Resolution and configurational assignment of 3,4,5,6-tetrahydro-2-methyl-2,6-methano-2H-1-benzoxocine derivatives

Article abstract of DOI:10.1039/b200751g

Chemical resolutions of the rigid, tricyclic 3,4,5,6-tetrahydro-2-methyl-2,6-methano-2H-1-benzoxocin-4-one (6) of biological and chiroptical interest have been performed. The configurational assignment was made by X-ray analysis and chemical correlation of its ketal (+)-9a synthesized with (2R,3R)-butane-2,3-diol. Kinetic resolution of the hydroxy derivative rac-10 by means of lipase from Pseudomonas cepacia proved the most effective procedure for resolution. The stereochemistry of the faster reacting enantiomer (+)-10 was deduced by chemical correlation with the ketal (+)-9. Model compounds (+)-1 and (-)-1 for chiroptical study of the chromane chromophore were also prepared.

Full text of DOI:10.1039/b200751g

View Article Online / Journal Homepage / Table of Contents for this issue
Resolution and configurational assignment of 3,4,5,6-tetrahydro-2-
methyl-2,6-methano-2H-1-benzoxocine derivatives
Tibor Kurtán,a Eszter Baitz-Gács,b Zsuzsa Májer,c Attila Bényeid and Sándor Antus*a
1
a Department of Organic Chemistry, University of Debrecen, P.O.B. 20, H-4010 Debrecen,
Hungary
b Institute of Chemistry, C. R. C., Hungarian Academy of Sciences, P.O.B. 32,
H-1515 Budapest, Hungary
c Institute of Organic Chemistry, L. Eötvös University, H-1518 Budapest 112, P.O.B. 32,
Hungary
d Department of Chemistry, Laboratory for X-Ray Diffraction, University of Debrecen,
P.O.B. 7, H-4010 Debrecen, Hungary
Received (in Cambridge, UK) 21st January 2002, Accepted 13th February 2002
First published as an Advance Article on the web 8th March 2002
Chemical resolutions of the rigid, tricyclic 3,4,5,6-tetrahydro-2-methyl-2,6-methano-2H-1-benzoxocin-4-one (6) of
biological and chiroptical interest have been performed. The configurational assignment was made by X-ray analysis
and chemical correlation of its ketal (ϩ)-9a synthesized with (2R,3R)-butane-2,3-diol. Kinetic resolution of the
hydroxy derivative rac-10 by means of lipase from Pseudomonas cepacia proved the most effective procedure for
resolution. The stereochemistry of the faster reacting enantiomer (ϩ)-10 was deduced by chemical correlation with
the ketal (ϩ)-9. Model compounds (ϩ)-1 and (Ϫ)-1 for chiroptical study of the chromane chromophore were also
prepared.
Introduction
3,4,5,6-Tetrahydro-2-methyl-2H-1-benzoxocine bridged by a
methylene unit across the 2,6 positions (1) is a rigid tricyclic
compound which contains a chirally perturbed chromane
chromophore with axial substituents on the heterocyclic ring.
This residue is a building block of gambogellic acid 2,1 which
has been found to be cytotoxic against HeLa cells, and the
racemic C-4 substituted derivatives 3–5 have shown central
nervous system activity.2 Moreover, it is noteworthy that the
oxygen analogs, which have an oxygen bridge across the 2,6
positions, have shown remarkable monoamine oxidase inhibi-
tor3 and antibacterial activity.4 Due to their rigid structure, this
type of optically active chromane derivative can be also used as
model compounds to study the chiroptical properties of the
chromane chromophore containing an axial carbon–carbon
bond at the benzylic position. We have found recently that P/M
helicity of the heterocyclic ring leads to a negative/positive CD
within the 1Lb band5,6 and this helicity rule is valid in its inverse
form if an axial carbon–oxygen bond is present at the benzylic
position.7,8 In order to study the scope and limitation of
our rule and elaborate a simple route for the synthesis of the
enantiomerically pure chromane derivatives 3–5, we set our
1
instead of two. The H NMR spectrum of the crude product
clearly indicated that the syn Z- and anti E-isomers of
both diastereomers [(Z)-(ϩ)-8/(E)-(ϩ)-8, (Z)-(Ϫ)-8/(E)-(Ϫ)-8]
were present in this mixture (Scheme 1). It was shown previ-
ously that similarly to oximes, the barrier of the syn–anti inver-
sion in hydrazones can be high enough to allow the isolation of
a pure non-inverting isomer14,15 whose inversion can be also
achieved by UV light irradiation.16 Although these isomers had
quite different retentions on TLC, they could not be separated
either by column chromatography or by preparative TLC
due to their rapid conversion on silica gel. Crystallization of
the crude product from methanol resulted in a mixture of
(Z)-(ϩ)-8 and (E)-(Ϫ)-8 with (Z)-(ϩ)-8 as the main com-
ponent whose repeated crystallization afforded the pure
dextrorotatory syn-isomer [(Z)-(ϩ)-8] while the levorotatory
sights on the enantioselective synthesis of 1 via 6.
Results and discussion
The racemic 2,6-methanobenzoxocinone derivative 6 was pre-
pared from salicylideneacetone and ethyl acetoacetate by using
the procedure of Kuhn and Weiser.9 Its chemical resolution was
carried out with (ϩ)-(1-phenylethyloxamoyl)hydrazine 7, a
carbonyl derivatizing agent10–13 (Scheme 1). In contrast to pre-
vious findings,10–13 where the resolution of chiral ketone deriv-
atives with 7 resulted in the two corresponding diastereomers,
which could be separated by fractional crystallization, the con-
densation of rac-6 with this reagent, monitored by TLC,
showed the formation of four products of comparable ratio
888
J. Chem. Soc., Perkin Trans. 1, 2002, 888–894
This journal is © The Royal Society of Chemistry 2002
DOI: 10.1039/b200751g

View Article Online
Scheme 1 Resolution of 3,4,5,6-tetrahydro-2-methyl-2,6-methano-2H-1-benzoxocin-4-one (rac-6) with (ϩ)-(1-phenylethyloxamoyl)hydrazine 7.
anti-isomer [(E)-(Ϫ)-8] was obtained from the mother liquor of
the first crystallization. The syn or anti arrangement of these
compounds was deduced from NOE experiments; the signifi-
cant effect between the hydrazone proton and H-5α in (E)-(Ϫ)-
8 unambiguously proved its anti arrangement about the C᎐N
᎐
double bond. It is interesting to note that when MeOD was
added to the CDCl3 solution of (E)-(Ϫ)-8, a new set of signals
appeared, which was assigned to the syn-isomer (Z)-(Ϫ)-8 on
the basis of its NOE between the hydrazone proton and H-3α.
