New approaches for the synthesis of erythrinan alkaloids

Article abstract of DOI:10.1039/b900189a

A concise asymmetric total synthesis of (+)-erysotramidine is described, using chiral base desymmetrisation of a meso-imide, N-acyliminium addition, retro-Diels-Alder cycloaddition and radical cyclisation as the key steps. A related route, starting from a cyclobutene-fused imide, was explored, and established a novel construction of the Erythrina alkaloid skeleton using a key ring-opening/ring-closing metathesis step. Completion of this synthesis was thwarted by problems with the removal of an unwanted vinylic side-chain. Complementary enantiospecific routes to Erythrina systems were explored, starting from (L)-malic acid. Some unexpected observations were made concerning the diastereocontrol in malic acid-derived N-acyliminium ion cyclisations, where changing the protecting group of the alcohol function from acetate to OTIPS resulted in a dramatic change in diastereocontrol. Products from these reactions could be transformed into known intermediates for natural alkaloids, and into (+)-demethoxyerythratidinone itself, by means of radical cyclisations or intramolecular aldol reactions as the key steps.

Full text of DOI:10.1039/b900189a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
New approaches for the synthesis of erythrinan alkaloids†
Fengzhi Zhang, Nigel S. Simpkins*‡ and Alexander J. Blake
Received 6th January 2009, Accepted 23rd February 2009
First published as an Advance Article on the web 24th March 2009
DOI: 10.1039/b900189a
A concise asymmetric total synthesis of (+)-erysotramidine is described, using chiral base
desymmetrisation of a meso-imide, N-acyliminium addition, retro-Diels–Alder cycloaddition and
radical cyclisation as the key steps. A related route, starting from a cyclobutene-fused imide, was
explored, and established a novel construction of the Erythrina alkaloid skeleton using a key
ring-opening/ring-closing metathesis step. Completion of this synthesis was thwarted by problems with
the removal of an unwanted vinylic side-chain. Complementary enantiospecific routes to Erythrina
systems were explored, starting from (L)-malic acid. Some unexpected observations were made
concerning the diastereocontrol in malic acid-derived N-acyliminium ion cyclisations, where changing
the protecting group of the alcohol function from acetate to OTIPS resulted in a dramatic change in
diastereocontrol. Products from these reactions could be transformed into known intermediates for
natural alkaloids, and into (+)-demethoxyerythratidinone itself, by means of radical cyclisations or
intramolecular aldol reactions as the key steps.
b-erythroidine 3, and alkenoids with only one double bond in the
Introduction
A-ring such as 3-demethoxyerythratidinone 2 and cocculolidine 4
(Fig. 2).3
The Erythrina alkaloids constitute a large group of compounds
isolated from the plants of the genus Erthyrina, which are widely
distributed in tropical and sub-tropical areas.1 These plants have
been widely used in folk medicine, and are known to produce
curare-like and hypnotic effects and to possess general CNS
activity. More specific pharmacological effects associated with
members of this alkaloid family include sedative, hypotensive and
neuromuscular blocking activities.2
Structurally, the Erythrina alkaloids are characterised by their
unique tetracyclic spiroamine framework (Fig. 1).
Fig. 1 Erythrina alkaloid skeleton.
Fig. 2 Structural classifications of Erythrina alkaloids.
According to the nature of the D-ring they are generally
classified into aromatic types such as erysotramidine 1 and
3-demethoxyerythratidinone 2, and non-aromatic types such as
b-erythroidine 3 and cocculolidine 4. According to the number
and position of the olefinic bonds they also may be subdivided into
dienoids with a conjugated diene unit such as erysotramidine 1 and
The Erythrina alkaloids have received considerable synthetic
attention in recent decades owing to their intriguing biological
activity and challenging characteristic polycondensed structures.3
Many different approaches to the core spirocyclic system present
in these alkaloids have been developed, and a number of syntheses
of members of this family, in racemic form, have been completed.
Recent examples include: synthesis of ( )-erythrocarine by Mori
et al.,4 using dienyne metathesis as the key step, synthesis of
( )-3-demethoxyerythratidinone by Padwa and Wang based on
furan intramolecular Diels–Alder chemistry,5 and the synthesis
of ( )-erysotramidine by Tu et al. using an N-acyliminium ion
cyclisation.6
The School of Chemistry, University of Nottingham, University Park,
Nottingham, NG7 2RD, U.K.
† Electronic supplementary information (ESI) available: Experimental
procedures and characterisation data. CCDC reference number 715824.
For ESI and crystallographic data in CIF or other electronic format see
DOI: 10.1039/b900189a
‡ Current address: School of Chemistry, University of Birmingham,
Edgbaston, Birmingham, B15 2TT U.K. Tel: +44 (0)121 4148905. E-mail:
n.simpkins@bham.ac.uk
By contrast, there are very few examples of asymmetric synthe-
ses of erythrinan alkaloids, and only very recently have research
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 1963–1979 | 1963
©

View Article Online
groups established enantioselective routes to compete with the
original efforts of Tsuda et al., in which L-DOPA derivatives
were converted into targets, including 1 and 2, by condensation
or Diels–Alder strategies.7 Notable efforts in this area include
the synthesis of 3-demethoxyerythratidinone 2 by Allin et al.,
using Meyer’s chiral lactam approach,8 a highly novel synthesis of
O-methylerysodienone using axially chiral biphenyl intermediates
by Matsumoto et al.,9 and access to (+)-b-erythroidine 3 using a
metathesis approach.10
Our own efforts in this area stemmed from our interest in access-
ing substituted polycyclic imides, amides, and derived substances
from reactions of symmetrical imides with chiral lithium amide
bases. In our previous work we established a new asymmetric
synthesis of (+)-erysotramidine 1 using this type of chiral base
approach, combined with a key N-acyliminium ion cyclisation,11
and later we accessed (+)-3-demethoxyerythratidinone 2, using
a chiral pool approach in which we probed some unexpected
diastereoselectivities in N-acyliminium ion cyclisations of malic
acid derivatives.12 Herein, we describe this work in full, including
substantial unpublished chemistry in which we have attempted to
further streamline our earlier syntheses, and developed rapid new
access to the erythrinan system.
of cyclopentadiene adduct 6. As will be seen, the use of this
framework serves as a device for controlling the stereochemistry
of the crucial C-5 spirocyclic centre in 1, which originates as one of
the carbonyl functions of a starting imide. In initial work we chose
a pentenyl side chain in lactam 6 to provide access to a functional
group in 5 capable of subsequent annulation, whilst not interfering
in earlier steps.
The crucial C-5 stereochemistry would be set up in an
N-acyliminium ion cyclisation of hydroxylactam 7, and control in
this step was anticipated to arise by exo-attack of the electron-rich
aromatic nucleus on the convex face of the reactive electrophile.15
Asymmetric construction of 7 required differentiation of the two
imide carbonyl functions of the readily available starting imide 9,
which would be achieved by temporary installation of a silicon
substituent using a chiral base to give 8. The first iteration of this
idea proved successful, and access to the polycyclic lactam 6 is
illustrated in Scheme 2.
Asymmetric silylation of imide 9 was most effective using
the bis-lithium amide 10 (this transformation was effected in
both enantiomeric series and also with racemic base), giving
(+)-8 in 94% ee.16 Grignard additions to imide 8 proved to
be completely regiocontrolled, C–C bond formation occurring
only at the carbonyl function distal to the installed trimethylsilyl
substituent. This appears to be a general result for such imides
in reactions with organometallics and metal hydride reducing
agents, and which we ascribe to steric inhibition of Lewis acid
co-ordination.17 Desilylation and acid-mediated N-acyliminium
ion cyclisation then proved to be completely diastereoselective,
Chiral base synthesis of erythrinan systems
The retro-Diels–Alder approach
Our initial synthesis of the erythrinan system came about due to an
ongoing interest in imide desymmetrisation using chiral bases, an
approach that we had used to access other natural product motifs,
including a proposed structure for an alkaloid called jamtine.13
Our approach is outlined in retrosynthetic form in Scheme 1.
Access to the complete alkaloid framework, and ultimately
natural products such as (+)–erysotramidine 1, would be accom-
plished by a cyclisation of an a,b-unsaturated lactam 5, in which a
functional group (FG) in the side-chain appendage would facilitate
a Michael-type cyclisation to form the required 6-membered ring.
