A surprising dipolar cycloaddition provides ready access to aminoglycosides

Article abstract of DOI:10.1021/ja0319238

This contribution describes the results of a new research effort in our laboratory aimed at the synthesis of novel aminoglycosides and amino-C-glycosides. Despite the importance of such compounds, and the previous development of some methodological soluti

Full text of DOI:10.1021/ja0319238

Published on Web 06/19/2004
A Surprising Dipolar Cycloaddition Provides Ready Access to
Aminoglycosides
Russell S. Dahl and Nathaniel S. Finney*
Department of Chemistry and Biochemistry, UniVersity of California, San Diego, 9500 Gilman DriVe,
La Jolla, California 92093-0358
Received December 23, 2003; E-mail: nfinney@chem.ucsd.edu
Scheme 1. Synthesis of Triazolines from Tri-O-acetyl-D-glucal
As appreciation for the biological importance of carbohydrates
has increased, so have efforts to develop methods for the synthesis
of biologically relevant oligosaccharides, glycoconjugates, and
analogues thereof.1 Among the more challenging synthetic targets
are aminoglycosides and their derivatives. Despite apparent similar-
ity, hexoses and 2-deoxy-2-aminohexoses pose very different
synthetic challenges, the aminosugars proving more difficult to
prepare or transform.1 Aminoglycosides are typically synthesized
by one of two approaches: modification of glucose or glucosamine,2
or introduction of a nitrogen substituent into a glycal derivative.3-5
The former method requires extensive protecting-group manipula-
tion, while the latter has fewer embodiments. We describe here an
approach to installing a nitrogen substituent at C2 of glycals that
is both novel and operationally simple.
a RN3 CH(OEt)3, reflux, 24-36 h. b acid, base, or Lewis acid. DMB )
3,5-dimethoxybenzyl.
Scheme 2. Transformation of Triazolines to Aminoglycosides
Among the existing methods for the direct introduction of a
nitrogen substituent at C2 of a glycal, the most extensively studied
are: (1) iodosulfonamidation followed by aziridine generation (eq
1).3a-c,4 (2) aziridine generation via a stoichiometric (nitrido)Mn(V)
oxidant (eq 1).3d,4 and (3) oxazoline generation via C2-sulfonium
intermediates.5,6 With the exception of dialkyl azodicarboxylate
hetero-Diels-Alder reactions (eq 2),7 methods that rely on cyclo-
addition for introduction of the nitrogen functionality are largely
absent despite the prevalence of nitrogen-containing dipoles. In the
course of synthesizing aminoglycoside inhibitors of chitin synthase,
we began exploring dipolar cycloadditions that could lead to
subsequent aziridine generation.
a hν, (CH3)2CO, 10-12 h. b NuH (Chart 1), Sc(OTf)3.
acid led, at best, to poor yields of the corresponding aminoglycoside
derivatives accompanied by extensive elimination to triazole. In
the case of 2c, small amounts of the aziridine (4c) could be
generated thermally, although this intermediate could not be
intercepted. Attempts at photochemical generation of aziridine 4b
from triazoline 2b led only to cycloreversion. Fortunately, triazoline
2a underwent clean, quantitative photochemical conversion to the
corresponding aziridine, 4a, provided the photolysis was carried
out in acetone (Scheme 2).13,14 The intermediate aziridine was not
isolated or purified, although photolysis in acetone-d6 allowed the
acquisition of a clean 1H NMR spectrum consistent with the
proposed aziridine intermediate.15
More importantly, solutions of the aziridine generated in acetone
could be transferred directly to solutions of nucleophile in THF
containing an appropriate Lewis acid. Again, careful selection of
reagent (Lewis acid) was critical. In the presence of 0.2 equiv of
Sc(OTf)3 and 2-3 equiv of nucleophile, aziridine 4a underwent
clean reaction with a range of nucleophiles to afford the corre-
sponding aminoglycoside derivatives (5-15, Scheme 2, Chart 1)
possessing exclusively the â-configuration.15,16
While all reactions were carried out in pure dry solvents, no
efforts were made to exclude oxygen or water; despite this, good
to excellent yields of almost all aminoglycosides were realized.
Among the notable successes are the reactions with a secondary
alcohols (7, 10), a nitrogen nucleophile (11), and more complex
alcohols (12-14). The preparation of 14 bodes well for developing
iterative versions of this reaction.
Two notable exceptions to the generally good yields are the
reactions of 4a with BnSH and glycine methyl ester. The former
gives only moderate yields of the adduct, while reaction with the
latter provides none of the desired product. In each case, the re-
maining material is a mixture of the products of aziridine hydrolysis
and condensation with the enol tautomer of acetone.17 This
observation led us to expose the crude aziridine to Sc(OTf)3 in the
After extensive experimentation, we have found that glycals with
electron-withdrawing protecting groups undergo efficient cyclo-
addition with electron-rich azides and that the cycloadducts can be
transformed to semi-stable aziridines which react with nucleophiles
under mild conditions to afford diverse aminoglycoside products.8-10
Tri-O-acetyl-D-glucal (1) readily engages in dipolar cycloaddi-
tions with alkyl azides, such as benzyl azide, 3,5-dimethoxybenzyl
azide, and 1-adamantyl azide, at elevated temperature.11 Judicious
choice of solvent is critical to the outcome of the reaction. If the
cycloaddition is carried out at high temperature in trimethyl- or
triethylorthoformate, the unstable triazoline intermediates (2a-c)
can be isolated in good yield with little or no purification. In all
other solvents the triazoline undergoes elimination to afford the
corresponding triazole (Scheme 1).12
Similarly, all efforts to convert isolated triazolines 2a-c to a
reactive aziridine in the presence of a nucleophile and acid or Lewis
9
8356
J. AM. CHEM. SOC. 2004, 126, 8356-8357
10.1021/ja0319238 CCC: $27.50 © 2004 American Chemical Society

