Stereocontrolled synthesis of polyhydroxylated hexahydro-1H-cyclopent[c]isoxazoles by intramolecular oxime olefin cycloadditions: An approach to aminocyclopentitols

Article abstract of DOI:10.1039/a905706d

A series of alk-5-enyl aldehydes derived from various carbohydrates (D-glucose, D-mannose, D-galactose, D-glucal) can be transformed into the corresponding oximes. Thermolysis of these oximes results in the isolation of hexahydro-1H-cyclopent[c]isoxazoles in good yields via intramolecular oxime olefin cycloadditions. Modest to excellent levels of diastereocontrol are observed in these cycloaddition reactions depending on the precise nature of the oxime precursor. In the best case, D-glucose-derived oxime 4 produces hexahydro-1H-cyclopent[c]isoxazole 5 as the sole product in quantitative yield. When the oxime possesses a substituent (OBn or OBz) adjacent to the oxime carbon atom, it is observed that reactions show a preference to produce the diastereomeric cycloadduct in which this substituent is located in an exo orientation relative to the newly formed hexahydro-1H-cyclopent[c]isoxazole ring system. The role of the solvent polarity on the diastereochemical outcome of these reactions is briefly discussed. Unsuccessful efforts to extend this chemistry to oximes derived from alk-4-enyl aldehydes are also presented. Finally, it is demonstrated that the hexahydro-1H-cyclopent[c]isoxazoles can be transformed into stereochemically defined aminocyclopentitols. The Royal Society of Chemistry 1999.

Full text of DOI:10.1039/a905706d

View Article Online / Journal Homepage / Table of Contents for this issue
Stereocontrolled synthesis of polyhydroxylated hexahydro-
1H-cyclopent[c]isoxazoles by intramolecular oxime olefin
cycloadditions: an approach to aminocyclopentitols
Paul J. Dransfield, Stéphane Moutel, Michael Shipman* and Vladimir Sik
School of Chemistry, Stocker Road, University of Exeter, Exeter, UK EX4 4QD
Received (in Cambridge, UK) 14th July 1999, Accepted 17th September 1999
A series of alk-5-enyl aldehydes derived from various carbohydrates (-glucose, -mannose, -galactose, -glucal)
can be transformed into the corresponding oximes. Thermolysis of these oximes results in the isolation of hexahydro-
1H-cyclopent[c]isoxazoles in good yields via intramolecular oxime olefin cycloadditions. Modest to excellent levels of
diastereocontrol are observed in these cycloaddition reactions depending on the precise nature of the oxime
precursor. In the best case, -glucose-derived oxime 4 produces hexahydro-1H-cyclopent[c]isoxazole 5 as the sole
product in quantitative yield. When the oxime possesses a substituent (OBn or OBz) adjacent to the oxime carbon
atom, it is observed that reactions show a preference to produce the diastereomeric cycloadduct in which this
substituent is located in an exo orientation relative to the newly formed hexahydro-1H-cyclopent[c]isoxazole ring
system. The role of the solvent polarity on the diastereochemical outcome of these reactions is briefly discussed.
Unsuccessful efforts to extend this chemistry to oximes derived from alk-4-enyl aldehydes are also presented. Finally,
it is demonstrated that the hexahydro-1H-cyclopent[c]isoxazoles can be transformed into stereochemically defined
aminocyclopentitols.
ing oxime, and subsequent thermally induced IOOC reaction to
Introduction
the hexahydro-1H-cyclopent[c]isoxazole skeleton (Scheme 1).
A number of biologically active natural products have been
In this key transformation, it is notable that two new stereo-
isolated and characterised which contain polyhydroxylated
genic centres are created at the ring fusion. While previous
aminocyclopentane (aminocyclopentitol) motifs. Illustrative
studies on IOOC reactions indicated that good levels of stereo-
examples include the selective mannosidase inhibitor man-
control can be accomplished in relatively simple systems, it
was unclear at the outset of our studies whether good, predict-
nostatin A 1,1 and the very strong and specific trehalase inhibi-
tor trehazolin 2.2 The biological activity of these and related
able levels of stereocontrol could be accomplished in the more
structures has stimulated considerable interest in the synthesis
stereochemically complex scenarios depicted in Scheme 1. Fur-
of aminocyclopentitols, and while a number of elegant strat-
ther conversion of the resulting cycloadduct into the desired
egies have been devised,3 new methods for the construction of
aminocyclopentitol was expected to be realised in a straight-
aminocyclopentitols in a stereo- and enantiocontrolled fashion
forward fashion by cleavage of the N–O bond and deprotection
are still in demand.
of the hydroxy groups (Scheme 1). In this paper, we present
Scheme 1
We wished to examine whether intramolecular oxime olefin
cycloaddition (IOOC) reactions could be used as the key step in
an approach to stereochemically defined aminocyclopentitols
bearing a hydroxymethyl substituent.4 The first example of an
IOOC reaction was reported by Oppolzer and Keller in 1970.5
Additional examples of IOOC reactions have been reported by
the groups of Wildman,6 Grigg,7 Heathcock8 and Hassner,9
among others.10,11 Until very recently, no examples of these
reactions had been reported in the field of carbohydrate chem-
istry.4,11 Our general strategy to the target aminocyclopentitol
system centres on conversion of a stereochemically defined,
carbohydrate-derived alk-5-enyl aldehyde into the correspond-
findings on the stereoselective IOOC reactions using a variety
of carbohydrate-derived precursors, indicate what factors con-
trol the stereochemical outcome of these reactions, and demon-
strate that the resultant cycloadducts can be transformed into
aminocyclopentitols.
Results and discussion
Our investigations began using aldehyde 3,12 in which the three
hydroxy groups along the backbone were protected as benzyl
ethers. This compound was readily prepared from methyl
J. Chem. Soc., Perkin Trans. 1, 1999, 3349–3355
This journal is © The Royal Society of Chemistry 1999
3349

View Article Online
Table
1
Vicinal coupling constants (in Hz) for 1H-cyclo-
pent[c]isoxazoles 5, 8 and 11
α--glucopyranoside in 5 steps using a Vasella fragmentation
as the key step.12a,13 Treatment of aldehyde 3 with hydroxy-
lamine hydrochloride in warm ethanol in the presence of pyri-
dine furnished oxime 4 as a 7:3 mixture of E and Z isomers as
judged by 1H NMR spectroscopy. Thermolysis of this oxime in
toluene at 110 ЊC for 15 hours yielded 5 in quantitative yield as
essentially a single stereoisomer. Identical results were obtained
in this cycloaddition reaction using either geometrically pure
E-4 or a mixture of the E- and Z-isomer.
5
8
11
J3a,4
J4,5
J5,6
J6,6a
J6a,3a
8.0
8.0
8.0
6.5
10.0
6.0
8.0
8.0
6.0
9.5
7.4
7.4
7.1
3.8
9.1
Next, we converted aldehyde 6,13,14 in which the three
hydroxy groups along the backbone were protected as benzoyl
esters, into oxime 7 and examined its IOOC reaction. In this
instance, a 60% yield was observed for the conversion of this
oxime into the diastereomerically pure 1H-cyclopent[c]-
1
isoxazole 8 after column chromatography. H NMR spectro-
scopic analysis of the crude reaction suggests that this reaction
produces traces of another product (≤ 10%), possibly a dia-
stereomeric cycloadduct, although our efforts to isolate and
characterise this minor product were unsuccessful.
