Synthesis of (+)-Casuarine

Article abstract of DOI:10.1021/jo991680v

The first synthesis of (+)-casuarine ((+)-6), a pentahydroxy pyrrolizidine alkaloid of the alexine/ australine subclass, is described. The key step is a tandem [4 + 2]/[3 + 2] nitroalkene cycloaddition involving nitrobenzoate 13, chiral vinyl ether 16c, and vinyl silane 10, which establishes five of the six stereocenters present in this potent glycosidase inhibitor. The completion of the synthesis requires only four additional steps to deliver the final product in 20% overall yield.

Full text of DOI:10.1021/jo991680v

J . Org. Chem. 2000, 65, 2875-2886
2875
Syn th esis of (+)-Ca su a r in e
Scott E. Denmark* and Alexander R. Hurd
Roger Adams Laboratory, Department of Chemistry, University of Illinois, Urbana, Illinois 61801
Received October 26, 1999
The first synthesis of (+)-casuarine ((+)-6), a pentahydroxy pyrrolizidine alkaloid of the alexine/
australine subclass, is described. The key step is a tandem [4 + 2]/[3 + 2] nitroalkene cycloaddition
involving nitrobenzoate 13, chiral vinyl ether 16c, and vinyl silane 10, which establishes five of
the six stereocenters present in this potent glycosidase inhibitor. The completion of the synthesis
requires only four additional steps to deliver the final product in 20% overall yield.
In tr od u ction a n d Ba ck gr ou n d
conjugates have also been isolated.7 Potential anti-cancer8
and anti-viral9 properties of polyhydroxyl pyrrolizidine
and indolizidine alkaloids stem from their resemblance
to sugars and, not surprisingly, many display significant
glycosidase inhibition activity.10 In a comparative study,
all australines and alexines are potent inhibitors of
fungal glucan 1,4-R-glucosidase and compare in activity
to the well-known glycosidase inhibitor castanospermine
(7).4 Additionally, both 7-epiaustraline (4) and 1-epiaus-
traline (5) are active in glucosidase activity mouse
models4 and have been patented.11
The alexines and australines are a growing subclass
of pyrrolizidine alkaloids, many of which display potent
glycosidase activity. Named after the first two identified
members of this new class, alexine (1)1 and australine
(2)2 bear a C(3) hydroxymethyl substituent, differentiat-
ing them from the more common necines which bear a
C(1) hydroxymethyl group, Chart 1. Since the isolation
Ch a r t 1
(+)-Casuarine (6) is the most recently isolated member
of this class. It was obtained from 75% aqueous ethanolic
extract of the bark of Casuarina equisetfolia L. (Casuari-
naceae) in 0.013% yield.6 A glycoside of casuarine, casu-
arine-6-R-D-glucoside, was also isolated from this plant
extract in nearly equal amount.7 The structure and the
absolute configuration of casuarine were established un-
ambiguously by single-crystal X-ray spectroscopy as (+)-
6. Casuarine is an effective inhibitor of glucosidase I (72%
inhibition at 5µg/mL) rivaling that of the indolizidine
alkaloid castanosperime (84% inhibition at 5µg/mL).12
The five hydroxyl groups make (+)-casuarine the most
highly oxygenated amino-sugar analogue yet identified.13
Although there are no syntheses of casuarine (6) on
record, the synthesis of four diastereomers was recently
disclosed.12 This work represents an effective method for
obtaining these diastereomers, but it is inflexible in its
design to access other diastereomers, as selective protec-
tion, deprotection, and activation schemes would render
the approach cumbersome and lengthy. The alternative,
of alexine (1) and australine (2) in 1988, the identification
of other members of this class quickly followed, including
3-epiaustraline (3),3 7-epiaustraline (4),4 1-epiaustraline
(5),4,5 and the hydroxylated derivative casuarine (6).6
Unlike the necine alkaloids, which are typically isolated
with an acid side chain, the alexine/australine pyrrolizi-
dine alkaloids are isolated as the base, although glyco-
(7) Wormald, M. R.; Nash, R. J .; Watson, A. A.; Bhadoria, B. K.;
Langford, R.; Sims, M.; Fleet, G. W. J . Carbohydr. Lett. 1996, 2, 169.
(8) (a) Dennis, J . W. Cancer Res. 1986, 46, 5131. (b) Humphries, K.
J .; Matsumoto, K.; White, S.; Olden, K. Cancer Res. 1986, 46, 5212.
(c) Ostrander, G. K.; Scribner, N. K.; Rohrschneider, L. R. Cancer Res.
1988, 48, 1091.
(9) (a) Gruters, R. A.; Neefjes, J . J .; Termette, M.; de Goede, R. E.
Y.; Tulp, A.; Huisman, H. G.; Miedema, F.; Ploegh, H. L. Nature 1987,
330, 74. (b) Walker, B. D.; Kowalski, M.; Goh, W. C.; Kozarsky, K.;
Krieger, M.; Rosen, C.; Rohrschneider, L.; Haseltine, W. A.; Sodroski,
J . Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 8210. (c) Ratner, L. AIDS
Res. Hum. Retroviruses 1992, 8, 165.
(10) Glycosidase inhibition by polyhydroxylated alkaloids has been
reviewed. Fellows, L. E.; Nash, R. J . Sci. Prog. (Oxford) 1990, 74, 245.
(11) (a) Elbein, A. D.; Tropea, J . E.; Molyneux, R. J . US pat. Appl.
US 289,907; Chem. Abstr. 1990, 113, 91444. (b) Fellows, L. E.; Nash,
R. J . PCT Int. Appl. WO 90, 12014; Chem. Abstr. 1991, 114, 143777.
(12) Bell, A. A.; Pickering, L.; Watson, A. A.; Nash, R. J .; Pan, Y.
T.; Elbein, A. D.; Fleet, G. W. J . Tetrahedron Lett. 1997, 38, 5869.
(13) The isolation of another pentahydroxylated pyrrolizidine alka-
loid, hyacinthacine C1, was recently reported. Kato, A.; Adachi, I.;
Miyauchi, M.; Ikeda, K.; Komae, T.; Kizu, H.; Kameda, Y.; Watson, A.
A.; Nash, R. J .; Wormald, M. R.; Fleet, G. W. J .; Asano, N. Carbohydr.
Res. 1999, 316, 95.
(1) Nash, R. J .; Fellows, L. E.; Dring, J . V.; Fleet, G. W. J .; Derome,
A. E.; Hamor, T. A.; Scofield, A. M.; Watkin, D. J . Tetrahedron Lett.
1988, 29, 2487.
(2) Molyneux, R. J .; Benson, M.; Wong, R. Y.; Tropea, J . E.; Elbein,
A. D. J . Nat. Prod. 1988, 51, 1198.
(3) Nash, R. J .; Fellows, L. E.; Plant, A. C.; Fleet, G. W. J .; Derome,
A. E.; Baird, P. D.; Hearty, M. P.; Scofield, A. M. Tetrahedron 1988,
44, 5959.
(4) Nash, R. J .; Fellows, L. E.; Dring, J . V.; Fleet, G. W. J .; Girdhar,
A.; Ramsden, N. G.; Peach, J . M.; Hegarty, M. P.; Scofield, A. M.
Phytochemistry 1990, 29, 111.
(5) Harris, C. M.; Harris, T. M.; Molyneux, R. J .; Tropea, J . E.;
Elbein, A. D. Tetrahedron Lett. 1989, 30, 5685.
(6) Nash, R. J .; Thomas, P. I.; Waigh, R. D.; Fleet, G. W. J .; Wormald,
M. R.; Lilley, P. M de Q.; Watkin,. D. J . Tetrahedron Lett. 1994, 35,
7849.
10.1021/jo991680v CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/21/2000

