An unexpected rearrangement during Mitsunobu epimerization reaction of sugar derivatives

Article abstract of DOI:10.1021/jo0000675

Mitsunobu reaction on the glucose derivative (3S,4R,5R,6R)-3,4,5,7- tetrabenzyloxy-6-hydroxy-1-heptene yielded an unexpected rearrangement major product. Its structure was determined as (3R,4R,5R,6S)-4,5,6,7- tetrabenzyloxy-3-hydroxy-1-heptene. The suggested rearrangement mechanism involves an initial intramolecular cyclization, followed by ring opening by the nucleophile p-nitrobenzoate. Product distribution of the Mitsunobu reaction was substrate-dependent, with the corresponding mannose derivative (the 3R epimer) giving less of the initial intramolecular reaction products and the corresponding galactose derivative (the 5S epimer) yielding almost exclusively the expected epimerization product. Varying the Mitsunobu reaction conditions (addition of base and using nonpolar solvent) led to the expected epimerization product of the glucose derivative.

Full text of DOI:10.1021/jo0000675

J . Org. Chem. 2000, 65, 3775-3780
3775
An Un exp ected Rea r r a n gem en t d u r in g Mitsu n obu Ep im er iza tion
Rea ction of Su ga r Der iva tives
Rachel Persky and Amnon Albeck*
The J ulius Spokojny Bioorganic Chemistry Laboratory, Department of Chemistry, Bar Ilan University,
Ramat Gan 52900, Israel
Received J anuary 18, 2000
Mitsunobu reaction on the glucose derivative (3S,4R,5R,6R)-3,4,5,7-tetrabenzyloxy-6-hydroxy-1-
heptene yielded an unexpected rearrangement major product. Its structure was determined as
(3R,4R,5R,6S)-4,5,6,7-tetrabenzyloxy-3-hydroxy-1-heptene. The suggested rearrangement mecha-
nism involves an initial intramolecular cyclization, followed by ring opening by the nucleophile
p-nitrobenzoate. Product distribution of the Mitsunobu reaction was substrate-dependent, with
the corresponding mannose derivative (the 3R epimer) giving less of the initial intramolecular
reaction products and the corresponding galactose derivative (the 5S epimer) yielding almost
exclusively the expected epimerization product. Varying the Mitsunobu reaction conditions (addition
of base and using nonpolar solvent) led to the expected epimerization product of the glucose
derivative.
Sch em e 1
In tr od u ction
The Mitsunobu reaction is a useful way of activating
alcohols for substitution reactions and for inversion of
their stereochemistry (in the case of secondary alcohols).1
The latter reaction is based on displacement of the
activated hydroxyl group by a carboxylic acid, thus
forming an ester of opposite absolute configuration. This
is then followed by solvolysis of the ester, yielding an
alcohol of the same chemical structure and opposite
configuration. As part of our interest in the development
of inhibitors and mechanistic probes for sugar-metaboliz-
ing enzymes,2 we were involved in the synthesis of cyclic
sugar analogues. In an attempt to utilize the Mitsunobu
reaction to epimerize a secondary alcohol of a sugar
derivative bearing benzyl protecting groups on the rest
of the hydroxyls, an unexpected rearrangement product
was isolated as the major product. Here, we describe the
reaction, its stereochemical and mechanistic analysis,
and its dependence on the substrate and the reaction
conditions.
Ta ble 1. Mitsu n obu Rea ction ; P r od u ct Distr ibu tion a s a
F u n ction of Su bstr a te
product distribution
sugar derivative
2 (cycl)
3 (epi)
4 (rearr)
gluco-1a
manno-1b
galacto-1c
1
0-4
2
2.5
9
10
1
1
product 3a , and a rearrangement product 4a (of uniden-
tified stereochemistry) in a 1:2:10 molar ratio (Scheme 1
and Table 1). The cyclization product 2a was identical
to that obtained by Tf2O activation of 1a .3
Resu lts
(3S,4R,5R,6R)-3,4,5,7-Tetrabenzyloxy-6-hydroxy-1-hep-
tene 1a (glucose derivative)3 was subjected to standard
Mitsunobu reaction conditions (entry 1, Table 2)4 to
invert the stereochemistry of its C6 (free hydroxy-bearing
carbon) configuration. The reaction, stopped at the p-
nitrobenzoate stage (prior to the solvolysis final step),
yielded a mixture of three products. These were identified
as a cyclization product 2a , the expected epimerization
The product distribution of the same reaction on other
sugar analogues was examined (Table 1). The corre-
sponding mannose derivative 1b (3R epimer) yielded
more of the expected epimerization product 3b than
rearrangement product 4b (2.5:1 molar ratio) and vari-
able amounts of cyclization product 2b (0-40% of total
products). The corresponding galactose derivative 1c (5S
epimer) gave primarily the desired epimerization product
3c and the rearrangement 4c product only as a minor
side-product (9:1 molar ratio).
* Tel: 972-3-5318862. Fax: 972-3-5351250. E-mail: albecka@
mail.biu.ac.il.
(1) Mitsunobu, O. Synthesis 1981, 1, 1-28.
(2) (a) Hassner, A.; Falb, E.; Nudleman, A.; Albeck, A.; Gottlieb, H.
Tetrahedron Lett. 1994, 35, 2397-2400. (b) Falb, E.; Bechor, Y.;
Nudelman, A.; Hassner, A.; Albeck, A.; Gottlieb, H. E. J . Org. Chem.
1999, 64, 498-506. (c) Persky, R.; Albeck, A., submitted for publication.
