Synthesis of carbasugar C-1 phosphates via Pd-catalyzed cyclopropanol ring opening

Article abstract of DOI:10.1021/ol801106r

(Chemical Equation Presented) The stereoselective syntheses of 2,3-dideoxy-4-oxo-5a-carba-α-D-rhamnopyanose 1-phosphate, 2,3-dideoxy-5a-carba-α-D-rhamnopyranose 1-phosphate, 5a-carba-α-D- rhamnopyranose 1-phosphate, 5a-carba-α-D-digitoxopyranose 1-phosphate, and 5a-carba-α-L-rhamnopyranose 1-phosphate have been achieved from D-quinic acid. The routes rely upon a Simmons-Smith cyclopropanation and diastereospecific ring opening of cyclopropanol under Pd/C hydrogenation condition to set up the α-methyl ketone. A sequence of diastereoselective reduction, dihydroxylation, and/or Myers' reductive 1,3-rearrangement were used to install the desired stereochemistry.

Full text of DOI:10.1021/ol801106r

ORGANIC
LETTERS
2008
Vol. 10, No. 16
3381-3384
Synthesis of Carbasugar C-1 Phosphates via
Pd-Catalyzed Cyclopropanol Ring Opening
Mingde Shan and George A. O’Doherty*
Department of Chemistry, West Virginia UniVersity, Morgantown, West Virginia 26506
george.odoherty@mail.wVu.edu
Received May 13, 2008
ABSTRACT
The stereoselective syntheses of 2,3-dideoxy-4-oxo-5a-carba-r-D-rhamnopyanose 1-phosphate, 2,3-dideoxy-5a-carba-r-D-rhamnopyranose
1-phosphate, 5a-carba-r-D-rhamnopyranose 1-phosphate, 5a-carba-ꢀ-D-digitoxopyranose 1-phosphate, and 5a-carba-r-L-rhamnopyranose
1-phosphate have been achieved from D-quinic acid. The routes rely upon a Simmons-Smith cyclopropanation and diastereospecific ring
opening of cyclopropanol under Pd/C hydrogenation condition to set up the r-methyl ketone. A sequence of diastereoselective reduction,
dihydroxylation, and/or Myers’ reductive 1,3-rearrangement were used to install the desired stereochemistry.
Carbohydrates are a recurring structural motif in many
bioactive natural products (e.g., antimicrobial and anticancer
agents).1 Often, the sugar portion of the molecule is a crucial
element for the activities of these compounds; for instance,
many aglycons of biologically active natural products are devoid
of activity.2 A cursory structure-activity relationship analysis
of the carbohydrate-containing natural products points to the
use of rare and deoxysugars (Figure 1).3 The removal or
glycosylated natural products for medicinal chemistry studies.4
An alternative approach is to alter the biosynthesis of the
glycosylated natural products.5
In an effort to understand and ultimately alter the
microbial biosynthesis of rare/deoxy-sugar containing
natural products, various C-glycoside analogues of the
NDP-sugars have been prepared.6 Most of these efforts
have been focused on the synthesis and use of C-glycoside
analogues with exoacetal oxygen replaced with a meth-
ylene group (Y ) CH2).6 In contrast, there is only one
report that describes the use of the endoacetal oxygen
C-glycoside (X ) CH2, cyclitol).7 To these ends, we have
initiated effort to develop an approach to these cyclitol
sugar phosphates (X ) CH2) in a manner that would mimic
our de novo approach to carbohydrates (i.e., one that is
amenable to both enantiomers of either R- or ꢀ-isomers).
Herein we report our successful efforts toward these aims
from quinic acid.
Figure 1. Biosynthesis of glycosyltransferase inhibitors.
stereochemical inversion of the various sugar hydroxyl groups
can impart significant in vitro stability by lessening the abilities
of naturally occurring glycosidase enzymes to degrade the
structure. In fact, our ongoing interest in the de novo synthesis
of carbohydrates is aimed at being able to provide new
(2) Thibodeaux, C.; Melancu¨on, C. E., III; Liu, H.-w. Nature 2007, 446,
1008–1016.
(3) Kren, V.; Martinkova, L. Curr. Med. Chem. 2001, 8, 1303–1328.
(4) Zhou, M.; O’Doherty, G. A. Curr. Top. Med. Chem. 2008, 8, 114–
125.
