82; N, 5.27. Found: C, 54.40;
H, 8.72; N, 5.12.
1-(Bromomethyl)dimethylsilyl-1-[2-O-benzoyl-3,5-
bis-O-(tert-butyldimethylsilyl)-D-erythro-pent-1-enofuran-
osyl]uracil (22). This compound was obtained as a foam in
80% yield from 13 (100 mg, 0.174 mmol) by the procedure
described for the preparation of 4: UV (MeOH) λmax 232 nm
(ꢀ 18200) and 263 nm (ꢀ 12800), λmin 218 nm (ꢀ 15300) and
1
250 nm (ꢀ 11900); H NMR (CDCl3) δ -0.02, 0.09, 0.12 and
0.13 (12H, each as s), 0.57 and 0.58 (6H, each as s), 0.84 and
0.93 (18H, each as s), 2.81 and 2.88 (2H, each as d, J ) 13.2
Hz), 3.82 (1H, dd, J ) 10.4 and 8.0 Hz), 3.91 (1H, dd, J ) 10.4
and 6.4 Hz), 4.41 (1H, ddd, J ) 8.0, 6.4 and 1.6 Hz), 5.31 (1H,
d, J ) 1.6 Hz), 5.98 (1H, s), 7.98 (1H, br); 13C NMR (CDCl3) δ
-5.4, -4.7, -3.3, -3.2, 14.8, 17.8, 18.5, 25.6, 25.9, 61.7, 76.7,
87.1, 112.7, 127.8, 128.2, 128.7, 130.1, 133.9, 137.7, 148.7,
156.9, 161.4, 163.1; FAB-MS m/z 725 and 727 (M+ + H). Anal.
Calcd for C26H48BrN2O7Si3: C, 51.29; H, 6.80; N, 3.86. Found:
C, 51.41; H, 7.02; N, 3.89.
1-(2-C-Acetoxymethyl-3,5-di-O-acetyl-2-deoxy-r-D-ribo-
furanosyl)uracil (27). A mixture of 5 (100 mg, 0.19 mmol),
30% aqueous H2O2 (73 µL, 0.76 mmol), 18-crown-6 (251 mg,
0.95 mmol), KHCO3 (81 mg, 0.95 mmol), and KF (55 mg, 0.95
mmol) in MeOH (3 mL) was stirred at refluxing temperature
for 22 h. The reaction mixture was evaporated and dried under
reduced pressure. To a pyridine (3 mL) solution of the resulting
residue was added Ac2O (0.45 mL, 4.75 mmol) and the mixture
was stirred at rt overnight. Evaporation of the solvent followed
by column chromatography (EtOAc) gave 27 (70 mg, 90%) as
a foam: UV (MeOH) λmax 261 nm (ꢀ 12000), λmin 231 nm (ꢀ
2600); 1H NMR (CDCl3) δ 2.02, 2.10, and 2.14 (9H, each as s),
3.23-3.30 (1H, m), 4.07-4.21 (4H, m), 4.58 (1H, dt, J ) 4.8
Radical Cyclization of 22: Formation of 23. This
reaction was carried out by using 22 (50 mg, 0.069 mmol) in
a similar manner as described for 20. After addition of
Bu3SnH, the reaction mixture was stirred for 3 h at rt. HPLC
separation (hexane/EtOAc ) 5/1) of the mixture gave 23 (tR
14.6 min, 35 mg, 75%, foam) and the reduced product (tR 11.3
min, 1.5 mg, 3%, foam). Physical data of 23: UV (MeOH) λmax
232 nm (ꢀ 15700) and 265 nm (ꢀ 10000), λmin 211 nm (ꢀ 10100)
and 249 nm (ꢀ 6800); 1H NMR (CDCl3) δ -0.04, 0.06, 0.07,
and 0.08 (12H, each as s), 0.39 and 0.43 (6H, each as s), 0.84
J. Org. Chem, Vol. 70, No. 5, 2005 1689

Ogamino et al.