(3) Martin, O. R.; Yang, F.; Xie, F. Tetrahedron Lett. 1995, 36, 47-
50.
The Mitsunobu reaction on the glucose derivative 1a
was studied to maximize the epimerization product at
the expense of the unexpected rearrangement product
(Table 2). Changing the solvent and reactant ratio finally
led to a 4:1 molar ratio in favor of the desired epimer-
ization product 4a . Under the same reaction conditions,
(4) Dodge, J . A.; Trujillo, J . I.; Presnell, M. J . Org. Chem. 1994, 59,
234-236.
10.1021/jo0000675 CCC: $19.00 © 2000 American Chemical Society
Published on Web 05/23/2000

3776 J . Org. Chem., Vol. 65, No. 12, 2000
Persky and Albeck
Ta ble 2. Mitsu n obu Rea ction ; P r od u ct Distr ibu tion a s a
F u n ction of Rea ction Con d ition s
product ratio
reactant molar ratio
alcohola Ph3P RCO2Hb Et3N DEAD solvent (epi) (rearr)
3a
4a
1
1
1
1
1
4
2
2
2
2
4
2
10
2
4
2
2
2
2
THF
THF
THF
toluene
toluene
1
2
1
4
1
5c
1
2
1
1
5
25
5
10
25
a
b
Glucose derivative 1a . p-Nitrobenzoic acid. c Minor cycliza-
tion product was also obtained (see Table 1).
F igu r e 2. NOESY data for cyclic olefins 2 and aldehydes 5.
(Arrows indicate weak (w) or medium (m) interactions (cross-
peaks in the 2-D spectrum) between sets of two protons).
F igu r e 1. Determination by NMR analysis of the structure
of the major product of Mitsunobu reaction on 1a .
the mannose derivative 1b yielded 3:1 epimerization:
rearrangement products, while the cyclization side-
product was not formed at all. The total yield was also
improved significantly, from 37% to 67% for the glucose
derivative and from 60% to 95% for the mannose deriva-
tive.
Discu ssion
F igu r e 3. Retroanalysis of the stereochemistry of rearrange-
ment product 4a .
P r od u ct An a lysis. The Mitsunobu reaction on 1a
yielded a major product that seemed, at first glance, to
be the expected epimerization product (at the p-nitroben-
zoate stage); it had the correct mass (by MS analysis),
and its 1H and 13C NMR spectra exhibited all of the
expected resonances (with the expected integrations).
Sch em e 2
1
Nevertheless, our curiosity was piqued by a H resonance
at 5.62 ppm, which seemed somewhat too low for the C6H
ester proton. Indeed, further analysis identified the
product as a rearrangement product on the basis of the
following NMR data (Figure 1):
1
(1) H COSY experiment showed that the 5.62 ppm
the analogous mannose derivative 1b, and its C6 epimer
1d (obtained by hydrolysis of the Mitsunobu epimeriza-
tion product 3b) (Scheme 3). The 1H and 13C NMR spectra
of the cyclization product derived from 1e were identical
to those of the cyclization product 2a , derived from the
glucose derivative 1a , confirming their identity.
The stereochemistry of cyclic compounds 2a ,b,d was
confirmed by NOESY experiments of the cyclic olefins
themselves and of their corresponding aldehydes 5a ,b,d
(Figure 2). The latter were obtained by OsO4, N-methyl
morpholine-N-oxide followed by NaIO4 oxidation.5 Ret-
roanalysis of the stereochemistry of the rearrangement
product 4a , based on the identification of its (Tf2O-
activated) cyclization product as 2a , revealed that it
underwent epimerization at both of the positions in-
volved, C3 and C6 (Figure 3).
resonance belongs to the allylic proton (C3H).
(2) 1H-1H long-range COSY experiment revealed a
cross-peak (interaction) between the allylic proton (C3H)
at 5.62 ppm and the aromatic protons of the p-nitroben-
zoate at 7.9-8.0 ppm.
1
(3) H-13C long-range HCOSY experiment exhibited
an interaction between the vinylic proton (C2H) at 6.00
ppm and the ester carbon at 163.3 ppm.
This set of experiments confirmed that the p-nitroben-
zoate ester is on C3, whereas C6 bears a benzyl ether
substituent. The absolute configuration of the two car-
bons involved in the rearrangement was not determined
at this stage, nor was the mechanism of this rearrange-
ment understood.
To determine the stereochemistry of the rearrange-
ment product, the benzoate ester was hydrolyzed, and
the resulting product 1e was cyclized by activation of the
hydroxy group as a triflate (Scheme 2).3 This cyclization
product was compared with the corresponding cyclization
products (2a ,b,d ) of the original glucose derivative 1a ,
Mech a n ism . We considered three possible mecha-
nisms for the rearrangement observed during the Mit-
(5) Barrett, A. G. M.; Edmunds, J . J .; Hendrix, J . A.; Malecha, J .
W.; Parkinson, C. J . J . Chem. Soc., Chem. Commun. 1992, 1240-1242.

Rearrangement during Mitsunobu Epimerization
J . Org. Chem., Vol. 65, No. 12, 2000 3777
F igu r e 4. Possible mechanisms for the formation of the rearrangement product 4a during the Mitsunobu reaction on 1a .
Sch em e 3
sunobu reaction (Figure 4). The first mechanism involves
a concerted nucleophilic attack of the two oxygens (the
C3 benzylic oxygen and the C6 phosphinoxy) on their
counterpart “skeleton” carbons via a bicyclic transition
state (Figure 4A). This is an “adjacent” substitution
reaction (well-known in phosphorus chemistry6), leading
to exchange of the two substituents (at C3 and C6) with
retention of configuration at both stereogenic centers.
