Total synthesis of meloscine by a [2+2]-photocycloaddition/ring-expansion route

Article abstract of DOI:10.1002/chem.200802383

The unusual monoterpenoid indole alkaloid meloscine was synthesized starting from a protected aminoethylquinolone in 15 steps and an overall yield of 9%, employing a [2+2]-photocycloaddition as the stereochemistry defining key step. After the initial plan of a Wagner-Meerwein type rearrangement of a [4.2.0]- into a [3.3.0]-bicyclic substructure could not be realized, the required ring enlargement of a cyclobutane was eventually achieved by a retro-benzilic acid rearrangement. Generation of the central pyrrolidine ring was possible by a three-step reductive amination domino sequence. The final ring was built up by a ring-closing metathesis after the last quaternary stereocenter had been constructed by a Johnson-Claisen rearrangement. The synthesis was concluded by a selenylation-elimination sequence to build up the exocyclic vinyl group of meloscine. Using our methodology for enantioselective [2+2]-photocycloaddition mediated by a chiral complexation agent, the experimentally very simple synthesis could be performed in an enantioselective fashion (7% overall yield). The enantioselective synthesis of (+)-meloscine represents the first example of a natural product synthesis employing an enantioselective [2+2]-photocycloaddition as its key step, and illustrates nicely the synthetic potential of photochemical transformations for the construction of complex heterocyclic structures.

Full text of DOI:10.1002/chem.200802383

FULL PAPER
DOI: 10.1002/chem.200802383
Total Synthesis of Meloscine by a [2+2]-Photocycloaddition/Ring-Expansion
Route
Philipp Selig,[a] Eberhardt Herdtweck,[b] and Thorsten Bach*[a]
Abstract: The unusual monoterpenoid
indole alkaloid meloscine was synthe-
sized starting from a protected amino-
ethylquinolone in 15 steps and an over-
all yield of 9%, employing a [2+2]-
photocycloaddition as the stereochem-
istry defining key step. After the initial
plan of a Wagner–Meerwein type rear-
rangement of a [4.2.0]- into a [3.3.0]-bi-
cyclic substructure could not be realiz-
ed, the required ring enlargement of a
cyclobutane was eventually achieved
by a retro-benzilic acid rearrangement.
Generation of the central pyrrolidine
ring was possible by a three-step reduc-
tive amination domino sequence. The
final ring was built up by a ring-closing
metathesis after the last quaternary ste-
reocenter had been constructed by a
Johnson–Claisen rearrangement. The
synthesis was concluded by a selenyla-
tion–elimination sequence to build up
the exocyclic vinyl group of meloscine.
Using our methodology for enantiose-
lective [2+2]-photocycloaddition medi-
ated by a chiral complexation agent,
the experimentally very simple synthe-
sis could be performed in an enantiose-
lective fashion (7% overall yield). The
enantioselective synthesis of (+)-melo-
scine represents the first example of a
natural product synthesis employing an
enantioselective [2+2]-photocycloaddi-
tion as its key step, and illustrates
nicely the synthetic potential of photo-
chemical transformations for the con-
struction of complex heterocyclic struc-
tures.
Keywords: asymmetric synthesis ·
natural products · photochemistry ·
rearrangement · total synthesis
Introduction
relationship, which most probably also represents the bioge-
netic origin of the Melodinus alkaloids is maybe best illus-
trated by comparison of (+)-scandine [(+)-2] with its indole
counterpart (ꢀ)-tabersonine [(ꢀ)-3], an alkaloid also pres-
ent in various Melodinus species (Figure 1).
The pentacyclic alkaloid (+)-meloscine [(+)-1, Figure 1]
represents the prototype of a small group of monoterpenoid
indole alkaloids commonly referred to as Melodinus alka-
loids or meloquinolines due to their exclusive occurrence in
the Apocynacea species Melodinus ssp.[1] As their distinct
structural feature the Melodinus alkaloids incorporate a six-
membered quinolone ring within a monoterpenoid Aspido-
sperma carbon skeleton, which is derived from the typical
structure of related indole alkaloids by the expansion of ring
B with a concomitant contraction of ring C. This structural
Figure 1. Structures of the Melodinus alkaloids (+)-meloscine [(+)-1],
(+)-scandine [(+)-2], and the closely related Aspidosperma alkaloid
(ꢀ)-tabersonine [(ꢀ)-3].
[a] Dr. P. Selig, Prof. Dr. T. Bach
Lehrstuhl fꢀr Organische Chemie I
Technische Universitꢁt Mꢀnchen, Lichtenbergstr. 4
85747 Garching (Germany)
Fax : (+49)89-289-13315
E-mail: thorsten.bach@ch.tum.de
Biosynthetically, all 14 meloquinolines known to date are
likely to be derived from the parent compound (+)-scandine
[(+)-2], which is the rearrangement product of 18,19-didehy-
drotabersonine, the compound thought to represent the
common origin of all the Melodinus alkaloids on the basis
of structural and chemotaxonomical data.[2] While the dis-
covery and structure elucidation of meloscine (1) and other
[b] Dr. E. Herdtweck
Lehrstuhl fꢀr Anorganische Chemie
Technische Universitꢁt Mꢀnchen, Lichtenbergstr. 4
85747 Garching (Germany)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200802383.
Chem. Eur. J. 2009, 15, 3509 – 3525
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3509

meloquinolines from the species Melodinus scandens Forst.
already took place in the late 1960s,[3] no data on any kind
of biological activity has been published ever since, except
for the observation that certain Melodinus species find use
in traditional chinese folk medicine as a treatment for men-
ingitis and rheumatic heart disease.[4] In contrast to the two
other main groups of B-ring expanded monoterpenoid
indole alkaloids, that is, the Cinchona and the Campthoteca
alkaloids, which are both well renowned for their pharmaco-
logical uses, the Melodinus alkaloids seem hitherto devoid
of any useful biological activity, maybe due to the fact that
the incorporated lactam moiety strongly impairs with the
passage of melodan structures through biological mem-
branes.
ent power of the [2+2]-photocycloaddition to build up mul-
tiply substituted carbon centers.[14] Moreover, given the pos-
sibility of the [2+2]-photocycloaddition being conducted in
an enantioselective fashion,[15] this photochemical key step
should also give rise to a first enantioselective total synthe-
sis. Our complete retrosynthetic analysis of (+)-meloscine
[(+)-1] based on these consideration is outlined in
Scheme 1. As in the following discussions, the absolute con-
figuration of all products and intermediates is initially omit-
ted for the sake of clarity, and descriptors (+)/(ꢀ) will be in-
troduced only at the stage of the enantioselective synthesis
(see below). Accordingly, substances were initially prepared
as racemates and will be discussed as such unless otherwise
indicated.
Despite the present lack of applications, Melodinus alka-
loids are attractive targets for chemical synthesis, given both
their structural uniqueness in the world of alkaloids and
their complex, highly substituted central cyclopentane ring
C as a challenge for synthetic organic chemistry. Two bio-
mimetic syntheses of the melodan skeleton were reported
almost simultaneously in 1984, both starting from deriva-
tives of (ꢀ)-tabersonine [(ꢀ)-3] or its dihydro-derivative
(ꢀ)-vincadifformine. Rearrangement of the B,C-ring system
was achieved by either a-ketol rearrangement,[5] or flow
thermolysis of an intermediate aziridine.[6] The latter
method was also applicable to the synthesis of natural (+)-
meloscine [(+)-1] and (+)-scandine [(+)-2] from 18,19-dide-
hydrotabersonine.[7] A different entry into the melodan skel-
eton was described starting from the alkaloid leuconolam by
an intramolecular Michael addition.[8] Besides these semi-
synthetic approaches to the melodan skeleton, only a few
total-synthetic studies have been published to date. The only
successful completion of the synthesis of racemic meloscine
(1) was reported by Overman et al. in 1989.[9] From a syn-
thetic point of view, the construction of the two quaternary
carbon centers C5 and C12 can be considered the pivotal
steps of any synthesis of the melodan skeleton.[10] In Over-
manꢃs synthesis, C12 was constructed as part of the tricyclic
core C,D,E by the powerful aza-Cope rearrangement, Man-
nich cyclization strategy, while C5 was already established at
an early stage of the synthesis by a ZnBr2-promoted silyl
enol ether alkylation. A different approach to the core
structure of racemic scandine (2) was reported recently by
Denmark et al., which relies on an intramolecular Heck cyc-
lization for the formation of carbon center C12.[11] However,
no substituent was introduced at C5, thus the synthesis re-
mained limited to the pentacyclic melodan core structure.
Finally, a single approach to an asymmetric synthesis of Mel-
odinus alkaloids has been reported, but this sequence re-
mained limited to the tricyclic core C,D,E.[12]
Scheme 1. Retrosynthetic analysis of (+)-meloscine [(+)-1].
As the final step of our synthetic plan, the ring closure of
the tetrahydropyridine E ring of meloscine would occur as
an intramolecular Heck reaction[16] after removal of the
benzyl (Bn) protective group (PG) and appropriate N-alkyl-
ation of the tetracyclic precursor 4 (Scheme 1). It is impor-
tant to note, that after b-hydride elimination this Heck reac-
tion would provide the vinyl side chain of meloscine in a
straightforward fashion, thus rendering its separate construc-
tion by elimination[9b] unnecessary.
Precursor 4 was in turn envisioned to be the direct prod-
uct of a Wagner–Meerwein-like 1,2-rearrangement[17] of the
iminium ion 5 followed by an exocyclic elimination, which
was already shown to take place selectively in earlier test
systems.[18] This rearrangement would convert the [4.2.0]-bi-
cyclic substructure incorporating the cyclobutane motif ac-
cessible from quinolone photochemistry into the [3.3.0]-bicy-
clic structure representing the rings C and D of meloscine in
one single step. Iminium ion 5 would be derived from hemi-
aminal 6 by elimination of the hydroxy group, and rear-
rangement precursor 6 should result from reduction and
ring closure of the [2+2]-photocycloaddition (PCA) product
7 of 4-aminoethylquinolone 8 and an a-ethyl acrylate. The
early intermediate 7 would already provide all the atoms of
the tetracyclic A,B,C,D substructure of meloscine as well as
In a fundamentally new approach to the synthesis of Mel-
odinus alkaloids, we planned to construct the quaternary
stereogenic center C12 by a [2+2]-photocycloaddition reac-
tion to the c-bond of a 4-substituted 2(1H)quinolone repre-
senting the lower two rings A and B of meloscine (1). Thus,
we wanted to employ both the well-known suitability of qui-
nolones for photochemical reactions[13] as well as the inher-
3510
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3509 – 3525

Total Synthesis of Meloscine
FULL PAPER
the C2 side chain to be converted into the final C5 vinyl sub-
stituent. Moreover, with the stereochemistry at the quater-
nary carbon center C8b established beforehand, all follow-
ing reactions should occur in a substrate-controlled, diaste-
reoselective fashion. It turned out that, indeed, the stereo-
chemical course of the whole synthesis could be controlled
by this single quaternary center, while the retrosynthetic
plan depicted in Scheme 1 had to be altered several times as
the synthesis progressed. Details of our synthetic endeavor
which eventually culminated in the first enantioselective
total synthesis of (+)-meloscine [(+)-1] are given in this full
account.[19]
actions were not optimized, but in later studies AlH3 was
eventually identified as a reductive agent even more selec-
tive for the reduction of the quinolone derived anilide.[22]
For the successful reductive formation of the required lactol
from a lactam precursor to take place, however, the pres-
ence of an N-acyl protective group like the tert-butyloxycar-
bonyl (Boc) group instead of the N-Bn protective group as
in lactam 9 seemed a necessary prerequisite.[23] Unfortunate-
ly, we were unable to perform the necessary protective
group interconversions, because lactam 9 was either unreac-
tive towards most N-Bn deprotection methods,[24] or showed
a reductive fragmentation of the cyclobutane ring[25] under
dissolved metal conditions.[26] As the desired reduction
turned out to be impossible with the tetracyclic lactam 9, we
investigated the possibility of establishing the correct oxida-
tion state (+I) of the rearrangement precursor before the
ring closure to the [4.2.0]-bicyclic system (Scheme 3).
Results and Discussion
Studies towards iminium ion rearrangement: The photo-
chemistry of protected aminoethylquinolone 8 and the syn-
thesis of 1-alkyl substituted cyclobutanes by [2+2]-photocy-
cloaddition reactions have already been investigated exten-
sively.[20] Indeed, the envisioned cyclobutane 7 was readily
accessible in both racemic as well as enantiomerically en-
riched form by the [2+2]-photocycloaddition of quinolone 8
and methyl a-ethylacrylate, although the diastereomeric
ratio with respect to C1 [d.r.
G
Scheme 3. Failed approaches to access the open-chain w-aminoaldehyde
13.
difficulties encountered in the separation of diastereoiso-
mers somehow limited the yield of the required 1-exo
isomer. Thus, more easily accessible 1-methyl substituted cy-
clobutanes were chosen as initial test substrates for the con-
struction of rearrangement precursors, while they still re-
tained the pivotal quaternary center at C1.
To our surprise, the up-to-now straightforward N-Boc de-
protection of cyclobutane photocycloaddition products [10
vol% trifluoroacetic acid (TFA) in CH2Cl2] failed in the
case of aldehyde 12,[20] and resulted exclusively in the frag-
The most straightforward entry to hemiaminals or lactols
such as 6 would certainly be the direct reduction of a corre-
sponding lactam. While in most cases the reduction of N-
alkyl lactams preferably yields the fully deoxygenated amine
products, several examples have been reported resulting in
lactols or the corresponding open-chain w-aminoalde-
hydes.[21] Lithium aluminum hydride (LAH) reduction of
lactam 9,[20] however, only resulted in moderate yields of the
two fully reduced amines 10 and 11, and no hint of the for-
mation of an intermediate lactol could be found (Scheme 2).
mentation of the cyclobutane unit.[25] Less acidic conditions
[27]
or milder deprotection agents, for example, ZnBr2
or
TMS-I[28] did not give any conversion. As an access to ami-
noaldehyde 13 or its ring-closed hemiaminal form remained
impossible also by the attempted reduction of free N-Boc
deprotected aminoesters, the oxidation (ox.) of free amino-
alcohol 14 was investigated as a final possibility. While it
was eventually possible to gain access to an O-ethylated
hemiaminal derivative by oxidation of 14 with o-iodoxyben-
zoic acid (IBX) in DMSO, this transformation turned out to
be hardly reproducible and the product suffered from both
low yields and purity. Moreover, the obtained N,O-acetal
showed no tendency towards the planned formation of a N-
Bn iminium ion such as 5, and no trace of any rearranged
product could be detected either.[25] Given the overall dis-
couraging results obtained in our approaches to a-alkyl sub-
stituted rearrangement precursors, the inaccessibility of N-
acyl substituted precursor structures and the seemingly very
low tendency of N-alkyl iminium ions towards rearrange-
ment, we finally abandoned our original synthetic plan. The
possibility of an oxidation[29] of tetracyclic piperidines, which
were also accessible by means of a [2+2]-photocycloaddi-
tion,[30] was left uninvestigated. Instead, we turned our focus
from a-alkyl to a-hydroxy substituted lactams giving rise to
Scheme 2. Reduction of tetracyclic lactam 9 with LAH (Et2O).
While both products 10 and 11 were of no use for our syn-
thetic plan towards meloscine, it is interesting to note, that
the lower, secondary lactam was unexpectedly more reactive
towards LAH reduction than the upper, tertiary lactam. Re-
Chem. Eur. J. 2009, 15, 3509 – 3525
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3511

T. Bach et al.
the envisioned 1,2-rearrangement as an a-ketol rather than
a Wagner–Meerwein-type rearrangement.
diol derivative of 14, attempts to realize an a-hydroxy assist-
ed iminium ion rearrangement were not undertaken. In-
stead, a different kind of rearrangement was discovered in
the fully protected cyclobutane 16.
Studies towards a-ketol rearrangement: The construction of
a tetracyclic lactam bearing a hydroxy instead of an alkyl
substituent at the a-position required the [2+2]-photocy-
cloaddition of an a-oxy substituted acrylic acid derivative.
As there was no precedence of any such [2+2]-photocy-
cloaddition, it came as a pleasant surprise that enol ether
15[31] derived from methyl pyruvate proved to be an excel-
lent substrate for the planned transformation.[32] Irradiation
of a toluene solution of quinolone 8 and an excess of enol
ether 15 with UV light (l=350 nm) gave the desired cis-
fused cyclobutane 16 as a single diastereoisomer in almost
quantitative yield. Obviously, the a-silyloxy substituent of
acrylate 15 and the resulting captodative electronic situation
at C-a exhibit a favorable effect on the course of the [2+2]-
photocycloaddition, as both yield and stereoselectivity sur-
passed the values obtained with any a-alkyl substituted acry-
late investigated earlier.[20] Using established conditions, cy-
clobutane 16 was readily converted into the a-hydroxy
lactam 17 by N-Boc deprotection and thermal cyclization of
the free aminoester in refluxing toluene. Thermal conditions
simultaneously affected the cleavage of the O-TMS (trime-
thylsilyl) protective group (Scheme 4).
Retro-benzilic acid rearrangement and reductive amination:
The rearrangement was discovered when the basic condi-
tions previously applied to lactam 17 were used for the reac-
tion of cyclobutane 16 itself. To our amazement, the yet im-
possible ring expansion from a cyclobutane to a cyclopen-
tane was readily realized by simply stirring 16 in methanolic
K2CO3 for several hours at room temperature. After O-
TMS cleavage was accomplished within a few minutes, the
free a-hydroxy ester 20 smoothly underwent a retro-benzilic
acid like rearrangement[33] to give a cyclopentane-1,2-dike-
tone which exclusively existed in its tautomeric enol form 21
(Scheme 5).
Scheme 5. Retro-benzilic acid rearrangement of cyclobutane 16.
While no attempts were made to isolate the free alcohol
20, its formation and thus the stepwise course of the reac-
tion was clearly evident from TLC and NMR analysis of the
reaction mixture. Ketone 21 could be isolated as a single
product in excellent yield and purity after simple hydrolytic
work-up of the reaction mixture. The fact, that the retro-
benzilic acid rearrangement of cyclobutane 20 took place
under very mild conditions in excellent yields, while the ini-
tially sought-after rearrangement of any closed [4.2.0]-bicy-
clic system had been impossible even under drastic condi-
tions is maybe best explained by stereoelectronic arguments.
Thus, the imine position in closed [4.2.0]-bicycles such as 5
or 17 is sterically fixed in a presumably unreactive position,
in which the migration of the tertiary benzylic center would
have to occur into the node-plane of the imine p* orbital.[17]
In contrast, cyclobutane 20 can access reactive conforma-
tions with the ester carbonyl group being positioned verti-
cally to the migrating bond by free rotation, thus enabling
perfect overlap of the migrating s-bond and the ester p* or-
bital, the target of migration (Figure 2). Regardless of any
enthalpic or entropic considerations, this observation con-
vinced us to abandon any further attempts of iminium ion
or a-ketol rearrangements.
Scheme 4. Synthesis of a-hydroxylactam 17 and the failed plan to access
intermediate 4 with an a-ketol rearrangement.
With lactam 17 in hand, we tried to carry out the skeletal
rearrangement to hemiaminal 18 under both base and acid
catalyzed conditions. Reduction of 18 to its corresponding
amine 19 and introduction of the C2 side chain by carbonyl
olefination should then give access to intermediate
4
(Scheme 1). However, the rearranged product 18 remained
unaccessible as lactam 17 showed no conversion under all
conditions investigated. It seems reasonable to assume, that
the enthalpic gain from the relief of ring strain plus the for-
mation of one carbonyl group is by far insufficient to make
up for the energetic loss connected with the sacrifice of the
six-membered lactam ring. Although it was possible to con-
vert cyclobutane 16 into an aminoalcohol, representing a
Enone 21 was configurationally stable under basic condi-
tions although the high acidity of bridgehead proton H3a
would render epimerization theoretically possible. This was
1
evident from H NMR analysis of 21 in a CD3OD solution,
which indicated a quick H,D-interchange at C3a but no al-
3512
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3509 – 3525