After three weeks the ratio of the syn and anti isomers was
approximately 2 to 3 in the solution. The syn–anti isomerism of
(E)-(Ϫ)-8 was also followed by CD spectroscopy which showed
a characteristic change in the CD transitions of (E)-(Ϫ)-8 in
acetonitrile when acid was added to the solution, indicating the
formation of the Z-isomer, as shown in Fig. 1(a). The CD
spectra of the isolated isomers (Z)-(ϩ)-8 and (E)-(Ϫ)-8 were
found to be almost mirror image. [Fig. 1(b)] Since their chiro-
ptical properties are dominated by the chirality of the tricyclic
skeleton, they must contain heterocyclic rings with opposite
configuration. Due to the lack of similar examples in the liter-
ature, the signs of the CD transitions could not be used to
assign safely their absolute configurations and our efforts to
obtain a suitable single crystal for X-ray analysis were also
unsuccessful.
Acidic hydrolysis of (Z)-(ϩ)-8 and (E)-(Ϫ)-8 provided the
optically active ketones (ϩ)-6 and (Ϫ)-6, respectively, whose
CD spectra proved that their precursors (Z)-(ϩ)-8 and (E)-(Ϫ)-
8 indeed had enantiomeric annulation points. It might be
assumed that their absolute configuration could be deduced
from the sign of the Cotton effect belonging to the n π* tran-
sition of the carbonyl chromophore on the basis of the well-
known octant rule17 and by comparison with 3-substituted
cyclohexanone derivatives.18,19 However, the following detailed
analysis of the contribution from the different parts of the
molecule to the n π* Cotton effect of the 3-methylcyclo-
hexanone moiety of (2S,6R)-6 has shown that our assumption
was too optimistic. The cyclohexanone residue of (2S,6R)-6
adopts the thermodynamically more stable chair conformation
and the C-2 equatorial methyl group is situated in the rear
upper left or lower right octant (Fig. 2), which has a positive
contribution to the n π* Cotton effects according to the
Fig. 1 (a) syn–anti isomerism of (E)-(Ϫ)-8 followed by CD spectro-
scopy in MeCN: A, CD spectrum of (E)-(Ϫ)-8; B, CD spectrum
measured immediately after the addition of acid to (E)-(Ϫ)-8, i.e a
mixture of (E)-(Ϫ)-8 and (Z)-(Ϫ)-8; C, CD spectrum measured 24 h
after the addition of acid to (E)-(Ϫ)-8. (b) CD spectra of (E)-(Ϫ)-8 and
(Z)-(ϩ)-8 in MeCN indicating the enantiomeric annulation points of
the tricyclic skeleton.
octant rule. However, β-axial substituents at C-2 and C-6 are
also present and they show antioctant behavior with weak con-
tributions according to the literature.18,20 Since the axial oxygen
and the phenyl residue at C-6 are located in sectors of opposite
sign, they can compensate each other to some extent. Further
difficulties arise from the fact the condensed phenyl group
points in the direction of the carbonyl group and hence toward
the front sectors, which means a part of the phenyl group is
J. Chem. Soc., Perkin Trans. 1, 2002, 888–894
889

View Article Online
Scheme 2 Resolution of rac-6 with (2R,3R)-butane-2,3-diol and transformation of the isolated diastereomer (ϩ)-9a.
known absolute configuration of the ketal moiety. The crystal
data also clearly showed that the cyclohexane moiety of the
molecule adopted a chair conformation while the O-hetero-
cyclic ring was fixed in a rigid envelope conformation, i.e. C-6,
C-7a, C-10a, O-1 and C-2 were located approximately in the
same plane, and C-11 was out of the plane. The O-heterocyclic
ring has P-helicity, which is defined by the positive C(10a)–
O(1)–C(2)–C(11) torsional angle (ϩ31.3Њ). Acidic hydrolysis of
(2R,6S)-(ϩ)-9a was carried out with dilute hydrochloric acid in
dioxane in order to correlate the resultant ketone with the pre-
viously synthesized optically active ketones (ϩ)-6 and (Ϫ)-6.
The 2R,6S ketone obtained was dextrorotatory which allowed
the configurational assignment of the oxamoylhydrazone pre-
cursors (ϩ)-(Z)-8 and (E)-(Ϫ)-8, as shown earlier in Scheme 1.
Since the resolution of the racemic ketone 6 by both (ϩ)-(1-
phenylethyl)oxamoylhydrazine (7) and (2R,3R)-(Ϫ)-butane-
2,3-diol required crystallization of the diastereomeric mixture,
which occurred in relatively low yields, the racemic ketone 6 was
reduced to the racemic alcohol 10 which was also expected to be
suitable for resolution (Scheme 3). According to our expect-
ation, the reduction of rac-6 with lithium aluminium hydride
proceeded diastereoselectively because the addition of the
hydride to the carbonyl carbon was sterically hindered from
the direction of the chromane residue of the molecule. Thus,
the addition of the hydride was favored from the opposite face
and afforded almost exclusively the axial alcohol rac-10. The
axial orientation of the hydroxy group in rac-10 was verified by
the coupling constants of the equatorial 4-H with the vicinal
protons (J 5.2, 4.0, 2.2 and 2.6 Hz), but the axial hydroxy group
could not be derivatized with chiral acids such as (S)-(ϩ)-O-
acetylmandelic acid and (Ϫ)-camphanic acid even in the
presence of N,NЈ-dicyclohexylcarbodiimide (DDC), which was
most likely due to its hindered position.