C O M M U N I C A T I O N S
Chart 1. Adducts of 4a and Various Nucleophilesa-c
of the corresponding aziridine. The approach is operationally simple,
with no purification required after the initial cyclization and mild
conditions for the formation of the aminoglycoside product. This
methodology complements existing methods, and there is clear
potential for development of iterative reaction protocols, expansion
of substrate scope and extension to solid-phase synthesis.
Acknowledgment. We thank the N.S.F. for infrastructure
support (CHE-9709183) and the UCSD Academic Senate for
financial support.
Supporting Information Available: Complete experimental and
spectroscopic details for all new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
References
b
a 2-3 equiv of NuH, 0.2 equiv of Sc(OTf)3, THF, 2-3 h. Yields for
(1) (a) Essentials of Glycobiology; Varki, A., Cummings, R., Esko, J., Freeze,
H., Hart, G., Marth, J., Eds.; Cold Spring Harbor Laboratory Press: Cold
Spring Harbor, NY, 1999. For recent reviews: (b) Dwek, R. A.; Butters,
T. D. Chem. ReV. 2002, 102, 283. (c) Bertozzi, C. R.; Kiessling, L. L.
Science 2001, 291, 2357. (d) Dove, A. Nat. Biotechnol. 2001, 19, 913.
(e) Kobata, A. Glycoconjugate J. 2001, 17, 443. (f) Koeller, K. M.; Wong,
C.-H. Nat. Biotechnol. 2000, 18, 835.
aziridine formation and opening. c2 equiv of TMSN3, 1 h, Sc(OTf)3 omitted.
Scheme 3. Expansion of Substrate Scopea
(2) For reviews of aminoglycoside synthesis and associated protecting group
manipulations, see: (a) Banoub, J.; Boullanger, P.; LaFont, D. Chem. ReV.
1992, 92, 1167. (b) Debenham, J.; Rodebaugh, R.; Fraser-Reid, B. Liebigs
Ann. 1997, 791.
(3) (a) Danishefsky, S. J.; Bilodeau, M. T. Angew. Chem., Int. Ed. Engl. 1996,
35, 1380. (b) Griffith, D. A.; Danishefsky, S. J. J. Am. Chem. Soc. 1991,
113, 5863. (c) Griffith, D. A.; Danishefsky, S. J. J. Am. Chem. Soc. 1990,
112, 5811. (d) Du Bois, J.; Tomooka, C. S.; Hong, J.; Carreira, E. M. J.
Am. Chem. Soc. 1997, 119, 3179. For earlier work on glycal aziridination,
see: (e) Driguez, H.; Vermes, J.-P.; Lessard, J. Can. J. Chem. 1978, 56,
119-X. (f) Koslowska-Gamsz, E.; Descotes, G. Can. J. Chem. 1982, 60,
558-X. (g) Lafont, D.; Descotes, G. Carbohydr. Res. 1988, 175, 35.
(4) Nucleophilic ring opening of such bicyclic aziridines has clear precedent
in the ring opening of the analogous epoxides. See: Halcomb, R. L.;
Danishefsky, S. J. J. Am. Chem. Soc. 1989, 111, 6661.
(5) (a) Di Bussolo, V.; Liu, J.; Huffman, L. G., Jr.; Gin, D. Y. Angew. Chem.,
Int. Ed. 2000, 39, 204. (b) Liu, J.; Gin, D. Y. J. Am. Chem. Soc. 2002,
124, 9789. (c) Liu, J. Di Bussolo, V.; Gin, D. Y. Tetrahedron Lett. 2003,
44, 4015.
(6) For halo- and azidonitration of glycals, see: (a) Lemieux, R. U.;
Nagabushan, T. L. Can. J. Chem. 1968, 46, 401. (b) Lemieux, R. U.;
Ratcliffe, R. M. Can. J. Chem. 1979, 57, 1244. For a recent modification,
see: (c) Seeberger, P. H.; Roehrig, S.; Schell, P.; Wang, Y.; Christ, W.
J. Carbohydr. Res. 2000, 328, 61.
(7) (a) Leblanc, Y.; Fitzsimmons, B. J.; Springer, J. P.; Rokacht, J. J. Am.
Chem. Soc. 1989, 111, 2995. (b) Fitzsimmons, B. J.; Leblanc, Y.; Rokacht,
J. J. Am. Chem. Soc. 1987, 109, 285. (c) Fitzsimmons, B. J.; Leblanc, Y.;
Chan, N.; Rokacht, J. J. Am. Chem. Soc. 1988, 110, 5229.
(8) For reviews of azide cycloadditions, see: (a) Sha, C.-K.; Mohanakrishnan,
A. K. Chem. Heterocycl. Comput. 2002, 59, 623. (b) Koumbis, A. E.;
Gallos, J. K. Curr. Org. Chem. 2003, 7, 771.
(9) For a review of triazolines as aziridine equivalents, see: Semenov, V. P.
Russ. J. Org. Chem. 1996, 32, 1576 (Zh. Org. Khim. 1996, 32, 1627).
(10) For recent work on the intramolecular cyclization of carbamoyl- and
azidoglycals, see: (a) Nicolaou, K. C.; Baran, P. S.; Zhong, Y.-L.;
Barluenga, S.; Hunt, K. W.; Kranich, R.; Vega, J. A. J. Am. Chem. Soc.
2002, 124, 2233 and references therein. (b) Churchill, D. G.; Rojas, C.
M. Tetrahedron Lett. 2002, 43, 7225 and references therein.
