New asymmetric routes to ajmaline and suaveoline indole alkaloids

Article abstract of DOI:10.1039/b005694o

The tetracyclic advanced intermediates 6,8 and 9 obtained from L-tryptophan via a cis-selective Pictet-Spengler reaction and a Dieckmann-Thorpe cyclisation are used in a range of new approaches to polycyclic monoterpenoid indole alkaloids such as ajmaline and suaveoline. Structural modifications designed to facilitate a key intermolecular addition to C15 (ajmaline numbering) are described, followed by two intramolecular routes based on the addition of a suitable carbon fragment to a remote nitrogen atom prior to bond formation at C15. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/b005694o

View Article Online / Journal Homepage / Table of Contents for this issue
New asymmetric routes to ajmaline and suaveoline indole alkaloids
Patrick D. Bailey,a Keith M. Morgan,a David I. Smithb and John M. Vernonc
a Chemistry Department, Heriot-Watt University, Riccarton, Edinburgh, UK EH14 4AS
b Sanofi-Synthelabo Research, Alnwick, Northumberland, UK NE66 2JH
c Chemistry Department, University of York, Heslington, York, UK YO1 5DD
1
Received (in Cambridge, UK) 14th July 2000, Accepted 1st September 2000
First published as an Advance Article on the web 19th October 2000
The tetracyclic advanced intermediates 6, 8 and 9 obtained from -tryptophan via a cis-selective Pictet–Spengler
reaction and a Dieckmann–Thorpe cyclisation are used in a range of new approaches to polycyclic monoterpenoid
indole alkaloids such as ajmaline and suaveoline. Structural modifications designed to facilitate a key intermolecular
addition to C15 (ajmaline numbering) are described, followed by two intramolecular routes based on the addition of
a suitable carbon fragment to a remote nitrogen atom prior to bond formation at C15.
Introduction
Indole alkaloids are a structurally diverse family of natural
products that attract considerable interest because of their
well established, wide-ranging medicinal properties. Clinically
important examples include the heteroyohimboid alkaloids
such as the vasodilatory ajmalicine,1 the macroline/sarpagine
alkaloids such as the anti-cancer bisindoles vincristine and
vinblastine,2 and the established anti-arrhythmic ajmaline 1.3
The intricate cage-like indole alkaloids typified by ajmaline and
the related alkaloids suaveoline 2a and raumacline 3, have
recently been the focus of much synthetic excitement.
Scheme 1
related indole alkaloids.8 Cook’s asymmetric synthesis of (Ϫ)-
suaveoline started from -tryptophan, and included modified
procedures to improve the overall yield considerably.9 Neverthe-
less, these syntheses use non-proteinogenic -tryptophan as the
starting material, and construction of the carbon skeletons is
lengthy and awkward.
We achieved the formal asymmetric synthesis of ajmaline and
suaveoline in 1993 starting from -tryptophan,10 by exploiting
the cis-selective kinetically controlled Pictet–Spengler reac-
tion;11 our route gave access to the bridged ketone 5 used by
Cook.7,9 We later developed an independent asymmetric route
to the α,β-unsaturated aldehyde 6 (ee у 97%) which is also an
intermediate in Cook’s syntheses, in which the aldehyde carbon
was introduced at the start of the synthesis by homologation
of -tryptophan to the nitrile 7. Further key steps included a
cis selective Pictet–Spengler reaction and a Thorpe cyclisation
to the ketone 8 (Scheme 2).12
However, we wished to develop an independent total syn-
thesis of one or more of these alkaloids, and our new synthetic
approaches are described here and in the following paper
(DOI: 10.1039/b005695m).
Suaveoline 2a was first isolated from the bark of Rauwolfia
suaveolens in 1972,4 but has since been found in other Rauwolfia
species along with structural analogues such as macrophylline
2b.5 Biotransformations in Rauwolfia serpentina cell cultures fed
with ajmaline produce a range of alkaloids including suaveoline
2a, raumacline 3 and alkaloid G 4.6 The biosynthetic inter-
conversion of members of the ajmaline family emphasises the
fact that advanced intermediates can often be exploited in the
synthesis of several alkaloids and analogues.