The Pseudomonas lipases have been found to be very effective
enzymes in the kinetic transesterification of numerous racemic
secondary alcohols with remarkable enantiopreference and
tolerance for a great variety of substrates, including hindered
ones.22 This prompted us to use the lipase from Pseudomonas
cepacia (PCL) for the kinetic resolution of rac-10 which was
succesfully performed in ether with vinyl acetate as an irrevers-
ible acyl donor (Scheme 3). The kinetic resolution was found to
be sensitive to both the carrier of PCL and the type of sol-
vent used. It proceeded smoothly with PCL immobilized on
Toyonite in ether and it was somewhat slower in dioxane but
there was no detectable conversion in acetonitrile and PCL on
Diatomite did not catalyze the esterification in ether at all. It
took 16 days to reach a conversion of 52% for the reaction of
rac-10 with PCL on Toyonite in ether which indicated the
hindered position of the axial hydroxy group. Besides the coup-
ling constant, the unusual upfield chemical shift of the acetyl
methyl protons (δH = 1.28, CDCl3) in the acetate (ϩ)-11 also
confirmed the axial orientation of the acetoxy group because it
was due to the magnetic anisotropy effect on the acetoxy group
situated above the benzene moiety, i.e. in an axial position.
Although the active site of PCL has been modelled23,24 exten-
sively and an empirical rule25–27 [Fig 4(a)] has been introduced
to predict the absolute configuration of the faster reacting
enantiomer of secondary alcohols, the prediction of the faster
reacting enantiomer of rac-10 was not straightforward. In
the empirical rule, the prediction is based on the difference in
the size of substituents flanking the stereogenic center of the
Fig. 2 Octant rule diagram for the ketone carbonyl n π* transition
of (2S,6R)-(Ϫ)-6; for clarity, the third nodal surface is represented as a
nodal plane. The signs depict the contribution of the octants.
most likely located in the front upper right or lower left sector
while the rest is in the rear upper right or lower left, which
makes the prediction even more ambiguous. Therefore, the
measured positive n π* CD transition (∆ε = 0.23 at 316 nm)
could not be used to deduce unequivocally the absolute
configuration of (Ϫ)-6.
Since the resolution with (ϩ)-(1-phenylethyloxamoyl)-
hydrazine provided very low yields due to the syn–anti isomer-
ism and the stereochemistry of the resolved compounds were
not determined, another resolving agent, the (2R,3R)-(Ϫ)-
butane-2,3-diol,21 was tried. The reaction of rac-6 with
(2R,3R)-(Ϫ)-butane-2,3-diol in toluene with a catalytic amount
of toluene-p-sulfonic acid (PTS) afforded a 1 : 1 mixture of the
two corresponding diastereomers (ϩ)-9a and 9b (Scheme 2),
as shown by the 1H NMR spectrum of the crude product.
Although these compounds [(ϩ)-9a, 9b] could not be separated
by TLC, their crystallization from concentrated methanolic
solution yielded colourless needles whose 1H NMR and optical
rotation proved that they consisted of a pure dextrarotatory
diastereomer. NOE experiments and the magnetic anisotropy
effect of the phenyl ring on the methyl groups of the ketal
moiety turned out to be insufficient to determine the absolute
configuration. However, a single crystal grown in methanol was
found suitable for X-ray measurement, from which the struc-
ture of (2R,6S)-(ϩ)-9a was deduced (Fig. 3), relying on the
Fig. 3 ORTEP view and numbering scheme of (ϩ)-9a at the 50%
probability level. Selected torsional angles (Њ): C(10a)–O(1)–C(2)–
C(11), ϩ31.3; O(1)–C(2)–C(11)–C(6), Ϫ60.3; C(2)–C(11)–C(6)–C(7a),
ϩ58.8; C(11)–C(6)–C(7a)–C(10a), Ϫ29.6; C(6)–C(7a)–C(10a)–O(1),
ϩ0.6; C(7a)–C(10a)–O(1)–C(2), Ϫ1.0.
890
J. Chem. Soc., Perkin Trans. 1, 2002, 888–894

View Article Online
Scheme 3 Kinetic resolution of the 4-hydroxy derivative rac-10 with lipase from Pseudomonas cepacia and configurational assignment of the
resolved enantiomers by chemical correlation.
skeletons (ϩ)-1 and (Ϫ)-1 with no substituents at C-4, the
optically active alcohols (Ϫ)-10 and (ϩ)-10 were tosylated and
reduced with LAH, respectively (Scheme 4). While the optical
Fig. 4 (a) Simple model to predict the faster reacting enantiomer
in the kinetic resolution of secondary alcohols with lipase from
Pseudomonas cepacia (PCL); M and L depict the medium and large
substituents flanking the stereogenic center based on differentiation by
the enzyme. (b) The faster reacting enantiomer [(2R,4S,6R)-(ϩ)-10]
in the kinetic resolution of rac-10 oriented in the same manner as in
Fig. 4(a); M, L assignment of the two sides flanking the C-4 stereogenic
center.
hydroxy group. In our case, the α-substituents are the same,
namely, methylene groups, and the effect of the C-2 methyl
group at the β position should be compared with the more
remote but bulkier phenyl ring that contributes to the other
side. In order to clarify this question and unequivocally deduce
the absolute configuration of the resolved molecules, the well
established 2R,6S absolute configuration of the acetal (ϩ)-9a
was chemically correlated with the products of the kinetic reso-
lution. The hydrolysis of (ϩ)-9a resulted in the dextrorotatory
ketone of 2R,6S absolute configuration [(ϩ)-6] whose reduction
with LAH in THF provided the dextrorotatory alcohol (ϩ)-10
with known absolute configuration (Scheme 3). Since the
remaining alcohol in the transesterification of rac-10 [(Ϫ)-10]
was levorotatory, it must be the enantiomer of the alcohol
obtained from the reduction of (2R,6S)-(ϩ)-6. The saponifi-
cation of the acetate (ϩ)-11 indeed afforded the dextrorotatory
alcohol (ϩ)-10 (Scheme 3). Thus, the absolute configuration of
the faster reacting alcohol (ϩ)-10 is 2R,4S,6R because the
Cahn–Ingold–Prelog priority of the groups flanking the C-6
stereogenic center is exchanged after the reduction of the
carbonyl group. Moreover, the addition of hydride took place
in a stereoselective manner yielding the axial alcohol as the
major product which defined the S absolute configuration at
C-4.