and 1.0 Hz), 5.40 (1H, dd, J ) 6.0 and 1.0 Hz), 5.76 (1H, d, J
) 8.2 Hz), 6.47 (1H, d, J ) 7.2 Hz), 7.41 (1H, d, J ) 8.2 Hz),
8.50 (1H, br); 13C NMR (CDCl3) δ 20.7, 20.8, 20.9, 45.2, 57.6,
63.5, 73.6, 84.5, 86.8, 102.0, 139.8, 150.5, 163.1, 169.5, 170.4;
FAB-MS m/z 385 (M+ + H). Anal. Calcd for C16H20N2O9‚1/
5H2O: C, 49.54; H, 5.30; N, 7.22. Found: C, 49.47; H, 5.16; N,
6.86.
1-(2-C-Acetoxymethyl-3,5-di-O-acetyl-2-deoxy-â-D-ara-
binofuranosyl)uracil (28). This compound was prepared as
a foam in 66% yield from 6 (100 mg, 0.19 mmol) by the
procedure described for the preparation of 27. The reaction
was continued for 16 h at the refluxing temperature of
MeOH: UV (MeOH) λmax 262 nm (ꢀ 9400), λmin 230 nm (ꢀ 2100);
1H NMR (CDCl3) δ 2.00, 2.13 and 2.14 (9H, each as s), 3.05-
3.09 (1H, m), 4.04 (1H, dd, J ) 12.0 and 3.6 Hz), 4.15 (1H, dd
J ) 12.0 and 5.2 Hz), 4.18 (1H, ddd, J ) 5.2, 4.4 and 3.2 Hz),
4.40 (1H, dd, J ) 12.2 and 4.4 Hz), 4.43 (1H, dd, J ) 12.2 and
3.2 Hz), 5.10 (1H, dd, J ) 5.2 and 4.8 Hz), 5.75 (1H, d, J ) 8.0
Hz), 6.17 (1H d, J ) 6.4 Hz), 7.55 (1H, d, J ) 8.0 Hz), 8.77
(1H, br); 13C NMR (CDCl3) δ 20.5, 20.7, 21.0, 46.6, 59.6, 62.6,
73.6, 80.6, 85.5, 101.9, 139.6, 150.2, 163.1, 170.1, 170.2, 170.3;
FAB-MS m/z 385 (M+ + H). Anal. Calcd for C16H20N2O9‚1/
5H2O: C, 49.54; H, 5.30; N, 7.22. Found: C, 49.42; H, 5.18; N,
6.90.
1-[1-C-Acetoxymethyl-3,5-bis-O-(tert-butyldimethyl-
silyl)-2-C-carbomethoxy-2-deoxy-â-D-arabinofuranosyl]-
uracil (32). To a solution of 21 (16.7 mg, 0.029 mmol) in
CF3CH2OH (1.5 mL) were added KHCO3 (14.5 mg, 0.145
mmol), 18-crown-6 (38 mg, 0.145 mmol), KF (8.4 mg, 0.145
mmol), and 30% aqueous H2O2 (13.1 µL, 0.116 mmol) at -20
°C. The reaction mixture was stirred for 15 h at rt, evaporated,
and partially purified by Florisil column chromatography
(hexane/EtOAc ) 2/1). The resulting crude product was treated
overnight with Ac2O (16.4 µL, 0.17 mmol) in pyridine (2 mL).
