2-Deoxyribose as a rich source of chiral 5-carbon building blocks

Article abstract of DOI:10.1021/jo0712143

(Chemical Equation Presented) We have developed concise routes to a number of useful chiral 5-carbon synthetic building blocks using readily available O-1-methyl-2-deoxyribose as starting material. Novel transformations include the use of indium triflate

Full text of DOI:10.1021/jo0712143

2-Deoxyribose as a Rich Source of Chiral 5-Carbon Building
Blocks§
Dengjin Wang* and William A. Nugent*,‡
Bristol-Myers Squibb Company, Process Research and DeVelopment, P.O. Box 4000,
Princeton, New Jersey 08543
william_nugent@Vrtx.com; dengjin.wang@bms.com
ReceiVed June 7, 2007
We have developed concise routes to a number of useful chiral 5-carbon synthetic building blocks using
readily available O-1-methyl-2-deoxyribose as starting material. Novel transformations include the use
of indium triflate to catalyze the oxidation of a methyl furanoside to the corresponding lactone with
MCPBA and the Vasella-type fragmentation of a 5-iodo furanoside using chromium(II) chloride when
zinc proved ineffective. In addition, 3,4-disubstitued piperidine derivatives were prepared without hydroxyl
group protection via a simple reductive amination reaction.
Introduction
propanediol via a route involving the Katsuki-Sharpless
asymmetric epoxidation and multiple chromatographic purifica-
tions. In the current paper we report an alternative synthesis of
2 starting with the commercially available O-methyl glycoside
of 2-deoxyribose. This improved route affords 2 in four simple
steps without the need for chromatography.
2-Deoxyribose (1) is unique among deoxysugars in that it is
available in commercial quantities at relatively low cost. This
reflects the fact that it is manufactured by alkaline degradation
of inexpensive glucose. Consequently synthetic organic chemists
have utilized 2-deoxyribose as a starting material in the total
syntheses of molecules as diverse as brevetoxin B,1 disparlure,2
and leukotriene B4.3 Nevertheless we suggest that, given the
availability and highly functional nature of 2-deoxyribose, it
remains underutilized in organic synthesis.
The related (S)-aldehyde 3 has been utilized in the synthesis
5
of 1R,25-dihydroxy-22-oxavitamin D3 and in a ring-closing
metathesis approach to dictyostatin.6 In both applications, 3 was
again prepared by Sharpless AE.7 Again, we find that use of
2-deoxyribose as ultimate starting material provides greatly
simplified access to 3.
In addition to these acyclic derivatives, 2-deoxyribose based
syntheses were also developed for piperidines 4 and 5. The
A case in point is carboxylic acid 2, a key intermediate in a
recent synthesis of (+)-diplodialides B and C.4 This excellent
chiral building block was prepared in 11 steps from 1,3-
(5) Hatakeyama, S.; Okano, T.; Maeyama, J.; Tomoyuki, E.; Hiyamizu,
H.; Iwabuchi, Y.; Nakagawa, K.; Ozono, K.; Kawase, A.; Kubodera, N.
Bioorg. Med. Chem. 2001, 9, 403.
(6) Kangani, C. O.; Brueckner, A. M.; Curran, D. P. Org. Lett. 2005, 7,
379.
§ Dedicated to Prof. Jay K. Kochi on the occasion of his 80th birthday.
‡ Current address: Vertex Pharmaceuticals, Ins., Cambridge, MA.
(1) Matsuo, G.; Kawamura, K.; Hori, N.; Matsukura, H.; Nakata, N. J.
Am. Chem. Soc. 2004, 126, 14374.
(2) Kang, S.-K.; Kim, Y.-S.; Lim, J.-S.; Kim, K.-S.; Kim, S.-G.
Tetrahedron Lett. 1991, 32, 363.
(3) Guindon, Y.; Delorme, D.; Lau, C. K.; Zamboni, R. J. Org. Chem.
1988, 53, 267.
(4) Sharma, G. V. M.; Reddy, K. L. Tetrahedron: Asymmetry 2006, 17,
3197.
(7) The (R)-enantiomer of aldehyde 3 has been prepared by a sequence
of enzymatic resolution of the tert-butyl â-ketoester, chromatographic
separation, DIBAL reduction, and silylation. Tan, C.-H.; Holmes, A. B.
Chem. Eur. J. 2001, 7, 1845. See also: Vrielynck, S.; Vandewalle, M.;
Garcia, A. M.; Mascarenas, J. L.; Mourino, A. Tetrahedron Lett. 1995, 36,
9023.
10.1021/jo0712143 CCC: $37.00 © 2007 American Chemical Society
Published on Web 08/09/2007
J. Org. Chem. 2007, 72, 7307-7312
7307

Wang and Nugent
SCHEME 1. Preparation of Crystalline Lactone 10
racemate of epoxide 5 is a precursor to the gastrointestinal
stimulant cisapride,8 while racemic 4 has been shown to be an
invaluable carbohydrate scaffold for drug discovery in the area
of extracellular signaling.9 Enantiopure (3R,4S)-4 and (3R,4S)-5
were readily synthesized from 2-deoxyribose. In preparing these
piperidine derivatives we took advantage of the reductive
amination chemistry developed by Wong and co-workers in
connection with the chemoenzymatic synthesis of deoxy aza
sugars.10
to utilize a nominally substoichiometric amount of tosyl chloride.
Any unreacted 7 is expected to be removed during the course
of extractive workup.
Results and Discussion
For convenience, commercial 1-O-methyl-2-deoxy-D-ribose
(6) was used as starting material throughout these studies. While
commercial 6 consists principally of the furanoside form shown,
in our experience it contains varying amounts (up to 15%) of
the pyranoside isomer 7. (Furanoside 6 is the kinetic product
of acetalization of 1 while 7 is the thermodynamic product.
