Linear fused pyran-dioxane-cyclohexane tricycles: Synthesis of the five linkage isomers and ensuing reactions

Article abstract of DOI:10.1002/ejoc.200400457

A systematic investigation of the mono-O-glycosylation of meso-, (S,S)-, and (R,R)-cyclohexane-1,2-diol with the D-glucose-derived 2-ketohexosyl bromide 5 is presented. In each instance, simple Koenigs-Knorr conditions not only elicit the exclusive formation of the respective β-2-ketoglycosides, but also effect their subsequent intramolecular hemiketalization to provide tricycles with a 1,5,10-trioxa-perhydroanthracene framework. The linkage geometries of the products - two each from the meso- (→ 10, 12) and (S,S)-diols (→ 16, 17), and only one from the (R,R)-isomer (→ 26) - are determined in the hemiketalization step by an interplay of the anomeric effect and steric factors, favoring those isomers in which the pyran ring oxygen and the ketal-OH are in a trans-diaxial disposition. The most propitious case with respect to uniformity of reaction and yield (87%) turned out to the (R,R)-diol-derived cis-cisoid-trans-fused 26 as steric and stereoelectronic factors operate concertedly in the hemiketalization step. In each of these pyran-dioxane-cyclohexane tricycles, the acetalic hydroxyl group could be removed by BF₃-mediated reduction with triethylsilane, the ¹H NMR spectroscopic data of the respective products 14, 15, 18 and 28 being instrumental in assigning their linkage geometries. Slightly basic conditions (nBu₄NOAc in acetonitrile) elicit highly stereoselective rearrangements in the pyran ring, for example 16 → 23 and 26 → 30; the tricycles have the structurally and stereochemically correct framework of a variety of cardenolides, in which a 2-ketosugar is doubly linked to a steroidal aglycon-diol. Thus, the methodology elaborated here shows high promise to provide a first, preparatively satisfactory access to this type of cardiac glycosides as well as to spectinomycin type antibiotics. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2004.

Full text of DOI:10.1002/ejoc.200400457

FULL PAPER
Linear Fused Pyran-Dioxane-Cyclohexane Tricycles: Synthesis of the Five
Linkage Isomers and Ensuing Reactions[‡]
Frieder W. Lichtenthaler*[a] and Eckehard Cuny[a]
Keywords: Carbohydrates / Glycosylation / Ulosyl bromides / Trioxa-perhydroanthracenes
A systematic investigation of the mono-O-glycosylation of
meso-, (S,S)-, and (R,R)-cyclohexane-1,2-diol with the -glu-
step. In each of these pyran-dioxane-cyclohexane tricycles,
the acetalic hydroxyl group could be removed by BF3-medi-
D
cose-derived 2-ketohexosyl bromide 5 is presented. In each ated reduction with triethylsilane, the 1H NMR spectroscopic
instance, simple Königs−Knorr conditions not only elicit the
exclusive formation of the respective β-2-ketoglycosides, but
also effect their subsequent intramolecular hemiketalization
to provide tricycles with a 1,5,10-trioxa-perhydroanthracene
framework. The linkage geometries of the products − two
each from the meso- (Ǟ 10, 12) and (S,S)-diols (Ǟ 16, 17),
and only one from the (R,R)-isomer (Ǟ 26) − are determined
in the hemiketalization step by an interplay of the anomeric
effect and steric factors, favoring those isomers in which the
data of the respective products 14, 15, 18 and 28 being instru-
mental in assigning their linkage geometries. Slightly basic
conditions (nBu4NOAc in acetonitrile) elicit highly stereo-
selective rearrangements in the pyran ring, for example
16 Ǟ 23 and 26 Ǟ 30; the tricycles have the structurally and
stereochemically correct framework of a variety of carden-
olides, in which a 2-ketosugar is doubly linked to a steroidal
aglycon-diol. Thus, the methodology elaborated here shows
high promise to provide a first, preparatively satisfactory
pyran ring oxygen and the ketal-OH are in a trans-diaxial access to this type of cardiac glycosides as well as to spec-
disposition. The most propitious case with respect to uniform-
tinomycin type antibiotics.
ity of reaction and yield (87%) turned out to the (R,R)-diol-
derived cis-cisoid-trans-fused 26 as steric and stereoelec- (© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
tronic factors operate concertedly in the hemiketalization
Germany, 2004)
Introduction
Various genera of the milkweed family (Asclepiadaceae)
as well as Streptomyces spectabilis generate glycosides
in which the sugar portion Ϫ a 4,6-dideoxy-d-hexos-2,3-
diulose in 1 and 2 or its epimeric C-3 reduction products in
3 and 4, respectively Ϫ is attached to the cyclohexanoid
aglycon-diol by both a β-glycosidic bond and a hemiketal
linkage, thereby engendering an additional 1,4-dioxane
ring (Figure 1).
[‡]
Sugar-Derived Building Blocks, 36. Part 35: See ref.[7]
Clemens-Schöpf-Institut für Organische Chemie und
[a]
Figure 1. Naturally occurring glycosides in which the sugar
portion is attached to a cyclohexanoid diol aglycon by a glycosidic
and a hemiacetal bond thereby engendering a central dioxane
ring
Biochemie, Technische Universität Darmstadt,
Petersenstrasse 22, 64287 Darmstadt, Germany
Fax: (internat.) ϩ 49-6151-166674
Eur. J. Org. Chem. 2004, 4911Ϫ4920
DOI: 10.1002/ejoc.200400457
© 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4911

F. W. Lichtenthaler, E. Cuny
FULL PAPER
For linear fused pyran-dioxane-cyclohexane tricycles of first explore its feasibility with (S,S)-, meso- and (R,R)-1,2-
this type Ϫ they de facto comprise the framework of 1,5,10- cyclohexanediol as the aglycons for gaining insight into the
trioxa-perhydroanthracenes Ϫ five basic linkage geometries stereochemical subtleties of linear-fused pyran-dioxane-
are possible, yet in the natural products only the cis-cisoid- cyclohexane tricycles. The results are the subject of this re-
trans form[6] is realized (Figure 2). This is obviously due to port.
the fact that in the biosynthetic key steps, the 2-ketohexose
(or a precursor thereof) is β-glycosidated by one of the
Results and Discussion
cyclohexane hydroxy groups and then undergoes intra-
molecular hemiketalization with the other, which, due to
operation of the anomeric effect, exclusively occurs from
the upper face of the pyran ring to generate an axially dis-
posed tertiary OH.
meso-1,2-Cyclohexanediol: Glycosidation and
Hemiketalization
As the silvercarbonate-promoted glycosidation of ulosyl
bromides proceeds in an essentially β-specific manner,[11]
the reaction of the most readily accessible[11b] 5 with meso-
1,2-cyclohexanediol is expected to generate the dia-
stereomeric β-d-glycosiduloses 8 and 9 (Scheme 2), de-
pending on which of the OH groups is glycosylated. Each
of these glycosiduloses can then undergo cyclo-hemiketaliz-
ation in two stereochemically different ways: addition of the
hydroxyl onto the pyran carbonyl from the axial side (β-
face) resulting in products 10 (from intermediate 8) and 12
(from 9), each having the hemiacetal-OH generated in trans-
diaxial disposition to the pyranoid ring oxygen (Scheme 2);
the alternate possibility comprises attack of the cyclohex-
ane-OH to the pyranoid carbonyl from the equatorial side
(α-face), i.e. 8 Ǟ 11 and 9 Ǟ 13, both products carrying
the hemiacetal-OH in a gauche disposition to either of the
anomeric oxygens.
