Second-generation synthesis of protected phosphonothiodifluoromethylene analogues of nucleoside-3′-phosphates

Article abstract of DOI:10.1016/j.tet.2004.03.082

A practical, eight-step synthesis of the key intermediate 14 in 19% overall yield from α-D-xylose is described. The preparation can be carried out on multi-gram scale and involves the use of the organomagnesium reagent 2d. The capability of derivative 14 to be transformed into the title compounds is examplified by the preparation of 15. Additionally, X-ray crystallography of intermediate 12 provided the first structural data on difluorophosphonothioates.

Full text of DOI:10.1016/j.tet.2004.03.082

Tetrahedron 60 (2004) 4895–4900
Second-generation synthesis of protected
phosphonothiodifluoromethylene analogues
of nucleoside-30-phosphates
*
Irina Kalinina, Arnaud Gautier, Carmen Salcedo, Jean-Yves Valnot and Serge R. Piettre
*
´ ´ ´ ´
Department of Chemistry, Laboratoire des Fonctions Azotees et Oxygenees Complexes, UMR CNRS 6014, IRCOF-Universite de Rouen,
`
Rue Tesnieres, F-76821 Mont Saint Aignan, France
Received 19 December 2003; revised 24 March 2004; accepted 29 March 2004
Abstract—A practical, eight-step synthesis of the key intermediate 14 in 19% overall yield from a-D-xylose is described. The preparation
can be carried out on multi-gram scale and involves the use of the organomagnesium reagent 2d. The capability of derivative 14 to be
transformed into the title compounds is examplified by the preparation of 15. Additionally, X-ray crystallography of intermediate 12
provided the first structural data on difluorophosphonothioates.
q 2004 Elsevier Ltd. All rights reserved.
Two decades ago Blackburn’s and McKenna’s pioneering
work on a,a-difluoromethylphosphonates aimed at the
development of an hydrolytically and enzymatically stable
mimic of the phosphate group.1 The strength of the P–C
bond as well as both the size and the electronic nature of the
fluorine atoms all contributed to reaching that goal.2
Numerous applications have since then been described
and, for instance, analogues of nucleosides mono-, di- and
triphosphates incorporating this functional group have been
prepared and reported to possess various bioactivities.3
However, the reactivity of reagents such as the lithium salt
of dialkyl difluoromethylphosphonate 1b has sometimes
restricted the generalized access to such analogues (Fig. 1).4
For instance, inertness of 1b toward secondary and tertiary
halides has long kept scientists from successfully preparing
difluorophosphonates bearing two substituents on the
carbon atom next to the CF2 unit. A number of solutions
to this problem have been published these past few years,
relying on either addition of phosphorus-centered radicals
onto gem-difluoroalkenes or reactions between organo-
metallic species and electrophilic sp2 carbon atoms.5,6
Additional approaches have involved exploiting cyclo-
addition processes with activated dienophiles, dipolaro-
philes or dienes substituted with a phosphonodifluoromethyl
unit, and coupling reactions with allylic substrates.7,8
Another elegant methodology based on a coupling reaction
with an halogenated sp2 carbon/[3,3]-sigmatropic rear-
rangement sequence has been described by Percy and
applied to the preparation of a 3,4-dideoxyribonolactone
bearing a phosphonodifluoromethyl unit on carbon 3.9 Our
interest to use this functional group as a replacement of the
phosphate in oligonucleotides (with implications in the
antisense and antigene strategies) led us to work out a first
and efficient approach to the preparation of the title
compounds, relying on the use of the lithium salt of diethyl
difluoromethylphosphonothioate 2b and the subsequent
transformation of the PvS bond into a PvO bond.10
Besides the stability of the C–P bond and the documented
isosteric and isoelectronic natures of difluorophosphonates
and phosphates, the effect of replacing the 30-oxygen atom
of a nucleotide with a CF2 unit may play a significant role on
the conformation of the ribose, in an oligomer. While it is
known that the replacement of the oxygen with a less
electronegative CH2 moiety induces a more pronounced
shift toward the required, usual 30-endo ring pucker, it has
been shown that simple phosphonates are not optimal, due,
among others, to an increased interbase distance. The two
fluorine atoms might thus have an indirect, positive effect on
the adopted conformation of the ribose and on the stability
Figure 1. Structures of 1a-1d and 2a-2d.
Keywords: Phosphonothiodifluoromethylene; Nucleotide analogue;
Organomagnesium reagent.
