The effect of electrostatic interactions on conformational equilibria of multiply substituted tetrahydropyran oxocarbenium ions

Article abstract of DOI:10.1021/jo8017846

The three-dimensional structures of dioxocarbenium ions related to glycosyl cations were determined by an analysis of spectroscopic, computational, and reactivity data. Hypothetical low-energy structures of the dioxocarbenium ions were correlated with both experimentally determined ¹H NMR coupling constants and diastereoselectivity results from nucleophilic substitution reactions. This method confirmed the pseudoaxial preference of C-3 alkoxy-substituted systems and revealed the conformational preference of the C-5 alkoxymethyl group. Although the monosubstituted C-5 alkoxymethyl substituent preferred a pseudoequatorial orientation, the C-5-C-6 bond rotation was controlled by an electrostatic effect. The preferred diaxial conformer of the trans-4, 5-disubstituted tetrahydropyranyl system underscored the importance of electrostatic effects in dictating conformational equilibria. In the 2-deoxymannose system, although steric effects influenced the orientation of the C-5 alkoxymethyl substituent, the all-axial conformer was favored because of electrostatic stabilization.

Full text of DOI:10.1021/jo8017846

The Effect of Electrostatic Interactions on Conformational Equilibria of
Multiply Substituted Tetrahydropyran Oxocarbenium Ions
Michael T. Yang and K. A. Woerpel*
Department of Chemistry, UniVersity of California, IrVine, California 92697-2025
kwoerpel@uci.edu
ReceiVed August 8, 2008
The three-dimensional structures of dioxocarbenium ions related to glycosyl cations were determined by
an analysis of spectroscopic, computational, and reactivity data. Hypothetical low-energy structures of
the dioxocarbenium ions were correlated with both experimentally determined 1H NMR coupling constants
and diastereoselectivity results from nucleophilic substitution reactions. This method confirmed the
pseudoaxial preference of C-3 alkoxy-substituted systems and revealed the conformational preference of
the C-5 alkoxymethyl group. Although the monosubstituted C-5 alkoxymethyl substituent preferred a
pseudoequatorial orientation, the C-5-C-6 bond rotation was controlled by an electrostatic effect. The
preferred diaxial conformer of the trans-4,5-disubstituted tetrahydropyranyl system underscored the
importance of electrostatic effects in dictating conformational equilibria. In the 2-deoxymannose system,
although steric effects influenced the orientation of the C-5 alkoxymethyl substituent, the all-axial conformer
was favored because of electrostatic stabilization.
Introduction
the treatment of influenza,15,16 diabetes,17,18 and cancer.19
Although transition state analogues of glycosyl transfer have
been identified20-22 and many inhibitors of glycosidases have
been developed,23-28 the three-dimensional structures of car-
bohydrate-derived oxocarbenium ions are not well understood,
and only recently have computational studies of these structures
emerged.14,29-32 A better understanding of the three-dimensional
structures of these cations would clarify the roles of these
intermediates and transition states and facilitate the design of
therapeutic glycosidase inhibitors.14,29,30,33,34
Oxocarbenium ions play important roles in both the synthetic
and bioorganic chemistry of carbohydrates. The synthesis of
oligosaccharides generally involves treatment of glycosyl acetals
with Lewis or Brønsted acids to form carbocationic intermedi-
ates, which are then trapped by sugar nucleophiles to form the
glycosidic linkage.1 Knowledge of the three-dimensional struc-
tures, relative energies, and reactivities of oxocarbenium ions
would facilitate the design of selective glycosylation reactions.2-7
In addition, understanding the conformational preferences of
oxocarbenium ions would impact glycobiology and medicinal
chemistry. Enzymes that catalyze glycosyl transfer, which can
operate by stabilizing transition states with substantial oxocar-
benium ion character,8-14 have emerged as potential targets for
Our studies of the conformational preferences of oxocarbe-
nium ions revealed that electronic effects exert strong influences
on the conformational preferences of these cations. For reactions
¨
(8) Unligil, U. M.; Rini, J. M. Curr. Opin. Struct. Biol. 2000, 10, 510–517.
(9) Rye, C. S.; Withers, S. G. Curr. Opin. Chem. Biol. 2000, 4, 573–580.
(10) Sinnott, M. L. Chem. ReV. 1990, 90, 1171–1202.
(11) Kempton, J. B.; Withers, S. G. Biochemistry 1992, 31, 9961–9969.
(12) Tull, D.; Withers, S. G. Biochemistry 1994, 33, 6363–6370.
(13) Zechel, D. L.; Withers, S. G. Acc. Chem. Res. 2000, 33, 11–18.
(14) Ionescu, A. R.; Whitfield, D. M.; Zgierski, M. Z.; Nukada, T. Carbohydr.
Res. 2006, 341, 2912–2920.
(15) Moscona, A. New Engl. J. Med. 2005, 353, 1363–1373.
(16) von Itzstein, M. et al. Nature 1993, 363, 418–423.
(17) Breuer, H.-W. M. Int. J. Clin. Pharmacol. Ther. 2003, 41, 421–440.
(18) Scott, L. J.; Spencer, C. M. Drugs 2000, 59, 521–549.
(19) Goss, P. E.; Baker, M. A.; Carver, J. P.; Dennis, J. W. Clin. Cancer
Res. 1995, 1, 935–944.
(1) Whitfield, D. M.; Douglas, S. P. Glycoconjugate J. 1996, 13, 5–17.
(2) Jensen, H. H.; Pedersen, C. M.; Bols, M. Chem. Eur. J. 2007, 13, 7576–
7582.
(3) Jensen, H. H.; Bols, M. Acc. Chem. Res. 2006, 39, 259–265.
(4) Roush, W. R.; Sebesta, D. P.; Bennett, C. E. Tetrahedron 1997, 53, 8825–
8836.
(5) Durham, T. B.; Roush, W. R. Org. Lett. 2003, 5, 1871–1874.
(6) Arne´s, X.; D´ıaz, Y.; Castillo´n, S. Synlett 2003, 2143–2146.
(7) Morales-Serna, J. A.; Boutureira, O.; D´ıaz, Y.; Matheu, M. I.; Castillo´n,
S. Carbohydr. Res. 2007, 342, 1595–1612.
10.1021/jo8017846 CCC: $40.75
Published on Web 12/10/2008
2009 American Chemical Society
J. Org. Chem. 2009, 74, 545–553 545

Yang and Woerpel
the pseudoaxial and pseudoequatorial conformers,42 the pseu-
doequatorial conformation 3 is stabilized by hyperconjugation
from the C-2 carbon-hydrogen bond.43
Knowledge of the contributions of individual substituents to
the conformational preferences of oxocarbenium ions, however,
is not sufficient to predict the reactivities of highly substituted
systems such as those formed from carbohydrates.44 For
instance, although the influences of the C-2, C-3, and C-4
substituents on the mannopyranosyl cation should reinforce each
other to favor the ꢀ-product, R-selectivity was observed upon
allylation of the mannosyl phosphate 4 (eq 1).45-48 This
FIGURE 1. Preferred conformations of the oxocarbenium ion inter-
mediates in nucleophilic substitution reactions of monosubstituted
tetrahydropyran oxocarbenium ions.
involving C-3 and C-4 alkoxy-substituted tetrahydropyran
cations,35 the pseudoaxial conformations of the oxocarbenium
ion intermediates 1 and 2 were consistent with the formation
of the 1,4-trans and the 1,3-cis products, respectively (Figure
1).36-38 The pseudoaxial conformations of the oxocarbenium
ions are favored because they are stabilized by electrostatic
interactions between the positively charged carbon atoms of the
cations and the partially negatively charged substituents.36,37,39-41
The results with substrates bearing an alkoxy group at C-2
revealed that although electrostatic effects are similar between
observation could be the result of a change in the conformational
equilibria of the oxocarbenium ion such that the cationic
intermediate now prefers the equatorial conformation 7 (eq 2)
to avoid 1,3-diaxial interaction between the C-3 and C-5
substituents. Alternatively, the reversal in facial selectivity could
be the result of a Curtin-Hammett scenario,49,50 where the
product distribution was dictated by the relative reactivities of
the low-energy conformers. In this case, nucleophilic addition
would occur via the higher energy conformer 7, given that the
kinetic barriers to the stereoelectronically disfavored twist
transition states were too high in energy and that the lower
energy conformer 6 was too sterically congested to react.