Figure 2. The linkage geometry in various cardiac glycosides and
in the antibiotic spectinomycin is invariably cis-cisoid-trans[7]
Some recent model experiments on the silver carbonate
promoted glycosidation of ulosyl bromide 5 with glycol[7]
have shown that this step not only proceeds in a β-specific
manner (5 Ǟ 6) but is spontaneously followed by an intra-
molecular hemiketalization from the upper face of the py-
ran ring (arrows in 6) to give the cis-annulated trioxadecalin
7, in which dipolar interactions between the pyranoid ring
oxygen and the ketal-OH are minimized by their trans-diax-
ial disposition (Scheme 1).
When performing the reaction, only two of the four theo-
retically possible β-linked pyran-dioxane-cyclohexane iso-
mers are obtained, namely 10 and 12, isolable in crystalline
form, and in yields of 38 and 35%, respectively. The 1H and
13C NMR spectroscopic data clearly show them to be py-
ran-dioxane-cyclohexane-fused products with the pyranoid
4
portion in a C1 chair geometry based on the large coup-
lings within the ring (Table 1). Their linkage geometries,
however, could be established by detailed NOE studies not
only on 10 and 12 but on the products obtained on re-
ductive removal of the hemiacetal-OH through brief expo-
sure to triethylsilane/BF3-etherate in dichloromethane
(Scheme 3), i.e. 14 (from 10) and 15 (from 12).
The linkage geometry assignments rest on the following
pieces of evidence: (i) 10 and 14 show distinct NOEs be-
tween the axially oriented pyranoid protons 2-H, 4-H and
10a-H (Scheme 3, left), yet none between the dioxane ring
hydrogen atoms; as the coupling constants J4,4a and J4a,10a
are both small (Table 1), pyran and dioxane rings are cis-
annulated, entailing the cis-transoid-cis fusion with all rings
in chair conformations; (ii) 12 and 14 exhibit the same
Scheme 1
As the molecular geometry of 7 corresponds to that of NOEs within the pyranoid ring, as expected (Scheme 3,
the pyrano-dioxane portion of the natural products 1Ϫ4, right), yet in 15 an additional one between 4a-H and 5a-H
this ‘‘ulosyl donor approach’’[8] offers the potential to pro- clearly indicates their cis-cisoid-cis linkage geometries Ϫ a
vide a straightforward methodology towards their total conclusion corroborated by the small J4,4a and J4a,10a coup-
syntheses, Ϫ especially, as none of the cardenolides has yet lings observed (Table 1). Additional validity for these as-
been synthesized and the strategies followed in two synth- signments is derived from the fact that the alternative prod-
eses of spectinomycin[9] are unsuited for general appli- ucts 11 and 13, respectively, should (upon triethylsilane-pro-
cation.[10] Before engaging in the total synthesis of the natu- moted removal of their acetalic OH) give J4,4 and J4a,10a
ral products by the ‘‘ulosyl donor approach’’, we opted to values in the range 9Ϫ10 Hz, which is clearly not the case.
4912
© 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 4911Ϫ4920

Linear Fused Pyran-Dioxane-Cyclohexane Tricycles
FULL PAPER
Scheme 2
The essentially complete stereoselection observed in the NMR), of which one crystallized well and was isolated in
reaction of ulosyl bromide 5 with meso-cyclohexanediol un- 44% yield. On the basis of its X-ray structural analysis (Fig-
doubtedly lies in the cyclo-hemiketalization step following ure 3), itwas confirmed to be the cis-transoid-trans-intercon-
the β-specific glycosidation: due to operation of the anom- nected tricycle 16 (Scheme 4). This linkage geometry
eric effect, the intermediate β-d-glycosiduloses 8 and 9 cycl- necessitates the central dioxane ring to adopt a skew-boat
ize such that the cyclohexane-OH Ǟ carbonyl addition oc- conformation, its unfavorable steric interactions obviously
curs from the β-face (axial side) of the pyran ring being overruled by minimization of the dipolar interactions
(Scheme 2) to exclusively generate the two products with through the trans-diaxial disposition of the pyranoid ring
the pyranoid ring oxygen and ketalic OH in a trans-diaxial oxygen and the ketalic OH group.
disposition. That each of the products can have the three
linear-fused six-membered rings in chair conformations
with the outer rings cis-annulated to the central 1,4-dioxane
is an additional stereochemical asset towards their forma-
tion.
The second product, isolable from the mother liquor by
chromatography, proved to be the trans-cisoid-trans-linked
isomer 17, obtained as a syrup only (33%). It appears to be
the energetically more stable isomer of the two, since 16, on
standing in chloroform at ambient temperature, undergoes
gradual equilibration to 17. After three weeks, a 5:1 mixture
had formed (1H NMR), from which the major product 17
could be isolated in 65% yield.
(S,S)-1,2-Cyclohexanediol
The all-trans-annulation of the three rings in 17 followed
Ag2CO3-promoted mono-glycosylation of (S,S)-1,2- Ϫ aside from being the only alternative to 16 Ϫ from a
cyclohexanediol with ulosyl bromide 5 under standard con- clear-cut NOE between 9a-H and 10a-H across the dioxane
ditions (CH2Cl2, 2 h, 40 °C) resulted in quantitative conver- ring, and the large coupling constants (J4,4a and J4a,10a) ob-
sion into an approximate 1:1-mixture of two products (1H served for the triethylsilane reduction product 18 (Table 1).
Eur. J. Org. Chem. 2004, 4911Ϫ4920
www.eurjoc.org
© 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4913

F. W. Lichtenthaler, E. Cuny
FULL PAPER
1
Table 1. Selected H NMR spectroscopic data (300 MHz, CDCl3) of the linear fused pyran-dioxane-cyclohexane isomers
Scheme 3
Figure 3. X-ray structure of the cis-transoid-trans-interconnected
That the trans-cisoid-trans fusion of the three rings in 17
is energetically more stable than the alternate cis-transoid-
trans geometry in 16, also finds its expression in the mono-
glycosylation of (S,S)-cyclohexanediol with the 3,4-unsatu-
tricycle 16[13] resulting from reaction of ulosyl bromide 5 with
(S,S)-1,2-cyclohexanediol; the chair conformations of the pyran
and cyclohexane rings are contrasted by the central dioxane
in
a
distorted boat form; selected torsional angles:
176.6, O1ϪC10aϪC4aϪO5 Ϫ63.4,
O1ϪC10aϪC4aϪO4a
rated analog of 5, the enolone bromide 19 (Scheme 5). Un- O5ϪC4aϪC10aϪO10 55.7, O5ϪC5aϪC9aϪO10 61.5
4914
© 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 4911Ϫ4920

Linear Fused Pyran-Dioxane-Cyclohexane Tricycles
FULL PAPER
detectable only in trace amounts in the reaction mixture
20 Ǟ 23 by TLC Ϫ could be prepared from 23 in 85% yield
by exposure to boron trifluoride in chloroform; the epimer-
ization at the C-4a ring junction occurring either through
BF3-promoted removal and subsequent re-addition of ben-
zoate, or by C-4aϪO-5 dioxane ring opening and closure.
This smooth epimerization also attests to the fact that the
all-trans-fusion of the three rings in 24 (or 17) is energeti-
cally more favored than the cis-transoid-trans arrangement
in 23 (or 16).