(5) Hong, L.; Zhao, Z., III; Zhang, H.; Liu, H.-w. J. Am. Chem. Soc.
2008, 130, 4954–4967.
(6) (a) Moffatt, J. G.; Khorana, H. G. J. Am. Chem. Soc. 1961, 83, 649–
658. (b) Melancu¨on, C. E., III; Hong, L.; White, J. A.; Liu, Y. N.; Liu,
H.-w. Biochemistry 2007, 46, 577–590.
(1) Johnson, D. A.; Liu, H.-w. ComprehensiVe Natural Products
Chemistry; Elsevier: Amsterdam, NY, 1999; Vol. 3, pp 311-365.
10.1021/ol801106r CCC: $40.75
Published on Web 07/18/2008
2008 American Chemical Society

Scheme 2. Cyclopropanation/Cyclopropanol Ring Opening
Figure 2. Target cyclitol sugar phosphates.
We targeted the synthesis of both enantiomers of three
R-sugars and one ꢀ-sugar. These structures are outlined in
Figure 2: 2,3-dideoxy-4-oxo-5a-carba-R-D-rhamnopyanose phos-
phate (3), 2,3-dideoxy-5a-carba-R-D-rhamnopyranose phosphate
(4), 5a-carba-R-D-rhamnopyranose phosphate (5), 5a-carba-ꢀ-
D-digitoxopyranose phosphate (6), and 5a-carba-R-L-rhamnopy-
ranose phosphate (ent-5). Retrosynthetically, we envisioned that
the desired carbasugar phosphate stereoisomers 7 could be
synthesized from both enantiomers of the cis- and trans-
diastereomers of the enone γ-phosphate 8. We thought 8 could
be derived through a ꢀ-elimination and phosphorylation from
acetonide 9 (Scheme 1). The critical methyl group could be
introduced by R-alkylation either through in situ generated
enolate or silylenol ether from enone 10 with the acetonide-
protecting group controlling the stereochemistry. The known
intermediate 10 could be easily synthesized from cheap starting
material D-(-)-quinic acid by known methods.8
Alternatively, we envisioned a cyclopropanation and ring-
opening process which could lead to the formation of R-methyl
ketone.12 Thus, one-pot of hydrosilylation of enone 10 (RhCl(P-
Ph3)3/PhMe2SiH, 60 °C) followed by Simmons-Smith cyclo-
propanation (ZnEt2/CH2I2) of the resulting silylenol ether gave
an inseparable mixture of cyclopropanol silyl ethers,13 which
when desilylated under acidic conditions (p-TsOH/MeOH) fur-
nished two readily separable diastereoisomers 13a and 13b in
83% overall yield (Scheme 2). The diastereomeric ratio of
cyclopropanes 13a to 13b rangedfrom 1:1 when the cyclopropa-
nation was performed at low temperature (-20 °C) to 2:1 at 0 °C.
We next investigated the subsequent ring opening of the
cyclopropanol 13a. Unfortunately, all of our attempts to promote
ring opening under acidic or basic conditions failed to give the
desired ketone 9.14 In contrast, the stochiometric Pd(OAc)2-
promoted ring opening of the cyclopropanol gave an R,ꢀ-
unsaturated ketone,15 which under hydrogenation conditions
resulted in an inseparable diastereomeric mixture of R-methyl
ketone 9 (dr 3:1 9a/9b). We wondered if the palladium
homoenolate intermediate could be trapped with hydrogen
before ꢀ-hydride elimination occurred. Because we found that
Pd/C also gave the same enone product, we decided to explore
typical hydrogenolysis conditions.16
Scheme 1. Retrosynthetic Analysis
Our syntheses commenced with the known four-step conver-
sion of D-(-)-quinic acid 11 to the R,ꢀ-unsaturated ketone 10,
which was routinely accomplished on 20 g scale in 60% overall
yield (Scheme 2). Application of the Tsuda-Saegusa 1,4-
reduction/alkylation via the aluminum enolate with methyl
iodide failed to cleanly give the desired R-methyl ketone 9.9
Similarly, the 1,4-hydrosilyation using Stryker’s reagent and
dimethylphenylsilane to form silylenol ether 12 was problem-
atic.10 Fortunately, the hydrosilylation occurred smoothly at 60
°C when Wilkinson’s catalyst and dimethylphenyl silane were
used.11 Unfortunately, various attempts to alkylate the silylenol
ether with methyl iodide via its enolate failed to cleanly provide
the R-methyl ketone 9.
