New stereoselective β-C-glycosidation by uncatalyzed 1,4-addition of organolithium reagents to a glycal-derived vinyl oxirane

Article abstract of DOI:10.1021/ol034662f

(Matrix presented) Epoxide 4 is generated in situ from D-glucal-derived hydroxy mesylate 3. Reaction of epoxide 4 with a series of alkyl- and aryllithium reagents affords 2,3-unsaturated β-C-glycosides with excellent 1,4-regioselectivity and complete ster

Full text of DOI:10.1021/ol034662f

ORGANIC
LETTERS
2003
Vol. 5, No. 12
2173-2176
New Stereoselective â-C-Glycosidation
by Uncatalyzed 1,4-Addition of
Organolithium Reagents to a
Glycal-Derived Vinyl Oxirane
Valeria Di Bussolo, Micaela Caselli, Mauro Pineschi, and Paolo Crotti*
Dipartimento di Chimica Bioorganica e Biofarmacia, UniVersita` di Pisa,
Via Bonanno 33, I-56126 Pisa, Italy
crotti@farm.unipi.it
Received April 18, 2003
ABSTRACT
Epoxide 4 is generated in situ from D-glucal-derived hydroxy mesylate 3. Reaction of epoxide 4 with a series of alkyl- and aryllithium reagents
affords 2,3-unsaturated â-C-glycosides with excellent 1,4-regioselectivity and complete stereoselectivity for the â-glycoside. Other organometallic
reagents demonstrate more complex behavior in their reactions with epoxide 4.
C-Alkyl and C-aryl glycosides are carbohydrate analogues
of O-glycosides in which the C-O glycosidic linkage is
substituted by a C-C bond. This modification results in a
greater stability of these compounds to both acid and
enzymatic cleavage, compared with the corresponding O-
derivatives, to the point that C-glycosides can be advanta-
geously used as mimics of the corresponding more com-
monly encountered O-glycosides.1 Moreover, the C-glycoside
moiety occurs in several natural products with important
pharmacological properties.2 However, for an effective use
of these compounds, methods for a complete stereoselective
introduction of the C-C glycosidic bond are decidedly
valuable.3 In this framework, 2,3-unsaturated C-glycosides
appear to be of interest because the presence of the
unsaturation allows further functionalization.
Several synthetic methods to these unsaturated compounds
by Pd- or Ni-catalyzed addition reaction of organometallics
to a glycosyl donor have been described.4 More recently,
methods of C-glycosidation using arylboronic acids and Pd-
(AcO)2,5a nucleophilic addition of organozinc compounds
to glycals,6 and a Lewis acid mediated cross-coupling
reaction between chiral titanium enolates and glycals have
(3) (a) Jaramillo, C.; Knapp, S. Synthesis 1994, 1. (b) Du, Y.; Linhardt,
R. J.; Vlahov, I. R. Tetrahedron 1998, 54, 9913. (c) Togo, H.; He, W.;
Waki, Y.; Yokoyama, M. Synlett 1998, 700.
(1) (a) Leavy, D. E.; Tang, C. The Chemistry of C-Glycosides; Perga-
mon: Oxford, 1995. (b) Postema, M. H. D. Tetrahedron 1992, 48, 8545.
(c) Nicotra, F. Top. Curr. Chem. 1997, 187, 55. (d) Weatherman, R. V.;
Mortell, K. H.; Chervenak, M.; Kiessling, L. L.; Toone, E. J. Biochemistry
1996, 35, 3619. (e) Ravishankar, R.; Surolia, A.; Vijayan, M.; Lim, S.;
Kishi, Y. J. Am. Chem. Soc. 1998, 120, 11297. (f) Garc´ıa-Herrero, A.;
Montero, E.; Mun˜oz, J. L.; Espinosa, J. F.; Via´n, A.; Garc´ıa, J. L.; Asensio,
J. L.; Can˜ada, F. J.; Jime´nez-Barbero, J. J. Am. Chem. Soc. 2002, 124,
4804.
(2) (a) Faul, M. M.; Huff, B. E. Chem. ReV. 2000, 100, 2407. (b) Hacksell,
U.; Daves, G. D., Jr. Prog. Med. Chem. 1985, 22, 1. (c) Townsend, L. B.;
Devivar, R. V.; Turk, S. R.; Nassiri, M. R.; Drach, J. C. J. Med. Chem.
1995, 38, 4098. (d) Gudmundsson, K. S.; Drach, J. C.; Townsend, L. B. J.
Org. Chem. 1997, 62, 3453
(4) (a) Kwok, D.-I.; Farr, R. N.; Daves, G. D., Jr. J. Org. Chem. 1991,
56, 3711. (b) Czernecki, S.; Dechavanne, V. Can. J. Chem. 1983, 61, 533.
(c) Daves, G. D., Jr. Acc. Chem. Res. 1990, 23, 201. (d) Dunkerton, L. V.;
Serino, A. J. J. Org. Chem. 1982, 47, 2812. (e) Yougai, S.; Miwa, T. J.
Chem. Soc., Chem. Commun. 1983, 68. (f) RajanBabu, T. V. J. Org. Chem.
1985, 50, 3642. (g) Brakta, M.; Le Borgne, F.; Sinou, D. J. Carbohydr.
Chem. 1987, 6, 307. (h) Dunkerton, L. V.; Euske, J. M.; Serino, A. J.
Carbohydr. Res. 1987, 171, 89. (i) Moineau, C.; Bolitt, V.; Sinou, D. J.
Org. Chem. 1998, 63, 582.
(5) Ramnauth, J.; Poulin, O.; Rakhit, S.; Maddaford, S. P. Org. Lett.
2001, 3, 2013.
(6) Steinhuebel, D. P.; Fleming, J. J.; Du Bois, J. Org. Lett. 2002, 4,
293.
10.1021/ol034662f CCC: $25.00 © 2003 American Chemical Society
Published on Web 05/16/2003