This means the alcohol (ϩ)-10 with 4S absolute configur-
ation was acylated faster than its enantiomer in the presence of
PCL, which rendered the C-5 side with the C-6 phenyl group
the large side (L) and the C-3 side with the C-2 methyl group the
medium side (M) according to the empirical rule25–27 [Fig 4(b)].
This could be interpreted that the large pocket of PCL, lined
with hydrophobic side-chains, attracts the phenyl group, while
the C-2 methyl group and the oxygen are oriented toward the
medium-sized pocket that contains both polar and hydrophobic
side chains. Since the optical purity of the crystallized ketal
(ϩ)-9 was found to be over 99%, and it was chemically correl-
ated with the products of the enzymatic resolution through its
hydrolysis to the ketone (ϩ)-6, the comparison of the optical
rotations gave the ee of the remaining alcohol (Ϫ)-10 as 91%
with a conversion of 52%. In order to produce the tricyclic
Scheme 4 Removal of the hydroxy group from the optically active
4-hydroxy derivatives (Ϫ)-10 and (ϩ)-10; the structures depict the
transformation of (Ϫ)-10.
rotations changed signs when the C-4 stereogenic center was
removed, i.e. (Ϫ)-10 was converted into (ϩ)-1 or (ϩ)-10 to (Ϫ)-
1
1, the sign of the Lb band CD at 275–285 nm did not change.
Since the CD spectra of (Ϫ)-1 and (ϩ)-1 are determined by the
chirality of the heterocyclic ring and there are no more remote
stereogenic centers to interfere, they are suitable model com-
pounds to study the effect of axial benzylic substituents on the
chromane chromophore. They can also be used for comparison
with compounds that contain stereogenic centers in the cyclo-
hexane ring to evaluate the contribution of remote spheres to
the 1Lb band CD. Further chiroptical studies of these and
related compounds, as well as their theoretical CD calculations,
are in progress.
Experimental
Melting points were determined with a Kofler hot-stage appar-
atus and are uncorrected. Microanalyses were performed on a
Carlo-Erba analyser Tpy 1106. Optical rotations were meas-
ured with a Carl Zeiss (Jena) Polamat-A polarimeter, and the
CD and UV spectra with a slightly modified Jobin-Yvon-Isa
Dichrograph-6. [α]D values are given in 10Ϫ1 deg cm2 gϪ1. 1H and
13C spectra were recorded with a Varian Unity-Inova spec-
trometer with TMS as the internal standard (δ = 0) for solutions
in CDCl3. Integrals were always in agreement with the assigned
number of protons. The coupling constants J are quoted in Hz.
J. Chem. Soc., Perkin Trans. 1, 2002, 888–894
891

View Article Online
Flash chromatography was carried out using Merck Kieselgel
oil, a mixture of the two diastereomers. A concentrated methan-
60 (0.040–0.063 mm).
olic solution of the mixture was prepared from which 89 mg
(14%) (ϩ)-9a crystallized overnight as colorless needles. A few
crystals were recrystallized from methanol which afforded
single crystals suitable for X-ray analysis.
Isomeric (؉)-2-(1-phenylethyloxamoyl)hydrazones of 5,6-
dihydro-2-methyl-2,6-methano-2H-1-benzoxocin-4(3H)-one
[(E )-(؊)-8, (Z )-(؉)-8 and (Z )-(؊)-8]
X-ray parameters:† colorless prism crystals (0.42 × 0.3 ×
0.22 mm) of C17H22O3, M = 274.35, orthorhombic, a = 6.062(1)
Å, b = 15.112(1) Å, c = 16.992(1) Å, V = 1556 Å3, Z = 4, space
group: P212121, ρcalc = 1.171 g cmϪ3. Data were collected at
293(1) K, Enraf–Nonius MACH3 diffractometer, Mo-Kα
radiation λ = 0.71073 Å, ω-2θ motion, θmax = 27Њ, 2047 unique
reflections were measured, of which 1041 were with I > 2σ(I ),
decay: 4%. The structure was solved using the SIR-92 soft-
ware28 and refined on F 2 using the SHELX-9729 program. Pub-
lication material was prepared with the WINGX-97 suite,30
R(F) = 0.0501 and wR(F 2) = 0.1335 for 2047 reflections, 182
parameters.
rac-6 (436 mg, 2.15 mmol) and 7 (522 mg, 2.52 mmol) were
dissolved in methanol (30 ml) and refluxed for 2 h after the
addition of an iodine crystal. TLC monitoring of the reaction
showed the formation of four new compounds [(Z)-(ϩ)-8/(E)-
(ϩ)-8, (Z)-(Ϫ)-8/(E)-(Ϫ)-8]. After 2 h, the reaction mixture was
cooled and stored for overnight crystallization at rt, which
resulted in a mixture of (Z)-(ϩ)-8 and (E)-(Ϫ)-8 (216 mg).
Repeated crystallization of this mixture from methanol yielded
the pure (Z)-(ϩ)-8 (64 mg, 15%). (E)-(Ϫ)-8 (57 mg, 14%) was
obtained from the mother liquor of the first crystallization
which was concentrated and recrystallized. Hexane and toluene
were also tested as solvents for crystallization but they were
found less effective in the separation.