*
Corresponding authors. Tel.: þ33-235-52-29-70; fax: þ33-235-52-29-
71; e-mail addresses: serge.piettre@univ-rouen.fr;
arnaud.gautier@univ-rouen.fr
0040–4020/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2004.03.082

4896
I. Kalinina et al. / Tetrahedron 60 (2004) 4895–4900
Scheme 2. Reagents and conditions: (i) acetone, CuSO4, H2SO4;
(ii) MeOH/HCl, or MeCN/H2O, CAN (3 mol%); (iii) 4-ClBzCl, TEA,
CH2Cl2; (iv) C3N3O3Cl3, TEMPO (4 mol%), CH2Cl2; (v) isopropMgCl,
2c, THF, 245 8C; (vi) isopropMgCl, THF, 0 8C, then ClCOCO2Me;
(vii) (n-Bu)3SnH/AIBN; toluene; (viii) AcOH/Ac2O/H2SO4; (ix) (U)TMS/
TMSOTf, 1,2-dichloroethane.
Scheme 1.
of the duplex RNA-oligomer.11 The need for large amounts
of the nucleotide analogues led us to develop a second-
generation, more efficient preparation of the title com-
pounds, the account of which is reported in this note.
according to literature procedure (Scheme 2).12 Selective,
room temperature deprotection of the 3,5-isopropylidene
unit was smoothly performed using cerium ammonium
nitrate (CAN) and quantitatively afforded diol 9.13 The
choice of a 4-chlorobenzoyl group to protect the primary
alcohol stemmed from the known crystalline nature of both
alcohol 10 and ketone 11, and from the general propensity
of such derivatives to be crystalline (vide infra).14 Alcohol
10 was oxidized by a modification of Matsuda’s procedure
calling for the use of 2,2,6,6-tetramethyl-1-piperidinyloxy
(TEMPO).14,15
Two substrates, namely a-D-glucose 3 and a-D-xylose 6,
had been selected in the first-generation synthesis (Scheme
1). These choices, while allowing us to get the target
molecules, ultimately turned out to feature several draw-
backs for a large scale preparation of a protected ribose
encom0passing a phosphonodifluoromethylene unit on car-
bon 3 . Thus, using the readily available a-D-glucose
implied unavoidable functional group manipulation and
oxidation of the 5,6-diol unit to get the 5-carbon ribose
skeleton. In addition, the overall transformation included
the use of the Dess–Martin reagent, and of tert-butyl-
lithium, and required several separative purifications by
chromatography. The six-step synthesis nevertheless
delivered the key difluorophosphonothioate 5a in 12%
yield from 4 and translated into a global yield of 8% from
a-D-glucose 3. The synthetic route from a-D-xylose 6
suffered from the same use of tert-butyllithium and
displayed similar efficiency, compound 5b being isolated
in 8% overall yield.10
The production of difluorophosphonothioate 12 was then
tackled. The lithiated reagent 2b is known to be efficiently
generated from bromide 2c via an halogen–metal exchange
reaction involving tert-butyllithium.10 The need to avoid the
latter reagent led us to reinvestigate the alternative provided
by the use of n-butyllithium (n-BuLi). Sequential addition
of (i) n-BuLi and (ii) benzaldehyde to a solution of bromide
2c in tetrahydrofuran (THF) at 278 8C invariably led to
mixtures of 16 and 17, the varying amount of 16 being
directly proportional to the stirring time separating the
additions of n-BuLi and benzaldehyde; not unexpectedly,
the n-butyl bromide formed by the halogen–metal exchange
process competitively reacted with 2b (Scheme 3). Addition
of n-BuLi to a mixture of 2c and benzaldehyde reproducibly
gave alcohol 17, in fair isolated yield (60%) (Table 1, entry
1). Using acetophenone led to an essentially identical
This last preparation nevertheless constituted the basis for
the second-generation synthesis. Classical protection of
a-D-xylose 6 quantitatively furnished bisacetonide 8,

I. Kalinina et al. / Tetrahedron 60 (2004) 4895–4900
4897
Scheme 3. Reagents and conditions: (i) n-BuLi, 278 8C; (ii) benzaldehyde
or acetophenone; H3Oþ.
Table 1. Yields of adducts from the reaction between 2b or 2d and
benzaldehyde, acetophenone or ketone 11
Entry
Reagent
Electrophile
Product
Yields (%)a
1
2
3
4
5
6
2b
2b
2b
2d
2d
2d
C6H5CHO
C6H5C(O)Me
11
C6H5CHO
C6H5C(O)Me
11
17
18
12
17
18
12
60
57
0
75
54
65–72
Figure 2. ORTEP drawing of adduct 12.
a
Isolated yields.
angle formed by these two bonds: it is found to be 120.38
and is thus wider than those reported for both the phosphate
(118.58) and the difluorophosphonate (116.58).
result and alcohol 18 was isolated in 57% yield (entry 2).
