New ventures in the construction of complex heterocycles: Synthesis of morphine and hasubanan alkaloids

Article abstract of DOI:10.1055/s-1998-5927

Synthetic approaches to morphine and hasubanan alkaloids involving novel reactions for the formation of their dihydrobenzofuran, piperidine and pyrrolidine ring, respectively, are described. The importance of conformational constraints for the success of these reactions is demonstrated. The syntheses of new dioxepanes, diazepanes, dithiocanes, triazenes and dioxonanes are also reported.

Full text of DOI:10.1055/s-1998-5927

SYNTHESIS
April 1998
653
New Ventures in the Construction of Complex Heterocycles: Synthesis of Morphine
and Hasubanan Alkaloids
Dirk Trauner,a Stefanie Porth,b Till Opatz,b Jan W. Bats,b Gerald Giester,c Johann Mulzera*
a Institut für Organische Chemie der Universität Wien, Währingerstraße 38, A-1090 Vienna, Austria
Fax +43(1)313672280; E-mail: dixi@felix.orc.univie.ac.at; mulzer@felix.orc.univie.ac.at
b Institut für Organische Chemie der Johann Wolfgang Goethe-Universität, Marie-Curie-Straße 11, D-60439 Frankfurt, Germany
c Institut für Mineralogie und Kristallographie der Universität Wien, Althahnstraße 14, A-1090 Vienna, Austria
Received 29 August 1997
Abstract: Synthetic approaches to morphine and hasubanan alka-
loids involving novel reactions for the formation of their dihy-
drobenzofuran, piperidine and pyrrolidine ring, respectively, are
described. The importance of conformational constraints for the
success of these reactions is demonstrated. The syntheses of new
dioxepanes, diazepanes, dithiocanes, triazenes and dioxonanes are
also reported.
Key words: morphine alkaloids, hasubanan alkaloids, total syn-
thesis, ketenes
Introduction
(–)-Morphine (1), the main alkaloid of the opium poppy,
remains one of the most important drugs for the treatment
of severe pain and a true delight for synthetic chemists. In
spite of its small size and long history, it is still considered
to be an attractive target for total synthesis.1 Its molecular
architecture features a complicated network of three car-
bocycles and two heterocycles containing five vicinal ste-
reocenters, four of which define ring junctures. Among
these, the benzylic quaternary carbon (C-13) deserves cept: the basic idea was to provide the carbocyclic
special attention. The dihydrobenzofuran and the piperi- phenanthrene core bearing the correct oxygenation pattern
dine ring offer additional challenges which must be care- first, possibly in enantiomerically pure form, and then
fully addressed in any planned approach.
mount the ethanamine bridge onto this platform. The ri-
gidity of the tricyclic hydrophenanthrene was expected to
facilitate the reactions chosen for the construction of the
benzylic quaternary carbon and the introduction of the C2-
residue: either conjugate additions of organocuprates or
[3,3]-sigmatropic rearrangements. These reactions pro-
ceeded with limited success in preliminary studies with
conformationally more flexible substrates.3 Following
completion of the carbocyclic core and adaptation of the
functionality pattern, the heterocyclic rings would be
closed. Again, the formation of these rings was expected
to profit from conformational constraints imposed by the
rigidity of the polycyclic substrates. Despite the mole-
cule's early identification as a phenanthrene derivative, a
similar overall strategy has previously only been used
once in a successful synthesis of morphine.4
The hasubanan alkaloids, represented by (–)-hasuba-
nonine (2) and (–)-cepharamine (3), have a lot in common
with the morphine alkaloids. Although their physiological
effects are of little importance, their interesting structures
have stimulated numerous synthetic studies. Several total
syntheses have been reported.2 It will be noted that the
morphine and hasubanan alkaloids contain enantiomeric
but otherwise identical carbon skeletons, consisting of 14
phenanthrene carbons and a C2-residue connected with the
quaternary carbon. This C2-residue forms an ethanamine
bridge together with the N-CH3 moiety. In the case of the
hasubanans, the tertiary amine is located at an angular car-
bon (C-14), thus forming a 1,2-bridged pyrrolidine ring.
In contrast, the morphine alkaloids feature a piperidine
ring which includes a secondary stereogenic carbon (C-9)
and is 1,3-bridged across the phenanthrene skeleton. Any
approach primarily aiming at the construction of the car-
bon skeleton and deferring the closure of the nitrogen-
containing heterocycle to a later stage might therefore be
applicable to the synthesis of both classes of natural prod-
ucts.
Formal Synthesis of (–)-Morphine
Our first approach to morphine along these lines started
with commercially available 4-(3,4-dimethoxyphenyl)bu-
tyric acid (4) which was converted to phenanthrenone 5 in
three straightforward steps (Scheme 1).5 Chromatograph-
ic resolution on cellulose triacetate provided 5 g quantities
of the optically pure compound.
Over the past three years, we have developed a unified
strategy for the synthesis of the morphine and hasubanan
alkaloids which is founded on a seemingly simple con-

SYNTHESIS
Papers
654
Figure 1. Conformations of 6–8
Reagents and conditions: (a) Cl2/AcOH, 99%; (b) i) (COCl)2/benzene,
reflux, ii) SnCl4/benzene, 0°C, 71%, (c) i) HCO2Me/NaOMe/benzene,
ii) Methyl vinyl ketone/Et3N/MeOH, iii) KOH/dioxane/H2O, 77%; (d)
i) (H2C=CH2)2CuMgCl/THF, –78°C, ii) TMSCl/Et3N, 0°C to 25°C;
(e) NBS/THF, –100°C, 63% (combined yield from 5)
The enormous influence of the α-haloketone-effect on 7a
is further illustrated by its crystal structure at two different
temperatures (Figure 2).6 At room temperature, the mole-
cule adopts a conformation wherein the cyclohexanone
ring is more or less chair-shaped. The dihedral angle φ be-
tween the C=O and C–Br bond amounts to approximately
108˚, thus considerably deviating from the ideal value of
90˚ required by both the electrostatic and orbital-overlap
models. On cooling down to 100 K, the cyclohexanone
ring rearranges to a twist-form (Figure 2, Scheme 2). The
C–Br and C=O bonds are almost perpendicular in this
Scheme 1
Conjugate addition of (H2C=CH)2CuMgCl to 5, followed
by addition of TMSCl afforded silyl enol ether 6 which
was purified by recrystallization from pentane. In con-
trast, simple hydrolysis of the intermediary enolate fur-
nished ketone 7c which served as a key intermediate in
our studies directed toward hasubanan alkaloids (see be-
low). Treatment of silyl enol ether 6 with N-bromosuccin-
imide (NBS) in THF at low temperature produced a 3:1
mixture of diastereomeric bromoketones 7a,b (81% over-
all yield). This vicinal difunctionalization not only estab-
lished the benzylic quaternary carbon and introduced the
C2-residue but also provided the electrophilic functional-
ization of the neopentylic position required for the subse-
quent formation of the ether bridge.
The epimeric bromoketones 7a and 7b adopt completely
different conformations as demonstrated by X-ray
diffraction6 and 1H NMR spectroscopy (Figure 1). These
conformations can be readily explained in terms of the “α-
haloketone effect”. It is well known that α-bromocyclo-
hexanones prefer a conformation with an axial bromine in
order to minimize electrostatic interaction of the dipolar
C–Br and C=O bonds or maximize overlap between the
corresponding σ* and π-orbitals, respectively.7 Accord-
ingly, 7a is forced into the conformation shown on the left
(Figure 1). Obviously, the effect is strong enough to com-
pensate for the unfavorable axial position of the bulky ar-
omatic residue. In 7b, the α-haloketone effect and the
conformational preference of the aromatic moiety rein-
force each other and strongly stabilize the extended con-
formation shown. It is important to note that in the
absence of an additional conformational bias, hydro-
phenanthrenes of this type (e.g. compounds 6, 7c and 8, a
close derivative of 7c) clearly prefer the extended confor-
mation as can be seen in their X-ray crystal structures.6
Figure 2. X-Ray Crystal Structure of Bromoketone 7a at 295 K (abo-
ve) and 100 K (below). (Hydrogen atoms are omitted for clarity)