Scheme 4
der conditions analogous to those for the 5 Ǟ 16/17 conver-
sion (Ag2CO3/CH2Cl2, 2 h, room temp.), however, only β-
glycosidation took place, allowing the isolation of the en-
olone glycoside 20 in crystalline form (81%) with no other
product detectable in the reaction mixture (TLC, 1H
NMR). Thus, the steric predispositions in 20 for an ensuing
cyclo-hemiketalization are considerably impaired by the
half-chair conformation of the dihydropyranone relative to
the 4C1 chair geometries of the pyran rings in the analogous
uloside precursors, which spontaneously cyclized to 16 and
17. Only on extended standing of 20 in chloroform solution,
gradual cycloketalization to the trans-cisoid-trans-fused
Scheme 5
(R,R)-1,2-Cyclohexanediol
Unlike the reactions of ulosyl bromide 5 with meso- and
product 21 occurred; the conversion was essentially quanti- with (S,S)-cyclohexanediol, which gave two tricyclic prod-
tative after 30 days. ucts each due to opposing steric and polar effects in the
Cyclo-hemiketalization of enolone glycoside 20 could intramolecular hemi-acetalization step, (R,R)-cyclohexane-
also be effected by stirring with n-butylammonium acetate diol gave only one: Ag2CO3-promoted β-glycosidation gen-
in acetonitrile/chloroform (room temp., 15 h), leading to the erated the glycosidulose intermediate 25, which exclusively
cis-transoid-trans-fused 23, isolable in crystalline form in cyclized by OH Ǟ CϭO attack from the axial (upper) face
57% yield, by a remarkable series of successive reactions. In of the pyran ring (arrows in 25a) to elaborate the cis-cisoid-
analogy to similar transformations in enolone esters of this trans-fused product 26, isolable in 87% yield (Scheme 6).
type,[7,9a,15] this conversion is surmised to be initiated by Here, steric and electronic effects obviously augment each
base-induced elimination of benzoic acid from the terminal other since polar interactions are minimized by the trans-
positions to the dienone intermediate 22, followed by cis- diaxial disposition of the pyran ring oxygen and acetalic
hemiketalization and migration of the enolic benzoyl group OH Ϫ a consequence of the anomeric effect Ϫ and steric
to the acetalic oxygen (arrows in 22), thereby liberating the interactions are at a minimum due to chair conformations
C-4-carbonyl group with concomitant shifts of the two of the three rings. The alternative possibility, cycloketaliz-
double bonds. The degree of stereocontrol exercised in in- ation from the equatorial (lower) face of the pyran ring (ar-
termediate 22 is remarkable, as attack of the cyclohexane- rows in 25b), is not realized as this would lead to the trans-
OH on the carbonyl group occurs Ϫ obviously steered by transoid-trans interconnected 27, in which the central diox-
the anomeric effect Ϫ such that the acetalic OH and the ane ring is forced into a sterically unfavorable boat confor-
pyranoid ring oxygen come into a trans-diaxial disposition, mation, and the ketalic OH is in an a,e-arrangement to the
hence resulting in product 23. The alternative product 24 Ϫ pyran as well as dioxane ring oxygen.
Eur. J. Org. Chem. 2004, 4911Ϫ4920
www.eurjoc.org
© 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4915

F. W. Lichtenthaler, E. Cuny
FULL PAPER
Figure 4. X-ray structure of the cis-cisoid-trans-fused 29 mono-eth-
erate: due to the presence of a double bond, the pyran ring adopts
the typical half-chair geometry versus nearly perfect chair confor-
mations of the dioxane and cyclohexane parts. The acetalic 4a-OH
Scheme 6
and the pyranoid oxygen are in
a trans-diaxial disposition
(O1ϪC10aϪC4aϪO4a: 164.8°)[16]
The pronouncedly uniform course of the conversion
5 Ǟ 26 was paralleled by the Ag2CO3-promoted reaction of
enolone bromide 19 with (R,R)-cyclohexanediol: β-glycos-
idation and subsequent hemiacetal formation gave a single
product, which could be isolated as well-formed prisms in
a yield of 84%. It proved to be cis-cisoid-trans-fused 29 in
the form of a mono-etherate with the ether oxygen uniquely
hydrogen-bonded to the acetalic hydroxy group (Figure 4).
The enolic benzoyloxy group in 29, expectedly sensitive
towards mild basic conditions, was cleaved already by stir-
ring with tetrabutylammonium acetate in moist acetonitrile
(3 h, room temp.) to give the respective tricyclic ketone 30
(85%). Interestingly, the same product could be obtained
from the ulosyl bromide-derived 26 on exposure to the
same treatment, here hemiketal opening and elimination of
benzoic acid leading to 29 and then 30.
The cis-cisoid-trans linkage geometry of 26 could readily
be delineated from its 1H NMR spectroscopic data and
those of the triethylsilane reduction product 28 (Table 1).
The latter exhibits small, hence, e,a-couplings for the py-
ranoid ring hydrogens 4H/4a-H and 4a-H/10a-H versus
Scheme 7
large ones for the others, and shows a particularly relevant orientations relative to the dioxane ring. A characteristic
NOE across the dioxane ring (4a-H o 5-H). Further un- feature of the structure is the location of the ether molecule:
equivocal proof is provided by an X-ray structural analysis it is hydrogen bonded to the tertiary hydroxy group with
of 29, crystallizing as the mono-etherate (Figure 4): the a comparatively short distance between 4a-OH and OEt2
˚
junction of the pyran and dioxane rings is cis, with the ter- (1.995 A), thus providing a unique ‘‘horse-saddle’’ arrange-
tiary hydroxy group and the pyranoid ring O atom in axial ment.
4916
© 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 4911Ϫ4920

Linear Fused Pyran-Dioxane-Cyclohexane Tricycles
FULL PAPER
(2R,3R,4S,4aS,5aR,9aS,10aR)-3,4-Bis(benzoyloxy)-2-(benzoyloxy-
methyl)-4a-hydroxy-decahydro-2H-pyrano[2,3-b][1,4]benzodioxin
(cis-cisoid-cis Isomer) (12): The methanolic mother liquor remain-
ing after isolation of 10 was evaporated to dryness, the syrup was
applied to a silica gel column (2 ϫ 15 cm) and eluted with toluene/
EtOAc (4:1). The first fractions (Rf ϭ 0.58, toluene/EtOAc, 4:1)
contained residual 10, those eluted next 12 (Rf ϭ 0.37). Evapor-
ation of the respective eluates in vacuo, dissolution of the syrup in
a little methanol and standing for several days eventually resulted
Conclusion
The trioxa-perhydroanthracene framework of 26, 29 and
30 with the cis-cisoid-trans fusion of the three rings and
their substituent pattern (most notably in 30) has close
structural analogies to the natural products 1Ϫ4 with their
doubly attached sugar portions. The remarkable ease and
efficiency with which these tricyclic products are elaborated
from a simple 2-ketohexosyl donor renders this method- in crystallization of 2.10 g (35%) of 12 as colorless crystals. M.p.
140Ϫ142 °C. [α]2D0 ϭ Ϫ8.3 (c ϭ 0.7, CHCl3). 1H NMR (300 Hz,
CDCl3): δ ϭ 1.1Ϫ2.1 (m, 8 H), 3.73 (m, 1 H), 3.98 (ddd, J ϭ 2.8,
ology highly promising for the syntheses of these natural
products, inasmuch as this approach may be considered bi-
omimetic.[17] These prospects call for the elaboration of ulo-
3.1, 12.2 Hz, 1 H), 4.56 (m, 1 H), 4.92 (s, 1 H), 5.24 (d, J ϭ 9.9 Hz,
1 H), 5.36 (s, OH), 5.97 (t, J ϭ 9.9 Hz, 1 H) ppm. 13C NMR
syl donors Ϫ or their 3,4-unsaturated analogs Ϫ from 6-
(75.5 MHz, CDCl3): δ ϭ 20.3, 24.4, 29.5, 30.2 (4 CH2), 62.0
deoxy-d-glucose, and their use for β-selective mono-O-gly-
(CH2OBz), 65.6 (C-5a), 67.2 (C-2), 71.3 (C-3), 72.5 (C-9a), 79.6 (C-
cosylation of steroidal diols or of an N-blocked actinamine,
4), 92.5 (C-4a), 95.4 (C-10a), 165.1, 166.0, 168.6 (3 C6H5CO) ppm.
as the resulting β-d-glycosiduloses are expected to self-elab-
orate the stereoelectronically most favorable cis-cisoid-trans-
annulated cyclo-hemiketal scaffold realized in the natural
cardenolides and in spectinomycin. Efforts towards this end
are to be reported.