However, the use of the structurally and functionally more
complex ketone 11 invariably led to intractable mixtures
(entry 3). These unsatisfactory results prompted us to
investigate the scope of the corresponding magnesium
species 2d. So far, this reagent has only been reported in
the context of a relatively unsuccessful interaction with
ketone 4.16 Attempts to directly generate 2d from 2c and
activated magnesium at room temperature resulted in red–
black solutions of unpredictable behavior and characterized
by competing decomposition processes (19F an 31P NMR
spectrometries). Halogen–metal exchange proved to be
more effective: addition of a freshly prepared solution of
isopropylmagnesium chloride in diethyl ether to a solution
of 2c in THF, followed by interaction of the resultant
organomagnesium reagent 2d with benzaldehyde led to the
isolation of alcohol 17 in 75% yield (entry 4).17 The use of
acetophenone also led to the expected adduct 18, albeit in
a somewhat lower yield (entry 5).18 To our delight,
interaction between 2d and ketone 11 delivered a single
adduct in 65–72% isolated yield (entry 6).
The deoxygenation radical process allowing the conversion
of alcohol 12 into phosphonate 13 is based on our previous
work.10 This time however, isopropylmagnesium chloride
was used as base instead of n-BuLi (Scheme 2). Quenching
the alcoholate with methyl oxalyl chloride yielded the
oxalate intermediate, which was treated with tri-n-butyltin
hydride and a catalytic amount of azobisisobutyronitrile
(AIBN), without further purification. Thus refluxing this
mixture in toluene for 2 h led to the formation of 13 as the
exclusive product. Again, hydrogen quenching of the
intermediate radical occurred exclusively from the convex
face to cleanly produce the stereochemical inversion at C-3.
It is noteworthy that the tin by-products could be distilled
off under vacuum from the crude mixture to leave the
product virtually free from tin impurities.19 Removal of the
1,2-isopropylidene group and protection of the hydroxyl
functions as acetates was carried out in a single operation
and delivered difluorophosphonothioate 14 in 77% isolated
yield. Here again, silica was used to filter the product as no
by-product was present in the crude material.
The choice of the 4-chlorobenzoyl unit as protecting group
turned out to be critical as (i) no product arising from a
competitive reaction on the ester group was detected, and
(ii) alcohol 12 crystallized upon standing, affording material
suitable for X-ray analysis. The ORTEP drawing unam-
biguously demonstrates the total stereoselectivity of the
process, resulting from the exclusive addition of 2d on the
convex face of 11 (Fig. 2). Some additional interesting
features can be selected from the X-ray data. Thus, the
C3 –CF2 and CF2–P bond lengths (1.88 and 1.51 A,
˚
respectively) compare favorably with values determined
by Chambers and O’Hagan on a,a-difluorophosphonates
˚
(1.85 and 1.50 A, respectively): the distances between these
Finally, the capability of compound 14 to be trans-
formed into a modified nucleoside was verified by using
Vorbruggen’s procedure.20 Treatment of 14 with the bis-
trimethylsilyl derivative of uracil under Lewis acid catalysis
afforded the fully protected difluorophosphonothioate 15
(76% yield).
The key intermediate14was thuspreparedfroma-D-xylose in
eight steps and 19% overall yield. The synthesis is easily
amenable to multigram quantity scale as no separation by
chromatography is required; the need for tert-BuLi has been
eliminated by using the organomagnesium reagent 2d and the
use of the 4-chlorobenzoyl group allows easy Kugel–Rohr
distillation of the tin by-products resulting from the
atoms are little influenced by the replacement of the double-
bonded oxygen atom with a sulfur. The most interesting
feature of difluorophosphonothioate seems to concern the

4898
I. Kalinina et al. / Tetrahedron 60 (2004) 4895–4900
deoxygenation process. Additionally, it provided with a
crystalline adduct 12 whose X-ray data yielded the first
structural comparison between the difluorophosphonothioate
and the corresponding fully oxygenated functional group.
removed by Kugel–Rohr distillation (90 8C/0.2 mbar) to
leave 15.8 g of a residual brown oil. This product is then
filtered through a 15 cm wide and 5 cm long silica path.