Although the selectivities exhibited by multiply substituted
cations suggested that the observation shown in eq 1 was the
result of reaction through intermediate 7,51 we required ad-
ditional evidence to confirm this hypothesis.
(20) Lillelund, V. H.; Jensen, H. H.; Liang, X.; Bols, M. Chem. ReV. 2002,
102, 515–554.
(21) Gloster, T. M.; Meloncelli, P.; Stick, R. V.; Zechel, D.; Vasella, A.;
Davies, G. J. J. Am. Chem. Soc. 2007, 129, 2345–2354.
(22) Wicki, J.; Williams, S. J.; Withers, S. G. J. Am. Chem. Soc. 2007, 129,
4530–4531.
(23) Svansson, L.; Johnston, B. D.; Gu, J. H.; Patrick, B.; Pinto, B. M. J. Am.
Chem. Soc. 2000, 122, 10769–10775.
(24) Heck, M.-P.; Vincent, S. P.; Murray, B. W.; Bellamy, F.; Wong, C.-H.;
Mioskowski, C. J. Am. Chem. Soc. 2004, 126, 1971–1979.
(25) Davis, B. G.; Hull, A.; Smith, C.; Nash, R. J.; Watson, A. A.; Winkler,
D. A.; Griffiths, R. C.; Fleet, G. W. J. Tetrahedron: Asymmetry 1998, 9, 2947–
2960.
(26) Tong, M. K.; Papandreou, G.; Ganem, B. J. Am. Chem. Soc. 1990, 112,
6137–6139.
(42) In the pseudoaxial and pseudoequatorial conformers, the position of the
C-2 substituent changes through a C-2-C-1 bond rotation. Consequently, the
distance between the C-2 oxygen atom and the cationic C-1 carbon atom should
remain essentially the same. Previous calculation shows that the distance between
the C-2 oxygen atom and the C-1 carbon atom is 2.30 Å when the substituent
is axial and 2.33 Å when the substituent is equatorial (ref 40).
(43) Alabugin, I. V. J. Org. Chem. 2000, 65, 3910–3919.
(27) Ganem, B.; Papandreou, G. J. Am. Chem. Soc. 1991, 113, 8984–8985.
(28) Williams, S. J.; Notenboom, V.; Wicki, J.; Rose, D. R.; Withers, S. G.
J. Am. Chem. Soc. 2000, 122, 4229–4230.
(29) Nukada, T.; Be´rces, A.; Wang, L. J.; Zgierski, M. Z.; Whitfield, D. M.
Carbohydr. Res. 2005, 340, 841–852.
(30) Ionescu, A. R.; Whitfield, D. M.; Zgierski, M. Z. Carbohydr. Res. 2007,
342, 2793–2800.
(31) Amat, L.; Carbo´-Dorca, R. J. Chem. Inf. Comput. Sci. 2000, 40, 1188–
1198.
(44) Other intermediates, aside from discrete oxocarbenium ions, could also
lead to the nucleophilic substitution products. For an example involving ion-
pair intermediates, see: Crich, D.; Chandrasekera, N. S. Angew. Chem. Int. Ed.
2004, 43, 5386–5389. For the type of substrates presented in this work, neither
the Lewis acid nor the solvent has significant influence upon product distribution
(as described in ref 37).
(45) Allylation of mannosyl phosphate 4 has been shown to be highly
R-selective with Me3SiOTf: Plante, O. J.; Palmacci, E. R.; Andrade, R. B.;
Seeberger, P. H. J. Am. Chem. Soc. 2001, 123, 9545–9554.
(32) Winkler, D. A. J. Med. Chem. 1996, 39, 4332–4334.
(33) Jung, K. H.; Schmidt, P. R. In Carbohydrate Based Drug DiscoVery;
Wong, C.-H., Ed.; Wiley VCH: Weinheim, Germany, 2003; Vol. 2, pp. 609-
660.
(34) Davies, G. J.; Ducros, V. M.-A.; Varrot, A.; Zechel, D. L. Biochem.
Soc. Trans. 2003, 31, 523–527.
(35) Throughout this work, carbohydrate numbering will be used for
tetrahydropyran derivatives.
(36) Romero, J. A. C.; Tabacco, S. A.; Woerpel, K. A. J. Am. Chem. Soc.
2000, 122, 168–169.
(37) Ayala, L.; Lucero, C. G.; Romero, J. A. C.; Tabacco, S. A.; Woerpel,
K. A. J. Am. Chem. Soc. 2003, 125, 15521–15528.
(38) The transition states for simple half-chair conformers are rationalized
based on the principles of stereoelectronics: Deslongchamps, P. Stereoelectronic
Effects in Organic Chemistry; Pergamon: New York, 1983; pp. 209-221.
(39) Miljkovic´, M.; Yeagley, D.; Deslongchamps, P.; Dory, Y. L. J. Org.
Chem. 1997, 62, 7597–7604.
(40) Woods, R. J.; Andrews, C. W.; Bowen, J. P. J. Am. Chem. Soc. 1992,
114, 859–864.
(41) Larsen, C. H.; Ridgway, B. H.; Shaw, J. T.; Smith, D. M.; Woerpel,
K. A. J. Am. Chem. Soc. 2005, 127, 10879–10884.
(46) C-Mannosylation reactions are generally R-selective: (a) Lee, Y. J.; Baek,
J. Y.; Lee, B.-Y.; Kang, S. S.; Park, H.-S.; Jeon, H. B.; Kim, K. S. Carbohydr.
Res. 2006, 341, 1708–1716. (b) Beignet, J.; Tiernan, J.; Woo, C. H.; Kariuki,
B. M.; Cox, L. R. J. Org. Chem. 2004, 69, 6341–6356. (c) Brunel, F. M.; Taylor,
K. G.; Spatola, A. F. Tetrahedron Lett. 2003, 44, 1287–1289. (d) Nishikawa,
T.; Ishikawa, M.; Isobe, M. Synlett 1999, 123–125. (e) Roche, D.; Ba¨nteli, R.;
Winkler, T.; Casset, F.; Ernst, B. Tetrahedron Lett. 1998, 39, 2545–2548. (f)
Bertozzi, C.; Bednarski, M. Carbohydr. Res. 1992, 223, 243–253. (g) Panek,
J. S.; Sparks, M. A. J. Org. Chem. 1989, 54, 2034–2038. (h) Hosomi, A.; Sakata,
Y.; Sakurai, H. Carbohydr. Res. 1987, 171, 223–232. (i) Hosomi, A.; Sakata,
Y.; Sakurai, H. Tetrahedron Lett. 1984, 25, 2383–2386.