The gross chemical structures of 1H-cyclopent[c]isoxazoles 5
and 8 were determined using 1D- and 2D-NMR spectroscopy
in conjunction with mass spectrometry. The relative stereo-
chemical relationships within both these bicycles was eluci-
1
dated using H NMR vicinal coupling constants and selective
NOE difference measurements. In the case of tribenzyl-
substituted derivative 5, significant NOE enhancements were
observed between H-3a and H-6a (H-3a → H-6a {9.4%}; H-
6a → H-3a {9.6%}) and between these ring-junction hydro-
gens and H-5 (H-3a → H-5 {2.2%}; H-6a → H-5 {2.8%}).
Again, for tribenzoyl-substituted derivative 8, NOE enhance-
ments were observed between H-3a and H-6a (H-3a → H-6a
{10.5%}; H-6a → H-3a {11.0%}) and between these ring-
junction hydrogens and H-5 (H-3a → H-5 {3.8%}; H-
5 → H-3a {2.3%}; H-6a → H-5 {4.6%}; H-5 → H-6a
{2.8%}). Additional enhancements were measured for cyclo-
adduct 8 between H-4 and H-6 (H-4 → H-6 {3.4%}; H-
6 → H-4 {5.1%}). These experiments suggest that H-3a, H-5
and H-6a all reside on the exo face while H-4 and H-6 reside on
the endo face of these bicycles. In addition, the vicinal coup-
ling constants measured for 5 and 8 are in good agreement for
those reported for 11 (Table 1) which was synthesised by Ferrier
et al. and whose structure was unambiguously solved by X-ray
crystallography.15 Furthermore, the spectroscopic and physical
data obtained for the 1H-cyclopent[c]isoxazole 5 is in close
agreement with that reported by the group of Vasella (mp
108 ЊC {lit.,12a 104–106 ЊC}; [α]D Ϫ3 (c 0.93, CHCl3) {lit.,12a [α]D
Ϫ3.6 (c 0.9)†}, who synthesised it in a less direct fashion using
an intramolecular nitrone olefin cycloaddition.
Scheme 2 Reagents, conditions and yields: (i) HONH2ؒHCl, pyridine,
EtOH, 60 ЊC, 79% (for 3 → 4); 85% (for 6 → 7); (ii) toluene,
110 ЊC, 15 h, 100% (for 4 → 5); 60% (for 7 → 8); (iii) NaOMe,
MeOH, rt, 60%; (iv) 10% Pd/C, H2, MeOH, 91%.
isoxazole 5, as Vasella and co-workers have already successfully
demonstrated that this material can be transformed into a
variety of aminocyclopentitol and cyclopentene products.12a
To probe the remarkable diastereoselectivity observed in the
conversion of oxime 4 into the 1H-cyclopent[c]isoxazole 5,
we decided to undertake a series of related oxime olefin cyclo-
addition reactions using substrates bearing different benzyl
ether configurations along the backbone of the tethering chain.
2-Deoxy aldehyde 1213,16 was prepared from -glucal again
using the Vasella fragmentation protocol. Formation of the
corresponding oxime 13 was straightforward and subsequent
thermolysis produced the two inseparable diastereomeric
adducts 14 and 15 in a 2:1 ratio (Scheme 3). Our inability to
In the case of the 1H-cyclopent[c]isoxazole 8, we have
demonstrated that conversion to a stereochemically defined
aminocyclopentitol can be accomplished in a straightforward
fashion. Thus, cleavage of the benzoyl esters using methoxide
yields triol 9,12a which can be further converted into amino-
cyclopentitol 10 by cleavage of the N–O bond by catalytic
hydrogenation (Scheme 2). No efforts have been made to effect
such transformations using benzyl-protected 1H-cyclopent[c]-
Scheme 3 Reagents, conditions and yields: (i) HONH2ؒHCl, pyridine,
EtOH, 60 ЊC, 93%; (ii) toluene, 110 ЊC, 36 h, 71%.
† [α]D-Values are given in units of 10Ϫ1 deg cm2 gϪ1
.
3350
J. Chem. Soc., Perkin Trans. 1, 1999, 3349–3355

View Article Online
physically separate 14 and 15, or to adequately resolve their 1H
NMR signals in a variety of deuterated solvents, meant that we
have been unable to confidently assign which of these cycload-
ducts is produced as the major product. However, the most
significant finding from this study is that the removal of the C-2
benzyl ether group from the oxime (13 cf. 4) has a very detri-
mental effect on the diastereoselectivity observed in the IOOC
reaction. It is also of interest to note that the IOOC reaction
of oxime 13 is significantly slower than that of oxime 4 (36 h cf.
15 h) although the origin of this rate difference is not entirely
clear.
Further investigations into the role of the C-2 substituent on
the stereoselectivity of the oxime cycloaddition have been
undertaken using oxime 17, which was prepared from -
mannose-derived aldehyde 16,16,17 using the same general
approach (Scheme 4). Again, thermolysis yielded two diastereo-
Scheme 5 Reagents, conditions and yields: (i) HONH2ؒHCl, pyridine,
EtOH, 60 ЊC, 91%; (ii) toluene, 110 ЊC, 24 h, 71% (22), 12% (23).
nished 1H-cyclopent[c]isoxazoles 22 and 23 in 71% and 12%
isolated yield after column chromatography.
NMR spectroscopy was used to elucidate the relative stereo-
chemistry of the substituents within cycloadducts 22 and 23
although this was hampered by line broadening of the signals in
1
their H and 13C NMR spectra. This problem was especially
pronounced for the major adduct 22 for which the spectra had
to be acquired at 80 ЊC (Cl2DCCDCl2). The most significant
NOE enhancements were observed between H-3a and H-6a (H-
3a → H-6a {10.3%}) and between H-3a and H-4 (H-
3a → H-4 {9.4%}; H-4 → H-3a {8.4%}). For minor
adduct 23, large NOE enhancements were measured (60 ЊC in
C6D6) between H-3a and H-6a (H-3a → H-6a {7.0%}; H-
6a → H-3a {4.8%}); and between H-6a and H-6 (H-
6a → H-6 {7.1%}; H-6 → H-6a {7.0%}). These compara-
tive studies led us to conclude that both cycloadducts possess
cis-fused bicyclic skeletons, and that major adduct 22 possesses
H-4 cis to the ring junction hydrogens, whereas minor product
23 has H-6 cis to the ring junction hydrogens.
Scheme 4 Reagents, conditions and yields: (i) HONH2ؒHCl, pyridine,
EtOH, 60 ЊC, 87%; (ii) toluene, 110 ЊC, 60 h, 64% (18), 12% (19).