Washing is carried out with a 1:7 mixture of ethyl acetate/
cyclohexane (750 mL) to yield 10.6 g of 12 as a colorless oil
that spontaneously crystallizes at room temperature (yield:
65%). Another run performed on a 5.0 g scale gave 72%
yield of 12. 1H NMR (200 MHz) d 7.96 (d, J¼8.8 Hz, 2H);
7.37 (d, J¼8.8 Hz, 2H); 5.92 (d, J¼4.4 Hz, 1H); 5.29 (d, J¼
4.4 Hz, 1H); 4.70 (td, J¼5.8, 12.4 Hz, 1H); 4.6 (td, J¼
3.0 Hz, 1H); 4.6–4.1 (m, 5H); 3.48 (s, 1H); 1.58(s, 3H); 1.4
(s, 3H); 1.34 (td, J¼3.7 Hz, 6H). 13C NMR (50 MHz) d
165.0; 139.1; 131.0; 128.4; 128.1; 120.0 (ddd, J¼174, 258,
276 Hz); 113.0; 104.4; 82.7; 80.9 (ddd, J¼12.2, 21.2,
24.3 Hz); 78.6; 64.8 (d, J¼6.0 Hz); 64.4 (d, J¼6.0 Hz); 62.7
(d, J¼12.1 Hz); 26.5; 26.2; 15.9; 15.8. 31P NMR (81 MHz)
d 71.8 (t, J¼99.0 Hz). 19F NMR (188 MHz) d 50.7 (dd,
J¼101, 113 Hz); 47.0 (dd, J¼101, 113 Hz). IR (neat) 3519,
1. Experimental
1.1. 1,2-O-Isopropylidene-a-D-xylofuranose (9)
To a solution of 1,2-3,5 di-O-isopropylidene-a-D-xylose
(66 g, 0.287 mol) in HPLC grade water (265 mL) and
acetonitrile (250 mL) is added CAN (4.7 g, 8.6 mmol,
3 mol%). The solution is stirred for 18 h at room
temperature until monitoring by thin-layer chromatography
(tlc) indicates completion. Aqueous ammonium hydroxide
(NH4OH) (20 mL) is added, the resultant yellow–orange
suspension is filtered through a mixture of celite and silica
(9/1 w/w) which is then washed with MeOH (150 mL).
Evaporation of the volatiles under reduced pressure and
lyophilisation of the resultant aqueous layer yield 9 as a pale
yellow oil. (53.4 g, 98%), identical in every respect with a
sample prepared according to literature procedures.14
2989, 2939, 1726, 1592, 1378, 1273, 1090, 1013 cm21
.
[a]5D89¼þ21.6 (c¼1.1, CH2Cl2). Mp¼80–81 8C. Anal.
Calcd for C20H26ClF2O8PS: C, 45.24; H, 4.90; S, 6.03.
Found: C, 45.37; H, 4.96; S, 5.88.
1.2.2. 3-Deoxy-3-(O,O-diethylphosphonothio)difluoro-
methyl-1,2-O-isopropylidene-5-O-(4-chlorobenzoyl)-a-
D-ribofuranose (13). Isopropylmagnesium chloride
(9.7 mL of a 1.3 M solution in ether, 12.6 mmol) is added
at 0 8C to THF (55 mL) under inert atmosphere. Alcohol 12
(5.7 g, 10.5 mmol) in THF (55 mL) is added dropwise and
the resultant solution is stirred for 1 h at the same
temperature before adding freshly distilled methyl oxalyl
chloride (2.9 mL, 31.5 mmol). The solution is stirred for an
additional 20 min at 0 8C, warmed up to room temperature
and evaporated. The mixture is poured into water (35 mL)
and extracted with AcOEt (100 mL). The separated organic
layer is dried (MgSO4) and evaporated under reduced
pressure at 60 8C to give 6.56 g of a pale yellow oil. This
crude material is dissolved in toluene (66 mL) and the
solution is refluxed under nitrogen for 20 min. Tri-n-butyltin
hydride (3.8 mL, 12.1 mmol) and AIBN (1.0 g, 5.4 mmol)
are then added and the solution is refluxed for two more
hours. The solution is evaporated and the residue is Kugel–
Rohr distilled (130–140 8C, 0.2 mbar) to eliminate the tin
by-products. The resultant, undistilled brownish oil is then
filtered through silica (15 cm wide and 5 cm high) and
washed with ethyl acetate/cyclohexane (1:20; 500 mL) to
1.2. 5-O-(4-Chlorobenzoyl)-1,2-O-isopropylidene-a-D-
pentofuranose-3-urose (11)
Dichloromethane (950 mL), alcohol 10 (101 g, 0.31 mol)
and trichlorocyanuric acid (110 g, 0.47 mol) are sequen-
tially added under nitrogen into a 2 L round bottomed
flask equipped with a mechanical stirrer and a condenser.