The first total synthesis of racemic suaveoline was achieved
by Trudell and Cook in 1989; the key steps in this ingenious
synthesis from /-tryptophan (overall yield 2.3%) include
a trans-specific Pictet–Spengler reaction to gain access (after
epimerisation) to the bridged ketone 5, an unusual homo-
logation to the α,β-unsaturated aldehyde 6 (Scheme 1), and
an anionic Cope rearrangement in order to introduce the final
carbons of the suaveoline skeleton.7 Related anionic Cope
rearrangements have since been employed by Cook et al. in
the asymmetric synthesis of ajmaline3 and other structurally
Results and discussion
With a total synthesis of ajmaline and suaveoline in mind
and with homochiral α,β-unsaturated aldehyde 6 available,12
our initial synthetic efforts were directed towards the inter-
molecular 1,4-addition of a suitable four-carbon fragment
to C15 (ajmaline numbering). Earlier work by ourselves13 and
others14 had hinted at the difficulty of this 1,4-addition; for
example, simple alkyl Grignards, enamines and organocuprates
were all known to fail. One set of conditions that is successful
3566
J. Chem. Soc., Perkin Trans. 1, 2000, 3566–3577
This journal is © The Royal Society of Chemistry 2000
DOI: 10.1039/b005694o

View Article Online
with other α,β-unsaturated aldehydes is a Cu() catalysed
Grignard 1,4-addition in which the intermediate enolate is
trapped as a TMS enol ether.15 However, this procedure did
not lead to any 1,4-addition and the only products isolated were
the result of direct 1,2-attack on the aldehyde (Scheme 3).
We therefore decided to look back into our established
synthetic route (Scheme 2) for earlier synthetic intermediates
that might be amenable to bond formation at C15. As a first
approach, the ketone 8 was proposed as a substrate for direct
nucleophilic addition. This material was known to be readily
enolisable, as confirmed by our earlier observation that reduc-
tion with sodium borohydride leads to a single diastereoismer
of the alcohol 9 via the interconvertion of ketone stereo-
isomers.16 However, a suitably non-basic nucleophile would
lead to a much shortened synthetic route and the Reformatsky
addition of 2-bromobutyronitrile to the ketone could lead to
an ideal advanced intermediate with the carbon skeletons
of ajmaline and suaveoline largely in place (Scheme 3). How-
ever, no reaction between the ketone 8 and 2-bromopropio-
nitrile (employed as a model reagent) was observed under
standard Reformatsky conditions, or even under the acceler-
ating influence of sonication17 or when using a highly reactive
form of dispersed zinc, namely a zinc/silver/graphite matrix.18
A simple lithium cuprate addition also failed with this ketone,
again presumably because of the acidity of the α-proton.
We reasoned that conversion of the electron-withdrawing
nitrile to an aldehyde masked as an acetal (as in 13) would
reduce the acidity of the α-proton and thus allow nucleophilic
addition of a suitable four-carbon unit to the ketone func-
tionality. Not surprisingly, direct DIBAL-H reduction of the
ketone 8 gave complex product mixtures, from which no
aldehyde could be isolated. Temporary masking of the ketone
functionality as a silyl enol ether 14 was achieved in high yield
with TBDMSCl and imidazole in DCM (Scheme 4). These
conditions also allowed the selective protection of the ketone 8
in the presence of the alcohol 9, and thus eased the chromato-
graphic separation of these two established intermediates. The
silyl enol ether 14 proved surprisingly inert towards reduction
with DIBAL-H, with no reaction being observed even with
several equivalents of the reducing agent. For another approach
to the acetal 13, DIBAL-H reduction of the nitrile moiety of
the alcohol 9 followed by acetal formation and oxidation to the
Scheme 3 Reagents and conditions: i, MeMgI, CuBrؒMe2S, TMS–Cl,
HMPA, THF, Ϫ78 ЊC (30% for 11a/b and 29% for 12a/b).
Scheme 2
Scheme 4 Reagents and conditions: i, DIBAL-H, DCM, Ϫ78 ЊC; ii, TBDMS–Cl, imidazole, DCM, room temp., 24 h (92%); iii, TBDMS–OTf,
2,6-dimethylpyridine, room temp., 24 h (80%); iv, HO(CH2)2OH, PTSA, reflux, 2 h (65%).
