Stereoselectivity in the epoxidation of carbohydrate-based oxepines

Article abstract of DOI:10.1021/jo800979a

(Figure Presented) The facial selectivity in the DMDO epoxidation of carbohydrate-based oxepines derived from glucose, galactose, and mannose has been determined by product analysis and density functional theory (DFT, B3LYP/6-31+G**//B3LYP/6-31G*) calcula

Full text of DOI:10.1021/jo800979a

Stereoselectivity in the Epoxidation of Carbohydrate-Based
Oxepines
Shankar D. Markad,† Shijing Xia,‡ Nicole L. Snyder,§,† Bikash Surana,†
Martha D. Morton,† Christopher M. Hadad,*,‡ and Mark W. Peczuh*,†
Department of Chemistry, The UniVersity of Connecticut, 55 North EagleVille Road, Storrs, Connecticut
06269, and Department of Chemistry, The Ohio State UniVersity, 100 West 18th AVenue,
Columbus, Ohio 43210
hadad.1@osu.edu; mark.peczuh@uconn.edu
ReceiVed May 12, 2008
The facial selectivity in the DMDO epoxidation of carbohydrate-based oxepines derived from glucose,
galactose, and mannose has been determined by product analysis and density functional theory (DFT,
B3LYP/6-31+G**//B3LYP/6-31G*) calculations. Oxepines 3 and 4, derived from D-galactose and
D-mannose, largely favor R- over ꢀ-epoxidation. The results reported here, along with selectivities in the
DMDO-mediated epoxidation of D-xylose-based oxepine 1 and D-glucose-based oxepines 2 and 5 reported
earlier, support a model in which electronic effects, guided by the stereochemistry of the oxygens on the
oxepine ring, largely determine the stereoselectivity of epoxidation. Other contributing factors included
conformational issues in the oxepine’s transition state relative to the reactant, the asynchronicity in bond
formation of the epoxide, and the overall steric bulk on the R- and ꢀ-faces of the oxepine. Considered
together, these factors should generally predict facial selectivity in the DMDO-epoxidation of cyclic
enol ethers.
1. Introduction
stereochemistry of the cyclic enol ether, high selectivity in the
formation of one anhydrosugar in preference to another is
common.2–5
Our approach to the formation of glycosidic bonds in the
septanose series has similarly relied on 1,2-anhydroseptanoses
as a donor both directly4,6,7 or indirectly through the formation
of a thiophenyl septanoside.8 Carbohydrate-based oxepines9–11
have proven to be useful starting materials for preparing 1,2-
anhydroseptanoses, with the oxepines acting as ring-expanded
glycals. Selectivity observed in the epoxidation of carbohydrate-
The formation of glycosidic linkages with use of 1,2-
anhydrosugars as donors has enjoyed wide application in the
syntheses of oligosaccharides and natural products.1 Access to
these donors from cyclic enol ether (glycals, for example) via
epoxidation has been instrumental to their utilization. Because
the cyclic enol ethers are often chiral, their reactions with achiral
epoxidizing reagents can follow distinct stereochemical path-
ways which need not be equivalent. In fact, based on the
† The University of Connecticut.
(3) (a) Majumder, U.; Cox, J. M.; Johnson, H. W. B.; Rainier, J. D. Chem.
Eur. J. 2006, 12, 1736. (b) Kim, Y. J.; Wang, P. F.; Navarro-Villalobos, M.;
Rohde, B. D.; Derryberry, J.; Gin, D. Y. J. Am. Chem. Soc. 2006, 128, 11906.
(c) Rainier, J. D.; Allwein, S. P.; Cox, J. M. J. Org. Chem. 2001, 66, 1380.
(4) Peczuh, M. W.; Snyder, N. L.; Fyvie, W. S. Carbohydr. Res. 2004, 339,
1163.
‡ The Ohio State University.
§ Current address: Department of Chemistry, Hamilton College, Clinton, NY
13323.
(1) (a) Seeberger, P. H.; Bilodeau, M. T.; Danishefsky, S. J. Aldrichim. Acta
1997, 30, 75. (b) Danishefsky, S. J.; Bilodeau, M. T. Angew. Chem., Int. Ed.
Engl. 1996, 35, 1380.
(5) Honda, E.; Gin, D. Y. J. Am. Chem. Soc. 2002, 124, 7343.
(6) DeMatteo, M.; Snyder, N. L.; Morton, M.; Baldisseri, D. M.; Hadad,
C. M.; Peczuh, M. W. J. Org. Chem. 2005, 70, 24.
(2) (a) Boulineau, F. P.; Wei, A. J. Org. Chem. 2004, 69, 3391. (b) Boulineau,
F. P.; Wei, A. Org. Lett. 2002, 4, 2281. (c) Gervay, J.; Peterson, J. M.; Oriyama,
T.; Danishefsky, S. J. J. Org. Chem. 1993, 58, 5465. (d) Marzabadi, C. H.;
Spilling, C. D. J. Org. Chem. 1993, 58, 3761.
(7) DeMatteo, M.; Mei, S.; Fenton, R.; Morton, M.; Baldisseri, D. M.; Hadad,
C. M.; Peczuh, M. W. Carbohydr. Res. 2006, 341, 2927.
10.1021/jo800979a CCC: $40.75
Published on Web 07/24/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 6341–6354 6341

Markad et al.
SCHEME 1
SCHEME 2
based oxepines has been variable and was related to the specific
starting material. For example, oxepine 1 was shown to
selectively form the ꢀ-1,2-anhydroseptanose 6 when using
dimethyldioxirane (DMDO) as the oxidant (Scheme 1). Selec-
tivity in this system was determined by 1H and 13C NMR
spectroscopy and by analysis of the methyl septanoside product
that resulted from methanolysis of the 1,2-anhydroseptanose.4
DMDO epoxidation of oxepine 2, on the other hand, proved to
be relatively unselective (7:8 ) 3:1, Scheme 1). In this case,
the ratio of epoxides formed was quantified by characterizing
the products of methanolysis.6,7 Combined, these experiences
prompted us to ask the question: What factors determine the
selectivity in the epoxidation of carbohydrate-based oxepines?
Herein, we analyze the outcomes of the DMDO oxidation of
three related oxepines (2-4) by product analysis and density
functional theory (DFT) calculations on the transition states in
these reactions. Oxepines 2, 3, and 4 are derived from glucose,
galactose, and mannose, respectively. The variation of stereo-
chemistry about the ring allowed for an evaluation of the
parameters that control the stereochemistry of epoxidation.
the systems they evaluated.14 We were interested in evaluating
these models in terms of the DMDO epoxidations of carbohy-
drate-based oxepines 1-5.
Recent reports on the epoxidation stereoselectivity of cyclic
enol ethers from the Wei12 and Rainier13 laboratories influenced
our perspective. Wei described a polarized-π frontier molecular
orbital (PPFMO) model that explained selectivity in DMDO
epoxidation of a series of 4-deoxypentenosides.12 In their
system, the reactant geometry was similar to that of the transition
state, and C-O bond formation was asynchronous in the
transition state. Repulsive electronic interactions between the
heteroatoms of the ring substituents and the π-bond of the cyclic
enol ether polarized the π-bond so that the nucleophilic π-bond
had greater electron density on the face with fewer electroneg-
ative substituents. This gave rise to the idea of the facial
selectivity being due to a “majority rule”. The Rainier model
similarly emphasized transition state geometry and asynchronous
bond formation during the epoxidation of six- and seven-
membered-ring cyclic enol ethers. Avoiding unfavorable inter-
actions between DMDO and the allylic C-H bond in the
transition state formed the basis for epoxidation selectivity in
2. Results and Discussion
2.1. DMDO Epoxidations of Carbohydrate-Based Oxepines.
Oxepines 1-5 served as the starting materials for the DMDO
epoxidation reactions. We have previously reported on the
details for the synthesis of 1, 2, and 5 by a ring-closing
methathesis (RCM)-based strategy.4,6,11 Two new oxepines, 3
and 4, were prepared by the same route but with a minor
modification. Scheme 2 shows the synthesis of 3 and 4 from
tetra-O-benzyl-D-galacto-pyranose 15 and tetra-O-benzyl-D-
manno-pyranose 16. Wittig methylenation15 of the lactols gave
δ-hydroxy alkenes 17 and 18 in 78% and 30% yields, respec-
tively.16 The change in synthetic sequence relative to our
previous route came in the formation of vinyl ethers 19 and
20. Vinyl etherification of the hydroxyl group on 17 was carried
out with Pd(OAc)2 and ethyl vinyl ether17 to give diene 19 with
(14) Ando, K.; Green, N. S.; Li, Y.; Houk, K. N. J. Am. Chem. Soc. 1999,
121, 5334.
(8) Castro, S.; Fyvie, W. S.; Peczuh, M. W. Org. Lett. 2005, 7, 4709.
(9) Castro, S.; Peczuh, M. W. J. Org. Chem. 2005, 70, 3312.
(10) Alca´zar, E.; Pletcher, J. M.; McDonald, F. E. Org. Lett. 2004, 6, 3877.
(11) Peczuh, M. W.; Snyder, N. L. Tetrahedron Lett. 2003, 44, 4057.
(12) Cheng, G.; Boulineau, F. P.; Liew, S.-T.; Shi, Q.; Wenthold, P. G.;
Wei, A. Org. Lett. 2006, 8, 4545.
(15) Martin, O. R.; Saavedra, O. M.; Xie, F.; Liu, L.; Picasso, S.; Vogel, P.;
Kizu, H.; Asano, N. Bioorg. Med. Chem. 2001, 9, 1269.
(16) Attempts to optimize the olefination of tetra-O-benzyl-D-manno-pyranose
were unsuccessful. Elimination of benzyloxy substituents may be a competing
reaction. See, for example: Callam, C. S.; Lowary, T. L. J. Org. Chem. 2001,
66, 8961.
(13) Orendt, A. M.; Roberts, S. W.; Rainier, J. D. J. Org. Chem. 2006, 71,
5565.
(17) Handerson, S.; Schlaf, M. Org. Lett. 2002, 4, 407.
6342 J. Org. Chem. Vol. 73, No. 16, 2008