1
meric cycloadducts 18 and 19 in a 4:1 ratio as judged by H
In an effort to produce functionalised cyclobutanes, we have
attempted an IOOC reaction using a shorter linking tether
between the oxime and the olefin reaction partners. Oxime 25
was produced from -ribose-derived aldehyde 2419 in an identi-
cal fashion to that described earlier. In this case, the E- and Z-
isomers of the oxime were readily separable. However, attempts
to facilitate the IOOC reaction of (E)-25 met with failure. Even
in high boiling solvents (o-xylene or 1,2-dichlorobenzene) no
reaction could be induced and the oxime was recovered
unchanged (Scheme 6).
NMR analysis of the crude reaction mixture. Careful column
chromatography allowed isolation of 1H-cyclopent[c]isoxazoles
18 and 19 in 64% and 12% yield, respectively.
The stereochemical outcome of this reaction was again
resolved using NOE difference experiments. For major adduct
18, large enhancements were observed between H-3a and H-6a
(H-3a → H-6a {8.5%}; H-6a → H-3a {8.2%}); between H-
3a and H-4 (H-3a → H-4 {11.4%}; H-4 → H-3a {8.7%});
and between H-5 and H-6 (H-5 → H-6 {8.2%}; H-6 → H-
5 {8.4%}). Significantly smaller NOEs were observed between
H-6 and H-6a (H-6 → H-6a {3.8%}; H-6a → H-6
{4.2%}).18 From these NOE studies, we conclude that H-3a, H-
4 and H-6a all reside on the exo face; while H-5 and H-6 reside
on the endo face of bicycle 18. In addition, a small vicinal coup-
ling constant (1.5 Hz) is observed between H-6 and H-6a in this
cycloadduct, suggesting that these two hydrogens are orientated
at a dihedral angle close to 90Њ. This observation lends further
support to this stereochemical assignment. Unfortunately,
attempts to independently verify the stereochemistry of minor
cycloadduct 19 by performing similar NOE experiments have
not proven fruitful. In a variety of deuterated NMR solvents,
too many of the signals remain coincident to allow meaningful
NOE experiments to be performed.
Scheme 6 Reagents, conditions and yields: (i) HONH2ؒHCl, pyridine,
EtOH, 60 ЊC, 82%.
From our experimental results, it would appear that the
asymmetric centre adjacent to the oxime carbon atom on the
linking tether exerts the greatest influence on the stereochemical
outcome of these reactions. In major cycloadducts 5, 8, 18 and
22, the oxygen substituent at C-6 (cycloadduct numbering sys-
tem) is positioned in an exo orientation relative to the bicyclic
ring junction. Further evidence in support of the importance
of this substituent is provided by the observation that low
Using aldehyde 20,16,17 derived from -galactose, we have
investigated the role of the C-4 stereochemistry on the selectiv-
ity of these IOOC reactions. Conversion of this aldehyde into
the corresponding oxime 21 proceeded in an uneventful fashion
(Scheme 5). Subsequent thermolysis in refluxing toluene fur-
J. Chem. Soc., Perkin Trans. 1, 1999, 3349–3355
3351

View Article Online
stereoselectivity is observed in the reaction of oxime 13 which
only possesses a methylene group next to the oxime carbon
atom. Interestingly, similar observations have been made by
researchers studying the intramolecular nitrone olefin cyclo-
addition reactions of carbohydrate-derived substrates.20
To account for the stereochemical outcome of these reac-
tions, we have analysed the most likely transition states leading
to the observed products using scale models. For the conversion
of -glucose-derived oximes 4 and 7 into cycloadducts 5 and 8
respectively, we postulate that after initial 1,2-prototropic shift,
these reactions proceed via transition state A in which the 1,3-
dipole and olefin components adopt exo orientations and the
oxygen substituents orientate themselves in pseudo-equatorial
positions about the forming 5-membered ring (structure A,
X = OR). For -galactose-derived oxime 21, a closely related
transition state is proposed in which the C-4 substituent adopts
a pseudo-axial orientation (transition state B).
Experimental
General
Reactions requiring anhydrous conditions were performed
using oven-dried glassware and conducted under a positive
pressure of nitrogen. Anhydrous solvents were prepared in
accordance with standard protocols, or alternatively purchased
from Aldrich in Sure/SealTM bottles. IR spectra were recorded
(4000–600 cmϪ1) on a Nicolet Magna-550 FT-IR spectrometer
with internal calibration. Spectra were recorded for samples
as thin films or as KBr discs. NMR spectra were recorded
on Bruker ACF-300 and DRX 400 spectrometers with either
TMS or residual protic solvent as internal reference. NOE differ-
ence experiments were carried out using the Bruker micro-
programme ‘noemul’ with presaturation to irradiate different
frequencies within one multiplet. The preirradiation time was
4 s in total and average degree of saturation 80%. The numeri-
cal values of enhancements quoted were not corrected for
incomplete saturation. Elemental analyses were carried out on
a Perkin-Elmer 2400 CHN elemental analyser. Mass spectra
and accurate masses were recorded under EIϩ or CIϩ conditions
on a VG Analytical ZAB-E instrument at the EPSRC Mass
Spectrometry Centre, University College, Swansea or under EIϩ
conditions on a Kratos Profile HV-3 mass spectrometer.
Optical rotations were determined on the sodium -line (589.3
nm) using an AA-1000 polarimeter. Light petroleum refers to
the fraction with distillation range 40–60 ЊC.
Postulated transition states for IOOC cycloadditions involving -
glucose- and -galactose-derived oximes.
To account for the stereochemical outcome in the -
mannose series, one must consider the relative merits of tran-
sition states C and D. We know from our experimental results
that 18 is formed as the major product (vide supra). Hence, we
suggest that transition state C must represent a lower-activation-
energy pathway to product than transition state D (Scheme 7).
General procedure for the synthesis of oximes
A
stirred solution of requisite aldehyde, hydroxylamine
hydrochloride (3 equiv.), pyridine (3 equiv.) and ethanol was
heated at 60 ЊC for 45 min. On cooling, the ethanol was
removed under reduced pressure and the residue taken up
in diethyl ether and washed with water. The organic layer
was dried over MgSO4, filtered and the solvent removed in
vacuo. Purification of the oxime was accomplished by column
chromatography.