Which of these isomers predominates is a critical function of
reaction time and temperature and the nature and concentration
of the acid catalyst.11) Thus part of any apparent yield loss in
the present studies must be attributed to this contaminant. In
this regard it is noteworthy that both acetals are now available
via improved syntheses, which allow isolation of pure 612 or
7.13
Tosylate 8 so produced was of sufficient purity to be carried
on to the subsequent silylation and oxidation steps. Protection
as the tert-butyldiphenylsilyl ether afforded 9. However, we
were disappointed to find that oxidation under the usual
conditions14 (BF3 etherate as Lewis acid promoter) afforded only
recovered starting material. The reason for this failure is not
clear but seems to be associated with the presence of the silyl
protecting group. The problem was circumvented by use of
indium triflate as an alternative Lewis acid. By using 5% of
In(OTf)3, the desired lactone 10 was obtained as a snow-white
crystalline solid (mp 102 °C). As we had hoped, 10 was readily
purified by crystallization from heptane/ethyl acetate.
A simple one-pot procedure was then developed to convert
lactone 10 into the carboxylic acid 2. A solution of 10 and
sodium iodide was heated overnight to convert the tosylate into
the corresponding iodide. At this point zinc flakes were added
to the solution, which was further stirred for 90 min at room
temperature. After extractive workup, 2 was obtained in
essentially quantitative yield.
In the case of carboxylic acid 2 we were especially desirous
of developing a route that was scaleable and hence chromatog-
raphy-free. Given the heterogeneous starting material, this
seemed to necessitate a route proceeding through a solid
intermediate, which could be purified by crystallization. A
reasonable candidate seemed to be the lactone 10 (Scheme 1),
which we envisioned could be prepared via the Grieco-type
oxidation14 of the corresponding methyl furanoside 9.
Selective tosylation of 6 on the C5 oxygen is known in the
literature.15 In our experience, use of excess tosyl chloride
promotes tosylation at the 2° hydroxyl. For this reason (and
recognizing the presence of 7 in the starting material) it is better
Alternatively, if desired, the intermediate iodide 11 could be
separately isolated (see the Supporting Information).
(8) Kim, B. J.; Pyun, D. K.; Jung, H. J.; Hyun, J. K.; Kim, J. H.; Kim,
E. J.; Jeong, W. J.; Lee, C. H. Synth. Commun. 2001, 31, 1081.
(9) Tamaruya, Y.; Suzuki, M.; Kamura, G.; Kanai, M.; Hama, K.;
Shimizu, K.; Aoki, J.; Arai, H.; Shibasaki, M. Angew. Chem., Int. Ed. 2004,
43, 2843.
(10) Kajimoto, T.; Chen, L.; Liu, K. K.-C.; Wong, C.-H. J. Am. Chem.
Soc. 1991, 113, 6678.
(11) Deriaz, R. E.; Overend, W. G.; Stacey, M.; Wiggins, L. F. J. Chem.
Soc. 1949, 2836.
(12) For example, Grinstaff and co-workers have reported a protocol
that affords 6 in 96% isolated yield. Hashmi, S. A. N.; Hu, X.; Immoos, C.
E.; Lee, S. J.; Grinstaff, M. W. Org. Lett. 2002, 4, 4571.
(13) Crotti, P.; Di Bussolo, V.; Favero, L.; Macchia, F.; Pineschi, M.
Tetrahedron: Asymmetry 1996, 7, 779.
To further enhance the synthetic utility of 2, we also converted
it into the corresponding Weinreb amide, i.e., compound 12.
T3P16 was used as a coupling reagent and the product was
purified by flash chromatography to separate ca. 15% of silanol,
which was cleaved under these conditions, affording pure 12
in 77% yield.
(15) (a) Zamboni, R.; Milette, S.; Rokach, J. Tetrahedron Lett. 1984,
51, 5835. (b) Aly, Y. L.; Abdel-Megied, A. E.-S.; Pedersen, E. B.; Nielsen,
C. Liebigs Ann. Chem. 1992, 127.
(14) Grieco, P. A.; Oguri, T.; Yokoyama, Y. Tetrahedron Lett. 1978,
19, 419.
(16) T3P is n-propylphosphonic acid cyclic anhydride. For a brief review
of its chemistry, see: Llanes Garcia, A. L. Synlett. 2007, 1328.
7308 J. Org. Chem., Vol. 72, No. 19, 2007

Concise Routes To 5-Carbon Synthetic Building Blocks
SCHEME 2. Preparation of Aldehyde 3
SCHEME 3. Preparation of Dihydroxypiperidine 4
We note that an alternative route to compound 10 can be
imagined based on the well-known oxidation of 2-deoxyribose
to 2-deoxyribonolactone 13 using elemental bromine.17 For
small-scale synthesis, this approach would eliminate the need
to convert 1 to methyl furanoside 6 and also circumvents the
requirement to deal with mixtures of anomers. Indeed, Hol-
lingsworth has reported the synthesis of the C3-unprotected
analogue of 2 using such a strategy.18 Our reason for avoiding
this chemistry reflects our interest in a scaleable route: the
oxidation of 1 to 13 is carried out in dilute aqueous solution.
This requires the distillation of a large amount of water to isolate
13.
newly released aldehyde species is trapped in situ with a vinyl
or propargyl organometallic reagent.21In cases where isolation
of the aldehyde is required, Fuerstner has pointed out the
desirability of carrying these reactions out under anhydrous
conditions and has utilized highly active zinc/silver-graphite for
this purpose.19c As an alternative, we considered the use of
chromium(II) chloride22 as reductant. The mechanistic studies
of Kochi on the reduction of vicinal disubstituted compounds
(including â-halohydrins) with CrCl2 suggested that this reagent
might be suitable for Vasella-type reductions under mild, aprotic
conditions.23 Consistent with these proposals, chromium(II)
chloride in THF cleanly reduced 15 to the ring-opened aldehyde
3 in 83% yield after flash chromatography (Scheme 2).