TEMPO (2.0 g, 14 mmol, 4.5 mol%) is then added in 10
portions (caution: exothermic reaction) and the reaction is
stirred for 12 h. The reaction mixture is filtered through
celite and eluted with CH2Cl2 (2£250 mL). The organic
layer is sequentially washed with saturated NaHCO3
(100 mL), 1.0 M aqueous HCl (200 mL) and brine
(200 mL), dried over MgSO4 and evaporated under reduced
pressure. The crude brownish oil is dissolved in chloroform
(100 mL), and added dropwise to vigorously stirred
n-heptane (1 L). After completion of the addition, the
flask is cooled down to 4 8C overnight and the pale yellow
solid (67.0 g) is collected by filtration. Concentration of the
mother liquor affords 27 g of brown oil that is purified as
above from chloroform (27 mL) and n-heptane (270 mL) to
give an additional 10.5 g batch (total yield: 77.5 g, 77%).
Analytical data identical with those from the literature.14
1
afford 3.4 g of 13 as a pale yellow liquid (64%). H NMR
(200 MHz) d 7.95 (d, J¼8.8 Hz, 2H); 7.37 (d, J¼8.8 Hz,
2H); 5.81 (d, J¼3, 4 Hz, 1H); 4.92 (t, J¼3.4 Hz, 1H); 4.7
(m, 2H); 4.2 (m; 5H); 3.0 (m, 1H); 1.53 (s, 3H); 1.31 (s, 3H);
1.31 (t, J¼7.0 Hz, 6H). 13C NMR (50 MHz) d 165.1; 139.4;
131.0; 128.6; 128.2; 119.6 (ddd, J¼276, 258, 174 Hz);
113.0; 104.7; 79.5; 79.4 (d, J¼7.6 Hz); 74.7 (dd, J¼7.8,
3.3 Hz); 65.4; 65.2; 64.4 (dd, J¼16.7, 4.5 Hz); 47,0 (td,
J¼21.2, 18.2 Hz); 26.7; 26.3; 16.1 (d, J¼4.6 Hz); 15.9 (d,
J¼3.0 Hz). 31P NMR (81 MHz) d 73,8 (dd, J¼102, 113 Hz).
19F NMR (188 MHz) d 60 (ddd, J¼298, 112, 6.8 Hz); 49,0
(ddd, J¼298, 102, 6.8 Hz). IR (Neat) 2986, 1724, 1595,
1381, 1274, 1117, 1020 cm21. [a]5D89¼þ42.4 (c¼2.51;
CH2Cl2). Anal. Calcd for C20H26ClF2O7PS: C, 46.64; H,
5.05; S, 6.22. Found: C, 46.13; H, 4.72; S, 5.94. HRMS
(E.I., 70 eV). Calcd for C19H23F2O7PSCl: 499.0537. Found:
499.0558.
1.2.1. 3-(O,O-Diethylphosphonothio)difluoromethyl-1,2-
O-isopropylidene-5-O-(4-chlorobenzoyl)-a-D-allofura-
nose (12). Isopropylmagnesium chloride (47 mL of a 1.1 M
solution in ether, 51.7 mmol) is added at room temperature
to anhydrous THF (150 mL) under inert atmosphere and the
solution is cooled down to 278 8C. A solution of 2c (13 g,
46.0 mmol) in THF (100 mL) is slowly added. After 5 min,
ketone 11 (10 g, 30.6 mmol) in THF (50 mL) is added
dropwise over 10 min. After another 10 min period of time,
the solution is warm up to 245 8C and stirred for 5 h.
Saturated aqueous NH4Cl (70 mL) is added and the resultant
mixture is warmed up to room temperature. The solvent is
then evaporated and the residue extracted with AcOEt
(2£200 mL). The organic layer is dried over MgSO4 and
evaporated. Compound 2a and unreacted substrate 2c are

I. Kalinina et al. / Tetrahedron 60 (2004) 4895–4900
4899
1.2.3. 5-O-(4-Chlorobenzoyl)-1,2-di-O-acetyl-3-deoxy-3-
(O,O-diethylphosphonothio)difluoromethyl-b-D-ribo-
furanose (14). Difluorophosphonothioate 13 (3.3 g,
6.4 mmol) is dissolved in a mixture of acetic acid (30 mL)
and acetic anhydride (30 mL). The solution is cooled down
to 0 8C and sulfuric acid (1 mL) is added. The solution is
stirred at room temperature overnight. The stirring mixture
is then added over 30 min to a cooled mixture of water
(100 mL), NaHCO3 (84 g), ice (100 g) and ethyl acetate
(200 mL). After completion of the addition, the solution is
warmed up to room temperature and stirred until the pH
reaches 8–9. The organic layer is separated and the aqueous
phase is extracted with ethyl acetate (250 mL). The
combined organic layers are extracted with saturated
NaHCO3 (120 mL), washed with brine (25 mL) and dried
over Na2SO4. Filtration through silica (6 cm wide and 3 cm
long) and washing of the solid layer with ethyl acetate/
cyclohexane (1:5) (300 mL) give 2.77 g (77%) of colorless,
Found: C, 45.19; H, 4.37; S, 5.14. HRMS (DCI, 200 eV)
m/z. Calcd for C23H27N2O9F2PSCl 611.0787. Found:
611.0831.