2876 J . Org. Chem., Vol. 65, No. 10, 2000
Denmark and Hurd
beginning with materials possessing the correct hydroxyl
configurations, may be limited by the availability of
carbohydrate sources.
followed by Tamao-Fleming oxidation18 of the silyl
moiety. Thus, nitroso acetal 8 possesses all the required
stereocenters and functionality for the synthesis of (+)-
casuarine. With the exception of C(7) (nitroso acetal
numbering), all the remaining stereocenters are estab-
lished in the tandem cycloaddition. Ketone 9 was targeted
as the key intermediate, as it would allow a substrate
controlled reduction for the creation of the center at C(7)
and would arise from a tandem inter [4 + 2]/inter
[3 + 2] cycloaddition of components 10, 12, and 13.
A predictable stereochemical outcome of the tandem
event is critical for the successful synthesis of (+)-
casuarine. The desired approach of the dipolarophile
1014g,19 in the [3 + 2] cycloaddition with nitronate 11 was
difficult to predict and could hinge on the placement of
the C(6) chiral auxiliary. In cases where the C(4) ben-
zoate and the C(6) acetal center are positioned on the
same side of the nitronate ring, as in the synthesis of
hastanecine14a and 7,7a-diepiaustraline,14g,19 the approach
of the dipolarophile is exclusively on the side opposite
these substituents. However, in the absence of a C(4)
substituent, the C(6) acetal center alone can direct the
facial attack of the dipolarophile to the opposite side.20
Additionally, the presence of a C(5) benzoate group could
exert considerable influence on the reactive conformation
of the nitronate. In fact, studies on heavily substituted
nitronates show a strong sensitivity of facial approach
to nitronate conformation.21 Since the configuration of the
C(6) auxiliary substituent is of no consequence to the
synthesis of the natural product, as it is destroyed
ultimately, the placement of the C(6) alkoxy ligand could
be used advantageously to direct the facial attack of the
dipolarophile in the desired manner, if necessary.
Recent publications from these laboratories have docu-
mented the use of tandem [4 + 2]/[3 + 2] nitroalkene
cycloadditions for the efficient syntheses of several poly-
hydroxylated pyrrolizidine and indolizidine alkaloids.14
These studies clearly demonstrated the ability of the
tandem cycloaddition strategy to assemble these skel-
etons with excellent control over as many as four
contiguous stereogenic centers. (+)-Casuarine (6) with its
densely oxygenated periphery represented a still greater
challenge to install five of the six-contiguous stereo-
centers by the cycloaddition construct. Of further interest
to the development of the [4 + 2]/[3 + 2] nitroalkene
cycloaddition method was that the synthesis of (+)-
casuarine required an intermolecular [3 + 2] cycloaddi-
tion of a suitable substituted dipolarophile and a flexible,
heavily substituted nitronate. The study of intermolecu-
lar [3 + 2] dipolar cycloadditions of conformationally
flexible nitronates has been limited. The successful
resolution of this challenge in the synthesis of (+)-
casuarine is detailed below.15
Syn th esis Design
To accommodate the installation of so many hydroxyl
groups in a stereodefined fashion, a suitably functional-
ized nitroalkene, dieneophile, and dipolarophile were
needed. From previous endeavors we knew that oxygen-
ated nitroalkenes14 and dieneophiles16 were amenable to
[4 + 2] cycloaddition, but oxygenated dipolarophiles gave
regioreversed cycloadducts.17 Thus, we made recourse to
the phenyldimethylsilyl substituent as a surrogate for
the hydroxyl group at C(1).18
One option employs a silyl substituent as a hydroxyl
surrogate on the dieneophile in the [4 + 2] cycloaddition.
The obvious choice to secure high facial selectivity in the
[3 + 2] cycloaddition was to use the vinyl silane (E)-1414c
in an endo (alkoxy) cycloaddition. This would produce
nitronate 15 with the necessary C(4)/C(5) trans arrange-
ment and with the C(4) and (C6) substituents on the
same side of the nitronate ring, Scheme 2. Currently, the
only available Lewis acid to achieve endo (alkoxy) selec-
tive cycloadditions22 is TiCl2(Oi-Pr)2, and its compat-
ibility with silane (E)-14 is unknown. Alternatively,
[4 + 2]cycloaddition of nitroalkene 1314a and acyloxy vinyl
ether (Z)-1616 should provide nitronate 17. The required
C(4)/C(5) trans relationship was expected, given the high
exo selectivity previously demonstrated with this chiral
vinyl ether.16 As a consequence of the Z-dieneophile, the
R-orientation of the C(6) auxiliary would be expected.
Since the C(4) benzoate and the C(6) acetal center are
on opposite faces of the nitronate, the dominant control-
ling element for the [3 + 2] cycloaddition a priori was
Sch em e 1
(15) A preliminary report of this work has appeared: Denmark, S.
E.; Hurd, A. R. Org. Lett. 1999, 1, 1311.
(16) Denmark, S. E.; Schnute, M. E. J . Org. Chem. 1994, 59, 4576.
(17) Denmark, S. E.; Seierstad, M.; Herbert B. J . Org. Chem. 1999,
64, 884.
This synthetic design is outlined in Scheme 1. It is
expected that (+)-casuarine would be produced from
nitroso acetal 8 by N-O bond cleavage and N-alkylation
(18) (a) Tamao, K.; Ishida, N.; Tanaka, T.; Kumada, M. Organome-
tallics 1983, 2, 1694. (b) Tamao, K.; Ishida, N. J . Organomet. Chem.
1984, 269, C37. (c) Fleming, I.; Henning, R.; Plaut, H. J . Chem. Soc.,
Chem. Commun. 1984, 29. (d) Fleming, I.; Henning, R.; Parker, D. C.;
Plaut, H. E.; Sanderson, P. E. J . J . Chem. Soc., Perkin Trans. 1 1995,
317.
(19) Denmark, S. E.; Herbert, B. J . Org. Chem. 2000, 65, 2887.
(20) Denmark, S. E.; Hurd, A. R. J . Org. Chem. 1998, 63, 3045.
(21) Schnute, M. E. Ph.D. Thesis, University of Illinois, Urbana,
1995.
(14) For previous natural product synthesis using tandem nitroalk-
ene cycloadditions, see: (a) Denmark, S. E.; Thorarensen, A. J . Org.
Chem. 1994, 59, 5672. (b) Denmark, S. E.; Thorarensen, A.; Middleton,
D. S. J . Am. Chem. Soc. 1996, 118, 8266. (c) Denmark, S. E.; Thorar-
ensen, A. J . Am. Chem. Soc. 1997, 119, 125. (d) Denmark, S. E.; Parker,
D. L.; Dixon, J . A. J . Org. Chem. 1997, 62, 435. (e) Denmark, S. E.;
Hurd, A. R.; Sacha, H. J . J . Org. Chem. 1997, 62, 1668. (f) Denmark,
S. E.; Marcin, L. R. J . Org. Chem. 1997, 62, 1675. (g) Denmark, S. E.;
Herbert, B. J . Am. Chem. Soc. 1998, 120, 7357. (h) Denmark, S. E.;
Martinborough, E. A. J . Am. Chem. Soc. 1999, 121, 3046.
(22) Denmark, S. E.; Seierstad, M. J . Org. Chem. 1999, 64, 1610.

Synthesis of (+)-Casuarine
Sch em e 2
J . Org. Chem., Vol. 65, No. 10, 2000 2877
Sch em e 3
uncertain. Additionally, previous studies have shown that
this dieneophile is poorly selective with nitroalkenes
lacking a C(1) substituent, and the use of SnCl4, neces-
sary for this dieneophile, has little precedent for oxygen-
ated nitroalkenes. Nonetheless, the acyloxy dieneophiles
would be the most fitting choice, as the desired hydroxy
group is liberated by simple saponification of the ester,
avoiding an additional Tamao-Fleming oxidation.
faces of the vinyl ether. Nitronate 17 proved to be too
unstable for isolation and purification, and was im-
mediately treated with a solution of the â-silyl enone 1019
to afford a 8/2/1/1 mixture (1H NMR analysis) of nitroso
acetals 18 in 55% yield over two steps. This ratio was
determined by integration of the silyl bearing methine
resonance and was judged to be of the same regiochem-
istry. Since this mixture contained diastereomers created
in the [4 + 2] cycloaddition, the [3 + 2] cycloaddition was
judged to be facially selective, on the order of 8/1. After
separation of the major diastereomer by preparative
HPLC, the structure of this component was ultimately
confirmed by conversion to the natural product. The
important finding in these sets of experiments was that
the C(4) benzoate was the dominant controlling element
in the [3 + 2] cycloaddition with vinyl silane dipolarophile
10. This was surprising given the placement of the C(4)
benzoate and the C(6) auxiliary on opposite sides of the
nitronate ring. Nonetheless, the approach for the syn-
thesis was now deemed feasible using the acyloxy di-
eneophiles.
Resu lts
Both options for the assembly of a nitronate suitable
for the synthesis of casuarine were initially explored. The
vinyl silane (E)-14, which was developed for the synthesis
of (+)-crotanecine14c first was examined in combination
with Ti(Oi-Pr)2Cl2, as this Lewis acid has been used for
cycloadditions with oxygenated nitroalkenes.14 The ad-
dition of Ti(Oi-Pr)2Cl2 to a cold (-78 °C) solution of 14
and nitroalkene 13 resulted in an immediate, deep red
coloration. Upon workup, the only isolated material was
the chiral alcohol, phenylcyclohexanol (19), Scheme 3.
Variations in order of addition gave the same result.
Since Ti(Oi-Pr)2Cl2 is the only endo selective Lewis acid
for the nitroalkene cycloaddition method, the use of silane
14 was abandoned.23
Orienting experiments were also conducted using the
acyloxy vinyl ethers (Z)-1616 and nitroalkene 1314a under
previously optimized conditions for this dieneophile for
the purpose of assaying the facial selectivity in the
[3 + 2] cycloaddition, Scheme 3. The acetoxy vinyl ether
(Z)-16a , and the pivaloyloxy vinyl ether (Z)-16b were
both unsuitable for a productive cycloaddition. Instead,
extensive polymerization was noticed. The benzoyloxy
vinyl ether (Z)-16c produced nitronate 17 as a 4/1
mixture (1H NMR analysis) of diastereomers, Scheme 3.
Examination of proton NMR coupling constants indicated
that both diastereomers possessed the desired C(4)/C(5)
trans relationship and were thus diastereomers resulting
from attack of the nitroalkene to the two diastereotopic
In addition to the obvious improvements needed for
both yield and facial selectivity in the [4 + 2] cycloaddi-
tion, the existing synthesis16 of the benzoate (Z)-16c was
poor yielding (17%), and a new route had to be developed,
Scheme 4. Low temperature and neutral conditions were
Sch em e 4
(23) The combination of the vinyl silane (Z)-14 with the Lewis acid
MAPh did produce a nitronate, but this decomposed before [3 + 2]
cycloaddition with vinyl silane 10.