Epoxidation of Carbohydrate-Based Oxepines
TABLE 1. Experimental Epoxidation Selectivities and Yields for
1-5
SCHEME 3
epoxides
(R:ꢀ)
products combined
C1 13C δ
(ppm)
oxepine
(R:ꢀ)
yield (%)
1
2
3
4
5
6 (1:25)
7:8 (3:1)
9:10 (25:1)
11:12 (>25:1) 26:27 (7:3)
13:14 (ND) 24
21
89a
63b
67
106.2 (21b)
106.2 (22b); 110.0 (23b)
110.1 (25b)
100.7 (26b)
112.2 (24b)
22:23 (1:3)
25
84
61c
a These data are taken from ref 4. b These data are taken from ref 6.
c These data are taken from ref 20.
2.2. Stereochemical Characterization of the Reaction
Products. Assignment of accurate structures to the products
previously obtained in the epoxidation/methanolysis sequence
was critical to understanding the stereoselectivity of the epoxi-
dation reaction. The C2 group consistently reports on this
selectivity and is insensitive to any subsequent epimerization/
anomerization. The stereochemistry of 21 was assigned by
comparison to a known compound from Stevens.21 Assignment
of structures 22 and 23 was concurrent with a conformational
analysis of the completely deprotected methyl septanosides
corresponding to them.6 The conformational analysis was in part
used to define the stereochemistry in these two structures.
Evidence supporting our structural assignment of the new
products has come from X-ray crystallography on methyl 3,4:
5,7-di-O-isopropylidene ꢀ-D-glycero-D-galacto-septanoside 29
(Figure 1). This molecule was prepared from the di-O-
isopropylidene protected oxepine 28 via the DMDO/metha-
nolysis sequence in 84% yield (Scheme 4). Deprotection of the
acetonide groups in 29 provided methyl ꢀ-D-glycero-D-galacto-
septanoside 29b in 93% yield. Hydrogenolysis of 25 and 26
provided methyl ꢀ-D-glycero-L-manno-septanoside 25b and
methyl R-D-glycero-D-galacto-septanoside 26b in 95% and 94%
yield (Scheme 3). In addition to the protected septanosides 25,
26, and 29, these deprotected methyl septanosides were utilized
in the overall assignment of stereochemistry for the new
molecules reported here.
very poor conversion even after refluxing the reaction mixture
for several days. To improve the yield in the vinyl etherification
reaction, an alternative method with an iridium catalyst as
reported by Okimoto et al.18 was used. Thus, treatment of
δ-hydroxy alkene 17 with 10 mol % of [Ir(COD)Cl]2 and vinyl
acetate gave diene 19 in 93% yield. The mannose-derived diene
20 was prepared in 74% yield from 18 with use of the same
reaction conditions. RCM, using 25 mol % of Schrock catalyst,19
with 19 and 20 gave oxepines 3 and 4 in 89% and 84% yield,
respectively.
As had previously been done for oxepines 1, 2, and 5, the
products from basic methanolysis of anhydroseptanoses derived
from 3, 4, and 5 were analyzed to determine the selectivity of
the DMDO epoxidation reaction. For example, methyl R-D-ido-
septanoside 21 was the product of the epoxidation/methanolysis
sequence for 1, indicating that epoxidation occurred on the
ꢀ-face of the oxepine. Under the same reaction conditions,
oxepine 2 gave rise to methyl R-D-glycero-D-ido-septanoside
22 and methyl ꢀ-D-glycero-D-gulo-septanoside 23 (1:3 R:ꢀ)
indicating only a slight preference for R-epoxide formation. C3
deoxygenated oxepine 5 provided ꢀ-septanoside 2420 through
R-1,2-anhydroseptanose 13, even in the absence of an allylic
substituent. Subjecting galactose-derived oxepine 3 to the
standard reaction conditions provided only the methyl ꢀ-sep-
tanoside product 25 in 67% overall yield (Scheme 3). Oxepine
4 showed more complex reactivity, giving methyl R-septanoside
26 and the cyclized product 27 in a combined yield of 84%.
On the basis of these product structures, epoxidation of 3 and
4 occurred with high selectivity for the R-face (>25:1) of these
oxepines. A complete rationale for the product structures in
Scheme 3 is provided below. Results for all of the experimental
epoxidation/methanolysis reactions are given in Table 1.
In general, we have observed a partitioning of C1 13C
chemical shifts in methyl septanosides based on the stereo-
chemistry at the anomeric position, especially for completely
deprotected methyl septanosides (Table 1).7,22,23 The C1 13C
chemical shifts for methyl R-septanosides range from 100 to
106 ppm whereas 13C chemical shifts for methyl ꢀ-septanosides
range from 110 to 112 ppm. On the basis of this trend, 25b and
29b fit into the range for ꢀ-septanosides with 13C δ for C1 of
110.1 and 110.8 ppm, respectively. The NMR analysis was in
accordance with the X-ray structure in the case of 29/29b.
Septanoside 26b, on the other hand, gave a C1 δ of 100.7 ppm,
indicating that its anomeric stereochemistry was R. With
confidence in the C1 stereochemical assignment, we used an
3
analysis of the J coupling constants to assign the remaining
C2 stereochemistry in the product structures.
(18) Okimoto, Y.; Sakaguchi, S.; Ishii, Y. J. Am. Chem. Soc. 2002, 124,
1590.
(21) Tran, T. Q.; Stevens, J. D. Aust. J. Chem. 2002, 55, 171.
(22) We have also reported the crystal structure of an R-septanoside. See
(19) Schrock, R. R. Tetrahedron 1999, 55, 8141.
(20) Castro, S.; Duff, M.; Snyder, N.; Morton, M.; Kumar, C. V.; Peczuh,
M. W. Org. Biomol. Chem 2005, 3, 3869.
ref 9.
(23) See the Supporting Information for a complete table of C1 chemical
shifts.
J. Org. Chem. Vol. 73, No. 16, 2008 6343

Markad et al.
anhydroseptanose 11, three that invoke the anhydrosugar directly
and two that proceed via cyclic oxonium ion 33. The paths are
the following: (a) attack of the C3 OBn oxygen to form a
C2-C3 epoxide 30; (b) attack of the C4 OBn oxygen atom to
form the bicyclic product 31; (c) attack of the C7 OBn oxygen
atom to form the bicyclic product 32; (d) attack by methoxide/
methanol on the R-face of oxonium ion 33 to form methyl
R-septanoside 26; and (e) attack of the C5 OBn oxygen to the
R-face of 33 to form 27.
Product analysis showed pathway d to be the preferred course
1
of the reaction, giving 26 in 55% yield. Both the H and 13C
NMR spectra of the other product, 27 (25%), clearly showed
that it was not a methyl septanoside. Notable differences were
the lack of an indicative methyl glycoside singlet and also the
absence of signals for one of the benzyl groups present in 4
and 11. Further, the 1H NMR and 13C NMR did not show signals
in the epoxide region, ruling out 30 as the product. HMBC NMR
of 27a showed correlation of H1 to carbons C3, C5, and C6
(all are three-bond couplings), and similarly, C1 correlated to
H5 and H3. Also, C4 and C7 both show correlations with benzyl
protons in the HMBC while C5 does not. This suggests that
the benzyloxy substituent at C5 was the nucleophile, and not
the corresponding groups at either C4 or C7. In total, correlations
are consistent with 27, but inconsistent with 31 or 32.26
Structure 27 arises from attack of the C5 OBn oxygen onto
the C1 of oxonium ion 33. We have reported on similar
reactivity of a C5 OBn oxygen in the NIS-mediated intramo-
lecular cyclization of oxepine 1. The bicyclic structure formed
in that reaction (34) is reminiscent of 27.25 The nucleophilicity
of the C5 OBn group likely relies on the electronic effects of
the group, the stability of the corresponding benzyl/tropylium
cation released in the reaction, and the proximity of the reacting
groups. Both of the observed products in the epoxidation/
methanolysis of 4 (26 and 27) arise from a presumed cyclic
oxonium ion intermediate 33. This is in contrast to the other
products of methoxide attack on 1,2-anhydroseptanoses, where
the products have arisen from SN2 ring-opening of the epoxide.
Although invoking oxonium ion 33 is speculative considering
the reaction conditions of the methanolysis (NaOCH3/HOCH3),
it is consistent with both products that have been characterized.27
FIGURE 1. ORTEP representation of 29 from X-ray data.
SCHEME 4
In addition to the X-ray and 13C data, analysis of COSY/
NOESY spectra of 25b, 26b, and 29b, along with their
AMBER-minimized structures, further supports the stereochem-
ical assignment.24 Diagnostic NOESY cross-peaks in the
spectrum of 25b are the H1-H4 and H1-H6 cross-peaks. These
two transannular cross-peaks are estimated to be 2.4 and 2.1 Å
in the calculated structure. For 26b, the COSY spectrum shows
only weak coupling between H1-H2 (3JH1,H2 ) 3 Hz from 1H
NMR). The NOESY spectrum, however, clearly shows an
H1-H2 cross-peak. The calculated structure for 26b puts the
H1-H2 protons approximately 2.3 Å apart and on the same
side/face of the ring. The magnitude of the H1-H2 cross-peak
is similar to that of the H1-aglycon methyl (-OCH3) NOE,
which is also calculated to be in close proximity (2.3 Å). Other
important NOEs observed for 26b are H2 to H3 and H4 to H6.
Calculated distances between these protons were 2.3 and 2.6
Å. Combined, these data strongly suggest the structure of 26b
to the 1,2-cis structure as shown. 1H NMR signals for 29b were
less disperse than those for 25b, but assignments could still be
made. The NOESY spectrum of 29b showed cross-peaks
between H1 and the aglycon methyl (-OCH3), H1 and H6, and
H4 and H6. Interproton distances from the AMBER-minimized
structure for 29b again agree with these observations (2.4-2.6
Å). Overall, the collected crystallographic and spectroscopic data
render a consistent picture for the structures of methyl septano-
sides 25, 26, and 29.
2.3. Transition State Calculations for DMDO Epoxida-
tions. The product analyses above established the selectivity in
epoxide formation for oxepines 1-5 and 28. We were interested
in developing a model for predicting the stereoselectivity of
epoxidation in these systems and turned to computational
chemistry for insight.
In this computational study, the selectivity of the DMDO-
mediated epoxidations of the R- and ꢀ-faces of the oxepines
was investigated at the B3LYP/6-31+G**//B3LYP/6-31G* level
of theory.28–31 Solvation effects were investigated by using the
We arrived at the structure for 27 by consideration of the
reaction mechanism in combination with NMR experiments. Our
previous observation of the intramolecular cyclization of an
activated oxepine via its C5 benzyloxy group gave us additional
perspective.25 By considering the potential reaction mechanisms
(Scheme 5), there could be five possible reaction paths for
(26) Full NMR spectral characterization (1H, 13C, COSY, HMQC, HMBC)
for 27a can be found in the Supporting Information.
(27) An investigation aimed at understanding the facility in forming oxonium
33 from 4 is currently underway.
(28) Parr, R. G.; Yang, W. Density Functional Theory in Atoms and
Molecules; Oxford University Press: New York, 1989.
(24) Spectra (COSY/NOESY), tabulated interproton distances, and the
AMBER minimized structures for 25b, 26b, and 29b are in the Supporting
Information.
(25) Fyvie, W. S.; Morton, M.; Peczuh, M. W. Carbohydr. Res. 2004, 339,
2363.
6344 J. Org. Chem. Vol. 73, No. 16, 2008