Florisil column chromatography (hexane/EtOAc ) 10/1) of the
acetylation mixture gave 32 (14.8 mg, 87%) as a foam: UV
(MeOH) λmax 261 nm (ꢀ 10600), λmin 230 nm (ꢀ 1900); 1H NMR
(CDCl3) δ 0.03, 0.07, 0.09, and 0.10 (12H, each as s), 0.87 and
0.92 (18H, each as s), 2.05 (3H, s), 3.36 (1H, d, J ) 6.2 Hz),
3.76 (1H, dd, J ) 11.2 and 3.2 Hz), 3.89 (1H, dd, J ) 11.2 and
4.4 Hz), 3.93-3.97 (1H, m), 4.53 (1H, t, J ) 6.2 Hz), 4.71 and
4.77 (2H, each as d, J ) 10.4 Hz), 5.65 (1H, dd, J ) 8.4 and
2.4 Hz), 8.00 (1H, d, J ) 8.4 Hz) 8.36 (1H, br); 13C NMR
(CDCl3) δ -5.4, -5.0, -4.9, 0.0, 17.8, 18.3, 20.7, 25.5, 25.9,
52.4, 58.3, 60.3, 64.8, 74.1, 85.8, 95.1, 101.0, 141.2, 150.0, 162.9,
169.9, 170.0; FAB-MS m/z 587 (M+ + H), 625 (M+ + K). Anal.
Calcd for C26H46N2O9Si2: C, 53.22; H, 7.90; N, 4.77. Found:
C, 53.62; H, 8.05; N, 4.72.
Hz), 7.38-7.42 (2H, m), 7.54-7.57 (1H, m), 7.78-7.80 (2H,
m), 7.91 (1H, d, J ) 8.4 Hz), 8.12 (1H, br); 13C NMR (CDCl3)
δ -5.6, -5.5, -5.2, -4.8, 17.8, 18.2, 20.6, 25.6, 25.7, 62.2, 63.0,
76.8, 79.9, 88.2, 97.3, 100.7, 128.7, 129.3, 133.7, 128.8, 141.2,
149.2, 163.0, 163.9, 170.0; FAB-MS m/z 687 (M+ + K). Anal.
Calcd for C31H48N2O9Si2: C, 57.38; H, 7.46; N, 4.32. Found:
C, 57.24; H, 7.55; N, 4.30.
1-[3,5-Bis-O-(tert-butyldimethylsilyl)-2-chloro-2-deoxyl-
1-C-hydroxymethyl-â-D-arabinofuranosyl]uracil (34). This
compound was prepared in 93% yield as a foam from 25 (23.5
mg, 0.041 mmol) by the procedure described for the prepara-
tion of 32. After stirring at rt for 3 h, the oxidation mixture
was quenched by adding saturated aqueous Na2S2O3. Extrac-
tion with EtOAc followed by HPLC purification (hexane/EtOAc
) 1/2) gave 34 (tR ) 12.3 min, 20.2 mg, 93%) as a foam: UV
(MeOH) λmax 261 nm (ꢀ 12700), λmin 230 nm (ꢀ 3300); 1H NMR
(CDCl3) δ 0.10, 0.13, and 0.16 (12H, each as s), 0.91 and 0.92
(18H, each as s), 3.14 (1H, br), 3.82 (1H, dd, J ) 10.7 and 6.6
Hz), 3.87 (1H, dd, J ) 10.7 and 4.8 Hz), 4.07 (1H, ddd, J )
6.6, 4.8, and 3.2 Hz), 4.19 (1H, s), 4.20 (1H, d, J ) 3.2 Hz),
4.47 (1H, dd, J ) 3.2 and 2.0 Hz), 4.57 (1H, d, J ) 2.0 Hz),
5.67 (1H, dd, J ) 8.3 and 1.8 Hz), 7.84 (1H, d, J ) 8.3 Hz),
8.71 (1H, br); 13C NMR (CDCl3) δ -5.4, -5.0, 4.7, 17.8, 18.4,
25.6, 25.9, 61.8, 63.1, 65.7, 79.8, 86.7, 99.9, 100.5, 142.5, 149.8,
165.0; FAB-MS m/z 559 (M+ + K). Anal. Calcd for C22H41-
ClN2O6Si2: C, 50.70; H, 7.93; N, 5.37. Found: C, 50.75; H, 8.17;
N, 5.23.