2,3,4-Tri-O-benzyl-5,6-dideoxy-D-xylo-hex-5-enose oxime 4
Treatment of aldehyde 3 (100 mg, 0.24 mmol) with hydroxy-
lamine hydrochloride (50 mg, 0.72 mmol) and pyridine (58 µl,
0.72 mmol) in ethanol (0.5 ml) according to the general pro-
cedure above and subsequent column chromatography (light
petroleum–ethyl acetate; 4:1) provided oxime 4 (82 mg, 79%) as
a colourless, oily, 7:3 mixture of E and Z isomers; νmax (thin
film)/cmϪ1 3357 (OH), 3030, 2868, 1497, 1454, 1086, 1069, 1027,
934, 735, 697; δH (400 MHz; CDCl3) 8.55 (0.3H, br s, OHZ),
8.30 (0.7H, br s, OHE), 7.46 (0.7H, d, J1,2 8.0, H-1E), 7.37–7.27
(15H, m, ArH), 6.93 (0.3H, d, J1,2 6.0, H-1Z), 5.82–5.77 (1H,
m, H-5), 5.27–5.18 (2H, m, 2 × H-6), 4.87–4.54 (4.3H, m,
4 × CH2Ph, H-2Z), 4.45–4.32 (2H, m, 2 × CH2Ph), 4.21 (0.3H,
t, J4,3 = J4,5 = 7.5, H-4Z), 4.18–4.11 (1.4H, m, H-4E, H-2E), 3.79
(0.3H, dd, J2,3 3.0, J4,3 7.5, H-3Z), 3.65 (0.7H, t, J4,3 = J3,2 = 8.0,
H-3E); δC (100 MHz; CDCl3) for major, E-isomer: 150.31 (C-1),
138.3 (C), 138.2 (C), 137.6 (C), 134.9 (C-5), 128.4 (CH), 128.34
(CH), 128.31 (CH), 128.2 (CH), 128.0 (CH), 127.8 (CH), 127.7
(CH), 127.6 (CH), 127.4 (CH), 119.2 (C-6), 82.7 (C-3), 81.1
(C-2), 76.6 (C-4), 75.0 (CH2), 71.2 (CH2), 70.9 (CH2); m/z 431
(Mϩ), 414, 91 (Found: Mϩ, 431.2091. C27H29NO4 requires M,
431.2096).
Scheme 7
Since, transition state D places fewer of the benzyl ether sub-
stituents in pseudo-axial orientations, we are drawn to the
conclusion that polar effects are playing a key role in the
stereochemical outcome of these IOOC reactions. Additional
support for this hypothesis comes from our observations on the
conversion of -glucal-derived oxime 13 into cycloadducts 14
and 15. If simple steric factors were controlling the outcome of
this reaction, we might have expected to observe good levels of
diastereocontrol in this transformation in favour of 14 by pos-
tulating the intermediacy of transition state A (X = H). Of
course, this was not the case, and quite low levels of diastereo-
selectivity were observed (vide supra). Interestingly, we have
observed that the diastereoselectivity observed in the conver-
sion of oxime 21 into cycloadducts 22 and 23 is improved
(22:23 ≈ 10:1) by switching to a more polar solvent such as
DMF (132 h, 110 ЊC). At the present time, the precise nature of
the dipolar or stereoelectronic effects that are exerting a strong
influence on the stereochemical outcome of these reactions
remain unclear and work to unravel the subtleties of this reac-
tion are ongoing.
2,3,4-Tri-O-benzoyl-5,6-dideoxy-D-xylo-hex-5-enose oxime 7
In summary, we have demonstrated that intramolecular
oxime olefin cycloaddition reactions of carbohydrate-derived
precursors can be used to prepare polyhydroxylated hexa-
hydro-1H-cyclopent[c]isoxazoles with good to excellent levels
of stereocontrol which can be further transformed into
aminocyclopentitols.
Treatment of aldehyde 6 (532 mg, 1.16 mmol) with hydroxy-
lamine hydrochloride (242 mg, 3.48 mmol) and pyridine (281 µl,
3.47 mmol) in ethanol (3 ml) according to the general procedure
above and subsequent column chromatography (light
petroleum–ethyl acetate; 3:1) provided oxime 7 (470 mg, 85%)
as a colourless, oily, 7:3 mixture of E and Z isomers; νmax (thin
3352
J. Chem. Soc., Perkin Trans. 1, 1999, 3349–3355

View Article Online
(E)-2,3-O-Isopropylidene-4,5-dideoxy-D-erythro-pent-5-enose
oxime (E)-25 and (Z)-2,3-O-isopropylidene-4,5-dideoxy-D-
erythro-pent-5-enose oxime (Z)-25
film)/cmϪ1 3416 (OH), 3067, 1721 (C᎐O), 1603, 1449, 1260,
᎐
1101, 1065, 1019, 937, 707; δH (300 MHz; CDCl3) 8.15–8.00
(7H, m, ArH, OH), 7.60–7.35 (9.7H, m, ArH, H-1E), 6.79
(0.3H, d, J1,2 5.5, H-1Z), 6.52 (0.3H, dd, J1,2 5.5, J3,2 4.0, H-2Z),
6.13 (0.3H, dd, J3,4 7.0, J3,2 4.0, H-3Z), 6.07 (0.7H, t,
J3,2 = J1,2 = 5.5, H-2E), 6.00–5.90 (2.7H, m, H-3E, H-4, H-5),
5.53–5.32 (2H, m, 2 × H-6); m/z 473 (Mϩ), 334, 308, 281, 105
(Found: Mϩ, 473.1472. C27H23NO7 requires M, 473.1474).
Treatment of aldehyde 24 (280 mg, 1.79 mmol) with hydroxy-
lamine hydrochloride (374 mg, 5.38 mmol) and pyridine (435 µl,
5.38 mmol) in ethanol (4 ml) according to the general pro-
cedure above and subsequent column chromatography (light
petroleum–ethyl acetate; 6:1) provided less polar oxime (Z)-25
(77 mg, 25%) as a white solid, mp 98 ЊC; νmax (KBr)/cmϪ1 3236
(OH), 3097, 1456, 1386, 1373, 1306, 1256, 1217, 1156; δH (300
MHz; CDCl3) 7.86 (1H, br d, OH), 6.82 (1H, d, J1,2 5.0, H-1),
5.77 (1H, m, H-4), 5.42–5.18 (2H, m, 2 × H-5), 5.31 (1H, dd,
J1,2 5.0, J3,2 7.0, H-2), 4.83 (1H, m, H-3), 1.55 (3H, s, CH3), 1.41
(3H, s, CH3); δC (75.4 MHz; CDCl3) 150.4, 133.2, 118.2, 109.6,
78.5, 72.7, 27.6, 25.2; m/z 172 (MHϩ), 156, 115, 98 (Found:
MHϩ, 172.0972. C8H14NO3 requires m/z, 172.0974); and more
polar oxime (E)-25 (174 mg, 57%) as a colourless oil; νmax (thin
film)/cmϪ1 3385 (OH), 1375, 1218, 1053; δH (300 MHz; CDCl3)
7.79 (1H, br s, OH), 7.31 (1H, d, J1,2 8.0, H-1), 5.76 (1H, m, H-
4), 5.47–5.28 (2H, m, 2 × H-5), 4.76 (1H, m, H-3), 4.69 (1H, dd,
J1,2 8.0, J3,2 7.0, H-2), 1.54 (3H, s, CH3), 1.42 (3H, s, CH3); δC
(75.4 MHz; CDCl3) 149.0, 132.2, 119.4, 110.0, 79.1, 75.9, 27.9,
25.4; m/z 172 (MHϩ), 156, 114, 98 (Found: MHϩ, 172.0975).