2878 J . Org. Chem., Vol. 65, No. 10, 2000
Denmark and Hurd
important for the new O-acylation method in order to
minimize polymerization of the base sensitive R-alkoxy-
aldehyde. A two step protocol was developed which
ensured that the deprotonated aldehyde was not present
in the reaction mixture at the same time as the free
aldehyde. The chiral alkoxy aldehyde 2016 was converted
to silyl enol ether 21 in 99% yield as a 10/1 (Z/E) mixture.
O-Acylation with benzoyl fluoride and a catalytic amount
of TBAF (2 mol %),24 afforded a separable mixture of the
Z-vinyl ether (Z)-16c in 81% yield along with 6% of the
undesired E-vinyl ether (E)-16c.
2.5 equiv of SnCl4 (NaOH/MeOH, 10 equiv) partially
converted the nitronate (up to 29%) to oxime 22. When
the reaction was conducted in toluene instead of CH2Cl2
and NaOH/MeOH solution was freshly prepared (entry
2), the yield was increased to 63% along with 18% of the
oxime. This indicated that an accurate titration of the
NaOH/MeOH solution was important for diverting the
reaction away from the oxime pathway. In fact, quench-
ing the [4 + 2] cycloaddition with a slight excess of
NaOH/MeOH (15 equiv) and stirring at -78 °C for an
additional 30 min led to complete conversion of the
nitronate to oxime 22, entry 3. Quenching the reaction
with Et3N/MeOH (entry 4) at low temperature led to the
isolation of the nitroso acetals in 76% yield with none of
the oxime detected by NMR. Although the changes in
solvent and quench were significant with the benzoate
vinyl ether (Z)-16c, the effect on the acetate vinyl ether
(Z)-16a was even more pronounced. In CH2Cl2, with the
standard NaOH/MeOH quench (entry 5), no nitroso
acetals were isolated. By changing the solvent to toluene
with NaOH/MeOH quench, the nitroso acetals were
isolated in 23% yield as approximately 5-6/1 ratio of
diastereomers (entry 6). In toluene with the modified
Et3N/MeOH quench, the nitroso acetals were isolated in
55% yield (entry 7). Despite the advances in the yield of
acetate cycloaddition and seemingly simplified diaste-
reoselectivity, the benzoate vinyl ether cycloaddition was
more efficient overall and was selected for the synthesis
of casuarine. Under optimized conditions (Scheme 5),
nitroso acetal 18 was isolated as a 45/7/3/2/1/1 mixture
of nitroso acetals in 76% yield.25 The major nitroso acetal
was enriched to the extent of 41/0/0/2/1/1 by preparative
HPLC for an overall yield of 55%.
With an improved synthesis of the dieneophile in hand,
optimization of the [4 + 2] cycloaddition could be under-
taken. It was discovered that changing two features of
the reaction was crucial for increasing its yield, Table 1.
Ta ble 1. Op tim iza tion of th e Ta n d em [4 + 2]/[3 + 2]
Cycloa d d ition
entry
R
solvent
quench
18, %
22, %
1
2
3
4
5
6
7
16c, Bz
16c, Bz
16c, Bz
16c, Bz
16a , Ac
16a , Ac
16a , Ac
CH2Cl2
toluene
toluene
toluene
CH2Cl2
toluene
toluene
A
A
B
C
A
A
C
55b
63
0
29
18
81
0
76c
0
Configurational assignment of the major diastereomer
of nitroso acetal 18 was assisted by examining the
cycloaddition of nitronate 17 with dimethyl maleate,
Scheme 6. Stirring a mixture of nitronate 1726a (dr ) 4.8/
1) and dimethyl maleate in CHCl3 for 18 h afforded
nitroso acetal 23 in 87% yield as a 12.0/4.8/3.0/1/1
mixture of diastereomers. The structural assignment of
the major diastereomer of 23 was established by single-
crystal X-ray analysis. Surprisingly, the major diastere-
23
55d
a
Quench method A ) NaOH/MeOH (10 equiv), B ) NaOH/
MeOH (15 equiv), C ) Et3N/MeOH. (10 equiv) dr ) 8/2/1/1. c dr
) 45/7/3/2/1/1. dr ) 5-6/1.
b
d
The identification of a byproduct 22 by 1H NMR analysis
of the [4 + 2] reaction prompted the examination of the
quench. The standard method (entry 1) for quenching the
Sch em e 5

Synthesis of (+)-Casuarine
Sch em e 6
J . Org. Chem., Vol. 65, No. 10, 2000 2879
stants (the C(6) auxiliary has been replaced with a
methyl group for clarity). Proton NMR coupling analysis
indicated that the two nitroso acetals (18 and 23) existed
in very different solution conformations about the oxazine
1
ring. In the H NMR spectrum of nitroso acetal 23, both
the C(4) and C(5) protons appeared as doublets of
1
doublets. However, in the H NMR spectrum of nitroso
acetal 18, both the C(4) methine and the C(5) methine
resonances are singlets, indicative of a near 90° dihedral
angle. Thus, four of the five sterocenters in (+)-casuarine
were now set.
omer in this reaction was derived from â-face approach
to nitronate 17. Figure 1 shows the X-ray crystal struc-
The final stereocenter was installed by a selective
ketone reduction, Scheme 5. Treatment of keto nitroso
acetal 18 with L-Selectride led to a 10/1 mixture of
epimeric alcohols 24 in 87% yield.25 Additionally, the
three minor diastereomers present in the starting ketone
mixture were chromatographically removed. These minor
diastereomers were isolated as unreacted ketones, which
was supported by spectroscopic evidence (MS, 13C NMR).
1
Furthermore, H NMR analysis (vide supra) indicated
that two of the three diastereomers were derived from
the opposite [3 + 2] facial approach. The configurational
assignment of the product alcohol was based on previous
examples14g and examination of models. Ultimately, the
assignment was verified by conversion to the natural
product.
The resulting mixture of epimeric alcohols was acti-
vated in nearly quantitative yield with methanesulfonic
anhydride in pyridine to form 25, Scheme 5. Mesylate
25 then was reduced in the presence of Raney nickel and
hydrogen. It was discovered that the C(5) benzoate
suffered partial saponification under the reaction condi-
tions. Thus, to facilitate purification, K2CO3 was added
at the end of the reaction to affect complete removal of
both benzoate groups. The hydrogenolysis of 25 showed
a slight pressure and time dependence, Table 2. At 160
F igu r e 1. X-ray crystal structure of nitroso acetal 23, with
indicated dihedral angles and 1H NMR coupling constants.
ture of nitroso acetal 23 with crystallographically deter-
Ta ble 2. Hyd r ogen olysis Op tim iza tion of Nitr oso
Aceta l 25
1
mined indicated dihedral angles and measured H NMR
coupling constants.26b Figure 2 shows the ground-state
entry
PH , psi
T, °C
time, h
27, %
26, %
2
1
2
3
4
5
160
160
260
260
500
23
50
23
23
23
40
48
44
64
45
65
60
64
64
50
64
90
99
99
93
psi the yield of the pyrrolizidine 27 and the auxiliary 26
were the same (entry 1). When the hydrogenolysis was
conducted at a higher temperature (entry 2) the yield of
(24) Limat, D.; Schlosser, M. Tetrahedron 1995, 51, 5799.
(25) Diastereomeric ratio was determined by chiral stationary phase
supercritical fluid chromatography (SFC).
(26) (a) The synthesis of 23 is described in the Experimental Section
using a chiral nonracemic dieneophile, but the X-ray crystal structure
was obtained from diffraction of racemic material. (b) The crystal-
lographic coordinates of 23 have been deposited with the Cambridge
Crystallographic Data Centre; deposition no. CCDC 136434.
F igu r e 2. MM2-minimized conformation of nitroso acetal 18
with indicated dihedral angles and 1H NMR coupling con-
stants.
calculated (MM2) conformation of nitroso acetal 18 with
indicated dihedral angles and 1H NMR coupling con-

2880 J . Org. Chem., Vol. 65, No. 10, 2000
Denmark and Hurd
the pyrrolizidine 27 was not significantly altered, but the
recovery of the auxiliary 26 was increased to 90%.
Similarly, at 260 psi and at room temperature the yield
of the pyrrolizidine 27 was again unchanged and the
recovery of the auxiliary 26 was near quantitative (entry
3). Longer reaction times at 260 psi had no effect (entry
4). At 500 psi hydrogen the recovery of the auxiliary 26
was again high, but the yield of the pyrrolizidine 27 was
reduced (entry 5). Under optimized conditions, treatment
of the mesylate 25 with Raney nickel in MeOH under
260 psi of H2 for 44 h provided crystalline pyrrolizidine
27 in 64% yield along with a 99% recovery of the chiral
auxiliary 26, Scheme 5.
All physical and spectral properties of synthetic (+)-
casuarine (6) were obtained from an analytically pure
sample and are in agreement with the reported measure-
ments6 of the natural product. However, the optical
rotation of synthetic (+)-casuarine (6) was low (+10.8
versus +16.9 (H2O) reported for the natural material).
This discrepancy prompted us to vouchsafe the enantio-
meric purity of the synthetic material by independently
establishing the enantiomeric composition of a late stage
synthetic intermediate. The 2,6,7-trisbenzoate 29, ob-
tained in 85% yield from 27 (pyridine, Bz2O, 12 h), was
determined to have an enantiomeric excess of ca. 98%
by chiral stationary phase supercritical fluid chromatog-
raphy (SFC). We thus conclude that the reported optical
rotation for natural for (+)-casuarine is erroneously high.
The completion of the synthesis required only the
transformation of the C(1) silyl group to the final hy-
droxyl substituent and deprotection. Initial investigations
were aimed at avoiding the use of mercury salts for this
process due to complications in the isolation of the final
product previously encountered.14g,19 The dearylation can
be conducted under acidic, basic or electrophilic condi-
tions. Initially, a strong acid was sought to favor pro-
todesilylation over Peterson olefination,27 an obvious
possibility given the trans relationship of the C(2)-
hydroxyl and C(1) silane groups. Trifluoroacetic acid at
reflux surprisingly returned starting material. Dry HCl
in CHCl3 slowly produced the olefin 28 (not shown), along
with the insoluble HCl salt of the starting silane. Sulfonic
acids were next attempted. Methanesulfonic acid at room
temperature slowly produced the Peterson olefinated
product, while trifluoromethanesulfonic acid cleanly and
rapidly produced the olefin trifluoroacetate salt. Basic
oxidation conditions as described by Woerpel28 were
ineffective, instead producing an unidentified mixture of
products. Electrophilic desilylation with iodine monochlo-
ride cleanly dearylated silane 27, but also seemed to
iodinate the tertiary nitrogen and was abandoned. Since
acidic, basic or electrophilic conditions were not encour-
aging, the use of mercury salts was reinvestigated.
In the synthesis of 7-epiaustraline,14g,19 mercury salts
were effective for the dearylation process, but failure to
remove all the mercury by ion exchange chromatography
complicated the subsequent N-oxide reduction step.
Conditions which limit the production of the N-oxide were
sought. It was discovered that N-oxide production could
be controlled by conducting the oxidation at room-
temperature rather than the standard 50 °C protocol.
This process began with dearylation of the silyl group of
27 with mercuric trifluoroacetate in trifluoroacetic acid
and acetic acid for 1 h, followed by room-temperature
oxidation (16 h) with peracetic acid to provide (+)-
casuarine (6), Scheme 5.29 If the oxidation were conducted
at a higher temperature (50 °C), a significant amount of
N-oxide was observed. Under the room-temperature
oxidation conditions less than 5% N-oxide was detected
by 1H NMR analysis. After ion exchange chromatography
(BioRad AG 50W-X8), casuarine was isolated by crystal-
lization from EtOH. The N-oxide present in the mother
liquor was reduced (H2, MeOH, Pd/C, HOAc, 16 h) and,
following ion-exchange chromatography and crystalliza-
tion, provided additional amounts of casuarine to bring
the combined yield to 84%.
Discu ssion
[4 + 2] Cycloa d d ition . In addition to providing
insight into the optimization of the [4 + 2] cycloaddition,
the isolation and identification of oxime 22 also provided
information about the effectiveness and diastereoselec-
tivity of the first step of the tandem process, the [4 + 2]
cycloaddition. The [4 + 2] cycloaddition was judged to
be high yielding (81%) and occurred with modest facial
selectivity (4.8/1).30 The rearrangement of nitronate 17
to oxime 22 is envisioned to proceed by a base-catalyzed
deprotonation of HC(3) creating a nitrile oxide intermedi-
ate, which collapses to form the oxime product, Scheme
7. This fragmentation has precedent in nitronate chem-
istry.31 Interestingly, by this mechanism, the hemiacetal
is captured by the nitrile oxide before acetal breakdown
can occur.
Sch em e 7
There are two factors that influence the stereochemical
outcome of the [4 + 2] cycloaddition. The first factor
concerns the relative topicity of π components in the
cycloaddition and this sets the relative configuration
between the C(4) substituent and the C(6) substituent.
The endo/exo descriptor has been traditionally employed
to delineate this relative diastereoselection. An exo
cycloadduct possesses a C(4)/C(6) trans, while an endo
cycloadduct has a C(4)/C(6) cis relationship. As a direct
consequence of the vinyl ether geometry, the relative
configuration between C(4) and C(5) is also established.
The second factor involves the differentiation of one of
the diastereotopic π faces (Re or Si) of the vinyl ether by
the chiral auxiliary and is termed internal diastereose-
lection. This element determines the configuration at C(6)
relative to the resident stereocenter(s) of the chiral
(30) Diastereomeric ratio was determined by NMR. The two major
diastereomers are assumed to be facial diastereomers. It is uncertain
whether the third diastereomer is an oxime geometric isomer or an
acetal epimer.
(31) (a) Denmark, S. E.; Moon, Y.-C.; Cramer, C. J .; Dappen, M. S.;
Senanayake, C. B. W. Tetrahedron 1990, 46, 7373. (b) Colvin, E. W.;
Robertson, A. D.; Seebach, D.; Beck, A. K. J . Chem. Soc., Chem.
Commun. 1981, 952.
(27) (a) Peterson, D. J . J . Org. Chem. 1968, 33, 780. (b) Ager, D. J .
Synthesis 1984, 384. (c) Ager, D. J . Org. React. 1990, 38, 1.
(28) Smitrovich, J . H.; Woerpel, K. A. J . Org. Chem. 1996, 61, 6044.
(29) Kolb, H. C.; Ley, S. V.; Slawin, A. M. Z.; Williams, D. J . J . Chem.
Soc., Perkin Trans. 1 1992, 2735.