Acknowledgements
I.K. is grateful to the C. N. R. S. for support during her
stay in France (poste rouge), and C.S. is indebted to the
ERASMUS program (European Union). We thank Pr
A. Matsuda (Hokkaido University) and Dr. M. Nomura
(Taiho Pharmaceutical Co) for helpful discussions.
References and notes
1. (a) Blackburn, G.; England, D. A.; Kolkmann, F. J. Chem.
Soc., Chem. Commun. 1981, 930–932. (b) Blackburn, G.;
Kent, D. E.; Kolkmann, F. J. Chem. Soc., Chem. Commun.
1981, 1188–1190. (c) McKenna, C. E.; Shen, P.-D. J. Org.
Chem. 1981, 46, 4573–4576.
1
oily 14. H NMR (200 MHz) d 8.0 (d, J¼8.4 Hz, 1H); 7.4
(d, J¼8.4 Hz, 1H); 6.05 (d, J¼1.5 Hz, 1H); 4.48 (d, J¼
4.4 Hz, 1H); 4.9 (m, 1H); 4.75 (d, J¼12.4 Hz, 1H); 4.4–4.1
(m, 5H); 3.6–3.4 (m, 1H); 2.09 (s, 3H); 1.90 (s, 3H); 1.32 (t,
J¼7.0 Hz, 3H); 1.31 (t, J¼7.0 Hz, 3H). 13C NMR (50 MHz)
d 169.4; 168.7; 164.9; 139.6; 131.0; 128.7; 128.1; 119.3
(ddd, J¼276, 259, 179 Hz); 98.2; 77.5 (t, J¼3.0 Hz); 75.0
(d, J¼7.6 Hz); 65.2 (d, J¼16.7 Hz); 65.2 (d, J¼16.7 Hz);
64.6 (d, J¼7.6 Hz); 43.1 (td, J¼22.7, 19.7 Hz); 20.8; 20.7;
16.0; 15.9. 31P NMR (81 MHz) d 74.1 (dd, J¼111.3,
102.3 Hz). 19F NMR (188 MHz) d 57.2 (ddd, J¼298, 111.7,
10.2 Hz); 47.8 (ddd, J¼298, 111.7, 20.3 Hz). IR (neat)
2959, 2927, 1753, 1728, 1595, 143, 1371, 1273, 1235, 1092,
1018 cm21. [a]D589¼þ5.18 (c¼2.8; CH2Cl2). Anal. Calcd
for C21H26ClF2O9PS: C, 45.12; H, 4.66. Found: C, 45.01; H,
4.80.
2. (a) Chambers, R. D.; Jaouhari, R.; O’Hagan, D. Tetrahedron
1989, 45, 5101–5108. (b) Chambers, R. D.; O’Hagan, D.;
Lamont, R. B.; Jain, S. C. J. Chem. Soc., Chem. Commun.
1990, 1053–1054. (c) Nieschalk, J.; Batsanov, A. S.;
O’Hagan, D.; Howard, J. A. K. Tetrahedron 1996, 52,
165–176. (d) Tatcher, G. R.; Campbell, A. S. J. Org. Chem.
1993, 58, 2272–2281. (e) Berkowitz, D. B.; Bose, M.
J. Fluorine Chem. 2001, 112, 13–33.
3. (a) Blackburn, G. M.; Kent, D. E.; Kolkmann, F. J. Chem.
Soc., Perkin Trans. 1 1984, 1119–1125. (b) Blackburn, G. M.;
Guo, M.-J.; Langston, S. P.; Taylor, G. E. Tetrahedron Lett.
1990, 31, 5637–5640. (c) Blackburn, G. M.; Langston, S. P.
Tetrahedron Lett. 1991, 32, 6425–6428. (d) Matulic-Adamic,
J.; Usman, N. Tetrahedron Lett. 1994, 35, 3227–3230. (e)
Matulic-Adamic, J.; Haeberli, P.; Usman, N. J. Org. Chem.