Based on some related work by Lete and co-workers,14 we expected
that the lactam 5 would be available by retro-Diels–Alder reaction
◦
provided that the latter reaction was conducted at 0 C to room
temperature. In our earliest attempts this reaction was carried out
in CH2Cl2 at reflux, and small amounts of a minor diastereomer
were formed that ultimately eroded the enantiomeric excess of
later compounds to about 85% ee.18
Further transformation of this series commenced with a retro-
Diels–Alder reaction (Scheme 3), to give the lactam (-)-11 in
93% ee, the absolute configuration of which had been established
earlier,18 and which was further secured by subsequent conversion
into (+)-erysotramidine 1. This transformation was carried out
by the simple expedient of brief heating of 6, placed in an
Scheme 1
1964 | Org. Biomol. Chem., 2009, 7, 1963–1979
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Scheme 2
Scheme 3
outsized round-bottom flask under high vacuum, using a Bunsen
burner.
probably due in part to the lack of a competing 1,5-hydrogen
atom abstraction process. Further transformation of 13 was
then possible by alcohol elimination, to give 14, followed by
selenium mediated dehydrogenation to give 15, which was an
intermediate in the synthesis of racemic erysotramidine, described
by Padwa and Wang .5 Diene 15 was then converted into (+)-
erysotramidine following the Padwa protocol, which involves
completely diastereocontrolled allylic hydroxylation, using SeO2
(we recorded 21% yield, 79% based on recovered starting material),
followed by alcohol methylation using Tsuda’s method—KOH,
MeI, Et4NBr, THF (95%).22
The side chain of (-)-11 was oxidised using a one-pot
dihydroxylation–diol cleavage procedure to give aldehyde (-)-12.19
Although we considered a number of possibilities for conversion
of aldehyde 12 into a tetracyclic product, the first method tried
worked extremely well. Thus, exposure of 12 to tributyltin hydride,
under conditions described by Muller et al. for cyclisation of
5¢-aldehyde nucleosides,20 led efficiently to the desired hydroxy-
lactam 13 as a mixture of diastereomers at the newly formed
carbinol centre (ca 3:1 ratio). At this stage, as disclosed previously,
full structure determination was possible for the major epimer
of 13 by X-ray crystallography.21 This type of radical 6-exo-trig
ring closure is comparatively rare compared to the 5-exo-trig
variants, and the very high yield in our successful example is
Whilst our initial total synthesis was complete, we were aware
of shortcomings in the chemical yields in later transformations,
particularly the dehydrogenation step and the allylic oxidation.
Believing that such functional group transformations were in
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 1963–1979 | 1965
©

View Article Online
principal solvable issues, we focused on developing a second
generation synthesis that would avoid the other shortcoming of
the synthesis, namely the requirement for the retro-Diels–Alder
process.
is off-set by the gain in molecular complexity. The problem here
is the rather protracted access to 19 in enantio-pure form, which
required in excess of 20 steps.
Bearing these aspects in mind, and wishing to further probe
synthesis sequences that combined our asymmetric deprotona-
tion protocol with a diastereoselective N-acyliminium cyclisa-
tion, we conceived the modified route to alkene 14 shown in
Scheme 5.
According to this approach, desymmetrisation of a cyclobutene-
fused imide 24 would be carried out in analogous fashion to that
achieved previously with 9, except that we now require addition of
a butenyl (rather than pentenyl) side-chain, to give 23. Cyclisation
of hydroxylactam 23 to give 22 appeared reasonable, although the
diastereocontrol available could not be guaranteed. Tetracyclic
cyclobutene 22 would be primed for the key step, which involves
ring-opening/ring-closing metathesis (RORCM) to give 21, which
has the complete erythrinan skeleton. Conversion of 21 into
the previous intermediate 14 would then require removal of the
unwanted vinyl side-chain, which we envisaged would be possible
via oxidative cleavage.
The metathesis approach
Whilst considering alternative variants of the imide desymmetri-
sation shown in Scheme 1, we were attracted to the idea of
using metathesis for the key 6-membered ring-forming process.
As mentioned above, the application of metathesis in the area of
Erythrina alkaloid synthesis has been explored by other groups,10,23
this work being largely contemporaneous with our own studies. Of
particular relevance to our work was the ring-closure of diene 16
to give 17, described by Lete and co-workers,23 and the metathesis
of 19 to give (+)-3 and 20 in the aforementioned access to (+)-
b-erythroidine,10 both using Grubbs’ first generation catalyst 18,
Scheme 4.
The efficient formation of 17 represents the key step in an
efficient access to the Erythrina skeleton, but the route does not
look easily amenable to asymmetric synthesis and no natural
products have been accessed to date. The dienyne metathesis of
19 is clearly more challenging and the moderate chemical yield
The required starting cyclobutene-imide 24 was synthesised
via a three-step sequence in 22% overall yield from cheap and
commercially available maleic anhydride 25 (Scheme 6).
Scheme 4
Scheme 5
1966 | Org. Biomol. Chem., 2009, 7, 1963–1979
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Scheme 6
Scheme 7
Irradiation of a dilute solution of maleic anhydride 25 and
1,2-dichloroethylene in degassed EtOAc in a pyrex photoreactor
using light from a 400 W medium pressure Hg lamp led to a
mixture of diastereomeric dichlorides 26 in 30% yield.24 Conden-
sation of anhydride 26 with amine 27 followed by dehalogenation
using zinc/acetic anhydride delivered the cyclobutene product 24
successfully in multi-gram quantities. Although for preliminary
studies we were content to use racemic compounds, we were
interested to check that the chiral base mediated silylation of imide
24 could be conducted with high levels of selectivity.
To our satisfaction, application of the chiral base conditions
used previously for imide 9 to the cyclobutene-fused imide 24
proved even more effective than before, the required product being
isolated in high yield, and as a single enantiomer according to the
limits of our HPLC assay, Scheme 7.
not be achieved using TFA, which we had employed previously; in
fact use of this reagent in CH2Cl2 at reflux gave a 1:1 mixture of 22
and the C-5 epimer. Use of TIPSOTf gave promising results, with
diastereoselectivities of up to 10:1 at low temperature (-78 ◦C), but
also low conversions due to a sluggish reaction. TMSOTf provided
excellent results provided that the reaction mixture was initially
cooled to -78 ◦C and then warmed slowly to -4 ◦C. Under such
conditions 22 was formed in 95% yield and with a diastereomer
ratio of 20:1, favouring the isomer shown (the stereochemistry of
this compound, arising from the expected exo-mode of attack was
proven later).
With the acyliminium cyclisation product 22 in hand, we next
sought to establish conditions for its conversion into 21 by
ring-opening/ring-closing metathesis. The reaction was initially
conducted using 4 mol% of Grubbs’ first generation catalyst 18
but only a very small amount of the desired product 21 was
formed (ca. 5%).25 During their research on diastereoselective ring
rearrangement metathesis, Blechert and co-workers had found
that an ethylene atmosphere could prevent dimerisation and
oligomerisation and could enhance the catalyst initiation rate.26
Following their procedure the catalyst 18 (10 mol%) was added to
Having established that an asymmetric synthesis might be
viable, we continued the route by Grignard addition to 24 to give
racemic adduct 23, followed by N-acyliminium cyclisation to give
tetracyclic lactam 22 in racemic form, Scheme 8.
The cyclisation of 23 required significant fine-tuning in order
to achieve high levels of diastereocontrol. Good selectivity could
Scheme 8
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 1963–1979 | 1967
©

View Article Online
Scheme 9
Fig. 3 X-ray crystal structure of metathesis product 21. Displacement ellipsoids are drawn at the 50% probability level.
a degassed solution of cyclobutene 22 in CH2Cl2 which was then
stirred under an ethylene atmosphere. Combined with an increase
in the dilution of the reaction mixture, this protocol allowed us to
achieve rearrangement of 22 into the erythrinan system 21 in 62%
yield, Scheme 9.
to be remarkably difficult, and which unearthed some interesting
chemistry.
We envisioned that if the terminal alkene of 21 could be dihy-
droxylated selectively to give diol 30, then subsequent oxidative
cleavage and decarbonylation of the corresponding aldehyde 31
would give the known intermediate 14 which is a precursor of
erysotramidine. The aldehyde 31 might alternatively be oxidised
to give acid 32, which would lead to 14 by decarboxylation
(Scheme 10).