Mp 74–76 ЊC, [α]2D0 = ϩ44.34 (c 0.20, chloroform) (Found: C,
74.35; H, 8.07; C17H22O3 requires C, 74.42, H 8.08%); λmax(CH3-
CN)/nm 200.0 (ε × 10Ϫ4/MϪ1 cmϪ1 3.53), 213.6 (4.61), 220.0
(0.50), 246.2 (0.02), 277.2 (0.16); CD in CH3CN nm (∆ε) 204.6
(Ϫ7.52), 226.0 (5.96), 277.2 (3.92), 283.8 (3.64); δH(400 MHz,
C6D6) 0.65 (3H, d, J 8.5, C-3Ј CH3), 0.85 (3H, d, J 8.5, C-2Ј
Me), 1.21 (3H, s, 2-Me), 1.34 (1H, dd, J 12.8 and 2.9, 11-Hax),
1.54 (1H, m, 11-Heq), 1.67 (1H, d, J 14.0, 3-Hax), 1.71 (1H, dd,
J 13.7 and 4.1, 5-Hax), 1.88 (1H, m, 5-Heq), 2.17 (1H, m, 3-Heq),
2.70 (1H, m, 6-H), 2.99 (1H, dd, J 8.50 and 6.8, 3Ј-H), 3.20 (1H,
dd, 2Ј-H), 6.68 (1H, td, J 8.1, 7.1 and 1.4, 9-H), 6.80 (1H, dd,
J 8.1 and 1.3, 10-H), 6.84 (1H, dd, J 7.3 and 1.4, 7-H), 6.97 (1H,
m, 8-H); δc(100 MHz, CDCl3) 16.25 (C-3Ј Me), 17.19 (C-2Ј
Me), 29.39 (2-Me), 32.44 (C-6), 35.52 (C-5), 42.12 (C-11), 49.20
(C-3), 74.99 (C-2), 76.70 (C-2Ј), 78.07 (C-3Ј), 107.01 (C-4),
115.12 (C-10), 118.79 (C-8), 126.23 (C-7a), 127.31 and 127.32
(C-7 and C-9), 155.30 (C-10a).
(Z)-(ϩ)-8: mp 209–211 ЊC, [α]2D0 = ϩ224.09 (c 0.08, chloro-
form) (Found: C, 70.56; H, 6.42; C23H25N3O3 requires C, 70.57,
H 6.44%); λmax(CH3CN)/nm 199.6 (ε × 10Ϫ4/MϪ1 cmϪ1 3.11),
228.2 (0.97), 251.4 (0.61), 263.6 (0.64), 277.4 (0.50), 285.0
(0.40); CD in CH3CN nm (∆ε) 227.4 (26.79), 264.2 (Ϫ15.17),
281.0sh (4.17), 286.8 (8.44); δH (200 MHz, CDCl3) 1.51 (3H, s,
2-Me), 1.55 (3H, d, J 6.9, 6Ј-Me), 2.10 (2H, m, 11-H), 2.15 (1H,
d, J 16.0, 3-Hax), 2.67 (1H, m, 5-Heq), 2.80 (1H, m, 3-Heq), 2.91
(1H, m, 6-H), 5.00 (1H, m, 6Ј-H), 6.68–6.78 (2H, m, 10-H and
8-H), 6.94–7.09 (2H, m, 7-H and 9-H), 7.26–7.38 (5H, m, Ph),
7.68 (1H, d, J 8.0, NH), 10.03 (1H, s, ᎐N–NH).
᎐
(E)-(Ϫ)-8: mp 215–216 ЊC, [α]2D0 = Ϫ7.79 (c 0.08, chloroform)
(Found: C, 70.57; H, 6.42; C23H25N3O3 requires C, 70.57, H
6.44%); λmax(CH3CN)/nm 215.2 (ε × 10Ϫ4/MϪ1 cmϪ1 1.5), 227.2
(1.23), 251.6 (0.76), 263.8 (0.79), 276.8 (0.77), 284.4 (0.53); CD
in CH3CN nm (∆ε) 214.8sh (Ϫ6.20), 226.2 (Ϫ16.44), 263.6
(25.96), 280.8sh (Ϫ3.83), 286.2 (Ϫ9.82), 300.2 (0.91); δH (400
MHz, CDCl3) 1.50 (3H, s, 2-Me), 1.51 (3H, d, J 6.9, 6Ј-Me),
2.10 (2H, m, 11-H), 2.33 (1H, m, J 4.0, 5-Hax), 2.49 (1H, d,
3-Hax), 2.90 (1H, m, J 2.5 and 2.0, 5-Heq), 3.00 (1H, m, J 2.0 and
1.5, 3-Heq), 3.30 (1H, m, 6-H), 5.01 (1H, m, 6Ј-H), 6.71 (1H, m,
10-H), 6.74 (1H, m, 8-H), 6.96 (1H, m, 7-H), 7.04 (1H, m, 9-H),
Enantiomeric 5,6-dihydro-2-methyl-2,6-methano-2H-1-benz-
oxocin-4(3H)-ones [(؊)-6 and (؉)-6] from oxamoylhydrazones
To a solution of (E)-(Ϫ)-8 (20 mg, 0.051 mmol) in methanol, a
few drops of sulfuric acid (25%) were added and the reac-
tion mixture was stirred for 5 hours. It was neutralized with
NaHCO3 and then extracted with dichloromethane. The
organic layer was collected and concentrated, and its purifi-
cation by preparative TLC gave 7.6 mg (74%) of a white crystal-
line product (Ϫ)-6. (ϩ)-6 was prepared in 65% yield in the same
manner starting from (Z)-(ϩ)-8.
7.09–7.70 (5H, m, Ph), 7.68 (1H, d, J 8.0, NH), 9.88 (1H, s, ᎐N–
᎐
NH); δc(100 MHz, CDCl3) 21.60 (6Ј-Me), 28.79 (C-6 Me), 32.50
(C-6), 34.89 (C-5), 35.10 (C-11), 47.47 (C-3), 49.60 (C-6Ј), 75.35
(C-2), 116.53 (C-10), 120.22 (C-8), 123.26 (C-7a), 126.09,
128.52, 128.77 and 140.63 (Ph), 127.95 (C-7), 128.52 (C-9),
153.48 (C-10a), 155.14 and 158.59 (C-3Ј and C-4Ј), 160.11
(C-4).