been reported.7 In these C-glycosidation methods involving
the addition of carbon nucleophiles to anomeric carbon
electrophiles, R-stereoselectivity is well documented.4a-h,5,6
On the other hand, a general synthesis of â-C-glycosides
that is not dependent on functionality in the substrate7 or on
the influence of a catalyst4i is still a significant challenge,
especially in consideration of the importance of these
compounds in naturally occurring glycoconjugates.1a,8
protocol at -40 °C12 afforded the monobenzyl derivative 7
with selective alkylation at the primary alcoholic functional-
ity. Monoprotection of 7 with TBDMS-Cl gave the 3-O-
TBS derivative 8, which was treated with MsCl/Py to give
the mesylate 9. The 3-O-TBS group of 9 was removed by
TBAF to give the hydroxy mesylate 3, the necessary
precursor of epoxide 4 (Scheme 3).
Recently, we have introduced into the synthesis a simple,
completely â-stereoselective O-glycosylation by means of a
regioselective 1,4-addition of O-nucleophiles (alcohols and
phenol) to the previously undescribed 6-O-trityl vinyl oxirane
1, derived from D-glucal (Scheme 1).9
Scheme 3. Synthesis of Hydroxy Mesylate 3
Scheme 1. O-Glycosylation of the in Situ-Formed Epoxide 1
with Alcohols and Phenol
This result prompted us to examine the possibility of
obtaining similar regio- and stereoselectivity also with
C-nucleophiles, to address the long-standing goal of syn-
thesizing â-C-glycosides.
We now report a new process that allows direct synthetic
access to â-C-glycosides from vinyl oxirane 4, the 6-O-Bn
analogue of 1.10 As in the case of 1, vinyl oxirane 4 is not
sufficiently stable to be isolated and can only be prepared
in situ (t-BuOK/benzene or Et2O) from the corresponding
hydroxy mesylate 3, immediately prior to use (Scheme 2).11
To our delight, when vinyl oxirane 4 (the glycosyl donor)
was exposed to organolithium reagents (the glycosyl accep-
tor, 3 equiv) in Et2O at 0 °C, a clear, regioselective 1,4-
addition reaction (conjugate addition) was obtained affording
the corresponding C-glycosides in good yield and with
complete â-stereoselectivity (compounds 10-14â, Table 1).13
The reaction is amenable to a variety of alkyllithium reagents,
and also a sterically hindered glycosyl acceptor like t-BuLi
can be effectively used. It should also be noted that the sp2
hybridization of the glycosyl acceptor, as in the case of PhLi,
does not change the regio- or stereoselectivity of the addition
process (entry 5, Table 1).
The protocol necessarily requires a separate generation in
situ of the reacting species, the vinyl oxirane 4, by
independent cyclization of the corresponding hydroxy me-
sylate 3 with t-BuOK prior to the addition of the organo-
metallic reagents (protocol A). In fact, when hydroxy
mesylate 3 was left to react directly with alkyllithium and
PhLi (3 equiv), with no pretreatment with t-BuOK (protocol
B),14 only complex reaction mixtures were obtained, indicat-
ing that no formation of the corresponding epoxide had
occurred under these conditions.15
Scheme 2. In Situ Preparation of Epoxide 4
Hydroxy mesylate 3 was prepared starting from tri-O-
acetyl-D-glucal (5), which was initially deprotected to D-
glucal (6). Monoalkylation of 6 with the LHMDS/BnBr
The â-C-glycoside configuration of the addition products
10-14â was firmly established by the presence in these
(7) Larrosa, I.; Romea, P.; Urp´ı, F.; Balsells, D.; Villarasa, J.; Font-
Bardia, M.; Solans, X. Org. Lett. 2002, 4, 4651.
(8) (a) Rohr, J.; Thiericke, R. Nat. Prod. Rep. 1992, 103. (b) Hansen,
M.; Hurley, L. H. Acc. Chem. Res. 1996, 29, 249.
(12) Deshpande, P. P.; Kim, H. M.; Zatorski, A.; Park, T.-K.; Ragupathi,
G.; Livingston, P. O.; Live, D.; Danishefsky, S. J. J. Am. Chem. Soc. 1998,
120, 1600.
(9) Di Bussolo, V.; Caselli, M.; Pineschi, M.; Crotti, P. Org. Lett. 2002,
4, 3695.
(10) In the present context, the 6-O-trityl group of the previous epoxide
1 was replaced with the benzyl group because of the greater stability of the
latter to a wide range of reaction conditions.
(11) As previously performed in the case of epoxide 1, the formation of
epoxide 4 by t-BuOK cyclization of hydroxy mesylate 3 was established
by running the reaction in C6D6 in an NMR tube and recording the 1H
NMR spectrum a few minutes after the addition of the base.
(13) Less than 5% of the corresponding anti 1,2-addition products were
present in some cases. Products from an R-1,4 addition process were not
observed at all.
(14) Under protocol B, it was hoped that the organometallic reagents
might behave as a base and thereby promote epoxide formation from
hydroxy mesylate 3 and then undergo an addition reaction.
(15) This observation indicates that alkyllithium and PhLi do not behave,
in the present case, as a base, which is necessary for the cyclization of
hydroxy mesylate 3 to epoxide 4.
2174
Org. Lett., Vol. 5, No. 12, 2003