Synthesis of (+)-Casuarine
J . Org. Chem., Vol. 65, No. 10, 2000 2881
auxiliary. The descriptor for this stereoinduction22,32
combines the configuration at C(1′) of the auxiliary (R
or S) with the observed face of attack (Si or Re). Attack
of a nitroalkene to the Re face of the vinyl ether bearing
an R configuration at C(1′) would be termed like (1,3-lk).
Similarly, attack on the Si face of a vinyl ether bearing
an R configuration at C(1′) would be unlike (1,3-ul).
In nitronate 17, the C(4) benzoate is trans to both the
C(5) benzoate and the C(6) auxiliary, Figure 3. This
assignment is supported by the 1H NMR analysis of
nitronate 17 and the X-ray crystal structure of [3 + 2]
cycloadduct 23. This relative configuration can arise only
from an exo orientation of the vinyl ether to the nitroalk-
ene. Additionally, the configuration at C(6) is of the S
configuration, which arises from attack of the nitroalkene
on the Si face of the vinyl ether. The X-ray crystal
structure of 23 also confirms that the C(1′) position of
the auxiliary is of the R configuration, making this
internal stereoinduction of the (1,3-ul) variety. Nitronate
17 arises from an exo, (1,3-ul) cycloaddition of nitroalkene
13 and vinyl ether 16c.
F igu r e 4. Representations of vinyl ether (Z)-16c (a) s-cis (b)
s-trans conformation from MM2 calculations.
F igu r e 3. Exo approach of Si face of vinyl ether (Z)-16c to
nitroalkene 13.
An understanding of the forces responsible for the
stereodifferentiation of the π faces of the vinyl ether by
the chiral auxiliary may be understood by first consider-
ing the reactive conformation of vinyl ethers in general.
There are two limiting reactive conformations of a vinyl
ethers, s-cis and s-trans. Recent calculations and experi-
mental work suggest that although the s-cis conformation
is typically the lowest energy conformation in the ground
state, the s-trans conformation is the more reactive of
the two.33 However, in tandem nitroalkene cycloadditions,
reactions through both s-cis and s-trans conformations
have been invoked to explain the observed facial selectiv-
ity.22,32
Examination of molecular models of the ground state
provides some indication of facial shielding in each of the
limiting reactive conformations. Both the s-trans and the
s-cis ground state conformations (MM2) of vinyl ether (Z)-
16c are shown in Figure 4. The s-cis conformation
provides little facial discrimination for an approaching
heterodiene, as one of the phenyl rings partially blocks
the Si face and the cyclopentyl ring partially obstructs
the Re face. To explain the observed product by reaction
through this conformation, the heterodiene would have
to approach the vinyl ether on the same face as the
phenyl substituent, while the cyclopentyl ring is the
stereocontrolling element. On the other hand, in the
s-trans conformation, one of the phenyl rings provides
considerable shielding to the Re face, leaving the Si face
available. Approach of the Re face of the nitroalkene-
Lewis acid complex to the Si face vinyl ether in the
s-trans configuration would provide the experimentally
observed nitronate product.
The current analysis stands in sharp contrast to a
recent study of cycloadditions promoted by SnCl4 in which
it was proposed that phenylcyclohexanol modified vinyl
ethers reacted in an s-cis conformation.22,32 The switch
in internal stereoinduction in these cases was rational-
ized by considering a nonbonded steric interaction be-
tween the auxiliary of the s-trans configured vinyl ether
and SnCl4 which is relieved by reaction through an s-cis
vinyl ether conformation. Experimentally, reaction via
the s-cis conformation seems to be a general phenomenon
only with phenylcyclohexanol modified vinyl ethers, but
the larger trans-2-cumylcyclohexyl (TCC) vinyl ether does
not exhibit this trait.34
The disparity between the reactive conformations of
phenylcyclohexanol modified vinyl ethers and vinyl ether
(Z)-16c may lie in the size and shape of the auxiliary.
J ust as steric interactions between the Lewis acid and
the auxiliary can alter a reactive conformation, the same
could be stated about steric interactions of the auxiliary
with the â-substituents of the vinyl ether. Clearly, the
greater the size of the auxiliary, the greater the non-
bonded interactions will be with both the â-substituent
and the Lewis acid. The reactive conformation of the vinyl
ether then depends on the interaction which incurs a
greater penalty. An alternative explanation is that the
shape of the diphenylcyclopentyloxy auxiliary is simply
different than that of phenylcyclohexanol and no interac-
tion with the Lewis acid occurs. To illustrate this pos-
(32) Denmark, S. E.; Dixon, J . A. J . Org. Chem. 1998, 63, 6178.
(33) Liu, J .; Niwayama, S.; You, Y.; Houk, K. N. J . Org. Chem. 1998,
63, 1064.
(34) Dixon, J . A. Ph.D. Thesis, University of Illinois, Urbana, 1998.

2882 J . Org. Chem., Vol. 65, No. 10, 2000
Denmark and Hurd
sibility, a model of the nitrobenzoate 13-SnCl4 complex
was assembled using modified X-ray crystallographic
data,35 as in the previous analysis,32 and was docked with
the s-trans vinyl ether (Z)-16c, itself obtained from MM2
calculations, Figure 5. As before, the tin atom is pre-
sumed to be hexacoordinate (with four chlorines and the
nitroalkene as the ligands) in an octahedral geometry.36
The final coordination site is occupied by a methanol
molecule for simplicity, but in the reaction this likely
would be another vinyl ether, nitroalkene or the product
nitronate. The nitroalkene is docked at the distances
determined by transition state calculations for the Lewis
acid-catalyzed cycloaddition of nitroethylene, and methyl
vinyl ether.37 This representation suggests that there are
no significant nonbonded steric interactions between the
chiral auxiliary and the Lewis acid. Therefore, the vinyl
ether may not be forced into reaction through the s-cis
conformation as is seen in SnCl4 promoted reactions with
phenylcyclohexanol modified vinyl ethers.
nyldimethylsilyl group. This suggests that the â-position
of the dipolarophile is involved in a critical nonbonded
steric interaction with the nitronate.
There are two likely reactive conformations of the ni-
tronate 17 in the [3 + 2] cycloaddition: a twist-chair with
an axially disposed C(6) alkoxy group (A) and one with
an equatorially disposed C(6) alkoxy group (B), Figure
6. The top side of nitronate conformer A is expected to
F igu r e 6. Preferred approach of dipolarophiles to 17.
be the more reactive face, as it leads directly to a favor-
able chair conformation, and the approaching dipolaro-
phile should encounter little steric interference from the
two equatorially disposed benzoate groups at C(4) and
C(5). Since the top face of nitronate conformer B is
significantly shielded by the C(4) benzoate group, reac-
tion is expected to take place on the bottom face of this
conformer. Whereas dimethyl maleate adds preferentially
to the top face of nitronate A, the silyl dipolarophile 10
adds to the bottom face of either conformer A or con-
former B. Examination of models constructed from AM1
semiempirical calculations of conformer A of nitronate
17 indicate that the top face approach of silyl dipolaro-
phile is disfavored by a severe nonbonded steric interac-
tion between the equatorially oriented C(4) benzoate of
the nitronate and the phenyldimethylsilyl moiety of 10,
Figure 7. As a consequence, reaction occurs on the bottom
F igu r e 5. Approach of s-trans conformer of (Z)-16c to SnCl4‚
13 complex.
[3 + 2] Cycloa d d ition . The stereochemical analysis
of the [3 + 2] cycloaddition step in this synthesis was
complicated by the difficulty in establishing the identity
of the minor diastereomers in the product. The [4 + 2]
cycloaddition was deemed to occur in 81% yield with 4.8/1
facial selectivity, as determined by its conversion to the
oximino ether 22. The selectivity in the [3 + 2] cycload-
dition step is less clear. In the worst instance, if all of
the other diastereomers arose from the â-face approach
to the nitronate, the [3 + 2] selectivity would be 3.2/1.
However, since this diastereomeric ratio must include the
[4 + 2] cycloaddition diastereomers (4.8/1), the [3 + 2]
ratio is probably closer to 9/1. The sense and magnitude
of this preference is astonishing because the same
nitronate undergoes [3 + 2] cycloaddition with dimethyl
maleate to the extent of 3/1 favoring the opposite face.
Assuming that both dipolarophiles react with similar
timing and geometry, it is surprising that the face
selectivity switches by the simple substitution of the exo
oriented â-substituent from a carbomethoxy to a phe-
(35) Lewis, F. D.; Oxman, J . D.; Huffman, J . C. J . Am. Chem. Soc.
1984, 106, 466.
(36) (a) Denmark, S. E.; Almstead, N. G. J . Am. Chem. Soc. 1993,
115, 3133. (b) Shambayati, S.; Crowe, W. E.; Schreiber, S. L. Angew.
Chem., Int. Ed. Engl. 1990, 29, 256. (c) Reetz, M. T. In Selectivities in
Lewis Acid Promoted Reactions; Schiner, D., Ed.; Kluwer: Dordrecht,
1989; pp 107-125.
F igu r e 7. Disfavored [3 + 2] cycloadditions indicating non-
bonded steric interactions between the C(4) benzoate of
nitronate 17 and the PhMe2Si group of the dipolarophile 10.
(37) Domingo, L. R.; Arno´, M.; Andre´s, J . J . Org. Chem. 1999, 64,
5867.