SYNTHESIS
April 1998
655
case (φ ≈ 92°). Notably, the rest of the molecule remains
virtually unchanged. Although these data refer to the solid
state, where intermolecular interactions play an important
role, it seems likely that this conformational rearrange-
ment is triggered by the predominance of the α-halo ke-
tone effect at low temperatures.
Scheme 2
With bromo ketone 7a in hand, we explored the formation
of the dihydrobenzofuran ring. This transformation was
first tested with the closely related “model compound” 9
lacking the 1-chloro substituent. Twofold demethylation
with excess boron tribromide followed by refluxing in tri-
ethylamine solution afforded a good overall yield of phe-
nol ketone 106 (Scheme 3).
Scheme 4
less efficient in the case of a substrate containing a confor-
mationally more flexible aromatic ring (Scheme 5). The
analogous ring closure of bromoketone 13 which differs
from 7a only in the missing ethylene bridge, required sev-
eral hours at 140°C and proceeded with modest yield due
to side reactions (Scheme 5). The resultant ketone 14 is a
potential synthetic precursor of galanthamine (15), an
amaryllidiaceae alkaloid which has recently attracted
much interest due to its anti-Alzheimer activity.10
Scheme 3
It was soon found that this ring closure might be achieved
in one synthetic step, leaving the second phenol ether in-
tact (Scheme 4). Following a precedent given by Grieco,8
7a was heated in DMF at 140°C. To our considerable sat-
isfaction, these conditions afforded ketone 12 in virtually
quantitative yield within 20 minutes. Mechanistically, the
reaction most likely proceeds via SN2-type displacement
of bromide by the ortho-methoxy group. The resulting
methyl oxonium ion 11 is rapidly demethylated by a
“counter-nucleophile”, the actual nature of which remains
to be determined (Br– vs. DMF). A four-membered tran-
sition state leading directly to methyl bromide and 12 is
considered to be unlikely for stereoelectronic reasons.
We attribute this particularly facile cyclic ether formation
to the conformational preorientation of 7a. As can be seen
in the solid state structures of 7a (Figure 2), the axial bro-
mine virtually folds the hydrophenanthrene framework
into a conformation similar to morphine which brings the
centers of nucleophilicity and electrophilicity into close
contact. Thus, the transition entropy of the ring closure is
less negative. The distance between C-5 and the proxi-
mate methoxy oxygen only amounts to approximately
2.7 Å in the crystal conformation. Not surprisingly, this
“enforced propinquity” then “leads to greater intimacy”.9
In accordance with this hypothesis, the reaction proved far
Scheme 5
The proposed SN2-mechanism of the reaction is further
supported by the fate of the undesired diastereomer
(Scheme 6). When heated under identical conditions, the
“wrong” isomer 7b remained unchanged for more than
half an hour. Addition of lithium bromide failed to effect

SYNTHESIS
Papers
656
the desired inversion of the secondary bromide and subse-
quent ring closure. Instead of the expected ketone 12, we
isolated the α,β-unsaturated ketone 17 in low yield.
Enone 17 obviously results from a “vicarious elimination”
of hydrogen bromide. This reaction probably follows the
Favorskii-type mechanism shown in Scheme 5. Intramo-
lecular displacement of bromine leads to the cyclopro-
panone intermediate 16. Abstraction of the β-proton and
fragmentation, followed by reprotonation of the resulting
dienolate then affords the α,β-unsaturated enone 17. The
identity of 17 was unambiguously confirmed by indepen-
dent synthesis via Saegusa oxidation of 7c and single
crystal X-ray structure determination.6
Reagents and conditions. (a) TMSCl/(CH2OH)2/CH2Cl2, 92%; (b) i)
BH3·SMe2/THF, ii) H2O2/OH–, 70%; (c) Raney Ni/MeOH/KOH,
98%; (d) PhSO2NHMe/ADDP/Bu3P, 81%; (e) NBS/(PhCO2)2/CCl4,
reflux, Et3N, 65%; (f) Li/NH3/THF/t-BuOH, –78°C, 79%; (g) 3 N
HCl, 90°C, 95%
Scheme 7
Scheme 6
We continued our synthesis with tetracyclic ketone 12
(Scheme 7). Protection of the carbonyl group, followed by
hydroboration/oxidation and dechlorination afforded the
corresponding primary alcohol which was transformed to
N-methylsulfonamide 18 via Mitsunobu reaction. At the
first glance, this intermediate seems to be difficult to func-
tionalize at C-9 which is a prerequisite for piperidine ring
formation. Closer inspection reveals that 18 contains a
benzylic site (C-10) which is activated towards radicals
and easily oxidized. Hence, it proved possible to directly
introduce a ∆9,10-double bond: benzylic bromination with
NBS followed by in situ dehydrobromination furnished
styrene 19 in acceptable yield along with the unchanged
starting material. An X-ray structure analysis confirmed
the benzylic position of the double bond beyond any
doubt and provided insight in the cup-like geometry of the
rigid tetracyclic framework of 19 (Figure 3).6
Figure 3. X-Ray Crystal Structure of Enantiomerically Pure Styrene
19 (Most hydrogen atoms are omitted for clarity)
tic” view of this interesting cyclization and Figure 4
shows the structure of the product in the crystal. Mecha-
nistically, the reaction probably proceeds via single elec-
tron reduction of the sulfonamide moiety to generate a
benzenesulfinate anion and a nitrogen radical 22 which
adds to the ∆9,10-double bond. The resultant stabilized
benzylic radical is then further reduced to the benzylic an-
ion 23 which is quenched by the tert-butanol present in the
reaction mixture. It cannot be ruled out that the sulfona-
mide is directly reduced to an N-anion which in turn un-
dergoes the corresponding addition/protonation sequence.
To the best of our knowledge, this is only the second re-
ported example of a successful “Parker-Focas” ring
closure which has been suggested to require conforma-
tionally restricted substrates.11
The stage was now set for the second heterocyclization.
Styrene 19 was designed to serve as a substrate for the
novel piperidine ring-closure developed by Parker and
Focas in their formal total synthesis of (±)-morphine.11 In-
deed, exposure of 19 to the action of excess lithium metal
in liquid ammonia afforded a good yield of dihydroco-
deinone ethyleneketal 20. Scheme 8 offers a more “realis-