Our approach to the piperidine cis-diol 4 also began with
the previously mentioned monotosylate 8,24 which was now
converted to the corresponding C5 azide 16 without protection
following a literature procedure25 as shown in Scheme 3.
Interestingly, conversion of the methyl furanoside moiety of
16 to the free lactol 17 proved the most difficult step of the
synthetic sequence. A variety of reagents and conditions have
been reported for the selective hydrolysis of methyl furano-
sides.26 From the standpoint of operational simplicity, we found
that use of strongly acidic Amberlyst 15 resin to catalyze the
hydrolysis was advantageous. Hydrolysis was complete after 4
h at room temperature. Polar impurities were removed by
passing the product over a bed of silica gel.
We next turned our attention to the aldehyde building block
3. The literature indicates that 3 can be prepared via DIBAL
reduction of the corresponding ester.7 However, we wished to
avoid unnecessary oxidation-reduction steps, taking advantage
of the fact that 6 already exists in the aldehyde oxidation state.
For this reason we prepared iodide 14, which we hoped to open
under Vasella-type conditions.19
Methyl furanoside 6 was first converted to the primary iodide
14 following the literature procedure20 as shown in Scheme 2.
It was then protected as the tert-butyldimethylsilyl ether 15.
Vasella-type reductive opening of 15 was attempted with a
variety of zinc sources including zinc dust, zinc flake, zinc
amalgam, zinc-copper couple, and diisopropylzinc containing
NiBr2 as catalyst. Complete consumption of iodide 15 was
generally observed. These procedures afforded the aldehyde 3
in yields ranging from 11% to 50% but none was judged
satisfactory.
The reductive amination of 17 was carried out in ethanol with
5% palladium on carbon as catalyst. The resulting freebase could
be isolated (see the Supporting Information) and was shown
(21) (a) Madsen, R.; Hyldtoft, L. J. Am. Chem. Soc. 2000, 122, 8444.
(b) Poulsen, C. S.; Madsen, R. J. Org. Chem. 2002, 67, 4441.
(22) Review: Hanson, J. R. Synthesis 1974, 1.
(23) Kochi, J. K.; Singleton, D. M. J. Am. Chem. Soc. 1968, 90, 1582.
(24) Tosylate 8 used in the preparation of diol 4 was purified by flash
chromatography. Unlike the preparation of 2, there is no crystalline
intermediate in the sequence leading to compound 4. Thus the yields in
Scheme 1 are based on crude 8 while those in Scheme 3 are based on
chromatographed 8 as starting material.
(25) (a) Tulshian, D. B.; Gundes, A. F.; Czarniecki Bioorg. Med. Chem.
Lett. 1992, 2, 525. (b) Schmidt, L.; Pedersen, E. B.; Nielson, C. Acta Chim.
Scand. 1994, 48, 215. (c) El-Barbary, A. A.; Khodair, A. I.; Pedersen, E.
B.; Nielsen, C. Nucleosides Nucleotides 1994, 13, 707.
(26) For example: (a) HCl/HOAc: Gadikota, R. R.; Keller, A. I.; Callam,
C. S.; Lowary, T. L. Tetrahedron: Asymmetry 2003, 14, 737. (b) 80%
aqueous HOAc: Vandendriessche, F.; Snoeck, R.; Janssen, G.; Hoog-
martens, J.; Van, Aerscshot, A.; De Clercq, E.; Herdewijn, P. J. Med. Chem.
1992, 35, 1458. (c) 0.5 M formic acid: Moustakis, C.; Viala, J.; Capdevila,
J.; Falck, J. R. J. Am. Chem. Soc. 1985, 107, 5283. (d) BCl3‚Me2S:
Congreve, M. S.; Holmes, A. B.; Hughes, A. B.; Looney, M. G. J. Am.
Chem. Soc. 1993, 115, 5815.
While disappointing, this was not entirely unexpected.
Aldehydes produced via the Vasella reaction often prove
unstable under the reaction conditions. Madsen has recently
demonstrated an elegant solution to this problem in which the
(17) Deriaz, R. E.; Overend, W. G.; Stacey, M.; Teece, E. G.; Wiggins,
L. F. J. Chem. Soc. 1949, 1879. For updated versions of this reaction, see:
Han, S.-Y.; Joullie, M. M.; Petasis, N. A.; Bigorra, J.; Corbera, J.; Font, J.;
Ortuno, R. M. Tetrahedron 1993, 49, 349. Wang, Z.-X.; Wiebe, L. I.;
DeClercq, E.; Balzarini, J.; Knaus, E. E. Can. J. Chem. 2000, 78, 1081.
(18) Song, J.; Hollingsworth, R. I. Tetrahedron: Asymmetry 2001, 12,
387.
(19) (a) Bernet, B.; Vasella, A. HelV. Chim. Acta 1979, 62, 1990. (b)
Nakane, M.; Hutchinson, C. R.; Gollman, H. Tetrahedron Lett. 1980, 21,
1213. (c) Fuerstner, A.; Jumbam, D.; Teslic, J.; Weidmann, H. J. Org. Chem.
1991, 56, 2213.
(20) Skaanderup, P. R.; Poulsen, C. S.; Hyldtoft, L.; Jorgensen, M. R.;
Madsen, R. Synthesis 2002, 1721.