Total Synthesis of Meloscine
FULL PAPER
mation of open-chain cyclopentane-1,2-diols, clearly being
the product of an opened 1,2-dione form of 18. In order to
overcome the problems associated with the 1,2-dione struc-
ture, rearrangement product 21 was chemically fixed in its
enol form by acetylation. Indeed, the obtained enol acetate
22 showed a profoundly different reactivity as compared to
the free enol 21 after N-Boc deprotection and basic work-
up. No 1,2-addition to the carbonyl group could by detected
at all, instead, the free amine rapidly underwent a 1,4-addi-
tion to the b-position at C3, resulting in the [3.2.1]-bicyclic
substructure of product 23 (Scheme 6).
Figure 2. Stereochemical conditions in the rearrangement precursors 5,
17, and 20.
Interestingly, piperidine 23 was obtained as a single dia-
stereoisomer with the shown exo-configuration at C15. At-
tempted purification by silica gel chromatography led to
widespread decomposition as 23 was unstable under acidic
conditions. Accordingly, the formation of 23 was completely
suppressed as long as the nucleophilic amine was kept pro-
tonated, that is, no basic work-up of the N-Boc deprotection
mixture was undertaken. When this acidic mixture was sub-
jected directly to reductive amination conditions using hy-
drogen gas and Pearlmanꢃs catalyst [Pd(OH)2/C] in MeOH,
a clean reaction took place, and after renewed N-Boc pro-
tection of the crude product mixture a single pyrrolidine
product 28 could be isolated in 78% yield over a total of
five synthetic transformations (Scheme 7).
teration of any other proton signal. The established cis anne-
lation of rings B and C was kept intact by the quaternary
benzylic center even though the adjacent position C3a was
strongly prone to deprotonation. Following N-Boc deprotec-
tion and basic work-up, the free aminoketone derived from
21 instantaneously cyclized to give the hemiaminal 18 in
Scheme 6. Further conversions of rearrangement product 21.
quantitative yield (Scheme 6). Using this preparatively
simple three-step sequence of [2+2]-photocycloaddition,
retro-benzilic acid rearrangement, and deprotection/ring clo-
sure we were able to access the tetracyclic product 18 in
95% overall yield on a multi-gram scale with no chromato-
graphic purification step being necessary. Following this first
breakthrough, extensive efforts were made to further con-
vert hemiaminal 18 into its corresponding deoxygenated
amine 19 (Scheme 4). However, none of the applied reduc-
tive conditions, for example, by employing Na(CN)BH3, Na-
Scheme 7. Formation of pyrrolidine 28 under hydrogenation conditions.
(Only the free amines are shown for the sake of clarity.)
A
tions with Pd or Pt catalysts, gave a defined product. If any
conversion of starting material was detectable at all, it had
generally resulted in a rapid decomposition giving a com-
plex mixture unsuitable even for rough, preliminary analysis.
Attempts to achieve carbonyl olefination prior to reduction
were equally unsuccessful, as were any attempts of reduc-
tion or olefination with the N-Boc protected open-chain pre-
cursor 21. In summary, the immediate rearrangement prod-
ucts 21 and 18 turned out to be dead ends, probably due to
the presence of the latent 1,2-dione moiety. Although hemi-
aminal 18 was one single, stable compound in NMR meas-
urements, it seems likely that open-chain and/or tautomeric
aminoketone forms played a decisive role under reductive
conditions. This was underlined by the observation, that the
NaBH4 reduction of hemiaminal 18 only resulted in the for-
The course of the reaction as it is shown in Scheme 7 was
elucidated by means of 1H NMR analysis of the reaction
mixture. Thus, benzylamine 24 was first hydrogenated at its
enol double bond to give cyclopentanone 25 in a diastereo-
selective fashion. This hydrogenation proceeded very quick-
ly and was basically finished after less than 10 minutes. The
diastereoselectivity of this hydrogenation showed a remark-
able strong dependence on the reaction temperature. While
at <158C 25, and thus the final product 28, was formed as
the C2 exo-isomer exclusively, temperatures of >258C led
to the formation of a 50:50 mixture of C2 diastereoisomers.
Thus, the reaction mixture was best cooled with an ice-bath
before initiating the reductive sequence. Subsequent N-de-
Chem. Eur. J. 2009, 15, 3509 – 3525
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3513

T. Bach et al.
benzylation to the free primary amine 26 probably repre-
sents the rate-limiting step of the overall sequence and re-
quired reaction times of more than 12 h. Attempts to speed
up the debenzylation step by heating to 508C quickly led to
widespread decomposition. Intramolecular reductive amina-
tion giving the free pyrrolidine 27 concluded the reductive
sequence. N-Boc protection to 28 finally allowed for an easy
chromatographic purification of the product. The sequence
outlined in Scheme 7 was conveniently performed as a one-
pot procedure by simply diluting the TFA-acidic mixture of
24 with MeOH, adding the catalyst and stirring under a H2
atmosphere over night. With 28 in hand, we finally had
access to an intermediate displaying both the [3.3.0]-bicyclic
substructure and the correct oxidation state at the nitrogen-
adjacent carbon center. What was left to do, was the addi-
tion of the missing C2 unit by a carbonyl olefination. Thus,
acetate 28 was readily saponified to its free alcohol 29 and
oxidized to give ketone 30 (Scheme 8). At this stage, the rel-
Studies towards intramolecular ring closure: N-Boc depro-
tection of 31/32 and subsequent N-alkylation allowed for the
straightforward synthesis of a number of precursors 33–36
for the final intramolecular ring closure in good yields
(Figure 3).
Figure 3. Precursors 33–36 for an intramolecular ring closure.
Especially the Z-iodopropenes 33/34, which were avail-
able from literature-known (Z)-3-bromo-1-iodopropene[37]
and which already incorporated the desired double-bond
configuration of the melodan skeleton, were investigated in-
tensively under a variety of cyclization conditions. A promis-
ing option seemed to be the metalation of 33 with subse-
quent Michael addition of the alkenyl metal species to the
a,b-unsaturated ester moiety. Much to our surprise, howev-
er, ester 33 was basically inert to many metalating agents
like Mg or Li metal, iPrMgBr, MeLi, nBuLi or even tBuLi.
In fact, the iodovinyl group of substrate 33 remained com-
pletely untouched upon the addition of up to three equiva-
lents of tBuLi at room temperature and even greater excess-
es of reagent or heating only led to an unspecific decompo-
sition. It seems that due to the poor reactivity of the vinyl
iodide, basic organometallic reagents led to an efficient,
multiple deprotonation of the substrate rather than giving
the desired halogen-metal exchange. Only under very drastic
conditions (>5 equiv tBuLi at room temperature) some
traces of proto-deiodination product could be observed.
Large excesses of SmI2 (THF)[38] resulted in the reduction of
the a,b-unsaturated ester double bond, but again the vinyl
iodide remained intact. Consequently, any attempted trans-
metallation reactions to form a soft nucleophile suitable for
addition to the a,b-unsaturated ester remained futile.[39] At-
tempted radical cyclizations of 33/34 were equally unsuc-
cessful[40] and exclusively resulted in unspecific decomposi-
tion. Finally, the attempted Heck-type ring closure[41] of 33/
34 failed as either no oxidative addition of Pd could be real-
ized, or a rapid N-deallylation occurred if more drastic con-
ditions (T > 1008C) were applied. The same problem was
encountered in preliminary studies concerning an ene-yne-
cyclization of propargylamine 35 with Au, Ag or Pt salts. Fi-
nally, acetal 36 was planned to serve as a precursor for a
SmI2 induced carbonyl radical cyclization,[42] yet this ap-
proach failed due to the extreme instability of the inter-
mediate free b-aminoaldehyde.[43] In summary, no intramo-
lecular ring closure could by realized under any conditions
in any of the precursors 33–36. However, an unusual photo-
Scheme 8. Formation and carbonyl olefination of N-Boc protected
ketone 30.
ative exo-configuration of the hydroxy group at C2 was cru-
cial, as the diastereomeric alcohol failed to yield the desired
ketone and decomposed into a complex, dark-red mixture.
N-Boc protected ketone 30 could not be converted into its
free amine or any other N-alkylated products due to the
high intramolecular nucleophilicity of the pyrrolidine nitro-
gen, which led to the formation of an unstable aziridine-
aminal. However, carbonyl olefination using the stabilized
reagent Ph3P=CHCOOEt in refluxing THF went very
smoothly, giving the a,b-unsaturated ester 31 as a single dia-
stereoisomer, which could be converted into the allylic alco-
hol 32 with DIBAL-H in CH2Cl2 at ꢀ458C.
In sharp contrast to the Wittig reaction giving ester 31,
any attempts to achieve a carbonyl olefination resulting in a
plain alkyl, for example, a methyl substituent remained
fruitless. Simple Wittig salts or phosphonic acid esters did
not lead to any conversion of ketone 30, presumably due to
a competing deprotonation of the adjacent methine proton
and the formation of an unreactive enolate. Less acidic re-
agents like cyclic phosphonamides[35] or a dimetallic chromi-
um-samarium reagent[36] were also ineffective. In the end,
we decided to rely on a,b-unsaturated ester 31 and allylic al-
cohol 32 as precursors for the final ring closure although
this implied sacrificing the Heck-promoted formation of the
vinyl side-chain of meloscine as planned earlier (Scheme 1).
3514
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3509 – 3525

Total Synthesis of Meloscine
FULL PAPER
chemical oxidation was discovered when the radical-cycliza-
tion of 33 was attempted under UV irradiation (Scheme 9).
Claisen rearrangement and ring-closing metathesis: Due to
the continuous failure of the attempted intramolecular clo-
sure of the final tetrahydropyridine E ring of meloscine, we
decided to decouple the two synthetic steps of ring closure
and formation of the quaternary carbon center and to build
up the final ring of meloscine by a combination of Claisen
rearrangement[45] and ring-closing metathesis.[46] Indeed, the
final quaternary carbon center could be established without
problems by performing a thermal Johnson–Claisen rear-
rangement[47] on the N-Boc protected allylic alcohol 32.
Using triethylorthoacetate or trimethylorthoacetate with hy-
droquinone as acidic catalyst, alcohol 32 could be converted
into the corresponding rearrangement products 38/39, both
of which were formed as mixtures in almost identical diaste-
reomeric ratios of d.r. ꢁ 70:30 (Scheme 10).
Scheme 9. Photochemical oxidation of (Z)-3-iodoallylamines to b-amino-
acroleines.
The formation of b-aminoacroleine 37 was in fact most ef-
ficient when using air as an oxidant, sunlight as the UV
source and NEt3 as a scavenger for liberated iodide, thus
representing a nice example of green photochemistry.[44] The
reaction was also applicable to simpler iodoallylamines as
shown for the diethylamine analogue (Scheme 9). However,
the synthetic value of this newly found photooxidation re-
mains limited, as the b-aminoacroleine products are also
available from the simple 1,4-addition of amines to propiol-
aldehyde. In spite of its continuous failure in all ring-closing
attempts, ester 33 still proved to be valuable, as it spontane-
ously crystallized from a CDCl3 solution to form big, off-
white plates suitable for single crystal X-ray structure analy-
sis (Figure 4, see also Experimental Section and Supporting
Information).
Scheme 10. Claisen rearrangement of 32 and N-allylation of the product
mixture.
Separation of the diastereoisomeric mixtures was conven-
iently achieved after the subsequent step of N-Boc deprotec-
tion and N-allylation during which the minor isomer quanti-
tatively cyclized to give the pentacyclic lactam 40 (>90%
rel. to d.r.). Allylamines 41 and 42 were obtained in 58 and
65% yield, respectively, and were thus clearly derived from
the major diastereoisomers. The configuration assignment of
1
the obtained products was initially based on H NOE meas-
urements, and was confirmed later on in the course of the
synthesis (Scheme 11). In order to obtain satisfactory yields
in the rearrangement step, however, some optimization
work turned out to be necessary, as standard conditions, that
is, refluxing in 7–10 equivalents of CH3CACTHNUGTRENNG(U OEt)3 at 1458C,
only provided mediocre yields of 30–40% of product 38.
Moreover, no full conversion of the starting alcohol could
be achieved even under prolonged heating and with exces-
sive amounts of acid catalysts, and the reaction mixture
always contained significant amounts (20–30%) of the sub-
strate acetate arising from hydrolysis of the intermediate or-
thoester. Yields were slightly better when using CH3C-
Figure 4. Structure of compound 33 in the crystal.
This obtained structure served as an unambiguous proof
for the stereochemical assignments made previously on the
1
basis of H NOE measurements, but it also indicated an un-
favorable steric proximity of the vinyl moiety and the acidic
proton at C11a.
ACHTUNGRTEN(NUNG OMe)3, but still no more than 45% of 39 could be isolated
even when a complete conversion of the substrate was
forced by further heating to 1508C. Microwave assisted ther-
molysis[48] and other catalytic methods[49] to overcome these
Chem. Eur. J. 2009, 15, 3509 – 3525
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3515

T. Bach et al.
common problems of the Johnson–Claisen rearrangement
seemed quite incompatible with the N-Boc pyrrolidine struc-
ture of the substrate. High yields could eventually be realiz-
ed when the reaction mixture was concentrated into a thick
gum with c(32) ꢁ 2–3 molLꢀ1 after the initial formation of
a homogeneous mixture. Using these high concentration
conditions a total of 85% of 39 was obtained after only a
few hours while keeping the temperature below 1358C. The
diastereomeric ratio (d.r. ꢁ 70:30) with respect to C1 re-
mained completely unchanged under all conditions, and no
difference was observed when using the different orthoace-
tate reagents either. Thus, no attempts were made to im-
prove the diastereomeric ratio by variation of the orthoester
substituent.[50] Fortunately, the required configuration at C1
was realized in the major diastereoisomer, and satisfactory
yields of 41/42 could be obtained. Ring closing metathesis
went smoothly for both allylamines 41/42 using Grubbs’ 2nd
generation catalyst[51] in toluene and provided the ester 43/
44, establishing the complete pentacyclic carbon skeleton of
meloscine (Scheme 11). From this point onwards, only the
methyl carboxylate 44 was used for further transformations.
times, high concentrations and the repeated addition of sele-
nide anion solution in order to achieve full conversion. Oxi-
dation of selenide 48 with meta-chloroperbenzoic acid
(MCPBA) at ꢀ788C and thermal decomposition of the sele-
nium oxide finally gave the desired product meloscine (1) in
86% yield, provided that the basic nitrogen atom had been
protonated previously by
(Scheme 12).
a
slight excess of TFA
Scheme 12. Completion of the synthesis of racemic meloscine (1).
Spectroscopic data of racemic meloscine (1) (m.p. 216–
2208C)[9a] were in all aspects identical to the enantiopure
natural product data.[3,5,55] Interestingly, when the sequence
depicted in Scheme 12 was conducted according to literature
precedence[9b] without the prior addition of TFA, yields of
isolated meloscine (1) never exceeded 20–30%. Using one
equivalent of MCPBA, only very little conversion of the sel-
enide was achieved, and more than two equivalents of
MCPBA lead to widespread decomposition of the product.
Amazingly, occasionally present residual contaminants of to-
sylate 47 could always be recovered quantitatively. Thus, the
observed decomposition does not seem to result from an ox-
idation of the basic nitrogen atom, but rather from its reac-
tion with the intermediate selenium oxide. Also, the quality
of NaBH4 was of utmost importance. When aged, partially
hydrolyzed NaBH4 was employed, reduction of the aromatic
nitro group giving aniline 49 was a significant side reaction.
Another difficulty arose from the formation of very stable
aminoborane adducts at the bridgehead nitrogen atom, a
problem which had already been encountered in the NaBH4
reduction to alcohol 46 (Scheme 11). The exact nature of
the NaBH4 contaminants leading to aminoborane formation
remained unclear, but the identical adducts were obtained
by the treatment of pentacyclic melodan products with
BH3·THF, as was shown for ester 44 (Figure 5). Both side
reactions could be completely suppressed, however, when
using fresh NaBH4 under anhydrous conditions. Starting
from quinolone 8, racemic meloscine (1) was thus available
in 15 linear steps and 9% overall yield. Moreover, the
newly established synthesis exclusively consists of experi-
mentally simple, robust and reliably reproducible reactions.
With the exception of NaBH4 and DIBAL-H reductions, an-
hydrous conditions were not necessary and none of the re-
agents or intermediates needed to be freshly prepared. In a
single experiment, alcohol 46 was converted into the hexacy-
Scheme 11. Construction of the melodan skeleton by ring-closing meta-
thesis and further conversions of the ester side-chain.
What remained to be done was the conversion of the
ester group of 44 into the vinyl moiety of meloscine (1). Re-
duction with DIBAL-H in CH2Cl2 gave aldehyde 45 in good
yields and subsequent NaBH4 reduction provided the known
alcohol 46 already used by Overman et al.[9] Purification of
46 was somehow problematic, however, due to its very high
polarity and unfavorable chromatographic behavior on silica
gel. Thus, purification of intermediates 45 and 46 was best
omitted and the crude product was directly converted into
tosylate 47 well suitable for chromatographic purification.
Several attempts were made to achieve a direct elimination
of the tosylate,[52] but no conversion could be achieved.
The good availability of aldehyde 45 led us to investigate
a possible reduction via its enol triflate.[53] While the latter
could indeed be synthesized as a mixture of E and Z iso-
mers, no subsequent oxidative addition of Pd could be real-
ized, and we finally had to fall back to the selenium oxide
elimination[54] already established in Overmanꢃs synthesis.
Tosylate 47 was thus converted to the o-nitrophenylselenide
48 by treatment with an excess (20 equiv) of o-nitrophenyl
selenide anion obtained from o-nitrophenyl selenocyanate
and NaBH4 in EtOH. This reaction required long reaction
clic ether 50 by oxymercuration with Hg
ACHTUNGRTEN(UNNG OOCCF3)2 and re-
ductive workup with NaBH4 according to Overmanꢃs syn-
3516
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3509 – 3525