Unfortunately, dihydroxylation reactions of diene 21 were
poorly regiocontrolled under standard conditions with both
alkenes reacting, and furnishing mixtures of products. After some
experimentation with the Sharpless asymmetric hydroxylation
conditions,29 we were able to obtain one product quite selectively,
Although we also tested Grubbs’ second generation catalyst,27
and the Hoveyda–Grubbs catalyst,28 these systems offered no
advantage over catalyst 18.
The stereochemical assignment of tetracyclic system 21 was
established by x-ray crystallographic analysis, which confirmed
the outcome of the previous key N-acyliminium ion cyclisation,
Fig. 3.†
At this stage, the selective oxidative removal of the vinylic side-
chain present in lactam 21 was our next task, which turned out
Scheme 10
1968 | Org. Biomol. Chem., 2009, 7, 1963–1979
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Scheme 11
Scheme 12
but this proved to be diol 33, resulting from reaction at the internal
alkene, Scheme 11.
reactions) by cross metathesis to give a more reactive, electron-rich
alkene.
Alternative oxidation conditions were explored in our search
for useful regioselectivity. Thus treatment of substrate 21 with
SeO2 in dioxane under reflux resulted in formation of the tertiary
alcohol 34 along with a small amount of dehydration product 35
(Scheme 12).
The transformations shown in Schemes 11 and 12 were not
helpful to us, and at this stage it seemed quite difficult to selectively
cleave the less electron rich terminal alkene directly. Instead
we envisioned modifying the reactivity of the terminal alkene
(and therefore changing the overall regioselectivity of subsequent
Treatment of diene 21 with Grubbs’ catalyst 18 and vinyl acetate
or vinyl trimethylsilane gave none of the desired cross metathesis
product. However, a mixture of the desired product 37 and another
one-carbon homologated by-product 38 were obtained in a
combined 85% yield when substrate 21 was reacted with Hoveyda–
Grubbs catalyst 36 and 2-methyl-2-butene (Scheme 13).30
Alternatively, the trans phenyl substituted alkene 40 was ob-
tained in 51% yield when substrate 21 was treated with styrene and
Grubbs’ second generation catalyst 39. We expected that either 37
or 40 might be available directly from cyclobutene 22 if required.
Scheme 13
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 1963–1979 | 1969
©

View Article Online
Scheme 14
With the new diene substrates 37 and 40 available we again
examined dihydroxylation reactions. However under standard
dihydroxylation conditions (cat.OsO4/NMO/acetone–H2O) both
of these dienes gave only complex, polar mixtures of products.
Treatment of compound 40 with O3 at low temperature for several
minutes, followed by work-up with Me2S at room temperature led
only to extensive degradation of the substrate.
it had become clear that the problems involved in removal of the
vinyl side-chain in diene 21 made the route too long to pursue
further.
Review and outlook for the imide desymmetrisation
strategy to erythrinans
After considering alternative reactions that might prove selec-
tive for a terminal alkene over an internal cyclohexene we selected
the Wacker oxidation reaction as a candidate for exploration,
since this would be expected to convert the mono-substituted
alkene into a methylketone, which might then prove amenable to
cleavage via oxidative conditions (or perhaps a retro-acylation).31
To our surprise, the reaction of diene 21 under standard Wacker
oxidation conditions (PdCl2–CuCl–O2) gave aldehyde 41 in good
yield, Scheme 14.
This type of anomalous regiochemistry has been observed
previously in the Wacker reactions of a-vinyl-b-lactams, and it
was postulated that palladium coordination to the Lewis basic
lactam function was responsible.32 This transformation was at least
selective for the terminal alkene, and provided us with one final
opportunity to cleave this appendage by enol silane formation,
followed by oxidation. Although we formed the sensitive enol
silane 42 in modest yield, we were unable to effect subsequent
cleavage using standard dihydroxylation conditions.
In a final effort to achieve our goal of cleaving the unwanted
vinyl group in 21 we resorted to an alkene protection strategy.
Treating the substrate 21 with pyridinium tribromide resulted in
the formation of dibromide 43 in moderate yield (Scheme 15).
Reaction of this dibromide with O3 at -78 ◦C and then with
zinc/acetic acid at 40 ◦C to decompose the ozonide, gave the
primary alcohol 44 in excellent yield, rather than the expected
aldehyde. Preliminary attempts to remove the remaining hydrox-
ymethyl group by a retro-aldol process, by heating under basic
conditions, proved ineffective.
Inevitably, our perspective on this total synthesis project changed
and developed whilst the work progressed, and it is appropriate
to both assess the work carried out and to sketch out some
opportunities that have been identified for further study.
Our initial synthesis of (+)–erysotramidine proceeded in 12
chemical steps from the readily available imide 9 and employed
chiral base-mediated imide silylation as a way of distinguishing
=
the two imide C O functions in an addition reaction employing
a Grignard addition. It is clear that a process involving direct
asymmetric organometallic addition to the imide 9 would be
superior to our three-step approach, which involves silylation,
addition and then de-silylation. However, to our knowledge
no established methodology exists for this type of asymmetric
addition, and our preliminary efforts to identify useful leads in
this area, for example using Grignard reagents in the presence of
sparteine, were not successful.33
As an interim measure it appeared to us that we might be able to
streamline the chiral base three-step desymmetrisation if, instead
of silylating the enolate, we could add organometallic reagents
directly to the lithium enolate itself, Scheme 16.
The lithium enolate in intermediate 45 would act as a pro-
=
tecting group for one of the two imide C O functions, enabling
selective addition at the other position. Unfortunately, it proved
impossible to realize this concept using imides. At the low
temperatures employed for the chiral base reactions the addition
of Grignard reagents to enolates like 45 did not proceed, whereas
organolithiums simply converted 45 into the well-known bis-
enolates. At higher temperatures (above -60 ◦C) we identified
a tendency for racemisation of the intermediate imide mono-
enolates, which undermined the process proposed in Scheme 16.
Subsequently, we were able to realize the principle of using chiral
At this stage our patience and materials were exhausted.
Although ultimately it may be possible to employ one or more
of the intermediates prepared here in a natural product synthesis,
Scheme 15
1970 | Org. Biomol. Chem., 2009, 7, 1963–1979
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Scheme 16
base generated enolates as electrophiles in this type of process but
with rather different types of substrate, namely 2,2-disubstituted
1,3-diketones.34
The success of the metathesis process shown in Scheme 9
was tempered by the severe problems in removing the unwanted
vinylic side-chain appendage. Also, although the metathesis route
avoided the clumsy retro-Diels–Alder process, the requirement
for an unusual cyclobutene-fused imide added synthetic steps.
From commercial starting materials the metathesis route actually
worked out longer than the retro-Diels–Alder route.
At present it appears to us that an optimal asymmetric synthesis
of the erythrinan system has yet to emerge. The types of metathesis
ring closure shown in Schemes 4 and 9 appear attractive if concise
asymmetric access to correctly substituted metathesis precursors
can be achieved.
it became clear that complementary access to such intermediates
should be possible starting from (L)-malic acid. Lee and co-
workers had observed that the cyclisation of hydroxylactam 47
gave the tricyclic pyrrolidinoisoquinoline product 49 as a single
diastereoisomer, Scheme 17.35
The stereochemical outcome results from neighbouring group
participation by the acetoxy group, which is known to control this
type of overall anti-addition.36
Adopting this synthesis to our needs, we proposed that effective
access to natural product targets might be gained by ring
closure of hydroxylactam intermediate 50, which we expected
would give lactam 51 as the major (or sole) diastereoisomer,
Scheme 18.
We could then envisage elimination of acetic acid from 51 to give
unsaturated lactam 52, which is a generic structure that includes
our previously established intermediate 11 for (+)-erysotramidine
synthesis (R = pentenyl). As will be seen, rather than converge
with the previous synthesis, we chose an alternative R group in 51,
which then enabled modified access to a known erysotramidine
intermediate. Alternatively, it appeared possible that functional
group manipulation could lead to an aminoketone of general
Chiral pool synthesis of erythrinan systems using
(L)-malic acid
During the course of our studies on the chiral base route to
N-acyliminium precursors for assembling the erythrinan system,
Scheme 17
Scheme 18
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 1963–1979 | 1971
©

View Article Online
structure 53, which could then be usefully advanced via known
aldol chemistry, provided an appropriate ketone appendage were
installed in the R group.