Displacement of the triphenylphosphinoxy substitution
at C3 by p-nitrobenzoate completes the rearrangement
reaction with an overall inversion of the configuration
at C3 and retention of configuration at C6.
The second mechanism (Figure 4B) is based on
nucleophilic attacks on the substituents rather than the
skeleton carbons. Thus, the C3 benzylic oxygen attacks
the C6 phosphorus, releasing a free alkoxide that can
attack the C3 benzylic carbon. This mechanism leaves the
C3 and C6 oxygens in their original positions, with full
retention of configuration. Therefore, completion of the
reaction by p-nitrobenzoate displacement of the triph-
enylphosphinoxy leads to an overall inversion of config-
uration at C3 and retention of configuration at the C6
position.
These two mechanisms were ruled out since their
stereochemical consequence does not match the observed
stereochemistry of the rearrangement product 4a .
Finally, the third mechanism (Figure 4C) involves an
initial attack of the C3 benzylic oxygen on the C6 carbon,
bearing the activated alcohol. This reaction is similar to
the cyclization reaction, driven by triflate-activated
alcohol, previously described by Martin and co-workers.3
The common cyclic oxonium intermediate thus formed
can interact with a common nucleophile, p-nitrobenzoate,
in three different ways. The nucleophilic p-nitrobenzoate
can attack the benzyl ring substituent, yielding the
cyclization product 2 (path a in Figure 4C). Alternatively,
it can attack at the (original) C6 position, yielding the
corresponding ester that, after solvolysis, will afford the
starting material (path b in Figure 4C). The third
possibility involves a similar attack at the C3 position,
yielding the rearrangement product 4 with inversion of
configuration at both C3 and C6 positions (path c in Figure
4C). This stereochemistry coincides with that determined
for the rearrangement product 4a .
Therefore, only the third mechanism (Figure 4C)
explains the formation of the observed rearrangement
product 4a . This mechanism also accounts for the ob-
served cyclization side-product 2a . It should be noted that
two of the alternative routes of the third mechanism, path
(6) (a) Knowles, J . Annu. Rev. Biochem. 1980, 49, 877-919. (b)
Benkovic, S. J .; Schray, K. J . In Transition States of Biochemical
Processes; Plenum Press: New York, 1978; p 493-527.

3778 J . Org. Chem., Vol. 65, No. 12, 2000
Persky and Albeck
a and path c, involve nucleophilic attack at activated
carbons (benzylic and allylic, respectively). Therefore they
are much more favorable than path b. Indeed, the former
two products were observed, in addition to the epimer-
ization product, while the path b product was not
observed in any of the sugar derivative Mitsunobu
reactions.
Su bstr a te Dep en d en ce. The outcome of the Mit-
sunobu reaction as a function of small variations of the
substrate was studied (Table 1). A majority of 85% of the
total product of the Mitsunobu reaction on the glucose
derivative originated from the initial cyclization reaction
(Figure 4C). Varying amounts of 25-70% of the product
arose from this cyclization reaction when the correspond-
ing mannose derivative was the substrate of the Mit-
sunobu reaction. Finally, only 10% of such products were
observed for the galactose derivative. This can be ex-
plained on steric hindrance grounds. The initial cycliza-
tion reaction is strongly suppressed in reaction of the
galactose analogue 1c by steric repulsion between the C5
benzyloxy and the C6 triphenyl phosphinoxy substituents
when properly oriented for displacement by the C3
benzyloxy oxygen. This interaction is unique to the
galactose derivative 1c only, as a result of its 5S absolute
configuration (in contrast to the 5R configuration of the
glucose and the mannose derivatives, 1a and 1b, respec-
tively).
Rea ction Con d ition Dep en d en ce. The distribution
of rearrangement and epimerization products depends
on the kinetics of intramolecular and intermolecular
nucleophilic attacks on the 6-phosphinoxy intermediate,
respectively. Initial attempts to increase the rate of the
intermolecular reaction by the addition of 20 equiv of the
carboxylic acid stopped all reactions entirely, and only
starting material was isolated. This was probably due
to diminishing even the small concentration of alkoxide
necessary for the Mitsunobu reaction to proceed (step 3,
Figure 5). Further analysis suggested addition of a base
to increase the concentration of the nucleophilic carboxy-
late to accelerate the intermolecular nucleophilic attack
(step 4, Figure 5).7 This was tested for the Mitsunobu
reaction on the glucose derivative 1a . The results pre-
sented in Table 2 show that this indeed helped to improve
the ratio of epimerization to rearrangement products.
Hydrophobic solvent was also superior to polar solvent,
probably because of better stabilization by the latter of
the oxonium intermediate in the intramolecular initial
cyclization reaction leading to the rearrangement prod-
uct. The conditions affording the best epimerization:
rearrangement ratio (entry 4, Table 2) also significantly
improved the overall yield of the reaction.
F igu r e 5. The Mitsunobu reaction mechanism.
of benzyl ether oxygen in nucleophilic cyclization reac-
tions, either as a desired reaction or as a side reaction,
is well documented.3,8 In the present study we have
demonstrated that, by varying the reaction conditions,
we can shift from an intramolecular reaction, leading to
the rearrangement product, to an intermolecular one,
yielding the Mitsunobu epimerization product.