Total Synthesis of Meloscine
FULL PAPER
thesis of deoxoapodine (51)[9b] (Figure 5). However, yields
were low (<15%) and 50 could only be isolated as a crude
product with about 80% purity according to NMR analysis.
amount of obtainable photoproduct, as due to the overall
smaller spectral overlap only low concentration solutions of
quinolone 8 (c=5–10 mm) could be converted within a rea-
sonable amount of time. Long irradiation times of more
than 10 h necessary for highly concentrated substrate solu-
tions or large volume irradiation vessels resulted in the for-
mation of by-products and a drop of yield to 70–75% of
(ꢀ)-16. The enantiomeric excess of (ꢀ)-16 could be further
improved by either recrystallization of the racemate from n-
hexane or by semipreparative HPLC separation (ChiralPak
AD). While racemate recrystallization involved the sacrifice
of 10% of the desired enantiomer, HPLC separation of the
enantiomers was basically quantitative, thus giving a total of
76% of enantiomerically pure product (ꢀ)-16 (>99% ee,
HPLC). With (ꢀ)-16 in hand, the previously established se-
quence of rearrangement (Scheme 5), reductive amination
(Scheme 7), Wittig reaction (Scheme 8), Claisen rearrange-
ment (Scheme 10), ring-closing metathesis (Scheme 11) and
construction of the vinyl group (Scheme 11,12) could be per-
formed without any experimental changes. No trace of race-
mization was observed at any step (chiral HPLC, see Sup-
porting Information), and all transformations gave yields
very close to those obtained in the racemic reactions, despite
some slight differences in the extraction and chromato-
graphic behaviors. Finally, the enantiomerically pure (+)-
meloscine [(+)-1] (m.p. 172–1768C) obtained from (ꢀ)-16
well matched with the natural product with regard to both
optical rotation and melting point.[3a,b] The latter is signifi-
cantly lower than for the racemic product probably due to
sterically impaired intermolecular hydrogen bonding.
Figure 5. Aminoborane adduct of ester 44 and compound 50 closely re-
sembling the aspidosperma alkaloid deoxoapodine (51).
Enantioselective total synthesis of (+)-meloscine [(+)-1]:
For the established synthesis of racemic meloscine (1) to be
conducted in an enantioselective fashion, the initial key step
of the [2+2]-photocycloaddition was performed in the pres-
ence of the chiral complexing agent (ꢀ)-52.[56] Using low-
temperature conditions and an irradiation wavelength of l=
370 nm (Rayonet RPR-3500 ꢄ lamps), the photocycloaddi-
tion product (ꢀ)-16 was obtained with an enantiomeric
excess (ee) of 79% ee in 87% yield (Scheme 13).
Conclusion
In summary, we have developed a new and efficient synthe-
sis of the unusual monoterpenoid indole alkaloid meloscine
(1) in 15 steps and 7–9% overall yield. Although our origi-
nal, very short synthetic plan could not be realized and sev-
eral additional, partially unprecedented reactions had to be
included, the newly established sequence gives for the first
time smooth access to the pentacyclic melodan carbon skel-
eton on a gram-scale. We especially want to emphasize the
experimental simplicity and robustness of the individual
steps. The established reaction sequence was easily scaled
up without any experimental changes to give 1.4 grams of
intermediate 44 in a single run starting from quinolone 8
without further optimization. Using the chiral complexing
agent (ꢀ)-52 already used for a variety of enantioselective
transformations in our group,[15,20,57] the synthesis could also
be conducted in an enantioselective fashion, thus being the
first natural product synthesis employing an enantioselective
[2+2]-photocycloaddition as its key step.[58] In addition, the
newly found combination of [2+2]-photocycloaddition and
retro-benzilic acid rearrangement may serve as an attractive
new possibility for the construction of annelated cyclopenta-
1,2-diones and as a starting point for various further reac-
tions also described in this paper.
Scheme 13. Photochemical key step of the enantioselective total synthe-
sis.
The obtained degree of enantioselectivity was in good ac-
cordance with the results found earlier in enantioselective
[2+2]-photocycloaddition reactions of quinolone 8 and a-
alkyl substituted acrylates.[20] Also, the enantiomeric excess
was completely unaffected by variations in substrate concen-
trations. Considering that only 5.0 mol% of complexing
agent (ꢀ)-52 were actually lost in this photocycloaddition
reaction, the reaction may reasonably be considered as
“quasi-catalytic” with respect to (ꢀ)-52. In contrast to the
corresponding racemic reaction (Scheme 4), a slightly longer
irradiation wavelength was necessary to achieve these high
recovery rates of the complexing agent (ꢀ)-52. With l=
350 nm (LZC-UVA lamps) only 60–70% of (ꢀ)-52 could be
re-isolated, although the complexing agent shows no signifi-
cant UV absorption at l > 320 nm. Unfortunately, the re-
quired longer irradiation wavelength somehow limited the
Chem. Eur. J. 2009, 15, 3509 – 3525
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3517

T. Bach et al.
(101 mg, 183 mmol, 87%) was isolated with 79% ee. 98% of the complex-
ing agent were recovered after the reaction. The enantiomeric excess of
(ꢀ)-16 could be further improved by either recrystallization of the race-
mate from n-hexane or by semipreparative HPLC (Chiralpak AD, 250ꢅ
20 mm I.D., 10 mm, n-hexane/iPrOH 90:10, 19 mLminꢀ1). Using the latter
method, enantiopure cyclobutane (ꢀ)-16 (>99% ee) was obtained as a
white foam in 76% overall yield. [a]2D0 =ꢀ14.7 (c=1.0 in CH2Cl2); chiral
HPLC (Chiralpak AD, n-hexane/iPrOH 90:10): tR =9.6 min, k=1.7; rac:
tR =10.6/15.8 min, k=1.9/3.3, a=1.7.
Experimental Section
General: All commercially available chemicals were used as received
without further purification. Reactions involving the water-sensitive
chemicals DIBAL-H or NaBH4 were performed in dried glassware under
argon using anhydrous solvents. Photochemical transformation were per-
formed in Duran-glass tubes (d=1.0 cm) in a merry-go-round apparatus
using 16 fluorescent lamps LZC-UVA (l=350 nm) or RPR-3500 ꢄ (l=
370 nm) at 358C. TLC: Merck glass sheets (0.25 mm silica gel 60, F254),
eluent given in brackets; detection by UV. Flash chromatography was
performed on silica gel 60 (Merck, 230–400 mesh) (dimensions of col-
umns given as [diameter]ꢅ[height]) with the indicated eluent. Common
solvents for chromatography [pentane (P), EtOAc, CH2Cl2, MeOH] were
(ꢂ)-(1aS,7bS,11aR)-10-Benzyl-11a-hydroxy-1,9,10,11a-tetrahydro-1aH-
pyrido[4’,3’:2,3]cyclobutaACTHNUTRGENN[GU 1,2-c]quinoline-2,11HCATUNGTREN(NGUN 3H,8H)dione (17): Cyclo-
butane photoproduct 16 (110 mg, 199 mmol) was stirred in 10 vol% TFA
in CH2Cl2 (10 mL) for 1.0 h at RT. Sat. aqueous NaHCO3 (15 mL) was
added and the mixture was extracted with CH2Cl2 (3ꢅ10 mL). The com-
bined organic extracts were dried over Na2SO4, filtered and concentrated
to yield the N-Boc deprotected aminoester (90 mg, quant) as a yellow
foam which was used without further purification. This aminoester was
refluxed in PhMe (10 mL) for 18 h. After evaporation of the solvent and
purification of the residue by flash chromatography (2.0ꢅ5 cm, CH2Cl2/
MeOH 20:1), hydroxylactam 17 (54 mg, 155 mmol, 78%) was obtained as
a white, crystalline solid. Rf =0.20 (EtOAc, [UV]); m.p. 195–1978C;
1H NMR (360 MHz, CDCl3): d=8.93 (br, 1H, NH), 7.42–7.29 (m, 5H,
Har), 7.25–7.17 (m, 2H, Har), 7.06 (virt. t, 3J=7.4 Hz, 1H, H6), 6.81 (d,
3J=7.9 Hz, 1H, H4), 4.79 (d, 2J=14.5 Hz, 1H, CHHPh), 4.68 (d, 2J=
14.5 Hz, 1H, CHHPh), 4.33 (br, 1H, OH), 3.68 (ddd, 2J=13.3, 3J=12.9,
3J=3.0 Hz, 1H, CHHN), 3.37 (ddd, 2J=13.3, 3J=5.3, 3J=2.0 Hz, 1H,
CHHN), 3.08 (dd, 3J=10.2, 3J=8.8 Hz, 1H, C1aH), 2.88 (dd, 2J=12.8,
3J=10.2 Hz, 1H, CHHexo), 2.58 (dd, 2J=12.8, 3J=8.8 Hz, 1H, CHendoH),
2.18 (ddd, 2J=13.8, 3J=12.9, 3J=5.3 Hz, 1H, RCHH), 1.93 ppm (ddd,
2J=13.8, 3J=3.0, 3J=2.0 Hz, 1H, RCHH); 13C NMR (90 MHz, CDCl3):
d=172.9 (s, CO), 169.9 (s, CO), 136.2 (s, C1Bn/C3a), 136.1 (s, C1Bn/C3a),
129.5 (d, CarH), 129.0 (d, CBnH), 128.7 (d, CarH), 128.1 (d, CarH), 128.0
(d, CBnH), 123.2 (d, C6H), 120.5 (s, C7a), 116.1 (d, C4H), 72.9 (s, C11aOH),
51.6 (t, CH2Ph), 49.3 (s, C7b), 43.2 (t, CH2N), 40.5 (t, C1HH), 32.8 (d,
C1aH), 32.4 ppm (t, RCH2); IR (KBr): n˜ =3205 (s, br, OH), 3060 (w,
CarH), 2918 (w, CalH), 1666 (vs, CONH), 1633 (vs, CONBn), 1594(s),
1493 (s), 1405 (m), 1194 (m), 755 (m), 701 cmꢀ1 (m); MS (70 eV): m/z
(%): 348 (2) [M+], 320 (91) [MꢀCO+], 265 (30), 159 (29) [RCH3+], 120
(19) [BnNCH2+], 91 (100) [C7H7+], 40 (37); HRMS (70 eV): m/z: calcd
for C20H20N2O2: 320.1525 [MꢀCO+], found 320.1527.
1
distilled prior to use. H and 13C NMR spectra were recorded in the indi-
cated solvent at ambient temperature unless otherwise indicated. Chemi-
cal shifts are reported relative to solvent residue signals as internal stan-
dard. Apparent multiplets, which occur as a result of the accidental
equality of coupling constants of magnetically non-equivalent protons are
marked as virtual (virt.). The multiplicities of the 13C NMR signals were
determined by DEPT experiments. Signals in the 13C NMR spectra which
may be interchanged are marked with an asterisk (*). The preparation
and analytical data of compounds 10, 11, 14, 34–38, 41, 43, 49, 50 and the
aminoborane adduct of compound 44 are reported in the Supporting In-
formation. 13C NMR spectra of all new compounds, chiral HPLC data of
(ꢀ)-16 and selected intermediates, full spectroscopic data of (+)-melo-
scine [(+)-1] and emission spectra of the fluorescent lamps used in this
work are also presented in the Supporting Information.
(ꢀ)-(1R,2aS,8bR)-8b-{2-[Benzyl(tert-butoxycarbonyl)amino]ethyl}-3-oxo-
1-[(trimethylsilyl)oxy]-1,2,2a,3,4,8b-hexahydrocyclobuta[c]quinoline-1-
methyl carboxylate (16)
Racemic [2+2]-photocycloaddition: A solution of quinolone 8 (6.81 g,
18.0 mmol) and enol ether 15 (17.0 mL, 15.1 g, 90.0 mmol, 5 equiv) in tol-
uene (582 mL) was irradiated with LZC-UVA lamps (l=350 nm) at
358C in three batches (200 mL each) for 3.0 h. After evaporation of the
solvent, the residue was recrystallized from petroleum ether (80–1108C)
to give the desired product as an off-white crystalline solid. The remain-
ing mother liquid was concentrated and subjected to flash chromatogra-
phy (5.0ꢅ30 cm, P/EtOAc 4:1!1:1) to give some additional product as a
white foam. In total, cyclobutane 16 (9.72 g, 17.6 mmol, 98%) could be
isolated in nearly quantitative yield. Rf =0.78 (EtOAc, [UV]); m.p. 128–
1308C; 1H NMR (500 MHz, [D6]DMSO, 808C): d=9.77 (br, 1H, NH),
7.29–7.20 (m, 3H, HBn), 7.16 (virt. t, 3J=8.1 Hz, 1H, H6), 7.05 (d, 3J=
7.1 Hz, 2H, HBn), 7.02 (d, 3J=7.6 Hz, 1H, H8), 6.96 (virt. t, 3J=7.3 Hz,
(+)-N-2-[(3aS,9bR)-2-Hydroxy-1,4-dioxo-1,3a,4,5-tetrahydro-9bH-cyclo-
penta[c]quinoline-9-yl]ethyl-N-benzyl-tert-butylcarbamate (21): Cyclobu-
tane photoproduct 16 (5.53 g, 10.0 mmol) was dissolved in MeOH
(60 mL), 0.5 gK2CO3 added, and the mixture stirred at RT for 6.0 h. Sat.
aqueous NH4Cl (250 mL) was added, the MeOH evaporated and the
aqueous residue extracted with EtOAc (1ꢅ100 mL + 2ꢅ50 mL). The
combined organic phases were washed with H2O (100 mL) and brine
(100 mL), dried over Na2SO4, filtered and concentrated to give enone 21
(4.38 g, 9.77 mmol, 98%) as a light yellow foam. Additional purification
by flash chromatography was ineffective. Rf =0.58 (EtOAc, [UV]);
1H NMR (360 MHz, CDCl3, 608C): d=9.35 (s, 1H, NH), 7.62 (d, 3J=
7.8 Hz, 1H, H9), 7.30–7.19 (m, 4H, Har), 7.12–7.05 (m, 3H, Har), 6.82 (d,
3J=7.9 Hz, 1H, H5), 6.63 (d, 3J=3.1 Hz, 1H, H3), 4.34 (d, 2J=15.2 Hz,
1H, CHHPh), 4.27 (d, 2J=15.2 Hz, 1H, CHHPh), 3.68 (brm, 1H, Hal),
3.20 (brm, 1H, Hal), 2.87 (virt. t, Jt =10.6 Hz, 1H, Hal), 2.33 (virt. dt, Jt =
12.2, Jd =4.3 Hz, 1H, Hal), 1.84 (brm, 1H, Hal), 1.45 ppm [s, 9H, C-
3
2
1H, H7), 6.87 (d, J=7.9 Hz, 1H, H5), 4.23 (d, J=15.4 Hz, 1H, CHHPh),
4.17 (d, 2J=15.4 Hz, 1H, CHHPh), 3.73 (s, 3H, OCH3), 2.97 (dd, 2J=
11.3, 3J=9.3 Hz, 1H, CHHexo), 2.91 (ddd, 2J=13.8, 3J=11.9, 3J=4.3 Hz,
1H, CHHN), 2.79 (dd, 3J=9.3, 3J=9.2 Hz, 1H, C2aH), 2.53 (ddd, 2J=
13.8, 3J=11.0, 3J=4.7 Hz, 1H, CHHN), 2.02 (dd, 2J=11.3, 3J=9.2 Hz,
1H, CHendoH), 1.90 (ddd, 2J=12.1, 3J=11.9, 3J=4.7 Hz, 1H, RCHH),
1.60 (ddd, 2J=12.1, 3J=11.0, 3J=4.3 Hz, 1H, RCHH), 1.38 [s, 9H, OC-
(CH3)3], ꢀ0.16 ppm [s, 9H, Si
(CH3)3]; 13C NMR (90 MHz, CDCl3): d=
ACHTUNGTRENNUNG
173.1 (s, COOMe), 170.3 (s, CONH), 155.5 (s, COBoc), 138.1 (s, Car), 136.9
(s, Car), 131.2 (d, C8H), 128.4 (d, CarH), 128.2 (d, CarH), 127.9 (d, CarH),
127.2 (d, CarH), 122.7 (d, C7H), 119.5 (s, Car), 115.7 (d, C5H), 80.2 (s, C1),
79.8 [s, C
CH2N), 37.3 (t, CHH), 36.9 (t, RCH2), 36.8 (d, COCH), 28.3 [q, C-
(CH3)3], 0.7 ppm [q, Si(CH3)3]; IR (KBr): n˜ =3212 (w, br, CONH), 3065
ACHTUNGTRENNUNG(CH3)3], 53.4 (s, C8b), 52.1 (q, OCH3), 50.4 (t, CH2Ph), 42.5 (t,
AHCTUNGTRENNUNG
(CH3)3]; 13C NMR (90 MHz, CDCl3, 608C): d=202.8 (s, CO), 169.0 (s,
E
ACHTUNGTRENNUNG
(w, CarH), 2975 (m, CalH), 1720 (s, COOMe), 1692 (vs, CON), 1594 (s),
1251 (m), 1172 (m), 843 (m), 754 (m), 700 cmꢀ1 (m); MS (70 eV): m/z
(%): 552 (1) [M+], 378 (31), 322 (18), 220 (46) [BocNBnCH2+], 159
(100) [RCH3+], 120 (80) [BnNCH2+], 91 (38) [C7H7+], 57 (54) [tBu+];
HRMS (70 eV): m/z: calcd for C30H40N2O6Si: 552.2656 [M+], found
552.2658; elemental analysis calcd (%) for C30H40N2O6Si: C 65.19, H
7.29; found: C 65.05, H 7.34.
CONH), 155.8 (s, COBoc), 153.1 (s, C2OH), 138.3 (s, Car), 135.6 (s, Car),
129.1 (d, CarH), 128.9 (d, CarH), 128.3 (d, CarH), 128.1 (d, CarH), 127.7 (d,
CarH), 124.6 (d, C8H), 119.2 (s, C9a), 116.2 (d, C6H), 80.7 [s, C
51.1 (brt, CH2Ph), 50.7 (s, C9b), 47.7 (d, C3aH), 43.2 (brt, CH2N), 38.5
(brt, RCH2), 28.8 ppm [q, C(CH3)3]; (d, C3H) is not visible due to super-
ACHTUNGTRENNUNG(CH3)3],
AHCTUNGTRENNUNG
imposition; IR (KBr): n˜ =3243 (s, br, NH), 2976 (m, CalH), 2926 (w,
CalH), 1673 (vs, CO), 1593 (m), 1494 (m), 1245 (w), 1163 (s), 756 (m),
700 cmꢀ1 (m, CHBn); MS (70 eV): m/z (%): 448 (2) [M+], 492 (18)
[MꢀtBu+], 348 (100) [MꢀBoc+], 214 (23), 120 (49) [BnNCH2+], 107
(36), 91 (73) [C7H7+], 57 (77) [tBu+]; HRMS (70 eV): m/z: calcd for
C26H28N2O5: 448.1998 [M+], found 448.1985.
Enantioselective [2+2]-photocycloaddition: According to the above
procedure, but with the addition of 2.5 equiv of the chiral complexing
agent (ꢀ)-52 (186 mg, 528 mmol, 2.5 equiv) and conducting the irradiation
with RPR-3500 ꢄ lamps (l=370 nm) at ꢀ608C, cyclobutane (ꢀ)-16
3518
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3509 – 3525