absence of added nucleophile. Under these conditions, the acetone-
derived C-glycoside (15) could be isolated in good yield (74%).
With regard to the glycal and azide components, matching the
electronic character of the components is essential. Glycal 1 failed
to undergo cycloaddition with electron-deficient azides such as TsN3
or BocN3. Condensation of BnN3 with tri-O-benzyl glucal was
similarly ineffective under our conditions, as was reaction with
dihydropyran. While alkynes react readily with numerous azides,
alkene/azide cycloadditions most commonly involve either an
electron-rich azide/electron-poor alkene pair or an electron-rich
alkene/electron-poor azide pair.18 The reaction of 1 and BnN3
appears to fall into the former category, suggesting that 1 acts as
an electron-deficient alkene despite the fact that enol ethers are
typically regarded as nucleophilic/electron rich. (While unusual,
this is entirely consistent with the requirement for the C3/C4 acetate
groups; vide infra.)
In contrast, and as expected on the basis of the reactivity of 1,
tri-O-acetyl-D-galactal (16) undergoes ready cycloaddition with
benzyl azide to form triazoline 17; subsequent photolysis and ring
opening efficiently provide the corresponding methylglycoside (18,
94%; Scheme 3; conditions as Scheme 2). Equally important, serial
replacement of the acetate groups has revealed that although the
C3/C4 acetates are required, 2,3-di-O-acetyl-6-O-tert-butyldi-
methylsilyl-D-glucal (19) reacts with benzyl azide to form triazoline
20 in moderate (57%) yield.19 Subsequent photochemical ring
contraction and opening cleanly afford methyl glycoside 21 (88%).
A final observation suggests further expansion of the substrate
scope: benzyl azide and vinyl acetate react readily under our
standard conditions, although rapid in situ elimination leads to
formation of N-benzyltriazole as the only isolable product. Thus,
while the issue of elimination must be addressed, the methodology
can be extended to acyclic alkenes.
(11) While we have encountered no problems in this respect, CAUTION should
always be used when heating solutions of azides.
(12) We believe the acid- and base-lability of the triazoline lead to the re-
quirement for trialkyl orthoformate as a non-basic acid-scavenging solvent.
(13) For early examples of the photochemical generation of aziridines from
triazolines, see: (a) Scheiner, P. J. Org. Chem. 1967, 32, 2022. (b)
Scheiner, P. J. Am. Chem. Soc. 1968, 90, 988. (c) Scheiner, P. Tetrahedron
1968, 24, 2757.
(14) The solvent specificity and requirement for a quartz reaction vessel are
evidence that acetone serves both as solvent and triplet sensitizer for
aziridine formation. For a previous example of sensitized photochemical
conversion of a triazoline to an aziridine, see ref 12b.
(15) See Supporting Information for experimental details and 1H NMR spectra.
(16) C2 configurations were assigned based on X-ray crystallographic analysis
of 4b. The C1 configuration was assigned based on 1H coupling constants.
(17) Formation of the acetone adduct could also proceed through O-alkylation
of acetone followed by rearrangement. We thank Prof. C. Rojas (Barnard
College) for bringing this possibility to our attention. See: Zhang, Y.;
Reynolds, N. T.; Manju, K.; Rovis, T. J. Am. Chem. Soc. 2002, 124, 9720.
(18) For examples and discussion, see: (a) Gothelf, K. V.; Jørgensen, K. A.
Chem. ReV. 1998, 98, 863. (b) Scriven, E. F. V.; Turnbull, K. Chem. ReV.
1988, 88, 298.
(19) Loss of the TBS group during the cycloaddition accounts for the majority
of material loss. Optimization is currently under way in the context of
identifying an appropriate silyl linker for solid-phase synthesis.
In conclusion, we have discovered a new route to aminoglyco-
sides based on the formation of triazolines from a readily available
glycal precursor followed by efficient formation and ring-opening
JA0319238
9
J. AM. CHEM. SOC. VOL. 126, NO. 27, 2004 8357