Our exploration of these ideas led to some unexpected observa-
tions concerning the cyclisations of N-acyliminium ions derived
from malic acid, and enabled us to establish two distinct synthetic
routes to 3-demethoxyerythratidinone (2).
acids employed, we decided to change the ring OAc substituent
for a bulky OTIPS group, Scheme 20.
Silylation of 56, followed by Lewis-acid treatment of hydroxy-
lactam 59, as described by Lee (Scheme 17) but initially at low tem-
perature, then furnished lactams 60 and 61 in good yield and over
9:1 selectivity. Surprisingly, the major product of this reaction was
the syn-adduct 60, in which aromatic attack on the N-acyliminium
intermediate had occurred from the same side of the lactam ring
as the OTIPS substituent. The stereochemical assignment of the
lactam 60 followed an X-ray structure determination, as described
previously.12
Observation of unexpected selectivity in an N-acyliminium
ion cyclisation
Lactams 60 and 61, were easily separated by chromatography,
and deprotection of 60, acetylation of the so-formed alcohol and
elimination of acetic acid as before, gave enantiomerically pure
(+)-58.
This type of syn selectivity has been reported previously for
intermolecular reactions of acetoxylactam 62 in both the N–H and
N–Bn series. Reactions with allyl, propargyl and allenylsilanes,
and stannanes, under Lewis acid catalysis, gave product ratios
ranging from 3.8:1 to >100:1.39
In order to probe these ideas we explored Grignard addition
reactions to the imide 54, which was readily prepared from
(L)-malic acid by standard procedures, Scheme 19.
Addition of but-3-enylmagnesium bromide 55 to imide 54 gave
hydroxylactam 56 as a mixture of inseparable diastereomers (about
3:1) in a completely regiocontrolled fashion. Selective addition
to the carbonyl group proximal to the acetoxy function was
expected, based on a closely related Grignard addition described
previously, in which the ring acetate was retained.37 In our case we
employed an excess of Grignard reagent (5 equiv.) to maximize
chemical yields, which also deacylated the secondary alcohol.
The high regioselectivity of Grignard reagent addition to the
=
a-oxygenated C O function has been attributed to both the
oxygen inductive effect and to complex induced proximity effects
(CIPE).38
This cis-selectivity was attributed to a stereoelectronic effect
(Cieplak effect) involving a preferred syn alignment of the
(relatively electron-rich) C–H at C-4 bond and the emerging s*
orbital at the reacting centre. This effect appears rather dependent
upon the nitrogen substituent, the nucleophile and the reaction
conditions, and the few selectivities reported to date are mainly
modest.40,41 This effect appears not to have been observed before
in additions of aromatics to N-acyliminium ions, and we were
interested to test the stereoelectronic explanation by testing an
additional lactam series.
Reaction of 56 with acetic anhydride re-installed the secondary
acetate without affecting the tertiary hydroxyl function, and
subsequent treatment with TIPSOTf then gave the tricyclic lactam
57 in good yield but with a surprisingly low diastereomer ratio of
3:1 (major isomer shown), compared to the precedent illustrated
in Scheme 17. Acetoxy-lactam 57 was then converted into the
unsaturated counterpart 58 (the butenyl homologue of 11) by
treatment with NaH, and the modest enantiomeric excess of
this product (47% ee) mirrored the low diastereomer ratio of the
precursor 57.
We selected for further study the parent system analogous to
that shown in Scheme 17, in which the C-5 substituent is hydrogen,
Scheme 21.
Following unsuccessful attempts to enhance the selectivity of
the key cyclisation by altering the reaction conditions or Lewis
Scheme 19
1972 | Org. Biomol. Chem., 2009, 7, 1963–1979
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Scheme 20
Scheme 21
Initially, for comparison with the result in Scheme 17 we
converted 54 into 65 by reduction with NaBH4 in CH2Cl2–MeOH,
and then c◦arried out the cyclisation using BF3-OEt2 in CH2Cl2
between 0 C and room temperature. As expected, only a single
diastereomeric lactam product (67), resulting from the expected
anti-attack, was obtained.
At present we consider the most likely explanation for the
selective formation of 60 is based on a conformational relay effect
as illustrated in Fig. 4.
In order to test for the ‘syn-effect’, we then prepared OTIPS
derivative 66, by routine transformations from 54, and then
effected cyclisation using BF3-OEt2 in CH2Cl2 at the lowest viable
temperature, -40 ◦C to RT, so as to mimic the conditions in
Scheme 20, and maximise selectivity. The cyclisation gave two
diastereomeric tricyclic products 68 and 70 in a 5:1 ratio, which is
the same sense of selectivity as the acetate series. This result seems
to suggest a steric argument for the syn-selectivity leading to the
butenyl derivative 60, since the effect does not seem to operate for
the parent system where the C-5 substituent is simply a proton.
Fig. 4
Examination of molecular models suggests that in the interme-
i
diate N-acyliminium ion the bulky TIPS group (R = Pr) pushes
the butenyl group to the opposite (top) face of the lactam ring,
which results in attack of the aromatic group from the same side
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 1963–1979 | 1973
©

View Article Online
as the OTIPS substituent. In the simpler system, where the C-5
substituent is a proton, the relay effect is lost and the reaction
displays modest anti-selectivity.
3:1.44 No regioisomeric hydroxyselenide was isolated. For the
radical cyclisations, solutions of triphenyltin hydride and AIBN in
benzene were added to a refluxing solution of the hydroxyselenide
72 in the same solvent over 8–10 hours using a syringe pump.
Reflux was continued for another 4 hours after the end of addition.
The desired cyclised product 73 was then isolated in 85% yield and
as a 1:1 mixture of diastereoisomers. Thus by a two-step procedure
the unsaturated lactam (+)-58 was successfully converted into the
Erythrina alkaloid skeleton 73, which is a known intermediate in
Tu and co-worker’s synthesis of ( )-erysotramidine.6
The cyclized secondary alcohol 73, was also treated with Dess-
Martin periodinane reagent to give another known compound,
ketone (+)-74, Both this ketone, and the derived dioxolane, were
synthesized in both enantiomeric forms by Allin and co-workers,
and data for our ketone matched their spectroscopic and specific
rotation data extremely well.8 Thus, our synthesis of 74 constituted
a formal synthesis of the natural product, since the dioxolane
derived from ketone 74 had been previously converted into
3-demethoxyerythratidinone (2) by Tsuda et al., although in race-
mic form.45
Access to known Erythrina intermediates using Clive’s radical
cyclisation protocol
With ready access to unsaturated lactam 58, we considered various
options to complete the construction of the erythrinan skeleton.
One possibility was to effect Michael addition of a vinyl group to
58, followed by ring closing metathesis which, starting with (+)-58
would then lead to (-)-14 (depending upon the stereochemistry of
71), Scheme 22.
This idea was soon abandoned since we found the starting
lactam to be resolutely unreactive to a range of vinyl cuprate
reagents, and also to rhodium catalysed reaction with potassium
vinyltrifluoroborate. We were also aware that, in the aforemen-
tioned access to lactam 16, Lete and co-workers had found
very similar problems and had resorted to incorporation of a
further carboxymethyl group to activate the Michael addition
process.42
Instead we returned to the use of radical chemistry, and adopted
a method described by Clive and co-workers for the synthesis
of functionalised ring-fused bicyclic compounds.43 The method
involves the radical cyclisation of a b-hydroxyselenide, which in
our case was readily available by addition to the unsaturated side-
chain appendage, Scheme 23.
Treatment of unsaturated lactam (+)-58 with phenylselenyl
chloride in aqueous acetonitrile afforded the hydroxylselenide
72 efficiently as a mixture of diastereoisomers in a ratio of
Total synthesis of 3-demethoxyerythratidinone
by aldol ring-closure
As outlined in Scheme 18 we also considered that the availability
of various lactams, arising from diastereoselective N-acyliminium
ion cyclisation, might enable access to erythrinan alkaloid natural
products via intramolecular aldol reaction. Our initial attempts in
this direction commenced by reduction of lactam 60 with LiAlH4
in the presence of AlCl3, Scheme 24.