SYNTHESIS
April 1998
657
amide 24. Reduction to the primary alcohol followed by
Mitsunobu-reaction gave N-methylbenzenesulfonamide
25. Based on precedents found in literature,13 we antici-
pated that epoxidation of the double bond and protonation
might trigger attack of the proximate methoxy group fol-
lowed by demethylation of the intermediary oxonium ion
(28) in an event closely reminiscent of the transetherifica-
tion described above. Indeed, diastereoselective epoxida-
tion of 25 with dimethyldioxirane followed by treatment
of the resultant β-epoxide 26 with trifluoroacetic acid
(TFA) in anhydrous THF resulted in ring closure with
concomitant demethylation to provide a good yield of sec-
omorphinane 29. It may be noted that this process can
only proceed through conformer 27b and not through 27a
which corresponds to the conformation of 24 in the crys-
tal. In the absence of a stereoelectronic bias comparable to
the α-halo ketone effect, it seems unlikely that 27b repre-
sents the ground state conformation of the molecule.
Considering our previous experiences with the conforma-
tional flexibility of the system, we assumed that the ener-
Scheme 8
Figure 4. X-Ray Crystal Structure of Dihydrocodeinone Ethylenketal
20 (Hydrogen atoms are omitted for clarity)
All that remained was the removal of the only protecting
group (Scheme 7). Hydrolysis of 20 under rather harsh
conditions afforded (-)-dihydrocodeinone (21), a standard
precursor of morphine, in excellent yield. When com-
bined with the efficient procedures for the conversion of
(–)-dihydrocodeinone to (–)-codeine and the facile O-
demethylation of (–)-codeine to (–)-morphine, our synthe-
sis represents an asymmetric formal synthesis of (–)-mor-
phine.
In addition to the cuprate chemistry we also used sigma-
tropic rearrangements to establish the benzylic quaternary
stereocenter (Scheme 9).12 This modification of our orig-
inal plan entailed a different tactical sequence for the clo-
sure of the dihydrobenzofuran ring. Diastereoselective
1,2-reduction of 5 to the corresponding allylic alcohol was
followed by an Eschenmoser–Claisen rearrangement
which afforded an acceptable overall yield of dimethyl
Reagents and conditions: (a) DIBAH/THF, –78°C, 80%; (b) N,N-di-
methylacetamide dimethyl acetal/toluene, reflux, 64%; (c) LiBHEt3/
THF, r.t., 96%; (d) PhSO2Me/ADDP/Bu3P, r.t, 90%; (e) dimethyl
dioxirane/CH2Cl2, 0°C to r.t., 80%; (f) TFA/THF, r.t., 88%; (g) H2/
Pd-C/Et3N/MeOH, r.t., 88%; (h) Swern oxidation, 90%; (i) TMSCl/
(CH2OH)2/CH2Cl2, r.t., 92%
Scheme 9

SYNTHESIS
Papers
658
gy difference between the two conformers would be small (H2C=CH)2CuMgCl furnished the entire carbon skeleton
enough to allow rapid interconversion and ring closure of the hasubanans in the form of ketone 7c (shown here as
which indeed proved to be the case. Dechlorination, the opposite enantiomer). This crucial intermediate was
followed by Swern oxidation and ketalization finally af- transformed to protected primary alcohol 30 through a
forded ethylene ketal 18, thus intercepting our previous straightforward series of functional group manipulations
route.
in high overall yield. Saegusa-oxidation14 then afforded
α,β−unsaturated ketone 31, placing the ∆7,8-double bond
into ring C. The functionalization of the angular position
(C-14), necessary for pyrrolidine ring closure, was
achieved by ketalization with concomitant double bond
migration into the ∆8,14−position. Removal of the silyl
protecting group finally gave an almost quantitative yield
of alcohol 32.
Synthesis of the Hasubanan Skeleton
Our studies on the synthesis of hasubanonine (3) and
cepharamine (4), respectively, also involved the conju-
gate addition to the sterically and electronically very chal-
lenging enone 5 (Scheme 10). Reaction of 5 with
At this point, several options for the introduction of the ni-
trogen and amination of the ∆8,14-double seemed possible.
Although some solutions to this problem involving elec-
trophilic activation of the alkene (e.g. by Hg2+) have been
described in the literature,15 we decided to explore a novel
sequence which would allow restitution of the original
∆7,8-double bond (Scheme 11). An intramolecular 1,3-di-
polar cycloaddition of an azide16 to a (masked) β,γ-unsat-
urated enone (38) was presumed to afford a triazene of
type 39. Subsequent elimination of N2 would furnish β-
aminoenone 41, perhaps via aziridine 40.
Scheme 11
Accordingly, primary alcohol 32 was converted to the
corresponding primary azide 34. This transformation
proved more challenging than anticipated. Activation of
the OH-function as methanesulfonate 33 followed by
treatment of the crude material with sodium azide fur-
nished a low yield of 34. The tertiary azide 35 and the
nine-membered dienol ether 36 were isolated as byprod-
ucts. These interesting compounds obviously arose from
intramolecular attack of a dioxolane oxygen on the meth-
anesulfonyl function in 33. The resultant cyclic carboxo-
nium ion either recombines with the azido-anion or
eliminates a proton to afford the tertiary azide and the di-
enol ether, respectively. On heating, 35 underwent ther-
mal elimination of HN3 to furnish a quantitative yield of
36. A single crystal X-ray structure determination con-
firmed the proposed structure of this unusual 9-membered
heterocycle (Figure 5).6
Reagents and conditions: (a) (H2C=CH)2CuMgCl/THF, –78˚ to 0°C;
(b) TMSCl/(CH2OH)2/CH2Cl2, 88%; (c) i) BH3·SMe2/THF, 0°C to
r.t., ii) H2O2/OH–, 73%; (d) 2 M HCl/THF, reflux, 98%; (e) TPSCl/
imidazole/DMAP/THF, 96%; (f) i) KHDMS/TMSCl/THF, –78°C; ii)
Pd(OAc)2/MeCN, 70%; (g) (CH2OH)2/TsOH/benzene, reflux, 77%;
(h) TBAF/THF, 99%; (j) i) MsCl/Et3N/CH2Cl2, ii) SiO2, 40%; (k)
NaN3/DMF, 23% 33 + 23% 34 + 26% 35; (l) benzene, reflux (76%);
With primary azide 34 in hand, we investigated the envis-
aged intramolecular 1,3-dipolar cycloaddition. Molecular
models suggested that the azido function and the double
Scheme 10