J. Org. Chem, Vol. 72, No. 19, 2007 7309

Wang and Nugent
SCHEME 5. Effect of Charge on Regioselectivity
SCHEME 4. Preparation of (3R,4S)-Epoxide 5
by 1H and 13C NMR to be identical with material produced by
the chemoenzymatic procedure of Wong.10 However, for the
preparation of 4, it was advantageous to treat the filtered product
solution in situ with di-tert-butyl dicarbonate to afford 4 directly,
thus circumventing the isolation step. By using this expedient,
the yield of 4 from 17 was 76% despite some loss of hydrophilic
product during extractive workup.
For conversion of cis-diol 4 to the epoxide 5, two reagents
can be considered, namely Viehe’s salt27 (dichloromethylene-
dimethyliminium chloride) and the Moffatt reagent28 (1-chlo-
rocarbonyl-1-methylethyl acetate, 18). We initially examined
the use of Viehe’s salt but found that the vigorous conditions
for hydrolysis of the intermediate chloro carbamate were
problematic. In contrast, reaction of 4 with the Moffatt reagent
(Scheme 4) gives the chloro acetate 20, which was converted
under mild conditions to epoxide 5. Regioselective formation
of 20 as the 4-chloro regioisomer was demonstrated by 2D NMR
studies, while chiral HPLC confirmed that the epoxide product
5 was >99% ee.
ions with dialkylamines, which again is completely regio-
specific.30,31 These results suggest that the directing effect of
the ring heteroatom in 3,4-disubstituted heterocycles is more
pronounced when nucleophilic addition proceeds by way of a
cationic intermediate. This is relevant to the selective formation
of 20 since the reaction of the Moffatt reagent with cis-diols is
proposed28 to proceed through a cationic intermediate of type
19.
Conclusion
In the foregoing discussion we have shown that 2-deoxy-D-
ribose is readily converted into a series of chiral 5-carbon
intermediates of demonstrated synthetic utility. Most of this
chemistry should be amenable to larger scale manufacture of
these chiral building blocks. The principal limitation in this
approach is that 1 is commercially available only as the
D-stereoisomer. However, the recent development of straight-
forward technology to convert 1 to into 2-deoxy-L-ribose32 may
remove this restriction as well. We hope that these procedures
will further enhance the attractiveness of 2-deoxyribose as an
available, enantiopure, and highly functional starting material.
The highly regioselective formation of chloro acetate 20 is
noteworthy.29 This result is consistent with literature observa-
tions8,30,31 on the nucleophilic opening of related 3,4-disubsti-
tuted heterocycles under conditions where the intermediate bears
a significant positive charge (Scheme 5).
As shown in Scheme 5, opening of 5 with HBr affords a
single â-bromohydrin while opening the same epoxide with
dialkylamines gives a 1:1 mixture of regioisomers.8 The latter
reaction also contrasts with the opening of dialkylaziridinium
Experimental Section
2-Deoxy-1-O-methyl-3-O-tert-butyldiphenylsilyl-5-O-(p-tolu-
enesulfonyl)-D-ribofuranose, 9. Tosylate 8 was first prepared
following the literature procedure.15 Toluenesulfonyl chloride (11.6
g, 60.7 mmol, 0.9 equiv) was added to a solution of 6 (10.0 g,
67.5 mmol) in pyridine (200 mL). The mixture was stirred overnight
after which the solvent was removed at reduced pressure. The
residue was taken up in ethyl acetate (300 mL) and was washed
sequentially with 1 N HCl, saturated NaHCO3 solution, water, and
brine (100 mL each). The solution was dried (Na2SO4) and the
solvent was removed to afford 8 (17.2 g, 94% based on TsCl) for
(27) (a) Sunay, U.; Mootoo, D.; Molino, B.; Fraser-Reid, B. Tetrahedron
Lett. 1986, 27, 4697. (b) Fraser-Reid, B.; Molino, B. F.; Magdzinski, L.;
Mootoo, D. R. J. Org. Chem. 1987, 52, 4505.
(28) Greenberg, S.; Moffatt, J. G. J. Am. Chem. Soc. 1973, 95, 4016.
(29) Our results using Viehe’s salt are not directly comparable since the
protecting group on nitrogen was benzyl rather than Boc. Nevertheless, it
is noteworthy that this reaction gave a nearly 1:1 mixture of regioisomeric
chloro carbamates. This is consistent with the proposal (ref 26) that the
positive charge in the case of Viehe’s salt is delocalized and resides mainly
on the iminium nitrogen.
which NMR data were consistent with those in the literature.15 1
H
(30) See also: (a) D’Andrea, S. V.; Michalson, E. T.; Freeman, J. P.;
Chidester, C. G.; Szmuszkovicz, J. J. Org. Chem. 1991, 56, 3133. (b) Zhao,
S.; Ghosh, A.; D’Andrea, S. V.; Freeman, J. P.; VonVoigtlander, P. F.;
Carter, D. B.; Smith, M. W.; Blinn, J. R.; Szmuszkovicz, J. Heterocycles
1994, 39, 163. (c) Thomas-Xavier, M.; Appenzeller, J.; Pardo, D. G.; Cossy,
J. Org. Lett. 2006, 8, 3509.
(31) Similarly, Crotti has studied addition of nucleophiles to 3,4-
epoxytetrahydropyran (the oxygen analogue of 5) and finds that selectivity
for attack at the 4-position is enhanced by the addition of Lewis acids.
These results were rationalized in terms of “chelation effects” but seem to
us to reflect the enhanced directing effect of the ring oxygen as binding to
the Lewis acid makes the epoxide more electron deficient. Chini, M.; Crotti,
P.; Gardelli, C.; Macchia, F. Tetrahedron 1994, 50, 1261.
