Stereoselective synthesis of vancosamine and saccharosamine glycals via tungsten-catalyzed alkynol cycloisomerization

Article abstract of DOI:10.1021/ol017195f

(equation presented) A stereoselective synthesis of the C-3 branched amino glycals o vancosamine and saccharosamine is described that features a tungsten carbonyl catalyzed cycloisomerization of the corresponding alkynyl alcohol.

Full text of DOI:10.1021/ol017195f

ORGANIC
LETTERS
2002
Vol. 4, No. 5
749-752
Stereoselective Synthesis of
Vancosamine and Saccharosamine
Glycals via Tungsten-Catalyzed Alkynol
Cycloisomerization
William W. Cutchins and Frank E. McDonald*
Department of Chemistry, Emory UniVersity, 1515 Pierce DriVe,
Atlanta, Georgia 30322
fmcdona@emory.edu
Received December 8, 2001
ABSTRACT
A stereoselective synthesis of the C-3 branched amino glycals of vancosamine and saccharosamine is described that features a tungsten
carbonyl catalyzed cycloisomerization of the corresponding alkynyl alcohol.
Partially deoxygenated carbon branched amino sugars are
important components of several classes of medicinally
useful compounds with demonstrated antibiotic and anti-
cancer activity. In particular, the 3-amino-3-methyl-2,3,6-
trideoxy lyxo-hexose L-vancosamine 1 (Figure 1) is an
component of saccharomicin,2 an oligosaccharide antibiotic
that is active against bacteria resistant to vancomycin.
The synthesis of these carbon-branched sugars and their
derivatives present challenges primarily in the construction
of the amine-bearing C-3 quaternary center. Several syntheses
of vancosamine have appeared in the literature.3 Although
no specific syntheses of saccharosamine 2 have been
reported, its nitro analogue, D-decilonitrose, has been known
for some time, and several syntheses that pass through a
saccharosamine intermediate have appeared.4 We wished to
(2) Kong, F.; Zhao, N.; Siegal, M. M.; Janota, K.; Ashcroft, J. S.; Koehn,
F. E.; Borders, D. B.; Carter, G. T. J. Am. Chem. Soc. 1998, 120, 13301.
(3) (a) Dyong, I.; Friege, H. Chem. Ber. 1979, 112, 3273. (b) Ahmad,
H. I.; Brimacombe, J. S.; Mengech, A. S.; Tucker, L. C. N. Carbohydr.
Res. 1981, 93, 288. (c) Giovanni, F.; Fuganti, C.; Grasselli, P.; Pedrocchi-
Fantoni, G. Tetrahedron Lett. 1981, 22, 5073. (d) Brimacombe, J. S.;
Mengech, A. S.; Rahman, K. M. M.; Tucker, L. C. N. Carbohydr. Res.
1982, 110, 207. (e) Hamada, Y.; Kawai, A.; Shioiri, T. Tetrahedron Lett.
1984, 25, 5413. (f) Hauser, F. M.; Ellenberger, S. R.; Glusker, J. P.; Smart,
C. J.; Carrell, H. L. J. Org. Chem. 1986, 51, 50. (g) Klemer, A.; Wilbers,
H. Liebigs Ann. Chem. 1987, 10, 815. (h) Greven, R.; Juetten, P.; Scharf,
H. D. Carbohydr. Res. 1995, 275, 83. (i) Nicolaou, K. C.; Mitchell, J. J.;
Jain, N. F.; Bando, T.; Hughes, R.; Winssinger, N.; Natarajan, S.; Koumbis,
A. E. Chem. Eur. J. 1999, 5, 2648. (k) Nicolaou, K. C.; Baran, P. S.; Zhong,
Y.-L.; Vega, J. A. Angew. Chem., Int. Ed. 2000, 39, 2525.
Figure 1. Representative 3-amino-2,3,6-trideoxyhexoses.
essential constituent of vancomycin, an antibiotic generally
considered to be the last line of defense for many severe
bacterial infections.1 A diastereomer of vancosamine, D-
saccharosamine 2, has more recently been isolated as a
(1) Ritter, T. K.; Wong, C. H. Angew. Chem., Int. Ed. 2001, 40, 3508.
10.1021/ol017195f CCC: $22.00 © 2002 American Chemical Society
Published on Web 02/13/2002