MS (FD, 15 mA): m/z (%) ϭ 588 (100) [Mϩ]. C33H32O10 (588.6):
calcd. C 67.33, H 5.48; found C 67.30, H 5.38.
(2R,3R,4S,4aS,5aS,9aR,10aR)-3,4-Bis(benzoyloxy)-2-(benzoyloxy-
methyl)-decahydro-2H-pyrano[2,3-b][1,4]benzodioxin (14): To a co-
oled (0 °C) solution of 10 (1.0 g, 1.7 mmol) in CH2Cl2 (80 mL)
was added triethylsilane (5.4 mL, 34 mmol) and BF3Ϫdiethyl ether
(4.3 mL, 34 mmol) consecutively, and the mixture was stirred for
30 min at 0 °C and then allowed to come to room temperature.
Dilution with CH2Cl2 (200 mL), washing with saturated NaHCO3
solution (50 mL) and water (2 ϫ 50 mL), drying (Na2SO4), and
evaporation to dryness in vacuo gave 0.79 g (81%) of 14 as colorless
crystals. M.p. 175Ϫ177 °C. [α]2D0 ϭ Ϫ15.1 (c ϭ 0.7, CHCl3). 1H
NMR: Table 1. 13C NMR (75.5 MHz, CDCl3): δ ϭ 19.5, 23.0, 24.0,
29.8 (4 CH2), 63.4 (CH2OBz), 64.9 (C-9a), 66.0 (C-4a), 66.3 (C-3),
72.4 (C-2), 72.8 (C-5a), 73.1 (C-4), 94.2 (C-10a), 165.4, 166.0, 166.2
(3 C6H5CO) ppm. C33H32O10 (572.6): calcd. C 69.22, H 5.63; found
C 68.98, H 5.57.
Experimental Section
General Methods: Melting points were determined with a Bock hot-
stage microscope and are uncorrected, optical rotations on a
PerkinϪElmer 241 polarimeter at 20 °C with a cell of 1 dm path
length; concentration (c) in g/100 mL and solvent are given in par-
entheses. 1H and 13C NMR spectra were recorded in the solvent
indicated at 300 and 75.5 MHz, respectively. Chemical shifts are
expressed in parts per million downfield from TMS. Mass spectra
were acquired on Varian MAT 311 and MAT 212 spectrometers,
microanalyses on a PerkinϪElmer 240 elemental analyzer. TLC
was performed on precoated Merck plastic sheets (0.2 mm silica
gel F254) with detection by UV (254 nm) and/or spraying with
H2SO4 (50%) and heating. Column and flash chromatography was
carried out on Fluka silica gel 60 (70Ϫ230 mesh) using the speci-
fied eluents.
(2R,3R,4S,4aS,5aR,9aS,10aR)-3,4-Bis(benzoyloxy)-2-(benzoyloxy-
methyl)-decahydro-2H-pyrano[2,3-b][1,4]benzodioxin (15): A mix-
ture of 12 (500 mg, 0.85 mmol), Et3SiH (2.7 mL, 17 mmol) and
BF3Ϫdiethyl ether (2.1 mL, 17 mmol) in CH2Cl2 (50 mL) was
stirred at 0 °C for 1.5 h and processed as described for 10 Ǟ 14.
The syrup obtained was purified by fast elution from a silica gel
column (2 ϫ 15 cm) with toluene/EtOAc (8:1). In vacuo removal
of the solvents from the respective eluates gave 219 mg (45%) of 15
as a colorless foam (Rf ϭ 0.66, toluene/EtOAc, 4:1). [α]2D0 ϭ Ϫ20.1
(c ϭ 0.8, CHCl3). 1H NMR: Table 1. 13C NMR (75.5 MHz,
CDCl3): δ ϭ 20.2, 24.7, 29.4, 31.0 (4 CH2), 62.3 (CH2OBz), 66.0
(C-3), 72.2 (C-2), 73.0 (C-4, C-9a), 74.3 (C-4a), 74.6 (C-5a), 92.8
(C-10a), 165.4, 166.1, 166.2 (3 C6H5-CO) ppm. MS (FD: m/z ϭ 572
[Mϩ]. C33H32O10 (572.6): calcd. C 69.22, H 5.63; found C 69.15,
H 5.59.
(2R,3R,4S,4aS,5aS,9aR,10aR)-3,4-Bis(benzoyloxy)-2-(benzoyloxy-
methyl)-4a-hydroxy-decahydro-2H-pyrano[2,3-b][1,4]benzodioxin
(cis-transoid-cis Isomer) (10): A mixture of meso-1,2-cyclohexane-
diol (1.16 g, 10 mmol), Ag2CO3 (2.76 g, 10 mmol), molecular sieves
˚
(4 A, 3 g) and CH2Cl2 (150 mL) was stirred for 15 min, followed
by addition of ulosyl bromide 5[11b] (5.53 g, 10 mmol) and refluxing
for 3 h. Filtration and removal of the solvent from the filtrate left
a syrup consisting of an approximate 1:1 mixture (1H NMR, TLC
with toluene/EtOAc, 4:1) of 10 (Rf ϭ 0.58) and 12 (0.37). Tritu-
ration with methanol resulted in crystallization of 10, which was
filtered with suction (mother liquor containing 12 vide infra) to
give 2.21 g (38%) of 10. M.p. 117Ϫ118 °C. [α]2D0 ϭ Ϫ1.9 (c ϭ 1.2,
(2R,3R,4S,4aS,5aS,9aS,10aR)-3,4-Bis(benzoyloxy)-2-(benzoyl-
oxymethyl)-4a-hydroxy-decahydro-2H-pyrano[2,3-b][1,4]benzodioxin
(cis-transoid-trans Isomer) (16): A slurry of (S,S)-1,2-cyclohexane-
diol (1.16 g, 10 mmol), Ag2CO3 (2.76 g, 10 mmol), molecular sieves
CHCl3). 1H NMR (300 Hz, CDCl3): δ ϭ 1.1Ϫ2.1 (m, 8 H), 3.78 (4 A, 3 g) and ulosyl bromide 5 (5.53 g, 10 mmol) in CH2Cl2
(m, 1 H), 4.13 (ddd, J ϭ 2.9, 4.9, 10.0 Hz, 1 H), 4.57 (m, 1 H), (150 mL) was gently refluxed for 2 h ( at approximately 40 °C) and
˚
5.03 (s, 1 H), 5.19 (d, J ϭ 9.8 Hz, 1 H), 5.30 (s, OH), 5.78 (dd, J ϭ
then filtered through kieselguhr. The filtrate, on removal of the
9.8, 10.0 Hz, 1 H) ppm. 13C NMR (75.5 MHz, CDCl3): δ ϭ 20.2, solvent in vacuo, gave a colorless foam, consisting of an approxi-
24.5, 29.2, 30.0 (four CH2), 63.4 (CH2OBz), 65.9 (C-9a), 67.9 (C-
2), 71.7 (C-3), 72.6 (C-5a), 81.1 (C-4), 90.5 (C-4a), 97.2 (C-10a),
mate 1:1-mixture (1H NMR) of the cis-transoid-trans (16) and the
all-trans isomer 17 with nearly identical Rf values (0.55 in toluene/
165.3, 167.5, 168.2 (3 C6H5CO) ppm. MS (FD, 15 mA): m/z (%) ϭ EtOAc, 8:1). Trituration with methanol resulted in crystallization
588 (100) [Mϩ]. C33H32O10 (588.6): calcd. C 67.33, H 5.48; found of 2.59 g (44%) of 16 (for mother liquor containing 17, see below).
C 67.23, H 5.43.