(Z)-(Ϫ)-8: δH(400 MHz, CDCl3) 1.52 (3H, s, 2-Me), 1.54
(3H, d, J 6.9, 6Ј-Me), 2.10 (2H, m, 11-H), 2.15 (1H, d, 3-Hax),
2.68 (1H, m, 5-Hax), 2.86 (m, 1H, H-5eq), 3.02 (1H, m, J 2.0 and
1.6, 3-Heq), 3.29 (1H, m, 6-H), 5.00 (1H, m, 6Ј-H), 6.69 (1H, m,
10-H), 6.80 (1H, m, 8-H), 6.99 (1H, m, 7-H), 7.06 (1H, m, 9-H),
7.09–7.70 (5H, m, Ph), 7.96 (1H, d, J 8.0, NH), 10.07 (1H, s,
(Ϫ)-6: mp 121–122 ЊC, [α]2D0 = Ϫ15.6 (c 0.15, chloroform)
(Found: C, 70.15; H, 6.96; C13H14O2 requires C, 77.20, H
6.98%); λmax(hexane)/nm 276.8 (ε × 10Ϫ3/MϪ1 cmϪ1 1.96), 283.8
(1.91); CD in hexane nm (∆ε) 277.0 (Ϫ2.91), 285.6 (Ϫ2.27),
308.4 (0.25), 316.0 (0.23); NMR data are the same as given in
reference 2.
(ϩ)-6: Mp 123–125 ЊC, [α]2D0 = ϩ19.7 (c 0.11, chloroform);
λmax(hexane)/nm 276.8 (ε × 10Ϫ4/MϪ1 cmϪ1 1.96), 283.8 (1.91);
CD in hexane nm (∆ε) 205.0 (20.47), 224.8 (3.16), 246.0
(Ϫ0.23), 277.2 (2.84), 284.6 (2.66) 308.8 (Ϫ0.27), 315.2 (Ϫ0.26);
CD in CH3CN nm (∆ε) 205.8 (22.35), 227.2 (2.65), 246.8
(Ϫ0.32), 278.2 (2.56), 285.0 (2.89), 313.8 (Ϫ0.22).
᎐N–NH); δ (100 MHz, CDCl ) 21.47 (6Ј-Me), 29.04 (2-Me),
᎐
c
3
32.69 (C-6), 35.34 (C-11), 39.45 (C-3), 42.45 (C-5), 49.67 (C-6Ј),
74.75 (C-2), 115.97 (C-10), 120.46 (C-8), 123.84 (C-7a), 126.06,
128.20, 128.61 and 141.65 (Ph), 127.73 (C-7), 128.76 (C-9),
152.88 (C-10a), 155.20 and 158.48 (C-3Ј and C-4Ј), 160.44
(C-4).
(؉)-6 from the ketal (؉)-9
(ϩ)-9 (64 mg, 0.23 mmol) was refluxed in dioxane with 5 ml
dilute hydrochloric acid (10%). The mixture was neutralized
with NaHCO3 solution and extracted with dichloromethane.
The crude product was purified by preparative TLC in 8 : 1
toluene–ethyl acetate to give 32 mg (68%) of a crystalline
substance, mp 124–126 ЊC, [α]2D0 = ϩ26.1 (c 0.10, chloroform).
The ketal of 5,6-dihydro-2-methyl-2,6-methano-2H-1-benz-
oxocin-4(3H)-one with (2R,3R)-(؊)-butane-2,3-diol [(؉)-9a]
rac-6 (472 mg, 2.33 mmol) was dissolved in toluene, 230 mg
(2.55 mmol) (2R,3R)-(Ϫ)-butane-2,3-diol and 10 mg PTS were
added and the mixture was refluxed for 12 h. Afterwards it was
diluted with dichloromethane and extracted with water and
saturated NaHCO3 solution. The crude product was purified by
flash chromatography to give 507 mg (79%) of a light-yellow
† CCDC reference number 178166. See http://www.rsc.org/suppdata/
p1/b2/b200751g/ for crystallographic files in cif. or other electronic
format.
892
J. Chem. Soc., Perkin Trans. 1, 2002, 888–894

View Article Online
(2R,4S,6R)-(؉)-3,4,5,6-Tetrahydro-2-methyl-2,6-methano-2H-
1-benzoxocin-4-ol [(؉)-10] from the reduction of the ketone
(؉)-6
(2S,6S,4R)-(Ϫ)-12: mp 132–133 ЊC, [α]2D0 = Ϫ19.15 (c 0.09,
chloroform); (Found: C, 66.95; H, 6.15; C15H18O3 requires C,
67.01, H 6.19%); δH(200 MHz, CDCl3) 1.39 (3H, s, 2-Me), 1.74–
1.98 (4H, overlapping m, 11-H, 3-Hax and 5-Hax), 2.20–2.37
(2H, m, 3-Heq and 5-Heq), 2.40 (3H, s, tosyl Me), 3.03 (1H, m,
6-H), 5.07 (1H, m, 4-H), 6.60 (1H, m, 7-H).
(2R,6R,4S)-(ϩ)-12 was prepared from (ϩ)-10 as described
above for (Ϫ)-12, mp 133–135 ЊC, [α]2D0 = ϩ19.80 (c 0.12,
chloroform).
(ϩ)-6 (14 mg, 0.07 mmol) was reduced with 10 mg LAH in dry
THF. After 30 minutes, a few drops of water were added to
decompose the excess LAH. After the addition of brine the
extraction was carried out with methylene chloride. Concen-
tration of the organic layer and the subsequent purification by
preparative TLC with 4 : 1 toluene–ethyl acetate gave 11.8 mg
(84%) of a slowly crystallizing oil (ϩ)-10.