SYNTHESIS
April 1998
659
Figure 5. X-Ray Crystal Structure of Dioxonane 36 (Hydrogen atoms
are omitted for clarity)
bond can readily approach each other in the transition
state of the reaction to form a relatively unstrained prod-
uct. Indeed, refluxing compound 34 in benzene solution
accomplished formation of the triazene 37 which contains
the entire hasubanan framework. The result of a single
crystal structure determination performed with 37 is
shown in Figure 6.6
Reagents and conditions: (a) i) NsCl/Et3N/CH2Cl2, ii) SiO2, iii) NaN3/
DMF, 55%; (b) benzene, reflux, 76%; (c) Py, reflux, 73%
Scheme 12
the intermediacy of the corresponding aziridine. All at-
tempts to simultaneously introduce the N-methyl group
via quaternization of the triazene followed by elimination
have failed so far.
Compound 44 contains the entire skeleton of the hasuba-
nan alkaloids in stereochemically correct and highly func-
tionalized form. Its conversion to hasubanonine only
requires dimethoxylation of the enone moiety, dechlorina-
tion and N-methylation which are currently being pursued
in our laboratories. The refinement of the sequence lead-
ing to azido enone 42 is also under investigation. Al-
though the experiments described herein were conducted
in the racemic series, optically pure hasubanans might be
obtained easily in that fashion since the starting phenan-
threnone 5 can be resolved by chiral chromatography on a
preparative scale.5
Figure 6. X-Ray Crystal Structure of Triazene 37 (Hydrogen atoms
are omitted for clarity)
Unfortunately, it proved impossible to cleave the ethylene
ketal in the presence of the basic nitrogens without com-
Ketene Insertions
plete destruction of the material. All attempts to obtain the Ethylene ketals are more reliable protecting groups and
corresponding aziridine by thermal elimination of N2 are usually regarded as “innocent bystanders”.17 Their nu-
were also met with limited success. In the end it was nec- cleophilicity and their tendency to form carboxonium ions
essary to resort to the unprotected primary azide 42 should not, however, be underestimated. The unexpected
(Scheme 12). This compound was obtained unexpectedly formation of dioxonanes 35 and 36 was not the only occa-
upon conversion of the primary alcohol 32 to the corre- sion when an ethylene ketal interfered with our efforts to
sponding p-nitrobenzenesulfonate followed by treatment synthesize hasubanans. An alternative approach to these
of the crude product with sodium azide and silica gel chro- alkaloids was based on ketal 45, readily obtained by ket-
matography. For reasons which remain to be clarified, this alization of phenanthrenone 5 with concomitant double
also led to cleavage of the ethylene ketal. Remarkably, bond migration (Scheme 13). With its ∆13,14-double bond,
none of the dioxonanes 35 and 36 were observed under this achiral compound 45 might potentially give rapid ac-
these conditons. When heated in benzene, 42 underwent cess to the hasubanan skeleton. Unfortunately, a 1,3-dipo-
the desired 1,3-dipolar cycloaddition to afford triazene 43 lar synthon which might directly add the missing
in good yield. Gratifyingly, 43 proved more compliant ethanamine bridge is not available. We therefore turned
with the envisaged β-elimination of molecular nitrogen: our attention to the [2+2]-cycloaddition of dichlo-
refluxing triazene 43 in pyridine solution afforded a good roketene.18 It was hoped that the considerable nucleophi-
yield of γ-aminoenone 44. It remains to be investigated licity of the tetrasubstituted double bond would
whether this novel reaction proceeds directly or through compensate for the expected steric hindrance and that the

SYNTHESIS
Papers
660
This conceptually simple, yet largely unknown reaction
was systematically explored by us.19 Using the experi-
mentally convenient Staudinger conditions, we investi-
gated the reactivity of a range of acetals and ketenes.
Acyclic acetals underwent the reaction as well as 5- or 6-
membered cyclic ones. Reactive ketenes included di-
phenylketene, methylketene, monochloroketene and
ketene itself. It turned out that the main limiting factors
are the substituents at the acetal center: in accordance with
the mechanism put forward in Scheme 14, only aryl- or
doubly alkyl-substituted acetals giving rise to stabilized
carboxonium ions efficiently underwent the reaction.
Our experiments involving chiral acetals which were per-
formed to investigate the simple and induced diastereose-
lectivity of the reaction, are exemplified by Scheme 15.
Thus, exposure of 5α-androstanone ethyleneketal 51 to
dichloroketene and zinc chloride furnished a 2:1 mixture
of spiro-dioxepanones 52a,b in good overall yield. So far,
these efforts have culminated in the highly diastereoselec-
tive insertions of mono-substituted ketenes into the pseu-
do-C2 symmetrical dioxolane 53 which proceeded with
more than 90 % diastereomeric excess in all cases inves-
tigated.19 The predominant diastereomers, 54a-c, are dis-
tinguished by the pseudo equatorial position of all four
substituents. This energetically favorable situation is
probably reflected in the transition state leading to their
formation. Further investigations will show whether this
kind of chemistry can be developed into a useful asym-
Reagents and conditions: (a) (CH2OH)2/p-TsOH/benzene, 86%; (b)
Cl3COCl/Zn-Cu/Et2O, 54%
Scheme 13
regiochemistry would be governed by the very electron
rich aromatic ring. After successful cycloaddition and re-
ductive removal of the halogens, the hasubanan skeleton
would be formed by a Beckmann-rearrangement.
When 45 was exposed to dichloroketene, generated in situ
by reduction of trichloroacetyl chloride with zinc-copper
couple, a single product was isolated along with unreacted
starting material. The expected dichlorocyclobutanone 46
was immediately ruled out by IR spectroscopy and the
material was identified as the novel dichlorodioxepanone
47 which obviously arose from insertion of dichlo-
roketene into the dioxolane. The mechanism of this inser-
tionprobably involves the nucleophilic attack of an acetal-
oxygen on the ketene which is activated by ZnCl2
(Scheme 14). The resulting zwitterion 48 undergoes C,O-
bond cleavage and rearranges under C,C-bond formation
via transition state 49 to afford the dioxepanone (50).
Scheme 14
Scheme 15