Exp er im en ta l Section
Gen er a l. 1H and 13C NMR spectra were recorded at 600
and 150 MHz, respectively, in CDCl3, with TMS as an internal
standard. Multiplicity “d” refers to a doubletlike second-order
1
peak. H NMR assignments were supported by COSY experi-
ments, and 13C NMR assignments were supported by hetero
COSY (HMQC and HMBC) experiments. Mass spectra were
recorded in DCI mode with methane as the reagent gas. TLC
was performed on E. Merck 0.2 mm precoated silica gel F-254
plates and viewed by UV, vanillin.9 Chromatography refers
to flash column chromatography,10 carried out on silica gel 60
Con clu sion s
In this work, sugar derivative 1a was exposed to
standard Mitsunobu epimerization reaction conditions.
Surprisingly, the major product proved to be the result
of a rearrangement involving substituents at the C3 and
C6 positions. The suggested rearrangement mechanism
implicates an initial intramolecular displacement of the
C6 activated hydroxyl by the C3 benzyl ether oxygen. The
common cyclic intermediate can further react in different
ways to give the observed products. Such a participation
(8) (a) White, D.; Zemribo, R.; Mead, K. T. Tetrahedron Lett. 1997,
38, 2223-2226. (b) Reed, L. A., III; Huang, J . T.; McGregor, M.;
Goodman, L. Carbohydr. Res. 1994, 254, 133-140. (c) Dehmlow, H.;
Mulzer, J .; Seilz, C.; Strecker, A. R.; Kohlmann, A. Tetrahedron Lett.
1992, 33, 3607-3610. (d) Ermert, P.; Vasella, A. Helv. Chim. Acta 1991,
74, 2043-2053. (e) Mootoo, D. R.; Fraser-Reid, B. J . Chem. Soc., Chem.
Commun. 1986, 1570-1571. (f) Gray, G. R.; Hartman, F. C.; Barker,
R. J . Org. Chem. 1965, 30, 2020-2024.
(9) Stahl, E. Thin-Layer Chromatography, 2nd ed.; Springer-Ver-
lag: Berlin, 1969; p 217-218.
(10) Still, W. C.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923-
2925.
(7) Campbell, D. A.; Bermak, J . C. J . Org. Chem. 1994, 59, 658-
660.

Rearrangement during Mitsunobu Epimerization
J . Org. Chem., Vol. 65, No. 12, 2000 3779
(230-400 mesh ASTM, E. Merck). Anhydrous solvents were
dried and freshly distilled (THF and toluene from sodium/
benzophenone, pyridine and triethylamine from CaH2, CH2-
Cl2 from CaCl2, and DMF from 4 Å molecular sieves).
The linear olefins 1a -c were synthesized according to
Martin and co-workers3 and purified by chromatography
(eluted with either CH2Cl2 or ether/hexane 1:4).
3.743 (dd, J ) 5.7, 4.7 Hz, 1H, C4H), 4.094 (dd, J ) 8.0, 5.7
Hz, 1H, C3H), 4.112 (dd, J ) 5.6, 4.7 Hz, 1H, C5H), 4.237 (d,
J ) 11.7 Hz, 1H, CH2Ph(C3)), 4.362 (d, J ) 12.3 Hz, 1H, CH2-
Ph(C7)), 4.469 (d, J ) 12.1 Hz, 1H, CH2Ph(C7)), 4.542 (d, J )
11.3 Hz, 1H, CH2Ph(C4)), 4.565 (d, J ) 11.6 Hz, 1H, CH2Ph-
(C5)), 4.597 (d, J ) 11.8 Hz, 1H, CH2Ph(C3)), 4.697 (d, J )
11.3 Hz, 1H, CH2Ph(C5)), 4.794 (d, J ) 11.3 Hz, 1H, CH2Ph-
(C4)), 5.410 (dt, J ) 10.4, 1 Hz, 1H, C1H2), 5.417 (ddd, J )
17.1, 1.6, 0.9 Hz, 1H, C1H2), 5.580 (td, J ) 5.7, 4.2 Hz, 1H,
C6H), 5.975 (ddd, J ) 17.1, 10.5, 8.1 Hz, 1H, C2H), 7.177-
7.287 (m, 20H, Ph), 8.097 (“d”, J ) 8.9 Hz, 2H, Ph-NO2), 8.212
(“d”, J ) 8.9 Hz, 2H, Ph-NO2); 13C NMR δ 68.14 (C7), 69.92
(CH2Ph(C3)), 72.96 (CH2Ph(C7)), 73.70 (CH2Ph(C4)), 74.33
(C6), 75.0 (CH2Ph(C5)), 77.67 (C5), 80.38 (C3), 80.63 (C4),
120.08 (C1), 123.35, 127.43-128.31, 130.76 (Ph), 135.47 (C2),
135.49, 138.14, 138.19, 138.24, 138.27, 150.40 (Ph), 164.07
(CO2).
The cyclic olefins 2a ,b,d were obtained by Tf2O activation
of the corresponding linear olefins 1a ,b,d , respectively, ac-
cording to Martin and co-workers3 and purified by chroma-
tography (eluted with either CH2Cl2 or ether/hexane 1:4).
Mitsu n obu Rea ction . (a ) The olefin 1 (2.69 g, 5 mmol),
Ph3P (5.24 g, 20 mmol), and p-nitrobenzoic acid (3.34 g, 20
mmol) were dissolved in dry THF at 0 °C. DEAD (3.15 mL, 20
mmol) was then added dropwise, and after 10 min the reaction
mixture was allowed to warm to room temperature. After 5-18
h, the solvent was evaporated, and CH2Cl2 was added and
extracted consecutively with 1 M HCl, saturated NaHCO3, and
H2O. The organic phase was dried over MgSO4, filtered, and
evaporated to dryness. The product was chromatographed
(ether/hexane 1:5), yielding a mixture of cyclization (2),
epimerization (3), and rearrangement (4) products in overall
37%, 60%, and 40% yield for the glucose (2a -4a ), mannose
(2b-4b), and galactose (2c-4c) derivatives, respectively. (See
Table 1 for product distribution).