(10) (a) Lipshutz, B. H.; Chrisman, W.; Noson, K.; Papa, P.; Sclafani,
J. A.; Vivian, R. W.; Keith, J. M. Tetrahedron 2000, 56, 2779–2788. (b)
Mahoney, W. S.; Stryker, J. M. J. Am. Chem. Soc. 1989, 111, 8818–8823.
(11) (a) Ojima, I.; Kogure, T. Organometallics 1982, 1, 1390–1399. (b)
Tojo, S.; Isobe, M. Synthesis 2005, 8, 1237–1244.
(7) Ko, K.-S.; Zea, C. J.; Pohl, N. L. J. Am. Chem. Soc. 2004, 126,
13188–13189.
(12) Kulinkovich, O. G. Chem. ReV. 2003, 103, 2597–2632.
(8) (a) Federspiel, M.; Fischer, R.; Hennig, M.; Mair, H.-J.; Oberhauser,
T.; Rimmler, G.; Albiez, T.; Bruhin, J.; Estermann, H.; Gandert, C.; Gockel,
V.; Gotzo, S.; Hoffmann, U.; Huber, G.; Janatsch, G.; Lauper, S.; Rockel-
Stabler, O.; Trussardi, R.; Zwahlen, A. G. Org. Process Res. DeV. 1999, 3,
266–274. (b) Trost, B. M.; Romero, A. G. J. Org. Chem. 1986, 51, 2332–
2342. (c) Audia, J. E.; Boisvert, L.; Patten, A. D.; Villalobos, A.;
Danishefsky, S. J. J. Org. Chem. 1989, 54, 3738–3740.
(9) Tsuda, T.; Satomi, H.; Hayashi, T.; Saegusa, T. J. Org. Chem. 1987,
52, 439–443.
(13) For a review on Simmons-Smith cyclopropanation, see: (a)
Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. ReV. 1993, 93, 1307. (b)
Lautens, M.; Klute, W.; Tam, W. Chem. ReV. 1996, 96, 49. (c) Lebel, H.;
Marcoux, J.-F.; Molinaro, C.; Charette, A. B. Chem. ReV. 2003, 103, 977.
(14) For acid-catalyzed opening, see: (a) Schreiber, S. L.; Smith, D. B.;
Schulte, G. J. Org. Chem. 1989, 54, 5994–5996. (b) Paquette, L. A.; Ross,
R. J.; Shi, Y. J. J. Org. Chem. 1990, 55, 1589–1598, and ref 12. For base-
catalyzed opening, see: (c) Barnier, J. P.; Blanco, L.; Rousseau, G.; Guibe-
Jampel, E.; Fresse, I. J. Org. Chem. 1993, 58, 1570–1574.
3382
Org. Lett., Vol. 10, No. 16, 2008

To our delight, when cyclopropanol 13a was exposed to Pd/C
under a balloon of H2, ketone 9a was produced in good yield
(80%) and as a single diastereomer (dr >100:1, 1H NMR). Next,
the acetonide was removed under acidic conditions (10% HCl/
THF, 0 °C to rt, 0.5-1 h) to afford diol 14 in good yield (80%).
Because great care needed to be taken to avoid epimerization
of either 9a or 14 under acidic conditions, we preferred to first
deprotect the acetonide 13a to give triol 15 in 82% yield. As
before, exposure of the triol 15 under hydrogenation conditions
(Pd/C, H2) afforded the diol 14 in 82% yield as a single
diastereomer.
Scheme 4. R-D-rhamno-Phosphate 5 and Deoxy Analogue 4
With the successful synthesis of diol 14 we next turned our
attention to make phosphate targets. Carbonate formation of
14 using triphosgene followed by ꢀ-elimination under basic
conditions resulted in allylic alcohol 16 in 73% yield (Scheme
3). Phosphorylation of the alcohol 16 with N,N-diethyl diben-
and diastereoselectivity (dr 11:1). Diimide reduction (NBSH/
Et3N) of 20 resulted in saturated alcohol 21 in 60% yield with
20% starting material back. Hydrogenation of 21 on Pd/C
furnished 2,3-dideoxy-5a-carba-R-D-rhamnopyranose phosphate
4 in 83% yield. Alternatively, dihydroxylation of allylic alcohol
20 under Upjohn conditions18 resulted in a single diastereomeric
triol 22 in 86% yield, which similarly underwent hydrogenolysis
to afford 5a-carba-R-D-rhamnopyranose phosphate 5 in 87%
yield.