(47) R-Selectivity was also observed with acetate-protected mannose deriva-
tives in which the protecting group could be involved in the selectivity: (a)
McDevitt, J. P.; Lansbury, P. T. J. Am. Chem. Soc. 1996, 118, 3818–3828. (b)
Marron, T. G.; Woltering, T. J.; Weitz-Schmidt, G.; Wong, C.-H. Tetrahedron
Lett. 1996, 37, 9037–9040. (c) Nicolaou, K. C.; Hwang, C.-K.; Duggan, M. E.
J. Am. Chem. Soc. 1989, 111, 6682–6690. (d) Giannis, A.; Sandhoff, K.
Tetrahedron Lett. 1985, 26, 1479–1482.
(48) Allylation of mannosyl phosphate with no selectivity has been
reported: Singh, G.; Vankayalapati, H. Tetrahedron: Asymmetry 2001, 12, 1727–
1735.
546 J. Org. Chem. Vol. 74, No. 2, 2009

Multiply Substituted Tetrahydropyran Oxocarbenium Ions
diacetate 15. Deacetylation followed by benzylation, hydrolysis,
and oxidation yielded the requisite lactone (eq 5).59 Monosub-
stituted lactones 17 and 18 were prepared according to literature
procedures.36,51
In this paper, we describe spectroscopic studies that illuminate
the three-dimensional structures of highly substituted oxocar-
benium ions. Although oxocarbenium ions are too reactive52 to
be observed in many cases,53 the related dioxocarbenium
ions54,55 serve as isolable structural homologues, with confor-
mational preferences that are also sensitive to electronic
effects.56 By comparing the H NMR coupling constants of
1
dioxocarbenium ions with those predicted by computational
methods, we have established the conformational preferences
of both monosubstituted and multiply substituted dioxocarbe-
nium ions with structures related to glycosyl cations. In the
absence of severe steric interactions, electrostatic forces dictate
the conformational preferences of dioxocarbenium ions. This
conclusion underscores the importance of considering electro-
static influences as well as steric effects when predicting
conformational preferences in structurally flexible compounds.
FIGURE 2. Monosubstituted lactones.
The C-2 benzyloxy-substituted dioxocarbenium ions were not
examined because it was anticipated that they might give
misleading results. The presence of an alkoxy group at C-2
would not only destabilize the cationic center inductively,60 but
it would also magnify a destabilizing eclipsing interaction in
the dioxocarbenium ion (X ) OEt, eq 6) as compared to the
oxocarbenium ion (X ) H).
Experimental Approach
To probe the solution-phase conformational equilibria of the
mannosyl oxocarbenium ion, we prepared a series of mono-
substituted and multiply substituted tetrahydropyran dioxocar-
benium ions and analyzed their conformational preferences. This
approach was employed successfully for determining that the
C-4 alkoxy-substituted dioxocarbenium ion preferred the pseudo-
axial conformation.56
Because dioxocarbenium ions can be formed from lactones
by alkylation with Meerwein salts,54,55 a series of lactones were
prepared. The lyxo and mannosyl lactones 10 and 11 were
prepared by methanolysis of the corresponding sugars, followed
by benzylation,51 hydrolysis,51 and oxidation (eq 3). The
The conformational preferences of the dioxocarbenium ions
were established by comparing experimental 1H NMR coupling
constants to theoretical coupling constants of the low-energy
conformers determined by computational methods.61 The low-
energy conformers of methyl analogues62 of each cation (to
simplify the system) were found by a systematic search of the
conformational space using semi-empirical methods (PM3) in
Spartan.63,64 The energies of all minima were determined using
low-level ab initio calculations (HF/3-21G). The structures of
conformers with energies within 10 kcal/mol of the lowest
energy conformer were further optimized with ab initio calcula-
tions, using a larger basis set (HF/6-31G*) and also density
functional methods (B3LYP/6-31G*). All computed structures
(49) Seeman, J. I. Chem. ReV. 1983, 83, 83–134.
(50) Seeman, J. I. J. Chem. Educ. 1986, 63, 42–48.
(51) Lucero, C. G.; Woerpel, K. A. J. Org. Chem. 2006, 71, 2641–2647.
(52) Amyes, T. L.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111, 7888–7900.
(53) Alkyl oxocarbenium ions have previously been characterized by various
means. Superacidic solution: (a) Olah, G. A.; Dunne, K.; Mo, Y. K.; Szilagyi,
P. J. Am. Chem. Soc. 1972, 94, 4200–4205. (b) Olah, G. A.; Berrier, A. L.;
Prakash, G. K. S. J. Am. Chem. Soc. 1982, 104, 2373–2376. (c) Prakash, G. K. S.;
Rasul, G.; Liang, G.; Olah, G. A. J. Phys. Chem. 1996, 100, 15805–15809.
Electrochemical oxidation: (d) Suga, S.; Suzuki, S.; Yamamoto, A.; Yoshida, J.
J. Am. Chem. Soc. 2000, 122, 10244–10245. (e) Suzuki, S.; Matsumoto, K.;
Kawamura, K.; Suga, S.; Yoshida, J. Org. Lett. 2004, 6, 3755–3758. (f) Yoshida,
J.; Suga, S. Chem. Eur. J. 2002, 8, 2650–2658.
trisubstituted lactone 13 was prepared in one step by oxidation
of the commercially available glycal 12 (eq 4).57 Lactone 16
(54) Deslongchamps, P.; Cheˆnevert, R.; Taillefer, R. J.; Moreau, C.; Saunders,
J. K. Can. J. Chem. 1975, 53, 1601–1615.
(55) Childs, R. F.; Kostyk, M. D.; Lock, C. J. L.; Mahendran, M. Can.
J. Chem. 1991, 69, 2024–2032.
(56) Chamberland, S.; Ziller, J. W.; Woerpel, K. A. J. Am. Chem. Soc. 2005,
127, 5322–5323.
(57) Squarcia, A.; Vivolo, F.; Weinig, H.-G.; Passacantilli, P.; Piancatelli,
G. Tetrahedron Lett. 2002, 43, 4653–4655.
was prepared from the commercially available acetylated glycal
14 through a six-step sequence beginning with a Ferrier
reaction58 and hydrogenation of the resulting olefin to yield the
J. Org. Chem. Vol. 74, No. 2, 2009 547

Yang and Woerpel
Results and Discussion
Although understanding the conformational preferences of
monosubstituted systems alone was not enough to predict the
conformational preferences of multiply substituted systems,
studies of the simplified dioxocarbenium ions indicated that
correlation of spectroscopic data with theoretical calculations
could identify low-energy conformers. Spectroscopic studies of
the C-3 alkoxy-substituted dioxocarbenium ion 2171 confirmed
our earlier proposal37 that the 3-alkoxy oxocarbenium ion
adopted a pseudoaxial orientation. Alkylation of lactone 17
provided the dioxocarbenium ion 21 (eq 7), which decomposed
FIGURE 3. Relative energy of conformers at different levels of theory.
were confirmed to be minima using frequency calculations. The
relative energies between different conformers do not vary
substantially by the method used to calculate them, and the
energies converge at higher levels of theory (Figure 3). All
energy values are expressed as total energies (∆E). The
theoretical coupling constants were calculated de novo by using
the Gaussian03 modeling package,65 using density functional
theory (B3LYP/6-31G*).66 All calculated structures, relative
energies, and coupling constants were evaluated with the
Integral-Equation-Formalism Polarizable Continuum Model
(IEFPCM)67 to account for the effect of solvation in dichloro-
methane.