Epoxidation of Carbohydrate-Based Oxepines
SCHEME 5
polarizable continuum model (PCM),32–35 using single-point
energy (B3LYP/6-31+G**) calculations with the gas-phase
(B3LYP/6-31G*) optimized geometries. (Note: For the subse-
quent discussion, we will refer to the B3LYP/6-31+G**//B3LYP/
6-31G* leVel simply as B3LYP, unless otherwise noted.) Despite
changes in the relative ordering of different conformers or
transition states, the general trends between the free energies
and enthalpies are relatively consistentsall of the energies are
provided in the Supporting Information. In the subsequent
discussion, the free energies will be presented.
To simplify the computational study, the model systems (2a,
3a, and 4asTable 2) with four methoxy substituents were
evaluated instead of the experimental systems (2, 3, and 4) that
had benzyloxy substituents at the C3, C4, C5, and C7 positions.
By replacing a phenyl group with a hydrogen atom, we
anticipate that the 2a, 3a, and 4a models would underestimate
the steric effect between the substituents and DMDO; however,
this replacement was significantly more computationally trac-
table for locating and characterizing transition states.
We began the investigation of the facial selectivity by first
considering the conformational flexibility of the three model
compounds 2a-4a. Using the OPLS-AA force field,36 we
performed a conformational search using a Monte-Carlo
protocol available in MacroModel,37 and for the different
compounds, many conformations within ∼3.5 kcal/mol of the
global minimum were obtained: 20 conformers for 2a, 25
conformers for 3a, and 43 conformers for 4a. These unique
conformations were then fully optimized at the B3LYP/6-31G*
level of theory with use of Gaussian03.38 After this subsequent
refinement, there were 4 unique conformers for 2a, 3 unique
conformers for 3a, and 5 unique conformers for 4a; all of these
had relative free energies within ∼1 kcal/mol of the respective
global minimum at the B3LYP level in each case.
The calculated energetic parameters for the low-energy
conformers of 2a, 3a, and 4a are shown in Table 2 at the OPLS-
AA and the B3LYP level. The B3LYP level predicted a different
global minimum from that predicted by the OPLS-AA force
field for 2a and 3a. For 4a, even though these two methods
predicted the same global minimum, the ordering of the low-
energy conformations was different based on these two meth-
ods.39 A similar tendency was observed in our previous study.7
The geometrical parameters for the OPLS-AA and B3LYP/6-
31G* optimized structures are very similar, and obviously
structural differences are not critical for the energetic trends by
the different computational methods. Our previous study stated
that the nature of the difference between molecular mechanics
(AMBER*) and electronic structure (HF/6-31G*) methods
appeared to be their treatment of important hydrogen-bonding
interactions.7 However, in these systems, hydrogen-bonding
interactions are not possible. Therefore, the nature of the
difference between the OPLS-AA and the B3LYP methods
arises from other interactions.
The low-energy conformers of the 2a and 3a had very similar
5
geometries with a C1,2 conformation being preferred for the
ring (Table 2). The differences between these conformers are
related to the orientation of the methoxy substituent on the
C3-C7 positions. The low-energy conformers of 4a, however,
possessed unique conformations of the seven-membered ring,
5
including C1,2
,
1,2C5, and 5,6TC3,4 isomers (Table 2).
(29) Labanowski, J. W.; Andzelm, J. Density Functional Methods in
Chemistry; Springer: New York, 1991.
(30) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
To predict the facial selectivity for epoxidation as suggested
by Wei,12 we computed the electrostatic potential (ESP) of the
most stable conformer of each oxepine (Table 3), in order to
assess a preference for reactivity based on this easily computed
diagnostic for the oxepine. The R-face and ꢀ-face views of the
electron density surface, mapping the values of the electrostatic
potential (red-blue corresponding to ESP values of -0.01 to
+0.01 au), are depicted for the most stable conformer (2a-A,
3a-A, and 4a-A) in each model system.
(31) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(32) Miertus, S.; Scrocco, E.; Tomasi, J. Chem. Phys. 1981, 55, 117.
(33) Cammi, R.; Tomasi, J. J. Comput. Chem. 1995, 16, 1449.
(34) Tomasi, J.; Persico, M. Chem. ReV. 1994, 94, 2027.
(35) Cramer, C. J.; Truhlar, D. G. Chem. ReV. 1999, 99, 2161.
(36) Jorgensen, W. L.; Maxwell, D. S.; Tirado-Rives, J. J. Am. Chem. Soc.
1996, 118, 11225.
(37) (a) Osawa, E.; Musso, H. Angew. Chem., Int. Ed. Engl. 1983, 22, 1. (b)
Halgren, T. A. J. Am. Chem. Soc. 1992, 114, 7827. (c) Halgren, T. A. J. Comput.
Chem. 1996, 17, 616. (d) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.;
Liskamp, R.; Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C.
J. Comput. Chem. 1990, 11, 440.
As shown in Table 3 for 2a-A, 3a-A, and 4a-A, the R-face
had a smaller magnitude of negative charge than the ꢀ-face for
(38) Frisch, M. J. et al. Gaussian 03, Revision B.04; Gaussian, Inc.:
Pittsburgh, PA, 2003.
(39) See the Supporting Information for details.
J. Org. Chem. Vol. 73, No. 16, 2008 6345

Markad et al.
TABLE 2. Calculated Energetic Parameters for the Low-Energy Conformers of 2a, 3a, and 4a: The Relative Energies (∆Erel, kcal/mol) with
the OPLS-AA Force Field and the Relative Free Energies at 298 K (∆G298,rel, kcal/mol) at the B3LYP/6-31+G**//B3LYP/6-31G* Level of
Theorya
a All of the energies are relative to the corresponding global minimum for each molecule.
all three oxepines, although the difference was not large. It is
well accepted that epoxidation appears to involve electrophilic
addition to the alkene.40,41 The ESP results would suggest a
ꢀ-face preference for the DMDO epoxidation. This prediction
is opposite from the experimental results. While asserting
transition state preferences from ground state properties is
obviously questionable, it is often synthetically convenient. The
disparity between the calculated ESP results and the experi-
mentally observed selectivities indicated that the facial prefer-
ences of the transition states were different from the ground
state preferences for these carbohydrate-based oxepines. There-
fore, to quantitatively predict the facial selectivity, the thermo-
dynamics of the reaction pathways, including comprehensive
structural and energetic evaluations of the diastereomeric
transition states, were completed. These transition state calcula-
tions also considered the different conformations available for
each oxepine.
systems (2a, 3a, and 4a) were obtained by using a natural
population analysis (NPA)42 at the B3LYP level of theory. For
all of the DMDO epoxidations of the various conformers, the
NPA results show that negative charge accumulates on the
transferring O of DMDO upon progressing from the reactant
complexes to the transition states. This indicates that DMDO
epoxidations involve electrophilic addition to the double bond
of the oxepine, a trend that is consistent with previous studies.41
On the basis of the NPA results, C2 has a negative charge, in
general, and C1, due to the bonded ring oxygen, has a positive
charge, and suggests that the C2 · · ·O bond would form first
for the eventual epoxide in an asynchronous transition state for
epoxidation.39
The DMDO epoxidations on the R- and ꢀ-faces of the most
stable conformers of 2a, 3a, and 4a were studied at the gas-
phase B3LYP level. The transition state for each pathway was
located and then connected to its respective reactant complexes
(RC) and products (P). Energetic values for all of the stationary
points for R/ꢀ facial attack of DMDO with the different
The atomic charges of the transition states and reactant
complexes for all of the DMDO epoxidations with the model
(40) Lynch, B. M.; Pausacker, K. H. J. Chem. Soc. 1955, 1525.
(42) (a) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985,
83, 735. (b) Reed, A. E.; Weinhold, F.; Curtiss, J. A. Chem. ReV. 1988, 88, 899.
(41) Miaskiewicz, K.; Smith, D. A. J. Am. Chem. Soc. 1998, 120, 1872.
6346 J. Org. Chem. Vol. 73, No. 16, 2008

Epoxidation of Carbohydrate-Based Oxepines
TABLE 3. The r-Face and ꢀ-Face Views of the Electrostatic Potential (Red to Blue Corresponding to Charge Values of -0.01 to +0.01 au)
for the Most Stable Conformers of 2a, 3a, and 4a at the B3LYP/6-31G* Level of Theory in the Gas Phase
a The double bond of the oxepine is located at the bottom of each contour plot.
FIGURE 2. Relative free energies at 298 K (∆G298, kcal/mol) of the most favorable DMDO epoxidations on the R- and ꢀ-faces of 2a at the
B3LYP/6-31+G**//B3LYP/6-31G* level of theory. All of the energies are relative to the infinitely separated DMDO and the most stable conformer
of 2a.
conformers are listed in the Supporting Information. The
energies are presented as relative to the infinitely separated
DMDO and the most stable conformer. Considering the effect
of implicit solvation with the PCM did not significantly alter
the calculated results relative to experiment, therefore, only the
gas-phase results are presented here.43
The calculated results of the most favorable pathways for
the DMDO epoxidations on the R- and ꢀ-faces of 2a are shown
in Figure 2. Among these low-energy conformers (2a-A, 2a-B,
2a-C, and 2a-D), 2a-C underwent the most favorable pathway
for the DMDO epoxidation on the R-face of 2a, while 2a-B
provided that on the ꢀ-face of 2a.39 The DMDO epoxidation
on either face was highly exoergic. As shown in Table 4, the
free energy of activation for the R-face epoxidation was about
(43) All of the PCM results are presented in the Supporting Information.
J. Org. Chem. Vol. 73, No. 16, 2008 6347

