Unique reactions of glycosyl iodides with oxa- and thiocycloalkane acceptors

Article abstract of DOI:10.1021/ol049966w

(Equation presented) Glucosyl-, galactosyl-, and mannosyl iodides efficiently react with strained oxacycloalkane acceptors to afford O-glycosides with high β-selectivity. The mechanism of ring opening was investigated by reacting mannosyl iodides with pur

Full text of DOI:10.1021/ol049966w

ORGANIC
LETTERS
2004
Vol. 6, No. 6
973-975
Unique Reactions of Glycosyl Iodides
with Oxa- and Thiocycloalkane
Acceptors
Darrin R. Dabideen and Jacquelyn Gervay-Hague*
Department of Chemistry, UniVersity of California DaVis, DaVis, California 95616
gerVay@chem.ucdaVis.edu
Received January 5, 2004
ABSTRACT
Glucosyl-, galactosyl-, and mannosyl iodides efficiently react with strained oxacycloalkane acceptors to afford O-glycosides with high â-selectivity.
The mechanism of ring opening was investigated by reacting mannosyl iodides with pure enantiomers of propylene oxide and styrene oxide.
Competition experiments between three- and five-membered oxacycloalkanes were also investigated. Finally, â-thiomannosides were synthesized
from thiocycloalkane acceptors.
In recent years, we have demonstrated the widespread utility
of glycosyl iodides en route to synthetically challenging
Scheme 1
glycoconjugates.1 Early in these investigations, difficulties
were encountered when performing the reactions in THF.
Under conditions that promote in situ anomerization, the THF
adduct (3) was obtained rather than the desired R-glycoside
(Scheme 1).2 The R-glycoside likely results from attack of
the acceptor on the more reactive â-iodide (2) to yield an
oxonium intermediate that undergoes ring opening to yield
3. This is essentially the reverse reaction to that invoked by
Fraser-Reid to explain the remarkable reactivities of pent-
4-enyl glycosides.3 Although mechanistically interesting, we
initially viewed THF addition as a limitation and avoided
using it as a solvent in subsequent studies. We have now
revisited this anomalous result and report our findings as
they relate to the unique reactivity of glycosyl iodides with
(1) (a) Lam, S. N.; Gervay-Hague, J. Org. Lett. 2002, 4, 2039-2042.
both oxa- and thiocyclic ethers.
(b) Lam, S. N.; Gervay-Hague, J. Carbohydr. Res. 2002, 337, 1953-1965.
Glycosyl iodides are quantitatively prepared from the
corresponding glycosyl acetate upon reaction with TMSI.4
Trimethylsilyl acetate (TMSOAc) is liberated, and it too can
act as an acceptor, giving rise to the â-glycosyl acetate.
(c) Bhat, A. S.; Gervay-Hague, J. Org. Lett. 2001, 3, 2081-2084. (d)
Gervay, J. Organic Synthesis: Theory and Applications; JAI Press:
Greenwich, 1998; pp 121-153. (e) Hadd, M. J.; Gervay, J. Carbohydr.
Res. 1999, 320, 61-69. (f) Gervay, J.; Hadd, M. J. J. Org. Chem. 1997,
62, 6961-6967.
(2) Hadd, M. J. Ph.D. Thesis, University of Arizona, Tucson, AZ, 1998.
(3) Fraser-Reid, B.; Konradsson, P.; Mootoo, D. R.; Udodong, U. J.
Chem. Soc., Chem. Commun. 1988, 823-825.
(4) Thiem, J.; Meyers, B. Chem. Ber. 1980, 113, 3075-85.
10.1021/ol049966w CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/13/2004