(b) The olefin 1 (2.69 g, 5 mmol), Ph3P (2.62 g, 10 mmol),
p-nitrobenzoic acid (1.67 or 8.35 g, 10 or 50 mmol), and Et3N
(3.5 or 17.5 mL, 25 or 125 mmol) were dissolved in dry THF
or dry toluene at 0 °C. DEAD (1.6 mL, 10 mmol) was then
added dropwise, and after 10 min the reaction mixture was
allowed to warm to room temperature. After 5-18 h, the
solvent was evaporated, and the reaction mixture worked up
as above, affording the glucose derivative products in 67% yield
and the mannose derivatives in 95% yield. (See Table 2 for
conditions and product distribution).
1
Ma n n o-r ea r r a n ged ester 4b: H NMR δ 3.593 (dd, J )
10.2, 5.2 Hz, 1H, C7H2), 3.675 (dd, J ) 10.2, 4.4 Hz, 1H, C7H2),
3.813 (q, J ) 4.7 Hz, 1H, C6H), 3.854 (t, J ) 5 Hz, 1H, C5H),
3.927 (t, J ) 5.3 Hz, 1H, C4H), 4.376 (d, J ) 12 Hz, 1H, CH2-
Ph(C7/C5)), 4.382 (d, J ) 12 Hz, 1H, CH2Ph(C7/C5)), 4.532
(d, J ) 11.9 Hz, 1H, CH2Ph(C6)), 4.657 (d, J ) 11.1 Hz, 2H,
CH2Ph(C4,C7/C5)), 4.709 (d, J ) 11.4 Hz, 1H, CH2Ph(C5/C7)),
4.722 (d, J ) 11.4 Hz, 1H, CH2Ph(C4)), 4.728 (d, J ) 11.8 Hz,
1H, CH2Ph(C6)), 5.211 (bt, J ) 10.5 Hz, 1H, C1H2), 5.229 (bd,
J ) 17.3, 1H, C1H2), 5.663 (t, J ) 5.8 Hz, 1H, C3H), 5.869
(ddd, J ) 17.2, 10.5, 6.5 Hz, 1H, C2H), 7.177-7.287 (m, 20H,
Ph), 8.101 (“d”, J ) 8.9 Hz, 2H, Ph-NO2), 8.176 (“d”, J ) 8.9
Hz, 2H, Ph-NO2); 13C NMR δ 69.31 (C7), 69.92 (CH2Ph(C3)),
72.65 (CH2Ph(C7/C5)), 73.16 (CH2Ph(C5/C7)), 74.40 (C4), 74.56
(CH2Ph(C6)), 76.14 (C3), 77.48 (C6), 78.03 (C5), 80.06 (C4),
118.59 (C1), 123.39, 127.56-128.24, 130.60 (Ph), 133.04 (C2),
135.49, 137.75-138.11, 150.40 (Ph), 163.61 (CO2).
HRMS (for the 3b:4b mixture) calcd for C42H42NO8 (MH+)
688.2910, found 688.2960; calcd for C35H34NO7 (MH+ - BnOH)
580.2335, found 580.2290.
Glu co-ep im er ized ester 3a : 1H NMR δ 3.504 (dd, J ) 10.6,
5.9 Hz, 1H, C7H2), 3.554 (dd, J ) 10.5, 3.7 Hz, 1H, C7H2),
3.558 (t, J ) 5.1 Hz, 1H, C4H), 4.031 (t, J ) 5.2 Hz, 1H, C5H),
4.044 (dd, J ) 7.4, 5.0 Hz, 1H, C3H), 4.241 (d, J ) 12.0 Hz,
1H, CH2Ph (C7)), 4.308 (d, J ) 12.0 Hz, 1H, CH2Ph (C7)), 4.517
(d, J ) 11.8 Hz, 1H, CH2Ph (C5)), 4.550 (d, J ) 11.7 Hz, 1H,
CH2Ph (C6)), 4.580 (d, J ) 11.6 Hz, 1H, CH2Ph (C3)), 4.631
(d, J ) 11.6 Hz, 1H, CH2Ph (C3)), 4.638 (d, J ) 11.6 Hz, 1H
CH2Ph (C6)), 4.658 (d, J ) 12.0 Hz, 1H, CH2Ph (C5)), 5.177
(ddd, J ) 17.2, 1.7, 0.9 Hz, 1H, C1H2), 5.191 (ddd, J ) 10.6,
1.7, 0.9 Hz, 1H, C1H2), 5.441 (q, J ) 5.1 Hz, 1H, C6H), 5.789
(ddd, J ) 17.2, 10.5, 7.4 Hz, 1H, C2H), 7.06-7.19 (m, 20H,
Ph), 7.981 (“d”, J ) 9.0 Hz, 2H, Ph-NO2), 8.072 (“d”, J ) 9.0
Hz, 2H, Ph-NO2); 13C NMR δ 68.12 (C7), 72.86 (CH2Ph(C7)),
74.08 (CH2Ph(C3)), 74.23 (CH2Ph(C6), C6), 74.66 (CH2Ph(C5)),
77.30 (C5), 80.22 (C3), 80.54 (C4), 119.00 (C1), 123.29, 127.43-
128.21, 130.65 (Ph), 135.02 (C2), 135.41, 137.72-138.22,
150.32 (Ph), 163.96 (CO2).