NMR (400 MHz, CDCl3): δ 7.75 (2H, m), 7.32 (2H, m), 4.99 (1H,
m), 4.37-4.01 (3H, m), 3.31 (1.5H, s), 3.18 (1.5H, m), 2.90 (0.5H,
d, J ) 11 Hz), 2.46 (0.5H, d, J ) 5.3 Hz), 2.41 (3H, s), 2.22-1.92
(2H, m). 13C NMR (100 MHz, CDCl3): δ 145.0, 132.6, 129.9,
127.9, 105.6, 105.3, 84.3, 82.9, 72.6, 72.4, 70.2, 69.4, 55.0, 41.2,
40.9, 21.6.
A solution of 8 (5.30 g, 17.6 mmol) in anhydrous DMF (200
mL) was cooled to 10 °C in an ice bath. To the cold solution was
(32) Ji, Q.; Pang, M.; Han, J.; Feng, S.; Zhang, X.; Ma, Y.; Meng, J.
Synlett 2006, 2498.
7310 J. Org. Chem., Vol. 72, No. 19, 2007

Concise Routes To 5-Carbon Synthetic Building Blocks
added imidazole (1.44 g, 21.1 mmol, 1.2 equiv) after which tert-
butyldiphenylsilyl chloride (5.4 mL, 22 mmol, 1.2 equiv) was added
at once. The mixture was stirred overnight at room temperature.
Heptane (100 mL) was added to the mixture, which was washed
with water (3 × 100 mL) and brine (100 mL). The organic layer
was dried over Na2SO4, and the solvent was removed under reduced
pressure affording 9 (9.31 g, 85%) as a pale yellow oil in adequate
purity to be carried on in the following step. (Note that under these
conditions, excess tBuPh2SiCl is converted to the silanol, which is
retained as an impurity in 9 but is shed during the crystallization
of 10.) A sample of 9 was purified by flash chromatography as a
2 equiv). A solution of triethylamine (3.82 g, 37.8 mmol, 4 equiv)
in anhydrous acetonitrile (35 mL) was added. The coupling agent
T3P16 was added as the commercially available 50% solution in
ethyl acetate (12.0 g, 18.9 mmol) and the weighing vessel was rinsed
in with additional acetonitrile (12 mL). The mixture was stirred
for 23 h at room temperature and rinsed into a separatory funnel
with water (75 mL) and ethyl acetate (75 mL). The organic layer
was separated and washed with 10% citric acid (100 mL) and finally
with water (50 mL). After removal of the solvent at reduced
pressure, the residue was purified by flash chromatography with
80:20 heptane/diethyl ether as eluant to afford 12 (3.02 g, 77%) as
a colorless liquid. 1H NMR (400 MHz, CDCl3): δ 7.75-7.63 (6H,
m), 7.45-7.30 (4H, m), 5.89 (1H, ddd, J ) 17, 10, 6 Hz), 5.03
(1H, d, J ) 17 Hz), 4.93 (1H, d, J ) 10 Hz), 4.74 (1H, m), 3.54
1
mixture of R and â anomers for characterization. H NMR (400
MHz, CDCl3): δ 7.8-7.2 (14H, m), 4.98 (0.5H, m), 4.81 (0.5,
m), 4.26-3.61 (4H, m), 3.31 (1.5H, s), 3.12 (1.5H, s), 2.42 (3H,
s), 2.04 (1H, m), 1.85 (1H,m), 1.03 (9H, s). 13C NMR (100 MHz,
CDCl3): δ 144.6, 135.6, 133.6, 130.0, 127.7, 105.6, 104.8, 83.6,
81.8, 73.5, 72.6, 70.4, 69.5, 55.0, 54.8, 41.8, 41.4, 31.99, 26.7, 22.6,
21.5, 19.0, 14.1. Anal. Calcd for C29H36O6SSi: C, 64.41; H, 6.71.
Found: C, 64.51; H, 6.65.
(3H, s), 3.08 (3H, s), 2.79 (1H, m), 2.52 (1H, m), 1.07 (9H, s). 13
C
NMR (100 MHz, CDCl3): δ 171.2, 139.9, 135.9, 134.0, 129.5,
127.3, 114.8, 71.5, 61.1, 40.6, 31.8, 26.9. 19.2. Anal. Calcd for
C23H31NO3Si: C, 69.48; H, 7.86; N, 3.52. Found: C, 69.56; H,
7.91; N, 3.33.
2,5-Dideoxy-1-O-methyl-3-O-tert-butyldimethylsilyl-5-iodo-D-
ribofuranose, 15. The C3-unprotected iodide 14 was first prepared
following the procedure of Madsen.20 A solution of 6 (5.00 g, 33.7
mmol), triphenylphosphine (13.3 g, 50.6 mmol, 1.5 equiv), imi-
dazole (4.60 g, 67.5 mmol, 2 equiv), and iodine (13.0 g, 50.6 mmol,
1.5 equiv) in anhydrous THF (100 mL) was stirred overnight at
room temperature. A white precipitate of imidazole hydroiodide
was formed, which was filtered off by using a bed of Celite 545.
The filtrate was dried on 10 g of silica gel and purified by flash
chromatography, using 70:30 hexanes/ethyl acetate as eluant.
Evaporation of the solvents afforded 14 (7.1 g, 81%) as a colorless
oil. NMR data matched those in the literature for the mixed R and
â anomers.15b 1H NMR (400 MHz, CDCl3): δ 5.07 (1H, s), 4.39
(0.5H, m), 4.03 (1.5H, m), 3.33 (3H, m), 3.19 (1H, m), 3.14 (1H,
m), 2.96 (1H, br), 2.22 (1H, m), 2.05 (0.5H, m), 1.95 (0.5H, m).