affording completely different results as regards the regio-
and stereoselectivity. In some cases (phosphoramidite-
catalyzed addition of dialkylzinc reagents17 and Pd-catalyzed
addition of boronic acids),5 only very complex mixtures were
obtained that did not contain traces of the corresponding
addition products.
Table 1. Regio- and Stereoselectivity of C-Glycosidation of
the in Situ-Formed Epoxide 4 with Organolithium Reagents13
As for the Grignard reagents, their behavior depends on
the protocol (A or B). In the presence of a preventive
cyclization (t-BuOK) of hydroxy mesylate 3 to epoxide 4
(protocol A), the use of MeMgBr and PhMgBr (3 equiv)
resulted in an exclusive 1,4-addition of t-BuOH,18 whereas
the use of PhMgCl led to a 3:7 mixture of the corresponding
R- (14R) and â-C-glycosides (14â) (1,4-adducts). On the
contrary, when the Grignard reagents (MeMgBr and Ph-
MgBr, 3 equiv) were directly brought into contact with
hydroxy mesylate 3 (protocol B), an almost 1:1 mixture of
the corresponding diastereoisomeric 4,5-dihydrofuran-derived
alcohols 16 and 17 turned out to be the only reaction product
(Scheme 5).
Scheme 5. Regio- and Stereoselectivity of the Addition
Reaction of Grignard Reagents and Cuprates to Hydroxy
Mesylate 3 and Epoxide 4
compounds of a clear NOE between protons H(1) and H(5),
which is completely absent in the corresponding R-stereo-
isomer, as demonstrated in the case of the phenyl-substituted
pair 14â and 14R (Scheme 4). Moreover, the presence of
Scheme 4. NOE in â-C-Glycoside 14â Compared with
R-C-Glycoside 14r.
Alcohols 16 and 17 reasonably derive from an isomer-
ization process, with ring contraction, of epoxide 4, formed
under the basic reaction conditions of the reagent (RMgX),
to the intermediate aldehyde 20 (Scheme 6).19 This isomer-
ization is reasonably promoted by the oxophilic character
of the magnesium salts present in the reaction mixture
(16) The only exception is given by the â-methyl derivative 10â [δ C(5)
) 73.79 ppm]. However, the â-C-glycosidic configuration for this compound
was firmly established by corresponding NOE experiment.
(17) Bertozzi, F.; Crotti, P.; Feringa, B. L.; Macchia, F.; Pineschi, M.
Synthesis 2001, 483 and pertinent references therein.
(18) From t-BuOK used for the cyclization of 3 to epoxide 4.
(19) As suggested by a Rewiever, epoxide 4 is not a necessary
intermediate for the generation of 16 and 17, via protocol B. A Grob
fragmentation process on hydroxy mesylate 3 could lead directly to the
observed products.
chemical shift values for C(5) higher than 75 ppm in the
13C NMR spectra of compounds 10-14â is diagnostic for a
1,5-cis relationship between substituents at C(1) and C(5)
in 2,3-unsaturated C-glycosides.5,16
Other organometallics, such as Grignard reagents and
cuprates, showed a more complex behavior with epoxide 4,
Org. Lett., Vol. 5, No. 12, 2003
2175