Markad et al.
TABLE 4. Difference in the Free Energies of Activation at 298 K
(δ∆Gq, kcal/mol) and the Ratios of the DMDO Epoxidations on the
r- and ꢀ-Faces of 2a, 3a,and 4a from Our Computational and
Experimental Studies
structure probably rationalizes the energetic preference for the
R-face epoxidation in 2a. The epoxidation on either face
proceeded through a transition state with an asynchronous spiro
geometry for which the C2 · · ·O bond forms first (Figure 3).
This transition state could minimize the electrostatic interaction
between the lone-pair electrons on the incoming DMDO oxygen
and those on O6 of the oxepine. We thus postulate that O6 plays
an important role in the transition state and also for dictating
the R-facial kinetic selectivity.
The calculated results of the most favorable pathway for the
DMDO epoxidations on either face of 3a are shown in Figure
4. Among three low-energy conformers (3a-A, 3a-B, and 3a-
C), the 3a-B conformer underwent the most favorable pathway
for the DMDO epoxidation on either face of 3a.39 Similar to
2a, the epoxidation on either face of 3a was highly exoergic.
The calculated free energy of activation for the epoxidation of
the R-face of 3a was 4.6 kcal/mol lower than that for the ꢀ-face;
thus, the R-facial selectivity was more significant for 3a (130:1
R:ꢀ) than for 2a (4:1 R:ꢀ), as shown in Table 4. Once again,
the calculated ratio for R:ꢀ facial selectivity for methyl-
substituted 3a is very consistent with our experimental results
(>25:1 R:ꢀ) for benzyloxy-substituted 3.
theoretical (B3LYP/6-31+G**//B3LYP/6-31G*)a
exptlb
Facial
Facial
selectivity selectivity
c
R-TS (conf.)
ꢀ-TS (conf.)
δ∆G‡ꢀ-R
(R:ꢀ)
(R:ꢀ)
2a 2a-C (5C1,2
)
)
)
2a-B (5,6TB3,4
)
1.3
4.6
9.0
80:20
99.2:0.8
100:0
3:1
>25:1
>25:1
3a 3a-B (5C1,2
3a-B (envelope)
4a 4a-D (5C1,2
4a-A (5C1,2
)
a Gas-phase calculations at the B3LYP/6-31+G **//B3LYP/
6-31G*level of theory (kcal/mol). b In this study. c Based on calculated
δ∆Gq values.
For the three low-energy conformers of 3a, the transition
states of the R-face epoxidation reactions had very similar
5
conformations as their corresponding conformers with a C1,2
ring structure. During the ꢀ-facial attack, the oxepines again
adopted novel ring structures, specifically envelope-like con-
formations.39 In the R-face transition state (Figure 5), the C1
and C2 positions did not deviate much from the plane, which
consists of C3, C4, C6, and O6. Only C5 was displaced from
the plane of the other six ring atoms. Similar to 2a, CH· · ·O
hydrogen bonding did not contribute to the preference for the
R-face epoxidation. On the other hand, the ꢀ-face epoxidation
of 3a proceeded through a transition state with an envelope-
like ring conformation, which appears to have sacrificed some
significant torsional strain to avoid the electrostatic interaction
between the lone-pair electrons on the incoming oxygen of
DMDO and those on O6 of the oxepine. However, for this
envelope conformation, there are still some electrostatic interac-
tions between the unpaired electrons on O5 and on the incoming
O of DMDO, thereby contributing to the R-facial selectively.
Similar to 2a, the epoxidation on either face of 3a proceeded
through a transition state with an asynchronous spirocyclic
geometry. For 3a, O6 appears to play an important role in the
formation of the asynchronous transition state, while O5 and
O6 seem to control the facial selectivity.
FIGURE 3. Front and side views of the transition state geometries of
the most favorable DMDO epoxidations on the R- and ꢀ-faces of 2a,
based on the B3LYP/6-31G* calculations.
1.3 kcal/mol lower than that for the ꢀ-face. Thus the formation
of the R-epoxide is favored kinetically over the ꢀ-epoxidation
(4:1 R:ꢀ). From the calculated δ∆Gq values, the calculated ratio
for the R/ꢀ facial selectivity by DMDO for 2a is very similar
to our experimental results (3:1 R:ꢀ) for the benzyloxy-
substituted 2.
The R-face epoxidation transition states for the four conform-
5
ers of 2a have very similar C1,2 ring conformations to the
The calculated results of the most favorable pathways for
the DMDO epoxidation on the R- and ꢀ-faces of 4a are shown
in Figure 6. Among the five low-energy conformers (4a-A
through 4a-E), the 4a-D conformer underwent the most favor-
able pathway for the R-facial epoxidations, while the 4a-A
conformer provided that preference for the ꢀ-face.44 In a similar
manner as for 2a and 3a, the epoxidation reactions on either
face of 4a were highly exoergic. The calculated free energy of
activation for epoxidation of the R-face of 4a was 9.0 kcal/mol
lower than that for the ꢀ-face; thus, the DMDO epoxidation
should only occur to the R-face of 4a (100:0 R:ꢀ) as shown in
Table 4. The computational results for the model compound
4a are once again very consistent with our experimental ratio
starting oxepine; however, those on the ꢀ-face adopt a novel
(
5,6TB3,4) conformation (Figure 3). In all cases, the geometric
5
search for each transition state started with a C1,2 ring
structure.39 No significant differences in the distances between
the axial hydrogens and the approaching oxygen could be
determined for the epoxidation transition states of the R- and
ꢀ-faces, and CH · · ·O hydrogen bonding did not contribute to
the preference for the R-face epoxidation. As shown in Figure
3, the transition state geometries for the most favorable DMDO
epoxidations on the R- and ꢀ-faces of 2a showed a significant
difference in torsional strain. The ꢀ-face epoxidation of 2a
proceeded through a transition state with a 5,6TB3,4 ring
conformation, which most likely sacrificed some torsional strain
to avoid the electrostatic interaction between lone-pair electrons
on the incoming oxygen and those on O6 of the oxepine. The
more demanding torsional strain for the ꢀ-face transition
(44) The transition states of the DMDO epoxidation of 4a-B and 4a-E could
not be found.
6348 J. Org. Chem. Vol. 73, No. 16, 2008

Epoxidation of Carbohydrate-Based Oxepines
FIGURE 4. The relative free energies at 298 K (∆G298, kcal/mol) of the most favorable DMDO epoxidations on the R- and ꢀ-faces of 3a at
the B3LYP/6-31+G**//B3LYP/6-31G* level of theory. All of the energies are relative to the infinite separated DMDO and most stable
conformer of 3a.
analyzing the geometries of the R-TS and ꢀ-TS structures, we
believe that the most significant contribution for the R-facial
preference was the electrostatic interaction between the lone-
pair electrons on the incoming oxygen of DMDO and those on
O3 and O6 of the oxepine. To avoid the electrostatic interaction
between the unpaired electrons on O3 and the incoming O of
5
DMDO, a C1,2 ring conformation is adopted in the transition
state for ꢀ-facial epoxidation on 4a. As shown in Figure 7, the
O-O3 distance for 4a-A-ꢀ-TS was only 2.79 Å, which was
0.3 Å shorter than the O-O5 distance in the 3a-B-ꢀ-TS case.
Similar as for 2a and 3a, the epoxidation on either face
proceeded through a transition state with an asynchronous spiro
geometry (Figure 7) in which the forming C1 · · ·O and C2 · · ·O
distances are 2.41 and 1.99 Å, respectively, for the favored
R-TS, while they are 2.25 and 1.92 Å, respectively, for the ꢀ-TS.
For 4a, it appears that O6 plays an important role in the
formation of the asynchronous transition state, and O3 controls
the facial selectivity.
2.4. Factors that Govern the Stereoselectivity of DMDO
Epoxidations. The original impetus for this investigation was
to determine the factors that govern selectivity in the epoxidation
of carbohydrate-based oxepines. The agreement between the
FIGURE 5. Front and side views of the transition state geometries of
experimentally determined epoxidation selectivities for 2-4 and
the computational results on 2a-4a provides the basis for
enumerating these factors. While there are both steric and
electronic contributions to the observed selectivities, the ability
of the less sterically demanding oxepines 2a-4a to computa-
tionally reproduce the experimental data for the parent com-
pounds 2-4 suggested that electronic factors predominate. By
incorporating factors of the Wei “majority rule”12 and the
Rainier “asynchronicity rule”,13 a normative model for cyclic
enol ether epoxidation develops. A list of key considerations
includes the following: (i) the number and orientation of
electronegative substituents on the ring relative to the reacting
alkene [these factors are a reiteration of the majority rule]; (ii)
reorganization of the transition state, driven by repulsive
interactions between the ring oxygen of the oxepine (O6) and
the O of DMDO, relative to the reactant; (iii) the overall
the most favorable DMDO epoxidations on the R- and ꢀ-faces of 3a,
based on the B3LYP/6-31G* optimizations.
for the R:ꢀ facial selectivity (>25:1 R:ꢀ) for benzyloxy-
substituted 4.
For the five low-energy conformers of 4a, the epoxidation
transition states on either face had very similar conformations
as their corresponding starting conformers (5C1,2 1,2C5, and
,
5,6TC3,4).39 This was different from the cases of the other two
oxepines, 2a and 3a. Similar to 2a and 3a, however, CH · · ·O
hydrogen bonding did not contribute to the preference for the
R-facial epoxidation. Because the transition state geometries for
the DMDO epoxidations on the R- and ꢀ-faces of 4a did not
show any obvious difference in their structures (Figure 3),
torsional strain does not appear to be the main contribution to
the high preference for the R-facial selectivity. By carefully
J. Org. Chem. Vol. 73, No. 16, 2008 6349

Markad et al.
FIGURE 6. The relative free energies at 298 K (∆G298, kcal/mol) of the most favorable DMDO epoxidations on the R- and ꢀ-faces of 4a at
the B3LYP/6-31+G**//B3LYP/6-31G* level of theory. All of the energies are relative to the infinite separated DMDO and most stable
conformer of 4a.
shows that there are more ꢀ substituents (C4, C5, and C6 for 3
and C3, C4, and C6 for 4) giving rise to R-selectivity for
epoxidation. The C3 deoxygenated oxepine 5 also preferentially
formed the R-epoxide as evidenced by the formation of the
methyl ꢀ-3-deoxy-D-glycero-D-gulo-septanoside 24. For 5, there
are two ꢀ substituents (C4 and C6) and one R substituent (C5).
The observed R-selective epoxidation for this system is in
accordance with the majority rule. Oxepine 1, which is the only
oxepine that gives rise to ꢀ-facial selectivity, follows this
analysis too. It has two R-benzyloxy substituents (C3 and C5)
and only one ꢀ-benzyloxy group at C4. The majority rule is an
electronic effect; the associated steric bulk of the benzyl
protecting groups in these systems should also make a contribu-
tion to the observed selectivities. Wei12 also used the PPFMO
method to predict the facial selectivity for epoxidation of
4-deoxypentenosides; however, the validity of the predictions
from PPFMO analysis depends strongly on the similarity
between the reactant and transition state structuressa relation-
ship that does not apply to the epoxidation of carbohydrate-
FIGURE 7. Front and side views of the transition state geometries of
based oxepines (2a and 3a). In fact, the observed R-selectivity
for epoxidation of 2 may solely reflect the increased steric bulk
on the ꢀ-face of the oxepine due to the size of the protected
hydroxymethyl group at C7 plus one benzyloxy group at C4
versus the two R-benzyloxy groups at C3 and C5.
the most favorable DMDO epoxidations on the R- and ꢀ-faces of 4a,
based on the B3LYP/6-31G* optimizations.
synchronicity in epoxide bond formation; and (iv) the difference
in the overall steric bulk on one side of the ring relative to the
other. There is some overlap in these factors, but each is
qualified in the proceeding paragraphs.
Transition state reorganization and asynchronicity of bond
formation in the epoxidation process also play a significant role
in determining selectivities. The conformational reorganization
observed in the computationally derived transition states for
ꢀ-epoxidation of 2 and 3 contribute to the fact that these are
energetically unfavorable routes; that reorganization for ꢀ-ep-
oxidation of 4 was not seen is somewhat of a surprise. However,
as a rationalization, we should note that in the ꢀ-epoxidation
Application of the Wei majority rule to carbohydrate-based
oxepines 2-4 allows qualitatively accurate prediction of ep-
oxidation selectivities. This result suggests that the majority rule
may be general for substituted cyclic enol ethers. For oxepines
2-4, the orientation of the (protected) C7 hydroxymethyl group
was considered among the electronegative substituents that
would contribute to the majority rule. The overall electronegative
effect may be attenuated due to its remote attachment to the
ring, but it should contribute nonetheless. With this in mind,
collecting substitutents on the oxepines shows that 2 should give
low selectivity, having two “R” substituents at C3 and C5, and
two “ꢀ” substituents at C4 and C6; that is, for 2, there should
be no electronic majority rule. A similar tabulation for 3 and 4
5
of 4, the ring adopted a C1,2 ring conformation to avoid
repulsive electrostatic interactions between unpaired electrons
on O6 and the incoming O of DMDO. In fact, the conforma-
tional reorganization in the ꢀ transition states for 2 and 3 is
likely due to the minimization of the repulsive interaction
between O6 and the O of DMDO. Anticipation of this repulsive
6350 J. Org. Chem. Vol. 73, No. 16, 2008