Ga la cto-ep im er ized ester 3c: 1H NMR δ 3.736 (t, J ) 5.4
Hz, 1H, C4H), 3.886 (dd, J ) 11.1, 6.7 Hz, 1H, C7H2), 3.914
(dd, J ) 11.2, 3.5 Hz, 1H, C7H2), 4.069 (dd, J ) 5.6, 3.5 Hz,
1H, C5H), 4.120 (dd, J ) 7.7, 5.1 Hz, 1H, C3H), 4.329 (d, J )
11.9 Hz, 1H, CH2Ph(C3)), 4.382 (d, J ) 12.1 Hz, 1H, CH2Ph-
(C7)), 4.459 (d, J ) 11.3 Hz, 1H, CH2Ph(C5)), 4.493 (d, J )
12.0 Hz, 1H, CH2Ph(C7)), 4.593 (d, J ) 11.7 Hz, 1H, CH2Ph-
(C5)), 4.613 (d, J ) 12.3 Hz, 1H, CH2Ph(C3)), 4.720 (d, J )
11.4 Hz, 1H, CH2Ph(C4)), 4.790 (d,J ) 11.4 Hz, 1H, CH2Ph-
(C4)), 5.336 (ddd, J ) 10.4, 1.6, 0.7 Hz, 1H, C1H2), 5.383 (ddd,
J ) 17.3, 1.5, 1.1 Hz, 1H, C1H2), 5.800 (dt, J ) 6.8, 3.5 Hz,
1H, C6H), 5.919 (ddd, J ) 17.4, 10.4, 7.6 Hz, 1H, C2H), 7.22-
7.32 (m, 20H, Ph), 8.069 (“d”, J ) 8.9 Hz, 2H, Ph-NO2), 8.227
(“d”, J ) 8.9 Hz, 2H, Ph-NO2); 13C NMR δ 68.66 (C7), 70.48
(CH2Ph(C3)), 72.87 (CH2Ph(C7)), 73.14 (CH2Ph(C5)),74.47
(CH2Ph(C4)), 74.63 (C6), 78.69 (C5), 80.66 (C3), 81.60 (C4),
119.16 (C1), 123.43, 127.47-128.33, 130.70 (Ph), 135.55 (C2),
137.92-138.22, 150.41 (Ph), 163.78 (CO2).
Glu co-r ea r r a n ged ester 4a : 1H NMR δ 3.549 (dd, J )
10.1, 5.4 Hz, 1H, C7H2), 3.593 (dd, J ) 10.1, 4.3 Hz, 1H, C7H2),
3.696 (dd, J ) 6.4, 4.5 Hz, 1H, C5H), 3.748 (dt, J ) 5.4, 4.5
Hz, 1H, C6H), 3.913 (dd, J ) 6.4, 3.8 Hz, 1H, C4H), 4.226 (d,
J ) 11.7 Hz, 1H, CH2Ph (C4)), 4.300 (d, J ) 11.9 Hz, 1H, CH2-
Ph (C7)), 4.329 (d, J ) 12.0 Hz, 1H, CH2Ph (C7)), 4.485 (d, J
) 11.7 Hz, 1H, CH2Ph (C4)), 4.509 (d, J ) 11.8 Hz, 1H, CH2-
Ph (C6)), 4.530 (d, J ) 12.0 Hz, 1H, CH2Ph (C5)), 4.575 (d, J
) 11.3 Hz, 1H, CH2Ph (C5)), 4.643 (d, J ) 11.6 Hz, 1H, CH2-
Ph (C6)), 5.248 (dt, J ) 10.5, 1.1 Hz, 1H, C1H2), 5.286 (dt, J
) 17.3, 1.2 Hz, 1H, C1H2), 5.616 (ddt, J ) 6.9, 3.8, 1.0 Hz,
1H, C3H), 6.001 (ddd, J ) 17.4, 10.5, 6.9 Hz, 1H, C2H), 7.06-
7.19 (m, 20H, Ph), 7.890 (“d”, J ) 9.0 Hz, 2H, Ph-NO2), 8.009
(“d”, J ) 9.0 Hz, 2H, Ph-NO2); 13C NMR δ 69.49 (C7), 70.60
(CH2Ph(C4)), 72.71 (CH2Ph(C6)), 73.19 (CH2Ph(C7)), 74.52
(CH2Ph(C5)), 76.26 (C3), 77.69 (C6), 78.36 (C5), 80.23 (C4),
119.78 (C1), 123.07, 127.43-128.21, 130.59 (Ph), 132.40 (C2),
135.41, 137.72-138.22, 150.32 (Ph), 163.28 (CO2).
Ga la cto-r ea r r a n ged ester 4c: 1H NMR δ 5.272 (dt, J )
17.3, 1.3 Hz, 1H, C1H2), 5.303 (dt, J ) 10.5, 1.2 Hz, 1H, C1H2),
5.933 (ddt, J ) 7.2, 2.8, 1.0 Hz, 1H, C3H), 6.052 (ddd, J )
17.2, 10.5, 7.2 Hz, 1H, C2H); 13C NMR δ 77.46 (C3), 119.80
(C1).
Meth a n olysis. To the ester (3 and 4, a mixture from the
Mitsunobu reaction) in methanol were added 10 equiv of
NaOH (solid). After 1 h, the solvent was partially removed
under reduced pressure. CH2Cl2 was added and extracted twice
with water. The organic phase was dried over MgSO4, filtered,
and evaporated to dryness. Chromatography (ether/hexane 1:3)
afforded the clean product (as a mixture of isomers derived
from 3 and 4).