J. Chem. Soc., Perkin Trans. 1, 2000, 3566–3577
3567

View Article Online
ketone was proposed. Direct reduction of the nitrile moiety
of the alcohol 9 again failed, but masking of the alcohol
functionality was achieved by reaction with TBDMSOTf and
2,6-dimethylpyridine in DCM. This protected alcohol was then
treated with DIBAL-H in DCM at Ϫ78 ЊC to afford the corre-
sponding aldehyde which was isolated and immediately pro-
tected as the acetal. However, the forcing conditions required
for complete acetal formation also caused desilylation and
dehydration to afford the unsaturated acetal 17 in low overall
yield, as outlined in Scheme 4. Even the milder conditions
required to form a thioacetal led solely to the formation of the
unsaturated thioacetal 18.
These initially disappointing results led us to propose
two possible applications for the unsaturated acetal 17; firstly,
hydroboration followed by Swern oxidation to afford the
desired protected ketone 13 and secondly, epoxidation of the
double bond. In fact, the acetal 17 was found to be readily
available from the established α,β-unsaturated aldehyde inter-
mediate 6 in 89% yield (Scheme 5), thus offering a potentially
Scheme 6 Reagents and conditions: i, MCPBA, NaHCO3, DCM, 0 ЊC,
30 min (76%); ii, alkaline peroxide, acetone, room temp., 48 h (79%).
group such as butyryl could not be employed as a Nb protecting
group in the early stages of the synthesis because of problems
of carboline ring-opening in the Thorpe cyclisation step.12
Thus, the addition of a butyryl group to a tetracyclic advanced
intermediate such as the nitrile 22 was proposed, as outlined in
Scheme 7.
Scheme 5 Reagents and conditions: i, HO(CH2)2OH, BF3ؒOEt2, DCM,
room temp., 4 h (89%); ii, 9-BBN, THF, room temp., 24 h then alkaline
peroxide, 15 min (55%).
high overall yield for the production of the desired ketone 13.
Hydroboration of the acetal 17 with BH3ؒTHF gave a complex
product mixture by TLC and the reaction was not pursued
further.
Reaction of the acetal 17 with 9-BBN in DCM followed
by an alkaline peroxide work-up afforded a single product.
Unfortunately, this material was not derived from the expected
hydroboration reaction but simply from acetal ring opening.
The thioacetal 18 failed to react under similar (and more
forcing) conditions.
Scheme 7
We were of course concerned that the stereo-electronic
factors might disfavour the desired Michael attack, but the
presumed thermodynamic stability of the cyclised product and
the brevity of the route encouraged us to attempt it. The free
amine 22 proved difficult to obtain: hydrogenation of the Nb-
benzyl α,β-unsaturated nitrile 10 led to reduction of the double
bond prior to benzyl cleavage, whilst cleavage of the benzyl
group from the alcohol 9 led to the formation of the highly
insoluble amino alcohol 25 which could only be dehydrated to
the required amine 22 in poor overall yield (Scheme 8).
Since the amine 22 was not readily available and to avoid the
solubility difficulties encountered with the free amino alcohol
25 the versatile silyl enol ether 14 was again employed en
route to the tetracyclic intermediate 23 proposed in Scheme 7.
Initially, the required debenzylation proved troublesome, but
dissolving the TBDMS enol ether 14 in MeOH, followed
by rigorous rotary evaporation afforded an amorphous solid
which reacted very cleanly to afford a quantitative yield of
the free amine 26 (Scheme 9). This evaporation procedure
presumably removed small traces of chlorinated solvents which
otherwise poisoned the palladium catalyst. The free amine
26 was found to be unstable and was not fully characterised.
Next the acetal 17 was treated with MCPBA in DCM in
an effort to produce the corresponding epoxide which might
itself be open to nucleophilic attack of a suitable four carbon
fragment. TLC analysis of the reaction mixture indicated the
presence of two products, which were separable by careful flash
chromatography. However, instead of the expected epoxides,
both diastereoisomers of the N-oxide 20a/b in the ratio of
ca. 3:1 were isolated. (Scheme 6). Extended reaction times
and/or excess MCPBA produced complex mixtures of products
from which no trace of an epoxide could be isolated. Related
attempts to epoxidise another established synthetic inter-
mediate, the nitrile 10, with alkaline hydrogen peroxide19
also failed, with nitrile hydrolysis20 leading solely to the corre-
sponding α,β-unsaturated amide 21. At this point, the inter-
molecular addition of a suitable four-carbon fragment to
advanced intermediates was abandoned in favour of two new
intramolecular approaches.