SYNTHESIS
April 1998
661
slowly crystallized on standing. The material was recrystallized from
pentane/CHCl3 to afford crystals suitable for X-ray structure determi-
nation; Rf 0.58 (hexanes/EtOAc, 1:3); mp 165–167°C (subl.).
Anal. C16H16O3 (256.3): C, 74.98; H, 6.29; found C, 75.12; H, 6.32.
HRMS: m/z calcd for C16H16O3 256.1099, found 256.1071.
1H NMR (250 MHz, CDCl3): δ =1.31–1.52 (m, 1 H), 1.69–1.79 (m, 1
H), 1.86–1.96 (m, 1 H), 2.03–2.18 (m, 1 H), 2.37–2.51 (m, 4 H),
2.61–2.71 (dd, 1 H, J = 7.5, 17.5 Hz), 2.90 (br s, 0.3 H), 4.62 (dd, 1
H, J = 0.9, 17.0 Hz), 4.78 (s, 1 H), 5.20 (dd, 1 H, J = 0.9, 10.2 Hz),
6.07 (dd, 1 H, J = 10.2, 17.0 Hz), 6.20 (br s, 0.7 H), 6.60 (d, 1 H, J =
8.2 Hz), 6.77 (d, 1 H, J = 8.2 Hz).
metric variant of the aldol reaction which requires
straightforward removal of the chiral tether.
The ketene insertion is not limited to O,O-acetals
(Scheme 16). Reaction of dichloroketene with 2-phenyl-
1,3-dithiane (55) under the usual conditions furnished the
8-membered dithiocane 56 in excellent yield. Insertion of
diphenylketene into aminal 57 afforded the interesting
triphenyldiazapine 58. Finally, it proved possible to ex-
tend the reaction of non-acetalic substrates, provided they
meet the necessary structural requirements: a nucleophilic
heteroatom bound to a carbon which is sufficiently stabi-
lized towards positive charge. Thus, treatment of styrene
oxide (59) with dichloroketene gave rise to γ-butyrolac-
tone 60.20 The scope and limitations of this chemistry are
under active investigation in our laboratories and will be
reported in due course.
13C NMR (63 MHz, CDCl3): δ = 20.8, 23.2, 26.3, 36.5, 39.5, 55.0,
91.7, 117.0, 117.9, 121.4, 125.1, 128.3, 138.1, 142.2, 145.0, 208.3.
MS (FT-ICR, DIP-EI): m/z (%) = 256 (M+·), 241, 227, 213, 199, 185,
153, 127, 115, 73, 58, 43.
IR (KBr): ν = 3415, 2929, 1721, 1634, 1615, 1502, 1452, 1307 cm–1.
(3aR,9aR,9bS)-7-Chloro-9b-ethenyl-1,3a,8,9,9a,9b-hexahydro-5-
methoxyphenanthro[4,4a,4b,5-bcd]furan-3(2H)-one (12):
A solution of 7a (3.99 g, 10 mmol) in anhyd DMF (50 ml) was heated
to 140°C for 20 min. After cooling to r.t., the solvent was removed in
vacuo. The crystalline crude product was purified by chromatography
on silica gel, eluting with hexanes/EtOAc (1:1), to afford the dihydro-
furan 12 (2.99 g, 9.8 mmol, 98 %) as colorless crystals; Rf 0.25 (hex-
anes/EtOAc, 3:1); mp 147–148°C (racemic), 178–179°C (enantio-
merically pure); [α]D –35.7 (c = 2, CHCl3).
Anal. C17H17ClO3 (304.8): calcd C, 67.00; H, 5.62; found C, 66.92;
H, 5.55.
HRMS: m/z calcd for C17H17ClO3 304.08662, found: 304.08662 ±
0.0015.
1H NMR (250 MHz, CDCl3): δ =1.29 (dq, 1 H, J = 4.9, 12.6 Hz),
1.70–2.20 (m, 3 H), 2.23–2.45 (m, 4 H), 2.61 (ddd, 1 H, J = 0.9, 7.3,
18.0 Hz), 3.82 (s, 3 H), 4.52 (dd, 1 H, J = 0.7, 17.0 Hz), 4.71 (s, 1 H),
5.14 (dd, 1 H, J = 0.7, 10.2 Hz), 5.97 (dd, 1 H, J =10.2, 17.0 Hz), 6.71
(s, 1 H).
13C NMR (63 MHz, CDCl3): δ =19.9, 22.9, 26.5, 36.1, 39.4, 55.2,
56.6, 91.7, 114.7, 117.1, 123.9, 125.0, 126.8, 141.8, 142.9, 145.6,
206.5.
MS (EI, 70 eV, 30°C): m/z (%) = 306 (57), 304 (M+·, 99), 244 (87),
174 (76), 149 (28), 131 (41), 95 (33), 69 (100).
IR (Si): ν = 2938, 1729, 1631, 1491, 1436, 1186, 1147, 1096, 1037,
973, 873 cm–1.
(4aR*,10aR*)-8-Chloro-4a-ethenyl-4a,9,10,10a-tetrahydro-5,6-
dimethoxyphenanthren-3(4H)-one (17):
A solution of 7b (1 mmol) in DMF (20 mL) was heated to 140°C. As
the mixture remained unchanged over a 40 min period according to
TLC analysis, anhyd LiBr (1g, 11.5 mmol) was added. After 1 h, the
mixture was cooled to r.t. and the solvent was removed in vacuo.
Chromatography of the residue (hexanes/EtOAc, 5.1) afforded 17
(36 mg, 0.11 mmol, 11%) as colorless crystals; Rf 0.15 (hexanes/
EtOAc, 5:1); mp 149°C.
Scheme 16
Optical rotations were measured at 20°C.
(3aR,9aR*,9bS*)-9b-Ethenyl-1,3a,8,9,9a,9b-hexahydro-5-
hydroxyphenanthro[4,4a,4b,5-bcd]furan-3(2H)-one (10):
Anal. C18H19ClO3 (318.8): calcd C, 67.82; H, 6.01; found C, 67.84;
H, 6.05.
To a solution of 9 (100 mg, 0.28 mmol) in anhyd CH2Cl2 (3 mL) at
–78°C was added dropwise a solution of BBr3 in CH2Cl2 (1 M,
1.5 mL, 1.5 mmol). The mixture was warmed to r.t. within 1.5 h and
then poured into ice-water (25 mL). The aqueous phase was saturated
with NaCl and extracted with CH2Cl2 (5 × 10 mL). The combined or-
ganic phases were dried (MgSO4) and evaporated in vacuo. The re-
sultant residue was treated with Et3N (10 mL), whereupon a white
precipitate was immediately formed, and heated to reflux for 1 h. Sub-
sequently, the solvent was evaporated in vacuo and the residue was
partitioned between Et2O (10 mL) and aq HCl (2 N, 10 mL). The or-
ganic phase was separated and the aqueous phase was extracted with
Et2O (2 × 10 mL). The combined organic phases were dried (MgSO4)
and evaporated in vacuo. The residue was purified by HPLC [Mach-
erey-Nagel Nucleosil 50–10 (250 × 20), hexanes/EtOAc, 1:1, 2 mL/
min] to afford 10 (54 mg, 0.21 mmol, 75%) as pale brown oil which
1H NMR (400 MHz, CDCl3): δ = 1.74 (mc, 1 H), 2.04 (mc, 1 H), 2.60
(mc, 1 H), 2.72 (mc, 1 H), 2.85 (mc, 1 H), 3.02 (d, 1 H, J = 16.2 Hz),
3.13 (d, 1 H, J = 16.2 Hz), 3.74 (s, 3 H), 3.78 (s, 3 H), 4.81 (d, 1 H, J
= 17.2 Hz), 5.08 (d, 1 H, J = 10.3 Hz), 6.00 (d, 1 H, J = 10.1 Hz), 6.08
(dd, 1 H, J = 10.3, 17.2 Hz), 6.78 (dd, 1 H, J = 10.1, 4.4 Hz), 6.87 (s,
1 H).
13C NMR (100 MHz, CDCl3): δ = 23.8, 27.1, 43.4, 44.7, 46.2, 55.9,
60.3, 113.0, 113.9, 126.7, 128.1, 129.2, 134.7, 145.0, 151.2, 151.5,
147.2, 198.9.
MS (EI, 70 eV, 200°C): m/z (%) = 318 (M+·, 100), 287 (7), 275 (6),
249 (2), 237 (14), 211 (15), 195 (18), 165 (19), 152 (23), 139 (18), 115
(60), 91 (20), 81 (19), 65 (12), 55 (24), 53 (49).
IR (KBr): ν = 2938, 1675, 1588, 1467, 1425, 1291 cm–1.