Epoxidation of Carbohydrate-Based Oxepines
by column chromatography (hexanes/ethyl acetate ) 9/1) to give
4 (0.35 g, 84%) as a thick liquid. Rf 0.52 (hexanes:EtOAc ) 8:2);
[R]D22 -17.55 (c 1.0, CHCl3). 1H NMR (400 MHz, CDCl3) δ 3.55
(dd, J ) 10.8, 2.6 Hz, 1H), 3.59 (dd, J ) 10.8, 5.6 Hz, 1H), 3.66
(dd, J ) 9.8, 2.2 Hz, 1H), 3.92-4.01 (m, 2H), 4.18 (d, J ) 11.4
Hz, 1H), 4.30 (d, J ) 11.4 Hz, 1H), 4.47 (d, J ) 12.3 Hz, 1H),
4.50 (d, J ) 12.5 Hz, 1H), 4.56 (d, J ) 12.3 Hz, 1H), 4.57 (d, J )
12.2 Hz, 1H), 4.60 (d, J ) 1.8 Hz, 1H), 4.67 (d, J ) 12.6 Hz, 1H),
4.73 (dt, J ) 9.6, 1.8 Hz, 1H), 4.83 (d, J ) 12.6 Hz, 1H), 6.42 (dd,
J ) 6.9, 2.2 Hz, 1H), 7.04-7.12 (m, 2H), 7.22-7.37 (m, 16H),
7.38-4.43 (m, 2H); 13C NMR (100 MHz, CDCl3) δ 70.2, 71.2,
72.1, 72.6, 73.3, 76.0, 77.7, 78.8, 105.4, 127.5, 227.5, 127.6, 127.8,
127.9, 127.9, 128.0, 128.3, 128.3, 128.3, 128.4, 137.5, 138.2, 138.4,
138.8, 146.5. ESI-MS m/z (M + Na)+ calcd for C35H36O5Na+
559.2455, found 559.2441.
1,2-Anhydro-3,4,5,7-tetra-O-benzyl-r-D-glycero-L-manno-sep-
tanoside (9). Oxepine 3 (0.22 g, 0.41 mmol) was dried azeotropically
with toluene (3 × 10 mL) under reduced pressure and then dissolved
in dry DCM (12 mL) and cooled in an ice bath to 0 °C. DMDO
(0.55 M) in DCM (2 × 0.88 mL) was added portion-wise and the
mixture was stirred at 0 °C for 1 h. The solvent was removed under
reduced pressure. 1H NMR of the resulting material showed
quantitative conversion (400 MHz, CDCl3) δ 3.12 (d, J ) 2.3 Hz,
1H), 3.44 (dd, J ) 9.0, 8.4 Hz, 1H), 3.53 (dd, J ) 8.7, 5.0 Hz,
1H), 3.75 (dd, J ) 9.5, 1.3 Hz, 1H), 3.86 (dd, J ) 8.1, 5.6 Hz,
1H), 3.96 (d, J ) 0.9 Hz, 1H), 4.36-4.45 (m, 3H), 4.65-4.97 (m,
7H), 7.28-7.43 (m, 20H); 13C NMR (100 MHz, CDCl3) δ 58.7,
69.5, 70.7, 73.4, 74.0, 74.2 (s), 74.7, 76.8, 79.1, 84.1, 127.8, 127.9
(s), 128.0, 128.3, 128.5, 128.6, 128.7, 129.2, 138.2, 138.5, 138.8.
ESI-MS m/z (M + Na)+ calcd for C35H36O6Na+ 575.2404, found
575.2409.
electronic interation between these oxygens for a given face of
a cyclic enol ether allows predictive power in new systems.
The asynchronous nature of the bond formation in the reaction
plays a significant role for R- over ꢀ-epoxidation selectivity of
4. That is because the pseudoaxial allylic substituent (C3)
disfavors epoxidation on the same face based on electronic
repulsion. While reorganization and asynchronicity are primarily
electronic effects, the steric effects of the benzyl protecting
groups also contribute here.
3. Conclusions
We have utilized density functional theory (B3LYP/6-
31+G**//B3LYP/6-31G*) to explain the facial selectivity in
DMDO-mediated epoxidation of carbohydrate-based oxepines.
Through these combined experimental and theoretical treat-
ments, a general model for predicting epoxidation outcomes for
highly substituted cyclic enol ethers has been developed. The
computational method can quantitatively predict the facial
preference of these reactions. It can also provide the transition
state structures that can be used to investigate the details of the
interactions. In this study, the DMDO epoxidation reactions
proceed through a transition state with an asynchronous spiro
geometry, and proceed in an electrophilic manner for which
the most electron-rich carbon center of the double bond attacks
the transferring O of the DMDO oxidant. The lone-pair electrons
on O6 prove to play an important role in the formation of the
asynchronous transition state. Selectivity is determined by two
major electronic factors. First, the collected electronegative
substituents on one side or the other of the double bond favors
attack on DMDO from the side with greater electron density
(the side opposite to the majority of electronegative substituents).
Second, the repulsive interaction between the O6 lone-pair
electrons and the O of DMDO disfavors one transition over
another. This repulsion can even distort the ring conformation
of the oxepine in a disfavored transition state. The factors
governing facial selectivity enumerated here should be ap-
plicable to substituted cyclic enol ethers generally.
1,2-Anhydro-3,4,5,7-tetra-O-benzyl-r-D-glycero-D-galacto-sep-
tanoside (11). Oxepine 4 (0.20 g, 0.37 mmol) was dried azeotro-
pically with toluene (3 × 10 mL) and then dissolved in DCM (20
mL). DMDO (2.44 mL, 0.22 M) was added at 0 °C and the resulting
reaction mixture was stirred for 1 h. The solvent was removed under
reduced pressure to give a thick liquid in quantitative yield
22
according to 1H NMR. Rf ) 0.35 (hexanes:EtOAc 6:4); [R]D
1
+8.07 (c 2.1, CHCl3). H NMR (300 MHz, CDCl3) δ 3.19 (br s,
1H), 3.65 (dd, J ) 10.5, 3.3 Hz, 1H), 3.67 (dd, J ) 8.1, 3.0 Hz,
1H), 3.83 (dd, J ) 10.5, 5.7 Hz, 1H), 3.92-3.98 (m, 1H), 3.99-4.03
(m, 1H), 4.21 (d, J ) 2.7 Hz, 1H), 4.30 (d, J ) 11.4 Hz, 1H), 4.39
(d, J ) 11.7 Hz, 1H), 4.54 (d, J ) 12.0 Hz, 1H), 4.65 (d, J ) 12.0
Hz, 1H), 4.66 (d, J ) 12.3 Hz, 1H), 4.74 (d, J ) 12.3 Hz, 1H),
4.82 (d, J ) 12.0 Hz, 1H), 4.85 (d, J ) 12.3 Hz, 1H), 4.96 (d, J )
2.7 Hz, 1H), 7.11-7.18 (m, 2H), 7.25-7.50 (m, 18H); 13C NMR
(75 MHz, CDCl3) δ 59.2, 69.1, 71.6, 71.7, 71.8, 72.9, 73.1, 75.5,
77.8, 77.9, 80.0, 127.5 (s), 127.55, 127.58, 127.6 (s), 127.7 (s),
127.8, 128.28 (s), 128.3 (s), 128.33 (s), 128.9, 129.6, 137.5, 137.8,
138.0, 138.3.
4. Experimental Methods
1,6-Anhydro-3,4,5,7-tetra-O-benzyl-2-deoxy-D-galacto-sept-1-eni-
tol (3). The synthesis was perfomed in a glovebox. To a solution
of diene 19 (0.26 g, 0.46 mmol) in toluene (115 mL) was added
Schrock catalyst (0.088 g, 0.115 mmol) and the resulting reaction
mixture was stirred in the glovebox for 4 h. The reaction vessel
was taken out from the glovebox and toluene was removed under
reduced pressure to obtain a black residue, which gave 3 (0.22 g,
89%) as a thick liquid after column chromatography (hexanes/ethyl
acetate ) 9/1). Rf 0.52 (hexanes:EtOAc 7:3); [R]D22 +24.8 (c 3.88,
CHCl3). 1H NMR (400 MHz, CDCl3) δ 3.42 (dd, J ) 9.0, 7.8 Hz,
1H), 3.56 (dd, J ) 9.3, 6.4 Hz, 1H), 3.60 (d, J ) 9.3 Hz, 1H), 3.80
(dd, J ) 6.6, 6.4 Hz, 1H), 4.04 (s, 1H), 4.37 (AB, JAB ) 11.9 Hz,
2H), 4.48 (dd, J ) 6.5, 2.2 Hz, 1H), 4.66 (d, J ) 11.8 Hz, 1H),
4.71-4.67 (m, 4H), 4.81 (d, J ) 11.8 Hz, 1H), 4.91 (d, J ) 11.8
3,4,5,7-Tetra-O-benzyl-1,2-dideoxy-D-galacto-hept-1-ene (17). To
a suspension of H3CPPh3Br (0.46 g, 1.29 mmol) in dry THF (5
mL) at 0 °C was added n-BuLi (0.76 mL, 1.6 M in hexane) slowly
over a period of 10 min and the resulting reaction mixture was
stirred for 30 min at 0 °C and 30 min at room temperature. To a
solution of galacto lactol 15 (0.2 g, 0.37 mmol) in THF (5 mL)
was added n-BuLi (0.21 mL, 1.6 M in hexane) at 0 °C and the
reaction mixture was stirred for 30 min. This solution was
transferred to the ylide solution under N2 at 0 °C and the resulting
reaction mixture was stirred at room temperature for 55 h. The
reaction mixture was quenched with saturated NH4Cl (3 mL), THF
was removed under reduced pressure, and then extracted with DCM
(3 × 10 mL). Purification by column chromatography (hexanes/
ethyl acetate ) 9.5/0.5) gave a thick liquid (0.155 g, 78%). Rf 0.48
(hexanes:EtOAc 8:2). [R]D22 -4.03 (c 1.6, CHCl3). 1H NMR (400
MHz, CDCl3) δ 3.07 (d, J ) 8.0 Hz, D2O exchangeable, 1H),
3.49-3.3.60 (m, 2H), 3.80-3.87 (m, 2H), 4.12 (dd, J ) 8.0, 4.0
Hz, 1H), 4.16 (dd, J ) 8.0, 4.0 Hz, 1H), 4.35-4.45 (m, 2H), 4.46
(d, J ) 12.0 Hz, 1H), 4.53 (d, J ) 12.0 Hz, 1H), 4.68 (d, J ) 11.8
Hz, 1H), 6.22 (dd, J ) 8.0, 2.2 Hz, 1H), 7.23-7.38 (m, 20H); 13
C
NMR (100 MHz, CDCl3) δ 70.0, 72.9, 73.5, 73.7, 74.2, 74.7, 79.0,
85.4, 107.3, 127.8 (s), 127.9, 128.0, 128.1 (s), 128.5, 128.6 (s),
138.1, 138.6, 138.8, 144. ESI-MS m/z (M + Na)+ calcd for
C35H36O5Na+ 559.2455, found 559.2445.
1,6-Anhydro-3,4,5,7-tetra-O-benzyl-2-deoxy-D-manno-sept-1-eni-
tol (4). The synthesis was perfomed in a glovebox. To a solution
of diene 20 (0.44 g, 0.78 mmol) in toluene (190 mL) was added
Schrock catalyst (0.149 g, 0.195 mmol) in toluene (10 mL) and
the resulting solution was stirred in the glovebox for 4 h. Toluene
was removed under reduced pressure and the residue was purified
J. Org. Chem. Vol. 73, No. 16, 2008 6351