3,4-Di-O-benzyl-2,5,6-trideoxy-D-threo-hex-5-enose oxime 13
Treatment of aldehyde 12 (360 mg, 1.16 mmol) with hydroxy-
lamine hydrochloride (242 mg, 3.48 mmol) and pyridine (281 µl,
3.48 mmol) in ethanol (3 ml) according to the general procedure
above and subsequent column chromatography (light
petroleum–ethyl acetate; 2:1) provided oxime 13 (352 mg, 93%)
as a colourless, oily, 50:50 mixture of E and Z isomers;
νmax (thin film)/cmϪ1 3324, 3262 (OH), 3027, 2858, 1495, 1449,
1076, 927, 732, 691; δH (300 MHz; CDCl3) 7.42 (0.5H, t,
J1,2 = J1,2Ј = 6.0, H-1E), 7.38–7.25 (10H, m, ArH), 6.79 (0.5H, t,
J1,2 = J1,2Ј = 5.0, H-1Z), 5.84 (1H, m, H-5), 5.40–5.30 (2H, m,
2 × H-6), 4.76–4.56 (3H, m, 3 × CH2Ph), 4.44–4.37 (1H, m,
CH2Ph), 3.94 (1H, m, H-4), 3.75 (0.5H, m, H-3Z), 3.68 (0.5H,
m, H-3E), 2.74–2.33 (2H, 2 × H-2); δC (75.4 MHz; CDCl3) 149.7,
149.5, 138.27, 128.27, 138.22, 134.7, 134.6, 128.4, 128.0, 127.9,
127.8, 127.6, 119.6, 119.5, 82.1, 81.8, 78.9, 78.0, 73.1, 73.0, 70.7,
70.6, 31.2, 27.1, not all signals resolved; m/z 326 (M ϩ Hϩ), 325,
202, 96 (Found: MHϩ, 326.1760. C20H24NO3 requires m/z,
326.1756).
(3aR,4R,5S,6S,6aR)-4,5,6-Tribenzyloxyhexahydro-1H-
cyclopent[c]isoxazole 512a
A stirred solution of oxime 4 (87 mg, 0.20 mmol) in dry toluene
(3 ml) under nitrogen was heated at reflux for 15 h. On cooling,
the solvent was removed in vacuo to give 5 (87 mg, 100%) as a
white solid, mp 108 ЊC (from diethyl ether–hexane) (lit.,12a 104–
106 ЊC), [α]D Ϫ3 (c 0.93, CHCl3) {lit.,12a [α]D Ϫ3.6 (c 0.9)}; νmax
(KBr)/cmϪ1 3206 (NH), 3028, 2874, 1497, 1453, 1360, 1117,
1094, 1069, 1029; δH (400 MHz; C6D6) 7.46–7.12 (15H, m,
ArH), 4.99 (1H, d, J 12.0, CH2Ph), 4.90 (1H, d, J 12.0, CH2Ph),
4.88 (1H, d, J 12.0, CH2Ph), 4.76 (1H, d, J 12.0, CH2Ph), 4.59
(1H, d, J 12.0, CH2Ph), 4.52 (1H, br s, NH), 4.47 (1H, d, J 12.0,
CH2Ph), 4.07 (1H, t, J5,4 = J5,6 = 8.0, H-5), 4.00 (1H, dd, J5,6 8.0,
J6a,6 6.5, H-6), 3.70 (1H, t, J4,5 = J4,3a = 8.0, H-4), 3.65 (1H, d, J3,3Ј
8.5, H-3), 3.35 (1H, dd, J6a,3a 10.0, J6a,6 6.5, H-6a), 2.87 (1H, m,
HЈ-3), 2.49 (1H, m, H-3a); δC (100 MHz; C6D6) 139.3 (C), 139.1
(C), 138.9 (C), 128.3 (CH), 128.27 (CH), 128.24 (CH), 127.74
(CH), 127.68 (CH), 127.63 (CH), 127.5 (CH), 127.4 (CH), 127.3
(CH), 86.1 (C-5 and -6), 85.3 (C-4), 75.4 (C-3), 72.4 (CH2),
72.0 (CH2), 71.8 (CH2), 66.5 (C-6a), 49.7 (C-3a); m/z 431 (Mϩ),
340, 218, 91 (Found: Mϩ, 431.2103; C, 75.16; H, 6.72; N, 3.22%.
Calc. for C27H29NO4 M, 431.2097; C, 75.15; H, 6.77; N, 3.24%).
2,3,4-Tri-O-benzyl-5,6-dideoxy-D-lyxo-hex-5-enose oxime 17
Treatment of aldehyde 16 (168 mg, 0.40 mmol) with hydroxy-
lamine hydrochloride (84 mg, 1.21 mmol) and pyridine (98 µl,
1.24 mmol) in ethanol (0.8 ml) according to the general pro-
cedure above and subsequent column chromatography (light
petroleum–ethyl acetate; 4:1) provided oxime 17 (151 mg, 87%)
as a colourless, oily, 80:20 mixture of E and Z isomers; νmax
(thin film)/cmϪ1 3348, 3031, 2869, 1454, 1089, 1068, 697; δH (300
MHz; CDCl3) 8.80 (0.2H, br s, OHZ), 8.49 (0.8H, br s, OHE),
7.42 (0.8H, d, J1,2 8.0, H-1E), 7.37–7.16 (15H, m, ArH), 6.88
(0.2H, d, J1,2 7.0, H-1Z), 5.86 (1H, m, H-5), 5.40–5.28 (2H, m,
2 × H-6), 5.02 (0.2H, dd, J2,3 4.0, J1,2 7.0, H-2Z), 4.75–4.26
(6H, m, 6 × CH2Ph), 4.22 (0.8H, dd, J2,3 5.0, J1,2 8.0, H-2E),
4.07–3.98 (1H, m, H-4), 3.81–3.73 (1H, m, H-3); δC (75.4 MHz;
CDCl3) for major, E-isomer: 149.9 (C-1), 138.4 (C), 138.3 (C),
137.8 (C), 135.3 (C-5), 128.4 (CH), 128.3 (CH), 128.2 (CH),
128.1 (CH), 127.9 (CH), 127.8 (CH), 127.7 (CH), 127.6 (CH),
127.5 (CH), 119.2 (C-6), 82.6 (C-3), 80.6 (C-4), 76.6 (C-2), 74.8
(CH2), 70.9 (CH2), 70.8 (CH2); m/z 449 (MNH4ϩ), 432 (MHϩ),
106 (Found: MHϩ, 432.2178. C27H30NO4 requires m/z,
432.2175).