Synthesis of (+)-Casuarine
J . Org. Chem., Vol. 65, No. 10, 2000 2883
face likely through conformer B, as this leads directly to
a chair conformation of the oxazine six-membered ring.
Keton e Red u ction . The stereochemical course of
hydride reduction of chiral ketones depends on the
reactive conformation of the ketone, the nonbonding
interactions between the approaching hydride reagent
and the shielding imparted by resident stereocenters of
the ketone. In the reduction of keto nitroso acetal 18, the
phenyldimethylsilyl substituent is believed to provide the
steric shielding. In studies on the related nitroso acetals
(30 and 31), the diastereomeric ratio was 14/1 when the
phenyldimethylsilyl substituent was present (30), Chart
2.19 In its absence (nitroso acetal 31) the reduction was
unselective (1/1).
Hyd r ogen olysis. The successful implementation of
the tandem nitroalkene cycloaddition method requires
the hydrogenolytic unmasking of nitroso acetals. Al-
though this step is crucial for the synthesis of pyrrolizi-
dine and indolizidine natural products, it remains a
transformation about which very little is known. The
hydrogenolysis of nitroso acetal 25 is no exception. Since
there were no other products isolated except the pyr-
rolizidine silane 27 and the auxiliary 26, very little
insight into the reaction mechanism is possible. Exami-
nation of the optimization study does allow for some
conjecture as to the timing of the hydrogenolysis events.
At 160 psi, the yield of the silane 27 and the recovery
of the auxiliary 26 are the same (64%), but at higher
pressures and higher temperatures only the recovery of
the auxiliary is effected (99%). One interpretation of these
results is that at 160 psi and above the first N-O bond
cleavage is rapid and the intermediate partitions between
internal alkylation (32) and the second N-O bond
cleavage (33), Scheme 8. At 160 psi, the second N-O bond
of oxazine 32 is unreactive and does not release the
auxiliary, such that the hemiacetal 33 is the only source
of both product silane 27 and auxiliary 26, explaining
the identical recoveries of these materials. At 260 psi (23
°C) psi and at 160 psi (50 °C), the partitioning between
32 and 33 remains the same, but in these cases 32 does
suffer the second N-O bond cleavage and releases the
remainder of the auxiliary 26, but no product. At 500 psi
of hydrogen the partitioning between 32 and 33 was
again unchanged, but the yield of the product silane was
lower, probably due to competitive reduction of the
intermediate aldehyde 34. To complete the transforma-
tion, aldehyde 34 forms a Schiff’s base with the tethered
amine, is saturated, and is alkylated affording pyrrolizi-
dine silane 27. The actual timing of the reductive
amination relative to the alkylation is uncertain, but is
supported by the isolation of mesylate containing prod-
ucts in other studies.14h
Ch a r t 2
Examination of reactive conformations of 18 where the
phenyldimethylsilyl substituent imparts carbonyl facial
shielding indicates that the Cornforth model38 best
explains the selective reduction of the keto nitroso acetal.
Figure 8 depicts a ground state MM2 calculated structure
in which the dihedral angle [O(24), C(18), C(6), O(4)] is
165° (the benzoate aryl rings and the chiral auxiliary
have been replaced by methyl groups, and thexyldim-
ethylsilyl methyl groups have been removed for clarity).
The phenyldimethylsilyl substituent effectively shields
the Si face of the ketone. Other models for the addition
of nucleophiles to ketones do not explain the observed
facial selectivity, but all are within 1 kcal/mol of the
global minimum. Only in the Cornforth conformation is
the Re face of the carbonyl exposed.
Sch em e 8
F igu r e 8. Reactive conformation of ketone 18 (MM2).
Con clu sion
The first synthesis of the (+)-casuarine ((+)-6) has been
accomplished in only eight steps from the chiral R-alkoxy
(38) Cornforth, J . W.; Cornforth, R. H.; Mathew, K. K. J . Chem. Soc.
1959, 112.