SYNTHESIS
Papers
662
(3aR,9aR,9bS)-3-Ethylenedioxy-1,2,3,3a,9a,9b-hexahydro-5-
methoxy-9b-{2-[N-methyl-N-(phenylsulfonyl)amino]ethyl}phen-
anthro[4,4a,4b,5-bcd]furan (19):
A solution of 18 (486 mg, 1 mmol), NBS (187 mg, 1.05 mmol) and a
catalytic amount of benzoyl peroxide (5 mg, 0.02 mmol, recrystal-
MS (EI, 70 eV, 120°C): m/z (%) = 343 (M+·, 100), 328 (11), 286 (23),
284 (13), 244 (18), 198 (13), 168 (17), 115 (16), 99 (40), 59 (17).
IR (Si): ν = 2924, 1638, 1614, 1503, 1446, 1276, 1246, 1194, 1058,
958, 794.
Triazoline (43):
lized from CHCl3/MeOH) in anhyd CCl (150 mL) was kept at reflux
4
A solution of 42 (170 mg, 0.47 mmol) in anhyd benzene (25 mL) was
refluxed for 1 h. Subsequently, the solvent was removed in vacuo. Re-
crystallization of the resultant crude product from pentane afforded
triazoline 43 (130 mg, 0.36 mmol, 76 %) as colorless crystals; mp
160°C.
for 30 min. Subsequently, Et N (10 mL) was added and the mixture
3
was refluxed for 10 min. After cooling to r.t., the mixture was washed
with satd aq NaHCO3 solution (2 × 50 mL) and satd aq Na2S2O3 so-
lution (50 mL). The organic phase was dried (MgSO4) and concen-
trated in vacuo. The residue was purified by HPLC [Macherey-Nagel
Nucleosil 50–10 (250 × 20), hexane/MeOAc, 10:4, 10 mL/min;
Merck Si-60, hexane/MeOAc, 10:4, 2 mL/min] to afford the starting
sulfonamide 18 (102 mg, 0.21 mmol, 21%) and styrene 19 (324 mg,
0.67 mmol, 67%, 85% based on recovered starting material) as color-
less crystals; Rf 0.15 (hexanes/EtOAc, 5:1); mp 115–117°C (enantio-
merically pure); [α]D –24.1 (c = 1, CHCl3), [α]D –23.8 (c = 0.5, CHCl3).
Anal. C26H29NO6S (483.6): calcd C, 64.58; H, 6.04; N, 2.90; found C,
64.61; H, 6.08; N, 2.75.
1H NMR (250 MHz, CDCl3): δ = 1.87 (d, 1 H, J = 16.4 Hz), 1.95–1.98
(m, 1 H), 2.02–2.25 (m, 3 H), 2.54–2.78 (m, 2 H), 3.08–3.24 (m, 2 H),
3.21 (d, 1 H, J = 16.4 Hz), 3.81 (s, 3 H), 3.83–3.90 (m, 1 H), 3.86 (s,
3 H), 4.13 (dt, 1 H, J = 9.17, 11.9 Hz), 4.46 (dd, 1 H, J = 1.8, 6.4 Hz),
6.88 (s, 1 H).
13C NMR (63 MHz, CDCl3): δ = 25.2, 30.7, 37.3, 39.3, 44.6, 46.4,
47.3, 55.8, 60.6, 70.1, 80.4, 113.2, 125.4, 128.2, 133.5, 147.1, 151.8,
207.3.
MS (ESI): m/z (%) = 336 (35), 335 (22), 334 (M+·- N , 100), 282 (15),
HRMS: m/z calcd for C26H29NO6S 483.1715, found 483.1715 ±
0.0025 (± 5 ppm).
2
185 (6), 145 (6), 144 (7), 101 (6).
1H NMR (400 MHz, CDCl3): δ = 1.24 (dq, 1 H, J = 2.7, 12.8 Hz),
1.38–1.62 (m, 2 H), 1.66–1.93 (m, 3 H), 2.43 (mc, 1 H), 2.62 (s, 3 H),
2.86 (ddd, 1 H, J = 5.4, 10.7, 16.0 Hz), 3.18 (ddd, 1 H, J = 5.4, 10.7,
16.0 Hz), 3.66–3.90 (m, 2 H), 3.87 (s, 3 H), 4.00 (mc, 1 H), 4.20 (mc,
1 H), 4.69 (s, 1 H), 5.79 (dd, 1 H, J = 5.7, 9.6 Hz), 6.36 (dd, 1 H, J =
0.8, 9.6 Hz), 6.63 (d, 1 H, J = 8.1 Hz), 6.71 (d, 1 H, J = 8.1 Hz),
7.44–7.58 (m, 3 H), 7.67–7.72 (m, 2 H).
IR (film): ν = 2941, 1717, 1588, 1469, 1428, 1291, 1286, 1062,
1012 cm–1.
Hasubanan (44):
A solution of 43 (40 mg, 0.11 mmol) in anhyd pyridine (15 mL) was
refluxed for 5 h. The solvent was removed in vacuo and the resultant
residue was purified by HPLC [Merck CN, hexane/dioxane, 10:7] to
afford hasubanan 44 (26 mg, 0.08 mmol, 73%) as colorless crystals;
mp 135–137°C.
13C NMR (100 MHz, CDCl3): δ = 25.2, 31.3, 34.7, 36.5, 38.8, 45.3,
46.1, 56.3, 64.8, 66.4, 94.1, 108.0, 112.9, 117.6, 123.0, 123.1, 127.1,
128.9, 129.6, 132.3, 137.6, 143.8, 146.1.
1H NMR (600 MHz, CD3CN): δ = 1.76–1.92 (m, 3 H), 2.05–2.26 (m,
2 H), 2.57 (d, 1 H, J = 16.3 Hz), 2.73–2.78 (m, 3 H), 2.84–2.88 (m, 1
H), 3.17 (dd, 1 H, J = 0.7, 16.3 Hz), 3.80 (s, 3 H), 3.86 (s, 3 H), 5.90
(dd, 1 H, J = 0.9, 10.1 Hz), 6.58 (d, 1 H, J = 10.1 Hz), 6.99 (s, 1 H).
13C NMR (100 MHz, CDCl3): δ = 24.0, 27.9, 38.2, 43.5, 45.2, 49.3,
50.7, 61.1, 63.3, 113.4, 126.7, 128.3, 128.4, 138.8, 148.2, 152.6,
157.2, 199.8.
MS (EI, 70 eV, 80°C): m/z (%) = 483 (M+·, 44), 422 (5), 364 (4), 342
(79), 286 (74), 285 (53), 199 (67), 141 (28), 100 (27), 99 (100), 77
(58), 55 (32).
IR (KBr): ν = 2948, 1636, 1622, 1505, 1446, 1338, 1278, 1163, 1092,
691 cm–1.
MS (ESI): m/z (%) = 335 (23), 334 (15), 333 (M+·, 69), 291 (34), 290
(44), 289 (100), 288 (69), 276 (32), 274 (67), 260 (19), 258 (45), 254
(77), 239 (12), 45 (57).
(–)-Dihydrocodeinone Ethyleneketal (20):
To a mixture of liquid ammonia (20 mL) and anhyd THF (20 mL) at
–78°C was added t-BuOH (400 mL, 4.2 mmol) followed by lithium
metal (200 mg, 28.8 mmol). The resulting dark blue solution was
stirred for 10 min and then styrene 19 (200 mg, 0.41 mmol) in THF
(10 mL) was added dropwise. The blue color was rapidly discharched
and additional lithium (200 mg, 28.8 mmol) was added. After 30 min,
the reaction was quenched with a mixture of satd aq NH4Cl solution
(25 mL) and MeOH (25 mL). After dilution with EtOAc (50 mL) and
separation of the basic aqueous phase, the organic phase was washed
with satd NaHCO3 solution (10 mL) and brine (10 ml), dried
(MgSO4) and concentrated in vacuo. The residue was purified by
chromatography (silica gel, CHCl3/MeOH, 5:1) to give morphinane
IR (film): ν = 3321, 2934, 2871, 1676, 1589, 1467, 1427, 1291, 1268,
1238, 1213, 1122, 1050, 1008 cm–1.
Spirodioxepanones (52a,b):
To a mixture of 5α-androstane ethyleneketal 51 (2.15 g, 5 mmol) and
Zn/Cu-couple (1 g, 15 mmol) in refluxing anhyd Et2O (25 mL) was
added dropwise a solution of trichloroacetyl chloride (1.5 mL,
13 mmol) in anhyd Et2O (25 mL) during a 1 h period. The mixture was
refluxed for an additional hour and stirred at r.t. overnight. Subsequent-
ly, it was diluted with Et2O (50 mL), filtered and extracted with H2O
(2 × 50 mL), satd aq NaHCO3 solution (50 mL) and brine (2 × 50 mL).
The organic phase was dried (MgSO4) and evaporated. The resultant
residue was purified by preparative HPLC (Waters PAK 500, hexane/
MeOAc, 10:0.8, 100 mL/min) to afford 52a (1.23 g, 2.3 mmol, 45%)
and 52b (590 mg, 1.09 mmol, 22%) as colorless crystals.
20 (111 mg, 0.32 mmol, 79%) as colorless crystals; R 0.30 (CHCl3/
f
MeOH, 4:1); mp 173–175°C (enantiomerically pure); [α]D –173.3 (c
= 1, CHCl3).
Anal. C20H25NO4 (343.4): calcd C, 69.95; H, 7.34; N, 4.08; found C,
69.82; H, 7.32; N, 3.83.
HRMS: m/z calcd for C20H25NO4 343.1784, found 343.1784 ±
0.0017.
Major Isomer 52a: R 0.64 (hexanes/EtOAc, 8:1); mp 189–190°C;
f
[α]D +17.3 (c = 2, CHCl3).
1H NMR (400 MHz, CDCl3): δ = 1.16 (dq, 1 H, J = 2.3, 12.8 Hz),
1.50–1.56 (m, 2 H), 1.64–1.71 (m, 2 H), 1.88 (td, 1H, J = 4.9, 12.3
Hz), 2.19–2.24 (m, 2 H), 2.36 (dd, 1 H, J = 5.4, 18.3 Hz), 2.41 (s, 3
H), 2.53 (dd, 1 H, J = 3.8, 12.2 Hz), 3.00 (d, 1 H, J = 18.3 Hz), 3.11
(dd, 1 H, J = 2.6, 5.4 Hz), 3.79 (mc, 1 H), 3.87 (s, 3 H), 3.88 (mc, 1
H), 4.11 (mc, 1 H), 4.18 (mc, 1 H), 4.48 (s, 1 H), 6.62 (d, 1 H, J =
8.2 Hz), 6.74 (d, 1 H, J = 8.2 Hz).
HRMS: m/z calcd for C31H50Cl2O3 540.3137 (35Cl), found 540.3137
± 0.003.
1H NMR (250 MHz, CDCl3): δ = 0.65 (s, 3 H), 0.68–2.21 (m, 40 H),
0.81 (s, 3 H), 3.87–4.00 (m, 2 H), 4.56–4.70 (m, 2 H).
13C NMR (63 MHz, CDCl3): δ = 11.5, 12.0, 18.6, 20.9, 22.4, 22.7,
23.7, 24.0, 27.9, 28.1, 31.1, 33.1, 35.3, 35.4, 35.7, 36.0, 39.4, 39.8,
40.5, 42.5, 53.7, 56.2, 56.3, 62.0, 70.5, 79.8, 94.2, 163.4.
MS (EI, 70 eV, 190°C): m/z (%) = 540 (M+·, 10), 525 (7), 506 (6), 385
(15), 351 (11), 317 (4), 231 (6), 154 (100), 120 (26), 95 (23), 81 (28),
69 (35), 55 (62).
13C NMR (100 MHz, CDCl3): δ = 20.0, 22.2, 33.2, 36.3, 42.4, 42.7,
43.5, 47.0, 56.4, 59.5, 64.8, 66.3, 94.2, 108.4, 113.4, 118.5, 126.2,
129.0, 142.1, 146.4.