Scheme 22
Scheme 23
1974 | Org. Biomol. Chem., 2009, 7, 1963–1979
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Scheme 24
Lactam reduction to give TIPS protected hydroxy pyrrolidine
75 was reasonably efficient if the reaction time was limited to
about 2 hours, but an even more effective reduction and in-situ
deprotection ensued if the reaction was carried out overnight,
providing the free alcohol 76. This compound was efficiently
oxidised, using typical Swern oxidation conditions, to give ketone
77. Our plan with this compound was to effect Wacker oxidation
of the alkenyl side-chain, followed by intramolecular aldol and
dehydration to provide 2 directly. However, despite good precedent
for Wacker oxidation of very similar alkenes we could not effect
this transformation with 77.46
After further experimentation, we established that our earlier
intermediates 60 and 61, in which the ring nitrogen is part of a
lactam, rather than a basic tertiary amine, participated well in
Wacker oxidations. We were then in a position to conclude our
proposed new access to 3-demethoxyerythratidinone 4. Initially,
in order to match our synthetic material with the alkaloid
in the natural enantiomeric series we used the minor product 61
generated from the cyclisation shown in Scheme 20 for further
transformations, Scheme 25.
and then aldol cyclisation, gave (+)-3-demethoxyerythratidinone
(2) with spectroscopic data consistent with the structure shown
and the previous published data.48 As anticipated, the use of
61 as starting material resulted in the formation of the final
26
alkaloid as its natural antipode, with [a]D +316 (c, 0.4, CHCl3),
compared to [a]D +325 (c, 0.25, CHCl3).49 We also applied the
20
sequence shown in Scheme 25 to the diastereomeric starting lactam
60 and obtained similar yields for each step (in parentheses)
and were able to access the natural product as the unnatural
(-)-antipode.
The total synthesis of alkaloid 2 proceeds in only 7 steps from
the readily available starting imide 54. However, access to the
natural product as its natural antipode is hindered by the fact that
the key N-acyliminium ion cyclisations are either poorly selective
(O-acetyl series), or highly selective for the wrong epimeric series
(O-TIPS compounds).
Summary and Conclusion
We were successful in establishing a concise new asymmetric
total synthesis of (+)-erysotramidine (1) in highly enantiomeri-
cally enriched form by application of the chiral lithium amide
base approach. The type of imide desymmetrisation employed
should be a useful method for the synthesis of other naturally
occurring alkaloids, both erythrinans and other types, although
the requirement for a stoichiometric amount of chiral base is
Wacker oxidation of diastereomerically pure 61 gave ketolac-
tam 78 in good yield and subsequent reduction with LiAlH4–
AlCl3 provided diol 79, which had been prepared previously
by Wasserman and Amici in their synthesis of racemic
3-demethoxyerythratidinone.47 Application of the final two steps
from their synthesis to 79, involving Swern oxidation to give 80
Scheme 25
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 1963–1979 | 1975
©

View Article Online
at present a disadvantage. There remain opportunities for new
asymmetric (preferably catalytic) transformations of cyclic imides,
both in terms of enolisation and additions to the imide carbonyl
functions, which we hope to explore in the future.
retention times were 29.5 min (major), and 32.4 min (minor). The
absolute configuration of this product was assigned by analogy
with known examples.
Our alternative procedure for constructing the skeleton of
Erythrina alkaloids using the ring-opening/ring-closing metathe-
sis provides another example of this powerful ring-rearrangement
procedure. Although ultimately unsuccessful, the route also
provided examples of chiral base selectivity and diastereocon-
trolled N-acyliminium ion cyclisation in the unusual context of
a cyclobutene-fused imide system.
The final part of our study provided significant new in-
formation concerning the stereochemical issues related to
N-acyliminium ion reactions of malic acid derived lactams, and
particularly the unusual syn-effect seen with a TIPS protected
compound. Based on these new results the formal syntheses
of (+)-erysotramidine and (+)-3-demethoxyerythratidinone were
achieved using radical cyclisation reaction for the final bond
formation of the Erythrina skeleton. Also the total synthesis of
(+)-3-demethoxyerythratidinone using N-acyliminium cyclisation
and aldol condensation reactions as the key steps represents only
the third asymmetric access to this compound, and at eight steps
from commercial material is one of the shortest of the routes
established to date.
(4aS,5R,13bS)-11,12-Dimethoxy-5-vinyl-1,2,4a,5,8,9-
hexahydroindolo[1-a]-isoquinolin-6-one (21) (Scheme 9)
Cyclobutene imide 22 (233 mg, 0.72 mmol) was dissolved in dry
CH2Cl2 (50 mL) so that the concentration reached approximately
0.01 M. This solution was degassed under argon. Ethylene was
then bubbled through the solution for 5 min. Grubbs’ first gen-
eration catalyst 18 (59 mg, 0.07 mmol) was then added. Ethylene
was again passed through the solution for 5 min, and the solution
was then stirred under an ethylene atmosphere (fitted balloon) at
35 ◦C until TLC revealed completion of the reaction. The mixture
was cooled to room temperature and ethyl vinyl ether (2 mL) was
added. The reaction mixture was stirred for another 1 hour then
concentrated. The crude product was purified by silica column
chromatography (eluent: petroleum ether/AcOEt = 3/1 to 1/1)
to give the major product 21 (144 mg, 62%): nmax. (CHCl3)/cm-1
2996, 2846, 1662, 1463, 1362, 1106, 908; dH (400 MHz, CDCl3)
1.74–1.81 (m, 1 H), 2.00–2.06 (m, 2 H), 2.16–2.22 (m, 1 H), 2.72–
2.78 (ddd, 1 H, J 16.4, 6.0, 3.2, ArCHAHB), 2.94–3.02 (m, 1 H,
=
ArCHAHB), 3.10 (d, J 8.0, 1 H, CHCH CH), 3.29–3.36 (m, 2 H,
CHCO, NCHAHB), 3.82 (s, 3 H, OCH3), 3.86 (s, 3 H, OCH3), 4.10–
4.21 (ddd, J 13.2, 7.2, 3.2, 1 H, NCHAHB), 5.26–5.32 (m, 2 H,
Experimental part
=
CH CH2), 5.73–5.86 (m, 2 H, CH=CH2, CH=CHCH), 6.05–
=
6.12 (m, 1 H, CH CHCH), 6.61 (s, 1 H, ArH), 6.73 (s, 1 H, ArH);
General procedures - see ESI.† The preparation and spectral data
of compounds 6, 7, (+)-8, (-)-11, (-)-12, (-)-13, (+)-14, (+)-15,
can be found in our previous paper.11
dC (100 MHz, CDCl3) 22.2 (CH2), 27.1 (CH2), 33.3 (CH2), 35.1
(CH2), 43.3 (CH), 50.8 (CH), 55.9 (OCH3), 56.0 (OCH3), 61.0 (C),
108.1 (ArCH), 111.8 (ArCH), 119.3 (CH2=), 125.2 (ArC), 126.5
(CH=), 128.8 (CH=), 133.5 (ArC), 133.8 (CH=), 147.6 (ArC),
(1S,5R)-3-(3,4-Dimethoxyphenethyl)-1-(trimethylsilyl)-3-aza-
bicyclo[3.2.0]hept-6-ene-2,4-dione (29) (Scheme 7)
+
=
148.0 (ArC), 172.8 (NC O); ESIMS m/z (%) 326.2 ([M + H] ,
100); HRMS (ESI) found M+, 325.1670. C20H23NO3 requires M,
325.1678.
The bis-lithium chiral base 10 solution was prepared by addition
of n-BuLi (1.6 M in THF, 0.66 mL, 1.05 mmol) to the solution of
corresponding chiral amine (253 mg, 0.6 mmol) in THF (4 mL).