Scheme 1. Retrosynthetic Analysis
Scheme 3. Stereoselective Reductions of Ketones 13a,b
explore our tungsten-catalyzed alkynol cycloisomerization
methodology5 for the synthesis of vancosamine 1 and
stereoisomeric sugars including saccharosamine 2 in order
to evaluate the compatibility of our methodology with
nitrogen-containing substrates. With glycals 3 and 4 in hand,
we would then seek to construct naturally occurring oligo-
saccharides containing those amino sugars.
Our retrosynthetic analysis is presented in Scheme 1.
Deconstructing each of the amino sugar glycals using our
endo-cyclization methodology gave alkynyl alcohols 5 and
6. We hypothesized that either alcohol diastereomer could
be prepared by action of a judiciously selected reducing agent
on ketone 7, which would arise from nucleophilic opening
of â-lactam 8. We planned to employ a Staudinger cyclo-
addition6 between the appropriate ketene and imine to rapidly
build the â-lactam framework.
Our synthesis began with the condensation of p-anisidine
and 4-trimethylsilyl-3-butyn-2-one to give imine 10 in 80%
yield as a single diastereomer (Scheme 2). The stereochem-
istry is tentatively assigned as Z on the basis of the absence
of any NOE enhancement between the methyl and aromatic
hydrogens. Combination of 10 with the ketene derived from
benzyloxyacetyl chloride 11 gave â-lactam 12 as a single
diastereomer.7 The relative stereochemistry was again ten-
a All reactions run at -78 °C. b Ratios determined by 400 MHz
1H NMR, conversion >80%. c 31% recovered 13b.
tatively assigned as shown due to the absence of any NOE
enhancement between the methyl and methine protons
directly attached to the â-lactam. Analysis of subsequent
compounds eventually proved our assignment to be correct
(vide infra). The stereochemistry of the lactam is also
consistent with nucleophilic addition of the Z imine nitrogen
atom to the ketene sp-hybridized carbon followed by
conrotatory ring closure.8
Reaction with methyllithium gave ketone 13a as a single
product in 94% yield. Whereas literature precedent suggested
reaction of carbamate-protected lactams with methyl nucleo-
philes gave tertiary alcohols,9 we observed that the PMP-
protected â-lactam 13a was highly resistant to overalkylation
by methyllithium, so that even an excess (2 equiv) of
methyllithium could be used without generating any tertiary
alcohol byproduct.
Scheme 2. Diastereoselective Synthesis of â-Lactam 12 and
Conversion to 13
With ketone 13a in hand, we next explored a variety of
reducing agents to selectively prepare both alkynol diaster-
eomers 14 and 15, leading to vancosamine and saccharo-
samine, respectively (Scheme 3). We found that Felkin-
Anh selectivity could be observed for ketone reductions with
both PMP- and Cbz-protected aminoketones 13a,b favoring
formation of the corresponding diastereomer 14a,b, but the
(4) (a) Noecker, L.; Duarte, F.; Bolton, S. A.; McMahon, W. G.; Diaz,
M. T.; Guiliano, R. M. J. Org. Chem. 1999, 64, 6275. (b) Greven, R.;
Juetten, P.; Scharf, H. D. J. Org. Chem. 1993, 58, 3742. (c) Brimacombe,
J. S.; Rahman, K. M. M. Carbohydr. Res. 1985, 140, 163.
(5) (a) McDonald, F. E.; Reddy, K. S. Angew. Chem., Int. Ed. 2001, 40,
3653. (b) McDonald, F. E.; Reddy, K. S. J. Organomet. Chem. 2001, 617,
444. (c) McDonald, F. E.; Reddy, K. S.; D´ıaz, Y. J. Am. Chem. Soc. 2000,
122, 4304.
(6) (a) Staudinger, H. Liebigs Ann. Chem. 1907, 356, 51. (b) Georg, G.
I.; Ravinkumar, V. T. In The Organic Chemistry of â-Lactams; Georg, G.
I., Ed.; VCH: New York, 1993; pp 256-368. (c) Palomo, C.; Aizpurua, J.
M.; Ganboa, I.; Oiarbide, M. Eur. J. Org. Chem. 1999, 3223.
750
Org. Lett., Vol. 4, No. 5, 2002