13C NMR (100 MHz, CDCl3): δ 105.6, 105.3, 85.9, 85.7, 75.5,
75.4, 55.2, 54.9, 41.7, 40.8, 7.8, 6.7.
2-Deoxy-3-O-tert-butyldiphenylsilyl-5-O-(p-toluenesulfonyl)-
D-ribonolactone, 10. Commercial 3-chloroperbenzoic acid (MCP-
BA) contains a significant amount of water that could potentially
interfere with the reaction.33 Consequently commercial MCPBA
(6.05 g, 27 mmol, nominal 2 equiv) was dissolved in dichlo-
romethane (300 mL) and the aqueous layer was removed by using
a separatory funnel. To the solution were added 9 (7.30 g, 13.5
mmol) and indium(III) trifluoromethanesulfonate (0.38 g, 0.68
mmol, 5%) as catalyst. After being stirred overnight at room
temperature, the reaction mixture was washed with 1 N NaOH (3
× 100 mL), water (3 × 100 mL), and brine (100 mL). The resulting
solution was dried (Na2SO4) and the solvent was removed at
reduced pressure. The semisolid residue was crystallized from 80:
20 heptane/ethyl acetate to afford 10 (5.05 g, 71%) as a white
crystalline solid, mp 102 °C. 1H NMR (400 MHz, CDCl3): δ 7.62-
7.29 (14H, m), 4.37 (1H, m), 4.31 (1H, m), 3.83 (1H, dd, J ) 3,
11 Hz), 3.81 (1H, dd, J ) 3, 11 Hz), 2.59 (1H, dd, J ) 7, 18 Hz),
2.47 (1H, dd, J ) 3, 18), 2.41 (3H, s), 1.02 (9H, s). 13C NMR (100
MHz, CDCl3): δ 175.2, 144.6, 135.5, 132.6, 130.4, 130.0, 128.1,
127.9, 84.3, 70.0, 67.7, 38.1, 26.7, 21.6, 18.9. Anal. Calcd for
C28H32O6SSi: C, 64.09; H, 6.15. Found: C, 64.10; H, 6.27.
(3S)-3-tert-Butyldiphenylsilyloxypent-4-enoic Acid, 2. A solu-
tion of 10 (5.00 g, 9.53 mmol) and sodium iodide (3.57 g, 23.8
mmol, 2.5 equiv) in anhydrous DMSO (45 mL) were heated at 60
°C for 18 h. The mixture was cooled to room temperature
whereupon zinc flake (1.71 g, 26.2 mmol, 2.7 equiv) was added
and stirring was continued for 1.5 h at room temperature. The
mixture was filtered into a flask containing water (100 mL) and 1
N HCl (25 mL) with stirring. The contents of the flask were rinsed
into a separatory funnel with dichloromethane (75 mL). The organic
layer was separated and washed with water (2 × 50 mL). Removal
of solvent at reduced pressure and drying in high vacuum afforded
A solution of 14 (6.35 g, 24.6 mmol) in anhydrous DMF (100
mL) was cooled to 10 °C in an ice bath. To the cold solution was
added imidazole (2.0 g, 29.5 mmol, 1.2 equiv). A 1.0 M solution
of tert-butyldimethylsilyl chloride in dichloromethane (24.6 mL,
1.0 equiv) was slowly added to the cold solution with use of a
syringe. The reaction mixture was stirred overnight at room
temperature. Heptane (100 mL) was added and the mixture was
washed with water (3 × 100 mL) and brine (100 mL). The organic
layer was dried over Na2SO4, and the solvent was removed under
1
reduced pressure affording 15 (8.0 g, 87%) as an oil. H NMR
(400 MHz, CDCl3): 5.01 (0.5H, dd, J ) 2, 5 Hz), 4.91 (0.5H, dd,
J ) 3, 5 Hz), 4.27 (0.5H, m), 3.86 (1H, m), 3.49 (0.5H, m), 3.39
(0.5H, dd, J ) 4, 10 Hz), 3.31 (1.5H, s), 3.28 (1.5H, s), 3.22-3.14
(1.5H, overlapping m), 2.38 (0.5 H, m), 2.11 (0.5H, m), 2.00 (0.5H,
m), 1.76 (0.5H, m), 0.78 (9H, s), 0.02 (1.5H, s), 0.00 (3H, s), -0.01
(1.5H, s). 13C NMR (100 MHz, CDCl3): δ 105.2, 104.4, 85.8, 80.5,
75.5, 75.1, 55.2, 42.5, 41.9, 25.7, 17.9, 7.8, 7.7, -4.55, -4.65.
Anal. Calcd for C12H25IO3Si: C, 38.71; H, 6.77. Found: C, 38.46;
H, 6.59.
(3S)-3-tert-Butyldimethylsilyloxy-pent-4-enal, 3. Generation of
anhydrous CrCl2 solution was found to be more facile with use of
commercial chromium(III) chloride tris(tetrahydrofuran) complex34
rather than unsolvated CrCl3, which is slow to dissolve and
sometimes requires heating. A flask was charged with CrCl3‚3THF
(4.45 g, 11.9 mmol) and zinc flakes (3.34 g, 51.1 mmol) and was
thoroughly flushed with nitrogen. Anhydrous THF (60 mL) was
added and the mixture was stirred 30 min during which time the
solution changed color from violet to pale blue. A solution of 15
(3.47 g, 9.32 mmol) in THF (10 mL) was gradually added over
the course of 10 min with use of a syringe. After 24 h the mixture
2 (3.35 g, 99%) as a yellow oil. [R]25D -11.2 (c 1.00, CHCl3) {lit.4
1
[R]28 -10.0 (c 0.6, CHCl3)}. H NMR (400 MHz, CDCl3): δ
D
7.71-7.62 (4H, m), 7.48-7.30 (6H, m), 5.83 (1H, ddd, J ) 17,
10, 6 Hz), 4.98 (1H, d, J ) 17 Hz), 4.94 (1H, d, J ) 10 Hz), 4.57
(1H, apparent q, J ) 6 Hz), 2.56 (1H, d, J ) 7, 3 Hz), 2.42 (1H,
d, J ) 7, 3 Hz), 1.05 (9H, s). 13C NMR (100 MHz, CDCl3): δ
176.4, 139.2, 136.0. 133.6, 130.0, 127.8, 116.0, 71.4, 43.6, 27.3,
19.9. Anal. Calcd for C21H26O3Si: C, 71.14; H, 7.39. Found: C,
69.80; H, 7.44.