2884 J . Org. Chem., Vol. 65, No. 10, 2000
Denmark and Hurd
[(2R,3R,3a R,4S,5R,6S)-4,5-Bis(ben zoyloxy)-3-(d im eth -
ylp h en ylsilyl)h exa h yd r o-6-[[(1R)-2,2-d ip h en ylcyclop en -
tyl]oxy]isoxazolo[2,3-b][1,2]-2-oxazin yl]-2-[[dim eth yl(1,1,2-
tr im eth ylp r op yl)silyl]oxy]eth a n on e ((+)-18). Tin tetrachlo-
ride (290 µL, 2.50 mmol, 2.5 equiv) was added to a -78 °C
solution of nitroalkene 1314a (193.0 mg, 1.00 mmol) and
dieneophile (Z)-16c (421 mg, 1.10 mmol, 1.1 equiv) in 20.0 mL
of toluene. The resulting solution was stirred for 2 h at -78
°C and then quenched with 1 N Et3N in MeOH (10 mL, 10.0
mmol, 10.0 equiv). The mixture was diluted with CH2Cl2 (200
mL) and washed with H2O (3 × 50 mL). The aqueous washes
were back-extracted with CH2Cl2 (200 mL). The combined
organic extracts were dried (Na2SO4), filtered, and concen-
trated. A solution of dipolarophile 1014g (377.0 mg, 1.0 mmol,
1.0 equiv) in 5 mL CHCl3 was added, and the solution was
then stirred at room temperature for 16 h. The mixture was
concentrated and then was purified by silica gel chromatog-
raphy (hexane/EtOAc, 10/1) to afford 713.9 mg (76%) of keto
nitroso acetal (+)-18 as 45/7/3/2/1/1 ratio of diastereomers
(SFC). The mixture was further purified by preparative HPLC
[3 injections (340 mg, 370 mg, and mixed), 50 × 2.5 silica gel
column (Regis), 1.80% EtOAc in hexane, 25 mL/min] to afford
521.0 mg (55%) of ketone (+)-18 as a 40/0/0/2/1/1 ratio of
diastereomers. Data for (+)-18: 1H NMR (500 MHz, CDCl3) δ
8.04 (dd, J ) 8.4, 1.7, 2H), 7.97 (dd, J ) 8.3, 1.1, 2H), 7.49-
7.67 (m, 8H), 6.98-7.32 (m, 12H), 6.74 (t, J ) 7.4, 1H), 5.24
(s, 1H), 5.20 (d, J ) 10.4, 1H), 4.86 (d, J ) 4.0, 1H), 4.66 (d, J
) 1.3, 1H), 4.46 (s, 1H), 3.92-4.02 (ABq J AB ) 19.0, νA ) 1992.9
Hz, νB ) 1971.9 Hz, 2H), 3.58 (d, J ) 13.2, 1H), 2.93 (dd, J )
13.3, 10.4, 1H), 2.47-2.53 (m, 1H), 2.22-2.27 (m, 1H), 2.07-
2.11 (m, 1H), 1.84-1.89 (m, 1H), 1.56-1.61 (m, 2H), 1.23-
1.28 (m, 1H), 0.86 (d, J ) 7.0, 6H), 0.81 (s, 6H), 0.44 (s, 3H),
0.27 (s, 3H), 0.02 (s, 3H), 0.00 (s, 3H); 13C NMR (125 MHz,
CDCl3) δ 205.78, 164.83, 164.42, 146.33, 144.10, 137.67, 136.72,
133.98, 133.70, 133.35, 130.20, 130.00, 129.46, 129.03, 128.76,
128.47, 128.39, 128.17, 127.91, 127.57, 127.16, 125.88, 97.73,
87.27, 86.96, 72.89, 68.52, 68.40, 66.74, 59.92, 34.75, 34.08,
32.14, 29.00, 25.26, 20.23, 20.13, 18.49, -2.26, -3.37, -3.50,
-3.57; IR (neat) 1728 (s), 1451 (m), 1260 (s); MS (FAB) 941
aldehyde 20 in 20% overall yield. Five of the six stereo-
centers were created in the tandem [4 + 2]/[3 + 2]
cycloaddition, with the sixth provided by a selective,
substrate controlled ketone reduction. The Lewis acid
promoted [4 + 2] cycloaddition was successfully applied
with two highly oxygenated components and would not
have been viable without the improved synthesis of the
chiral acyloxy vinyl ether. The facial selectivity in the
[3 + 2] cycloaddition was found to be surprisingly
dependent upon the dipolarophile, apparently due to
nonbonded interactions with the â-silyl substituent.
Exp er im en ta l Section
Gen er a l Exp er im en ta l. See the Supporting Information
for details.
(Z)-2-[[(R)-(2,2-Dip h en ylcyclop en tyl)oxy]eth en eyloxy]-
tr im eth ylsila n e (21). A solution of aldehyde 2016 (2.80 g, 10.0
mmol), Et3N (5.6 mL, 40.2 mmol, 4.02, equiv), and TMSCl (2.60
mL, 20.5 mmol, 2.0 equiv) in CH3CN was heated to reflux for
2 h. The suspension was allowed to cool to room temperature
and the volatile components were removed in vacuo (3 h). The
resulting thick, white suspension was partitioned between H2O
(25 mL) and hexane (50 mL). The phases were separated and
the aqueous phase was extracted with hexane (50 mL). The
combined organic extracts were washed with H2O (25 mL),
dried (Na2SO4), filtered, and concentrated to afford crude silyl
enol ethers (7/1, Z/E (NMR)). This mixture was distilled to
afford 3.492 g (99%) of 21 (10/1, Z/ E) as a clear, colorless oil.
Data for 21: bp 220 °C (0.25 mmHg); 1H NMR (500 MHz,
CDCl3) δ 7.10-7.34 (m, 10H), 5.43 (d, J ) 3.4, 1H), 5.37 (d, J
) 3.4, 1H), 4.79 (t, J ) 3.8, 1H), 2.44-2.26 (m, 2H), 1.86-
1.98 (m, 3H), 1.62-1.68 (m, 1H), 0.07 (s, 9H); 13C NMR (125
MHz, CDCl3) δ 146.83, 145.00, 130.85, 128.74, 128.25, 127.66,
126.87, 125.88, 125.56, 122.65, 87.76, 59.00, 34.79, 29.34 19.93,
-0.52; IR (neat) 1665 (s); MS (70 ev) 52 (M+, 4); [R]26D ) -63.2
(CH2Cl2, c ) 1.00); TLC Rf ) 0.56 (hexane/EtOAc, 19/1). Anal.
Calcd for C22H28O2Si (384.47): C, 74.95; H, 8.01. Found: C,
74.74; H, 7.91.
(M+ + 1, 1); [R]26 ) +34.8 (CHCl3, c ) 1.00); TLC Rf ) 0.27
D
2-[(R)-(2,2-Dip h en ylcyclop en tyl)oxy]-(Z)-eth en ol Ben -
zoa te (Z-16c) a n d 2-[(R)-(2,2-Dip h en ylcyclop en tyl)oxy]-
(E)-eth en ol Ben zoa te (E-16c). Tetrabutylammonium fluoride
(170 µL, 1 M in THF, 0.17 mmol, 0.02 equiv) was added to a
0 °C solution of silyl enol ether 21 (3.00 g, 8.5 mmol) and
benzoyl fluoride (2.00 g, 16.1 mmol, 1.9 equiv) in 8.5 mL of
THF. The resulting solution was stirred for 2 h at 0 °C and
then was concentrated. The resulting clear oil was then
purified by silica gel chromatography (hexane/EtOAc, 19/1) and
distillation to afford 210.0 mg (6.4%) of enol benzoate (E)-16c
as a clear, thick oil, and 2.637 g (81%) of enol benzoate (Z)-
16c, which crystallized to a white solid after distillation. Data
(hexane/EtOAc, 6/1); SFC tR ) 12.15 min, (WELKO-1 column,
PCO ) 150 psi, 4.5 mL/min, 7.0% MeOH). Anal. Calcd for
2
C55H65NO9Si2 (1020.41): C, 70.26; H, 6.97; N, 1.49. Found: C,
70.57; H, 6.98; N, 1.43.
3,4-Bis(ben zoyloxy)-5-[[(1R)-2,2-d ip h en ylcyclop en tyl]-
oxy]tetr a h yd r ofu r a n -2-on e Oxim e ((+)-22). Tin tetrachlo-
ride (290 µL, 2.50 mmol, 2.5 equiv) was added to a -78 °C
solution of nitroalkene 1314a (193.0 mg, 1.00 mmol) and
dieneophile (Z)-16c (421 mg, 1.10 mmol, 1.1 equiv) in 20.0 mL
of toluene. The resulting solution was stirred for 2 h at -78
°C and then quenched with 1 N NaOH in MeOH (15 mL, 15
mmol, 15 equiv) and stirred at -78 °C for 30 min. The mixture
was partitioned between CH2Cl2 (200 mL) and pH )7.0
phosphate buffer (50 mL) and the organic layer then was
washed with H2O (200 mL). The aqueous washes were back-
extracted with CH2Cl2 (200 mL). The combined organic
extracts were dried (Na2SO4), filtered and concentrated. The
mixture then was purified by silica gel chromatography
(hexane/EtOAc, 2/1) to afford 469.3 mg (81%) of oxime 22 as
12.3/2.7/1.0 ratio of diastereomers (NMR). An analytical
sample (147.3 mg, 25%) of the major diastereomer as an Et2O
inclusion complex was obtained by crystallization from Et2O.
Data for 22: mp 120-121 °C (Et2O); 1H NMR (500 MHz,
CDCl3) δ 8.03 (d, J ) 8.5, 2H), 7.90 (d, J ) 8.5, 2H), 6.80-
7.60 (m, 16H), 6.33 (br s, 1H), 6.33 (d, J ) 7.60, 1H), 5.79 (d,
J ) 4.4, 1H), 5.39 (dd, J ) 7.8, 4.4, 1H), 4.85 (t, J ) 4.4, 1H),
3.85 (q, J ) 7.1, 1.6H), 2.53-2.60 (m, 1H), 2.20-2.31 (m, 2H),
1.88-1.97 (m, 2H), 1.54-1.63 (m, 1H), 1.21 (t, J ) 7.1, 2.4H);
13C NMR (125 MHz, CDCl3) δ 165.35, 165.21, 151.60, 146.48,
143.90, 133.70, 133.62, 130.20, 130.12, 128.69, 128.34, 128.25,
128.19, 127.61, 126.58, 125.91, 125.76, 101.41, 86.88, 74.17,
70.43, 65.91, 58.96, 34.86, 31.02, 19.59, 15.23; IR (KBr) 3462
(br s), 2970 (s), 1749 (s), 1284 (s); MS (FAB) 578 (M+ + 1, 16);
1
for (Z)-16c: bp 245 °C (0.3 mmHg); mp 86-88 °C; H NMR
(500 MHz, CDCl3) δ 7.97 (dd, J ) 8.2, 1.2, 2H), 7.58 (tt, J )
7.5, 1.2, 1H), 7.44 (t, J ) 7.9, 2H), 7.13-7.39 (m, 10H), 6.74
(d, J ) 3.6, 1H), 5.82 (d, J ) 3.6, 1H), 4.90 (t, J ) 4.7, 1H),
2.75 (ddd, J ) 12.8, 9.0, 7.7, 1H), 2.46 (ddd, J ) 12.8, 9.0, 4.0,
1H), 2.06-2.12 (m, 1H), 1.90-2.02 (m, 2H), 1.67-1.74 (m, 1H);
13C NMR (125 MHz, CDCl3) δ 163.18, 146.61, 144.55, 133.27,
132.31, 130.07, 129.32, 128.77, 128.39, 127.66, 126.86, 126.11,
125.81, 118.41, 88.06, 58.92, 35.08, 29.42, 19.81; IR (neat) 1728
(s); MS (70 eV) 384 (M+, 1); [R]26D ) -125.4 (CH2Cl2, c ) 1.00);
TLC Rf ) 0.28 (hexane/EtOAc, 9/1). Anal. Calcd for C26H24O3
(384.47): C, 81.22; H, 6.29. Found: C, 81.15; H, 6.30. Data
for (E)-16c: bp 250 °C (0.3 mmHg); 1H NMR (500 MHz, CDCl3)
δ 8.07 (dd, J ) 8.3, 1.2, 2H), 7.59 (tt, J ) 7.5, 1.2, 1H), 7.46 (t,
J ) 7.8, 2H), 7.13-7.33 (m, 10H), 7.30 (d, J ) 10.8, 1H), 6.77
(d, J ) 10.8, 1H), 4.89 (t, J ) 2.4, 1H), 2.54-2.63 (m, 2H),
1.91-2.04 (m, 3H), 1.55-1.72 (m, 1H); 13C NMR (125 MHz,
CDCl3) δ 164.12, 145.86, 144.98, 138.55, 133.71, 129.81, 129.32,
128.50, 128.44, 128.28, 127.93, 126.63, 126.14, 125.81, 123.67,
86.34, 59.51, 34.42, 28.52, 20.27; IR (neat) 1725 (s); MS (70
eV) 384 (M+, 1); [R]26 ) -13.2 (EtOH, c ) 1.00); TLC Rf )
D
0.37 (hexane/EtOAc, 9/1). Anal. Calcd for C26H24O3 (384.47)
C, 81.22; H, 6.29. Found: C, 80.95 H, 6.29.
[R]26 ) +43.2 (CHCl3, c ) 1.00); TLC Rf ) 0.21 (hexane/
D