Scheme 4. Synthesis of Vancosamine Glycal
Scheme 5. Synthesis of Saccharosamine Glycal
a (a) TBAF, THF. (b) 10% W(CO)6, THF, DABCO, hν, 35 °C.
step protocol that was employed for the preparation of ketone
13b. To our delight, substrate 16b readily underwent
cycloisomerization with only 5% W(CO)6 in less than 3 h,
to give protected vancosamine glycal 17 in 97% yield.12
Hydrolysis of the enol ether gave the known vancosamine
derivative 18 whose spectral data (1H and 13C NMR and IR)
were identical to that reported in the literature.3i This
correlation also corroborates our stereochemical assignment
for formation of â-lactam 12.
Having established the feasibility of the Cbz-protected
amine for the cycloisomerization methodology, we next
sought to apply this transformation to the synthesis of the
saccharosamine glycal (Scheme 5). However, we found that
19 furnished the desired glycal product 20 in only 74% yield
using the same conditions employed for the vancosamine
glycal. By replacing triethylamine with diaza[2.2.2]bi-
cyclooctane (DABCO), we could increase the yield of glycal
20 to 98%.5a,b
a (a) K2CO3, CH3OH. (b) CAN, CH3CN, H2O; CbzCl, K2CO3.
(c) 5% W(CO)6, THF, Et3N, hν, 55 °C. (d) CSA, H2O, THF.
complementary reduction with chelation control from the
adjacent benzyloxy group was achieved only with the Cbz-
protected ketone 13b. Chelate-controlled reduction of 13a
gave reduced selectivity, presumably due to competing
chelation with the basic anisidine substituent. We observed
that the standard Luche reduction conditions gave the best
Felkin-Anh selectivity, providing 14a from the PMP-
protected amino ketone 13a, whereas Zn(BH4)2 reduction
of the Cbz-protected amino ketone 13b gave the best
selectivity for chelation-controlled reduction to provide the
alcohol diastereomer 15b. Both reactions were highly
solvent-dependent. As expected, the Luche reduction did not
proceed with any appreciable rate even at room temperature
in the absence of methanol but also exhibited poor selectivity
if too much methanol was included (see Supporting Informa-
tion for exact reaction details). Likewise, the zinc borohy-
dride reduction proceeded extremely well in the nonchelating
solvent CH2Cl2, but with much slower rate and lower
diastereoselectivity in ethereal solvents.10
The tertiary amine base probably serves two roles: not
only does it act as a proton shuttle during the course of the
cycloisomerization, but it also stabilizes the catalytically
active “W(CO)5” species. DABCO is a better ligand than
triethylamine and may stabilize the tungsten species more
effectively than triethylamine, preventing it from degrading
to catalytically inactive species. Substrates in which cyclo-
isomerization occurs rapidly do not exhibit a pronounced
“amine effect” and proceed well regardless of the tertiary
amine base used, but sluggish cycloisomerization reactions
with lower turnover frequency benefit from DABCO ligation,
as the catalytically active “W(CO)5” enjoys a longer lifetime.
We note that simply increasing catalyst loading or adding
additional tungsten hexacarbonyl during the reaction gener-
ally did not improve the product yield, as larger quantities
of the spent catalyst proved rather difficult to remove from
the glycal products.
As our explorations of the tungsten-catalyzed cycloisomer-
ization of PMP-protected amino-alkynol substrate 16a11 were
unsatisfactory, we exchanged the PMP in 16a for the much
less basic Cbz-carbamate (Scheme 4) using the same two-
(7) For other examples of stereoselective Staudinger reactions, see:
Palomo, C.; Aizpurua, J. M.; Garc´ıa, J. M.; Galarza, R.; Legido, M.;
Urchegui, R.; Roma´n, P.; Luque, A.; Server-Carrio´, J.; Linden, A. J. Org.
Chem. 1997, 62, 2070.
(8) Although ketene-olefin cycloadditions generally occur by Wood-
ward-Hoffmann allowed [Π2s-Π2a] concerted processes (Snider, B. B.
Chem. ReV. 1988, 88, 793), ketene-imine cycloadditions have been
demonstrated to proceed by a stepwise mechanism. See: (a) Lynch, J. E.;
Riseman, S. M.; Laswell, W. L.; Tschaen, D. M.; Volante, R. P.; Smith, G.
B.; Shinkai, I. J. Org. Chem. 1989, 54, 3792. (b) Hegedus, L. S.;
Montgomery, J.; Narukawa, Y.; Snustad, D. C. J. Am. Chem. Soc. 1991,
113, 5784. (c) Palomo, C.; Aizpurua, J. M.; Ganboa, I.; Oiarbide, M. Eur.
J. Org. Chem. 1999, 3223.
In conclusion, a rapid entry to both vancosamine and
saccharosamine glycals has been achieved via tungsten-
catalyzed cycloisomerizations of acyclic alkynyl alcohols.
Studies directed toward the asymmetric synthesis of these
(9) Palomo, C.; Aizpurua, J. M.; Garc´ıa, J. M.; Iturburu, M.; Odriozola,
J. M. J. Org. Chem. 1994, 59, 5184.
(10) For a dramatic example of this phenomena with borohydride
reductions, see: Faucher, A. M.; Brochu, C.; Landry, S. R.; Duchesne, S.
H.; Hantos, S.; Roy, A.; Myles, A.; Legault, C. Tetrahedron Lett. 1998,
39, 8425.
(12) Despite extensive attempts to optimize cycloisomerization reaction
conditions for substrate 16a (W(CO)6, Et3N or DABCO, THF, hν ) 350
nm), we could not raise the yield of the reaction above ca. 30%, nor could
we separate the glycal from the many other unidentified byproducts. We
suspect that the p-methoxyphenylamine is poorly compatible with the
reaction, either from complexation of the amine with the tungsten catalyst
or photolytic degradation of the electron-rich aromatic system.
(11) Alkynol 16a was generated by base-promoted removal of the
acetylenic silyl group from compound 14a.
Org. Lett., Vol. 4, No. 5, 2002
751

glycals as well as their application to oligosaccharide
synthesis are in progress.
the National Institute of Health, National Science Foundation,
and the Georgia Research Alliance.
Supporting Information Available: Experimental details
and procedures for compounds 10 and 12-20. This material
is available free of charge via the Internet at http://pubs.acs.org.
Acknowledgment. We thank the National Institute of
Health (CA 59703) for support of this research. We also
acknowledge the use of shared instrumentation (NMR
spectroscopy, mass spectrometry) provided by grants from
OL017195F
752
Org. Lett., Vol. 4, No. 5, 2002