Markad et al.
Hz, 1H), 4.71-4.78 (m, 1H), 4.79 (s, 2H), 5.33 (d, J ) 8.0 Hz,
1H), 5.38 (d, J ) 16.0 Hz, 1H), 5.90 (ddd, J ) 16.0, 7.9, 2.5 Hz,
1H), 7.20-7.42 (m, 20H); 13C NMR (100 MHz, CDCl3) δ 69.7,
70.3, 71.2, 73.08, 73.14, 75.2, 76.6, 80.8, 82.1, 119.1, 127.5 (s),
127.6 (s), 127.7 (s), 128.0 (s), 128.1 (s), 128.3 (s), 135.7, 138.0,
138.1, 138.2. ESI-MS m/z (M + Na)+ calcd for C35H38O5Na+
561.2611, found 561.2594.
3,4,5,7-Tetra-O-benzyl-1,2-dideoxy-6-O-vinyl-D-manno-hept-1-
ene (20). To a solution of alcohol 18 (0.58 g, 1.08 mmol) in toluene
(20 mL) was added [Ir(cod)Cl]2 (0.072 g, 0.11 mmol), Na2CO3
(0.069 g, 0.65 mmol), and vinyl acetate (0.2 mL, 21.6 mmol). The
solution was heated to 100 °C and stirred for 15 h. Toluene was
removed under reduced pressure and the residue was extracted with
DCM (3 × 10 mL). Column chromatography of the residue
(hexanes/ethyl acetate ) 9.5/0.5) gave a thick liquid 20 (0.451 g,
3,4,5,7-Tetra-O-benzyl-1,2-dideoxy-D-manno-hept-1-ene (18). To
a solution of lactol 16 (1.4 g, 2.59 mmol) in THF (15 mL) was
added n-BuLi (1.7 mL, 1.6 M) at -10 °C and the resulting solution
was stirred for 30 min at 0 °C. To a solution of Ph3PCH3Br (3.24
g, 9.06 mmol) was added n-BuLi (1.6 M, 5.67 mL) at -20 °C and
the resulting solution was allowed to warm to 0 °C and stirred for
30 min. The ylide solution was added to the lactol via cannula at
-20 °C and stirred for 24 h at rt. Saturated NH4Cl (5 mL) was
added and THF was removed under reduced pressure. The residue
was dissolved in DCM (40 mL) and washed with water and brine.
The DCM layer was dried and solvent removed under reduced
pressure. Purification of the resulting material by column chroma-
tography (hexanes/ethyl acetate ) 9/1) gave a thick liquid (0.42 g,
30%). Rf 0.59 (hexanes:EtOAc 7:3). [R]D22 +6.38 (c 1.55, CHCl3).
1H NMR (400 MHz, CDCl3) δ 2.64 (d, J ) 5.8 Hz, 1H, D2O
exchangeable), 3.56 (dd, J ) 9.6, 5.4 Hz, 1H), 3.62 (dd, J ) 9.6,
3.4 Hz, 1H), 3.83-3.91 (m, 2H), 3.96-4.04 (m, 1H), 4.11 (d, J )
7.4 Hz, 1H), 4.24 (d, J ) 11.7 Hz, 1H), 4.47 (s, 2H), 4.48 (d, J )
11.1 Hz, 1H), 4.49 (d, J ) 11.2 Hz, 1H), 4.60 (d, J ) 11.3 Hz,
1H), 4.62 (d, J ) 11.1 Hz, 1H), 4.71 (d, J ) 11.2 Hz, 1H), 5.39
(dd, J ) 8.8, 1.5 Hz, 1H), 5.42 (d, J ) 17.3 Hz, 1H), 5.94 (ddd, J
) 17.6, 7.8, 2.4 Hz, 1H), 7.17-7.38 (m, 20H); 13C NMR (100
MHz, CDCl3) δ 69.9, 70.2, 71.3, 73.3, 73.9, 74.3, 78.6, 80.3, 80.9,
119.7, 127.47, 127.54, 127.66, 127.7, 127.8, 127.84, 128.2, 128.21,
128.3, 128.4, 136.2, 138.0, 138.4, 138.5. ESI-MS m/z (M + Na)+
calcd for C35H38O5Na+ 561.2611, found 561.2615.
22
74%). Rf 0.57 (hexanes:EtOAc 8:2). [R]D -1.56 (c 0.9, CHCl3).
1H NMR (400 MHz, CDCl3) δ 3.67 (dd, J ) 10.8, 5.0 Hz, 1H),
3.78 (dd, J ) 7.0, 3.5 Hz, 1H), 3.84 (dd, J ) 10.8, 2.6 Hz, 1H),
4.02 (dd, J ) 6.5, 1.6 Hz, 1H), 4.08 (t, J ) 7.7 Hz, 1H), 4.09 (dd,
J ) 7.0, 3.5 Hz, 1H), 4.16-4.20 (m, 1H), 4.24 (d, J ) 11.6 Hz,
1H), 4.34 (dd, J ) 14.0, 1.6 Hz, 1H), 4.49 (s, 2H), 4.51 (d, J )
11.3 Hz, 1H), 4.55 (d, J ) 11.4 Hz, 1H), 4.56 (d, J ) 11.5 Hz,
1H), 4.59 (d, J ) 12.3 Hz, 1H), 4.70 (d, J ) 11.0 Hz, 1H), 5.39
(d, J ) 8.6 Hz, 1H), 5.43 (d, J ) 15.7 Hz, 1H), 5.96 (ddd, J )
17.6, 7.9, 2.4 Hz, 1H), 6.30 (dd, J ) 14.1, 6.5 Hz, 1H), 7.20-7.32
(m, 20H); 13C NMR (100 MHz, CDCl3) δ 68.9, 69.9, 73.3, 74.4,
74.43, 77.8, 78.7, 80.4, 81.0, 89.2, 119.8, 127.37, 127.42, 127.59,
127.61, 127.7, 128.0, 128.1, 128.2, 128.3, 136.3, 138.2, 138.4,
138.6, 138.7, 151.0. ESI-MS m/z (M + Na)+ calcd for C37H40O5Na+
587.2768, found 587.2768.
Methyl-3,4,5,7-tetra-O-benzyl-ꢀ-D-glycero-L-manno-septano-
side (25). To a solution of epoxide 9 in CH3OH (5 mL) was added
NaOCH3 (20 mg) at 0 °C and the mixture was stirred overnight at
rt. The reaction was quenched by adding saturated NH4Cl (2 mL)
and volatiles were removed under reduced pressure. The residue
was dissolved in DCM (10 mL) and washed with H2O. The DCM
layer was dried and solvent was removed under reduced pressure.
Purification of this material by column chromatography (hexanes/
ethyl acetate ) 7/3) gave ꢀ-methyl septanoside 25 (0.16 g, 67%
overall). Rf 0.59 (hexanes:EtOAc 1:1). [R]D22 +41.9 (c 3.38, CHCl3).
1H NMR (400 MHz, CDCl3) δ 2.84 (d, J ) 1.6 Hz, 1H), 3.43 (s,
3H), 3.59-3.73 (m, 3H), 3.83 (dd, J ) 8.9, 1.5 Hz, 1H), 3.97-4.02
(m, 2H), 4.23 (d, J ) 6.4 Hz, 1H), 4.30 (dd, J ) 8.9, 4.4 Hz, 1H),
4.48 (AB, JAB ) 11.8 Hz, 2H), 4.58-4.68 (m, 3H), 4.82 (d, J )
12.0 Hz, 1H), 4.84 (d, J ) 12.0 Hz, 1H), 4.95 (d, J ) 11.6 Hz,
1H), 7.24-7.35 (m, 20H); 13C NMR (100 MHz, CDCl3) δ 55.9,
69.9, 73.6, 73.7, 74.1, 74.41, 74.44, 75.6, 77.2, 79.1, 79.5, 79.6,
108.1, 127.48 (s), 127.50 (s), 127.7 (s), 127.8 (s), 127.9 (s), 128.1
(s), 128.2 (s), 128.3 (s), 128.4 (s), 138.1, 138.2, 138.6, 138.7. ESI-
MS m/z (M + Na)+ calcd for C36H40O7Na+ 607.2666, found
607.2642.
3,4,5,7-Tetra-O-benzyl-1,2-dideoxy-6-O-vinyl-D-galacto-hept-1-
ene (19). Ethyl vinyl ether (75 mL) and DCM (5 mL) were
combined in a flame-dried round-bottomed flask. To this mixture
was added 1,10-phenanthroline (0.098 g, 0.54 mmol) followed by
Pd(OAc)2 (0.121 g, 0.54 mmol) with stirring for 15 min. To this
mixture was added alcohol 17 (1.95 g, 3.62 mmol) in DCM (20
mL) and the resulting reaction mixture was refluxed for 10 d. The
solvent was removed under reduced pressure. The residue was
directly loaded on a column and purified (hexanes/ethyl acetate )
9.5/0.5) to give a thick liquid 19 (0.27 g, 13%) in 59% yield based
on 0.46 g, conversion. Rf 0.64 (hexanes:EtOAc 8:2). Repeated
attempts at obtaining HRMS data for 19 provided only the product
2-O-Acetyl-3,4,5,7-tetra-O-benzylmethyl-ꢀ-D-glycero-L-manno-
septanoside (25a). To a solution of 25 (0.05 g, 0.086 mmol) in
pyridine (0.068 g, 6.5 mmol) was added Ac2O (0.26 g, 2.56 mmol)
and DMAP (2 mg) and the mixture was stirred for 15 h at rt.
Volatiles were removed under reduced pressure and the reaction
mixture was purified by column chromatography (hexanes/ethyl
acetate ) 9/1) to give 25a (0.05 g, 94%). Rf 0.51 (hexanes:EtOAc
22
1
of vinyl ether hydrolysis (17). [R]D +6.87 (c 5.39, CHCl3). H
NMR (400 MHz, CDCl3) δ 3.59 (dd, J ) 9.8, 6.4 Hz, 1H), 3.65
(dd, J ) 9.8, 5.9 Hz, 1H), 3.75 (dd, J ) 8.0, 3.2 Hz, 1H), 3.99 (dd,
J ) 6.5, 1.6 Hz, 1H), 4.03 (dd, J ) 8.0, 2.7 Hz, 1H), 4.18 (dd, J
) 7.8, 3.1 Hz, 1H), 4.32 (d, J ) 12.0 Hz, 1H), 4.30-4.34 (m,
1H), 4.37 (dd, J ) 14.1, 1.6 Hz, 1H), 4.42 (s, 2H), 4.43 (AB, JAB
) 12.0 Hz, 2H), 4.62 (d, J ) 12.0 Hz, 1H), 4.65 (d, J ) 12.0 Hz,
1H), 4.70 (d, J ) 12.0 Hz, 1H), 5.29 (d, J ) 11.3 Hz, 1H), 5.38
(d, J ) 17.4 Hz, 1H), 6.00 (ddd, J ) 17.8, 7.7, 2.6 Hz, 1H), 6.36
(dd, J ) 14.0, 6.5 Hz, 1H), 7.16-7.36 (m, 20H); 13C NMR (100
MHz, CDCl3) δ 68.6, 69.9, 72.9, 73.8, 74.4, 76.9, 77.2, 79.8, 80.9,
88.3, 118.4, 127.2, 127.23 (s), 127.3 (s), 127.33, 127.4 (s), 127.6
(s), 127.62 (s), 127.68 (s), 127.95 (s), 128.0 (s), 128.1, 128.2, 136.3,
137.7, 138.1, 138.2, 138.23, 151.6.
22
1
7:3); [R]D -29.13 (c 0.7, CHCl3). H NMR (300 MHz, CDCl3)
δ 2.08 (s, 3H), 3.39 (s, 3H), 3.62 (dd, J ) 9.8, 5.4 Hz, 1H), 3.68
(dd, J ) 9.8, 6.8 Hz, 1H), 3.77-3.82 (m, 1H), 3.84 (dd, J ) 8.4,
2.1 Hz, 1H), 3.97 (t, J ) 2.4 Hz, 1H), 4.31 (dd, J ) 8.5, 3.6 Hz,
1H), 4.35 (d, J ) 5.8 Hz, 1H), 4.45 (AB, JAB ) 11.8 Hz, 2H), 4.52
(d, J ) 11.7 Hz, 1H), 4.60 (d, J ) 12.0 Hz, 1H), 4.62 (d, J ) 12.0
Hz, 1H), 4.63 (d, J ) 11.7 Hz, 1H), 4.82 (d, J ) 12.0 Hz, 1H),
4.86 (d, J ) 11.7 Hz, 1H), 5.49 (dd, J ) 5.8, 3.6 Hz, 1H),
7.18-7.34 (m, 20H); 13C NMR (75 MHz, CDCl3) δ 21.1, 55.8,
70.0, 73.4, 73.6, 73.9, 74.0, 74.4, 75.2, 78.1, 78.4, 78.9, 105.6,
127.4, 17.6, 127.61, 127.8, 127.9, 128.1, 128.13, 128.2, 128.3,
169.6. ESI-MS m/z (M + Na)+ calcd for C38H42O8Na+ 649.2772,
found 649.2771.
3,4,5,7-Tetra-O-benzyl-1,2-dideoxy-6-O-vinyl-D-galacto-hept-1-
ene (19). To a solution of 17 (0.1 g, 0.186 mmol) in toluene (5
mL) was added [Ir(cod)Cl]2 (0.012 g, 0.018 mmol), Na2CO3 (0.013
g, 0.113 mmol), and vinyl acetate (0.33 g, 3.72 mmol). The solution
was heated to 100 °C and stirred for 12 h. Toluene was removed
under reduced pressure and the residue was extracted with DCM
(3 × 10 mL). Purification by column chromatography (hexanes/
ethyl acetate ) 9.5/0.5) gave a thick liquid 19 (0.098 g, 93%). The
physical and analytical data were found to be identical as given
above.
Methyl-3,4,5,7-tetra-O-benzyl-r-D-glycero-D-galacto-septano-
side (26) and 1,5-Anhydro-3,4,7-tri-O-benzyl-r-D-glycero-D-galacto-
septanose (27). To a solution of oxepine 4 (0.2 g, 0.37 mmol) in
DCM (20 mL) was added DMDO (2.44 mL, 0.22 M) at 0 °C and
the mixture was stirred for 1 h. Volatiles were removed under
6352 J. Org. Chem. Vol. 73, No. 16, 2008