over five hours at room temperature but could be analyzed
spectroscopically at 0 °C. The small magnitude of the vicinal
(3J) coupling constants of the hydrogen at C-3 suggested that
the alkoxy group was oriented pseudoaxially. This observation
is consistent with computational studies (B3LYP/6-31G*) of
the corresponding methyl derivative, which showed that the
pseudoaxial conformer 22-ax is favored by 1.4 kcal/mol over
the equatorial conformer 22-eq (eq 8). The coupling constants
In all cases, the stereoselective C-glycosylation reactions of
related oxocarbenium ions were compared with the conforma-
tional preferences of the dioxocarbenium ions. Nucleophilic
substitution reactions of the oxocarbenium ions were performed
with acetate precursors obtained by reductive acetylation68 of
the corresponding lactones prepared earlier (vide supra). Al-
though the acetates were typically obtained as mixtures of
anomers, the anomer distributions did not affect the outcome
of the substitution reactions.69 Allyltrimethylsilane was chosen
as the nucleophile because addition of this nucleophile to
oxocarbenium ions provided kinetic adducts reliably.70 The
stereochemistry of the allylated products was assigned using a
combination of 1H NMR coupling constants and NOE measure-
ments.
for the computationally preferred conformer, 22-ax, are in
agreement with experimental J-values (Figure 4) and the
observed stereoselectivities of additions to the corresponding
oxocarbenium ion.37 We attribute the preference for the pseudo-
axial conformer 22-ax to electrostatic attraction between the
positively charged carbon atom of the dioxocarbenium ion72
and the negatively charged oxygen atom of the alkoxy group,
in agreement with our earlier studies with C-4 benzyloxy-
substituted dioxocarbenium ions.36,37,56
The influence of an alkoxymethyl group at C-5 of a
carbohydrate-derived oxocarbenium ion is more complicated
than the influence of an alkoxy group at either C-3 or C-4. In
addition to the ability to adopt either a pseudoaxial or a
pseudoequatorial orientation, the substituent has additional
conformational flexibility about the exocyclic C-5-C-6 bond,
(58) Ferrier, R. J.; Overend, W. G. J. Chem. Soc. 1962, 3667–3670.
(59) Sasaki, M.; Ishikawa, M.; Fuwa, H.; Tachibana, K. Tetrahedron 2002,
58, 1889–1911.
(60) Namchuk, M. N.; McCarter, J. D.; Becalski, A.; Andrews, T.; Withers,
S. G. J. Am. Chem. Soc. 2000, 122, 1270–1277.
(61) Unpurified 1H NMR spectra are provided as Supporting Information.
(62) The C-OMe rotamers found during the conformational search process
were either retained or excluded based on the selection criteria described in the
Experimental Section (see also “Details of Computation” in the Suppporting
Information). If more than one rotamer exists after the HF/3-21G 10 kcal/mol
filter, the lowest energy rotamer was used for the final analysis. The relative
energy of the rotamers appears to be consistent with steric considerations.
(63) Deppmeier, B. J.; Driessen, A. J.; Hehre, T. S.; Hehre, W. J.; Johnson,
J. A.; Klunzinger, P. E.; Leonard, J. M.; Pham, I. N.; Pietro, W. J.; Yu, J.
Spartan’02, version 115a; Wavefunction, Inc.: Irvine, CA, 2002.
(64) Kong, J. et al. J. Comput. Chem. 2000, 21, 1532–1548.
(67) Mennucci, B.; Cance`s, E.; Tomasi, J. J. Phys. Chem. B 1997, 101,
10506–10517.
(68) Dahanukar, V. H.; Rychnovsky, S. D. J. Org. Chem. 1996, 61, 8317–
8320.
(69) Shenoy, S. R.; Smith, D. M.; Woerpel, K. A. J. Am. Chem. Soc. 2006,
128, 8671–8677.
(70) Hagen, G.; Mayr, H. J. Am. Chem. Soc. 1991, 113, 4954–4961.
(71) Although the C-3 alkoxy-substituted dioxocarbenium ion 21 could be
formed at 0 °C, rapid decomposition within hours of product formation hindered
our efforts to obtain X-ray quality crystals.
(72) Our calculated Mullikan populations for the C-1 atom of dioxocarbenium
ions 22-ax or 22-eq are approximately 0.7. The net atomic charges for various
simple oxocarbenium ions were also previously calculated and ranged from 0.5
to 0.8: Woods, R. J.; Andrews, C. W.; Bowen, J. P. J. Am. Chem. Soc. 1992,
114, 850–858.
(65) Pople, J. A. et al. Gaussian03, Revision B.4; Gaussian, Inc.: Pittsburgh,
PA, 2003.
(66) (a) The validity of using the DFT/B3LYP method for calculating
NMR spin-spin coupling constants have been investigated: Sychrovsky´, V.;
Gra¨fenstein, J.; Cremer, D. J. Chem. Phys. 2000, 113, 3530–3547, Recent
examples of calculations of NMR coupling constants using the DFT/B3LYP
´
method. (b) Anizelli, P. R.; Vilcachagua, J. D.; Neto, A. C.; Tormena, C. F. J.
Phys. Chem. A 2008, 112, 8785–8789. (c) Silva, A. M. S.; Sousa, R. M. S.;
Jimeno, M. L.; Blanco, F.; Alkorta, I.; Elguero, J. Magn. Reson. Chem. 2008,
46, 859–864.
548 J. Org. Chem. Vol. 74, No. 2, 2009

Multiply Substituted Tetrahydropyran Oxocarbenium Ions
FIGURE 4. Theoretical and experimental J-values for C-3 alkoxy
dioxocarbenium ion (B3LYP/6-31G*).
FIGURE 5. Relative low-energy conformers of C-5 alkoxymethyl
dioxocarbenium ion (B3LYP/6-31G*).
as illustrated in cations 23. This exocyclic dihedral angle can
significantly influence the reactivity of carbohydrates.73
Computational studies on monosubstituted dioxocarbenium
ion 23 (X ) OEt) indicated that the conformational distribution
could not be explained by simply invoking steric effects.
Although the pseudoequatorial half-chair conformer 26 (which
74
4
is in the H3 orientation) would minimize steric interactions,
it was not the lowest energy structure (Figure 5). Despite
incurring two gauche interactions, conformers 24 and 25
remained approximately isoenergetic with conformer 26.75 The
presence of low-energy structures 24 and 25 could not be easily
explained by consideration of steric effects. The relative energies
of these structures, however, are consistent with the proposal
that electrostatic effects stabilize these cations.37,39-41 The
decrease in distance between the electronegative oxygen sub-
stituent at C-5 and the carbocationic carbon C-176 compensated
for the steric penalties associated with additional gauche
interactions.
FIGURE 6. Comparison of theoretical and experimental J-values of
C-5 alkoxymethyl dioxocarbenium ion.