(3aR,4R,5S,6S,6aR)-4,5,6-Tribenzoyloxyhexahydro-1H-
cyclopent[c]isoxazole 8
A stirred solution of oxime 7 (456 mg, 0.96 mmol) in dry tolu-
ene (16 ml) under nitrogen was heated at reflux for 15 h. On
cooling, the solvent was removed in vacuo and the residue puri-
fied by column chromatography (light petroleum–ethyl acetate;
1:1) to provide bicycle 8 (273 mg, 60%) as a white solid, mp
207–209 ЊC (from toluene); [α]D22 Ϫ29 (c 1.09, CHCl3); νmax
(KBr)/cmϪ1 3242 (NH), 1721 (C᎐O), 1265, 1106, 702; δ (400
MHz; CDCl3) 8.15–7.95 (6H, m, ArH), 7.60–7.30 (9H, m,
ArH), 6.00 (1H, t, J5,4 = J5,6 = 8.0, H-5), 5.62–5.35 (2H, m, NH,
H-6), 5.34 (1H, dd, J4,5 8.0, J4,3a 6.0, H-4), 4.63 (1H, d, J3,3Ј 9.0,
H-3), 4.25 (1H, dd, J6a,3a 9.5, J6a,6 6.0, H-6a), 3.69 (1H, m, HЈ-3),
3.21 (1H, m, H-3a); δC (100 MHz; CDCl3) 166.2 (C), 165.6 (C),
133.5 (CH), 133.3 (CH), 133.26 (CH), 129.9 (CH), 129.87 (CH),
129.84 (CH), 129.5 (CH), 129.2 (CH), 128.5 (CH), 128.4
(CH), 128.3 (CH), 79.2 (C-4), 78.7 (C-6), 76.8 (C-5), 75.8 (C-
3), 66.0 (C-6a), 50.7 (C-3a), not all carbons resolved; m/z 473
(Mϩ), 105, 77 (Found: Mϩ, 473.1472; C, 68.35; H, 4.77; N,
2.69%. C27H23NO7 requires M, 473.1475; C, 68.49; H, 4.90; N,
2.96%).
2,3,4-Tri-O-benzyl-5,6-dideoxy-L-arabino-hex-5-enose oxime 21
Treatment of aldehyde 20 (265 mg, 0.64 mmol) with hydroxy-
lamine hydrochloride (133 mg, 1.90 mmol) and pyridine (154 µl,
1.90 mmol) in ethanol (1.3 ml) according to the general pro-
cedure above and subsequent column chromatography (light
petroleum–ethyl acetate; 6:1) provided oxime 21 (250 mg, 91%)
as a colourless, oily, 75:25 mixture of E and Z isomers; νmax
(thin film)/cmϪ1 3359 (OH), 3032, 2868, 1495, 1455, 1065, 937,
738, 692; δH (300 MHz; CDCl3) 7.42 (0.75H, d, J1,2 8.0, H-1E),
7.36–7.22 (15H, m, ArH), 6.91 (0.25H, d, J1,2 6.0, H-1Z), 5.99–
5.84 (1H, m, H-5), 5.47–5.35 (2H, m, 2 × H-6), 5.07 (0.25H, dd,
J1,2 6.0, J3,2 3.0, H-2Z), 4.75–4.36 (5H, m, 5 × CH2Ph), 4.25
(0.75H, dd, J2,3 5.0, J1,2 8.0, H-2E), 4.22–4.11 (1.25H, m, CH2Ph,
H-4Z), 4.06 (0.75H, dd, J4,5 7.5, J3,4 7.0, H-4E), 3.80 (0.25H, dd,
J2,3 3.0, J4,3 8.0, H-3Z), 3.66 (0.75H, dd, J4,3 7.0, J3,2 5.0, H-3E);
m/z 431 (Mϩ), 340, 284, 208, 91 (Found: Mϩ, 431.2092.
C27H29NO4 requires M, 431.2096).
᎐
H
J. Chem. Soc., Perkin Trans. 1, 1999, 3349–3355
3353

View Article Online
(3aR,4R,5R,6aS)-4,5-Dibenzyloxyhexahydro-1H-cyclopent-
[c]isoxazole 14 and (3aS,4R,5R,6aR)-4,5-dibenzyloxyhexa-
hydro-1H-cyclopent[c]isoxazole 15
cooling, the solvent was removed in vacuo and the residue puri-
fied by column chromatography (light petroleum–ethyl acetate;
2:1) to give less polar bicycle 22 (167 mg, 71%) as a white solid
mp 114 ЊC; [α]D20 ϩ5 (c 1.04, CHCl3); νmax (KBr)/cmϪ1 3160 (NH),
2924, 2853, 1495, 1444, 1347, 1152, 1091, 1024, 840, 735, 692;
δH (400 MHz; Cl2DCCDCl2; 80 ЊC) 7.40–7.33 (15H, m, ArH),
4.83 (1H, d, J 11.9, CH2Ph), 4.76–4.72 (3H, m, CH2Ph), 4.65
(1H, d, J 12.0, CH2Ph), 4.62 (1H, d, J 11.9, CH2Ph), 4.25 (1H,
dd, J3,3Ј 8.1, J3,3a 5.1, H-3), 4.13 (1H, dd, J4,5 4.1, J3a,4 6.7, H-4),
4.03 (1H, br t, J4,5 = J5,6 ≈ 4.1, H-5), 3.99–3.95 (1H, dd, J6,6a 3.2,
J5,6 5.0, H-6), 3.80 (1H, t, J3,3Ј = J3Ј,3a = 8.1, HЈ-3), 3.76 (1H, dd,
J6a,3a 8.6, J6a,6 3.2, H-6a), 3.20 (1H, m, H-3a); δC (100 MHz;
Cl2DCCDCl2; 80 ЊC) 138.6 (C), 138.5 (C), 138.4 (C), 128.5
(CH), 127.7 (CH), 127.65 (CH), 127.6 (CH), 85.2 (C-5), 83.4
(C-6), 78.4 (C-4), 73.4 (CH2), 72.9 (CH2), 72.2 (CH2), 71.2
(C-3), 69.3 (C-6a), 48.6 (C-3a), not all aromatic carbons
resolved; m/z 431 (Mϩ), 397, 369, 242, 91 (Found: Mϩ,
431.2106; C, 75.13; H, 6.77; N, 3.12%. C27H29NO4 requires M,
431.2097; C, 75.15; H, 6.77; N, 3.24%); and more polar bicycle
23 (27 mg, 12%) as a colourless oil; νmax (thin film)/cmϪ1 3221
(NH), 3062, 3037, 2868, 1501, 1449, 1357, 1209, 1132, 732, 697;
δH (400 MHz; C6D6; 60 ЊC) 7.45–7.10 (15H, m, ArH), 4.69 (1H,
d, J 12.0, CH2Ph), 4.68 (1H, d, J 12.0, CH2Ph), 4.62 (2H, m,
J 12.0, 2 × CH2Ph), 4.56 (1H, d, J 12.0, CH2Ph), 4.49 (1H, d,
J 12.0, CH2Ph), 4.30 (1H, dd, J5,6 8.0, J6,6a 6.7, H-6), 4.06 (1H,
dd, J4,5 4.0, J5,6 8.0, H-5), 3.74 (1H, dd, J3a,6a 8.0, J6,6a 6.7, H-6a),
3.59 (1H, dd, J4,5 4.0, J3a,4 1.9, H-4), 3.50 (1H, br t, H-3), 3.35
(1H, dd, J3,3Ј 8.8, J3Ј,3a 2.9, HЈ-3), 2.85 (1H, m, H-3a); δC (100
MHz; C6D6; 30 ЊC) 139.3 (C), 139.1 (C), 138.9 (C), 128.3 (CH),
128.0 (CH), 127.9 (CH), 127.8 (CH), 127.5 (CH), 127.4 (CH),
84.0 (C-5), 81.4 (C-6), 80.5 (br, C-4), 74.0 (br, C-3), 72.4
(2 × CH2), 72.1 (CH2), 61.0 (C-6a), 51.1 (C-3a), not all aromatic
carbons resolved; m/z 431 (Mϩ), 217, 91 (Found: Mϩ, 431.2092.