M.p. 154Ϫ156 °C. [α]2D0 ϭ Ϫ43.7 (c ϭ 1, CHCl3). 1H NMR
Eur. J. Org. Chem. 2004, 4911Ϫ4920
www.eurjoc.org
© 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4917

F. W. Lichtenthaler, E. Cuny
FULL PAPER
(300 Hz, CDCl3): δ ϭ 1.1Ϫ2.1 (m, 8 H), 4.08 (m, 3 H), 4.44 (dd,
trans isomer 17, which after 20 days resulted in a 5:1 mixture of 17
J ϭ 3.1, 12.1 Hz, 1 H), 4.66 (dd, J ϭ 4.4, 12.1 Hz, 1 H), 4.70 (s, 1 and 16. The major product (17) could be secured therefrom by
H, OH), 5.06 (s, 1 H), 5.35 (d, J ϭ 9.8 Hz, 1 H), 5.88 (dd, J ϭ 9.8,
9.9 Hz, 1 H), 7.2Ϫ8.1 (m, 15 H) ppm. 13C NMR (75.5 MHz,
CDCl3): δ ϭ 123.7, 23.8, 31.3, 31.6, 63.2, 68.5, 71.8, 74.9, 75.7,
78.6, 93.1, 97.4, 128Ϫ133, 165.3, 166.1, 167.8 ppm. MS (FD,
15 mA): m/z (%) ϭ 588 (100) [Mϩ]. C33H32O10 (588.59): calcd.
C 67.33, H 5.48; found C 67.19, H 5.39.
Keeping a methanolic solution of 16 at room temperature over-
night allowing slow evaporation gave good quality crystals suitable
for an X-ray analysis, which showed two independent, yet nearly
identical molecular forms. Only one is depicted in Figure 1.
elution from a silica gel column with toluene/EtOAc (10:1) as de-
scribed above. Yield from 1.0 g of 16: 650 mg (65%).
(2R,3R,4S,4aR,5aS,9aS,10aR)-3,4-Bis(benzoyloxy)-2-(benzoyloxy-
methyl)-decahydro-2H-pyrano[2,3-b][1,4]benzodioxin (18): To
a
solution of (500 mg, 0.85 mmol) of 17 in CH2Cl2 (50 mL) was ad-
ded Et3SiH (2.7 mL, 17 mmol) and boron trifluorideϪdiethyl ether
(2.1 mL, 17 mmol), and the mixture was stirred for 1.5 h at 0 °C.
Workup as described for 10 Ǟ 14 and trituration of the syrupy resi-
due with methanol resulted in crystallization of 390 mg (80%) of
18. M.p. 159Ϫ160 °C. [α]2D0 ϭ Ϫ14.6 (c ϭ 0.9, CHCl3). H NMR
1
(300 MHz, CDCl3): δ ϭ 1.1Ϫ2.1 (m, 8 H, 4 CH2), 3.24, 3.46 (m, 2
H, 5a-H, 9a-H), 3.60 (dd, J4,4a ϭ 9.8, J4a,10a ϭ 7.8 Hz, 4a-H), 4.21
(ddd, J2,CH ϭ 3.0, 5.2, J2,3 ϭ 9.6 Hz, 2-H), 4.47 and 4.61 (dd, 2
Table 2. Crystal data and structure refinement for 16
H, 2-CH2),24.70 (d, J4a,10a ϭ 7.8 Hz, 10a-H), 5.62 (t, J2,3 ϭ J3,4
ϭ
Empirical formula
Molecular mass
Temperature (K)
C33H32O10
588.59
293(2)
1.54180
9.6 Hz, 3-H), 5.71 (dd, J3,4 ϭ 9.6, J4,4a ϭ 9.8 Hz, 4-H), 7.2Ϫ8.1
(m, 15 H, 3 C6H5) ppm. C33H32O10 (572.6): calcd. C 69.22, H 5.63;
found C 69.15, H 5.51.
˚
Wavelength (A)
Crystal system, space group
Unit cell dimensions
orthorhombic, P 21 21 21
(2R,6S)-4-Benzoyloxy-2-(benzoyloxymethyl)-2-[(S,S)-(2-hydroxy-
cyclohexyl)oxy]-2H-pyran-3(6H)-one (20): Silver carbonate (2.75 g,
10 mmol) and molecular sieves (4 A, 2 g) were added to a solution
of (S,S)-cyclohexanediol (1.16 g, 10 mmol) in CH2Cl2 (20 mL), and
the mixture was stirred for 20 min, followed by dropwise addition
˚
a (A)
45.21(2)
13.622(10)
9.820(10)
90
90
90
6047(8)
8
1.293
0.797
2480
0.5 ϫ 0.05 ϫ 0.02
4.38 to 59.57
˚
˚
b (A)
˚
c (A)
α (°)
β (°)
γ (°)
of
(2R,6S)-4-benzoyloxy-6-(benzoyloxymethyl)-6-bromo-2H-py-
ran-3(6H)-one[14] (19) (2.60 g, 6 mmol) in CH2Cl2 (20 mL) over
5 min. After stirring for another 15 min (TLC showed absence of
starting material), the mixture was diluted with CH2Cl2 (150 mL)
and washed with water (3 ϫ 50 mL), dried (Na2SO4) and taken to
dryness in vacuo. Treatment with ether resulted in crystallization.
Recrystallization from ether gave unsaturated glycosiduloside 20 as
colorless needles (3.78 g, 81%). M.p. 168Ϫ169 °C. [α]2D0 ϭ Ϫ77.4
3
˚
Volume (A )
Z
Calculated density (Mg·mϪ3
)
Absorption coefficient (mmϪ1
F(000)
)
Crystal size (mm)
θmax. (°) for data collection
Limiting indices
Reflections collected/unique
Completeness to θ ϭ 59.57
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices [I Ͼ 2σ (I)]
R indices (all data)
1
(c ϭ 0.65, CHCl3). H NMR (300 Hz, CDCl3): δ ϭ 1.1Ϫ2.1 (m, 8
0 Յ k Յ 14, 0 Յ l Յ 8
4129/4129 [R(int) ϭ 0.0000]
82.8%
H, 4 CH2), 3.20 (m, 1 H, OH), 3.31 (m, 2 H, 5a-H, 9a-H), 4.69,
4.73 (dd, 2 H, CH2OBz), 5.10 (ddd, 1 H, 2-H), 5.41 (s, 1 H, 10a-
H), 6.94 (d, 1 H, 3-H), 7.3Ϫ8.2 (m, 10 H, 2 C6H5) ppm; J2,3 ϭ 3.3,
full-matrix least-squares on F2
4129/748/778
J3,CH
ϭ
6.5,
J
C,H OBzgem ϭ 11.2 Hz. 13C NMR (75.5 MHz,
2
CDC2l3): δ ϭ 23.9, 24.4 (C-7, C-8), 31.8, 32.2 (C-5, C-9), 66.3, 71.1,
74.2 (C-2, C-5a, C-9a, CH2OBz), 100.2 (C-10a), 142.5 (C-4), 166.1,
166.2 (2 BzCO), 183.9 (C-4a) ppm. MS (FD, 15 mA): m/z (%) ϭ
466 (10) [Mϩ], 320 (100). C26H26O8 (466.5): calcd. C 66.94, H 5.62;
found C 66.88, H 5.53.
1.269
R1 ϭ 0.0894, wR2 ϭ 0.2494
R1 ϭ 0.1036, wR2 ϭ 0.2809
0(10)
0.0033(6)
0.391 and Ϫ0.430
Absolute structure parameter
Extinction coefficient
Largest diff. peak and hole
(e·AϪ3
)
(2S,4aR,5aS,9aS,10aS)-4-Benzoyloxy-2-(benzoyloxymethyl)-4a-
hydroxy-octahydro-2H-pyrano[2,3-b][1,4]benzodioxin (trans-cisoid-
trans Isomer) (21): A solution of 20 (200 mg, 0.43 mmol) in CHCl3
(15 mL) was kept at ambient temperature for 30 days, whereafter
cycloketalization had occurred nearly quantitatively (1H NMR).