The computational predictions illustrated in Figure 5 are
consistent with the NMR coupling constant data obtained for
dioxocarbenium ion 2377 (X ) OEt). The comparison between
experimentally determined and calculated coupling constants
for conformers 24-26 is illustrated in Figure 6. Because
deviations between predicted and experimental values are
reflected on the y axis, a curve which lies close to the x axis
indicates a good correlation between theoretical and experi-
mental values. The experimental J-values of conformer 24,
where the dioxocarbenium ion adopted a half-chair conformation
(4H3) with the electronegative 5-alkoxymethyl group oriented
toward the positive charge, agreed most closely with the
predicted coupling constants. The substantial deviation between
the experimentally determined JH5,H6′ value and that calculated
for cation 26 indicates that the alkoxymethyl substituent does
FIGURE 7. Theoretical and experimental J-values of C-5 alkoxymethyl
rotamers.
not adopt this orientation. This discrepancy can be rationalized
3
3
by considering the expected magnitude of JH5,H6 and JH5,H6′
based on the Karplus equation (Figure 7). Both gt and tg
rotamers would require one large and one small J-value, while
two small coupling constants would be expected for the gg
rotamer. The experimental vicinal coupling constants of 4.1 and
2.8 Hz are consistent with a gg conformation as found in both
24 (4H3) and 25 (1C4) (Figure 5). Because the small energy
differences between conformers 24-26 suggested a mixture of
conformers in solution, a weighted average of their populations
(based upon their calculated relative energies) was employed
to determine statistically relevant averaged coupling constants.
The resulting curve resembles the experimental values, sug-
gesting that all three conformers are present in solution.
The conclusion from spectroscopic evidence that a dioxo-
carbenium ion with an alkoxymethyl group at C-5 adopts a
pseudoequatorial conformation is consistent with reactivity of
(73) Jensen, H. H.; Nordstrøm, L. U.; Bols, M. J. Am. Chem. Soc. 2004,
126, 9205–9213.
(74) The IUPAC nomenclature (OS2, 3H4, BO,3, etc.) defined for all conformers
presented in this paper was assigned based on the method developed by Whitfield
et al.: Be´rces, A.; Whitfield, D. M.; Nukada, T. Tetrahedron 2001, 57, 477–
491. The endocyclic dihedral angles required for this analysis are supplied as
Supplementary Information.
(75) Conformers 24-26 are considered to be isoenergetic, because errors
associated with the B3LYP method can approach 1 kcal/mol: (a) Weldon, A. J.;
Vickrey, T. L.; Tschumper, G. S. J. Phys. Chem. A 2005, 109, 11073–11079.
(b) Woodcock, H. L.; Moran, D.; Pastor, R. W.; MacKerell, A. D., Jr.; Brooks,
B. R. Biophys. J. 2007, 93, 1–10.
(76) Dudley, T. J.; Smoliakova, I. P.; Hoffmann, M. R. J. Org. Chem. 1999,
64, 1247–1253.
(77) Repeated attempts, including synthesizing analogues with different
protecting groups (such as benzyl, 3-bromobenzyl, and tosyl), did not yield
suitable crystals for X-ray crystallography.
J. Org. Chem. Vol. 74, No. 2, 2009 549

Yang and Woerpel
the related oxocarbenium ion. Lewis acid-catalyzed nucleophilic
substitution of acetate 27 afforded the 1,5-trans product 28 with
high diastereoselectivity (eq 9).78 This result can be explained
by stereoelectronically controlled addition to the half-chair
conformer analogous to 24 (or even 26) in which the alkoxy-
methyl substituent adopts a pseudoequatorial orientation. Nu-
cleophilic addition to the oxocarbenium ion analogous to 25
might give similar selectivity because the alkoxymethyl sub-
stituent would hinder approach of the nucleophile to the top
face.
FIGURE 8. Relative low-energy conformers of the 4,5-disubstituted
dioxocarbenium ion (B3LYP/6-31G*).
3
agreed well for most of the values expected for a H4 half-
chair conformation (30, 32, and 3481), but they corresponded
poorly with the H3 half-chair conformers (31 and 33). The
The importance of the exocyclic electrostatic effect revealed
by the C-5 alkoxymethyl-substituted dioxocarbenium ion 23
(vide supra) was reinforced by results for the trans-4,5-
disubstituted dioxocarbenium ion 2979 (eq 10). Computational
4
diaxial orientation of the 4,5-disubstituted dioxocarbenium ion
was further substantiated by the observed 1.5 Hz W-coupling
between the hydrogen atom at C-5 and the equatorial hydrogen
atom at C-3. This coupling is only possible when the C-5
substituent is pseudoaxial. The two half-chair conformers 30
and 32 can be distinguished by examination of the JH5,H6′ region
of the graph. The preference for conformer 30, with gauche
orientations of the hydrogen at C-5 to both hydrogens at C-6,
indicates that the alkoxy group at C-6 must be positioned over
the ring of the cation, presumably to maximize electrostatic
stabilization.
studies revealed that the dioxocarbenium ion strongly favored
the diaxial half-chair conformer 30 (a 3H4 conformer) over other
possible conformers (Figure 8). The preference for this diaxial
conformation demonstrates that the powerful electrostatic effect
of the C-4 alkoxyl group can overwhelm the steric preference
of the C-5 substituent to reside pseudoequatorially (vide supra).
As with the C-5 alkoxymethyl-substituted dioxocarbenium ion
23, the exocyclic alkoxymethyl group of 29 assumes a gauche
orientation to maximize electrostatic interactions between the
C-6 alkoxy group and the cationic carbon atom (as depicted by
conformer 30). Consistent with this argument, conformers 31,
32, and 34, which are stabilized by only one electrostatic
interaction, are considerably higher in energy. Although both
conformers 30 and 34 have the proper alignment for a σ C-6-H
to σ* C-5-O+ donation,80 a 2.7 kcal/mol difference in their
energy suggests that the stabilization is electrostatic rather than
hyperconjugative in nature. Without electrostatic benefits, the
diequatorial conformer 33 (4H3) is predicted to be higher in
energy than the diaxial conformer 30 (3H4).
The discrepancy between the conformational preference of
the dioxocarbenium ion 29 and the selectivity of the reaction
of the analogous oxocarbenium ion derived from 35 illustrates
the importance of the Curtin-Hammett principle.49,50 Although
it might be expected that nucleophilic substitution of acetate
35 via its oxocarbenium ion would occur with ꢀ-selectivity
through a conformer resembling 30, R-selectivity was observed
in this reaction (eq 11). Because interconversion among the
different conformers is fast relative to the rate of nucleophilic
addition,40 the major product would arise from the conformer
with the lowest transition state.49,50 Reaction through the diaxial
half-chair conformers analogous to 30, 32, and 34 would involve
higher energy transition states due to developing 1,3-diaxial
interactions. Therefore, the product likely resulted from R-attack
4
of the H3 diequatorial conformers (31 or 33).
In accordance with the calculated energy differences of the
low-energy conformers, experimental coupling constants con-
firmed that the diaxial conformation 30 (in the box in Figure 8)
3
is the predominant species in solution. The fact that all JH,H
coupling constants could be obtained for this molecule permitted
a thorough analysis of its conformational preferences. As is
evident from Figure 9, the experimental coupling constants
Examination of the conformational bias of the dioxocarbe-
nium ion derived from 2-deoxymannolactone 1383,84 (eq 12)
revealed the limits of the electrostatic stabilization by an
exocyclic alkoxy group. The lowest energy conformer identified
(78) This result is inconsistent with the product ratio (70:30) we previously
reported for this reaction (ref 51). In that paper, a minor product was observed
in the unpurified reaction mixture, and it was assumed that the minor material
was a diastereomer. We have since prepared the cis-isomer of 28 and confirmed
that it is not formed in the reaction mixture. The details of these experiments
are provided as Supporting Information.
(79) Repeated attempts, including synthesizing analogues with different
protecting groups (such as benzyl, 3-bromobenzyl, and naphthyl), did not yield
suitable crystals for X-ray crystallography. The benzylated dioxocarbenium ion
29 is stable for at least one month at -25 °C.