Epoxidation of Carbohydrate-Based Oxepines
reduced pressure and 1H NMR showed quantitative conversion with
formation of a mixture of products. The residue was dissolved in
CH3OH (10 mL) and NaOCH3 (15 mg) was added. The mixture
was stirred at rt for 12 h and then quenched by adding saturated
NH4Cl (3 mL). The solvent was removed under reduced pressure
and H2O (15 mL) was added to the mixture. Extraction with DCM
(3 × 10 mL), drying, removal of solvent, and purification by column
chromatography gave two products. The first eluted with hexanes/
ethyl acetate ) 9/1 giving 26 (0.12 g, 55%) as a thick liquid. Rf )
0.55 (hexanes:EtOAc 7:3). [R]D22 +55.91 (c 2.28, CHCl3). 1H NMR
(300 MHz, CDCl3) δ 2.67 (d, J ) 4.2 Hz, D2O exchangeable, 1H),
3.45 (s, 3H), 3.60 (d, J ) 4.1 Hz, 2H), 3.78 (dd, J ) 9.3, 5.8 Hz,
1H), 3.94 (d, J ) 7.7 Hz, 1H), 3.98-4.04 (m, 2H), 4.10 (dt, J )
7.6, 3.8 Hz, 1H), 4.31 (d, J ) 11.2 Hz, 1H), 4.44 (d, J ) 12.2 Hz,
1H), 4.54 (d, J ) 12.2 Hz, 1H), 4.55 (d, J ) 11.2 Hz, 1H), 4.64
(d, J ) 11.9 Hz, 1H), 4.71 (s, 2H), 4.73 (d, J ) 11.9 Hz, 1H), 4.88
127.85 (s), 127.9 (s), 128.1 (s), 128.3 (s), 128.4 (s), 128.5 (s), 137.3,
137.4, 138.1, 169.8. ESI-MS m/z (M + Na)+ calcd for C30H32O7Na+
527.2040, found 527.2042.
Methyl-3,4;5,7-diacetonide-ꢀ-D-glycero-D-galacto-septanoside (29).
To a solution of oxepine 28 (0.04 g, 0.156 mmol) in DCM (10
mL) was added DMDO (0.53 mL, 0.36 M) at 0 °C and the mixture
was stirred for 1 h. Volatiles removed under reduced pressure and
1H NMR of the residue showed quantitative conversion of 28. This
material was dissolved in CH3OH (5 mL) and NaOCH3 (7 mg)
was added. The reaction was stirred overnight and quenched by
adding saturated NH4Cl (2 mL). Volatiles were removed under
reduced pressure and H2O (15 mL) was added to the residue.
Extraction with DCM (3 × 10 mL), drying of the DCM layers,
and removal of solvent under reduced pressure was followed by
purification via column chromatography (hexanes/ethyl acetate )
4/6) to give 29 (0.04 g, 84% for 2 steps) as a white solid. Rf 0.49
(hexanes:EtOAc 2:8). Mp 96-98 °C. 1H NMR (400 MHz, CDCl3)
δ 1.35 (s, 6H), 1.41 (s, 3H), 1.48 (s, 3H), 2.42-2.59 (br s, D2O
exchangeable, 1H), 3.47 (m, 1H), 3.50 (s, 3H), 3.63 (d, J ) 8.0
Hz, 1H), 4.01-4.10 (m, 4H), 4.21 (d, J ) 8.0 Hz, 1H), 4.37 (q, J
) 8.0 Hz, 1H); 13C NMR (100 MHz, CDCl3) δ 25.4, 26.4, 26.9,
28.1, 56.9, 66.6, 73.2, 73.5, 73.7, 74.2, 78.7, 103.4, 109.4, 110.3.
ESI-MS m/z (M + Na)+ calcd for C14H24O7Na+ 327.1414, found
327.1410.
Methyl-ꢀ-D-glycero-L-manno-septanoside (25b). To a solution of
25 (0.15 g, 0.26 mmol) in CH3OH (15 mL) was added Pd(OH)2
(0.17 g, 1.23 mmol) and the solution was stirred under H2 (1 atm)
for 3 h. The catalyst was filtered through a pad of celite and the
filtrate was concentrated under reduced pressure to give a thick
liquid (0.054 g, 95%). Rf 0.54 (DCM:CH3OH 9:1). [R]D22 +6.0 (c
1.4, CH3OH). 1H NMR (400 MHz, CD3OD) δ 3.41 (s, 3H,
-OCH3), 3.61-3.66 (m, 3H), 3.70 (dd, J ) 12.3, 8.2 Hz, 1H),
3.86 (dd, J ) 2.8, 2.8 Hz, 1H), 3.91 (d, J ) 1.8 Hz, 1H), 3.99 (dd,
J ) 9.5, 3.7 Hz, 1H), 4.31 (d, J ) 4.8 Hz, 1H); 13C NMR (100
MHz, CD3OD) δ 56.8, 63.3, 68.8, 71.7, 72.8, 76.2, 80.3, 110.1.
ESI-MS (M + Na+) calcd for C8H16O7Na+ 247.0794, found
247.0796.
(d, J ) 3.5 Hz, 1H), 7.08-7.20 (m, 2H), 7.24-7.38 (m, 18H); 13
C
NMR (75 MHz, CDCl3) δ 55.7, 70.3, 71.0, 71.4, 73.1, 73.2, 73.4,
73.5, 77.4, 77.9, 79.7, 98.4, 127.4 (s), 127.5 (s), 127.54 (s), 127.6
(s), 127.7 (s), 127.8 (s), 128.0 (s), 128.2 (s), 128.23 (s), 128.3 (s),
138.0, 138.1, 138.3, 138.5. ESI-MS m/z (M + Na)+ calcd for
C36H40O7Na+ 607.2666, found 206.2671. Further elution with
hexanes:EtOAc (8.5:1.5) gave 27 (0.05 g, 29%) as a thick liquid.
22
1
Rf 0.59 (hexanes:EtOAc 7:3). [R]D +67.54 (c 2.0, CHCl3). H
NMR (300 MHz, CDCl3) δ 2.48 (d, J ) 4.7 Hz, D2O exchangeable,
1H), 3.56 (dd, J ) 10.8, 4.7 Hz, 1H), 3.62 (dd, J ) 10.8, 2.0 Hz,
1H), 3.69 (dd, J ) 9.7, 4.3 Hz, 1H), 3.85 (m, 1H), 4.07 (d, J ) 2.2
Hz, 1H), 4.25-4.31 (m, 1H), 4.36 (d, J ) 4.3 Hz, 1H), 4.40 (d, J
) 11.6 Hz, 1H), 4.45 (d, J ) 12.3 Hz, 1H), 4.49 (d, J ) 11.5 Hz,
1H), 4.51 (d, J ) 11.5 Hz, 1H), 4.60 (d, J ) 11.4 Hz, 1H), 4.61
(d, J ) 12.6 Hz, 1H), 5.27 (d, J ) 4.6 Hz, 1H), 7.14-7.35 (m,
15H); 13C NMR (75 MHz, CDCl3) δ 69.2, 70.3, 71.3, 72.2, 73.4,
73.7, 79.6, 79.7, 82.9, 98.7, 127.6 (s), 127.8 (s), 127.83 (s), 127.9
(s), 128.0, 128.3 (s), 128.4 (s), 128.5 (s), 137.5 (s), 137.6, 137.9.
ESI-MS m/z (M + Na)+ calcd for C28H30O6Na+ 485.1935, found
485.1929.
2-O-Acetyl-3,4,5,7-tetra-O-benzylmethyl-r-D-glycero-D-galacto-
septanoside (26a). To a solution of 26 (0.09 g, 0.154 mmol) in
pyridine (0.12 g, 1.54 mmol) was added Ac2O (0.47 g, 4.6 mmol)
and DMAP (5 mg). The mixture was stirred for 15 h at rt and
volatiles were removed under reduced pressure. The residue was
purified by column chromatography (hexanes/ethyl acetate ) 9/1)
giving 26a (0.09 g, 93%) as a thick liquid. Rf 0.49 (hexanes:EtOAc
Methyl-r-D-glycero-D-galacto-septanoside (26b). To a solution
of 26 (0.1 g, 0.17 mmol) in CH3OH (10 mL) was added Pd(OH)2
(0.115 g, 0.82 mmol) and the solution was stirred under H2 (1 atm)
for 4 h. Catalyst was filtered through a pad of celite and the filtrate
was concentrated under reduced pressure to give 26b as a thick
liquid (0.036 g, 94%). Rf 0.52 (DCM:CH3OH 9:1). [R]D22 +57.31
(c 1.5, CH3OH). 1H NMR (400 MHz, CD3OD) δ 3.43 (s, 3H), 3.60
(dd, J ) 8.8, 7.7 Hz, 1H), 3.65 (dd, J ) 12.0, 7.0 Hz, 1H), 3.74
(dd, J ) 12.0, 3.0 Hz, 1H), 3.74-3.77 (m, 1H), 3.84 (d, J ) 1.8
Hz, 1H), 3.86 (d, J ) 3.0 Hz, 1H), 3.96 (dd, J ) 6.8, 2.2 Hz, 1H),
4.67 (d, J ) 3.0 Hz, 1H); 13C NMR (100 MHz, CD3OD) δ 56.4,
64.3, 71.6, 73.2, 73.3, 75.3, 75.6, 100.1. ESI-MS (M + Na+) calcd
for C8H16O7Na+ 247.0794, found 247.0783.
22
1
7:3); [R]D +34.70 (c 1.5, CHCl3). H NMR (300 MHz, CDCl3)
δ 2.01 (s, 3H), 3.39 (s, 3H), 3.57 (dd, J ) 10.4, 2.6 Hz, 1H), 3.63
(dd, J ) 10.4, 5.1 Hz, 1H), 3.84 (dd, J ) 8.9, 6.1 Hz, 1H),
3.85-3.92 (m, 2H), 3.93-4.00 (m, 1H), 4.31 (d, J ) 11.2 Hz,
1H), 4.41 (d, J ) 12.1 Hz, 1H), 4.52 (d, J ) 11.2 Hz, 1H), 4.56
(d, J ) 11.2 Hz, 1H), 4.70 (s, 2H), 4.73 (s, 2H), 4.94 (d, J ) 3.2
Hz, 1H), 5.34 (dd, J ) 7.5, 3.2 Hz, 1H), 7.10-7.34 (m, 20H); 13
C
NMR (75 MHz, CDCl3) δ 21.0, 55.9, 70.9, 71.0, 72.1, 73.1, 73.3,
73.48, 73.53, 76.7, 76.8, 79.4, 97.2, 127.5, 127.52, 127.54, 127.6,
127.7 (s), 128.1 (s), 128.2 (s), 128.3 (s), 138.0, 138.1, 138.3 (s).
ESI-MS m/z (M + Na)+ calcd for C38H42O8Na+ 649.2772, found
649.2775.
Methyl-ꢀ-D-glycero-D-galacto-septanoside (29b). To a solution of
29 (0.16 g, 0.526 mmol) in EtOH/H2O (32 mL, 1:3) was added
Amberlite-IR-120 (260 mg) then the solution was refluxed at 110
°C for 3.5 h. The reaction mixture was cooled and filtered through
a pad of celite and the filtrate was concentrated under reduced
pressure to give 29b (0.11 g, 93%) as an oil. Rf 0.35 (DCM:CH3OH
8:2). [R]D22 -13.97 (c 1.4, CH3OH). 1H NMR (300 MHz, CD3OD)
δ 3.34-3.44 (m, 1H), 3.48 (s, 3H), 3.60-3.3.74 (m, 3H), 3.76-3.88
(m, 3H), 4.22 (d, J ) 7.2 Hz, 1H); 13C NMR (75 MHz, CD3OD)
δ 56.7, 64.6, 71.8, 76.2, 76.3, 76.4, 84.5, 110.8. ESI-MS (M +
Na+) calcd for C8H16O7Na+ 247.0794, found 247.0773.
Computational Methods. A Monte-Carlo conformational search
(torsional sampling: 5000 steps; energy window for saving struc-
tures: 50 kJ/mol) was used to generate the conformers of 2a, 3a,
and 4a. The calculations were performed by employing the OPLS-
AA force field36 as implemented in the MacroModel software
(version 8.5)37 from Schro¨dinger. This process resulted in 431
unique conformations for 2a, 655 unique conformations for 3a, and
750 unique conformations for 4a.
1,5-Anhydro-3,4,7-tri-O-benzyl-r-D-glycero-D-galacto-sep-
tanose (27a). To a solution of 27 (0.06 g, 0.13 mmol) in pyridine
(0.5 mL) was added Ac2O (2 mL) and DMAP (2 mg) and the
mixture was stirred at rt for 15 h. Volatiles were removed under
reduced pressure and the material was purified by column chro-
matography (hexanes:EtOAc 9:1) to give 27a (0.058 g, 89%) as a
22
thick liquid. Rf 0.47 (hexanes:EtOAc 8:2); [R]D +99.85 (c 1.06,
1
CHCl3). H NMR (300 MHz, CDCl3) δ 2.06 (s, 3H), 3.60 (br s,
2H), 3.75 (br d, J ) 1.3 Hz, 2H), 4.28 (d, J ) 2.2 Hz, 1H), 4.41
(br s, 1H), 4.43 (d, J ) 11.4 Hz, 1H), 4.44 (d, J ) 11.9 Hz, 1H),
4.51 (s, 2H), 4.57 (d, J ) 13.0 Hz, 1H), 4.61 (d, J ) 12.2 Hz, 1H),
5.08 (dd, J ) 4.4, 2.2 Hz, 1H), 5.58 (d, J ) 4.4 Hz, 1H), 7.05-7.23
(m, 2H), 7.25-7.46 (m, 13H); 13C NMR (75 MHz, CDCl3) δ 20.6,
68.9, 70.1, 71.4, 72.4, 73.3, 73.9, 79.1, 79.9, 80.2, 96.9, 127.8 (s),
J. Org. Chem. Vol. 73, No. 16, 2008 6353