Removal of the solvent in vacuo and filtration of the crystalline
residue with ether afforded 175 mg (87%) of 21. M.p. 155Ϫ158 °C.
[α]2D0 ϭ Ϫ53 (c ϭ 0.9, CHCl3). 1H NMR (300 Hz, CDCl3): δ ϭ
1.1Ϫ2.2 (m, 8 H), 3.23 (m, 1 H, 5a-H), 3.34 (m, 1 H, 9a-H), 4.52
and 4.62 (dd, 2 H, CH2OBz), 4.92 (ddd, 1 H, 2-H), 5.03 (d, 1 H,
10a-H), 6.05 (d, 1 H, 3-H), 6.14 (s, 1 H, OH) ppm; J2,CH ϭ 6.4
and 7.0, J2,3 ϭ 2.6, JC,H OBzgem ϭ 10.8 Hz; NOE between2signals
at 3.34 and 5.03 (9a-H2o 10a-H) ppm. 13C NMR (75.5 MHz,
CDCl3): δ ϭ 23.7, 23.9 (C-7, C-8), 29.6, 29.7 (C-5, C-9), 67.1, 71.0,
76.0, 78.2 (C-2, C-5a, C-9a, CH2OBz), 89.0 (C-4a), 96.6 (C-10a),
117.4 (C-3), 143.9 (C-4), 166.2, 167.3 (2 BzCO) ppm. C26H26O8
(466.47): calcd. C 66.94, H 5.62; found C 66.81, H 5.59.
(2R,3R,4S,4aR,5aS,9aS,10aR)-3,4-Bis(benzoyloxy)-2-(benzoyl-
oxymethyl)-4a-hydroxy-decahydro-2H-pyrano[2,3-b][1,4]benzodioxin
(trans-cisoid-trans Isomer) (17): The methanolic mother liquor re-
maining after isolation of 16 was evaporated to dryness in vacuo,
and the syrupy residue, mainly consisting of 17, was purified by
rapid elution from a silica gel column (2 ϫ 20 cm) with toluene/
EtOAc (10:1). Removal of the solvents from the respective eluates
gave 1.95 g (33%) of 17 as a colorless syrup. [α]2D0 ϭ Ϫ20.5 (c ϭ 1,
CHCl3). 1H NMR (300 Hz, CDCl3): δ ϭ 1.1Ϫ2.1 (m, 8 H), 3.52
(m, 1 H), 3.76 (m, 1 H), 4.23 (ddd, J ϭ 2.8, 4.9, 9.7 Hz, 1 H), 5.68
(d, J ϭ 9.7 Hz, 1 H), 5.90 (t, J ϭ 9.7 Hz, 1 H), 7.2Ϫ8.2 (m, 15 H)
ppm. 13C NMR (75.5 MHz, CDCl3): δ ϭ 23.9, 24.0, 29.3, 29.6,
63.1, 69.4, 72.4, 73.9, 81.1, 92.1, 98.6, 128Ϫ133 (C6H5), 165.3,
166.2, 167.7 ppm. MS (FD, 15 mA): m/z ϭ 588 [Mϩ]. C33H32O10
(588.59): calcd. C 67.33, H 5.48; found C 67.28, H 5.40. The stand-
ing of 16 in chloroform at room temperature, as noticed with an (4aS,5aS,9aS,10aR)-4a-Benzoyloxy-2-methyl-octahydro-4H-pyrano-
NMR sample, induced gradual equilibration to the more stable all- [2,3-b][1,4]benzodioxin-4-one (cis-transoid-trans Isomer) (23): Tetra-
4918
© 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 4911Ϫ4920

Linear Fused Pyran-Dioxane-Cyclohexane Tricycles
n-butylammonium acetate (2.60 g, 8.6 mmol) was added to a solu- (2R,3R,4S,4aS,5aR,9aR,10aR)-3,4-Bis(benzoyloxy)-2-(benzoyl-
FULL PAPER
tion of enolone glycoside 20 (2.0 g, 4.3 mmol) in a mixture of aceto-
nitrile (200 mL) and CHCl3 (50 mL), followed by stirring at ambi-
ent temperature for 15 h. Subsequent dilution with CH2Cl2
(300 mL), washing with water (3 ϫ 200 mL), drying (Na2SO4), and
evaporation to dryness in vacuo left a syrup, which was purified by
elution from a silica gel column with toluene/EtOAc (4:1), to give
oxymethyl)-decahydro-2H-pyrano[2,3-b][1,4]benzodioxin (28): To a
cooled (0 °C) solution of 26 (1.0 g, 1.7 mmol) in CH2Cl2 (50 mL)
was added triethylsilane (4.0 g, 34 mmol) and BF3Ϫdiethyl ether
(4.8 g, 34 mmol) consecutively. The mixture was stirred at 0 °C for
1 h and then allowed to come to room temperature. Workup as
described for 10 Ǟ 14, purification by elution from a short silica
gel column with toluene/EtOAc (20:1) gave a syrupy residue, which
1.08 g (73%) of 23 as a colorless syrup. [α]2D0 ϭ Ϫ103.6 (c ϭ 1,
1
CHCl3). H NMR (300 Hz, CDCl3): δ ϭ 2.11 (d, J3,CH ϭ 0.7 Hz, crystallized from MeOH to give 0.67 g (79%) of 28. M.p. 129Ϫ131
3
1
3 H, CH3), 4.02, 4.13 (m, 2 H, 5a-H, 9a-H), 5.48 (dd, J3,10a ϭ 0.4,
°C. [α]2D0 ϭ Ϫ17.1 (c ϭ 0.8, CHCl3). H NMR: Table 1. 13C NMR
J3,CH ϭ 0.7 Hz, 1 H, 3-H), 6.34 (d, J3,10a ϭ 0.4 Hz, 1 H, 10a-H) (75.5 MHz, CDCl3): δ ϭ 24.0 (C-8, C-7), 29.7, 30.0 (C-6, C-9),
ppm.3 13C NMR (75.5 MHz, CDCl3): δ ϭ 21.1 (CH3), 23.8 (C-7, 63.2 (CH2OBz), 66.4 (C-3), 72.5 (C-2), 72.9 (C-5a, C-9a), 73.1
C-8), 30.9, 31.3 (C-5, C-9), 78.4 (C-5a, C-9a), 95.1 (C-4a), 97.0 (C-
(C-4a), 79.9 (C-4), 93.8 (C-10a), 165.4, 166.0, 166.2 (3 BzCO)
10a), 103.4 (C-3), 164.7 (BzCO), 171.3 (C-2), 183.6 (C-4) ppm. MS ppm. C33H32O10 (572.6): calcd. C 69.22, H 5.63; found C 69.20,
(FD, 15 mA): m/z ϭ 344 [Mϩ]. C19H20O6 (344.4): calcd. C 66.27,
H 5.85; found C 66.19, H 5.78.
H 5.59.
Exposure of cis-transoid-trans-linked 16 to Bu4NOAc treatment in
MeCN/CHCl3 as described above followed by analogous workup
afforded 23 in 57% yield.
(2S,4aS,5aR,9aR,10aS)-4-Benzoyloxy-2-(benzoyloxymethyl)-4a-
hydroxy-octahydro-2H-pyrano[2,3-b][1,4]benzodioxin (29): Molecu-
lar sieves (4 A, 2 g) and Ag2CO3 (2.76 g, 10 mmol) were added to a
˚
solution of (R,R)-1,2-cyclohexanediol (1.16 g, 10 mmol) in CH2Cl2
(20 mL), and the mixture was stirred for 20 min with the exclusion
of moisture and light. A solution of bromo-enolone 19[14] (2.60 g,
6 mmol) in CH2Cl2 (20 mL) was then added dropwise over 5 min.