Scheme 6. Formation of the Intermediate Aldehyde 20
Scheme 7. Rationalization of the 1,4-Regio- and
â-Stereoselectivity of the Addition Reaction of RLi (R = Alkyl
and Ph) to Epoxide 4
(Schlenk equilibrium) and proceeds through a typical con-
certed, highly stereocontrolled mechanism with migration of
the C(4)-C(5) bond to the developing C(3)-carbocationic
center, as shown in structure 19 (Scheme 6).20 Subsequent
nonstereoselective nucleophilic addition of the reagent
(RMgX) to the intermediate aldehyde 20 affords alcohols
16 and 17, as found experimentally. Oxidation of the
separated alcohols 16 and 17 by PCC/CH2Cl2 yielded the
same ketone 18, indicating that 16 and 17 have the same
configuration at C(4) and C(5), as independently suggested
by examination of the reasonable reaction mechanism relative
to their formation (Schemes 5 and 6).
brought to the â-face of the glycal system and appropriately
disposed to nucleophilically attack C(1) to give the 1,4-regio-
and â-stereoselective result observed. The â-1,4-attack on
C(1) by the reagent engaged in the chelated structure 21
(internal nucleophile) (route a), rather than an R-1,4-attack
(route b) or an anti 1,2-attack (route c, Scheme 7) by an
external nucleophile, appears to be favored by entropic
factors to the point that products arising from routes b and
c are consequently not observed.13
Studies are underway to obtain more evidence in favor of
the proposed mechanism and to verify how the nature of
the nucleophile and changes in the branched chain of vinyl
oxiranes of type 4 can modify the regio- and stereoselective
result so far obtained both in the O-glycosylation and in the
present C-glycosidation reaction.
Under protocol A reaction conditions, cuprates such as
Me2CuLi or EtMgBr in the presence of CuCN afforded only
the corresponding anti 1,2-adduct 15 (Scheme 5). Interest-
ingly, this result is in contrast with the 1,4-regioselectivity
normally observed with cuprates in their addition reactions
to vinyl oxiranes.21
The regio- and stereoselective result obtained with orga-
nolithium reagents (RLi, R ) alkyl, Ph) in their reaction
with epoxide 4, under protocol A, may be tentatively
rationalized by admitting the incursion in the reaction
pathway of a coordination of the lithium counterpart of the
organometallic compound with the oxirane oxygen, probably
followed by a bidentate chelation of the metal with both the
endocyclic and exocyclic oxygen, as shown in structure 21
(Scheme 7). In this way, the reagent (RLi) is effectively
Acknowledgment. This work was supported by the
University of Pisa and MIUR (Ministero dell’Istruzione,
dell’Universita` e della Ricerca), Roma. P.C. gratefully
acknowledges Merck Research Laboratories for the generous
financial support derived from the 2002 ADP Chemistry
Award.
Supporting Information Available: Experimental details
and spectral and analytical data for all reaction products. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(20) Rickborn, B. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Fleming, I., Pattenden, G., Eds.; Pergamon: Oxford, 1991; Vol 3, pp 733-
775.
(21) Marshall, J. A. Chem. ReV. 1989, 89, 1503.
OL034662F
2176
Org. Lett., Vol. 5, No. 12, 2003