Recombination can be minimized by removing TMSOAc in
vacuo, but this maneuver is tedious because judicious
exclusion of water is required. We now find that the addition
of MgO effectively deactivates TMSOAc toward addition.
In all of the reactions reported in this study, MgO (2 equiv)
is added to the acetate prior to the introduction of TMSI.
MgO does not interfere with glycosyl iodide formation or
acceptor reactivity. Similar results were obtained in the
absence of MgO with in vacuo removal of TMSOAc,
supporting the hypothesis that MgO serves only as a Lewis
base that sequesters TMSOAc.
way around this problem is to utilize tethering strategies to
deliver the acceptor from the â-face.7 Other approaches
involve functionalizing mannose in a fashion that confor-
mationally biases â-attack.8
Initially, we reacted per-O-benzyl-R-mannosyl iodide with
THF, but elimination of HI was the major reaction manifold.
This difficulty was easily overcome by replacing the C-6
O-benzyl with an electron-withdrawing group (O-acetyl)
(Scheme 3).9 Mannosyl iodide (8) gave predominantly â
Initial studies focused on reacting per-O-benzylated glu-
cosyl iodide 1 with 1.5 equiv of THF and MgO in CH2Cl2
(Scheme 2). The reaction was sluggish at room temperature
Scheme 3
Scheme 2
products with both THF (9) and trimethylene oxide acceptors
(10), and as seen with 1 and 5, the more highly strained
oxacyclobutane afforded greater reactivity and stereoselec-
tivity.
To probe the mechanism of ring opening, 8 was indepen-
dently reacted with pure enantiomers of propylene oxide.
Unsymmetrical cyclic ether systems introduce the possibility
of both regio- and stereoisomers in the resulting products.
If ring opening proceeds through an SN2 mechanism, as sug-
gested in Scheme 1, then each enantiomer of propylene oxide
could afford four possible products, i.e., one set of regio-
isomers for each anomer. This was indeed the case as deter-
mined from analysis of the 1H and 13C NMR of the products
resulting from the reaction shown in Scheme 4. Reaction of
but proceeded to completion at reflux for 24 h. Upon workup
and purification, a 69% yield of 3 was obtained in a 3:1
R: â anomeric ratio. Encouraged by these results we next
reacted 1 with tetrahydropyran, but the reaction was far more
sluggish, giving only low yields of the expected product. In
contrast, reaction of 1 with trimethylene oxide occurred
within 7 h at room temperature giving predominantly the
â-glycoside 4 (1:8).
Previous mechanistic studies in our laboratory suggest that
â-glycosides result from SN2-like displacement of the
R-iodide. We believe that cyclic ethers, rather than ring-
opened iodo alcohols or iodo alcoholates, serve as the ac-
ceptors in these reactions as we saw no evidence of iodo
alcohol byproducts in any of our investigations. Moreover,
we were unable to effect ring opening under analogous
reactions where only the glycosyl iodide was omitted. The
relative reactivities of these cyclic ethers correlate well with
calculated strain energies with increased ring strain translat-
ing to increased reactivity.5 Analogous reactions with per-
O-benzylated galactosyl iodide (5) were equally productive,
as shown in Scheme 2.
Scheme 4
Having accomplished the aforementioned â-selective gly-
cosidations, we naturally turned our attention to mannosyl
iodides. Achieving â-selectivity with mannose-derived do-
nors is notoriously difficult as both the anomeric effect and
anchiomeric assistance favor R-glycoside formation.6 One
8 with (R)-propylene oxide was complete within 30 min,
and a 1:1 ratio of regioisomers was obtained. Reaction with
(6) Gridley, J. J.; Osborn H. M. I. J. Chem. Soc., Perkin Trans. 1 2000,
1471-1491.
(7) (a) Barresi, F.; Hindsgaul, O. Front. Nat. Prod. Res. 1996, 1, 251-
276. (b) Stork, G.; LeClair, J. J. J. Am. Chem. Soc. 1996, 118, 247-248.
(5) Vila, A.; Mosquera, R. A. Tetrahedron 2001, 57, 9415-9422.
974
Org. Lett., Vol. 6, No. 6, 2004