Enantiomeric 3,4,5,6-tetrahydro-2-methyl-2,6-methano-2H-1-
benzoxocines (2R,6S)-(؉)-1 and (2S,6R)-(؊)-1
Mp 81–83 ЊC, [α]2D0 = ϩ32.12 (c 0.22, chloroform) (Found:
C, 76.38; H, 7.91; C13H16O2 requires C, 76.44, H 7.90%);
λmax(CH3CN)/nm 198.8 (ε × 10Ϫ4/MϪ1 cmϪ1 4.14), 219.8 (0.63),
276.6 (0.29), 283.2 (0.28); CD in CH3CN nm (∆ε) 195.6 (4.40),
204.4 (Ϫ5.35), 214.0sh (0.82), 226.8 (6.00), 270.8 (1.16), 281.0
(1.06), 284.8 (0.92); δH(400 MHz, CDCl3) 1.42 (3H, s, 2-Me),
1.62 (1H, br s, OH), 1.75 (1H, dd, J 13.3 and 3.0, 11-Hax), 1.81
(1H, dd, J 15.2 and 5.2, 3-Hax), 1.88 (1H, m, 11-Heq), 1.95 (1H,
m, 5-Hax), 2.14 (1H, m, 5-Heq), 2.22 (1H, m, 3-Heq), 3.06 (1H, m,
6-H), 4.04 (1H, m, 4-H), 6.78 (1H, m, 8-H), 6.78 (1H, over-
lapped m, 10-H), 7.02 (1H, overlapped m, 7-H). 7.04 (1H, m,
9-H).
(2S,6S,4R)-(Ϫ)-12 (28 mg, 0.08 mmol) was reduced with LAH
(20 mg, 0.52 mmol) in dry 1 : 1 THF-ether under argon. After
5 hours, NH4Cl solution was added to the reaction mixture,
which was then extracted with dichloromethane. The crude
product was purified with 1 : 1 hexane–toluene by preparative
TLC to give (2R,6S)-(ϩ)-1 (6 mg, 41%) as an oil with character-
istic spicy odor.
(2R,6S)-(ϩ)-1: [α]2D0 = ϩ17.41 (c 0.12, chloroform) (Found:
C, 82.84; H, 8.51; C13H16 Orequires C, 82.94, H 8.57%);
λmax(CH3CN)/nm 199.0 (ε × 10Ϫ4/MϪ1 cmϪ1 2.41), 276.6 (0.13),
283.0 (0.12); CD in CH3CN nm (∆ε) 203.8 (5.79), 227.8
(Ϫ3.05), 276.2 (Ϫ0.67), 283.0 (Ϫ0.57); δH(200 MHz, CDCl3)
1.43 (3H, s, 2-Me), 1.5–2.0 (8H, overlapping m, 3-H, 4-H, 5-H,
11-H), 3.03 (1H, m, 6-H), 6.74–7.22 (4H, m, ArH); δc(50 MHz,
CDCl3) 18.19 (C-4), 29.35 (2-Me), 32.58 (C-5), 32.95 (C-6),
35.86 (C-3), 39.35 (C-11), 74.63 (C-2), 114.96 (C-10), 119.01
(C-8), 125.94 (C-7a), 127.28 (C-9), 128.99 (C-7), 156.35
(C-10a).
(2R,4S,6R)-(؉)-4-Acetoxy-3,4,5,6-tetrahydro-2-methyl-2,6-
methano-2H-1-benzoxocine [(؉)-11]: kinetic resolution of rac-10
rac-10 (190 mg, 0.93 mmol) was stirred in dry ether with 1 ml
vinyl acetate and 81 mg lipase from Pseudomonas cepacia (on
Toyonite). The reaction was followed by TLC and stopped after
16 days. The enzyme was filtered off and the solvent was evap-
orated. Column chromatography of the residue with 6 : 1
hexane–ethyl acetate gave 93 mg acetate (ϩ)-11 and 71 mg
alcohol (Ϫ)-10 (52% conversion).
(2S,6R)-(Ϫ)-1: [α]2D0 = Ϫ15.21 (c 0.2, chloroform); λmax(CH3-
CN)/nm 199.0 (ε × 10Ϫ4/MϪ1 cmϪ1 3.00), 276.6 (0.15), 283.4
(0.15); CD in CH3CN nm (∆ε) 203.0 (Ϫ5.75), 227.8 (4.21),
276.4 (0.83), 282.4 (0.79).
(Ϫ)-10: mp 75–76 ЊC, [α]2D0 = Ϫ38.88 (c 0.10, chloroform)
λmax(CH3CN)/nm 198.8 (ε × 10Ϫ4/MϪ1 cmϪ1 4.14), 219.8 (0.63),
276.6 (0.29), 283.2 (0.28); CD in CH3CN nm (∆ε) 195.0
(Ϫ6.41), 204.6 (5.25), 217.2sh (Ϫ2.39), 227.2 (Ϫ6.39), 276.8
(Ϫ1.93), 282.8 (Ϫ1.80).
Acknowledgements
(ϩ)-11: mp 101–102 ЊC, [α]2D0 = ϩ95.13 (c 0.10, chloroform)
(Found: C, 73.12; H, 7.35; C15H18O3 requires C, 73.15, H
7.37%); δH(400 MHz, CDCl3) 1.28 (3H, s, COCH3), 1.39 (3H, s,
2-Me) 1.72 (1H, dd, J 13.3 and 3.2, 11-Hax), 1.74 (1H, dd, J 15.2
and 5.0, 3-Hax), 1.84 (1H, m, 5-Hax), 1.87 (1H, m, 11-Heq), 2.13
(1H, m, 5-Heq), 2.14 (1H, m, 3-Heq), 3.01 (1H, m, 6-H), 4.95
(1H, m, 4-Heq), 6.66 (1H, m, 8-H), 6.67 (1H, m, 10-H), 6.96
(1H, m, 7-H), 7.02 (1H, m, 9-H); δc(100 MHz, CDCl3) 20.53
(COMe), 29.62 (2-Me), 30.56 (C-6), 35, 57 (C-11), 36.54 (C-5),
41.88 (C-3), 68.16 (C-4), 72.90 (C-2), 115.41 (C-10), 119.03
(C-8), 126.79 (C-7a), 127.47 (C-7), 127.71 (C-9), 155.43
The authors thank the National Science Foundation (OTKA,
Grant No. T034250) for financial support. Attila Bényei is
indebted to the Hungarian Academy of Sciences for a Bolyai
Postdoctoral Fellowship. This paper is dedicated to Professor
Nina Berova and Professor Koji Nakanishi.