1-[3,5-Bis-O-(tert-butyldimethylsilyl)-2-deoxyl-â-D-psico-
furanosyl]uracil (35). A mixture of 34 (38 mg, 0.073 mmol),
Et3B (1 M THF solution, 146 µL, 0.146 mmol), and Bu3SnH
(39 µL, 0.146 mmol) in benzene (3.0 mL) was stirred at rt. After
2 h, the same amount of Et3B was added to the reaction
mixture, and stirring was continued for further 1 h. Evapora-
tion of the solvents followed by HPLC purification (hexane/
EtOAc ) 1/2) gave 35 (tR ) 9.8 min, 27 mg, 77%) as a foam:
1
UV (MeOH) λmax 263 nm (ꢀ 12700), λmin 234 nm (ꢀ 7100); H
NMR (CDCl3) δ 0.03, 0.04, 0.09, and 0.10 (12H, each as s),
0.86 and 0.90 (18H, each as s), 2.55 (1H, dd, J ) 14.8 and 2.4
Hz), 2.62 (1H, dd, J ) 8.0 and 5.0 Hz), 2.69 (1H, dd, J ) 14.8
and 5.6 Hz), 3.61-3.68 (2H, m), 3.92 (1H, dd, J ) 11.6 and
8.0 Hz), 3.97 (1H, dd, J ) 11.6 and 5.0 Hz), 4.15 (1H, dd, J )
5.8 and 4.0 Hz), 4.34 (1H, dt, J ) 5.6 and 2.4 Hz), 5.62 (1H, d,
J ) 8.4 Hz), 7.90 (1H, d, J ) 8.4 Hz), 8.31 (1H, br); 13C NMR
(CDCl3) δ -5.7, -5.6, -5.0, -4.8, 17.5, 18.2, 25.6, 25.7, 44.1,
62.4, 65.9, 72.5, 89.6, 99.9, 100.2, 142.1, 150.1, 164.4; FAB-
MS m/z 487 (M+ + H). Anal. Calcd forC22H42N2O6Si2: C, 54.29;
H, 8.70; N, 5.76. Found: C, 54.44; H, 9.03; N, 5.62.
Acknowledgment. This study was supported by
Grants-in-Aid (KAKENHI) from JSPS (Japan Society
for the Promotion of Science): No. 15790075 to H.K. and
No. 15590020 to H.T. We are also grateful to Ms. K.
Shiobara and Y. Odanaka (Center for Instrumental
Analysis, Showa University) for technical assistance
with NMR, MS, and elemental analyses.
1-[1-C-Acetoxymethyl-3,5-bis-O-(tert-butyldimethyl-
silyl)-2-O-benzoyl-â-D-arabinofuranosyl]uracil (33). This
compound was prepared in 91% yield as a foam from 23 (13.8
mg, 0.021 mmol) by the procedure described for the prepara-
tion of 32, except that the oxidative cleavage was performed
at 0 °C for 15 h. Acetylation mixture was purified by Florisil
column chromatography (hexane/EtOAc ) 5/1): UV (MeOH)
λmax 233 nm (ꢀ 16500) and 262 nm (ꢀ 10400), λmin 212 nm (ꢀ
1
9300) and 249 nm (ꢀ 8300); H NMR (CDCl3) δ -0.06, -0.03,
Supporting Information Available: X-ray crystallo-
graphic coordinates for compound 1 (CIF) and General Ex-
perimental Section. This material is available free of charge
via the Internet at http://pubs.acs.org.
0.14, and 0.23 (12H, each as s), 0.80 and 0.94 (18H, each as
s), 2.06 (3H, s), 3.59 (1H, dd, J ) 7.6 Hz), 3.78 (1H, dd, J )
10.4 and 5.8 Hz), 4.24 (1H, ddd, J ) 7.6, 5.8, and 1.6 Hz), 4.33
(1H, t, J ) 1.6 Hz), 4.74 and 4.96 (2H, each as d, J ) 12.0
Hz), 5.61 (1H, d, J ) 1.6 Hz), 5.74 (1H, dd, J ) 8.4 and 2.4
JO040260P
1690 J. Org. Chem., Vol. 70, No. 5, 2005