Synthesis of (+)-Casuarine
J . Org. Chem., Vol. 65, No. 10, 2000 2885
EtOAc, 2/1). Anal. Calcd for C35H31NO7‚0.4Et2O (577.64): C,
72.39; H, 5.81; N, 2.39. Found: C, 72.35; H, 5.97; N, 2.51.
(2R,3a R,3S,4R,5R,6S)-3-Ca r bom eth oxy-4,5-bis(ben zo-
yloxy)-6-[((1R)-2,2-d ip h en ylcyclop en tyl)oxy]h exa h yd r o-
isoxa zole-[1,7-b][1,2]oxa zin e-2-ca r boxylic Acid Meth yl
Ester ((+)-23). Tin tetrachloride (290 µL, 2.50 mmol, 2.5 equiv)
was added to a -78 °C solution of nitroalkene 1314a (193.0 mg,
1.00 mmol) and dieneophile (Z)-16c (421 mg, 1.10 mmol, 1.1
equiv) in 20.0 mL of toluene. The resulting solution was stirred
for 2 h at -78 °C and then quenched with 1 N Et3N in MeOH
(10 mL, 10.0 mmol, 10.0 equiv). The mixture was diluted with
CH2Cl2 (200 mL) and then was washed with H2O (3 × 50 mL).
The aqueous washes were back-extracted with CH2Cl2 (200
mL). The combined organic extracts were dried (Na2SO4),
filtered and concentrated. A solution of dimethyl maleate
(125.0 µL, 1.0 mmol, 1.0 equiv) in 5 mL of CHCl3 was added,
and the solution then was stirred at room temperature for 18
h. The mixture was concentrated and then was purified by
silica gel chromatography (hexane/EtOAc; 3/1) to afford 628
mg (87%) of nitroso acetal 23 as 12.4/4.3/3.0/1.4/1.0 ratio of
diastereomers (SFC). An analytical sample of the major ketone
diastereomer was obtained by prep HPLC [50 × 2.5 cm silica
gel column (Regis), 15% EtOAc/hexane, 25 mL/min] to afford
57.0 mg (8%) of nitroso acetal 23 as a 45/0/1.0/2.9/0 ratio of
diasteromers. Data for 23: 1H NMR (500 MHz, CDCl3) δ 7.87
(d, J ) 8.4, 2H), 7.84 (d, J ) 7.1, 2H), 6.78-7.56 (m, 16H),
6.12 (dd, J ) 10.2, 6.0, 1H), 5.37 (d, J ) 10.4, 1H), 5.29 (dd, J
) 10.3, 3.9, 1H), 4.84 (t, J ) 5.3, 1H), 4.59 (dd, J ) 10.6, 6.0,
1H), 4.10 (t, J ) 10.5, 1H), 3.72 (s, 3H), 3.26 (s, 3H), 2.61-
2.63 (m, 1H), 2.24-2.31 (m, 3H), 1.90-1.92 (m, 1H), 1.48-
1.59 (m, 1H); 13C NMR (125 MHz, CDCl3) δ 169.68, 167.65,
165.67, 164.71, 146.99, 144.27, 133.59, 133.43, 130.09, 129.73,
128.88, 128.58, 128.44, 128.38, 128.33, 128.09, 127.54, 126.69,
125.71, 125.63, 97.77, 84.67, 73.81, 67.70, 65.80, 59.12, 52.82,
52.53, 47.75, 35.01, 30.60, 19.83; IR (KBr) 1737 (s); HRMS
+31.0 (CHCl3, c ) 1.00); TLC Rf ) 0.23 (hexane/EtOAc, 6/1);
SFC tR ) 1.862 min, (SiO2 column, PCO ) 150 psi, 4.5 mL/
2
min, 3.0% MeOH). Anal. Calcd for C55H67NO9Si2 (1020.41): C,
70.11; H, 7.17; N, 1.49. Found: C, 70.19; H, 7.23; N, 1.51.
(rS,2R,3R,3aR,4S,5R,6S)-4,5-Bis(ben zoyloxy)-3-(dim eth -
ylp h en ylsilyl)-R-[[[d im eth yl(1,1,2-tr im eth ylp r op yl)silyl]-
oxy]m et h yl]h exa n ea h yd r o-6-[[(1R)-2,2-d ip h en ylcyclo-
pen tyl]oxy]isoxazolo[2,3-b][1,2]oxazin e-2-m eth an ol Meth -
a n esu lfon a te Ester ((+)-25). Solid methanesulfonic anhy-
dride (1.05 g, 0.9 mmol, 6.0 equiv) was added to a solution of
alcohol 24 (942.0 mg, 1.00 mmol) in 8.0 mL of pyridine at room
temperature. The resulting brown suspension was stirred for
1 h and then was placed on a silica gel chromatography
column. The mixture was purified by silica gel chromatography
(hexane/EtOAc, 2/1) to afford 992.8 mg (97%) of mesylate 25
1
as a white foam. Data for 25: H NMR (500 MHz, CDCl3) δ
8.09 (dd, J ) 8.5, 1.2, 2H), 8.07 (dd, J ) 8.5, 1.5, 2H), 7.48-
7.70 (m, 8H), 7.18-7.32 (m, 10H), 7.08 (t, J ) 8.0, 2H), 6.85
(t, J ) 7.3, 1H), 5.46 (s, 1H), 5.03 (d, J ) 10.7, 1H), 4.93 (d, J
) 3.7, 1H), 4.74 (d, J ) 0.7, 1H), 4.59 (s, 1H), 4.48 (t, J ) 6.6,
1H), 3.92 (dd, J ) 10.6, 6.0, 1H), 3.71 (dd, J ) 10.5, 7.1, 1H),
3.69 (d, J ) 13.4, 1H), 3.01 (dd, J ) 10.7, 9.8, 1H), 2.97 (s,
3H), 2.53-2.59 (m, 1H), 2.26-2.32 (m, 1H), 2.13-2.17 (m, 1H),
1.91-1.97 (m, 1H), 1.55-1.67 (m, 2H), 1.29-1.45 (m, 1H), 0.87
(d, J ) 6.8, 6H), 0.80 (s, 6H), 0.64 (s, 3H), 0.31 (s, 3H), 0.05 (s,
3H), -0.01 (s, 3H); 13C NMR (125 MHz, CDCl3) δ 165.11,
164.82, 146.40, 144.21, 136.53, 133.73, 133.69, 133.31, 130.19,
130.07, 129.71, 129.02, 128.80, 128.44, 128.41, 128.25, 128.20,
128.14 127.61, 127.19, 126.00, 125.88, 97.65, 86.88, 84.13,
81.85, 72.05 68.76, 67.15, 61.54, 59.92, 39.67 34.79, 33.95,
32.14, 27.56, 25.13, 20.22, 20.19, 20.11, 18.49, -0.52, -3.56,
-3.60, -3.95; IR (KBr) 1735 (s), 1737 (s); HRMS (FAB) calcd
1020.420816, found 1020.420400; [R]26 ) +42.8 (CHCl3, c )
D
1.00); TLC Rf ) 0.43 (hexane/EtOAc, 4/1). Anal. Calcd for
C
56H69NO11Si2S (1020.41): C, 65.92; H, 6.82; N, 1.37; S, 3.14.
for
C41H40NO11 (721.79) (FAB) calcd 722.2601123, found
Found: C, 65.35; H, 6.73; N, 1.12.
722.260000; [R]26 ) +37.5 (CHCl3, c ) 0.89); TLC Rf ) 0.18
D
(1S,2S,5R,6R,7R,7a R)-7-(Dim eth ylp h en ylsilyl)-5-[[[d i-
m eth yl(1,1,2-tr im eth ylpr opyl)silyl]oxy]m eth yl]h exah ydr o-
1H-p yr r olizin e-1,2,6-tr iol ((-)-27). To a solution of mesylate
25 (918.0 mg, 0.9 mmol) in 120 mL of MeOH in a glassed lined
steel autoclave was added A5000 Raney nickel39 (washed 3 ×
100 mL MeOH). The autoclave was sealed, pressurized to 260
psi with H2, and stirred at room temperature for 42 h. H2 was
carefully released from steel autoclave and the reaction
mixture then was filtered through a 4 cm pad of Celite. The
filter cake was washed with MeOH (500 mL), and the filtrate
was concentrated to approximately 25 mL. Solid K2CO3 (5 g)
was added, and the resulting slurry was stirred at room
temperature for 4 h. The slurry was then dry loaded onto
activity III basic alumina (2 g) the mixture was purified by
activity III basic alumina chromatography (hexane/EtOAc,
8/1,2/1; CH2Cl2/MeOH; 0/1,50/1, 25/1, 10/1). to afford alcohol
(+)-26 and pyrrolizidine (-)-27. The alcohol 26 was distilled
to provide 212 mg (99%) of alcohol 26, and the pyrrolizidine
27 was recrystallized from hexane (2 crops) to afford 269.7 mg
(64%) of pyrrolizine 27 as a white crystalline solid. Data for
27: 1H NMR (500 MHz, CDCl3) δ 7.60-7.63 (m, 2H), 7.30-
7.48 (m, 3H), 4.17 (dt, J ) 6.3, 5.6, 1H), 4.01 (ddd, J ) 11.0,
7.8, 1.7, 1H), 3.81 (dd, J ) 9.5, 5.1, 1H), 3.56 (t, J ) 5.1, 1H),
3.52 (t, J ) 9.2, 1H), 3.26 (dd, J ) 10.6, 5.6, 1H), 3.15 (dd, J
) 11.0, 4.9, 1H), 2.83 (ddd, J ) 9.2, 7.8, 5.1, 1H), 2.76 (d, J )
1.7, 1H), 2.74 (dd, J ) 10.6, 6.6, 1H), 2.01 (br s, 1H), 2.66 (sept,
J ) 6.8, 1H), 1.64 (t, J ) 11.0, 1H), 0.921 (s, 3H), 0.918 (s,
3H), 0.889 (s, 3H), 0.886 (s, 3H), 0.480 (s, 3H), 0.473 (s, 3H),
0.155 (s, 3H), 0.147 (s, 3H); 13C NMR (125 MHz, CDCl3) δ
137.78, 133.67, 129.53, 128.18, 83.27, 80.44, 79.92, 73.97,
70.33, 67.14, 58.44, 38.94, 33.94, 25.01, 20.22, 20.08, 18.38,
18.30, -3.26, -3.69, -3.76, -6.23; IR (KBr) 3546 (br s), 3444
(hexane/EtOAc, 2/1).
(rS,2R,3R,3aR,4S,5R,6S)-4,5-Bis(ben zoyloxy)-3-(dim eth -
ylp h en ylsilyl)-R-[[[d im eth yl(1,1,2-tr im eth ylp r op yl)silyl]-
oxy]m eth yl]h exah ydr o-6-[[(1R)-2,2-diph en ylcyclopen tyl]-
oxy]-isoxa zolo[2,3-b][1,2]oxa zin e-2-m et h a n ol ((+)-24).
L-Selectride (4.00 mL, 0.76 M in THF, 3.04 mmol, 2.1 equiv)
was added to a -78 °C solution of ketone 18 (1.350 g, 1.44
mmol) in 30.0 mL of THF. The resulting solution was stirred
for 2 h at -78 °C and then was quenched slowly by careful
addition of glycerol/pH ) 7 buffer (15 mL). The mixture was
extracted with Et2O (200 mL) and washed with H2O (75 mL)
and brine (75 mL). The aqueous washes were back-extracted
with Et2O (200 mL). The combined organic extracts were dried
(Na2SO4), filtered, and concentrated on to silica gel (2 g). The
mixture then was purified by silica gel chromatography
(hexane/EtOAc, 6/1) to afford 1.169 g (86%) of alcohol 24 as
10/1 ratio of alcohol epimers (SFC) and 89.5 mg of unreacted
ketones. Residual ketone 18 present in the isolated unreacted
ketones was further reduced with L-Selectride (250 µL, 0.76
M in THF, 0.190 mmol) in 5 mL THF at -78 °C for 2 h.
Following quench, workup and silica gel purification afforded
70.9 mg of unreacted ketones (5.2%) as a 2/1/1 ratio of
diastereomers. Data for 24: 1H NMR (500 MHz, CDCl3) δ 8.10
(dd, J ) 8.3, 1.2, 2H), 8.02 (dd, J ) 8.3, 1.2, 2H), 7.46-7.70
(m, 8H), 7.14-7.28 (m, 10H), 7.06 (t, J ) 7.8, 2H), 6.79 (t, J )
7.3, 1H), 5.46 (d, J ) 1.0, 1H), 4.93 (d, J ) 3.9, 1H), 4.80 (d, J
) 10.3, 1H), 4.72 (d, J ) 1.2, 1H), 4.51 (br s, 1H), 3.72 (d, J )
13.4, 1H), 3.40-3.56 (m, 3H), 2.93 (dd, J ) 13.4, 10.0, 1H),
2.53-2.57 (m, 1H), 2.23-2.31 (m, 1H), 2.11-2.15 (m, 1H), 1.94
(d, J ) 5.4, 1H), 1.91-1.93 (m, 1H), 1.55-1.63 (m, 2H), 1.28-
1.32 (m, 1H), 0.87 (d, J ) 6.8, 6H), 0.81 (s, 6H), 0.55 (s, 3H),
0.31 (s, 3H), 0.04 (s, 3H), 0.02 (s, 3H); 13C NMR (125 MHz,
CDCl3) δ 165.99, 164.60, 146.45, 144.20, 137.67, 133.59, 133.23,
130.23, 129.21, 128.80, 128.31, 128.15, 127.98, 127.58, 127.21,
125.96, 125.87, 97.60 86.76, 86.07, 72.67, 71.82, 68.52, 67.07,
64.44, 59.96, 34.76, 34.14, 32.23, 27.49, 25.05, 20.31, 20.11,
18.51, 18.48, -1.05, -3.54, -3.56, -3.78; IR (KBr) 3567 (br,
(br s); MS (FAB) 466 (M+ + 1, 100); [R]26 ) -38.2 (CHCl3, c
D
) 1.00). Anal. Calcd for C24H43NO4Si2 (465.786): C, 61.89; H,
9.31; N, 3.01. Found: C, 61.80; H, 9.34; N, 2.70.
(39) A5000 Raney Nickel was purchased from Activated Metals and
Chemicals, Inc.
w), 3458 (br, w), 1729 (s); MS (FAB) 943 (M+ + 1, 11); [R]26
)
D