Ma n n o-ep im er ized a lcoh ol 1d : 79% yield from 3b; 1H
NMR δ 3.389 (dd, J ) 9.6, 6.0 Hz, 1H, C7H2), 3.429 (dd, J )
9.5, 6.2 Hz, 1H, C7H2), 3.721 (dd, J ) 6.1, 3.0 Hz, 1H, C5H),
3.831 (dd, J ) 6.1, 5.1 Hz, 1H, C4H), 3.921 (ddd, J ) 6.2, 6.0,
3.0 Hz, 1H, C6H), 4.036 (ddd, J ) 7.9, 5.0, 0.7 Hz, 1H, C3H),
4.306 (d, J ) 11.8 Hz, 1H, CH2Ph(C3)), 4.380 (d, J ) 11.9 Hz,
1H, CH2Ph(C7)), 4.423 (d, J ) 11.9 Hz, 1H, CH2Ph(C7)), 4.451
HRMS (for the 3a :4a mixture) calcd for C42H42NO8 (MH+)
688.2910, found 688.29.40.
1
Ma n n o-ep im er ized ester 3b: H NMR δ 3.568 (dd, J )
10.7, 5.7 Hz, 1H, C7H2), 3.683 (dd, J ) 10.7, 4.2 Hz, 1H, C7H2),

3780 J . Org. Chem., Vol. 65, No. 12, 2000
Persky and Albeck
(d, J ) 11.3 Hz, 1H, CH2Ph(C5)), 4.583 (d, J ) 11.1 Hz, 1H,
CH2Ph(C4)), 4.597 (d, J ) 11.8 Hz, 1H, CH2Ph(C3)), 4.684 (d,
J ) 11.1 Hz, 1H, CH2Ph(C5)), 4.731 (d, J ) 11.1 Hz, 1H, CH2-
Ph(C4)), 5.329 (ddd, J ) 17.5, 1.7, 0.8 Hz, 1H, C1H2), 5.354
(ddd, J ) 10.4, 1.6, 0.8 Hz, 1H, C1H2), 5.962 (ddd, J ) 17.5,
10.4, 7.9 Hz, 1H, C2H), 7.218-7.280 (m, 20H, Ph); 13C NMR
δ 70.14 (C6/CH2Ph(C3)), 70.17 (C6/CH2Ph(C3)), 71.21 (C7),
73.19 (CH2Ph(C7)), 74.32 (CH2Ph(C4)), 74.87 (CH2Ph(C5)),
78.64 (C5), 81.12 (C3), 81.75 (C4), 119.61 (C1), 127.48-129.56
(Ph), 135.53 (C2), 138.05-138.45 (Ph); HRMS calcd for C35H39O5
(MH+) 539.2798, found 539.2841; MS m/z 539 (MH+, 8), 431
(22), 323 (19), 91 (84).
(dd, J ) 4.8, 2.0 Hz, 1H, C2H), 4.533 (d, J ) 12.0 Hz, 1H,
CH2Ph(C6)), 4.575 (td, J ) 6.1, 3.6 Hz, 1H, C5H), 4.629 (d, J
) 12.0 Hz, 1H, CH2Ph(C6)), 7.2-7.4 (m, 15H, Ph), 9.673 (d, J
) 2.0 Hz, 1H, C1HO); 13C NMR δ 67.97 (C6), 72.35 (CH2Ph-
(C4)), 72.41 (CH2Ph(C3)), 73.53 (CH2Ph(C6)), 80.89 (C4), 81.06
(C5), 83.71 (C3), 84.72 (C2), 127.54-128.49, 136.87, 137.35,
137.97 (Ph), 201.88 (C1); HRMS calcd for C27H29O5 (MH+)
433.2015, found 433.2000; MS m/z 433 (MH+, 1), 341 (13), 107
(50), 91 (100).
Ma n n o-cyclic a ld eh yd e 5b: 48% yield;1H NMR δ 3.772
(dd, J ) 10, 5.5 Hz, 1H, C6H2), 3.791 (dd, J ) 10, 6.5 Hz, 1H,
C6H2), 3.973 (dd, J ) 3.4, 1.4 Hz, 1H, C4H), 4.194 (dd, J )
1.3, 1.1 Hz, 1H, C3H), 4.299 (d, J ) 11.8 Hz, 1H, CH2Ph(C4)),
4.368 (dd, J ) 1.4, 0.9 Hz, 1H, C2H), 4.443 (d, J ) 11.8 Hz,
1H, CH2Ph(C4)), 4.495 (ddd, J ) 6.5, 5.5, 3.4 Hz, 1H, C5H),
4.504 (d, J ) 11.7 Hz, 1H, CH2Ph(C6)), 4.532 (d, J ) 12.0 Hz,
1H, CH2Ph(C3)), 4.614 (d, J ) 12.0 Hz, 1H, CH2Ph(C3)), 4.614
(d, J ) 12.0 Hz, 1H, CH2Ph(C6)), 7.25-7.35 (m, 15H, Ph), 9.588
(d, J ) 0.9 Hz, 1H, C1HO); 13C NMR δ 68.28 (C6), 71.67 (CH2-
Ph(C4)), 71.76 (CH2Ph(C3)), 73.51 (CH2Ph(C6)), 80.19 (C4),
81.11 (C5), 84.87 (C3), 87.16 (C2), 127.62-128.49, 137.00,
137.06, 137.95 (Ph), 203.33 (C1); HRMS calcd for C27H29O5
(MH+) 433.2015, found 433.1978; MS m/z 433 (MH+, 2), 341
(51), 313 (54), 163 (80).