3
by density functional methods (B3LYP/6-31G*) was the H4
3
conformer 38 (Figure 10). The H4 conformation is similar in
structure to the 3E conformation, a calculated minimum for the
mannosyl oxocarbenium ion.29 The conformer 38 experiences
electrostatic stabilization from both C-3 and C-4 alkoxy groups.
It does not, however, bring the alkoxy group at C-6 close to
550 J. Org. Chem. Vol. 74, No. 2, 2009

Multiply Substituted Tetrahydropyran Oxocarbenium Ions
FIGURE 9. Comparison of theoretical and experimental J-values for 4,5-disubstituted dioxocarbenium ion.82
the cationic center, presumably to avoid interactions with the
pseudoaxial C-3 substituent. This conformer would be desta-
bilized not only by syn-butanol interactions,85 but also by
bringing two oxygen atoms into proximity.86 In contrast,
although both 40 (the OS2 conformer) and 41 (5H4) place all
three substituents in positions to maximize their electrostatic
influences, they remain higher in energy than the half-chair
conformer 38. The conformation predicted based solely on steric
4
grounds, 42 (the half-chair conformer H3), was considerably
less stable.
FIGURE 10. Calculated low-energy conformers of the 2-deoxyman-
nosyl oxocarbenium ion (B3LYP/6-31G*).
Experimental coupling constants for the 2-deoxymannosyl
dioxocarbenium ion 37 are more consistent with a weighted
average of the low-energy conformers than they are with any
particular conformer (Figure 11). On the basis of the energy
differences between the low-energy conformers, 38 (3H4) should
dominate the solution-phase equilibrium with some contributions
from the twist conformers 39 and 40 (conformers 41 and 4287
would, combined, represent less than 3% of the mixture at 298
K). The experimental coupling constants, however, do not match
(80) We did not consider the σ C-5-H to σ* C-6-O interaction to be a
significant source of stabilization due to prior observations that methoxycyclo-
hexane prefers to adopt an equatorial conformation and 1,2-dimethoxyethane
does not adopt a gauche conformation: Lii, J.-H.; Chen, K.-H.; Grindley, T. B.;
Allinger, N. L. J. Comput. Chem. 2003, 24, 1490–1503.
(81) To simplify the graph, the higher energy conformer 34 was omitted
from Figure 9. It was plotted along with its C-5-C-6 rotamers (30 and 33) in
the Supporting Information.
FIGURE 11. Comparison of theoretical and experimental J-values of
the 2-deoxymannosyl oxocarbenium ion.
(82) Because one dominant low-energy conformer (30) exists for this
substrate, we felt that taking the absolute value of the difference between
experimental and theoretical J-values best illustrates the differences between
theoretical values and experimental data. Other means of analyzing the data,
such as considering the value of the difference or proportional errors, were
considered, and these plots lead to the same conclusion. These plots are provided
as Supporting Information.
(83) This lactone may be considered to be 2-deoxygluconolactone because
mannose and glucose differ only in the configuration at C-2.
(84) Although the trisubstituted dioxocarbenium ion 37 could be formed,
rapid decomposition within hours of product formation hindered our efforts to
obtain X-ray quality crystals.
well with any individual conformer. A weighted average of their
populations was again employed to determine statistically
relevant averaged coupling constants. The resulting curve more
closely resembles the experimental values, suggesting that
multiple conformers (in the box in Figure 10) are present in
solution.
If conformations resembling the half-chair conformer 38 (3H4)
and the twist conformers 39 and 40 were also responsible for
(85) Ohno, K.; Yoshida, H.; Watanabe, H.; Fujita, T.; Matsuura, H. J. Phys.
Chem. 1994, 98, 6924–6930.
(86) Law, R. V.; Sasanuma, Y. J. Chem. Soc., Faraday Trans. 1996, 92,
4885–4888.
(87) To simplify the graph, higher energy conformers 41 and 42 were omitted
from Figure 11. They were plotted along with conformer 38 in the Supporting
Information.
J. Org. Chem. Vol. 74, No. 2, 2009 551

Yang and Woerpel
ions,27,28 alkoxy groups at C-3 and C-456 exhibited marked
preferences to reside in axial orientations. Although the C-5
alkoxymethyl group preferred to reside pseudoequatorially, an
electrostatic attraction oriented the C-6 alkoxy group toward
the oxocarbenium ion carbon atom. This electrostatic effect was
most pronounced in the trans 4,5-disubstituted system, where
the powerful electrostatic effect exerted by the C-4 alkoxyl group
caused the C-5 alkoxymethyl group to adopt a diaxial confor-
mation with the C-6 alkoxy group pointed over the ring. The
trisubstituted system preferred to adopt a half-chair (3H4)
conformation with three axial substituents. These studies
provided additional insight that should facilitate understanding
reactions that involve carbohydrate-derived oxocarbenium ions.
FIGURE 12. Nucleophilic addition to oxocarbenium ions derived from
2-deoxymannolactone.
the reactivity of the related oxocarbenium ion, nucleophilic
addition to the oxocarbenium ion should be R-selective. Nu-
cleophilic attacks from the ꢀ-faces to conformers analogous to
38-40 should be disfavored over the corresponding attacks from
the R-faces, because they would lead to transition states with
1,3-diaxial interactions (Figure 12). Alternatively, if any of these
modes of nucleophilic attack were slow relative to stereoelec-
tronically controlled addition to the higher energy half-chair
conformer 42, R-selectivity would still be predicted based upon
the Curtin-Hammett principle.49,50 This interpretation was
supported by the nucleophilic substitution reaction of acetate
47, which yielded the R-product (eq 13).
Experimental Section
General experimental details and the syntheses of lactones and
acetates are provided as Supporting Information.
1-Ethoxy-3-benzyloxytetrahydropyrylium Hexachloroanti-
monate (21). To lactone 1736 (0.031 g, 0.15 mmol) in 0.7 mL of
CD2Cl2 was added triethyloxonium hexachloroantimonate (0.066
1
g, 0.152 mmol). The reaction mixture was cooled at 0 °C. A H
1
NMR spectrum was taken of the reaction mixture after 3 h: H
NMR (500 MHz, CD2Cl2) δ 7.41-7.29 (m, 5H), 5.38-5.30 (m,
2H), 4.92-4.84 (m, 2H), 4.64 (d, J ) 11.5 Hz, 1H), 4.59 (d, J )
11.6 Hz, 1H), 4.30-4.27 (m, 1H), 3.41 (dd, J ) 19.9, 4.3 Hz, 1H),
3.28 (ddt, J ) 19.8, 2.7, 1.4 Hz, 1H), 2.52-2.39 (m, 2H), 1.55 (t,
J ) 7.0 Hz, 3H). The 4.3 Hz J-value at δ 3.41 ppm and the 2.7 Hz
J-value at δ 3.28 ppm suggest an axial disposition of the C-3
substituent.
1-Ethoxy-5-(benzyloxymethyl)tetrahydropyrylium Hexachlo-
roantimonate (23). To lactone 1851 (0.053 g, 0.24 mmol) in 0.7
mL of CD2Cl2 was added triethyloxonium hexachloroantimonate
(0.111 g, 0.254 mmol). The reaction mixture was protected from
light at room temperature in a nitrogen atmosphere dry box. A 1H
The strategy employed to discern the conformational prefer-
ence of the less substituted oxocarbenium ions described above
was not suitable for studying oxocarbenium ions with alkoxy
groups at C-2. When the lactone precursors 10 and 11 were
subjected to alkylation conditions, the dioxocarbenium ion
products 48 and 49 were not observed (eq 14). Instead, the
reaction yielded a complex mixture of byproducts.61 It is
possible that the dioxocarbenium ions were formed, but the
neighboring electronegative C-2 substituent60 made these in-
termediates much more reactive than their deoxy analogues
(such as 29 and 37, vide supra).