C27H29NO4 requires M, 431.2097).
A stirred solution of oxime 13 (87 mg, 0.27 mmol) in dry tolu-
ene (4 ml) under nitrogen was heated at reflux for 36 h. On
cooling, the solvent was removed in vacuo and the residue puri-
fied by column chromatography (light petroleum–ethyl acetate;
1:2) to give an inseparable 2:1 mixture of bicycles 14 and 15
(67 mg, 71%) as a colourless oil; stereochemistry of major
adduct could not be determined; νmax (thin film)/cmϪ1 3216,
2929, 2858, 1449, 1357, 1096, 691; δH (400 MHz; C6D6; 80 ЊC)
7.45–7.10 (10H, m, ArH), 4.62–4.40 (4.33H, m, 4 × CH2Ph, H-
3min), 4.13 (0.33H, dt, J5,4 = J5,6 = 7.0, J5,6Ј 9.0, H-5min), 3.85
(0.66H, dt, J5,4 = J5,6Ј 6.0, J5,6 7.0, H-5maj), 3.81–3.75 (1H, m,
H-4), 3.72 (0.66H, dd, J3,3Ј 8.5, J3,3a 4.0, H-3maj), 3.62–3.46
(1.32H, m, H-3maj, H-6amaj), 3.40 (0.33H, m, J6a,3a = J6a,6 = 9.0,
J6a,6Ј 2.0, H-6amin), 3.13 (0.33H, t, J3Ј,3a = J3Ј,3 = 8.0, HЈ-3min),
2.65–2.61 (1H, m, H-3a), 2.20–2.08 (1H, m, H-6), 1.78 (0.66H,
m, HЈ-6maj), 1.71 (0.33H, dt, J6,6Ј 14.0, J5,6Ј = J6a,6Ј = 9.0, HЈ-
6min); δC (100 MHz; C6D6; 30 ЊC) 139.4 (C), 139.0 (C), 138.9 (C),
138.8 (C), 128.36 (CH), 128.33 (CH), 128.2 (CH), 128.0 (CH),
127.9 (CH), 127.8 (CH), 127.7 (CH), 127.6 (CH), 127.5 (CH),
127.49 (CH), 127.46 (CH), 127.25 (CH), 127.19 (CH), 127.16
(CH), 88.0 (C-4maj), 84.0 (C-4min), 83.6 (C-5maj), 82.0 (C-5min),
75.7 (C-3maj), 71.96 (CH2), 71.86 (CH2), 71.7 (CH2), 71.5 (C-
3min), 71.1 (CH2), 62.1 (C-6amaj), 60.4 (C-6amin), 53.2 (C-3amaj),
48.5 (C-3amin), 36.2 (C-6min), 34.3 (C-6maj); m/z 326 (MHϩ), 106
(Found: MHϩ, 326.1757. C20H24NO3 requires m/z, 326.1756).
(3aS,4R,5S,6R,6aS)-4,5,6-Tribenzyloxyhexahydro-1H-
cyclopent[c]isoxazole 18 and (3aR,4R,5S,6R,6aR)-4,5,6-
tribenzyloxyhexahydro-1H-cyclopent[c]isoxazole 19
A stirred solution of oxime 17 (410 mg, 0.95 mmol) in dry
toluene (14 ml) under nitrogen was heated at reflux for 60 h.
On cooling, the solvent was removed in vacuo and the residue
purified by column chromatography (toluene–acetone;
10:1 → 5:1) to give less polar bicycle 18 (263 mg, 64%) as a
colourless oil; [α]D23 Ϫ18 (c 0.76, CHCl3); νmax (thin film)/cmϪ1
3219 (NH), 3063, 3030, 2863, 1496, 1453, 1366, 1135, 1072,
1027, 735, 697; δH (400 MHz; C6D6) 7.45–7.10 (15H, m, ArH),
4.70 (1H, br s, NH), 4.65 (1H, d, J 12.0, CH2Ph), 4.64 (1H, d,
J 12.0, CH2Ph), 4.59 (1H, d, J 12.0, CH2Ph), 4.55 (3H, m,
H-3, 2 × CH2Ph), 4.46 (1H, d, J 12.0, CH2Ph), 4.35 (1H, t,
J4,5 = J4,3a = 8.0, H-4), 4.11 (1H, dd, J5,4 8.0, J5,6 4.5, H-5), 3.98
(1H, dd, J5,6 4.5, J6a,6 1.5, H-6), 3.49 (1H, br d, J6a,3a 8.0, H-6a),
2.93 (1H, dd, J3Ј,3a 7.5, J3,3Ј 8.5, HЈ-3), 2.74 (1H, m, H-3a); δC
(100 MHz; C6D6) 139.3 (C), 139.1 (C), 139.0 (C), 128.33 (CH),
128.30 (CH), 128.2 (CH), 127.7 (CH), 127.6 (CH), 127.5 (CH),
127.4 (CH), 127.3 (CH), 83.8 (C-5), 81.9 (C-6), 81.3 (C-4), 72.38
(CH2), 72.35 (CH2), 72.26 (CH2), 71.0 (C-3), 66.1 (C-6a), 46.8
(C-3a); m/z 431 (Mϩ), 346, 91 (Found: Mϩ, 431.2091. C27H29-
NO4 requires M, 431.2097); and more polar bicycle 19 (50 mg,
12%) as a colourless oil; νmax (thin film)/cmϪ1 3231 (NH), 3062,
3032, 2924, 2863, 1495, 1449, 1357, 1091, 738, 691; δH (400
MHz; C6D6) 7.50–7.10 (15H, m, ArH), 4.78 (1H, m, CH2Ph),
4.62–4.49 (3H, m, 3 × CH2Ph), 4.30 (2H, m, 2 × CH2Ph), 4.07–
3.96 (2H, m, H-3, H-4), 3.93–3.81 (3H, m, H-5, H-6, H-6a),
3.55 (1H, m, H-3), 2.80 (1H, m, H-3a); δC (100 MHz; C6D6)
138.6 (C), 138.5 (C), 129.1 (CH), 128.3 (CH), 128.0 (CH), 127.9
(CH), 127.8 (CH), 127.7 (CH), 127.5 (CH), 125.4 (CH), 84.9
(C-6), 83.9 (C-5), 77.9 (C-4), 75.1 (C-3), 72.9 (CH2), 72.7 (CH2),
71.7 (CH2), 64.3 (C-6a), 53.2 (C-3a), not all aromatic signals
resolved; m/z 431 (Mϩ), 401, 91 (Found: Mϩ, 431.1727)
(3aR,4R,5S,6S,6aR)-4,5,6-Trihydroxyhexahydro-1H-
cyclopent[c]isoxazole 912a
A solution of sodium methoxide (produced from fresh sodium
shavings) in methanol (3 ml) was added to bicycle 8 (146 mg,
0.31 mmol) at 0 ЊC. The mixture was allowed to warm to room
temperature, stirred for 1 h, then neutralised with Dowex 50
(Hϩ) resin, filtered, and concentrated in vacuo. Column chrom-
atography (CH2Cl2–MeOH; 1:1 containing 2% of 32% aq.
ammonia) provided triol 9 (30 mg, 60%) as a white solid; [α]D20
Ϫ60 (c 0.50, MeOH) {lit.,12a [α]D Ϫ58.3 (c 1.2, MeOH)}; δH (300
MHz; CD3OD) 3.97 (1H, d, J 9.0), 3.61–3.37 (5H, m), 2.55
(1H, m); δC (75.4 MHz; CD3OD) 81.2, 79.7, 79.3, 75.8, 67.4,
51.9; m/z 161 (Mϩ), 122, 100, 70 (Found: Mϩ, 161.0687. Calc.
for C6H11NO4: M, 161.0688).