Total Synthesis of Meloscine
FULL PAPER
Enantiopure: According to the above procedure, enantiomerically pure
enone (+)-21 (340 mg, 758 mmol, quant.) was obtained in >99% ee when
starting from enantiomerically pure cyclobutane (ꢀ)-16. [a]2D0 =+60 (c=
1.0 in CH2Cl2); chiral HPLC (Chiralpak AD, n-hexane/iPrOH 70:30):
tR =6.9 min; rac: tR =7.2/13.0 min, k=1.0/2.6, a=2.6.
when starting from enantiomerically pure enone (+)-21. [a]2D0 =+28.9
(c=1.0 in CH2Cl2); chiral HPLC (Chiralpak AS-H, n-hexane/iPrOH
80:20): tR =11.4 min, k=2.3; rac: tR =14.2/19.6 min, k=2.3/3.6, a=1.6.
(ꢂ)-(1R,2R,11R,15S)-14-Benzyl-3,16-dioxo-4,14-diazatetracyclo-
[9.3.2.02,11.05,10]hexadeca-5,7,9-trien-15-yl acetate (23): Enolacetate 22
(50 mg, 102 mmol) was stirred in 7.5 vol% TFA in CH2Cl2 (5 mL) for
2.0 h at RT. The mixture was diluted with CH2Cl2 (20 mL) and H2O
(20 mL) and made basic with 5% aqueous NaOH. Brine (50 mL) was
added, phases were separated, and the aqueous phase extracted with
CH2Cl2 (3ꢅ30 mL). The combined organic phases were dried over
Na2SO4, filtered and the solvent evaporated to give acetate 23 (38 mg,
97 mmol, 95%) as an off-white solid. Attempted purification by flash
chromatography led to significant losses (60–70%) of product but no im-
provement of product purity. Rf =0.37 (P/EtOAc 1:1, [UV]); m.p. 175–
(+)-(2aS,8bR,11aS)-11-Benzyl-11a-hydroxy-2,2a,9,10,11,11a-hexahydro-
1H-pyrrolo[3’,2’:2,3]cyclopentaACTHUNRTGNEUNG[1,2-c]quinoline-1,3(4H)-dione (18): Enone
21 (250 mg, 557 mmol) was stirred in 10 vol% TFA in CH2Cl2 (10 mL) for
1.5 h at RT. The mixture was diluted with CH2Cl2 (20 mL) and H2O
(20 mL) and made basic with 15% aqueous NaOH. The resulting suspen-
sion is adjusted to pH 8–10 with sat. aqueous NH4Cl, phases were sepa-
rated, and the aqueous phase extracted with CH2Cl2 (3ꢅ25 mL). The
combined organic phases were dried over Na2SO4, filtered and the sol-
vent evaporated to give hemiaminal 18 (192 mg, 551 mmol, 99%) as a
light yellow, crystalline solid. Attempted flash chromatography was inef-
fective and led to widespread decomposition. Rf =0.55 (EtOAc, [UV]);
m.p. 166–1698C (decomp); 1H NMR (500 MHz, CDCl3): d=9.43 (s, 1H,
NH), 7.99 (d, 3J=7.6 Hz, 1H, H8), 7.37–7.32 (m, 2H, HBn), 7.31–7.24 (m,
4H, HBn, H6), 7.11 (t, 3J=7.6 Hz, 1H, H7), 6.88 (d, 3J=7.7 Hz, 1H, H5),
4.00 (d, 2J=14.2 Hz, 1H, CHHPh), 3.70 (d, 2J=14.2 Hz, 1H, CHHPh),
3.58 (s, 1H, OH), 3.35 (virt. t, 3J=9.4 Hz, 1H, H2a), 3.25–3.15 (m, 1H,
CHHN), 3.21 (dd, 2J=19.2, 3J=9.0 Hz, 1H, C2HH), 3.09 (ddd, 2J=10.2,
3J=10.5, 3J=4.0 Hz, 1H, CHHN), 2.52 (dd, 2J=19.2, 3J=9.6 Hz, 1H,
C2HH), 2.41 (ddd, 2J=12.5, 3J=8.6, 3J=4.0 Hz, 1H, RCHH), 2.27 ppm
(ddd, 2J=12.5, 3J=10.5, 3J=7.3 Hz, 1H, RCHH); 13C NMR (63 MHz,
CDCl3): d=208.6 (s, CO), 170.9 (s, CONH), 138.5 (s, C1Bn), 136.1 (s, C4a),
130.8 (d, C8H), 128.7 (d, CarH), 128.4 (d, CBnH), 127.9 (d, CBnH), 127.1
(d, CarH), 123.9 (d, C7H), 121.3 (s, C8aH), 115.8 (d, C5H), 96.7 (s, C11a),
56.6 (s, C8b), 50.3 (t, CH2Ph), 49.5 (t, CH2N), 43.4 (d, C2aH), 37.9 (t,
C2HH), 37.4 ppm (t, RCH2); IR (KBr): n˜ =3440 (vs, br, OH), 3059 (m,
CarH), 2913 (m, CalH), 2880 (m, CalH), 1743 (vs, CO), 1690 (vs, CON),
1588 (m), 1491 (s), 1420 (s), 1132 (m), 738 (m), 699 (m), 509 cmꢀ1 (m);
MS (70 eV): m/z (%): 348 (1) [M+], 330 (12), 320 (82) [MꢀCO+], 305
(11) [MꢀCOOH+] 289 (17), 265 (34), 91 (100) [C7H7+]; HRMS (70 eV):
m/z: calcd for C21H20N2O3: 348.1474 [M+], found 348.1481; elemental
analysis calcd (%) for C21H20N2O3: C 72.40, H 5.79; found: C 72.30, H
5.78.
1
3
1778C; H NMR (500 MHz, CDCl3): d=8.83 (brs, 1H, NH), 7.32 (d, J=
7.4 Hz, 1H, H9), 7.28–7.22 (m, 4H, H2/3Bn), 7.21–7.16 (m, 2H, H7, H4Bn),
7.06 (t, 3J=7.6 Hz, 1H, H8), 6.75 (d, 3J=7.7 Hz, 1H, H6), 5.40 (d, 3J=
5.5 Hz, 1H, CHOAc), 4.54 (d, 3J=5.5 Hz, 1H, CHN), 4.04 (d, 2J=
13.6 Hz, 1H, CHHPh), 3.79 (d, 2J=13.6 Hz, 1H, CHHPh), 3.17 (s, 1H,
C2H), 2.91 (dd, 2J=12.9, 3J=6.3 Hz, 1H, CHHN), 2.75 (ddd, 2J=12.9,
3J=12.9, 3J=4.0 Hz, 1H, CHHN), 2.41 (dd, 2J=12.4, 3J=4.0 Hz, 1H,
3
2
3
RCHH), 2.19 (s, 3H, CH3), 1.96 ppm (ddd, J=12.9, J=12.4, J=6.3 Hz,
1H, RCHH); 13C NMR (63 MHz, CDCl3): d=211.1 (s, CO), 169.4 (s,
CONH/COAc), 167.9 (s, CONH/COAc), 138.8 (s, C5/C1Bn), 135.6 (s, C5/
C1Bn), 129.2 (d, CarH), 128.4 (d, CarH), 128.3 (d, CarH), 127.3 (d, CarH),
126.7 (d, CarH), 124.0 (d, C8H), 121.1 (s, C10), 115.9 (d, C6H), 77.3 (d,
CHO), 60.6 (t, CH2Ph), 60.3 (d, CHN), 51.7 (s, C11), 48.4 (d, C2H), 44.5
(t, CH2N), 35.7 (t, RCH2), 20.7 ppm (q, CH3); IR (KBr): n˜ =3447 (w, br,
NH), 3198 (m), 3069 (w, CarH), 2908 (m, CalH), 1757 (vs, CO), 1672 (vs,
CO), 1591 (m), 1490 (s), 1369 (s), 1274 (s), 753 (m), 700 cmꢀ1 (w); MS
(70 eV): m/z (%): 390 (43) [M+], 347 (64) [MꢀAc+], 330 (6)
[MꢀAcOH+], 289 (24), 199 (27), 167 (28), 149 (55), 91 (100) [C7H7+], 57
(24) [tBu+], 43 (28) [Ac+]; HRMS (70 eV): m/z: calcd for C23H22N2O4:
390.1580 [M+], found 390.1578.
(+)-(1S,2aS,8bR,11aR)-1-(Acetyloxy)-3-oxo-1,2,2a,3,4,9,10,11a-octahy-
dro-11H-pyrrolo[3’,2’:2,3]cyclopentaACTHNUTRGENUGN[1,2-c]quinoline-11-tert-butylcarboxy-
late (28): Enolacetate 22 (346 mg, 705 mmol) was stirred in 10 vol% TFA
in CH2Cl2 (15 mL) for 1.0 h at RT. The solution was concentrated, redis-
solved in MeOH (120 mL) and cooled to 08C. Pd(OH)2/C (300 mg 10–
15% Pd with 50% H2O, ꢃ141 mmol Pd, 20 mol%) was added, and the
solution was stirred under a H2 atmosphere (1 bar) for 18 h while slowly
warming to RT. After removal of the catalyst by filtration and evapora-
tion of the solvent, the remaining crude TFA salt was suspended in
CH2Cl2 (25 mL), the solution adjusted to pH 9–10 with NEt3, Boc2O
(200 mg, 917 mmol, 1.3 equiv) was added and the mixture was stirred at
RT for 1.0 h. Sat. aqueous NH4Cl (40 mL) and H2O (40 mL) were added,
phases were separated and the aqueous phase extracted with CH2Cl2 (5ꢅ
30 mL). The combined organic phases were dired over Na2SO4, filtered
and concentrated. The crude product was purified by flash chromatogra-
phy (2.5ꢅ20 cm, P/EtOAc 1:1!EtOAc) to give acetate 28 (212 mg,
549 mmol, 78%) as a white solid. Rf =0.50 (EtOAc, [UV]); m.p. 130–
1388C (polymorph), 200–2038C (decomp); 1H NMR (360 MHz, CDCl3):
d=9.21 (s, 1H, NH), 7.32 (d, 3J=7.5 Hz, 1H, H8), 7.22 (t, 3J=7.5 Hz,
1H, H6), 7.10 (t, 3J=7.3 Hz, 1H, H7), 6.84 (d, 3J=7.9 Hz, 1H, H5), 5.16
(brm, 1H, CHOAc), 4.41/4.27 (br, 1H, CHN), 3.92/3.86 (brm, 1H,
CHHN), 3.47 (brm, 1H, CHHN), 2.90 (brm, 1H, C2aH), 2.61 (brm, 1H,
CHH), 2.25–2.10 (m, 3H, CHH, RCH2), 1.83 (s, 3H, CH3), 1.46 ppm
(+)-(3aS,9bR)-9b-2-[Benzyl(tert-butoxycarbonyl)amino]ethyl-1,4-dioxo-
3a,4,5,9b-tetrahydro-1H-cyclopenta[c]quinoline-2-yl acetate (22): Enone
21 (2.63 g, 5.86 mmol) and NEt3 (1.14 mL, 830 mg, 8.20 mmol, 1.40 equiv)
were dissolved in THF (40 mL), cooled to 08C, treated dropwise with
acetyl chloride (0.48 mL, 529 mg, 6.74 mmol, 1.15 equiv) and stirred at
08C for 30 min. H2O (100 mL) was added, the THF evaporated and the
aqueous residue extracted with EtOAc (3ꢅ100 mL). The combined or-
ganic extractes were dried over Na2SO4, filtered and concentrated. After
purification of the crude product by flash chromatography (5.0ꢅ15 cm,
P/EtOAc 1:1!2:3), enolacetate 22 (2.73 g, 5.57 mmol, 95%) was ob-
tained as a light yellow foam. Rf =0.35 (P/EtOAc 1:1, [UV]); 1H NMR
(360 MHz, CDCl3, 608C): d=8.94 (br., 1H, NH), 7.60 (brm, 1H, H9),
7.47 (br., 1H, H3), 7.30–7.15 (m, 4H, Har), 7.14–7.02 (m, 3H, Har), 6.74
(brm, 1H, H6), 4.40–4.20 (m, 2H, CH2Ph), 3.82 (br, 1H, Hal), 3.20 (br,
1H, Hal), 2.88 (br., 1H, Hal), 2.31 (br, 1H, Hal), 2.21 (s, 3H, CH3), 1.88
(br., 1H, Hal), 1.44 ppm [br, 9H, CACHTUNGTRNEUNG
(CH3)3]; 13C NMR (90 MHz, CDCl3,
608C): d=199.2 (s, CO), 166.9* (s, COAc), 166.8* (s, CONH), 155.3 (s,
COBoc), 149.5 (s, C2OAc), 141.6 (d, C3H), 138.1* (s, C5a), 135.2* (s, C1Bn),
128.8 (d, CarH), 128.5 (d, CarH), 128.1 (d, CarH), 127.8 (brd, CarH), 127.4
(d, CarH), 124.3 (d, CarH), 119.2 (s, C9a), 115.7 (d, C6H), 80.2 [s, OC-
[brs, 9H, CACTHNUTRGNEUNG
(CH3)3]; 13C NMR (90 MHz, CDCl3): d=170.7 (s, COAc),
169.6 (brs, CONH), 153.7 (s, COBoc), 135.1 (s, C4a), 128.3 (d, C6H), 127.1
(d, C8H), 125.5 (brs, C8a), 124.3 (d, C7H), 115.9 (d, C5H), 80.5 [brs, C-
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
to Boc coalescence; IR (KBr): n˜ =3229 (m, br, CONH), 3064 (w, CarH),
2977 (m, CalH), 1780 (s, CO), 1730 (s, CO), 1680 (vs, CONH), 1594 (m),
1367 (m), 1188 (m), 1042 (m), 757 (m), 701 cmꢀ1 (m); MS (70 eV): m/z
(%): 490 (2) [M+], 434 (14) [MꢀtBu+], 417 (7), 390 (78) [MꢀBoc+], 345
(25) [MꢀBocꢀAc+], 120 (62) [BnNHCH2+], 107 (36), 91 (100) [C7H7+],
57 (7) [tBu+], 43 (25) [Ac+]; HRMS (70 eV): m/z: calcd for C28H30N2O6:
490.2104 [M+], found 490.2112.
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
CalH), 2931 (w, CalH), 1743 (s, CO), 1686 (vs, CONH), 1391 (s, CH3),
1240 (s, COOMe), 1167 (m), 1042 (m), 758 cmꢀ1 (m); MS (70 eV): m/z
(%): 386 (2) [M+], 326 (24) [MꢀAcOH+], 270 (100) [MꢀAcOꢀtBu+],
226 (26) [MꢀAcOHꢀBoc+], 173 (19), 57 (81) [tBu+], 44 (13) [Ac+], 41
(14) [C3H5+]; HRMS (70 eV): m/z: calcd for C21H26N2O5: 386.1842 [M+],
found 386.1883.
Enantiopure: According to the above procedure, enantiomerically pure
enolacetate (+)-22 (353 mg, 720 mmol, 95%) was obtained in >99% ee
Chem. Eur. J. 2009, 15, 3509 – 3525
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3519