After cooling to -100 ◦C, the chiral base was added dropwise by
cannula to the solution of imide 24 (143 mg, 0.5 mmol) and TMSCl
(0.64 mL, 5 mmol) in THF (8 mL). The resulting reaction mixture
was stirred at the same temperature for 4 hours. Saturated aqueous
NaHCO3 (5 mL) solution was added and diluted with EtOAc
(100 mL). The organic phase was washed with water (10 mL), brine
(10 mL), dried over MgSO4 and concentrated in vacuo. The crude
product was purified by silica column chromatography (eluent:
petroleum ether/AcOEt = 2/1 to 1/1) to give the product 29
2-((4aS,5R,13bS)-11,12-Dimethoxy-6-oxo-1,2,4a,5,8,9-
hexahydroindolo-[1-a]isoquinolin-5-yl)acetaldehyde (41)
(Scheme 14)
A suspension of PdCl2 (17.4 mg, 0.1 mmol) and CuCl (40.3 mg,
0.4 mol) in DMF (2.0 mL) and water (0.4 mL) was degassed with
oxygen. The mixture was then stirred at room temperature for
4 hours. A solution of imide 21 (145 mg, 0.45 mmol) in DMF
(0.8 mL) was added and the resulting reaction mixture was stirred
under oxygen atmosphere for 20 h. The reaction mixture was
diluted with EtOAc (100 mL). The organic phase was washed
with water (3 ¥ 10 mL), brine (10 mL), dried over MgSO4 and
concentrated in vacuo. The crude product was purified by silica
column chromatography (eluent: petroleum ether/AcOEt = 1/1)
to give the product as a colourless oil 41 (119 mg, 78%): nmax.
(CHCl3)/cm-1 2935, 2835, 1723, 1694, 1456, 1392, 1361, 1120,
870; dH (500 MHz, CDCl3) 1.73–1.78 (m, 1 H), 1.94–2.13 (m, 3 H),
2.64–2.71 (m, 2 H), 2.83–3.02 (m, 2 H), 3.12–3.28 (m, 3 H), 3.79 (s,
3 H, OCH3), 3.81 (s, 3 H, OCH3), 4.10–4.19 (m, 1 H, NCHAHB),
5.57–5.60 (m, 1 H, CH=), 6.01–6.04 (m, 1 H, CH=), 6.56 (s,
1 H, ArH), 6.66 (s, 1 H, ArH), 9.82 (s, 1 H, CHO); dC (125 MHz,
CDCl3) 21.7 (CH2), 26.8 (CH2), 33.8 (CH2), 35.0 (CH2), 39.9 (CH),
41.4 (CH), 42.6 (CH2), 55.8 (OCH3), 56.0 (OCH3), 61.5 (C), 108.0
25
(88 mg, 88%) as a yellow oil: [a]D +131 (c 4.50, CHCl3); nmax.
(CHCl3)/cm-1 2938, 2838, 1755, 1682, 1593, 1387, 1357, 986; dH
(500 MHz, CDCl3) 0.09 (s, 9 H, Si(CH3)3), 2.77 (t, J 7.5, 2 H,
PhCH2), 3.47 (d, J 1.0, 1 H, NCOCH), 3.59–3.69 (m, 2 H, NCH2),
3.69 (s, 3 H, OCH3), 3.83 (s, 3 H, OCH3), 6.31 (dd, J 2.5, 1.0,
=
1 H, CH=CH), 6.39 (d, J 2.5, 1 H, CH CH), 6.69–6.74 (m, 3 H,
ArH); dC (125 MHz, CDCl3) -3.5 (SiCH3), 33.0 (ArCH2), 39.7
(NCH2), 50.0 (CH), 52.4 (CSi), 55.9 (OCH3), 55.9 (OCH3), 111.2
(ArCH), 112.1 (ArCH), 121.0 (ArCH), 130.3 (ArC), 136.0 (CH=),
=
142.6 (CH=), 147.7 (ArC), 148.8 (ArC), 174.3 (NC O), 176.6
+
=
(NC O); HRMS (ESI) found [M + H] , 360.1626. C19H25NO4Si
requires [M + H], 360.1590; The ee was determined as > 99% by
HPLC (chiral OD column, 4% iPrOH in hexane, 0.6 ml/min), the
1976 | Org. Biomol. Chem., 2009, 7, 1963–1979
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
(CH), 111.8 (CH), 125.1 (C), 125.4 (CH), 130.0 (CH), 133.0 (C),
CHAHBPh), 2.65 (dd, J 15.6, 3.2, 1 H, CHAHBCON), 2.80–2.88
(m, 1 H, CHAHBPh), 3.00–3.11 (m, 2 H, CHAHBCON, CHAHBN),
3.83 (s, 3 H, OCH3), 3.87 (s, 3 H, OCH3), 4.44 (dd, J 13.2,
=
147.6 (C), 148.0 (C), 174.7 (NC O), 200.4 (CHO); HRMS (ESI)
found M+, 341.1627. C20H23NO4 requires M, 341.1626.
=
5.6, 1 H, CHAHBN), 4.95–5.01 (m, 2 H, CH CH2), 5.67–5.73
(m, 2 H, CH=CH2, CHOAc), 6.51 (s, 1 H, ArH), 6.59 (s, 1 H,
ArH); dC (100 MHz, CDCl3) 20.7 (CH3), 28.6 (CH2), 28.8 (CH2),
36.4 (CH2), 38.6 (CH2), 39.8 (NCH2), 55.8 (OCH3), 56.0 (OCH3),
68.6 (C), 73.4 (CH), 108.8 (ArCH), 111.4 (ArCH), 115.5 (CH2=),
(S)-10b-(But-3-enyl)-8,9-dimethoxy-5,6-dihydropyrrolo[2,1-
a]isoquinolin-3-one (-)-(58) (Scheme 19)
To a stirred suspension of NaH (2.6 g, 64 mmol, 60% dispersion
in mineral oil) in THF (20 mL) was added dropwise a solution
of mixture 57 (1.6 g, 4.3 mmol) in THF (10 mL). The resulting
solution was then stirred overnight under nitrogen atmosphere.
The reaction mixture was added dropwise to the ice-water and
extracted with EtOAc (200 mL). The organic phase was washed
with water (10 mL), brine (10 mL), dried over MgSO4 and
concentrated in vacuo. The crude product was purified by silica
column chromatography (eluent: petroleum ether/AcOEt = 2/1)
126.5 (ArC), 127.1 (ArC), 137.0 (CH=), 147.4 (ArC), 147.8 (ArC),
+
=
=
169.9 (NC O), 171.0 (C O); ESIMS m/z (%) 382.2 ([M + Na] ,
100); HRMS (ESI) found [M + H]+, 360.1803. C20H25NO5 requires
[M + H], 360.1811.
To a stirred suspension of NaH (2.6 g, 64 mmol, 60% dispersion
in mineral oil) in THF (20 mL) was added dropwise a solution of
acetate (1.6 g, 4.3 mmol) in THF (10 mL). The resulting solution
was then stirred overnight under a nitrogen atmosphere. The
reaction mixture was added dropwise to ice-water and extracted
with EtOAc (200 mL). The organic phase was washed with water
(10 mL), brine (10 mL), dried over MgSO4 and concentrated
in vacuo. The crude product was purified by silica column
chromatography (eluent: petroleum ether/AcOEt = 2/1) to give a
viscous yellow oil (+)-58 (1.2 g, 95%): [a]D27 +242 (c 1.57, CHCl3).
Thespectraldatawere identical with(-)-58. The eewas determined
as >99% by HPLC (Chiracel OD Column, 20% iPrOH in hexane,
0.4 mL/min); the retention time was 23.4 min (major) and 31.2 min
(minor).
25
to give a viscous yellow oil (-)-58 (1.2 g, 95%): [a]D -108 (c 2.20,
CHCl3); nmax. (CHCl3)/cm-1 2937, 2847, 2936, 1681, 1360, 1108;
dH (400 MHz, CDCl3) 1.86–1.95 (m, 2 H, CCH2), 2.01–2.10 (m,
2 H, CH2CH=), 2.67 (dd, J 16.1, 4.0, 1 H, ArCHAHB), 2.94 (ddd,
J 16.1, 12.0, 6.6, 1 H, ArCHAHB), 3.18 (ddd, 1 H, J 13.3, 12.0,
4.0, NCHAHB), 3.83 (s, 3 H, OCH3), 3.88 (s, 3 H, OCH3), 4.40
(dd, J 13.3, 6.6, 1 H, NCHAHB), 5.00 (m, 2 H, CH2=), 5.70–5.80
(m, 1 H, CH=CH2), 6.18 (d, J 5.8 Hz, 1H, CH=), 6.60 (s, 1 H,
ArH), 6.70 (s, 1 H, ArH), 7.29 (d, J 5.8, 1 H, CH=); dC (100 MHz,
CDCl3) 27.6 (CH2), 29.1 (CH2), 34.8 (CH2), 37.9 (NCH2), 56.0
(OCH3), 56.3 (OCH3), 68.4 (C), 109.2 (CH=), 112.2 (CH=), 115.2
(CH2=), 125.3 (ArC), 126.4 (CH=), 129.4 (ArC), 137.5 (CH=),
(4aR,13bR)-11,12-Dimethoxy-1,2,4a,5,8,9-hexahydroindolo[1-a]-
isoquinoline-3,6-dione (+)-(74)
=
147.7 (ArC), 148.3 (ArC), 151.7 (CH=), 170.9 (C O); ESIMS m/z
(%) 322.1 ([M + Na]+, 100), 300.2 ([M + H]+, 76); HRMS (ESI)
found M+, 299.1522. C18H21NO3 requires M, 299.1521; The ee was
determined as 47% by HPLC (Chiracel OD Column, 20% iPrOH
in hexane, 0.4 mL/min); the retention times were 23.4 min (minor)
and 31.2 min (major).