Our first intramolecular approach was based on the initial
addition of a four-carbon unit to the Nb nitrogen. In earlier
synthetic studies we had shown that an electron-withdrawing
3568
J. Chem. Soc., Perkin Trans. 1, 2000, 3566–3577

View Article Online
We envisaged a second intramolecular route towards suaveo-
line 2a, again starting from the TBDMS enol ether 14. The first
step proposed was a simple LAH reduction of the nitrile moiety
of the silyl enol ether 14 to the corresponding amine. The amine
30 would then be ideally set up for condensation with butyr-
aldehyde, or an equivalent, and following desilylation, intra-
molecular enamine cyclisation to the ketone functionality
at C15 (Scheme 10). However, an initial small scale LAH
Scheme 8 Reagents and conditions: i, Pd–C, H2, MeOH (80% for 24
and 67% for 25); ii, POCl3, pyridine, benzene, reflux, 48 h (25%).
Reaction with butyryl chloride to afford the amide 27, pro-
ceeded in good yield. A small quantity of the ester 28 was
also isolated, presumably via desilylation and trapping of the
resultant ketone. Desilylation of the TBDMS enol 27 with
TBAF proceeded quantitatively to afford the corresponding
ketone 29. As expected, the NMR spectra of this compound
were very broad, presumably due to keto–enol equilibration
and the presence of amide rotamers about Nb. It was envisaged
that reduction–dehydration of this ketone would proceed
in high overall yield, based on our earlier work with the Nb-
benzyl series.12 However, sodium borohydride reduction of the
ketone 29 in methanol afforded complex mixtures of products
from which no corresponding alcohol was readily isolated.
Dehydration of the crude reaction mixtures with phosphoryl
chloride and pyridine in refluxing benzene again afforded
complex mixtures of products. However, some of the α,β-
unsaturated nitrile 23 could be obtained in low yield by careful
flash chromatography, along with some recovered ketone 29
(Scheme 9).
Scheme 10 Reagents (for conditions see text): i, LAH, THF.
reduction (over 24 h) gave a single product which was not the
expected amine 30, but was eventually identified as the vinylo-
gous amide 31 (45% yield). A larger scale reduction (over 72 h)
afforded some of the same product 31 in 20% yield, along
with a higher running component in 55% yield which was
assigned as the ketone 32. Unfortunately, this material rapidly
dimerised, thus preventing full characterisation. The structure
of the dimer (Mr = 684) remains unknown due to the com-
plexity of the available NMR data. Another larger scale reduc-
tion, carefully monitored to ensure that just enough LAH was
added to consume the starting material, afforded some of
the vinylogous amide 31 in 26% yield but mostly returned the
desilylated ketone 8 in 67% yield. Thus, a possible mechanism
for the formation of the vinylogous amide is via the hydride
promoted desilylation of the intermediate aldimine.
We reasoned that the vinylogous amide 31 was structurally
close to suaveoline and could also undergo the enamine
reaction postulated in Scheme 10. Unfortunately no reaction
with butyraldehyde was observed under standard conditions
for imine formation. Thus we attempted to increase the nucleo-
philicity of the nitrogen by reducing the double bond. However,
the vinylogous amide 31 proved surprisingly robust towards
hydrogenation in methanol; an excess of 10% Pd–C catalyst
along with a 24 h reaction time was required to effect complete
consumption of the starting material. A single product was
obtained in 66% yield (Scheme 11). Remarkably, NMR indi-
cated the presence of 3 methyl groups and this material was
identified as the ketone 33. This result confirms a report by
Cook and Fu14a where an Nb-benzyl in related systems is con-
verted to Nb-methyl under similar conditions. A full mechanism
to account for this transformation is elusive, but presumably
involves formaldehyde, which can be generated from the solvent
under similar reaction conditions.21
Scheme 9 Reagents and conditions: i, Pd–C, H2, MeOH, 45 min; ii,
PrCOCl, Cs2CO3, DCM, room temp., 30 min (61% overall for 27 and
17% for 28); iii, TBAF, MeOH, room temp., 30 min (100%); iv, NaBH4,
MeOH, room temp., 72 h; v, POCl3, pyridine, benzene, reflux, 24 h
(24% overall from 29).