T. Bach et al.
Enantiopure: According to the above procedure, enantiomerically pure
acetate (+)-28 (270 mg, 700 mmol, 76%) was obtained as a white foam
with >99% ee when starting from enantiomerically pure enolacetate
(+)-22. [a]2D0 =+23.4 (c=1.0 in CH2Cl2); chiral HPLC (Chiralpak AD-H,
n-hexane/iPrOH 80:20): tR =11.7 min, k=2.4; rac: tR =14.8/17.3 min, k=
2.4/3.0, a=1.25.
Enantiopure: According to the above procedure, enantiomerically pure
ketone (+)-30 (66 mg, 193 mmol, 95%) was obtained as a white foam
with >99% ee when starting from the enantiomerically pure alcohol (+)-
29. [a]2D0 =+123 (c=1.0 in CH2Cl2); chiral HPLC (Chiralpak AD-H, n-
hexane/iPrOH 50:50): tR =6.1 min, k=0.7; rac: tR =7.3/10.7 min, k=0.9/
1.7, a=1.9.
(+)-(1S,2aS,8bR,11aR)-1-Hydroxy-3-oxo-1,2,2a,3,4,9,10,11a-octahydro-
(+)-(2aS,8bR,11aS)-1-[(E)-2-Ethoxy-2-oxoethyliden]-3-oxo-1,2,2a,3,4,9,
11H-pyrrolo[3’,2’:2,3]cyclopentaACTHUNRTGNEUNG[1,2-c]quinoline-11-tert-butylcarboxylate
10,11a-octahydro-11H-pyrrolo[3’,2’:2,3]cyclopentaACTHNUGRTNEUNG[1,2-c]quinoline-11-tert-
(29): Acetate 28 (206 mg, 533 mmol) was stirred in K2CO3 in MeOH
(25 mL) for 3.5 h at RT. After evaporation of the solvent, the residue was
filtered through a short plug of silica (3.5ꢅ5 cm, CH2Cl2/MeOH 15:1)
and concentrated to give the free alcohol 29 (173 mg, 502 mmol, 94%) as
a white, crystalline solid. Rf =0.55 (CH2Cl2/MeOH 10:1, [UV]); m.p. 114–
butylcarboxylate (31): Ketone 30 (1.50 g, 4.38 mmol) and the Wittig re-
agent PH3P=CHCOOEt (1.98 g, 5.69 mmol, 1.30 equiv) were refluxed in
THF (75 mL) for 20 h. After remomal of the solvent, the residue was pu-
rified by flash chromatography (5.0ꢅ25 cm, P/EtOAc 1:1!1:2) to give
the a,b-unsaturated ester 31 (1.52 g, 3.68 mmol, 84%) as a white crystal-
line solid. Rf =0.60 (EtOAc, [UV]); m.p. 190–1938C (decomp); 1H NMR
(250 MHz, CDCl3): d=9.34/9.18 (br., 1H, NH), 7.30–7.14 (brm, 2H, H6,
H8), 7.07 (t, 3J=7.3 Hz, 1H, H7), 6.86 (d, 3J=7.7 Hz, 1H, H5), 6.41/6.19
(brm, 1H, =CH), 4.98/4.74 (brm, 1H, CHN), 4.12–3.81 (brm, 3H, OCH2,
Hal), 3.72–3.40 (brm, 2H, Hal), 3.35–2.95 (m, 2H, Hal), 2.20–2.00 (brm,
2H, RCH2), 1.48 [brs, 9H, (CH3)3], 1.25 ppm (t, 3J=7.1 Hz, 3H, CH3);
signal doubling and strong broadening due to Boc coalescence; 13C NMR
(63 MHz, CDCl3): d=171.0 (s, CONH), 166.3 (brs, COOEt), 159.3 (brs,
C1), 154.8/154.4 (brs, COBoc), 135.6 (s, C4a), 128.6 (d, C6H), 126.7 (brd,
C8H), 124.5 (brs, C8a), 124.1 (d, C7H), 118.9/118.2 (brd, =CH), 116.0 (d,
1
1158C (decomp); H NMR (250 MHz, CDCl3): d=9.49 (s, 1H, NH), 7.31
(d, 3J=7.4 Hz, 1H, H8), 7.18 (t, 3J=7.3 Hz, 1H, H6), 7.08 (t, 3J=7.4 Hz,
1H, H7), 6.85 (d, 3J=7.7 Hz, 1H, H5), 4.31 (d, 3J=3.8 Hz, 0.65H*,
CHNBoc), 4.20 (brm, 0.35H*, CHNBoc), 4.11 (virt. dt, 3Jt =10.4, 3Jd =
3.8 Hz, 1H, CHOH), 3.90/3.77 (brm, 0.35H*/0.65H*, CHHN), 3.76 (brs,
<1H, OH), 3.43 (ddd, 3/2J=10.5, 3/2J=10.4, 3J=6.9 Hz, 1H, CHHN),
2.81 (dd, 3J=11.8, 3J=6.3 Hz, 1H, H2a), 2.48 (virt. dt, 2J=12.1, 3J=
2
3
3
6.0 Hz, 1H, CHH), 2.12 (ddd, J=13.2, J=6.7, J=2.8 Hz, 1H, RCHH),
2.08–1.88 (m, 2H, CHH, RCHH), 1.51 ppm [s, 9H, CACHTNUGTRNEG(UN CH3)3]; *signals
are doubled due to Boc coalescence; 13C NMR (63 MHz, CDCl3): d=
170.9 (s, CONH), 155.4 (s, COBoc), 134.7 (s, C4a), 128.1 (d, C6H), 127.3 (d,
C5H), 80.8/80.4 [brs, C
55.3/55.0 (brs, C8b), 46.6 (brd, C2aH), 46.1 (brt, CH2N), 37.9/37.1 (brt,
RCH2), 35.5/34.9 (brt, C2HH), 28.4 [q, C(CH3)3], 14.2 (q, CH3); signal
ACHTNUGRTNE(NUGN CH3)3], 72.8/72.4 (brd, CHN), 60.0 (t, OCH2),
C8H), 126.8 (s, C8a), 124.4 (d, C7H), 116.0 (d, C5H), 80.6 [s, C
(d, CHNBoc), 78.9 (d, CHOH), 55.1 (s, C8b), 47.3 (d, C2aH), 46.3 (t,
CH2N), 40.1 (t, RCH2), 37.9 (t, C2HH), 28.4 ppm [q, C(CH3)3]; IR (KBr):
ACHTUGNTRNEN(UNG CH3)3], 79.0
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
doubling and strong broadening due to Boc coalescence; IR (KBr): n˜ =
3412 (w, br, CONH), 3066 (w, CarH), 2978 (m, CalH), 1687 (vs, br, CO),
n˜ =3422 (vs, br, OH), 2973 (w, CalH), 2927 (w, CalH), 1680 (vs, C=O),
ꢀ
1657 (vs, C=ONH), 1594 (m, Car), 1402 (vs, CH3), 1169 (m, C O), 1117
ꢀ
1593 (m, Car), 1490 (s), 1390 (vs, CH3), 1210 (s), 1165 (s, C O), 1034 (m),
ꢀ1
ꢀ
(m, C O), 903 (w), 754 (m, Car), 730 cm (m, Car); MS (70 eV): m/z
755 cmꢀ1 (s, Car); MS (70 eV): m/z (%): 412 (5) [M+], 356 (40) [MꢀtBu+
], 312 (49) [MꢀBoc+], 283 (25), 266 (25) [MꢀBocꢀOEt+], 239 (19)
[MꢀBocꢀCOOEt+], 159 (21) [RCH3+], 57 (100) [tBu+]; HRMS
(70 eV): m/z: calcd for C23H28N2O5: 412.1998 [M+], found 412.1993; ele-
mental analysis calcd (%) for C23H28N2O: C 66.97, H 6.84; found: C
66.90, H 6.75.
(%): 344 (36) [M+], 288 (55) [MꢀtBu+], 243 (29) [MꢀBoc+], 215 (26),
159 (20) [RCH3+], 144 (17), 84 (21), 57 (100) [tBu+], 41 (17) [C3H5+];
HRMS (70 eV): m/z: calcd for C19H24N2O4: 344.1736 [M+], found
344.1733; elemental analysis calcd (%) for C19H24N2O4: C 66.26, H 7.02;
found: C 66.30, H 6.93.
Enantiopure: According to the above procedure, enantiomerically pure
alcohol (+)-29 (405 mg, 1.18 mmol, 96%) was obtained as a white foam
with >99% ee when starting from enantiomerically pure acetate (+)-28.
[a]2D0 =+26.0 (c=1.0 in CH2Cl2); chiral HPLC (Chiralpak AD-H, n-
hexane/iPrOH 70:30): tR =9.1 min, k=1.8; rac: tR =11.6/14.6 min, k=1.7/
2.4, a=1.4.
Enantiopure: According to the above procedure, enantiomerically pure
a,b-unsaturated ester (+)-31 (157 mg, 381 mmol, 84%) was obtained as a
white foam with >99% ee when starting from the enantiomerically pure
ketone (+)-30. [a]2D0 =+46.1 (c=1.0 in CH2Cl2); chiral HPLC (Chiralpak
AD-H, n-hexane/iPrOH 80:20): tR =11.9 min, k=2.3; rac: tR =12.8/
14.9 min, k=2.3/3.0, a=1.3.
(+)-(2aS,8bR,11aR)-1,3-Dioxo-1,2,2a,3,4,9,10,11a-octahydro-11H-
pyrrolo[3’,2’:2,3]cyclopentaACTHNUGTRNEUNG[1,2-c]quinoline-11-tert-butylcarboxylate (30):
(+)-(2aS,8bR,11aS)-1-[(E)-2-Hydroxyethyliden]-3-oxo-1,2,2a,3,4,9,10,11a-
octahydro-11H-pyrrolo[3’,2’:2,3]cyclopentaACTHNUGRTENUNG[1,2-c]quinoline-11-tert-butyl-
Alcohol 29 (178 mg, 517 mmol) and IBX (434 mg, 1.55 mmol, 3.0 equiv)
were stirred in DMSO (7.5 mL) for 30 h at RT. H2O (150 mL) and sat.
aqueous NaHCO3 (30 mL) were added and the mixture extracted once
with EtOAc (100 mL). The organic extract was washed with H2O (3ꢅ
100 mL) and brine (2ꢅ100 mL), dried over Na2SO4 and filtered. After re-
moval of the solvent, ketone 30 (167 mg, 488 mmol, 94%) was obtained
as a white, crystalline solid, which sometimes contained an additional
carboxylate (32): a,b-Unsaturated ester 31 (200 mg, 485 mmol) was dis-
solved in dry CH2Cl2 (20 mL), cooled to ꢀ458C, treated dropwise with
DIBAL-H (1.1m in cyclohexane, 1.76 mL, 1.94 mmol, 4.0 equiv) and
stirred at ꢀ458C for 30 min. The reaction was quenched with a few drops
of MeOH, and the mixture concentrated to dryness. After purification of
the crude residue by flash chromatography (2.5ꢅ25 cm, CH2Cl2/MeOH
30:1!20:1), allylic alcohol 32 (145 mg, 391 mmol, 81%) was obtained as
1= equiv of co-crystallized EtOAc. Rf =0.48 (EtOAc, [UV]); m.p. 196–
2
1
1998C (decomp); 1H NMR (250 MHz, CDCl3): d=9.74 (br, 1H, NH),
a white foam. Rf =0.20 (EtOAc, [UV]); H NMR (360 MHz, CDCl3): d=
3
3
3
9.22 (brs, 1H, NH), 7.24 (brs, 1H, H8), 7.19 (t, J=7.6 Hz, 1H, H6), 7.06
7.35–7.20 (m, 2H, CarH), 7.10 (t, J=7.2 Hz, 1H, H7), 6.92 (d, J=7.7 Hz,
1H, H5), 4.46/4.25 (br, 1H, CHN), 3.94 (brm, 1H, CHHN), 3.49 (ddd,
2J=11.5, 3J=8.1, 3J=7.7 Hz, 1H, CHHN), 3.17 (virt. t, 3J=8.6 Hz, 1H,
C2aH), 2.88 (virt. d, 3J=8.6 Hz, 2H, C2HH), 2.32–2.10 (m, 2H, RCH2),
3
3
(t, J=7.5 Hz, 1H, H7), 6.84 (d, J=7.7 Hz, 1H, H5), 6.13/5.97 (brs, 1H, =
CH), 4.90/4.67 (brs, 1H, CHN), 4.12 (d, 3J=5.2 Hz, 2H, CH2O), 4.05–
3.80 (brm, 1H, CHHN), 3.43 (ddd, 2J=16.1, 3J=11.8, 3J=7.9 Hz, 1H,
CHHN), 3.02–2.88 (m, 2H, C2aH, C2HH), 2.85–2.55 (brm, 1H, C2HH),
1.46 ppm [brs, 9H, CACHTUNGTRENNUNG
(CH3)3]; 13C NMR (63 MHz, CDCl3): d=209.3 (brs,
2.10/1.95 (brm, 3H, RCH2, OH), 1.48 ppm [s, 9H, CACTHNUTRGNEUNG
(CH3)3]; 13C NMR
CO), 170.2 (s, CONH), 154.5 (s, COBoc), 135.3 (s, C4a), 129.0 (d, C6H),
(90 MHz, CDCl3): d=171.4 (s, CONH), 154.6 (brs, COBoc), 140.7 (s, C1),
135.4 (s, C4a), 128.3 (d, C6H), 127.1/126.9 (brd, =CH), 126.8 (brd, C8H),
125.8/125.4 (brs, C8a), 124.2 (d, C7H), 115.8 (d, C5H), 80.3/80.0 [brs, C-
126.8 (d, C8H), 124.5 (d, C7H), 124.0 (s, C8a), 116.2 (d, C5H), 80.7 [s, C-
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
broadening due to Boc coalescence; IR (KBr): n˜ =3204 (w, br, CONH),
3067 (w, CarH), 2979 (m, CalH), 2929 (m, CalH), 1756 (s, C=O), 1678 (vs,
ꢀ
AHCTUNGTRENNUNG
br, CON), 1592 (m, Car), 1493 (s), 1392 (vs, CH3), 1245 (m), 1170 (s, C
O), 754 cmꢀ1 (s, Car); MS (70 eV): m/z (%): 342 (18) [M+], 286 (49)
[MꢀtBu+], 269 (13) [MꢀtBuO+], 242 (15) [MꢀBoc+H+], 214 (25), 159
(77) [RCH3+], 57 (100) [tBu+]; HRMS (70 eV): m/z: calcd for
C19H22N2O4: 342.1580 [M+], found 342.1580.
due to Boc coalescence; IR (KBr): n˜ =3224 (w, br., OH/NH), 2973 (w,
CalH), 2931 (w, CalH), 1666 (vs, br., CO), 1592 (m, Car), 1488 (m), 1389 (s,
br.), 1164 (m), 755 cmꢀ1 (m, Car); MS (70 eV): m/z (%): 370 (4) [M+],
352 (15) [MꢀH2O+], 314 (60), 296 (77) [MꢀtBuOH+], 226 (82), 159 (35)
3520
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3509 – 3525