Scheme 3. 4-Keto-Sugar Phosphate 3
In order to achieve the ꢀ-D-digitoxose phosphate 6, we simply
switched the acetonide protecting group of the vicinal cis-diol
of quinic acid to the bis-ketal protecting group of its trans-
vicinal diol (Scheme 5). By reported methods,19 R,ꢀ-unsaturated
ketone 23 could be produced on a 20 g scale in five steps with
high overall yield (50%). Similarly, a one-pot hydrosilylation
and cyclopropanation, followed by desilylation under acidic
conditions, gave a diastereomeric mixture of cyclopropanols
24a/b (dr ∼1:1) in 85% overall yield. Although in this bis-
ketal case the two cyclopropanol isomers were inseparable, a
direct isomerization on Pd/C of this mixture under typical
hydrogenation conditions (1 atm of H2, 6 h, rt) afforded a
diastereomeric mixture of R-methyl ketones 25a/b (63% yield,
dr 2:1) along with some recovered cyclopropanol 24b (20%,
as a single diastereomer). Under equilibrating conditions, the
thermodynamically more stable isomer 25b could be prepared.
Thus, treatment of a mixture 25a and 25b with acid (concd
HCl(aq.)/THF) resulted in a diastereomerically enriched mixture
of 25b to 25a (dr >13:1), which after silica gel chromatography
provided pure 25b. TFA deprotection of 25b gave crude diol
(albeit with some C-5 epimerization, dr ∼15:1), which under-
went acylation with Boc anhydride and subsequent ꢀ-elimina-
tion to afford enone 26 (79% yield from 25b, dr ∼15:1). While
the major isomer could be purified by careful chromatography,
the diastereomeric mixture of 26 could be selectively reduced
with LiAlH4 at -78 °C to produced the allylic alcohol 27 in
82% yield with the minor diastereomers easily removed by
chromatography.
zylphosphoramidite in the presence of 1H-tetrazole gave a
phosphite intermediate which was oxidized by addition of 30%
H2O2 to form dibenzyl phosphate 17 in 76% yield. Hydroge-
native debenzylation and reduction of the double bond (H2, Pd/
C) failed to give the desired dihydrogen phosphate 3, presum-
ably due to hydrogenolysis of the allylic phosphate. Attempts
at hydrogenation of 17 with diimide produced in situ from
o-nitrobenzenesulfonyl hydrazide (NBSH) and Et3N proved
unsuccessful.17 Thus, we decided to reduce the double bond
before introduction of the phosphate functional group. Hydro-
genation of 16 (H2, Pd/C) gave saturated alcohol 18 in 72%
yield. Phosphorylation of 18 as previously gave dibenzyl
phosphate 19 in 70% yield. Debenzylation (H2, Pd/C) afforded
the 2,3-dideoxy-4-oxo-5a-carba-R-D-rhamnopyranose phosphate
3 in 73% yield. While analysis of the crude product showed
diastereomerically clean material, upon sitting, the product slowly
epimerized resulting in a ∼2:1 mixture of methyl diastereomers.
To synthesize the phosphates 4 and 5, we returned to enone
17 (Scheme 4). A diastereoselective reduction with LiAlH4 at
-78 °C gave the equatorial alcohol 20 in good yield (87%)
(15) Park, S.-B.; Cha, J. K. Org. Lett. 2000, 2, 147–149.
(16) For examples of the trapping of organometallic intermediates with
hydrogen, see: (a) Ngai, M.-Y.; Kong, J.-R.; Krische, M. J. J. Org. Chem.
2007, 72, 1063. (b) Iida, H.; Krische, M. J. Top. Curr. Chem. 2007, 279,
77.
(18) VanRheenen, V.; Kelly, R. C.; Cha, D. Y. Tetrahedron Lett. 1976,
17, 1973–1976.
(19) (a) Montchamp, J.-L.; Tian, F.; Hart, M. E.; Frost, J. W. J. Org.