(3S)-N-Methyl-N-methoxy-3-tert-butyldiphenylsilyloxypent-
4-enamide, 12. A flask was charged with 2 (3.35 g, 9.45 g) and
(N,O)-dimethylhydroxylamine hydrochloride (1.84 g, 18.9 mmol,
(33) Attempting to dry a CH2Cl2 solution of commercial MCPBA with
4A molecular sieves results in its conversion to dibenzoyl peroxide. If a
greater level of dryness is desired beyond that described in the text, we
recommend drying commercial MCPBA in high vacuum at room temper-
ature for a period of 1 to 2 h.
(34) Alternatively, CrCl3‚3THF can be prepared by a straightforward
literature procedure: Kern, R. J. J. Inorg. Nucl. Chem. 1962, 24, 1105.
J. Org. Chem, Vol. 72, No. 19, 2007 7311

Wang and Nugent
was filtered through a 1 cm bed of Celite into a flask containing a
stirred mixture of water (135 mL) and 1 N HCl (15 mL). The
product was extracted into hexanes (50 mL) and was washed with
water (50 mL) and dried (MgSO4). Removal of the solvent at
reduced pressure gave the product, which was purified by flash
chromatography affording 3 (1.66 g, 83%) as a colorless liquid.
[R]25 -3.2 (c 2.0, CHCl3) {lit.7 for (3R)-enantiomer [R]25 +3.0
dropwise with stirring. After 48 h, any excess (Boc)2O was
destroyed by addition of triethylamine (1.01 g, 10 mmol) and taurine
(1.25 g, 10 mmol) and the mixture was again stirred overnight.
The mixture was filtered to remove some undissolved taurine and
the solvent was distilled at reduced pressure. The residue was taken
up in dichloromethane (25 mL) and was washed with water (10
mL). Distillation of the solvent afforded 4 (2.53 g, 76%). 1H NMR
(400 MHz, CDCl3): δ 3.65-2.97 (8H, overlapping br m), 1.52
(1H, br m), 1.34 (1H, br m), 1.15 (9H, s). 13C NMR (100 MHz,
CDCl3): δ 155.8, 80.3, 68.8, 68.3, 45.9 (br), 40.6 (br), 29.9, 28.9.
Anal. Calcd for C10H19NO4: C, 55.28; H, 8.81; N, 6.45. Found:
C, 55.13; H, 8.70; N, 6.38.
D
D
(c 2.58, CHCl3)}. 1H NMR (400 MHz, CD2Cl2): δ 9.69 (1H, br t,
J ) 2.4 Hz), 5.82 (1H, ddd, J ) 5.7, 10.5, 17.0 Hz), 5.21 (1H, dt,
J ) 1.3, 17.0 Hz), 5.05 (1H, dt, J ) 1.3, 10.5 Hz), 4.60 (1H, br q,
J ) 5 Hz), 2.57 (1H, ddd, J ) 2.6, 6.5, 15.6 Hz), 2.49 (1H, ddd,
J ) 2.0, 4.9, 15.6 Hz), 0.86 (9H, s), 0.02 (3H, s), 0.01 (3H, s). 13
C
NMR (100 MHz, CD2Cl2): δ 202.3, 140.8, 115.5, 170.1, 51.8, 26.6,
18.7, -4.55, -4.65. Anal. Calcd for C11H22O2Si: C, 61.63; H,
10.34. Found: C, 61.67; H, 10.37.
The optical purity of 3 was further confirmed by separate
conversion into the diastereomeric propargylic sulfinamides 21 and
22, using the Ellman chiral auxiliary35 and the improved protocol
for acetylide addition reported by Hou and co-workers.36 (See the
Supporting Information for details.)
(3R,4R)-1-tert-Butoxycarbonyl-3-acetoxy-4-chloropiperi-
dine, 20. A solution of 1-chlorocarbonyl-1-methylethyl acetate (1.05
g, 6.38 mmol, 1.2 equiv) in dichloromethane (5 mL) was added
dropwise with stirring to a solution of 4 (1.16 g, 5.34 mmol) in
dichloromethane (10 mL). The mixture was stirred an additional 1
h at room temperature whereupon saturated NaHCO3 solution (15
mL) was added and stirring was continued an additional 15 min.
The organic layer was separated, washed with water (10 mL), and
dried (MgSO4). After removal of the solvent, the residue was
purified by flash chromatography to afford 20 (1.23 g, 83%) as a
1
colorless oil. H NMR (400 MHz, CDCl3): δ 4.80 (1H, m), 4.09
(1H, m), 3.76 (d, J ) 13.8 Hz), 3.6-3.4 (3H, overlapping m), 2.23
(1H, m), 2.09 (3H, s), 1.82 (1H, m), 1.47 (9H, s). 13C NMR (100
MHz, CDCl3): δ 169.9, 155.0, 80.4, 71.2 (br), 57.0, 44.1 (br), 40.5
(br), 31.5, 28.7, 21.7. To avoid complications from carbamate
rotamers, 2D NMR studies were carried out in DMSO-d6 at 100
°C. (See the Supporting Information.) Anal. Calcd for C12H20-
ClNO4: C, 51.89; H, 7.26. Found: C, 51.78; H, 7.29.