SYNTHESIS
3,3-Dichloro-4-phenyldihydrofuran-2-one (60):
April 1998
663
IR (Si): ν = 2933, 1742, 1455, 1374, 1306, 1211, 1093, 1067,
855 cm–1.
To a mixture of styrene oxide (1 g, 1 mL, 8.4 mmol) and Zn/Cu-cou-
ple (1 g, 15 mmol) in refluxing anhyd Et2O (25 mL) was added drop-
wise a solution of trichloroacetyl chloride (1.5 mL, 13 mmol) in
anhyd Et2O (25 mL) within 1 h. The mixture was further refluxed for
1 h and cooled to r.t. Subsequently, it was diluted with Et2O (50 mL),
filtered and extracted with H2O (50 mL), satd aq NaHCO3 solution
(50 mL) and with brine (2 × 50 mL). The organic phase was dried
(MgSO4) and evaporated. Purification of the residue by column chro-
matography (silica gel, hexanes/EtOAc, 40:1 → 20:1 → 10:1) afford-
Minor Isomer (52b): Rf 0.50 (hexanes/EtOAc, 8:1); [α]D +27.5 (c =
1, CHCl3).
HRMS: m/z calcd for C31H50Cl2O3 540.3137 (35Cl), found 540.3137
± 0.003.
1H NMR (250 MHz, CDCl3): δ = 0.65 (s, 3 H), 0.68–1.86 (m, 37 H),
0.84 (s, 3 H), 1.94 (mc, 1 H), 1.99 (mc, 1 H), 2.29 (mc, 1 H),
3.86–4.06 (m, 2 H), 4.53–4.70 (m, 2 H).
13C NMR (63 MHz, CDCl3): δ =11.9, 12.6, 18.5, 21.0, 22.4, 22.7,
23.7, 24.0, 27.8, 28.1, 28.4, 31.4, 34.3, 35.1, 35.4, 35.6, 36.0, 36.4,
39.3, 39.8, 41.9, 42.4, 53.9, 56.1, 56.2, 62.3, 70.1, 78.0, 95.5, 163.6.
MS (EI, 70 eV, 190°C): m/z (%) = 540 (M+·, 30), 525 (20), 506 (24),
385 (30), 351 (28), 317 (14), 231 (4), 154 (100), 120 (49), 95 (33), 81
(33), 69 (31), 55 (48).
ed 60 (1.6 g, 6.9 mmol, 82%) as colorless crystals; R 0.82 (hexanes/
f
EtOAc); mp 45°C.
HRMS: m/z calcd for C10H8Cl2O2 229.9901 (35Cl), found 229.9900 ±
0.002.
1H NMR (270 MHz, CDCl3): δ = 4.61–4.75 (m, 2 H), 5.17 (dd, 1 H,
J = 6.2, 7.4 Hz), 7.35–7.46 (m, 5 H).
IR (Si): ν = 2931, 1743, 1450, 1373, 1302, 1211, 1093, 1066,
13C NMR (68 MHz, CDCl3): δ = 58.5, 71.4, 89.4, 127.5, 129.0, 129.4,
136.5, 161.5 cm–1.
855 cm–1.
MS (EI, 70 eV, 30°C): m/z (%) = 201 (3), 138 (96), 125 (85), 103
(100), 89 (17), 77 (36).
3,3-Dichloro-4-phenyl-1,5-dithiocan-2-one (56):
IR (film): ν = 3425, 1770, 1651, 1494, 1455, 1247, 1216, 908,
To a mixture of 2-phenyl-1,3-dithiane 160 (982 mg, 5 mmol) and Zn/
Cu-couple (1 g, 15 mmol) in refluxing anhyd Et2O (25 mL) was added
dropwise a solution of trichloroacetyl chloride (1.5 mL, 13 mmol) in
anhyd Et2O (25 mL) within 1 h. The mixture was refluxed for 2 h and
cooled to r.t. Subsequently, it was diluted with Et2O (50 mL), filtered
and extracted with H2O (2 × 50 mL), satd aq NaHCO3 solution
(50 mL) and brine (2 × 50 mL). The organic phase was dried
(MgSO4) and evaporated. Purification of the residue by column chro-
matography (silica gel, hexanes/EtOAc, 10:1) afforded 56 (1.46 g,
733 cm–1.
(1) For recent reviews of the total synthesis of morphine alkaloids,
see:
(a) Hudlicky, T.; Butora, G.; Fearnley, S.P.; Gum, A.G.; Stabile,
M.R. In Studies in Natural Products Chemistry; Ed. Atte-Ur-
Rahmen; Elsevier: Amsterdam, 1996, Vol. 18, p 43.
(b) Maier, M. In Organic Synthesis Highlights II, Waldmann,
H., Ed.; VCH: Weinheim, 1995; p 357.
4.8 mmol, 95%) as soft, colorless crystals; R 0.65 (hexanes/EtOAc,
f
10:1); mp 89–90°C.
HRMS: m/z calcd for C12H12Cl2OS2 305.9707 (35Cl), found 305.9710
± 0.0015.
c) Szántay, G.; Dörnyei, G. In The Alkaloids; Cordell, G.A.;
Brossi, A., Eds.; Academic: New York; Vol. 45, p 127.
(2) For reviews, see:
1H NMR (250 MHz, CDCl3): δ = 2.01–2.18 (m, 1 H), 2.30–2.42 (m,
1 H), 2.61 (dd, 1 H, J = 10.2, 15.2 Hz), 2.99–3.24 (m, 2 H), 3.26 (ddd,
1 H, J = 2.0, 6.0, 14.2 Hz), 4.91 (s, 1 H), 7.34–7.45 (m, 5 H).
13C NMR (63 MHz, CDCl3): δ = 29.9, 31.6, 33.9, 66.7, 127.7, 128.3,
128.6, 130.4, 135.9, 192.0.
(a) Matsui, M. In The Alkaloids; Brossi, A., Ed.; Academic:
New York, 1988; Vol. 33, p 307.
(b) Anubishi, Y.; Ibuka, T. In The Alkaloids; Marske, R. F. H.,
Ed.; Academic: New York, 1977; Vol. 16, p 393.
(3) Trauner, D. Diplomarbeit, Freie Universität Berlin, 1994.
(4) Ginsburg, D.; Elad, D. J. Am. Chem. Soc. 1954, 76, 312;
J. Chem. Soc. 1954, 3052. For a recent synthesis along similar
lines, see:
MS (EI, 70 eV, 30°C): m/z (%) = 306 (M+·, 19), 243 (6), 196 (36), 172
(11), 106 (100), 74 (21).
IR (Si): ν = 3652, 3030, 1716, 1683, 1492, 1450, 1420, 1274, 1057 cm–1.
White, J.D.; Hrnciar, P.; Stappenbeck, F. J. Org. Chem. 1997,
62, 5250.
N,N' -Dimethyl-6,6,7-triphenyl-1,4-diazepan-5-one (58):
To a suspension of Zn/Cu-couple (1 g, 15 mmol) in anhyd Et2O
(20 mL) was added dropwise a solution of α-chloro-α,α-diphenyl-
acetyl chloride (2.65 g, 10 mmol) in Et2O (10 mL). After stirring for
1 h at r.t., a solution of aminal 57 (5 mmol) in anhyd Et2O (10 mL)
was added dropwise while the mixture was gradually heated to reflux.
The mixture was then refluxed for 1 h, cooled to r.t., diluted with Et2O
(50 mL), filtered and washed with satd aq NaHCO3 solution
(2 × 50 ml) and brine (50 mL). The organic phase was dried (MgSO4)
and evaporated. Purification of the residue by column chromatogra-
phy (silica gel, hexanes/EtOAc, 10:1) afforded triphenyldiazepanone
58 (1.17 g, 3.2 mmol, 64%) as large, colorless crystals; Rf 0.20 (hex-
anes/EtOAc, 10:1); mp 164°C.
(5) Mulzer, J.; Dürner, G.; Trauner, D. Angew. Chem. 1996, 108,
3046; Angew. Chem., Int. Ed. Engl. 1996, 35, 2830.
(6) Details of the crystal structure determinations of compounds 5,
6, 7a, 8, 10, 17, 18, 19, 20, 24, 36, 37, 54a,b and 58 may be ob-
tained from: The Director, Cambridge Crystallographic Data
Centre, 12 Union Road, Cambridge CB2 1 EZ, UK.
(7) Eliel, E.L.; Wilen, S.H. Stereochemistry of Organic Com-
pounds; Wiley-Interscience: New York, 1995; pp 731 ff.
(8) (a) Kawabata, T.K.; Grieco, P.A.; Sham, H.L.; Kim, H.; Jaw,
J.Y.; Tu, S. J. Org. Chem. 1987, 52, 3347.
(b) Grieco, P.A.; Sham, H.L.; Inanaga, J.; Kim, H.; Tuthill, P.A.
J. Chem. Soc., Chem. Commun. 1984, 1345.
HRMS: m/z calcd for C25H26N2O 370.2045, found 370.2045 ± 0.002.
1H NMR (250, MHz, CDCl3): δ = 2.12 (s, 3 H), 2.40 (dd, 1 H, J = 4.4,
13.2 Hz), 2.91–3.06 (m, 2 H), 3.22 (m, 3 H), 3.41 (mc, 1 H), 4.19 (s,
1 H), 6.45–6.50 (m, 2 H), 6.78–7.02 (m, 5 H), 7.20–7.50 (m, 8 H).
13C NMR (63 MHz, CDCl3): δ = 39.1, 44.3, 49.2, 50.5, 64.9, 65.7,
125.5, 125.6, 126.8, 127.2, 127.8, 127.9, 129.6, 130.7, 131.1, 134.3,
140.3, 144.4, 173.7.
(9) Woodward, R.B. Pure Appl. Chem. 1968, 17, 519.
(10) Kewitz, H.; Wilcock, G.; Davies, B. In Alzheimer Disease: The-
rapeutic Strategies; Giacobini, E.; Becker, R., Eds.; Birkhäuser:
Boston 1994; pp 140–144.
(11) Parker, K.A.; Fokas, D. J. Am. Chem. Soc. 1992, 114, 9688 (Re-
ference 18).
(12) Mulzer, J.; Bats, J.W.; List, B.; Opatz, T.; Trauner, D. Synlett
1997, 441.
(13) Labidalle, S.; Min, Z.Y.; Reynet, A.; Moskowitz, H.; Vierfond,
J.-M.; Miocque, M. Tetrahedron 1988, 44, 1171.
(14) Itô, Y.; Saegusa T. J. Org. Chem. 1978, 43, 1011.
MS (EI, 70 eV, 80°C): m/z (%) = 370 (M+·, 24), 251 (76), 194 (21),
165 (100), 132 (85), 91 (73), 51 (43).
IR (Si): ν = 3055, 2928, 1637, 1492, 1444, 1391, 1340, 1216, 1139,
1104, 976 cm–1.