After stirring for another 15 min, TLC (n-hexane/acetone/CHCl3,
4:3:3) indicated the absence of 19; the mixture was filtered through
(4aR,5aS,9aS,10aR)-4a-Benzoyloxy-2-methyl-octahydro-4H-
pyrano[2,3-b][1,4]benzodioxin-4-one (trans-cisoid-trans Isomer) (24):
BF3Ϫdiethyl ether (1.25 mL, 10 mmol) was added to a cooled (0
°C) CH2Cl2 solution of 23 (345 mg, 1 mmol, in 20 mL), and the
mixture was stirred at 0 °C for 30 min followed by 1 h at ambient
temperature, whereafter 23 (Rf ϭ 0.60 in toluene/EtOAc, 4:1) had kieselguhr, and the filtrate was diluted with CH2Cl2 (150 mL) fol-
isomerized completely to 24 (Rf ϭ 0.50). Dilution with CH2Cl2 lowed by washing with water (3 ϫ 50 mL), drying (Na2SO4), and
(200 mL), successive washing with saturated NaHCO3, solution
evaporation to dryness in vacuo. The syrupy residue crystallized on
(50 mL), and water (2 ϫ 50 mL), followed by drying (Na2SO4), and trituration with ether to give 2.73 g (84%) of 29 as the mono-ether-
evaporation to dryness in vacuo gave a syrup, which was purified
by elution from a silica gel column (3 ϫ 20 cm) with toluene/EtOAc
(20:1). Removal of the solvents from the eluates left colorless crys-
ate. M.p. 78Ϫ80 °C. [α]2D0 ϭ Ϫ7.3 (c ϭ 1, CHCl3). 1H NMR
(300 Hz, CDCl3): δ ϭ 1.1Ϫ2.2 (m, 8 H), 3.65 (s, 1 H, OH), 3.87,
3.94 (m, 2 H, 5a-H, 9a-H), 4.52 (m, 2 H, CH2OBz), 4.83 (ddd, 1
tals of 24 (295 mg, 85%). M.p. 132Ϫ134 °C. [α]2D0 ϭ Ϫ185.3 (c ϭ H, 2-H), 4.93 (s, 1 H, 10a-H), 5.97 (d, 1 H, 3-H), 7.3Ϫ8.2 (m, 10
0.8, CHCl3). 1H NMR (300 Hz, CDCl3): δ ϭ 2.32 (d, J3,Me
0.7 Hz, 3 H, CH3), 4.04 (m, 2 H, 5a-H, 9a-H), 5.46 (d, J3,Me
ϭ
H, 2 C6H5) ppm; J2,3 ϭ 1.7, J2,CH ϭ 5.1 Hz. 13C NMR (75.5 MHz,
2
ϭ
CDCl3): δ ϭ 24.2, 24.3 (C-7, C-8), 29.8, 29.9 (C-6, C-9), 65.9
0.7 Hz, 1 H, 3-H), 6.04 (s, 1 H, 10a-H) ppm; NOE between signals (CH2OBz), 70.5 (C-2), 72.3 (C-5a, C-9a), 88.2 (C-4a), 95.3 (C-10a),
at 4.04 and 6.04 (9a-H o 10a-H) ppm. 13C NMR (CDCl3, 115.5 (C-3), 128Ϫ134 (2 C6H5), 145.3 (C-4), 164.6, 166.4
75.5 MHz): δ ϭ 17.2 (CH3), 24.0, 24.1 (C-7, C-8), 29.6 (C-5, C-9), (2 C6H5CO) ppm. MS (FD, 5 mA): m/z (%) ϭ 466 (100) [Mϩ].
72.3, 76.1 (C-5a, C-9a), 88.5 (C-10a), 100.1 (C-4a), 104.1 (C-3), C26H26O8·Et2O (540.66): calcd. C 66.65, H 6.71; found C 66.70,
164.9 (BzCO), 195.1 (C-4) ppm. C19H20O6 (344.4): calcd. C 66.27,
H 6.69.
H 5.85; found C 66.23, H 5.81.
Recrystallization from diethyl ether afforded well-formed needles
suitable for X-ray structural analysis,[16] which proved the presence
of an ether molecule.
(2R,3R,4S,4aS,5aR,9aR,10aR)-3,4-Bis(benzoyloxy)-2-(benzoyloxy-
methyl)-4a-hydroxy-decahydro-2H-pyrano[2,3-b][1,4]benzodioxin
(cis-cisoid-trans Isomer) (26): A mixture of (R,R)-1,2-cyclohexane-
diol (116 mg, 1 mmol), Ag2CO3 (276 mg, 1 mmol), freshly desic-
(2S,4aR,5aR,9aR,10aS)-2-(Benzoyloxymethyl)-4a-hydroxy-deca-
hydro-2H-pyrano[2,3-b][1,4]benzodioxin-4-one (30): To a solution of
29·Et2O (270 mg, 0.5 mmol) in acetonitrile (10 mL) was added
tetrabutylammonium acetate (450 mg, 1.5 mmol) and 5 drops of
˚
cated molecular sieves (4 A) and CH2Cl2 (50 mL) was stirred for
15 min at ambient temperature, followed by addition of ulosyl bro-
mide 5[11b] (553 mg, 1 mmol) and gentle refluxing for 2 h with the water, and the mixture was stirred for 3 h at ambient temperature.
exclusion of light and moisture. Subsequent filtration through
kieselguhr and evaporation of the filtrate to dryness in vacuo af-
forded a syrup, which crystallized on trituration with methanol to
Dilution with CH2Cl2 (100 mL), washing with water (3 ϫ 50 mL),
drying (Na2SO4), and evaporation to dryness in vacuo left 155 mg
(85%) of 30 as a chromatographically uniform syrup (Rf ϭ 0.12,
give 510 mg (87%) of 26 as colorless crystals. M.p. 196Ϫ198 °C. toluene/EtOAc, 8:1). [α]2D0 ϭ Ϫ5.6 (c ϭ 1, CHCl3). 1H NMR
1
[α]2D0 ϭ Ϫ9.5 (c ϭ 1.1, CHCl3). H NMR (300 MHz, CDCl3): δ ϭ
(300 Hz, CDCl3): δ ϭ 1.1Ϫ2.0 (m, 8 H, 4 CH2), 2.59, 3.08 (dd, 2
1.1Ϫ2.0 (br. m, 8 H), 3.90 (m, 2 H), 4.11 (ddd, J ϭ 2.9, 4.6,
H, 3-H2), 3.91 (m, 1 H, 2-H), 4.05 (m, 2 H, 5a-H, 9a-H), 4.43 (s,
10.0 Hz, 1 H), 4.47, 4.68 (dd, J ϭ 2.9, 4.6 Hz, 2 H), 4.94 (s, 1 H), 1 H, OH), 4.51, 4.55 (dd, 2 H, CH2OBz), 4.73 (s, 1 H, 10a-H),
5.10 (s, OH), 5.31 (d, J ϭ 10.0 Hz, 1 H), 5.88 (t, J ϭ 10.0 Hz, 1 7.2Ϫ8.1 (m, 5 H, C6H5) ppm; J3,3 ϭ 14.2, J2,3 ϭ 2.0 and 12.3,
H), 7.2Ϫ8.1 (m, 15 H, 3 C6H5) ppm. 13C NMR (75.5 MHz,
J2,CH ϭ 4.6 and 5.0, JC,H OBzgem ϭ 11.7 Hz. 13C NMR (75.5 MHz,
2
CDCl3): δ ϭ 24.1 (C-7, C-8), 29.5, 29.8 (C-6, C-9), 63.1 (CH2OBz), CDC2l3): δ ϭ 24.1, 24.2 (C-7, C-8), 29.7, 29.8 (C-6, C-9), 39.9 (C-
67.9 (C-3), 71.9 (C-2), 72.2, 72.6 (C-5a, C-9a), 78.6 (C-4), 91.9 (C- 3), 65.7 (CH2OBz), 69.3 (C-2), 72.5, 72.8 (C-5a, C-9a), 91.3 (C-4a),
4a), 96.1 (C-10a), 165.3, 166.2, 168.2 (3 BzCO) ppm. MS (FD,
15 mA): m/z (%) ϭ 588 (100) [Mϩ]. C33H32O10 (588.59): calcd.