Glu co-r ea r r a n ged a lcoh ol 1e: 72% yield from 4a ; 1H
NMR δ 3.202 (d, J ) 5.2 Hz, 1H, OH), 3.522 (dd, J ) 10.3, 5.1
Hz, 1H, C7H2), 3.568 (dd, J ) 5.2, 4.1 Hz, 1H, C4H), 3.642
(dd, J ) 10.4, 3.7 Hz, 1H, C7H2), 3.884 (dd, J ) 5.4, 4.0 Hz,
1H, C5H), 3.907 (qd, J ) 5.2, 3.7 Hz, 1H, C6H), 4.35 (m, 1H,
C3H), 4.383 (d, J ) 12.1 Hz, 1H, CH2Ph(C7)), 4.433 (d, J )
12.2 Hz, 1H, CH2Ph(C7)), 4.475 (d, J ) 11.5 Hz, 1H, CH2Ph-
(C4)), 4.601 (d, J ) 11.7 Hz, 1H, CH2Ph(C6)), 4.625 (d, J )
11.2 Hz, 1H, CH2Ph(C5)), 4.642 (d, J ) 11.5 Hz, 1H, CH2Ph-
(C4)), 4.681 (d, J ) 11.2 Hz, 1H, CH2Ph(C5)), 4.717 (d, J )
11.7 Hz, 1H, CH2Ph(C6)), 5.206 (dt, J ) 10.5, 1.7 Hz, 1H,
C1H2), 5.348 (dt, J ) 17.2, 1.7 Hz, 1H, C1H2), 5.883 (ddd, J )
17.2, 10.6, 5.3 Hz, 1H, C2H), 7.248-7.31 (m, 20H, Ph); 13C
NMR δ 69.55 (C7), 71.94 (C3), 72.46 (CH2Ph(C5)), 72.72 (CH2-
Ph(C7)), 73.07 (CH2Ph(C6)), 74.04 (CH2Ph(C4)), 78.14 (C6),
78.65 (C5), 79.34 (C4), 116.04 (C1), 127.5-128.3 (Ph), 137.59
(C2), 137.7-138.2 (Ph); HRMS calcd for C35H39O5 (MH+)
539.2798, found 539.2800; MS m/z 539 (MH+, 35), 521 (20),
431 (22), 361 (16.5).
Oxid a tion . To the cyclic olefin (2a ,b,d ) (4.3 g, 10 mmol) in
H2O/THF (50 mL, 1:1 v/v) was added N-methyl morpholine-
N-oxide (2.34 g, 20 mmol), followed by the addition of a
catalytic amount (0.04 mmol) of OsO4. After 2-18 h, NaIO4
(5.8 g, 30 mmol) and methanol (50 mL) were added. After 2 h,
the organic solvents were removed under reduced pressure,
and the aqueous solution was extracted with CH2Cl2 (3 × 30
mL). This organic phase was dried over MgSO4, filtered, and
evaporated. The product was purified by chromatography,
using CH2Cl2 as eluent.
Ma n n o-ep im er ized cyclic a ld eh yd e 5d : 76% yield; 1H
NMR δ 3.585 (dd, J ) 9.9, 6.3 Hz, 1H, C7H2), 3.646 (dd, J )
9.9, 6.4 Hz, 1H, C7H2), 4.023 (dd, J ) 2.4, 1.7 Hz, 1H, C4H),
4.199 (t, J ) 1.8 Hz, 1H, C3H), 4.401 (td, J ) 6.4, 2.3 Hz, 1H,
C5H), 4.422 (d, J ) 11.8 Hz, 1H, CH2Ph), 4.453 (d, J ) 11.8
Hz, 1H, CH2Ph), 4.462 (dd, J ) 2.4, 0.9 Hz, 1H, C2H), 4.464
(d, J ) 11.8 Hz, 1H, CH2Ph), 4.523 (d, J ) 12.1 Hz, 1H, CH2-
Ph), 4.570 (d, J ) 11.8 Hz, 1H, CH2Ph), 4.586 (d, J ) 12.1 Hz,
1H, CH2Ph), 7.255-7.942 (m, 15H, Ph), 9.679 (d, J ) 1.0 Hz,
1H, C1HO); 13C NMR δ 69.87 (C6), 71.45 (CH2Ph), 71.90 (CH2-
Ph), 73.35 (CH2Ph), 82.60 (C4), 83.90 (C5), 84.75 (C3), 87.56
(C2), 127.70-128.52, 136, 137.11, 137.97 (Ph), 202.60 (C1);
HRMS calcd for C27H29O5 (MH+) 433.2015, found 433.2028.
Ack n ow led gm en t. This work was supported in part
by the “Marcus Center for Pharmaceutical and Medici-
nal Chemistry” at Bar Ilan University.
Glu co-cyclic a ld eh yd e 5a : 44% yield; 1H NMR δ 3.742 (dd,
J ) 10.8, 6.1 Hz, 1H, C6H2), 3.765 (dd, J ) 10.8, 6.1 Hz, 1H,
C6H2), 4.022 (dd, J ) 3.6, 1.3 Hz, 1H, C4H), 4.314 (dd, J )
4.8, 1.3 Hz, 1H, C3H), 4.372 (d, J ) 12.0 Hz, 1H, CH2Ph(C3)),
4.416 (d, J ) 12.1 Hz, 1H, CH2Ph(C4)), 4.432 (d, J ) 12.0 Hz,
1H, CH2Ph(C3)), 4.480 (d, J ) 12.1 Hz, 1H, CH2Ph(C4)), 4.487
Su p p or tin g In for m a tion Ava ila ble: Figures showing 1H
NMR and NOESY spectra of cyclic compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
J O0000675