SYNTHESIS
Papers
664
(15) (a) Wong, H.; Belleau, I.; Pachter, Y.; Perron, Y. Can. J. Chem.
1975, 53, 2515.
(17) For a brief discussion of protecting groups as not-so-innocent
bystanders, see: Kocienski. P.J. Protecting Groups; Thieme:
Stuttgart, 1994; pp 13 ff.
(b) ibid. 1976, 54, 883.
(c) Benhamou, M.C.; Speziale, V.; Lattes, A.; Cros, J. Eur. J.
Med. Chem. 1981, 16, 263.
(d) Monkovic, I.; Wong, H. Can. J. Chem. 1976, 54, 883.
(16) For the use of intramolecular azide cycloadditions in the total
synthesis of natural products, see:
(18) For a review of ketene cycloadditions, see:
Hyatt, J.; Raynolds, P.W. Org. React. 1994, 45, 159.
(19 Mulzer, J.; Trauner, D.; Bats, J.W. Angew. Chem. 1996, 108,
2093; Angew. Chem., Int. Ed. Engl. 1996, 35, 1970.
(20) The reaction of diphenyl ketene with epoxides was reported to
afford the corresponding ketene acetals:
Molander, G.; Hiersemann, M. Tetrahedron Lett. 1997, 38,
4347, and references cited therein.
Gulbins, K.; Hamann, K. Chem. Ber. 1961, 94, 328.