C 67.33, H 5.48; found C 67.23, H 5.43.
97.3 (C-10a), 128Ϫ133 (C6H5), 166.1 (BzϪCO), 201.2 (C-4) ppm.
MS (FD, 15 mA): m/z (%) ϭ 362 (100) [Mϩ]. C19H22O7 (362.4):
calcd. C 62.97, H 6.12; found C 62.88, H 6.17.
Eur. J. Org. Chem. 2004, 4911Ϫ4920
www.eurjoc.org
© 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4919

F. W. Lichtenthaler, E. Cuny
FULL PAPER
Exposure of ulosyl bromide-derived 26 (295 mg, 0.5 mmol) to Bu4-
NOAc in moist acetonitrile (as described above) afforded 145 mg
(80%) of 30.
[11] [11a]
F. W. Lichtenthaler, E. Kaji, S. Weprek, J. Org. Chem.
Acknowledgments
[11b]
1985, 50, 3505Ϫ3515.
F. W. Lichtenthaler, U. Kläres, M.
Lergenmüller, S. Schwidetzky, Synthesis 1992, 179Ϫ184.
The BF3-promoted reductive deoxygenation of 10 and 12 can-
not occur through SN2-type hydride attack on the tertiary car-
bon (C-4a) of the OH-activated intermediate (e.g. II), as the
rigidity of the tricyclic ring system precludes inversion at C-4a.
The reaction must thus proceed through a carboxonium ion
(e.g. III), which adds hydride, e.g. II Ǟ III Ǟ 15. Accordingly,
the OH groups in 10 and 12 are replaced by hydrogen with
retention of configuration, thereby verifying the linkage geo-
metry conclusions drawn from NOE effects of the cis-transoid-
cis (10/14) and the cis-cisoid-cis pairs 12/15 (Scheme 3).
The authors would like to thank Mrs. Ingrid Svoboda for collecting
the X-ray data and Prof. Dr. H. J. Lindner for elaborating on the
crystal structure.
[12]
[1] [1a]
G. Hesse, F. Reicheneder, H. Eysenbach, Justus Liebigs
Ann. Chem. 1939, 537, 67Ϫ86. [1b] F. Brüschweiler, W. Stöcklin,
K. Stöckel, T. Reichstein, Helv. Chim. Acta 1969, 52,
2086Ϫ2106. [1c] F. Büschweiler, K. Stöckel, T. Reichstein, Helv.
[1d]
Chim. Acta 1969, 52, 2276Ϫ2302.
H. T. A. Cheung, T. R.
Watson, J. Chem. Soc., Perkin Trans. 1 1980, 2162Ϫ2168.
[2]
[3] [3a]
W. Rosenbrook, Jr., J. Antibiotics 1979, 32, S211ϪS227.
H. T. A. Cheung, R. G. Coombe, W. T. L. Sidwell, T. R.
[3b]
Watson, J. Chem. Soc., Perkin Trans. 1 1981, 64Ϫ72.
H. T.
A. Cheung, F. C. K. Chiu, T. R. Watson, R. J. Wells, J. Chem.
Soc., Perkin Trans. 1 1983, 2827Ϫ2835.
P. Brown, J. v. Euw, T. Reichstein, K. Stöckel, T. R. Watson,
Helv. Chim. Acta 1979, 62, 412Ϫ441.
S. Nishio, M. S. Blum, J. V. Silverton, R. J. Highet, J. Org.
Chem. 1982, 47, 2154Ϫ2157.
The linkage notations used correspond to those established for
the five isomeric perhydroanthracenes (cf. E. Eliel, S. H. Wilen,
Stereochemistry of Organic Compounds, Wiley-Interscience,
New York, 1994, pp. 780Ϫ782), as the fused pyran-dioxane-
cyclohexane tricycles described are de facto 1,5,10-trioxa de-
rivatives thereof.
[4]
[5]
[6]
[7]
[8]
E. Cuny, F. W.Lichtenthaler, H. J. Lindner, Eur. J. Org. Chem.
2004, preceding paper.
For reviews see: [8a] E. Kaji, F. W. Lichtenthaler, Trends Glyco-
[8b]
sci. Glycotechnol. 1993, 5, 121Ϫ142.
J. J. Gridley, H. M. I.
[8c]
[13]
Osborn, J. Chem. Soc., Perkin Trans. 1 2000, 1471Ϫ1491.
V. Pozsgay, in Carbohydrates in Chemistry and Biology, Part I
(Eds.: B. Ernst, G. W. Hart, P. Sinay¨), Wiley-VCH, Weinheim,
CCDC-224773 contains the supplementary crystallographic
data for 16. These data can be obtained free of charge at
www.ccdc.cam.ac.uk/conts/retrieving.html [or from the Cam-
bridge Crystallographic Data Centre, 12, Union Road,
Cambridge CB2 1EZ, UK; Fax: ϩ44-1223-336-033; E-mail:
deposit@ccdc.cam.ac.uk].
F. W. Lichtenthaler, U. Kraska, Carbohydr. Res. 1977, 58,
363Ϫ377.
F. W. Lichtenthaler, S. Nishiyama, T. Weimer, Liebigs Ann.
Chem. 1989, 1163Ϫ1170.
E. Cuny, F. W. Lichtenthaler, H. J. Lindner, Acta Crystallogr.,
Sect. B 1994, 50, 1599Ϫ1601.
An attractive possibility for the biosynthesis of cardenolides of
type 1, 3 and 4, as well as of spectinomycin (2), implies mono-
glycosylation of the steroidal diol aglycon or of actinamine,
oxidation of the respective 6-deoxy-glycosides to the glycosid-
uloses which Ϫ in nonenzyme-mediated fashion Ϫ undergo in-
tramolecular hemiketalization.
Germany, 2000, part I, pp. 332Ϫ336. Recent applications
[8d]
towards β-d-mannosides:
M. Nitz, D. R. Bundle, J. Org.
[8e]
Chem. 2001, 66, 8411Ϫ8423.
F. W. Lichtenthaler, M.
[14]
[15]
[16]
[17]
Lergenmüller, S. Peters, Z. Varga, Tetrahedron: Asymmetry
2003, 14, 727Ϫ736. [8f] F. W. Lichtenthaler, M. Lergenmüller, S.
Schwidetzky, Eur. J. Org. Chem. 2003, 3094Ϫ3103.
[9] [9a]
D. R. White, R. D. Birkenmeyer, R. C. Thomas, S. A. Miz-
[9b]
sak, V. H. Wiley, Tetrahedron Lett. 1979, 2737Ϫ2740.
S.
Hanessian, R. Roy, Can. J. Chem. 1985, 63, 163Ϫ172.
[10]
The two total syntheses of spectinomycin, both comprising the
dioxanoid annulation of the sugar portion onto N,NЈ-bis(ben-
zyloxycarbonyl)actinamine (I), required nine steps from l-glu-
cose (3% overall yield),[9a] and 23 from d-glucose, the presumed
overall yield (step 9 proceeded with an efficiency of 114%)
being in the 2% range.[9b]
Received July 1, 2004
4920
© 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 4911Ϫ4920