Markad et al.
The 20 most stable conformers of 2a, 25 most stable conformers
of 3a, and 43 most stable conformers of 4a from the OPLS-AA
conformational searches were further optimized at the B3LYP/6-
31G* level of theory.28–31 All density functional theory (DFT)
calculations were performed with Gaussian0338 at the Ohio
Supercomputer Center. The optimized geometries and subsequently
the harmonic vibrational frequencies for all of these conformers
were obtained, and the relative energies were compared. A scaling
factor of 0.980645 was used for the zero-point vibrational energy
(ZPE) corrections. Enthalpies and free energies were obtained from
the calculated thermal and entropic corrections at 298 K, using the
unscaled, harmonic vibrational frequencies for the vibrational
contribution to the partition function. After this analysis, 4
conformations for 2a, 3 conformations for 3a, and 5 conformations
for 4a were obtained, and for which their relative free energies
were within ∼1 kcal/mol of the global minimum in each case.
The transition states for the epoxidation by DMDO with 10
different conformers of 2a, 3a, and 4a were located at the B3LYP/
6-31G* level of theory. Transition state structures were confirmed
to have one imaginary vibrational frequency. The desired reactants
and products were located by displacement (10%) of the transition
state geometries along the normal coordinate of the imaginary
vibrational frequency in the positive and negative directions,
followed by careful optimization with the analytical second
derivatives (opt ) calcfc).
The electrostatic potential for each of the most stable conformers
of 2a, 3a, and 4a (2a-A, 3a-A, and 4a-A, respectively) was
generated with Gaussview 3.046 at the B3LYP/6-31G* level of
theory in the gas phase. The isosurface value was 0.02 au.
Acknowledgment. The authors thank Chris Incarvito (Yale
University) for collection of the X-ray data. Financial support
to M.W.P. was provided by the National Science Foundation
(CAREER CHE-0546311) and the University of Connecticut.
S.X. acknowledges a Lubrizol fellowship at The Ohio State
University. C.M.H. is grateful to the Ohio Supercomputer Center
for generous computational resources as well as financial support
from the National Science Foundation and the National Institutes
of Health.
Supporting Information Available: NMR spectra for all
new compounds and a summary of enthalpic and free energies,
optimized geometries, vibrational frequencies, as well as NPA
results for each conformer, reactant complex, transition state
and product. This material is available free of charge via the
Internet at http://pubs.acs.org. Complete crystallographic data
for the structural analysis of methyl septanoside 29 have been
deposited in the Cambridge Crystallographic Data Centre
(CCDC), No. 655858; copies of this information may be
obtained free of charge from CCDC, 12 Union Road, Cambridge
CB2 1EZ, UK (Fax: +44-1223-336033; web: www.ccdc.cam.
ac.uk/conts/retrieving/html; e-mail: deposit@ccdc.cam.ac.uk).
Single-point energy and natural population analysis (NPA)42
calculations were then performed for these stationary points at the
B3LYP/6-31+G** level of theory with six Cartesian d functions.
The effect of solvation on the gas-phase single-point calculations
was investigated by using the polarizable continuum model
(PCM)32–35 for CH2Cl2 and CH3OH as solvents.
JO800979A
(45) National Institute of Standards and Technology, Computational Chem-
istry Comparison and Benchmark Database Release 12, 2005.
(46) Copyright 2000-2003 by Semichem, Inc.
6354 J. Org. Chem. Vol. 73, No. 16, 2008