(R)-10b-(3-Butenyl)-8,9-dimethoxy-5,6-dihydropyrrolo[2,1-
a]isoquinolin-3-one (+)-(58) (Scheme 20)
TBAF (0.2 mL, 0.2 mmol, 1 M in THF) was added to the solution
of lactam 60 (65 mg, 0.14 mmol) in THF (2 mL). The resulting
reaction mixture was then stirred at rt for 3 h. The mixture was
extracted with EtOAc (50 mL), the organic phase was washed
with H2O (10 mL), brine (10 mL), then dried over MgSO4 and
concentrated in vacuo. The crude product was purified by silica
chromatography (eluent: petroleum ether/AcOEt = 2/1 then pure
EtOAc) gave the alcohol corresponding to 60 (39 mg, 91%) as a
A solution of Dess-Martin periodinane (0.2 mL, 15 wt% in
CH2Cl2) was added to a solution of secondary alcohols 73 (30 mg,
0.09 mmol) in CH2Cl2 (2 mL) at room temperature. The resulting
solution was then stirred for 8 hours. Saturated Na2S2O3 solution
(6 mL) was added and the mixture was extracted with CH2Cl2
(10 mL). The organic phase was washed with water (6 mL)
and brine (6 mL). After drying (MgSO4), the organic layer was
concentrated to yield the crude product, which was further purified
by column chromatography (eluent: petroleum ether/AcOEt =
1/2, then 1/4) to give the desired product (+)-74 (21 mg, 75%):
25
colorless oil: [a]D +182 (c 1.89, CHCl3).
Et3N (76 mg, 0.75 mmol), Ac2O (55 mg, 0.53 mmol) and DMAP
(6.0 mg, 0.05 mmol) were added to the solution of secondary
alcohol (160 mg, 0.5 mmol) in CH2Cl2 (4 mL). The resulting
solution was then stirred at rt for 4 h. The reaction mixture was
then extracted with Et2O (100 mL). The organic phase was washed
with water (10 mL), brine (10 mL), then dried over MgSO4 and
concentrated in vacuo. The crude product was purified by silica
chromatography (eluent: petroleum ether/AcOEt = 1/1) to give
the acetate corresponding to 60 (130 mg, 78%) as a colourless
22
[a]D +45 (c 1.10, CHCl3); nmax. (CHCl3)/cm-1 2937, 1716, 1681,
1463, 1362, 1122; dH (500 MHz, CDCl3) 2.10–2.16 (m, 1 H, 5-HA),
2.24–2.35 (m, 3 H, 1-H, 2-HA), 2.38–2.43 (m, 1 H, 2-HB), 2.60–2.76
(m, 3 H, 4-HA, 5-HB, 9-HA), 2.96–3.11 (m, 4 H, 4a-H, 4-HB, 8-HA,
9-HB), 3.86 (s, 3 H, OCH3), 3.88 (s, 3 H, OCH3), 4.32–4.40 (m, 1 H,
8-HB), 6.58 (s, 1 H, ArH), 6.69 (s, 1 H, ArH); dC (125 MHz, CDCl3)
27.6 (ArCH2), 33.6 (CCH2), 34.8 (CH2CH2CO), 35.3 (NCH2),
37.5 (CCH), 37.8 (NCOCH2), 43.3 (COCH2CH), 56.0 (OCH3),
56.4 (OCH3), 62.5 (C), 107.2 (ArCH), 111.7 (ArCH), 125.5 (ArC),
oil: [a]D +112 (c 0.85, CHCl3); nmax. (CHCl3)/cm-1 2938, 2836,
25
=
1738, 1682, 1463, 1363, 1042, 907; dH (400 MHz, CDCl3) 1.58 (s,
3 H, COCH3), 1.92–2.14 (m, 4 H, CH2CH2), 2.38 (d, J 17.6, 1 H,
134.4 (ArC), 148.3 (ArC), 148.5 (ArC), 172.2 (NC O), 210.2
=
+
+
(C O); ESIMS m/z (%) 316.2 ([M + H] , 100), 338.1 ([M + Na] ,
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 1963–1979 | 1977
©

View Article Online
74); HRMS (ESI) found [M + H]+, 316.1554. C18H21NO4 requires
[M + H], 316.1532.
9 Y. Yasui, K. Suzuki and T. Matsumoto, Synlett, 2004, 619.
10 H. Fukumoto, K. Takahashi, J. Ishihara and S. Hatakeyama, Angew.
Chem. Int. Ed., 2006, 45, 2731.
11 A. J. Blake, C. Gill, D. A. Greenhalgh, N. S. Simpkins and F. Zhang,
Synthesis, 2005, 19, 3287.
12 F. Zhang, N. S. Simpkins and C. Wilson, Tetrahedron Lett., 2007, 48,
5942.
(S)-11,12-Dimethoxy-1,2,5,6,8,9-hexahydroindolo[1-a]isoquinolin-
3-one (+)-(2)
13 N. S. Simpkins and C. D. Gill, Org. Lett., 2003, 5, 535.
14 I. Gonzalez-Temprano, N. Sotomayor and E. Lete, Synlett, 2002, 593.
15 (a) B. E. Maryanoff, H. C. Zhang, J. H. Cohen, I. J. Turchi and C. A.
Maryanoff, Chem. Rev., 2004, 104, 1431; (b) W. N. Speckamp and M. J.
Moolenaar, Tetrahedron, 2000, 56, 3817; (c) A. Yazici and S. G. Pyne,
Synthesis, 2009, 339; (d) A. Yazici and S. G. Pyne, Synthesis, 2009,
513.
16 V. Rodeschini, N. S. Simpkins and F. Zhang, Organic Syntheses, 2007,
84, 306.
17 (a) D. J. Adams, A. J. Blake, P. A. Cooke, C. D. Gill and N. S. Simpkins,
Tetrahedron, 2002, 58, 4603; (b) C. D. Gill, D. A. Greenhalgh and N. S.
Simpkins, Tetrahedron, 2003, 59, 9213.
18 C. Gill, D. A. Greenhalgh and N. S. Simpkins, Tetrahedron Lett., 2003,
44, 7803.
19 W. S. Yu, Y. Mei, Y. Kang, Z. M. Hua and Z. D. Jin, Org. Lett., 2004,
6, 3217.
20 E. Muller, D. Gasparutto, M. Jaquinod, A. Romieu and J. Cadet,
Tetrahedron, 2000, 56, 8689.
21 This mixture of epimeric alcohols has been reported previously in
racemic form, see: A. Mondon, K. F. Hansen, K. Boehme, H. P. Faro,
H. J. Nestler, H. G. Vilhuber and K. Bottcher, Chem. Ber., 1970, 103,
615.
A solution of diketone 80 (34 mg, 0.107 mmol) and 20% KOH
(1.5 mL) in MeOH (30 mL) was heated at 120 ◦C under a nitrogen
atmosphere for 10 hours. The reaction mixture was concentrated
and extracted with CH2Cl2 (20 mL). The organic phase was washed
with water (5 mL), brine (5 mL) and dried over anhydrous MgSO4.