Total Synthesis of Meloscine
FULL PAPER
[RCH3+], 57 (100) [tBu+], 41 (28) [C3H5+]; HRMS (70 eV): m/z: calcd
for C21H26N2O4: 370.1893 [M+], found 370.1883.
2HCOOH+HCOONa+HCOOK+K+]; HPLC (ODS-A, MeCN/H2O
20:80!100:0, 30 min): tR =18.7 min (N-cis-vinyl, major), tR =19.3 min
(N-trans-vinyl, minor).
Enantiopure: According to the above procedure, enantiomerically pure
allylic alcohol (+)-32 was obtained with >99% ee when starting from
the enantiomerically pure a,b-unsaturated ester (+)-31. [a]2D0 = +64.8
(c=1.0 in CH2Cl2); chiral HPLC (Chiralpak AD-H, n-hexane/iPrOH
80:20): tR =12.6 min, k=2.7; rac: tR =15.2/30.2 min, k=2.8/6.5, a=2.3.
N-Boc deprotection and N-alkylation of rearragement product 39: Rear-
rangement product 39 (560 mg, 1.31 mmol, d.r. 70:30) was stirred in
15 vol% TFA in CH2Cl2 (20 mL) for 1.0 h and evaporated to dryness.
The residue was dissolved in H2O (30 mL) and the aqueous solution
made basic with aqueous NaOH (2.0m). The solution was extracted with
CH2Cl2 (4ꢅ25 mL), adjusted to pH 8–9 with sat. aqueous NH4Cl and ex-
tracted with CH2Cl2 (4ꢅ25 mL) again. The combined organic phases
were dried over Na2SO4, filtered and concentrated. The solid residue was
dissolved in MeCN (15 mL), K2CO3 (181 mg, 1.31 mmol, 1.0 equiv) and
allyl bromide (91 mL, 128 mg, 1.05 mmol, 0.8 equiv) were added and the
solution stirred at RT for 20 h. H2O (75 mL) was added and the mixture
extracted with CH2Cl2 (4ꢅ50 mL). The combined organic phases were
dried over Na2SO4, filtered and concentrated. The residue was purified
by flash chromatography (3.5ꢅ20 cm, CH2Cl2/MeOH 50:1!20:1) to give
allylamine 42 (312 mg, 851 mmol, 65%, 93% rel. to d.r.) as the unpolar
and pentacyclic lactam 40 (110 mg, 374 mmol, 29%, 95% rel. to d.r.) as
the polar fraction as light yellow foams or solids. Rf =0.16 (EtOAc,
[UV]); m.p. 195–2028C (decomp); 1H NMR (360 MHz, CDCl3): d=9.82
(ꢂ)-(2aS,8bR,11aS)-11-[(Z)-3-Iodo-2-propenyl]-3-oxo-2,2a,3,4,9,10,11,11a-
octahydro-1H-pyrrolo[3’,2’:2,3]cyclopentaACHTNUGTRNEUNG[1,2-c]quinoline-1-yliden-2-eth-
ylacetate (33): a,b-Unsaturated ester 31 (75 mg, 182 mmol) was stirred in
20 vol% TFA in CH2Cl2 (2.0 mL) for 1.0 h at RT and the solution con-
centrated to dryness. The residue was dissolved in MeCN (2.0 mL), made
basic with solid K2CO3 and treated with (Z)-3-bromo-1-iodopropene
(22 mL, 55 mg, 222 mmol, 1.2 equiv) for 6.0 h at RT. H2O (50 mL) was
added and the mixture extracted with EtOAc (2ꢅ25 mL). The combined
organic phases were washed with H2O (15 mL) and brine (15 mL), dried
over Na2SO4, filtered and the solvent evaporated to give vinyliodide 33
(80 mg, 167 mmol, 92%) as an off-white, crystalline solid. Rf =0.58
(EtOAc, [UV]); m.p. 177–1828C; 1H NMR (360 MHz, CDCl3): d=8.08
3
3
(brs, 1H, NH), 7.31 (d, J=7.5 Hz, 1H, H8), 7.18 (t, J=7.6 Hz, 1H, H6),
7.05 (t, 3J=7.6 Hz, 1H, H7), 6.74 (d, 3J=7.7 Hz, 1H, H5), 6.40 (d, 3J=
7.7 Hz, 1H, =CHI), 6.32 (virt. td, 3Jt =7.3, 3Jd =4.8 Hz, 1H, CH=CHI),
5.81 (virt. t, 4J=1.8 Hz, 1H, =CH), 4.12 (q, 3J=7.1 Hz, 2H, OCH2), 3.67
(brs, 1H, CHN), 3.61–3.48 (m, 2H, NCHH, C2HH), 3.42–3.32 (m, 2H,
3
3
(brs, 1H, NH), 7.32 (d, J=7.3 Hz, 1H, H8), 7.23 (t, J=7.6 Hz, 1H, H6),
7.10 (t, 3J=7.6 Hz, 1H, H7), 6.91 (d, 3J=7.9 Hz, 1H, H5), 5.77 (dd, 3JE =
3
3
3
17.4, JZ =10.7 Hz, 1H, =CH), 4.93 (d, J=10.7 Hz, =CHHE), 4.91 (d, J=
17.4 Hz, 1H, =CHZH), 4.48 (s, 1H, CHN), 3.97 (ddd, 2J=12.1, 3J=8.1,
3J=7.5 Hz, 1H, CHHN), 3.29 (ddd, 3J=12.1, 3J=9.9, 3J=3.0 Hz, 1H,
CHHN), 3.08 (d, 2J=17.3 Hz, 1H, CHHCON), 2.76 (virt. t, 3J=9.9 Hz,
1H, C2aH), 2.49 (d, 2J=17.3 Hz, 1H, CHHCON), 2.51–2.43 (m, 1H,
2
3
C2aH, CHHN), 3.15 (dd, J=14.3, J=7.0 Hz, 1H, NCHH), 2.91–2.78 (m,
2H, CHHN, C2HH), 2.31 (ddd, 2J=13.5, 3J=8.1, 3J=6.4 Hz, 1H,
RCHH), 1.88 (virt. td, 2/3Jt =11.7, 3Jd =4.5 Hz, 1H, RCHH), 1.24 ppm (t,
3J=7.1 Hz, 3H, CH3); 13C NMR (63 MHz, CDCl3): d=172.4 (s, CONH),
165.8 (s, COOEt), 159.9 (s, C1), 137.2 (d, HC=CHI), 136.2 (s, C4a), 128.2
(d, C6H), 126.9 (d, C8H), 124.2 (s, C8a), 124.0 (d, C7H), 116.8 (d, =CH),
116.1 (d, C5H), 83.6 (d, HC=CHI), 80.3 (d, CHN), 60.1 (t, OCH2), 56.7 (s,
C8b), 56.1 (t, CH2CH=), 53.6 (t, CH2N), 50.0 (d, C2aH), 37.9 (t, RCH2),
36.7 (t, C2HH), 14.2 ppm (q, CH3); IR (KBr): n˜ =3432 (w, br., CONH),
3058 (w, CarH), 2975 (w, CalH), 1712 (s, COOEt), 1675 (vs, CONH), 1592
2
3
3
RCHH), 2.28–2.20 (m, 2H, C2HH), 2.13 ppm (ddd, J=13.6, J=9.9, J=
7.5 Hz, 1H, RCHH); 13C NMR (90 MHz, CDCl3): d=175.7 (s, CONR2),
171.0 (s, CONH), 141.9 (d, =CH), 135.9 (s, C4a), 128.5 (d, C6H), 126.2 (d,
C8H), 124.2 (d, C7H), 122.8 (s, C8a), 116.3 (d, C5H), 113.5 (t, =CH2), 81.8
(d, CHN), 55.3 (s, C8b), 49.4 (s, C1), 48.1 (d, C2aH), 45.95 (t, CH2CON/
C2HH), 45.88 (t, CH2CON/C2HH), 42.5 (t, CH2N), 40.6 ppm (t, RCH2);
IR (ATR, neat): n˜ =3211 (w, NH), 3079/3063 (w, Car/olefH), 2966/2928 (w,
CalH), 1670 (vs, br., CON), 1592 (s, Car), 1487 (s), 1364 (s, br.), 1241 (m),
917 (m), 756 (m, Car), 726 cmꢀ1 (m, Car); MS (70 eV): m/z (%): 294 (100)
[M+], 172 (51), 123 (26); MS (ESI): m/z: 295 [M+H+], 611 [2M+Na+],
905 [3M+Na+]; HRMS (70 eV): m/z: calcd for C18H18N2O2: 294.1368
[M+], found 294.1371.
ꢀ
(s, Car), 1490 (m), 1388 (m, CH3), 1372 (m), 1209 (s), 1139 (m, C O),
1035 (m), 754 (s, Car); MS (70 eV): m/z (%): 478 (16) [M+], 449 (7)
[MꢀEt+], 433 (8) [MꢀOEt+], 405 (10) [MꢀCOOEt+], 351 (27) [MꢀI+],
311 (100) [MꢀC3H4I+], 277 (62), 167 (19) [C3H4I+]; HRMS (70 eV):
m/z: calcd for C21H23IN2O3: 478.0754 [M+], found 478.0755.
Single-crystal X-ray structure determination of compound (ꢂ)-33: Crys-
tal data and details of the structure determination: formula:
C21H23IN2O3; Mr =478.31; crystal color and shape: colorless fragment,
crystal dimensions=0.53ꢅ0.56ꢅ0.58 mm; crystal system: triclinic; space
Enantiopure: According to the above sequence consisting of Johnson–
Claisen rearrangement and N-Boc deprotection followed by lactamiza-
tion enantiomerically pure pentacyclic lactam (ꢀ)-40 (43 mg, 176 mmol,
31%, quant. rel. to d.r.) was obtained with >99% ee when starting from
the enantiomerically pure allylic alcohol (+)-32. [a]2D0 =ꢀ18.2 (c=1.0 in
CH2Cl2); chiral HPLC (Chiralpak AD-H, n-hexane/iPrOH 80:20): tR =
28.6 min, k=7.5; rac: tR =27.3/29.4 min, k=7.0/7.6, a=1.09.
¯
group P1 (no. 2); a=9.7252(11), b=9.8423(3), c=11.7678(15) ꢄ; a=
76.621(6), b=70.268(11), g=72.021(6)8; V=998.3(2) ꢄ3; Z=2; m-
ACHTUNGTRENNUNG
(MoKa)=1.627 mmꢀ1; 1calcd =1.591 gcmꢀ3; q range=2.67–25.368; data col-
lected: 21917; independent data [Io >2s(Io)/all data/Rint]: 3327/3640/
0.016; data/restraints/parameters: 3640/0/246; R1 [Io >2s(Io)/all data]:
0.0268/0.0308; wR2 [Io >2s(Io)/all data]: 0.0601/0.0622; GOF=1.054;
(ꢀ)-(1S,2aS,8bR,11aS)-2-[11-Allyl-3-oxo-1-vinyl-2,2a,3,4,9,10,11,11a-octa-
hydro-1H-pyrrolo[3’,2’:2,3]cyclopentaACTHNUGRTNEUNG[1,2-c]quinoline-1-yl]methylacetate
D1max/min: 0.81/ꢀ0.75 eꢄꢀ3
.
(42): Rf =0.55 (EtOAc, [UV]); 1H NMR (500 MHz, CDCl3): d=8.60
3
3
CCDC 717011 (33) contains the supplementary crystallographic data for
this paper. These data can be obtained free of charge from The Cam-
bridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
(brs, 1H, NH), 7.30 (d, J=7.5 Hz, 1H, H8), 7.17 (t, J=7.5 Hz, 1H, H6),
7.07 (t, 3J=7.6 Hz, 1H, H7), 6.78 (d, 3J=7.9 Hz, 1H, H5), 6.27 (dd, 3JE =
3
3
3
3
17.9, JZ =10.9 Hz, 1H, =CHVin), 5.94 (dddd, JE =16.9, JZ =10.2, J=6.6,
3J=5.9 Hz, 1H, =CHAll), 5.26 (dd, 3J=16.9, 2/4J=1.1 Hz, 1H, =CHHAll),
5.20–5.11 (m, 3H, =CHHAll, =CH2Vin), 3.56 (s, 1H, CHN), 3.57–3.47 (m,
1H, NCHH), 3.48 (s, 3H, OCH3), 3.22–3.12 (m, 2H, NCHH, CHHN),
2.94 (dd, 3J=10.6, 3J=6.1 Hz, 1H, C2aH), 2.83 (virt. dt, 2J=10.8, 3J=
(ꢀ)-(6bR,11aR,11bS,12aS)-11a-Vinyl-7,8,11a,11b,12,12a-hexahydropyrroli-
zino[1’,7’:2,3,4]cyclopentaACTHNUTRGENNU[G 1,2-c]quinoline-1,10ACHTUNGTNER(NUGN 2H,11H)-dione (40)
Johnson–Claisen rearrangement of allylic alcohol 32 with trimethylor-
thoacetate (39): Allylic alcohol 32 (5.00 g, 13.5 mmol) and hydroquinone
(1.00 g, 20 wt%, 9.08 mmol, 0.67 equiv) in trimethylorthoacetate
(20.0 mL, 18.9 g, 157 mmol, 12 equiv) were heated to 1308C for 4.0 h.
The mixture was concentrated to about 4 mL in vacuum yielding a thick
orange oil, which was stirred at 1308C for an aditional 16 h. After evapo-
ration to dryness, the residue was purified by flash chromatography (5.0ꢅ
25 cm, P/EtOAc 2:1!1:2) to give rearrangement product 39 (4.87 g,
11.4 mmol, 85%) as an inseparable mixture of diastereoisomers (d.r.
70:30) which was used directly in the next step. Rf =0.36 (P/EtOAc 1:2,
[UV]); MS (ESI): m/z: 449 [M+Na+], 659 [(3M+H+ +K+)/2+], 875
[2M+Na+], 1005 [2M+HCOOH+HCOONa+Na+], 1141 [2M+
2
3
6.7 Hz, 1H, CHHN), 2.53 (dd, J=13.8, J=6.1 Hz, 1H, CHHexo), 2.49 (d,
2
2J=14.2 Hz, 1H, CHHCOO), 2.30 (d, J=14.2 Hz, 1H, CHHCOO), 2.04
(dd, 2J=13.8, 3J=10.6 Hz, 1H, CHendoH), 1.96 (virt. dt, 2J=12.8, 3J=
6.7 Hz, 1H, RCHH), 1.81 ppm (virt. dt, 2J=12.8, 3J=6.7 Hz, 1H,
RCHH); 13C NMR (90 MHz, CDCl3): d=171.7 (s, CO), 171.5 (s, CO),
140.3 (d, =CHVin), 136.2 (d, =CHAll), 134.8 (s, C4a), 129.1 (s, C8a), 127.6 (d,
C8H), 127.5 (d, C6H), 123.9 (d, C7H), 117.0 (t, =CH2All), 115.7 (d, C5H),
112.2 (t, =CH2Vin), 87.2 (d, CHN), 58.8 (t, NCH2), 57.9 (s, C8b), 54.8 (t,
CH2N), 51.2 (q, OCH3), 49.5 (d, C2aH), 48.3 (s, C1), 44.5 (t, CH2COO),
42.1 (t, C2HH), 41.6 ppm (t, RCH2); IR (ATR, neat): n˜ =3200 (w, NH),
3074 (w, Car/olefH), 2949 (m, CalH), 1734 (s, COOEt), 1671 (vs, CONH),
Chem. Eur. J. 2009, 15, 3509 – 3525
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3521

T. Bach et al.
1591 (s, Car), 1488 (s), 1110 (m, br.), 915 (s), 753 cmꢀ1 (s, Car); MS
(70 eV): m/z (%): 366 (8) [M+], 335 (10) [MꢀOMe+], 307 (7)
[MꢀCOOMe+], 293 (10) [MꢀCH2COOMe+], 240 (50), 239 (81), 199
(100) [C13H16N2O+], 167 (15), 149 (20), 57 (14) [tBu+], 41 (21) [C3H5+];
HRMS (70 eV): m/z: calcd for C22H26N2O3: 366.1943 [M+], found
366.1940; elemental analysis calcd (%) for C22H26N2O3: C 72.11, H 7.15;
found: C 71.99, H 7.09.
1H, NCHH), 3.25–3.13 (m, 2H, NCHH, CHHN), 3.06 (virt. t, 3J=9.4 Hz,
1H, C3H), 2.94 (brvirt. dt, Jt =7.9, Jd =5.3 Hz, 1H, CHHN), 2.52 (dd,
2J=13.2, 3J=8.6 Hz, 1H, CHHexo), 2.28 (d, 3J=2.1 Hz, 2H, CH2CHO),
2.19 (ddd, 2J=13.0, 3J=7.0, 3J=5.2 Hz, 1H, RCHH), 2.02–1.91 ppm (m,
2H, CHendoH, RCHH); 13C NMR (90 MHz, CDCl3): d=200.9 (d, CHO),
171.9 (s, CONH), 135.2 (s, C18), 132.3 (d, =CHC), 128.0 (d, C16H), 127.5
(d, C14H), 126.7 (d, CH2CH=), 126.4 (s, C13), 124.3 (d, C15H), 115.8 (d,
C17H), 80.2 (d, CHN), 57.0 (s, C12), 53.4 (t, CH2CHO), 52.8 (t, CH2N),
50.6 (d, C3H), 46.9 (t, NCH2), 44.7 (t, C4HH), 43.1 (s, C5), 40.8 ppm (t,
RCH2); IR (ATR, neat): n˜ =3186 (w, NH), 3055 (w, Car/olefH), 2924 (m,
CalH), 1717 (s, CHO), 1668 (vs, CONH), 1593 (s, Car), 1492 (s), 1389 (s),
1250 (m), 756 (m, Car), 727 cmꢀ1 (w); MS (70 eV): m/z (%): 308 (16) [M+],
279 (13) [MꢀCHO+], 264 (100) [MꢀCH2CH2O+], 235 (12), 84 (44), 59
(23), 41 (12); HRMS (70 eV): m/z: calcd for C19H20N2O2: 308.1525 [M+],
found 308.1517.
Enantiopure: According to the above sequence consisting of Johnson–
Claisen rearrangement and N-alkylation, enantiomerically pure allyl-
amine (ꢀ)-42 was obtained with >99% ee when starting from the enan-
tiomerically pure allylic alcohol (+)-32. [a]2D0 =ꢀ36.2 (c=1.0 in CH2Cl2);
chiral HPLC (Chiralpak AD-H, n-hexane/iPrOH 90:10): tR =18.2 min,
k=4.5; rac: tR =15.6/19.6 min, k=3.5/4.6, a=1.3.
(+)-(6bR,12aS,12bS,13aS)-2-[1-Oxo-1,2,7,8,13,13a-hexahydro-10H-indol-
izino[1’,8’:2,3,4]cyclopentaACTHNUTRGENNU[G 1,2-c]quinoline-12HCATUNGTREN(NGUN 12bH)-yl]methylacetate
(+)-(6bR,12aS,12bS,13aS)-12a-(2-Hydroxyethyl)-7,8,12a,12b,13,13a-hexa-
hydro-10H-indolizino[1’,8’:2,3,4]cyclopentaACTHNUTRGENUG[N 1,2-c]quinolin-1(2H)-one
(44): Allylamine 42 (2.15 g, 5.87 mmol) and second generation Grubbs
catalyst (249 mg, 294 mmol, 5 mol%) in PhMe (1.47 L) were heated to
608C. After 2.5 h, a second load of Grubbs catalyst (249 mg, 294 mmol,
5 mol%) was added and the solution stirred at 608C for 6.0 h. After
evaporation of the solvent, the residue was purified by flash chromatog-
raphy (5.0ꢅ20 cm CH2Cl2/MeOH 40:1!20:1) to give recovered starting
material 42 (264 mg, 720 mmol, 12%) as the unpolar, and melodan ester
44 (1.65 g, 4.88 mmol, 83%, 95% b.r.s.m.) as the polar fraction as brown
to violet oils. Attempts to remove colored Ru contaminants by chroma-
tography were ineffective. Ester 44 could however be converted to the
equal amount of a light brown solid by precipitation with Et2O (200 mL).
Rf =0.50 (CH2Cl2/MeOH 10:1, [UV]); 1H NMR (360 MHz, CDCl3): d=
(46): Melodan ester 44 (200 mg, 591 mmol) was dissolved in dry CH2Cl2
(50 mL) and cooled to ꢀ708C. DIBAL-H (1.1m in cyclohexane, 1.13 mL,
1.24 mmol, 2.1 equiv) was added slowly and the solution stirred for 1.5 h
at ꢀ708C. The reaction was quenched with MeOH (5 mL) and the sol-
vent evaporated. The solid residue was suspended in dry EtOH (50 mL)
and cooled to 08C. Fresh NaBH4 (26.8 mg, 709 mmol, 1.2 equiv) was
added and the mixture stirred at 08C for 25 min. The reaction was
quenched with aqueous HCl (10%, 5 mL) and the solvent removed in
vacuum. The residue was dissolved in H2O (30 mL) and CH2Cl2 (30 mL),
the solution made basic with aqueous NaOH (2m), phases were separat-
ed and the aqueous phase was extracted with CH2Cl2 (5ꢅ30 mL). The
combined organic phases were dried over Na2SO4, filtered and concen-
trated to give melodan alcohol 46 (172 mg, 554 mmol, 94%) as a slightly
gray foam of sufficient purity (ca. 90% NMR). Attempted additional pu-
rification by flash chromatography was ineffective due to the very high
polarity of the product alcohol. Rf =0.28 (CH2Cl2/MeOH 8:2, [UV]);
1H NMR (500 MHz, CDCl3; c ꢁ 10 mgmLꢀ1): d=8.47 (brs, 1H, NH),
7.41 (brd, 3J=7.0 Hz, 1H, H14), 7.16 (dt, 3J=7.6, 4J=1.4 Hz, 1H, H16),
7.06 (dt, 3J=7.5, 4J=1.2 Hz, 1H, H15), 6.74 (dd, 3J=7.9, 4J=1.4 Hz, 1H,
H17), 5.96 (ddd, 3JZ =9.9, 3J=5.3, 3J=2.3 Hz, 1H, CH2CH=), 5.77 (dt,
3
3
9.62 (brs, 1H, NH), 7.39 (d, J=7.5 Hz, 1H, H14), 7.15 (t, J=7.2 Hz, 1H,
H16), 7.05 (t, 3J=7.4 Hz, 1H, H15), 6.84 (d, 3J=7.7 Hz, 1H, H17), 5.95
3
3
3
3
4
(ddd, JZ =9.8, J=5.3, J=1.6 Hz, 1H, CH2CH=), 5.86 (dd, JZ =9.8, J=
1.9 Hz, 1H, =CHC), 3.58 (s, 1H, CHN), 3.50 (s, 3H, OCH3), 3.26 (dd,
2J=16.3, 3J=5.3 Hz, 1H, NCHH), 3.21–3.08 (m, 2H, NCHH, CHHN),
3.01 (dd, 3J=10.3, 3J=8.6 Hz, 1H, C3H), 2.86 (virt. dt, Jt =8.3, Jd =
4.8 Hz, 1H, CHHN), 2.43 (dd, 2J=13.0, 3J=8.6 Hz, 1H, CHHexo), 2.25–
2.10 (m, 3H, CH2COO, RCHH), 2.03 (dd, 2J=13.0, 3J=10.3 Hz, 1H,
2
3
CHendoH), 1.89 ppm (virt. dt, J=13.0, J=7.3 Hz, 1H, RCHH); 13C NMR
(90 MHz, CDCl3): d=172.6 (s, CONH), 171.3 (s, COOMe), 135.4 (s, C18),
132.4 (d, =CHC), 127.7 (d, C16H), 127.5 (d, C14H), 127.1 (d, CH2CH=),
126.9 (s, C13), 123.9 (d, C15H), 115.8 (d, C17H), 80.0 (d, CHN), 56.9 (s,
C12), 52.7 (t, CH2N), 51.2 (q, OCH3), 50.4 (d, C3H), 46.7 (t, NCH2), 44.6
(t, CH2COO), 44.3 (t, C4HH), 43.2 (s, C5), 40.9 ppm (t, RCH2); IR (ATR,
neat): n˜ =3194 (w, NH), 3059 (w, Car/olefH), 2958 (m, CalH), 2929 (m,
CalH), 1727 (s, COOMe), 1671 (vs, CONH), 1594 (m, Car), 1494 (s), 1432
(m), 1391 (s), 1163 (m), 1113 (m), 770 cmꢀ1 (m, Car); MS (70 eV): m/z
(%): 338 (100) [M+], 323 (5) [MꢀMe+], 307 (13) [MꢀOMe+], 279 (19)
[MꢀCOOMe+], 265 (56) [MꢀCH2COOMe+], 264 (49), 214 (43), 205
(42), 180 (27), 132 (27), 91 (28) [C7H7+], 57 (24) [tBu+], 43 (15) [C3H5+];
HRMS (70 eV): m/z: calcd for C20H22N2O3: 338.1631 [M+], found
338.1621.
4
3JZ =9.9, J=1.9 Hz, 1H, =CHC), 3.60–3.47 (m, 3H, CHN, CH2OH), 3.26
(dd, 2J=16.2, 3J=5.3 Hz, 1H, NCHH), 3.21–3.13 (m, 2H, NCHH,
3
3
2
3
CHHN), 2.92 (dd, J=10.1, J=8.3 Hz, 1H, C3H), 2.88 (ddd, J=8.8, J=
8.2, 3J=4.8 Hz, 1H, CHHN), 2.29 (dd, 2J=12.7, 3J=8.3 Hz, 1H,
CHHexo), 2.12 (ddd, 2J=12.8, 3J=6.6, 3J=4.8 Hz, 1H, RCHH), 1.96–1.88
(m, 2H, RCHH, CHendoH), 1.51 ppm (virt. dt, 3Jt =7.0, Jd =3.3 Hz, 2H,
1
CCH2); the H NMR spectrum is very strongly dependent on the concen-
tration, giving a distinctly different set of signals at c ꢁ 50 mgmLꢀ1 due
to intra- and intermolecular hydrogen bonding; 13C NMR (90 MHz,
CDCl3): d=171.7 (s, CONH), 135.0 (s, C18), 133.7 (d, =CHC), 127.8 (d,
C16H), 127.7 (d, C14H), 127.3 (s, C13), 126.4 (d, CH2CH=), 124.2 (d, C15H),
115.5 (d, C17H), 80.3 (d, CHN), 59.9 (t, CH2OH), 57.1 (s, C12), 53.1 (t,
CH2N), 50.2 (d, C3H), 46.5 (t, NCH2), 45.1 (t, C4HH), 43.7 (s, C5), 43.6 (t,
CCH2), 41.1 (t, RCH2); IR (ATR, neat): n˜ =3350 (m, vbr, OH), 3213 (w,
NH), 3061 (w, Car/olefH), 2926 (m, CalH), 1665 (vs, CONH), 1592 (s, Car),
1488 (m), 1383 (s), 1242 (m), 1046 (m), 755 cmꢀ1 (s, Car); MS (70 eV): m/
z (%): 310 (100) [M+], 296 (17), 279 (33) [MꢀCH2OH+], 266 (52), 172
(32), 154 (61), 152 (52), 108 (33), 45 (23) [CH2CH2OH+]; HRMS
(70 eV): m/z: calcd for C19H22N2O2: 310.1681 [M+], found 310.1670.
Enantiopure: According to the above procedure, enantiomerically pure
melodan ester (+)-44 was obtained with >99% ee when starting from
the enantiomerically pure allyl amine (ꢀ)-42. [a]2D0 =+147 (c=0.1 in
CH2Cl2); chiral HPLC (Chiralpak AD-H, n-hexane/iPrOH 80:20): tR =
12.1 min, k=2.5; rac: tR =13.6/15.2 min, k=2.3/2.7, a=1.17.
(ꢂ)-(6bR,12aS,12bS,13aS)-2-[1-Oxo-1,2,7,8,13,13a-hexahydro-10H-indol-
izino[1’,8’:2,3,4]cyclopentaACTHNUTRGNENG[U 1,2-c]quinoline-12ACHTUNGTRENN(UGN 12bH)-yl]acetaldehyde
Enantiopure: According to the above procedure, enantiomerically pure
melodan alcohol (+)-46 (98 mg, 316 mmol, 58% after chromatography)
was obtained with >99% ee when starting from the enantiomerically
pure melodan ester (+)-44. [a]2D0 =+76 (c=0.1 in CH2Cl2); chiral HPLC
(Chiralpak AD-H, n-hexane/iPrOH 80:20): tR =12.0 min, k=2.5; rac:
tR =11.8/19.4 min, k=2.6/4.9, a=1.9.
(45): Melodan ester 44 (41 mg, 121 mmol) was dissolved in dry CH2Cl2
(10 mL) and cooled to ꢀ708C. DIBAL-H (1.1m in cyclohexane, 260 mL,
286 mmol, 2.4 equiv) was added and the solution stirred for 2.0 h at
ꢀ708C. The reaction was quenched with a few drops of MeOH and the
solvent evaporated. Purification of the solid residue by flash chromatog-
raphy (2.5ꢅ10 cm, CH2Cl2/MeOH 15:1!10:1) gave melodan aldehyde 45
(28 mg, 91 mmol, 75%) as a white foam. Rf =0.46 (CH2Cl2/MeOH 8:2,
[UV]); 1H NMR (360 MHz, CDCl3): d=9.55 (t, 3J=2.1 Hz, 1H, CHO),
(+)-(6bR,12aS,12bS,13aS)-12a-(2-Tosylethyl)-7,8,12a,12b,13,13a-hexahy-
dro-10H-indolizino[1’,8’:2,3,4]cyclopentaACTHNUGRTENUNG[1,2-c]quinoline-1(2H)-one (47):
Melodan alcohol 46 (115 mg, 370 mmol), NEt3 (180 mL, 131 mg,
1.30 mmol, 3.5 equiv) and TosCl (177 mg, 926 mmol, 2.5 equiv) in CH2Cl2
(5.0 mL) were stirred for 18 h at RT. Sat. aqueous NaHCO3 (50 mL) and
CH2Cl2 (50 mL) were added, phases separated and the aqueous phase ex-
3
3
9.15 (brs, 1H, NH), 7.42 (d, J=7.7 Hz, 1H, H14), 7.18 (t, J=7.6 Hz, 1H,
H16), 7.07 (t, 3J=7.5 Hz, 1H, H15), 6.82 (d, 3J=7.9 Hz, 1H, H17), 5.98
3
3
3
3
4
(ddd, JZ =9.9, J=5.4, J=1.8 Hz, 1H, CH2CH=), 5.89 (dd, JZ =9.9, J=
2.0 Hz, 1H, =CHC), 3.49 (brs, 1H, CHN), 3.35 (dd, 2J=16.6, 3J=5.4 Hz,
3522
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3509 – 3525