The base-induced ring closure postulated in Scheme 7 was
now attempted with 1, 2, 5 and 10 equivalents of LDA in
THF. No reaction was observed at Ϫ78 ЊC, and at higher
temperatures steady decomposition of starting material was
the only observed process. The amides 27 and 29 were also
subjected to the same conditions, but again no ring closure
was observed. The failure of these reactions was presumably
due to the conformational constraints placed upon the ring
system by the amide functionality and in light of the sur-
prisingly difficult and low yielding sequence to the Nb-butyryl
α,β-unsaturated nitrile 23, this approach was not investigated
further.
Nevertheless, by returning to the hydroxynitrile 9, reduction
of the nitrile to the primary amine was more straightforward,
J. Chem. Soc., Perkin Trans. 1, 2000, 3566–3577
3569

View Article Online
Experimental
Melting points were determined on a Reichert microscope
hot-stage apparatus, and are uncorrected. Elemental analysis
was performed by CHN Ltd. NMR spectra were recorded
on a Bruker WP80 at 80 MHz (1H), a JEOL FX90 at 90 MHz
(1H) and 22.5 MHz (13C), a Bruker MSL300 spectrometer at
300 MHz (1H) and 75 MHz (13C), a Bruker WP200 at 200 MHz
(1H) and 50 MHz (13C) or a Bruker DPX400 at 400 MHz (1H)
and 100 MHz (13C). Chemical shifts were measured in ppm
on the δ scale downfield from tetramethylsilane as internal
standard. The solvent employed was CDCl3, unless otherwise
stated. When pairs of diastereoisomers were not separated the
Scheme 11 Reagents and conditions: i, Pd–C, H2, MeOH, 24 h (66%).
diastereomeric NMR resonances are bracketed thus {}. All 13
C
1
data are quoted with H multiplicities (off resonance results)
in brackets thus (), although this multiplicity was inferred
from DEPT experiments. Infrared spectra were recorded on a
Perkin-Elmer 1420 ratio recording spectrophotometer. Optical
rotations were recorded on a Perkin-Elmer 141 polarimeter at
room temp. and are reported in units of 10Ϫ1 deg cm2 gϪ1
.
Mass spectra were obtained by electron impact at 70 eV on an
AEI MS-3074, or a VG AUTOSPEC spectrometer. Analytical
TLC was carried out on Merck aluminium sheet silica gel
60 F254 plates (thickness 0.2 mm). Spots were visualised with a
UV hand lamp. Flash chromatography was performed using
silica gel 60 (230–400 mesh) as the stationary phase, purchased
from Camlab.
THF was dried by distillation from sodium or obtained
anhydrous from Aldrich. DMF was obtained anhydrous from
Aldrich and was stored under dry argon. HPLC grade CHCl3,
DCM and MeOH used in the large scale preparations were
obtained from Rathburn Chemicals Ltd. For small scale work,
chlorinated solvents were distilled from phosphorus penta-
oxide. Benzene, toluene and ethoxyethane were distilled and
stored over sodium. Pyridine and 2,6-dimethylpyridine were
distilled and stored over potassium hydroxide pellets. Triethyl-
amine and phosphoryl chloride were distilled and stored under
dry argon. LAH 1 M solution in THF, DIBAL-H 1 M solution
in DCM, 9-BBN 0.5 M solution in THF, BH3ؒTHF 1 M
solution in THF, LDA 1.5 M solution in THF and n-BuLi
2.5 M solution in hexanes were obtained from Aldrich and
stored under dry argon.