(1R,2S,3S,4R,5R)-1-Amino-2,3,4-trihydroxy-5-
(hydroxymethyl)cyclopentane 1012a
A solution of bicycle 9 (25 mg, 0.15 mmol) in MeOH (2.5 ml)
was hydrogenated in the presence of 10% Pd on carbon (spatula
tip) at room temperature under one atmospheric pressure of
hydrogen overnight. The catalyst was removed by filtration
through Celite and the filtrate concentrated in vacuo to give
cyclopentane 10 (23 mg, 91%) as a white solid; [α]D20 ϩ2 (c 0.44,
MeOH); δH (300 MHz; D2O) 3.78 (3H, m), 3.70 (1H, t, J 8.0),
3.60 (1H, t, J 8.0), 3.30 (1H, dd, J 8.0, J 10.0), 2.15 (1H, m); δC
(75.4 MHz; D2O) 79.7, 79.4, 74.1, 58.5, 53.3, 44.8; m/z 163
(Mϩ), 122, 105, 73.
(3aR,4S,5S,6S,6aR)-4,5,6-Tribenzyloxyhexahydro-1H-
cyclopent[c]isoxazole 22 and (3aS,4S,5S,6S,6aS)-4,5,6-
tribenzyloxyhexahydro-1H-cyclopent[c]isoxazole 23
Acknowledgements
We are grateful to the Leverhulme Trust for financial sup-
port. We are also indebted to the National EPSRC Mass
Spectrometry Service for some of the high-resolution mass
spectra.
A stirred solution of oxime 21 (234 mg, 0.54 mmol) in dry
toluene (8 ml) under nitrogen was heated at reflux for 24 h. On
3354
J. Chem. Soc., Perkin Trans. 1, 1999, 3349–3355

View Article Online
10 (a) A. Arnone, M. Cavicchioli, A. Donadelli and G. Resnati,
Notes and references
1 T. Aoyagi, T. Yamamoto, K. Kojiri, H. Morishima, M. Nagai,
M. Hamada, T. Takeuchi and H. Umezawa, J. Antibiot., 1989, 42,
883.
Tetrahedron; Asymmetry, 1994, 5, 1019; (b) S. Baskaran and H. G.
Aurich, Synlett, 1998, 277.
11 G. V. M. Sharma, I. Srinivas Reddy, V. Goverdhan Reddy and A. V.
Rama Roa, Tetrahedron: Asymmetry, 1999, 10, 229.
12 (a) B. Bernet and A. Vasella, Helv. Chim. Acta, 1979, 62, 1990; (b)
J. K. Gallos, A. E. Koumbis and N. E. Apostolakis, J. Chem. Soc.,
Perkin Trans. 1, 1997, 2457.
13 All the aldehydes used in this study were made from the appropriate
6-deoxy-6-iodo carbohydrate derivative by treatment with activated
zinc dust in ethanol. The original method as described by Vasella
(ref. 12a) used the corresponding 6-bromo-6-deoxy compounds.
14 A. Fuerstner, D. Jumbam, J. Teslic and H. Weidmann, J. Org.
Chem., 1991, 56, 2213.
15 R. J. Ferrier, R. H. Furneaux, P. Prasit, P. C. Tyler, K. L. Brown,
G. J. Gainsford and J. W. Diehl, J. Chem. Soc., Perkin Trans. 1, 1983,
1621.
2 O. Ando, H. Satake, K. Itoi, A. Sato, M. Nakajima, S. Takahashi,
H. Haruyama, Y. Ohkuma, T. Kinoshita and R. Enokita,
J. Antibiot., 1991, 44, 1165.
3 For
a recent review, see A. Berecibar, C. Grandjean and
A. Siriwardena, Chem.Rev., 1999, 99, 779.
4 For a preliminary account of part of this work, see S. Moutel and
M. Shipman, Synlett, 1998, 1333.
5 W. Oppolzer and K. Keller, Tetrahedron Lett., 1970, 1117.
6 M. R. Slabaugh and W. C. Wildman, J. Org. Chem., 1971, 36, 3202.
7 (a) R. Grigg and S. Thianpantangul, J. Chem. Soc., Perkin Trans. 1,
1984, 653; (b) R. Grigg, J. Markandu, T. Perrior, S. Surendrakumar
and W. J. Warnock, Tetrahedron Lett., 1990, 31, 559; (c) R. Grigg,
F. Heaney, J. Markandu, S. Surendrakumar, M. Thornton-Pett and
W. J. Warnock, Tetrahedron, 1991, 47, 4007.
16 J. J. C. Grové, C. W. Holzapfel and D. B. G. Williams, Tetrahedron
Lett., 1996, 37, 1305.
8 M. H. Norman and C. H. Heathcock, J. Org. Chem., 1987, 52, 226.
9 (a) A. Padwa, U. Chiacchio, D. C. Dean, A. M. Schoffstall, A.
Hassner and K. S. K. Murthy, Tetrahedron Lett., 1988, 29, 4169; (b)
A. Hassner, R. Maurya and E. Mesko, Tetrahedron Lett., 1988, 29,
5313; (c) A. Hassner and R. Maurya, Tetrahedron Lett., 1989, 30,
2289; (d) A. Hassner and R. Maurya, Tetrahedron Lett., 1989, 30,
5803; (e) A. Hassner, R. Maurya, A. Padwa and W. H. Bullock, J.
Org. Chem., 1991, 56, 2775; ( f ) A. Hassner, R. Maurya, O.
Friedman, H. E. Gottleib, A. Padwa and D. Austin, J. Org. Chem.,
1993, 58, 4539.
17 J. Désiré and J. Prandi, Tetrahedron Lett., 1997, 38, 6189.
18 NOEs of similar magnitude were observed between such trans
hydrogens in some of the other cycloadducts. For example, for
bicycle 8, small NOEs were measured between H-6 and H-6a
(H-6 → H-6a {3.9%}; H-6a → H-6 {3.7%}).
19 L. A. Paquette and S. Bailey, J. Org. Chem., 1995, 60, 7849.
20 For a discussion and leading reference, see ref. 10a.
Paper 9/05706D
J. Chem. Soc., Perkin Trans. 1, 1999, 3349–3355
3355

View Article Online