References
1 J. Asano, K. Chiba, M. Tada and T. Yoshii, Phytochemistry, 1996,
41, 815.
2 F. Eiden and P. Gmeiner, Arch. Pharm. (Weinheim), 1987, 320,
213.
3 M. Yamazaki, Y. Satoh, Y. Maebayashi and Y. Horie, Chem. Pharm.
Bull., 1988, 36, 670.
4 K. Kojiri, S. Nakajima, A. Fuse, H. Suzuki and H. Suda,
J. Antibiot., 1995, 48, 1506.
5 S. Antus, E. Baitz-Gács, J. Kajtár, G. Snatzke and A. L. Tökés,
Liebigs Ann. Chem., 1994, 497.
6 T. Kurtán, E. Baitz-Gács, Z. Májer and S. Antus, J. Chem. Soc.,
Perkin Trans. 1, 2000, 453.
7 G. Snatzke, F. Snatzke, A. L. Tökés, M. Rákosi and R. Bognár,
Tetrahedron, 1973, 29, 909.
8 S. Antus, T. Kurtán, L. Juhász, L. Kiss, M. Hollósi and Z. Májer,
Chirality, 2001, 13, 493.
(C-10a), 170.58 (C᎐O).
᎐
(2R,4S,6R)-(؉)-3,4,5,6-Tetrahydro-2-methyl-2,6-methano-2H-
1-benzoxocin-4-ol [(؉)-10] from the hydrolysis of (؉)-11
70 mg (0.28 mmol) (ϩ)-11 was reacted with 5 ml 1 M NaOMe
in methanol at room temperature for 24 h. The mixture was
then concentrated under vacuum, and the residue was dissolved
in dichloromethane and washed with water. Separation by pre-
parative TLC in 4 : 1 toluene–ethyl acetate gave 50 mg (86%) of
(ϩ)-10, mp 81–83 ЊC, [α]2D0 = ϩ38.27 (c 0.11, chloroform).
9 R. Kuhn and D. Weiser, Chem. Ber., 1955, 88, 1601.
10 A. L. Tökés and L. Szilágyi, Synth. Commun., 1987, 17, 1235.
11 G. Nadler, I. Delimoge, P. Lahouratate, I. Leger, M. Morvan and
R. G. Zimmermann, Eur. J. Med. Chem., 1996, 31, 805.
12 Y. Xia, Z. Y. Yang, P. Xia, K. F. Bastow, Y. Tachibana, S. C. Kuo,
E. Hamel, T. Hackl and K. H. Lee, J. Med. Chem., 1998, 41,
1155.
13 A. L. Tökés, Liebigs Ann. Chem., 1987, 1007.
14 A. Srivastava, V. Srivastava, S. M. Verma and E. Subramanian,
J. Org. Chem., 1994, 59, 3560.
Enantiomeric 3,4,5,6-tetrahydro-2-methyl-2,6-methano-4-tosyl-
oxy-2H-1-benzoxocines (2R,4S,6R)-(؉)-12 and (2S,4R,6S)-
(؊)-12
Toluene-p-sulfonyl chloride (300 mg, 1.57 mmol) was added to
a stirred solution of (Ϫ)-10 (70 mg, 0.34 mmol) in dry pyridine.
After 48 hours, the reaction mixture was diluted with dichloro-
methane and washed with hydrochloric acid solution (10%).
The crude product was purified by preparative TLC in 8 : 1
toluene–ethyl acetate to yield (Ϫ)-12 (78 mg, 64%).
15 G. Giorgi, F. Ponticelli, L. Savini, L. Chiasserini and C. Pellerano,
J. Chem. Soc., Perkin Trans. 2, 2000, 2259.
J. Chem. Soc., Perkin Trans. 1, 2002, 888–894
893

View Article Online
16 T. Nakazawa, Y. Mizuta, A. Kawahara, E. Miyamoto and N. Muto,
Chem. Lett., 1992, 1125.
17 N. Berova and K. Nakanishi, in Circular Dichroism, Principles
and Applications, eds. N. Berova, K. Nakanishi and R. W. Woody,
Wiley-VCH, New York, 2000.
18 N. L. Allinger and C. K. Riew, J. Org. Chem., 1975, 40, 1316.
19 D. A. Lightner, T. D. Bouman, A. E. Hansen and W. M. D.
Wijekoon, J. Am. Chem. Soc., 1986, 108, 4484.
24 K. Lemke, M. Lemke and F. Theil, J. Org. Chem., 1997, 62,
6268.
25 R. J. Kazlauskas, A. N. E. Weissfloch, A. T. Rappaport and L. A.
Cuccia, J. Org. Chem., 1991, 56, 2656.
26 K. Nakamura, M. Kawasaki and A. Ohno, Bull. Chem. Soc. Jpn.,
1996, 69, 1079.
27 K. Burgess and L. D. Jennings, J. Am. Chem. Soc., 1991, 113,
6129.
20 G. Snatzke, B. Ehrig and H. Klein, Tetrahedron, 1969, 25, 5601.
21 W. D. Fessner and H. Prinzbach, Tetrahedron, 1986, 42, 1797.
22 U. T. Bornscheuer and R. J. Kazlauskas, Hydrolases in Organic
Synthesis, Wiley-VCH, Weinheim, 1999.
23 X. Grabuleda, C. Jaime and A. Guerrero, Tetrahedron: Asymmetry,
1997, 8, 3675.
28 A. Altomare, G. Cascarano, C. Giacovazzo and A. Guagliardi,
J. Appl. Crystallogr., 1993, 26, 343.
29 G. M. Sheldrick, SHELXL-97, Universität Göttingen, Germany,
1997.
30 L. J. Farrugia, WINGX-97, University of Glasgow, UK,
1996.
894
J. Chem. Soc., Perkin Trans. 1, 2002, 888–894