Scheme 12 Reagents and conditions: i, LAH, THF, room temp. 24 h
then reflux 5 h; ii, BOC2O, NaHCO3, CHCl3, room temp., 90 min (64%
overall from 9); iii, PCC–Florosil, DCM, room temp., 24 h (41%); iv,
1:9 water–TFA, room temp., 1 h then PrCHO, DCM, 72 h.
and selective protection to the N-Boc derivative 34 proceeded
in an overall yield of 65% from the alcohol 9 (Scheme 12). Next
we envisaged a simple oxidation from the alcohol back to the
ketone, so that this slightly circuitous route could proceed
in high overall yield. Standard Swern and TPAP oxidation con-
ditions failed, but the use of excess Swern reagent did yield
some of the ketone 35 (37%), along with a complex mixture
of rearrangement products which could not be characterised
(although we have since applied related Swern conditions in
synthetically valuable oxidations and alkylations of tryptamine
derivatives).22 Oxidation of the alcohol to the ketone 35 was
therefore more conveniently achieved using PCC. Unexpec-
tedly, this reaction also generated the 3Ј-cyclised derivative 36
as the major product,23 which bears an interesting structural
similarity to alkaloid G 4 (Scheme 12). Finally, deprotection of
35 with TFA afforded an unstable ketone (which could not be
isolated and characterised), but which was treated with butanal
in the presence of oxygen in an effort to promote the intra-
molecular enamine reaction proposed earlier. Although a
complex reaction mixture was generated, traces of Nb-benzyl-
suaveoline 2c could be identified by TLC/MS comparison with
authentic material.9,14a
Attempted Michael addition to the tetracyclic aldehyde 6
Methyl iodide (250 mg, 1.75 mmol) was added dropwise to
clean magnesium turnings (40.5 mg, 1.75 mmol) in dry THF
(1.5 ml) at room temp. When all the magnesium was consumed
(ca. 10 min) HMPA (540 mg, 3.0 mmol) and CuBrؒMe2S
(1.5 mg, 7 µmol) were added and a portion of the resulting
solution (200 µl, ca. 0.17 mmol, 1.5 eq. of Grignard reagent)
was added dropwise to a solution of the aldehyde 6 (40 mg,
0.12 mmol) and TMSCl (33 mg, 0.3 mmol) in THF (1 ml)
maintained at Ϫ78 ЊC. After 4 h TLC indicated complete con-
sumption of starting material and saturated aqueous NH4Cl
was added to quench the reaction. Upon warming to room
temperature, the solution was extracted 3 times with EtOAc
and the combined organic extracts were washed with water and
then dried over MgSO4. Filtration and evaporation afforded
a yellow oil (39 mg). Flash chromatography on silica eluted
with a solvent gradient (DCM to 1:9 ether–DCM) afforded
2 pairs of diastereoisomers (each ca. 1:1): the free alcohol 11a/b
(15 mg, 30%) and the TMS ether 12a/b (12 mg, 29%).
Conclusion
In this paper we present a range of transformations from
bridged intermediates such as 6, 8 and 9. A number of unusual
reactions were observed, in particular: (a) the hydride-promoted
desilylation leading to a range of compounds including the
vinylogous amide 31, (b) the reductive alkylations occurring
during the hydrogenation of 31, (c) the formation of the 3Ј-
cyclised derivative 36. We were successful in the production of
small amounts of Nb-benzylsuaveoline 2c; however, problems
with introducing the required carbons at C15 are a specific issue
that needs addressing. Our efficient solution to this long term
problem is described in the next paper.
Data for the alcohol 11a/b. Rf 0.11 and 0.07 (1:9 ether–
1
CHCl3); H NMR (300 MHz) δ 1.29 (3H, dd, J 6.5, 2.4 Hz,
CHCH3), 1.97–2.05 (1H, m, one of ArCHCH2), 2.67–2.81
(2H, m, one of ArCHCH2 and one of ArCH2CH), 3.07–3.15
(1H, m, one of ArCH2CH), {3.54, 3.56} (3H, s, indole NCH3),
3.62–4.01 (4H, m, ArCHCH2 and NCH2Ph and ArCH2CH),
3570
J. Chem. Soc., Perkin Trans. 1, 2000, 3566–3577

View Article Online
4.23 (1H, q, J 6.5 Hz, CHOH), {5.61, 5.71} (1H, br s,
ArCHCH2CH), 7.05–7.49 (9H, m, ArH); 13C NMR (75 MHz)
δ 20.8 (q), {22.3, 23.0} (t), 28.9 (q), {29.4, 30.3} (t), {48.6, 47.4}
(d), {52.2, 52.5} (d), 56.3 (t), 68.8 (d), 104.9 (s), 108.9 (d),
118.1 (d), 118.9 (d), 119.4 (d), 121.0 (d), 127.1 (d), 128.3 (d),
128.8 (d), 135.4 (s), 138.1 (s), 138.9 (s), 142.6 (s), 144.1 (s);
IR νmax/cmϪ1 (CHCl3) 1700, 1670, 1620, 1470; MS m/z 358
(Mϩ, 100%), 343 (15), 313 (32), 273 (21), 249 (64), 222 (26),
181 (40), 157 (19), 144 (21), 91 (96); found m/z 358.2048, calc.
for C24H26N2O 358.2045.
to Ϫ78 ЊC. Sufficient DIBAL-H to consume the starting
material (3.5 eq. of 1 M solution in DCM) was added dropwise.