After filtration and concentration, the crude product was purified
by flash silica chromatography (eluent: CH2Cl2/MeOH = 10/1)
26
to give a yellow oil (+)-2 (15.9 mg, 50%): [a]D +316 (c 0.40,
CHCl3); nmax. (CHCl3)/cm-1 2936, 2852, 1666, 1463, 1360, 1107;
dH (400 MHz, CDCl3) 2.24–2.33 (m, 2 H, 1-H), 2.43–2.68 (m, 4 H,
2-H, 5-HA, 9-HA), 2.68–2.92 (m, 2 H, 5-HB, 6-HA), 3.03–3.12 (m, 2
H. 9-HB, 6-HB), 3.26 (dd, 1 H, J 14.4, 7.6, 8-HA), 3.45–3.52 (m, 1 H,
8-HB), 3.76 (s, 3 H, OCH3), 3.88 (s, 3 H, OCH3), 6.12 (s, 1 H, 4-H),
6.56 (s, 1 H, ArH), 6.66 (s, 1 H, ArH); dC (100 MHz, CDCl3) 21.5 (9-
CH2), 28.6 (5-CH2), 32.8 (2-CH2), 35.9 (1-CH2), 40.1 (8-CH2), 45.7
(6-CH2), 55.9 (OCH3), 56.0 (OCH3), 63.7 (13b-C), 110.2 (ArCH),
112.8 (ArCH), 123.9 (4-CH=), 124.4 (ArC), 125.4 (ArC), 146.9
22 Y. Tsuda, S. Hosoi, N. Katagiri, C. Kaneko and T. Sano, Chem. Pharm.
Bull., 1993, 41, 2087.
23 I. Osante, N. Sotomayor and E. Lete, Lett. Org. Chem., 2004, 1,
323.
24 N. Gauvry, C. Comoy, C. Lescop and F. Huet, Synthesis, 1999,
574.
25 (a) R. H. Grubbs, Tetrahedron, 2004, 60, 7117; (b) K. C. Nicolaou, P. G.
Bulger and D. Sarlah, Angew. Chem., Int. Ed., 2005, 44, 4490.
26 V. Bohrsch, J. Neidhofer and S. Blechert, Angew. Chem., Int. Ed., 2006,
45, 1302.
27 M. Scholl, S. Ding, C. W. Lee and R. H. Grubbs, Org. Lett., 1999, 1,
=
(ArC), 148.4 (ArC), 168.5 (4a-C), 199.3 (C O); ESIMS m/z (%)
953.
300.2 ([M + H]+, 100); HRMS (ESI) found [M + H]+, 300.1586.
28 (a) J. S. Kingsbury, J. P. A. Harrity, P. J. Bonitatebus, Jr. and A. H.
Hoveyda, J. Am. Chem. Soc., 1999, 121, 791; (b) S. B. Garber, J. S.
Kingsbury, B. L. Gray and A. H. Hoveyda, J. Am. Chem. Soc., 2000,
122, 8168.
29 H. C. Kolb, M. S. Van Nieuwenhze and K. B. Sharpless, Chem. Rev.,
1994, 94, 2483.
C18H21NO3 requires [M + H], 300.1594.
Acknowledgements
We gratefully acknowledge the University of Nottingham and
the ORS scheme for support of this work through a student-
ship to FZ.
30 The mechanism of formation of by-product 38 is not clear, but seems to
involve initial metathesis with an in-situ generated ethylidene complex,
=
followed by C C isomerisation, and then metathesis with 2-methyl-2-
butene. See: A. K. Chatterjee, D. P. Sanders and R. H. Grubbs, Organic
Lett., 2002, 4, 1939; S. H. Hong, D. P. Sanders, C. W. Lee and R. H.
Grubbs, J. Am. Chem. Soc., 2005, 127, 17160.
References
31 J. Smidt, W. Hafner, R. Jira, J. Sedlmeier, R. Sieber, R. Ru¨ttinger and
H. Kojer, Angew. Chem., 1959, 71, 176.
32 A. K. Bose, L. Krishnan, D. R. Wagle and M. S. Manhas, Tetrahedron
Lett., 1986, 27, 5955.
33 For analogous reactions of anhydrides, see: R. Shintani and G. C. Fu,
Angew. Chem., Int. Ed., 2002, 41, 1057.
34 B. Butler, T. Schultz and N. S. Simpkins, Chem. Commun.., 2006, 3634.
35 Y. S. Lee, D. W. Kang, S. J. Lee and H. Park, J. Org. Chem., 1995, 60,
7149.
36 S. Louwrier, M. Ostendorf, A. Boom, H. Hiemstra and W. N.
Speckamp, Tetrahedron, 1996, 52, 2603.
37 T. Ohta, S. Shiokawa, R. Sakamoto and S. Nozoe, Tetrahedron Lett.,
1990, 31, 7329.
38 (a) B. Y. He, T. J. Wu, X. Y. Yu and P. Q. Huang, Tetrahedron-
Asymmetry, 2003, 14, 2101; (b) P. Beak and A. I. Meyers, Acc. Chem.
Res., 1986, 19, 356; (c) See alsoJ. L. Ye, P. Q. Huang and X. Lu, J. Org.
Chem., 2007, 72, 35.
1 (a) A. S. Chawla, and V. K. Kapoor, in The Alkaloids: Chemical and
Biological Perspectives, ed. S. W. Pelletier, Pergamon, 1995, Vol 9, p. 86;
(b) Y. Tsuda, and T. Sano, in The Alkaloids, ed. G. A. Cordo, Academic
Press, New York, 1996, Vol 48, p. 249; (c) K. W. Bentley, Nat. Prod.
Rep., 2006, 23, 444; (d) M. E. Amer, M. Shamma and A. J. Freyer,
J. Nat. Prod., 1991, 54, 329.
2 (a) S. F. Dyke, and S. N. Quessy, in The Alkaloids, ed. R. G. A. Rodrigo,
Academic Press, New York, 1981, Vol 18, p 1, see also references 1 and 3;
(b) For a recent medicinal chemistry related report, see: A. K. Ganguly,
C. H. Wang, D. Biswas, J. Misiaszek and A. Micula, Tetrahedron Lett.,
2006, 47, 5539.
3 E. Reimann, Prog. Chem. Org. Nat. Prod., 2007, 88, 1.
4 K. Shimizu, M. Takimoto, Y. Sato and M. Mori, J. Organomet. Chem.,
2006, 691, 5466.
5 Q. Wang and A. Padwa, Org. Lett., 2006, 8, 601.
6 S. H. Gao, Y. Q. Tu, X. D. Hu, S. H. Wang, R. B. Hua, Y. J. Jiang,
Y. M. Zhao, X. H. Fan and S. Y. Zhang, Org. Lett., 2006, 8, 2373.
7 Y. Tsuda, S. Hosoi, K. Ishida and M. Sangai, Chem. Pharm. Bull., 1994,
42, 204.
39 G. Keum and G. Kim, Bull. Korean Chem. Soc., 1994, 15, 278.
40 (a) R. Ben Othman, T. Bousquet, M. Othman and V. Dalla, Org. Lett.,
2005, 7, 5335; (b) R. Ben Othman, T. Bousquet, A. Fousse, M. Othman
and V. Dalla, Org. Lett., 2005, 7, 2825.
8 S. M. Allin, G. B. Streetley, M. Slater, S. L. James and W. P. Martin,
Tetrahedron Lett., 2004, 45, 5493.
41 M. Lennartz and E. Steckhan, Synlett, 2000, 319.
1978 | Org. Biomol. Chem., 2009, 7, 1963–1979
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
42 I. Osante, M. N. Abdullah, S. Arrasate, N. Sotomayor and E. Lete,
Arkivoc, 2007, 206.
43 D. L. J. Clive, D. R. Cheshire and L. Set, J. Chem. Soc., Chem. Commun.,
1987, 353.
46 W. D. Z. Li and Y. Q. Wang, Org. Lett., 2003, 5, 2931.
47 H. H. Wasserman and R. M. Amici, J. Org. Chem., 1989, 54,
5843.
48 Y. Tsuda, Y. Sakai, A. Nakai, M. Kaneko, Y. Ishiguro, K.
Isobe, J. Taga and T. Sano, Chem. Pharm. Bull, 1990, 38,
1462.
44 A. Toshimitsu, T. Aoai, H. Owada, S. Uemura and M. Okano, J. Chem.
Soc., Chem. Commun., 1980, 412.
45 Y. Tsuda, A. Nakai, K. Ito, F. Suzuki and M. Haruna, Heterocycles,
1984, 22, 1817.
49 D. H. R. Barton, A. Gunatila, R. M. Letcher, A. Lobo and D.
Widdowson, J. Chem. Soc., Perkin Trans. 1, 1973, 874.
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 1963–1979 | 1979
©