1
NMR spectrum was taken of the reaction mixture after 3 h: H
NMR (500 MHz, CD2Cl2) δ 7.42-7.29 (m, 5H), 5.51 (dtd, J )
9.6, 4.1, 2.8 Hz, 1H), 4.91-4.74 (m, 2H), 4.60 (s, 2H), 4.00 (dd,
J ) 12.0, 2.8 Hz, 1H), 3.86 (dd, J ) 12.0, 4.2 Hz, 1H), 3.26-3.17
(m, 1H), 3.11-3.02 (m, 1H), 2.32-2.03 (m, 4H), 1.54 (t, J ) 7.1
Hz, 3H); 13C NMR (125 MHz, CD2Cl2) δ 189.9, 137.5, 129.1,
128.7, 128.5, 94.8, 75.0, 74.1, 69.9, 29.1, 23.6, 16.3, 13.9. The 4.1
and 2.8 Hz J-values at δ 5.51 ppm suggest a gg orientation of the
C-5 substituent.
1-Ethoxy-4,6-di-O-benzyl-2,3-dideoxy-D-glucopyryliumHexachlo-
roantimonate (29). To a solution of lactone 16 (0.068 g, 0.21
mmol) in CD2Cl2 (1.0 mL) was added triethyloxonium hexachlo-
roantimonate (0.092 g, 0.21 mmol). The reaction mixture was
protected from light in ambient temperature in a nitrogen atmo-
1
sphere dry box. A H NMR spectrum was taken of the reaction
1
mixture after 6 h: H NMR (500 MHz, CD2Cl2) δ 7.44-7.27 (m,
10H), 5.54 (dtd, J ) 3.9, 2.5, 1.5 Hz, 1H), 4.89-4.75 (m, 2H),
4.67 (d, J ) 11.6 Hz, 1H), 4.62 (d, J ) 11.6 Hz, 1H), 4.60 (d, J )
11.8 Hz, 1H), 4.54 (d, J ) 11.7 Hz, 1H), 4.15 (dt, J ) 5.0, 2.7 Hz,
1H), 4.02 (dd, J ) 11.9, 3.9 Hz, 1H), 3.90 (dd, J ) 11.8, 2.6 Hz,
1H), 3.18 (ddd, J ) 20.5, 9.7, 7.4 Hz, 1H), 3.12 (ddd, J ) 20.5,
6.9, 3.4 Hz, 1H), 2.44 (dddd, J ) 14.4, 9.8, 7.0, 2.9 Hz, 1H),
2.36-2.27 (m, 1H), 1.53 (t, J ) 7.1 Hz, 3H); 13C NMR (125 MHz,
CD2Cl2) δ 189.9, 137.0, 136.9, 129.2, 129.1, 128.9, 128.8, 128.7,
128.6, 93.7, 75.5, 74.3, 71.9, 69.8, 67.3, 25.9, 20.9, 13.9. The 3.9,
2.5, and 1.5 Hz J-values at δ 5.54 ppm suggest both axial
dispositions of the C-4 and C-5 substituents and the gg orientation
of the C-5 substituent.
Conclusion
Insight into the conformational preferences of highly substi-
tuted tetrahydropyran oxocarbenium ions was obtained by
comparing spectroscopic and computational studies of related
dioxocarbenium ions with reactivity data of oxocarbenium ions.
Experimentally determined coupling constants for dioxocar-
benium ions agreed closely with those predicted by computa-
tional studies, particularly if relative populations of conformers
were considered. These studies revealed strong conformational
preferences. As demonstrated by reactions of oxocarbenium
1-Ethoxy-3,4,6-tri-O-benzyl-2-deoxy-D-glucopyrylium Hexachlo-
roantimonate (37). To a solution of lactone 13 (0.042 g, 0.097
mmol) in CD2Cl2 (0.7 mL) was added triethyloxonium hexachlo-
roantimonate (0.049 g, 0.112 mmol). The reaction mixture was
552 J. Org. Chem. Vol. 74, No. 2, 2009

Multiply Substituted Tetrahydropyran Oxocarbenium Ions
protected from light in ambient temperature in a nitrogen atmo-
Purification by flash column chromatography on silica gel (1:9 Et2O:
1
sphere dry box. A H NMR spectrum was taken of the reaction
hexanes) afforded 1,5-trans isomer 36 (0.142 g, 86%) and 1,5-cis
1
1
mixture after 2 h: H NMR (500 MHz, CD2Cl2) δ 7.41-7.23 (m,
isomer (0.007 g, 4%) as colorless oils. Major Isomer (36): H
15 H), 5.36 (dt, J ) 5.7, 3.9 Hz, 1H), 4.96-4.84 (m, 2H),
4.68-4.46 (m, 6H), 4.23 (q, J ) 3.7 Hz, 1H), 4.12 (td, J ) 3.9,
0.9 Hz, 1H), 3.92 (dd, J ) 11.7, 5.8 Hz, 1H), 3.88 (dd, J ) 11.6,
4.0 Hz, 1H), 3.30 (dd, J ) 18.3, 4.0 Hz, 1H), 1.54 (t, J ) 7.1 Hz,
3H). The 3.7 Hz J-value at δ 4.23 ppm and the 3.9 Hz J-value at
δ 5.36 ppm suggest axial dispositions of all three substituents.
General Procedure for the Nucleophilic Substitution of
Acetate Substrates. Allyltrimethylsilane (4.0 equiv) was added to
a solution of acetate (0.10 M) in CH2Cl2, and the mixture was
cooled to -78 °C and treated with the Lewis acid (4.0 equiv). After
5 min, the mixture was allowed to warm to -45 °C. After 4 h, the
reaction mixture was warmed to room temperature and treated with
saturated aqueous NaHCO3 (30 mL per mmol of acetate). The layers
were separated, and the aqueous layer was extracted three times
with CH2Cl2 (10 mL per mmol of acetate), dried over MgSO4,
filtered, and concentrated in vacuo.
NMR (500 MHz, CDCl3) δ 7.36-7.29 (m, 10H), 5.80 (ddt, J )
17.1, 10.3, 7.1 Hz, 1H), 5.08 (dt, J ) 1.8, 1.5 Hz, 1H), 5.04 (dt, J
) 2.2, 1.1 Hz, 1H), 4.59 (d, J ) 12.1 Hz, 1H), 4.57 (d, J ) 12.2
Hz, 1H), 4.52, (d, J ) 12.2 Hz, 1H), 4.45 (d, J ) 11.9 Hz, 1H),
3.88-3.82 (m, 1H), 3.80 (ddd, J ) 6.7, 4.6, 4.2 Hz, 1H), 3.69 (dd,
J ) 10.3, 5.0 Hz, 1H), 3.62 (dd, J ) 10.3, 3.8 Hz, 1H), 3.50-3.45
(m, 1H), 2.54-2.46 (m, 1H), 2.28-2.21 (m, 1H), 1.97-1.89 (m,
1H), 1.77-1.63 (m, 3H); 13C NMR (125 MHz, CDCl3) δ 138.7,
138.5, 135.2, 128.44, 128.41, 127.9, 127.8, 127.64, 127.63, 116.8,
73.4, 73.2, 72.9, 71.8, 70.8, 69.6, 36.7, 26.2, 24.3; IR (neat) 3064,
3030, 2938, 2862, 1951, 1871, 1812, 1641, 1496, 1454, 1364, 1101,
913, 736, 698 cm-1; HRMS (ESI) m/z calcd for C23H28O3Na (M
+ Na)+ 375.1936, found 375.1947. Anal. Calcd for C23H28O3: C,
78.38; H, 8.01. Found: C, 78.25; H, 8.13. [R]24 +57.5 (c 1.17,
D
CH2Cl2). Minor Isomer: 1H NMR (500 MHz, CDCl3) δ 7.37-7.22
(m, 10H), 5.84 (dddd, J ) 17.0, 10.2, 7.4, 6.9 Hz, 1H), 5.11-5.01
(m, 2H), 4.64 (d, J ) 12.3 Hz, 1H), 4.61-4.55 (m, 2H), 4.40 (d,
J ) 11.5 Hz, 1H), 3.77 (dd, J ) 10.5, 1.6 Hz, 1H), 3.67 (dd, J )
10.8, 4.8 Hz, 1H), 3.42-3.36 (m, 3H), 2.43-2.35 (m, 1H),
2.30-2.15 (m, 2H), 1.72-1.97 (m, 1H), 1.50-1.27 (m, 2H).