Total Synthesis of Meloscine
FULL PAPER
tracted with CH2Cl2 (2ꢅ25 mL). The combined organic phases were
dried over Na2SO4, filtered and concentrated. Purification of the residue
by flash chromatography (2.5ꢅ15 cm, CH2Cl2/MeOH 40:1!20:1) gave
melodan tosylate 47 (123 mg, 265 mmol, 72%) as a light brown solid. In a
three-step sequence with no purification of the intermediates, melodan
ester 44 (338 mg, 1.00 mmol) could be converted into melodan tosylate
47 (243 mg, 523 mmol, 52%) while avoiding the difficulties of the isola-
tion of the intermediate alcohol. Rf =0.45 (CH2Cl2/MeOH 10:1, [UV]);
m.p. 175–1788C (MeOH); 1H NMR (360 MHz, CDCl3): d=9.05 (brs,
1H, NH), 7.68 (d, 3J=8.4 Hz, 2H, H2’), 7.37–2.28 (m, 3H, H14, H3’), 7.15
(dt, 3J=7.6, 4J=1.3 Hz, 1H, H16), 7.03 (dt, 3J=7.5, 4J=1.1 Hz, 1H, H15),
6.79 (dd, 3J=7.8, 4J=1.0 Hz, 1H, H17), 5.93 (ddd, 3JZ =9.9, 3J=5.6, 3J=
2.0 Hz, 1H, CH2CH=), 5.77 (dd, 3JZ =9.9, 4J=2.0 Hz, 1H, =CHC), 3.97–
3.85 (m, 2H, CH2OTos), 3.36 (s, 1H, CHN), 3.24 (dd, 2J=16.3, 3J=
5.6 Hz, 1H, NCHH), 3.13 (ddd, J=8.6, J=7.2, J=6.8 Hz, 1H, CHHN),
3.03 (virt. dt, 2J=16.3, 3J=2.0 Hz, 1H, NCHH), 2.93 (dd, 3J=10.2, 3J=
8.5 Hz, 1H, C3H), 2.83 (ddd, 2J=8.6, 3J=7.9, 3J=4.8 Hz, 1H, CHHN),
2.43 (s, 3H, ArCH3), 2.28 (dd, 2J=12.9, 3J=8.5 Hz, 1H, CHHexo), 2.11
(ddd, 2J=12.7, 3J=6.8, 3J=4.8 Hz, 1H, RCHH), 1.86 (ddd, 2J=12.7, 3J=
7.9, 3J=7.2 Hz, 1H, RCHH), 1.79 (dd, 2J=12.9, 3J=10.2 Hz, 1H,
CHendoH), 1.65–1.45 ppm (m, 2H, CCH2); 13C NMR (90 MHz, CDCl3):
d=172.1 (s, CONH), 144.7 (s, C1’), 135.2 (s, C18), 133.0 (s, C4’), 132.7 (d, =
CHC), 129.8 (d, C3’H), 127.83 (d, CH2CH=), 127.79 (d, C2’H), 127.76 (d,
C16H), 127.4 (d, C14H), 127.0 (s, C13), 124.1 (d, C15H), 115.7 (d, C17H),
80.1 (d, CHN), 67.6 (t, CH2OTos), 56.9 (s, C12), 52.8 (t, CH2N), 50.3 (d,
C3H), 46.8 (t, NCH2), 45.0 (t, C4HH), 43.5 (s, C5), 41.1 (t, RCH2), 39.4 (t,
CCH2), 21.6 ppm (q, ArCH3); IR (ATR, neat): n˜ =3203 (w, NH), 3060
(w, Car/olefH), 2921 (w, CalH), 1669 (vs, CONH), 1593 (s, Car), 1489 (m),
1360 (s, OTos), 1174 (vs, OTos), 953 (m, br.), 813 (w), 754 cmꢀ1 (s, Car);
MS (70 eV): m/z (%): 464 (14) [M+], 309 (100) [MꢀTos +], 292 (7)
[MꢀTosOH+], 265 (10), 172 (10), 134 (13), 120 (10), 91 (10) [C7H7+];
HRMS (70 eV): m/z: calcd for C26H28N2O4S: 464.1770 [M+], found
464.1797.
135.2 (s, C18), 133.5 (d, C5’H), 133.3 (s, C1’), 132.6 (d, =CHC), 128.7 (d, C4’/
6’H), 127.82 (d, CH2CH=), 127.77 (d, C16H), 127.3 (d, C14H), 127.2 (s,
C13), 126.4 (d, C3’H), 125.3 (d, C4’/6’H), 124.0 (d, C15H), 115.8 (d, C17H),
79.9 (d, CHN), 57.0 (s, C12), 53.0 (t, CH2N), 50.3 (d, C3H), 47.0 (t,
NCH2), 45.8 (s, C5), 44.5 (t, C3HH), 41.4 (t, RCH2), 39.5 (t, CCH2),
21.3 ppm (t, CH2SeAr); IR (ATR, neat): n˜ =3194 (w, NH), 3052 (w, Car/
olefH), 2925 (w, CalH), 1668 (vs, CONH), 1591 (s), 1508 (vs, NO2), 1330 (s,
NO2), 1038 (m), 848 (m, NO2), 756 (s, Car), 724/700/670 cmꢀ1 (s, ArNO2);
MS (70 eV): m/z (%): 495 (23) [M
[M
(80Se)ꢀOH+], 476 (39) [M(78Se)ꢀOH+], 373 (40) [M
(19) [M
(80Se)+], 493 (14) [M
ACHUTGTNRENNUG CAHTUNGTRENNUNG
G
E
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
], 41 (17) [C3H7+]; HRMS (70 eV): m/z: calcd for C25H25N3O3Se:
495.1061 [M+], found 495.1051.
Enantiopure: According to the above procedure, enantiomerically pure
melodan selenide (+)-48 was obtained with >99% ee when starting from
the enantiomerically pure melodan tosylate (+)-47. [a]2D0 =+167 (c=0.1
in CH2Cl2); chiral HPLC (Chiralpak AD-H, n-hexane/iPrOH 80:20): tR =
11.2 min, k=2.4; rac: tR =12.3/18.7 min, k=2.7/4.7, a=1.74.
2
3
3
(+)-(6bR,12aS,12bS,13aS)-12a-Vinyl-7,8,12a,12b,13,13a-hexahydro-10H-
indolizino[1’,8’:2,3,4]cyclopentaACTHNUTRGNEU[GN 1,2-c]quinoline-1(2H)-one, (+)-meloscine
(1): Melodan selenide 48 (45 mg, 91 mmol) was dissolved in CH2Cl2
(10 mL), TFA (10.2 mL, 15.6 mg, 137 mmol, 1.5 equiv) was added and the
solution cooled to ꢀ788C. A solution of MCPBA (20.5 mg, 77% aq.,
15.8 mg, 91 mmol, 1.00 equiv) in CH2Cl2 (3.0 mL) was added and the mix-
ture stirred for 1.5 h at ꢀ788C. The reaction was quenched with Me2S
(0.32 mL, 283 mg, 4.55 mmol, 50 equiv) and NEt3 (0.38 mL, 276 mg,
2.73 mmol, 30 equiv) and the reaction mixture warmed to RT, at which it
was stirred for an additional 2.0 h. Sat. aqueous NaHCO3 (20 mL) and
CH2Cl2 (20 mL) were added, phases separated and the aqueous phase ex-
tracted with CH2Cl2 (3ꢅ25 mL). The combined organic phases were
dried over Na2SO4, filtered and concentrated. Purification of the crude
product by flash chromatography (1.0ꢅ15 cm, CH2Cl2/MeOH 50:1!
20:1) gave (ꢂ)-meloscine (1) (23.0 mg, 78.7 mmol, 86%) as an off-white
solid. Rf =0.33 (CH2Cl2/MeOH 10:1, [UV]); m.p. 216–2208C (Et2O/
Enantiopure: According to the above procedure, enantiomerically pure
melodan tosylate (+)-47 was obtained with >99% ee when starting from
the enantiomerically pure melodan alcohol (+)-46 or melodan ester (+)-
44. [a]2D0 =+58.5 (c=1.0 in CH2Cl2); chiral HPLC (Chiralpak AD-H, n-
hexane/iPrOH 80:20): tR =30.7 min, k=8.0; rac: tR =30.7/38.7 min, k=
8.3/10.7, a=1.3.
1
MeOH); H NMR (360 MHz, CDCl3): d=9.18 (brs, 1H, NH), 7.38 (brd,
3
4
3
3J=7.5 Hz, 1H, H14), 7.14 (td, J=7.6, J=1.3 Hz, 1H, H16), 7.04 (td, J=
7.5, 4J=1.4 Hz, 1H, H15), 6.79 (dd, 3J=7.8, 4J=1.0 Hz, 1H, H17), 6.01
(ddd, 3JZ =9.8, 3J=5.6, 3J=2.2 Hz, 1H, NCH2CH=), 5.73 (dd, 3JZ =9.8,
4J=2.2 Hz, 1H, =CHC5), 5.54 (dd, 3JE =17.3, 3JZ =10.6 Hz, 1H, CH=
CH2), 4.90 (dd, 3JE =17.3, 2J=0.5 Hz, 1H, =CHHpro-Z), 4.78 (dd, 3JZ =
10.6, 2J=0.5 Hz, 1H, =CHHpro-E), 3.55 (brs, 1H, CHN), 3.30 (dd, 2J=
(+)-(6bR,12aS,12bS,13aS)-12a-2-[(2-Nitrophenyl)selanyl]ethyl-7,8,12a,
12b,13,13a-hexahydro-10H-indolizino[1’,8’:2,3,4]cyclopenta
ACHTUNGTREN[NUNG 1,2-c]quino-
line-1(2H)-one (48): Melodan tosylate 47 (55 mg, 118 mmol) was dis-
ACHTUNGTRENNUNG
3
3
3
15.1, J=5.6 Hz, 1H, NCHHi), 3.20 (ddd, 2J=8.3, J=7.7, J=7.1 Hz, 1H,
solved in dry EtOH (5.0 mL) and treated with the deep red solution of
selenid anion (5.8 equiv) prepared from 2-nitrophenylselenocyanate
(166 mg, 731 mmol, 6.2 equiv) and NaBH4 (25.9 mg, 684 mmol, 5.8 equiv
NaBH4) in dry EtOH (3.0 mL) at RT. After 3.5 h an additional 5.2 equiv
of selenid anion were added and the solution stirred at RT for 18 h.
Again, 5.4 equiv of selenid anion were added and after 4.0 h a final load
of 5.4 equiv of selenid was added. The solution was stirred for an addi-
tional 32 h at RT during which time a light brown suspension was
formed. Sat. aqueous NaHCO3 (50 mL) was added and the mixture ex-
tracted with CH2Cl2 (4ꢅ40 mL). The combined organic phases were
dried over Na2SO4, filtered and concentrated. Purification of the residue
by flash chromatography (2.5ꢅ20 cm, CH2Cl2/MeOH 50:1!20:1) gave
the melodan nitrophenylselenide 48 (57 mg, 115 mmol, 98%) as a brightly
yellow foam containing no more than 5% (NMR) of the residual tosylate
substrate. Rf =0.55 (CH2Cl2/MeOH 10:1, [UV]); 1H NMR (360 MHz,
CDCl3): d=9.36 (brs, 1H, NH), 8.22 (dd, 3J=8.5, 4J=1.2 Hz, 1H, H3’),
7.42 (dt, 3J=7.7, 4J=1.3 Hz, 1H, H5’), 7.40 (brd, 3J=8.4 Hz, 1H, H14),
7.33–7.23 (m, 2H, H4’, H6’), 7.15 (dt, 3J=7.6, 4J=1.3 Hz, 1H, H16), 7.04
(dt, 3J=7.6, 4J=1.3 Hz, 1H, H15), 6.81 (dd, 3J=7.8, 4J=1.0 Hz, 1H, H17),
2
3
4
CHHiN), 3.15 (ddd, J=15.1, J=2.2, J=2.2 Hz, 1H, NCHaH), 2.95 (dd,
3
2
3
3
3J=9.6, J=8.3 Hz, 1H, C3H), 2.89 (ddd, J=8.3, J=7.8, J=4.5 Hz, 1H,
CHaHN), 2.31 (dd, 2J=12.7, 3J=8.3 Hz, 1H, CHHexo), 2.17 (dd, 2J=12.7,
3J=9.6 Hz, 1H, CHendoH), 2.11 (ddd, 2J=12.8, 3J=7.1, 3J=4.5 Hz, 1H,
RCHHi), 1.95 ppm (ddd, 2J=12.8, 3J=7.8, 3J=4.5 Hz, 1H, RCHaH);
13C NMR (90 MHz, CDCl3): d=172.4 (s, CONH), 143.2 (d, CH=CH2),
135.2 (s, C18), 132.5 (d, =CHC5), 127.7 (d, C14H), 127.6 (d, C16H), 127.5 (s,
C13), 127.0 (d, CH=), 123.9 (d, C15H), 115.5 (d, C17H), 112.6 (t, CH=CH2),
82.3 (d, CHN), 56.9 (s, C12), 53.0 (t, CH2N), 50.8 (d, C3H), 48.0 (s, C5),
46.7 (t, NCH2CH=), 43.4 (t, C2HH), 42.0 ppm (t, RCH2); IR (ATR,
neat): n˜ =3188 (w, CONH), 3054 (w, Car/olefH), 2982 (m, CalH), 2926 (m,
CH2), 2770 (m), 1656 (vs, CONH), 1593 (s), 1501 (m), 1397 (s), 1253 (m),
930 (m), 871 (m), 755 (s, Car), 679 cmꢀ1 (m, Colef); UV/Vis (EtOH): lmax
(e)=253 (10000), shoulders at 280 (2000), 290 nm (1600), lmin (e)=
230 nm (5250 molꢀ1 m3 cmꢀ1); MS (70 eV): m/z (%): 292 (100) [M+], 277
(17), 264 (26) [MꢀC2H4+], 210 (23), 172 (24) [C11H10NO+], 159 (17)
[C10H9NO+], 134 (74) [C9H12N+], 120 (24) [C8H10N+], 91 (21) [C7H7+],
77 (26) [C6H5+], 57 (14) [tBu+], 41 (21) [C3H5+]; HRMS (70 eV): m/z:
calcd for C19H20N2O: 292.1576 [M+], found 292.1574.
3
3
3
3
6.07 (ddd, JZ =9.9, J=5.3, J=1.9 Hz, 1H, CH2CH=), 5.81 (dd, JZ =9.9,
4J=2.0 Hz, 1H, =CHC), 3.49 (brs, 1H, CHN), 3.31 (dd, 2J=16.8, 3J=
5.5 Hz, 1H, NCHH), 3.23–3.14 (m, 2H, NCHH, CHHN), 2.99 (dd, 3J=
9.4, 3J=7.8 Hz, 1H, C3H), 2.91 (ddd, 2J=8.1, 3J=8.1, 3J=4.7 Hz, 1H,
CHHN), 2.73 (dd, 3J=9.4, 3J=7.4 Hz, 2H, CH2SeAr), 2.35 (dd, 2J=12.8,
3J=7.8 Hz, 1H, CHHexo), 2.14 (ddd, 2J=12.7, 3J=6.6, 3J=4.7 Hz, 1H,
RCHH), 2.00–1.86 (m, 2H, RCHH, CHendoH), 1.68–1.57 ppm (m, 2H,
CCH2); 13C NMR (90 MHz, CDCl3): d=172.4 (s, CONH), 146.7 (s, C2’),
Enantiopure: According to the above procedure, enantiomerically pure
meloscine (+)-1 was obtained with >99% ee when starting from the
enantiomerically pure melodan selenide (+)-48. m.p. 172–1768C (Et2O);
[a]2D0 =+90.6 (c=0.5 in CH2Cl2); chiral HPLC (Chiralpak AD-H, n-
hexane/iPrOH 95:5, T=58C): tR =24.0 min, k=6.1; rac: tR =26.8/
31.5 min, k=6.9/8.3, a=1.2.
Chem. Eur. J. 2009, 15, 3509 – 3525
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3523