2886 J . Org. Chem., Vol. 65, No. 10, 2000
Denmark and Hurd
(1S,2S,5R,6R,7R,7a R)-7-(d im et h ylp h en ylsilyl)-5-[[[d i-
m eth yl(1,1,2-tr im eth ylpr opyl)silyl]oxy]m eth yl]h exah ydr o-
1H-p yr r olizin e-1,2,6-tr iol Tr iben zoa te (28). Benzoyl chlo-
ride (90 µL, 0.74 mmol, 20 equiv) was added to a solution of
amino triol (-)-27 (1.7 mg, 0.0365 mmol) in pyridine (0.5 mL).
The solution was stirred at room temperature for 20 h and
then was concentrated. The residue was purified by silica gel
chromatography (toluene/Et2O, 49/1) to afford 2.5 mg of
trisbenzoate 28. Data for 28: 1H NMR (500 MHz, CDCl3) δ
8.11 (dd, J ) 8.3, 1.2, 2H), 7.98 (dd, J ) 8.3, 1.2, 2H), 7.88
(dd, J ) 8.3, 1.2, 2H), 7.60-7.64 (m, 4H), 7.43-7.59 (m, 6H),
7.39 (t, J ) 7.8, 2H), 7.05-7.09 (m, 2H), 5.49 (t, J ) 1.7, 1H),
5.46 (dd, J ) 9.3, 7.1, 1H), 5.34 (s 1H, H(7)), 3.71 (dd, J )
10.0, 2.2, 1H), 3.65 (dd, J ) 10.3, 4.4, 1H), 3.55 (dd, J ) 10.3,
7.3, 1H), 3.50 (m, 2H), 3.38 (dt, J ) 7.1, 4.4, 1H), 2.21 (t, J )
9.6, 1H), 1.48 (sept, J ) 6.8, 1H), 0.74 (d, J ) 6.8, 3H), 0.73
(d, J ) 6.8, 3H), 0.71 (s, 3H), 0.70 (s, 3H), 0.40 (s, 3H), 0.39 (s,
3H), 0.01 (s, 3H, -0.06 (s, 3H); MS (FAB) 779 (M++1, 63); SFC
BioRad AG 50W-X8 resin (10 g), then was washed with H2O
(40 mL) to remove acetic acid, and then with 2 M NH4OH (80
mL) to elute casuarine. Evaporation of the NH4OH eluant
afforded a pale-yellow powder which was recrystallized (2
crops) from EtOH to provide 40.3 mg of casuarine (27%). The
combined crops were treated with pentane and concentrated
to azeotropically remove the remaining EtOH for a combined
yield of 124.7 mg (84%) of casuarine as a white powder. Data
for (+)-6: mp 180-181 °C (EtOH); 1H NMR (500 MHz, D2O) δ
4.21-4.24 (m, 2H), 4.19 (t, J ) 8.1, 1H), 3.81 (t, J ) 8.8, 1H),
3.78 (dd, J ) 11.7, 3.7, 1H), 3.63 (dd, J ) 11.7, 6.6, 1H), 3.29
(dd, J ) 12.1, 4.3, 1H), 3.10 (dd, J ) 8.3, 2.9, 1H), 3.07 (ddd,
J ) 8.8, 6.6, 3.7, 1H), 2.94 (dd, J ) 12.1, 3.6, 1H); 13C NMR
(125 MHz, D2O) δ 79.55, 78.53, 78.29, 77.37, 73.13, 70.87,
62.99, 58.90; IR (KBr), 3362 (br, s); MS (FAB) 206 (M++1, 80);
[R]27 ) +10.8 (H2O, c ) 1.020). Anal. Calcd for C8H15NO5
D
(205.212): C, 46.82; H, 7.37; N, 6.83. Found: C, 46.59; H, 7.28;
N, 6.49.
tR ) 4.301 min, tR ) 4.908 min; (Chiralcel OD column PCO
)
2
150 psi, 5.5 mL/min, 5.0% MeOH).
Ack n ow led gm en t. We gratefully acknowledge the
gift of natural (+)-casuarine from R. J . Nash (Aberyst-
wyth). Financial support was provided by the National
Institutes of Health (GM 30938). A.R.H. gratefully
acknowledges the University of Illinois for a Graduate
Fellowship, Eastman Chemical Company for an East-
man Summer Fellowship and Zeneca Pharmaceuticals
for the Zeneca Graduate Fellowship in Synthetic Or-
ganic Chemistry. We thank Dr. Scott R. Wilson for
collecting the X-ray data for 23 and Mark Seierstad and
J eromy Cottell for solving the structure. We also thank
Mr. J . Vessels for obtaining the 13C NMR spectrum of
(+)-6.
(1R,2R,3R,6S,7R,7a R)-Hexa h yd r o-3-(h yd r oxym eth yl)-
1H-p yr r olizin e-1,2,6,7-tetr a ol (Ca su a r in e) ((+)-6). Mercu-
ric trifluoroacetate (462.7 mg, 1.08 mmol, 1.5 equiv) was added
to a solution of (-)-27 (337 mg, 0.723 mmol, 1 equiv) in CF3-
CO2H (4 mL), CHCl3 (4 mL), and HOAc (2 mL) and was stirred
at room temperature for 60 min. The reaction mixture was
then cooled to 0 °C and peroxyacetic acid (37% in HOAc, 12
mL) was added dropwise, after which the cooling bath was
removed and the solution was stirred at room temperature for
18 h. The reaction mixture was cooled to 0 °C and slowly
quenched with Me2S, maintaining the internal temperature
e 15 °C, until a negative starch/I2 test was achieved. The
contents of the reaction were transferred to a column of BioRad
AG 50W-X8 resin (10 g) and then was washed with H2O (40
mL) to remove nonamine containing products and then with
2 M NH4OH (80 mL) to elute casuarine. Evaporation of the
NH4OH eluant afforded a pale-yellow powder that was recrys-
tallized from EtOH to provide 84.4 mg of casuarine (57%). The
mother liquor was concentrated and redissolved in MeOH (5
mL). 10% Palladium on carbon and 10 drops of HOAc were
added and the mixture was stirred under 1 atm of H2 for 16
h. The catalyst was removed by filtration through Celite, then
was rinsed with MeOH (100 mL) and concentrated. The
resulting, pale-yellow oil was transferred to a column of
Su p p or tin g In for m a tion Ava ila ble: Full 1H NMR and
13C NMR with assignments, complete listing of IR peaks, and
MS fragments of compounds 6, (E)-16c, (Z)-16c, 18, 21-25,
and 27 along with comparison 1H NMR spectra of synthetic
(+)-casuarine and 1H NMR of natural (+)-casuarine. This
material is available free of charge via the Internet at
http://pubs.acs.org.
J O991680V