1995, 60, 2563–2569. (f) Levy, S. G.; Wasson, B.; Carson,
D. A.; Cottam, H. B. Synthesis 1996, 843–846.
1.2.4. 1-N-[2-O-Acetyl-5-(4-chlorobenzoyl)-3-deoxy-3-
(O,O-diethylphosphonothio)difluoromethyl-b-D-ribo-
furanosyl]uridine (15). TMSOTf (0.71 mL, 3.9 mmol) is
added dropwise to a cold solution (0 8C) of 14 (441 mg,
0.79 mmol)
and
bis(trimethylsilyl)uracil
(607 mg,
4. (a) Obayashi, M.; Ito, E.; Kondo, K. Tetrahedron Lett. 1982,
23, 2323–2326. (b) Burton, D. J.; Yang, Z.-Y. Tetrahedron
1992, 48, 189–275.
2.4 mmol) in dry 1,2-dichloroethane (15 mL). The mixture
is stirred at room temperature during 3 h, after which period
of time it is poured into CH2Cl2 (50 mL) and washed with
saturated aqueous NaHCO3 (20 mL). The organic layer is
dried over Na2SO4 and evaporated to give 420 mg of a thick
oil. Chromatography on silica and elution with dichloro-
methane/methanol (98:2) give 367 mg (76%) of 15 as a
5. (a) Piettre, S. R. Tetrahedron Lett. 1996, 37, 2233–2236.
(b) Herpin, T. F.; Houlton, J. S.; Motherwell, W. B.; Roberts,
B. P.; Weibel, J.-M. J. Chem. Soc., Chem. Commun. 1996,
613–614. (c) Herpin, T. F.; Motherwell, W. B.; Roberts, B. P.;
Roland, S.; Weibel, J.-M. Tetrahedron 1997, 53,
15085–15100. (d) Kovensky, J.; McNeil, M.; Sinay¨, P.
J. Org. Chem. 1999, 64, 6202–6205.
1
white solid. H NMR (200 MHz) d 8.96 (s, 1H); 7.98 (d,
J¼8.8 Hz, 1H); 7.41 (d, J¼8.8 Hz, 1H); 7.22 (d, J¼3.3 Hz,
1H); 5.72 (d, J¼4.7 Hz, 1H); 5.59 (dd, J¼8.0, 2.2 Hz, 1H);
5.0 (s, 1H); 4.9–4.7 (m, 2H); 4.51 (dd, J¼12.8, 4.4 Hz, 1H);
4.3–4.1 (m, 4H); 3.8 (m, 1H); 2.112 (s, 3H); 1.31 (t, J¼
7.3 Hz, 3H); 1.30 (t, J¼7.3 Hz, 3H). 13C NMR (50 MHz) d
169.8; 165.1; 164.0; 149.7; 141.7; 139.6; 131.0; 128.8;
127.8; 119.5 (ddd, J¼274, 262, 177 Hz); 102.5; 93.3; 77.1
(d, J¼9 Hz); 75.3 (d, J¼6 Hz); 65.0 (d, J¼11 Hz); 64.7 (d,
J¼11 Hz); 64.3; 43.4 (q, J¼18 Hz); 20.8; 16.0; 15.9. 31P
NMR (81 MHz) d 73.8 (t, J¼108 Hz). 19F NMR (188 MHz)
d 56.1 (ddd, J¼295, 108, 13.5 Hz); 49.9 (ddd, J¼295, 108,
20.3 Hz). IR (neat) 2991, 1717, 1693, 1456, 1377, 1272,
1231, 1091, 1014 cm21. UV–vis lmax (CH2Cl2) 245 nm.
[a]5D89¼þ16.28 (c¼1.4; CH2Cl2). Mp 62–64 8C. Anal.
Calcd for C20H26ClF2O7PS: C, 45.22; H, 4.29; S, 5.25.
6. (a) Lequeux, T. P.; Percy, J. M. Synlett 1995, 361–362.
(b) Berkowitz, D. B.; Eggen, M.; Shen, Q.; Shoemaker, R. K.
J. Org. Chem. 1996, 61, 4666–4675. (c) Blades, K.; Cockerill,
G. S.; Easterfield, H. J.; Lequeux, T. P.; Percy, J. M. Chem.
ˆ
Commun. 1996, 1615–1616. (d) Blades, K.; Lapotre, D.;
Percy, J. M. Tetrahedron Lett. 1997, 38, 5895–5898.
(e) Blades, K.; Percy, J. M. Tetrahedron Lett. 1998, 39,
9085–9088. (f) Murano, T.; Muroyama, S.; Yokomatsu, T.;
Shibuya, S. Synlett 2002, 1657–1660. (g) Murano, T.; Yuasa,
Y.; Muroyama, S.; Yokomatsu, T.; Shibuya, S. Tetrahedron
2003, 59, 9059–9073. (h) Murano, T.; Yuasa, Y.;
Kobayakawa, H.; Yokomatsu, T.; Shibuya, S. Tetrahedron
2003, 59, 10223–10230.