the (S)-enantiomer required 1 h for completion, and the regio-
isomeric distribution favored attack at the least-substituted
epoxide carbon (2:1). These epoxides are even more reactive
than trimethylene oxide, albeit somewhat less â-selective.
We reasoned that regiochemical scrambling could be
overcome by incorporating a stabilizing functionality proxi-
mal to the incipient positive charge. Again, this proved to
be true. However ring opening proceeded in a stepwise rather
than concerted process. For example, when 8 was treated
with (R)- and (S)-styrene oxide, the reaction was far more
sluggish (reflux, 20 h) and the only products resulted from
an SN1 mechanism at the benzylic position (Scheme 5).
iodide donors. First, reactions proceed without donor activa-
tion. Glycosyl iodides are unique in this respect, as analogous
reactions with glycosyl bromides were not productive. Highly
â-selective glycosidations occur with glucosyl, galactosyl,
and mannosyl iodides when combined with reactive acceptors
such as propylene oxide and trimethylene oxide. Less reactive
acceptors including THF and deactivated epoxides tend to
favor R-attack. Finally, in the absence of carbocation
stabilization, ring opening of the ether acceptor proceeds by
SN2 attack.
Having demonstrated the unique acceptor reactivity of
cyclic oxygenated ethers, we were naturally curious about
analogous thioether acceptors. Reaction of 8 with tetrahy-
drothiophene gave no reaction, which is perhaps not surpris-
ing considering it has little strain energy (3.1 kcal/mol).10
In contrast, reaction of trimethylene sulfide with 8 provided
predominantly the â-anomer after 6 h at reflux, and ethylene
sulfide afforded only the â-anomer after 18 h at reflux
(Scheme 7).11
Scheme 5
Scheme 7
Slightly different anomeric ratios were obtained for each
acceptor. Product distributions were determined from evalu-
ation of the reaction mixture 13C NMR, which showed four
anomeric signals (δc 98.2, 98.5, 101.7, 102.1) corresponding
to the four diastereomers resulting from attack of the iodide
at the benzylic position (δc 29.7, 29.9, 30.0, 30.9).
In an effort to access more highly functionalized systems
we also reacted 8 with 17. As anticipated, epoxide opening
predominated, but the reactivity of 17 more closely paralleled
styrene epoxide, requiring 14 h at reflux and yielding a 1.3:1
R: â anomeric mixture (Scheme 6).
The relative reactivities of these thioethers directly cor-
relate with reported strain energies of 22.2 and 19.1 kcal/
mol for thietan and thiirane, respectively.10 Further studies
probing the mechanism and extended applications of these
remarkably â-selective reactions are currently underway in
our laboratory.12
Scheme 6
Acknowledgment. We thank the NSF (CHE-0196482)
for generous financial support.
Supporting Information Available: General experimen-
tal procedure and NMR spectral data. This material is
available free of charge via the Internet at http://pubs.acs.org.
OL049966W
(10) Gronert, S.; Lee, J. M. J. Org. Chem. 1995, 60, 6731-36.
Several important features are evident from the combined
experiments using strained ether acceptors with glycosyl
1
(11) Anomeric stereochemistry was assigned on the basis of JC,H
coupling constant data as per: Bock, K.; Lundt, I.; Pedersen, C. Tetrahedron
Lett. 1973, 14, 1037-1040.
(12) For examples of conformationally biased â-selective thioglycosi-
dations, see: Crich, D.; Li, H. J. Org. Chem. 2000, 65, 801-805 and
references therein.
(8) (a) Hodosi, G.; Kova´c, P. J. Am. Chem. Soc. 1997, 118, 22335-2336.
(b) Crich, D.; Sun, S. J. Org. Chem. 1997, 62, 1198-99.
(9) Lam, S. N.; Gervay-Hague, J. Org. Lett. 2003, 22, 4219-22.
Org. Lett., Vol. 6, No. 6, 2004
975