1-Allyl-3,4,6-tri-O-benzyl-2-deoxy-D-glucopyranose (47). Un-
der standard allylation conditions with BF3 ·OEt2 as the Lewis acid,
acetate 46 (0.202 g, 0.425 mmol) afforded a 92:8 R:ꢀ mixture of
anomers.89 Purification by flash column chromatography on silica
gel (1:4 Et2O:hexanes) provided the R-isomer 47 as a colorless oil
(0.133 g, 68%). The spectral data correlate with the previously
reported data for 47:88 1H NMR (500 MHz, CDCl3) δ 7.33-7.19
(m, 15H), 5.07-5.03 (m, 1H), 5.03-5.01 (m, 2H), 4.78 (d, J )
11.3 Hz, 1H), 4.62-4.48 (m, 5H), 4.03 (tt, J ) 7.5, 4.4 Hz, 1H),
3.82-3.73 (m, 3H), 3.69-3.63 (m, 1H), 3.54 (t, J ) 7.1 Hz, 1H),
2.44 (dddt, J ) 14.3, 7.8, 6.7, 1.3 Hz, 1H), 2.21 (dtt, J ) 14.2, 7.2,
0.9 Hz, 1H), 1.98 (dt, J ) 13.4, 4.3 Hz, 1H), 1.76 (ddd, J ) 13.6,
9.4, 5.0 Hz, 1H); 13C NMR (125 MHz, CDCl3) δ 138.6, 138.5,
138.4, 134.8, 128.52, 128.48, 128.4, 128.1, 128.0, 127.77, 127.75,
127.74, 127.67, 117.1, 77.0, 76.5, 74.2, 73.5, 72.9, 71.4, 70.6, 69.1,
36.8, 32.5; HRMS (ESI) m/z calcd for C30H34O4Na (M + Na)+
For allylation of 5-substituted acetate 27, the unpurified product
1
ratios were determined by using GCMS and confirmed with H
NMR spectroscopy. For allylation of the disubstituted acetate 35,
the product ratios were based on isolated yield of the minor isomer
and confirmed by GCMS and 1H NMR spectroscopy of the
unpurified material. For allylation of trisubstituted acetate 46, the
1
product ratios were based on H NMR spectroscopy of the un-
purified material. The reported yields are of purified materials. For
allylated products 28 and 36, the relative configurations were
determined by nuclear Overhauser effect (NOE) enhancements. For
compound 47, the relative configurations were determined by
correlation to previously published spectral data.88
1-Allyl-5-benzyloxymethyltetrahydropyran (28). Under stan-
dard allylation conditions with BF3 ·OEt2 as the Lewis acid, acetate
2751 (0.144 g, 0.546 mmol) afforded 28 (0.123 g, 92%) as a >97:3
1,5-trans:cis mixture of isomers. The absence of the minor cis
isomer was confirmed by the synthesis of authentic minor material
(which is described in the Supporting Information). The oil was
purified by flash chromatography (1:4 Et2O:pentane) to afford the
1,5-trans product 28 as a colorless oil: 1H NMR (500 MHz, CDCl3)
δ 7.34-7.25 (m, 5H), 5.81 (ddt, J ) 17.0, 10.3, 7.0 Hz, 1H), 5.09
(dq, J ) 17.2, 1.6 Hz, 1H), 5.04 (dt, J ) 2.0, 1.0 Hz, 1H), 4.58 (d,
J ) 12.0 Hz, 1H), 4.54 (d, J ) 12.0 Hz, 1H), 3.98-3.92 (m, 1H),
3.84-3.78 (m, 1H), 3.58 (dd, J ) 9.9, 6.4 Hz, 1H), 3.44 (dd, J )
10.0, 5.3 Hz, 1H), 2.46 (qnt, J ) 7.0, 1.3 Hz, 1H), 2.23 (qnt, J )
7.1, 1.1 Hz, 1H), 1.70-1.55 (m, 1H), 1.50-1.34 (m, 2H),
1.73-1.54 (m, 4H), 1.50-1.34 (m, 2H); 13C NMR (125 MHz,
CDCl3) δ 138.5, 135.4, 128.4, 127.7, 127.6, 116.7, 73.2, 71.64,
71.59, 70.0, 37.7, 28.9, 27.1, 18.4; IR (neat) 3066, 3030, 2935,
2864, 1641, 1454 cm-1; HRMS (ESI) m/z calcd for C16H22O2Na
(M + Na)+ 269.1518, found 269.1520. Anal. Calcd for C16H22O2:
C, 78.01; H, 9.00. Found: C, 77.75; H, 8.92.
481.2355, found 481.2358. [R]24 +32.2 (c 1.29, CH2Cl2).
D
Acknowledgment. This research was supported by the
National Institutes of Health, National Institute of General
Medical Sciences (GM-61066). K.A.W. thanks Amgen and Lilly
for generous support for research. We thank Dr. John Greaves
and Dr. Shirin Sooroshian (UCI) for mass spectrometric data,
Dr. Phil Dennison (UCI) for assistance with NMR studies, Dr.
Nathan Crawford (UCI) for computational support, and Dr.
Joseph Ziller (UCI) for assistance with X-ray crystallography.
1-Allyl-5-benzyloxymethyl-4-benzyloxytetrahydropyran (36).
Under standard allylation conditions with BF3 ·OEt2 as the Lewis
acid, acetate 35 (0.205 g, 0.553 mmol) afforded 36 as a 96:4 1,5-
trans:cis mixture of isomers. The product distribution was deter-
mined based on isolated yield of the minor isomer and confirmed
with GCMS and 1H NMR spectroscopy of the unpurified material.
Supporting Information Available: Complete refs 16, 64
and 65, experimental procedures, product characterization,
stereochemical proofs, details of computations, and GC and
spectral data for selected compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
JO8017846
(88) Characterization data for the ꢀ-isomer: Crich, D.; Lim, L. B. L. J. Chem.
Soc., Perkin Trans. 1 1991, 2205–2208. Characterization data for the R-isomer:
Wang, Z.; Shao, H.; Lacroix, E.; Wu, S.-H.; Jennings, H. J.; Zou, W. J. Org.
Chem. 2003, 68, 8097–8105.
(89) The minor isomer was identified in the unpurified spectrum by
comparison to previously reported 1H NMR spectrum for the ꢀ-isomer.
J. Org. Chem. Vol. 74, No. 2, 2009 553