7. (a) Blades, K.; Lequeux, T. P.; Percy, J. M. Chem. Commun.

4900
I. Kalinina et al. / Tetrahedron 60 (2004) 4895–4900
1996, 1457–1458. (b) Yokomatsu, T.; Katayama, S.; Shibuya,
15. De Lucca, L.; Giacomelli, G.; Porcheddu, A. Org. Lett. 2001,
3, 3041–3043.
S. Chem. Commun. 2001, 1878–1879. (c) Yokomatsu, T.;
Suemune, K.; Murano, T.; Shibuya, S. Heterocycles 2002, 56,
273–282. (d) Butt, A. H.; Kariuki, B. M.; Percy, J. M.;
Spencer, N. S. Chem. Commun. 2002, 682–683.
16. The desired adduct was isolated in only 21% yield. See Ref.
10.
17. For the use of the corresponding oxygenated reagent, see:
Waschbu¨sch, R.; Samadi, M.; Savignac, P. J. Organomet.
Chem. 1997, 529, 267–278.
8. Yokomatsu, T.; Kato, J.; Sakuma, C.; Shibuya, S. Synlett
2003, 1407–1410.
9. Butt, A. H.; Percy, J. M.; Spencer, N. S. Chem. Commun. 2000,
1691–1692.
18. Presumably, 2d acted as a base in a competitive manner.
19. Attemps to eliminate the use of tin derivatives by substituting
tri-n-butyltin hydride with other hydrogen donors ((Me3Si)3
SiH, Et3SiH, Ph2SiH2, (EtO)2P(O)H, (EtO)2P(S)H or
H2PO2Na) gave lower yields or were unsuccessful.
20. (a) Niedballa, U.; Vorbru¨ggen, H. Angew. Chem., Int. Ed.
Engl. 1970, 9, 461–462. (b) Niedballa, U.; Vorbru¨ggen, H.
J. Org. Chem. 1974, 39, 3654–3660. (c) Niedballa, U.;
Vorbru¨ggen, H. J. Org. Chem. 1974, 39, 3660–3663.
(d) Niedballa, U.; Vorbru¨ggen, H. J. Org. Chem. 1974, 39,
3664–3667. (e) Niedballa, U.; Vorbru¨ggen, H. J. Org. Chem.
1974, 39, 3668–3671. (f) Niedballa, U.; Vorbru¨ggen, H.
J. Org. Chem. 1974, 39, 3672–3674. (g) Niedballa, U.;
Vorbru¨ggen, H. J. Org. Chem. 1976, 41, 2084–2086.
(h) Vorbru¨ggen, H.; Bennua, B. Tetrahedron Lett. 1978, 39,
1339–1342. (i) Vorbru¨ggen, H.; Krolikiewicz, K.; Bennua, B.
10. Lopin, C.; Gautier, A.; Gouhier, G.; Piettre, S. R. J. Am. Chem.
Soc. 2002, 124, 14668–14678.
11. (a) Altona, C.; Sundaralingam, M. J. Am. Chem. Soc. 1972, 94,
8205–8212. (b) Altona, C.; Sundaralingam, M. J. Am. Chem.
Soc. 1973, 95, 2333–2334. (c) Thibaudeau, C.; Plavec, J.;
Garg, N.; Papchikhin, A.; Chattopadhyaya, J. J. Am. Chem.
Soc. 1994, 116, 4038–4043. (d) Freier, S. M.; Altmann, K.-H.
Nucleic Acids Res. 1997, 25, 4429–4443.
12. Bozo, E.; Boros, S.; Kuszmann, J.; Gacs-Baitz, E.; Parkanyi,
L. Carbohydr. Res. 1998, 308, 297–310.
13. Marko, I. E.; Ates, A.; Gautier, A.; Leroy, B.; Plancher, J.-M.;
Quesnel, Y.; Vanherck, J.-C. Angew. Chem., Int. Ed. 1999, 38,
3207–3209.
14. Benzoate 10 was prepared in 79% isolated yield according to
literature procedure. See Nomura, M.; Sato, T.; Washinosu,
M.; Tanaka, M.; Asao, T.; Shuto, S.; Matsuda, A. Tetrahedron
2002, 58, 1279–1288.
¨
Chem. Ber. 1981, 114, 1234–1255. (j) Vorbru¨ggen, H.; Holfe,
G. Chem. Ber. 1981, 114, 1256–1268.