Chem. 1996, 61, 3897–3899. (b) Barros, M. T.; Maycock, C. D.; Ventura,
M. R. J. Chem. Soc., Perkin Trans. 1 2001, 166–173.
(17) Haukaas, M. H.; O’Doherty, G. A. Org. Lett. 2002, 4, 1771–1774.
Org. Lett., Vol. 10, No. 16, 2008
3383

Scheme 5
.
ꢀ-D-Digitoxo-Phosphate 6
Scheme 6. R-L-rhamno-Phosphate ent-5
35 at -78 °C resulted in allylic alcohol with good diastereo-
selectivity (dr 11:1), which after TBS-protection and ester
hydrolysis gave alcohol 36 in good overall yield (77% for three
steps). Phosphorylation of alcohol 36 gave dibenzyl phosphate
37 in 66% yield. Dihydroxylation of 37 under Upjohn condi-
tions produced diol in 81% yield as a single diastereomer, which
after acid deprotection (p-TsOH/MeOH) provided triol 38 in
excellent overall yield (79%, two steps). Finally hydrogenation
of dibenzylphosphate 38 afforded 5a-carba-R-L-rhamnopyranose
phosphate ent-5 in 81% yield.
In conclusion, five carbasugar phosphates have been suc-
cessfully synthesized in either D or L, R or ꢀ form from D-(-)-
quinic acid. This synthesis includes a novel highly diastereospe-
cific Pd/C-catalyzed cyclopropanol ring opening, which may
be a useful alternative for the introduction of methyl groups to
the R-position of ketones. Diastereoselective reduction, dihy-
droxylation, Mitsunobu reaction, and Myers’ reductive rear-
rangement were also used to install the desired sugar stereo-
chemistry. The biological evaluation of these cyclitol phosphates
is ongoing.
With the allylic alcohol in hand, we turned our attention to
the preparation of ꢀ-D-digitoxose phosphate 6. Treatment of
allylic alcohol 27 under the Myers’ reductive transposition
conditions provided olefin 28 in 89% yield.17,20 The olefin was
subjected to Upjohn dihydroxylation (OsO4/NMO) and
acetonide protection (dr 11:1), at which point the minor
diastereomer was removed. Finally, the Boc-carbonate reduc-
tion by LiAlH4 afforded alcohol 29 in 75% overall yield
(three steps). Alcohol 29 was then phosphorylated to give
dibenzyl phosphate 30 in 74% yield. The acetonide depro-
tection of 30 produced diol 31 with excellent yield (95%),
which under typical hydrogenation conditions (H2, Pd/C)
resulted in 5a-carba-ꢀ-D-digitoxopyranose phosphate 6 in
78% yield.
In order to access the L-series of carbasugar phosphate
(Scheme 6), we returned to the cyclopropanol 13b, which was
prepared as part of a 1:1 mixture with 13a (Scheme 2).
Deprotection of the acetonide 13b with 10% aqueous HCl in
THF furnished triol 32 with good yield (82%), which was
hydrogenolytically ring opened to produce R-methyl ketone 33
in 80% yield as a single diastereomer. Dehydration of 33 via
carbonate formation followed by elimination afforded allylic
alcohol 34 in 73% yield over two steps. Mitsunobu reaction21
of 34 with p-nitrobenzoic acid in the presence of DIAD and
PPh3 gave the ester 35 with stereochemical inversion in excellent
yield (95%). Diastereoselective LiAlH4 reduction of the enone
Acknowledgment. We thank Dr. Novruz Akhmedov (WVU)
for the help in carrying out 31P NMR experiments. We are
grateful to the NSF (CHE-0749451) and the ACS-PRF (47094-
AC1) for the support of our research program and NSF-
EPSCoR (0314742) for a 600 MHz NMR at WVU.
Supporting Information Available: Complete experimen-
tal procedures and spectral data for all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL801106R
(20) (a) Myers, A. G.; Zheng, B.; Movassaghi, M. J. Org. Chem. 1997,
62, 7507. (b) Zhou, M.; O’Doherty, G. A. Org. Lett. 2006, 8, 4339–4342.
(21) Mitsunobu, O. Synthesis 1981, 1–28.
3384
Org. Lett., Vol. 10, No. 16, 2008