(3R,4S)-1-tert-Butoxycarbonyl-3,4-epoxypiperidine, 5. A solu-
tion of potassium carbonate (1.93 g, 14.0 mmol, 3 equiv) in water
(2.5 mL) was added to a solution of 20 (1.30 g, 4.68 mmol) in
methanol (25 mL). The heterogeneous mixture was stirred for 2 h
at room temperature and was then filtered. The filtrate was diluted
with water (75 mL) and the product was extracted into chloroform
(2 × 25 mL). The organic phase was washed with water (10 mL)
and solvent was distilled at reduced pressure to afford 5 (0.82 g,
4.1 mmol, 88%) as a pale yellow oil. As expected, several NMR
resonances are broad (and the C2 and C6 resonances in the 13C
spectrum are bifurcate) due to carbamate rotamers. 1H NMR (400
MHz, CDCl3): δ 4.00-3.08 (6H, overlapping m), 2.02 (1H, br s),
1.94 (1H, br s), 1.46 (9H, s). 13C NMR (100 MHz, CDCl3): δ
155.2, 80.1, 51.0, 50.6 (br), 42.9 and 42.3 (br), 38.0 and 36.9 (br),
28.9, 24.8. Anal. Calcd for C10H17NO3: C, 60.28; H, 8.60; N, 7.03.
Found: C, 60.41; H, 8.82; N, 6.95.
2,5-Dideoxy-1-O-methyl-5-azido-D-ribofuranose, 16. Following
the literature procedure,25 sodium azide (5.86 g, 90.2 mmol, 4 equiv)
and tetrabutylammonium iodide (0.50 g, 1.4 mmol, 0.06 equiv) were
added to a solution of 824 (6.81 g, 22.5 mmol) in anhydrous DMF
(50 mL). The mixture was stirred for 36 h at 80 °C and after cooling
was added to ethyl acetate (50 mL) and heptane (50 mL). The
resulting organic phase was extracted with water (100 mL then 50
mL) and was dried (Na2SO4). Removal of the solvent at reduced
pressure afforded 16 (3.27 g, 78%) as a 1:1 mixture of R and â
anomers, which was purified by flash chromatography in 70:30
hexanes/ethyl acetate. 1H NMR (400 MHz, CDCl3): δ 5.07 (0.5H,
dd, J ) 4, 10 Hz), 5.02 (0.5H, dd, J ) 4, 13 Hz), 4.11 (0.5H, m),
4.05 (0.5H, m), 3.91 (0.5H, m), 3.41-3.18 (2H, m), 3.32 (1.5H,
s), 3.29 (1.5H, s), 2.97 (1H, br), 2.14 (1H, m), 1.95 (1H, m). 13C
NMR (100 MHz, CDCl3): δ 105.3, 85.3, 84.5, 73.1, 72.6, 55.2,
54.8, 53.6, 52.4, 41.4, 41.1.
(3R,4S)-1-tert-Butoxycarbonyl-3,4-dihydroxypiperidine, 4. Lac-
tol 17 was first released following the general literature precedent.25a
A mixture of 16 (3.50 g, 20.2 mmol), water37 (100 mL), and
Amberlyst 15 resin (10 g) was stirred 4 h at room temperature.
After filtration the mixture was stirred briefly with Amberlyst 21
(5.0 g) to scavenge any traces of acid and was re-filtered. Silica
was added to the resulting solution and water was removed at
reduced pressure keeping the temperature below 50 °C. The product
was then eluted with 92:8 dichloromethane/methanol to afford 17
(2.10 g, 65%) as an amber oil, which was immediately carried on
to the next step.
The enantiomeric excess of the product was shown to be >99%
by chiral capillary column gas chromatography (Rt-bDEXsm
column, J&W Scientific, 30 m × 0.25 mm i.d., 0.25 µm film, flow
rate 40 cm/s, programmed from 110 to 220 °C at 20 deg min-1).
Retention times for the enantiomers were established to be 24.3
min for (3S,4R)-5 and 25.4 min for (3R,4S)-5 by comparison with
a sample of authentic racemate prepared by the literature procedure.30b
Sample chromatograms are given in the Supporting Information.
A Fischer-Porter tube was charged with a solution of crude 17
(2.43 g, 15.3 mmol) in anhydrous ethanol (75 mL). After adding
5% palladium on carbon catalyst (1.00 g), the system was flushed
with hydrogen and stirred under 60 psi H2 for 24 h at room
temperature. The solution was filtered and a solution of di-tert-
butyl dicarbonate (3.06 g, 14 mmol) in ethanol (10 mL) was added
Acknowledgment. The authors thank Ms. Shirley Wong for
chiral gas chromatographic analyses and Ms. Sarah Traeger for
2D NMR studies.
Supporting Information Available: Synthetic details for
compounds 11, 21, and 22 and freebase of compound 4, 2D NMR
studies on 20, chiral gas chromatographic analysis of 5, DTC
(35) Tang, T. P.; Volkman, S. K.; Ellman, J. A. J. Org. Chem. 2001,
66, 8772.
1
analysis of 10 and 22, and H and 13C NMR spectra for all new
(36) Ding, C. H.; Chen, D.-D.; Luo, Z.-B.; Dai, L.-X.; Hou, X.-L. Synlett
2006, 1272.
(37) We find that the use of organic co-solvents diminishes the yield of
compounds. This material is available free of charge via the Internet
at http://pubs.acs.org.
17.
JO0712143
7312 J. Org. Chem., Vol. 72, No. 19, 2007