T. Bach et al.
[28] R. S. Lott, V. S. Chauhan, C. H. Stammer, J. Chem. Soc. Chem.
Commun. 1979, 495–496.
Acknowledgements
[29] T. Imanishi, K.-I. Miyashita, A. Nakai, M. Inoue, M. Hanaoka,
Chem. Pharm. Bull. 1982, 30, 1521–1524.
[30] S. Brandes, P. Selig, T. Bach, Synlett 2004, 2588–2590.
[31] X. Creary, P. A. Inocentio, T. L. Underiner, R. Kostromin, J. Org.
Chem. 1985, 50, 1932–1938.
Our research was supported by the Deutsche Forschungsgemeinschaft
(Schwerpunktprogramm 1179 Organokatalyse), by the Fonds der Chemi-
schen Industrie and by the Studienstiftung des Deutschen Volkes (PhD-
scholarship to P.S.). The help of Mr. Olaf Ackermann in obtaining chiral
HPLC data is gratefully acknowledged.
[32] For corresponding thermal [2+2]-cycloadditions see: a) C. De Cock,
S. Piettre, F. Lahousse, Z. Janousek, R. Merꢇyi, H. G. Viehe, Tetra-
hedron 1985, 41, 4183–4193; b) S. Kabanyane, A. Decken, C.-M.
Yu, G. M. Strunz, Can. J. Chem. 2000, 78, 270–274; c) Y. Horino, M.
Kimura, S. Tanaka, T. Okajima, Y. Tamaru, Chem. Eur. J. 2003, 9,
2419–2428.
[1] a) G. A. Cordell, Introduction to Alkaloids, Wiley Interscience, New
York, 1981; b) M. Hesse, Alkaloide, Helvetica Chimica Acta,
Zꢀrich, 2000.
[33] a) T. Hatsui, C. Nojima, H. Takeshita, Bull. Chem. Soc. Jpn. 1989,
62, 2932–2938; b) T. Hatsui, S.-y. Ikeda, H. Takeshita, Chem. Lett.
1992, 1891–1894; c) T. Hatsui, J.-J. Wang, H. Takeshita, Bull. Chem.
Soc. Jpn. 1995, 68, 2393–2399.
[2] L. F. Szabꢆ, ARKIVOK 2007, 7, 280–290.
[3] a) K. Bernauer, G. Englert, W. Vetter, Experientia 1965, 21, 374–
375; b) K. Bernauer, G. Englert, W. Vetter, E. Weiss, Helv. Chim.
Acta 1969, 52, 1886–1905; c) W. E. Oberhꢁnsli, Helv. Chim. Acta
1969, 52, 1905–1911.
[34] T.-L. Ho, Synth. Commun. 1979, 9, 665–668.
[35] S. Hanessian, Y. L. Bennai, Y. Leblanc, Heterocycles 1993, 35, 1411–
1424.
[36] S. Matsubara, M. Horiuchi, K. Takai, K. Utimoto, Chem. Lett. 1995,
250–260.
[37] a) S. Ma, X. Lu, Z. Li, J. Org. Chem. 1992, 57, 709–713; b) E. Piers,
J. Renaud, S. J. Rettig, Synthesis 1998, 590–602.
[38] D. P. Curran, M. J. Totleben, J. Am. Chem. Soc. 1992, 114, 6050–
6058.
[4] L.-W. Gou, Y.-L. Zhou, Phytochemistry 1993, 32, 563–566.
[5] G. Palmisano, B. Danieli, G. Lesma, R. Riva, S. Riva, F. Demartin,
N. Masciocchi, J. Org. Chem. 1984, 49, 4138–4143.
[6] G. Hugel, J. Lꢇvy, J. Org. Chem. 1984, 49, 3275–3277.
[7] G. Hugel, J. Lꢇvy, J. Org. Chem. 1986, 51, 1594–1595.
[8] a) S. H. Goh, A. R. M. Ali, Tetrahedron Lett. 1986, 27, 2501–2504;
b) S. H. Goh, A. R. M. Ali, W. H. Wong, Tetrahedron 1989, 45,
7899–7920.
[9] a) L. E. Overman, G. M. Robertson, A. J. Robichaud, J. Org. Chem.
1989, 54, 1236–1238; b) L. E. Overman, G. M. Robertson, A. J. Ro-
bichaud, J. Am. Chem. Soc. 1991, 113, 2598–2610.
[39] a) Y. Yamamoto, K. Maruyama, J. Am. Chem. Soc. 1978, 100, 3240–
3241; b) G. J. Leotta III, L. E. Overman, G. S. Welmakeer, J. Org.
Chem. 1994, 59, 1946.
[10] Quaternary Stereocenters: Challenge and Solutions for Organic Syn-
thesis (Eds.: J. Christoffers, A. Baro), Wiley-VCH, Weinheim, 2005.
[11] S. E. Denmark, J. J. Cottell, Adv. Synth. Catal. 2006, 348, 2397–2402.
[12] A. G. Schultz, M. Dai, Tetrahedron Lett. 1999, 40, 645–648.
[13] Examples: a) G. R. Evanega, D. L. Fabiny, J. Org. Chem. 1970, 35,
1757–1761; b) O. Buchardt, J. J. Christensen, N. Harrit, Acta Chem.
Scand. 1976, 30, 189–192; c) C. Kaneko, T. Naito, Heterocycles 1982,
19, 2183–2206; d) F. D. Lewis, G. D. Reddy, J. E. Elbert, B. E. Till-
berg, J. A. Meltzer, M. Kojima, J. Org. Chem. 1991, 56, 5311–5318.
[14] For reviews on the use of photochemistry in natural product synthe-
sis see: a) T. Bach, Synthesis 1998, 683–703; b) J. Iriondo-Alberdi,
M. F. Greaney, Eur. J. Org. Chem. 2007, 4801–4815; c) N. Hoff-
mann, Chem. Rev. 2008, 108, 1052–1103.
[15] For examples of enantioselective [2+2]-photocycloadditions in solu-
tion see: a) T. Bach, H. Bergmann, K. Harms, Angew. Chem. 2000,
112, 2391–2393; Angew. Chem. Int. Ed. 2000, 39, 2302–2304; b) T.
Bach, H. Bergmann, J. Am. Chem. Soc. 2000, 122, 11525–11526;
c) T. Bach, H. Bergmann, B. Grosch, K. Harms, J. Am. Chem. Soc.
2002, 124, 7982–7990.
[16] L. E. Overman, Pure Appl. Chem. 1994, 66, 1423–1430.
[17] P. J. Connolly, C. H. Heathcock, J. Org. Chem. 1985, 50, 4135–4144.
[18] S. Brandes, Ph.D. thesis, TU Mꢀnchen, 2004.
[19] P. Selig, T. Bach, Angew. Chem. 2008, 120, 5160–5162; Angew.
Chem. Int. Ed. 2008, 47, 5082–5084.
[20] P. Selig, T. Bach, J. Org. Chem. 2006, 71, 5662–5673.
[21] a) F. Galinovski, A. Wagner, R. Weiser, Monatsh. Chem. 1951, 82,
551–559; b) G. A. Swan, J. D. Wilcock, J. Chem. Soc. Perkin Trans. 1
1974, 885–891.
[22] N. M. Yoon, H. C. Brown, J. Am. Chem. Soc. 1968, 90, 2927–2938.
[23] Examples: a) M. J. Wanner, G.-J. Koomen, J. Org. Chem. 1995, 60,
5634–5637; b) R. K. Dieter, R. R. Sharma, J. Org. Chem. 1996, 61,
4180–4184; c) K. Tomooka, A. Nakazaki, T. Nakai, J. Am. Chem.
Soc. 2000, 122, 408–409; d) R. E. Gawley, G. Barolii, S. Madan, M.
Saverin, S. OꢃConnor, Tetrahedron Lett. 2004, 45, 1759–1761.
[24] T. W. Greene, P. G. M. Wuts, Protective Groups in Organic Synthesis,
4th ed., Wiley Interscience, New York, 2007.
[40] a) L. Capella, P. C. Montevecchi, Tetrahedron Lett. 1994, 35, 8445–
8448; b) M. Nishida, H. Hayashi, Y. Yamaura, E. Yanaginuma, O.
Yonemitsu, A. Nishida, N. Kawahara, Tetrahedron Lett. 1995, 36,
269–272; c) J. Cossy, L. Tresnard, D. G. Pardo, Eur. J. Org. Chem.
1999, 1925–1933.
[41] a) B. Patro, J. A. Murphy, Org. Lett. 2000, 2, 3599–3601; b) M. Mori,
M. Nakanishi, D. Kajishima, Y. Sato, J. Am. Chem. Soc. 2003, 125,
9801–9807.
[42] a) S.-I. Fukuzawa, M. Iida, A. Nakanishi, T. Fujinami, S. Sakai, J.
Chem. Soc. Chem. Commun. 1987, 920–921; b) G. A. Molander,
J. A. McKie, J. Org. Chem. 1992, 57, 3132–3139.
[43] A. Chesney, I. E. Markꢆ, Synth. Commun. 1990, 20, 3167–3180.
[44] M. Oelgemçller, C. Jung, J. Mattay, Pure Appl. Chem. 2007, 79,
1939–1947.
[45] The Claisen Rearrangement: Methods and Applications (Eds.: M.
Hiersemann, U. Nubbemeyer), Wiley-VCH, Weinheim, 2007.
[46] For applications of ring-closing metathesis in natural product synthe-
sis see: a) A. Gradillas, J. Pꢇrez-Castells, Angew. Chem. 2006, 118,
6232–6247; Angew. Chem. Int. Ed. 2006, 45, 6086–6101; b) A. Dei-
ters, S. F. Martin, Chem. Rev. 2004, 104, 2199–2238; c) K. C. Nico-
laou, P. G. Bulger, D. Sarlah, Angew. Chem. 2005, 117, 4564–4601;
Angew. Chem. Int. Ed. 2005, 44, 4490–4527.
[47] W. S. Johnson, L. Werthemann, W. R. Bartlett, T. J. Brocksom, T.-T.
Li, D. J. Faulkner, M. R. Petersen, J. Am. Chem. Soc. 1970, 92, 741–
743.
[48] a) R. S. Huber, G. B. Jones, J. Org. Chem. 1992, 57, 5778–5780;
b) G. B. Jones, R. S. Huber, S. Chau, Tetrahedron 1993, 49, 369–380.
[49] R. P. Lutz, Chem. Rev. 1984, 84, 205–247.
[50] D. Samain, C. Descoins, Bull. Soc. Chim. Fr. 1979, I–II, 71–76.
[51] M. Scholl, T. M. Trnka, J. P. Morgan, R. H. Grubbs, Tetrahedron
Lett. 1999, 40, 2247–2250.
[52] Examples: a) P. Eilbracht, E. Balß, M. Acker, Chem. Ber. 1985, 118,
825–839; b) T. Nozoye, Y. Shibanuma, T. Nakai, Y. Hatori, Chem.
Pharm. Bull. 1988, 36, 4980–4985; c) P. Phukan, M. Bauer, M. E.
Maier, Synthesis 2003, 1324–1328.
[53] a) W. J. Scott, J. E. McMurry, Acc. Chem. Res. 1998, 31, 47–54;
b) S. K. Pandey, A. E. Greene, J.-F. Poisson, J. Org. Chem. 2007, 72,
7769–7770.
[25] P. Selig, T. Bach, Synthesis 2008, 2177–2182.
[26] S. Sugasawa, T. Fujii, Chem. Pharm. Bull. 1958, 6, 587–590.
[27] S. C. Nigama, A. Mann, M. Taddei, C.-G. Wermuth, Synth.
Commun. 1989, 19, 3139–3142.
[54] K. B. Sharpless, M. W. Young, J. Org. Chem. 1975, 40, 947–949.
3524
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3509 – 3525

Total Synthesis of Meloscine
FULL PAPER
[55] M. Daudon, M. H. Meri, M. M. Plat, E. W. Hagamann, F. M. Schell,
E. Wenkert, J. Org. Chem. 1975, 40, 2838–2839.
[56] a) T. Bach, H. Bergmann, B. Grosch, K. Harms, E. Herdtweck, Syn-
thesis 2001, 1395–1405; b) Immobilized variant of 52: S. Breiten-
lechner, T. Bach, Angew. Chem. 2008, 120, 8075–8077; Angew.
Chem. Int. Ed. 2008, 47, 7957–7959.
[57] For applications of the chiral complexing agent apart from photocy-
cloaddition reactions see: a) T. Bach, T. Aechtner, B. Neumꢀller,
Chem. Eur. J. 2002, 8, 2464–2475; b) T. Bach, B. Grosch, T. Strass-
ner, E. Herdtweck, J. Org. Chem. 2003, 68, 1107–1116; c) B.
Grosch, C. N. Orlebar, E. Herdtweck, M. Kaneda, T Wada, Y.
Inoue, T. Bach, Chem. Eur. J. 2004, 10, 2179–2189; d) H. Bergmann,
B. Grosch, S. Sitterberg, T. Bach, J. Org. Chem. 2004, 69, 970–973;
e) T. Aechtner, M. Dressel, T. Bach, Angew. Chem. 2004, 116, 5974–
5976; Angew. Chem. Int. Ed. 2004, 43, 5849–5851; f) A. Bauer, T.
Bach, Tetrahedron: Asymmetry 2004, 15, 3799–3803; g) M. Dressel,
T. Bach, Org. Lett. 2006, 8, 3145–3147; h) M. Dressel, T. Aechtner,
T. Bach, Synthesis 2006, 2206–2214; i) P. Kapitꢈn, T. Bach, Synthesis
2008, 1559–1564.
[58] Recent review on template-mediated enantioselective photochemi-
cal reactions: C. Mꢀller, T. Bach, Aust. J. Chem. 2008, 61, 557–564.
Received: November 17, 2008
Published online: February 13, 2009
Chem. Eur. J. 2009, 15, 3509 – 3525
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3525