After stirring for 30 min the reaction was allowed to warm
to room temperature and stirred for a further 1 h. The solution
was then cooled to Ϫ78 ЊC and MeOH, saturated NH4Cl
(aq), and 0.1 M aqueous sulfuric acid were added sequentially
and the mixture allowed to warm to room temperature.
The organic layer was separated, washed with brine, dried
over MgSO4, filtered and evaporated to afford a yellow oil.
TLC indicated that these oils contained at least 4 products
and none of the desired aldehydes could be isolated from the
mixtures.
Data for 12a/b. Rf 0.48 and 0.38 (1:9 ether–CHCl3); 1H
NMR (300 MHz) δ 0.03 (9H, s, SiC(CH3)3), 1.24 (3H, d, J 6.4
Hz, CHCH3), 2.00 (1H, dd, J 13.5, 5.5 Hz, one of ArCHCH2),
2.67–2.78 (2H, m, one of ArCHCH2 and one of ArCH2CH),
3.05 (1H, dt, J 15.9, 7.4 Hz, one of ArCH2CH), 3.57 (3H, s,
indole NCH3), 3.65–3.82 (3H, m, NCH2Ph and ArCH2CH),
4.03 (1H, d, J 5.9 Hz, ArCHCH2), 4.19–4.27 (1H, m, CHOT-
MS), {5.52, 5.62} (1H, br s, ArCHCH2CH), 7.07–7.48 (9H, m,
ArH); 13C NMR (75 MHz) δ 0.1 (q), 21.7 (q), 22.4 (t), 28.8
(q), 29.6 (t), 48.4 (d), 50.9 (d), 57.6 (t), {68.6, 69.1} (d), 104.8
(s), 108.2 (d), 117.4 (d), 117.6 (d), 118.5 (d), 120.2 (d), 126.7
(d), 127.5 (d), 128.5 (d), 135.2 (s), 136.5 (s), 138.6 (s), 141.7 (s),
142.6 (s); IR νmax/cmϪ1 (CHCl3) 1525, 1425; MS m/z 430 (Mϩ,
65%), 415 (87), 339 (17), 313 (19), 273 (23), 249 (24), 181 (31),
144 (22), 91 (100); found m/z 430.2440, calc. for C27H34N2OSi
430.2426.
Preparation of (6S,10S)-12-benzyl-9-cyano-8-tert-butyldimethyl-
silyloxy-5-methyl-6,7,10,11-tetrahydro-5H-6,10-iminocyclo-
octa[b]indole 14 from the pure ketone 8
The ketone 8 (25 mg, 0.07 mmol) was dissolved in dry DCM
(10 ml) and stirred at 0 ЊC. Imidazole (10 mg, 0.14 mmol) and
TBDMSCl (22 mg, 0.14 mmol) were added and the resulting
solution allowed to warm to room temperature and stirred
for 24 h. The solution was then washed with brine, dried over
MgSO4, filtered and evaporated to afford a yellow foam.
Flash chromatography on silica eluted with DCM afforded the
TBDMS enol ether ether 14 as colourless foam (30 mg, 92%).
Rf 0.88 (1:9 ether–CHCl3); 1H NMR (300 MHz) δ 0.10–0.15
(6H, m, Si(CH3)2), 0.89–0.91 (9H, m, SiC(CH3)3), 2.03 (1H, d,
J 17.6 Hz, one of ArCHCH2), 2.77–2.90 (1H, m, one of
ArCHCH2), 2.86 (1H, d, J 15.8 Hz, one of ArCH2CH), 3.16
(1H, dd, J 15.8, 5.3 Hz, one of ArCH2CH), 3.57 (3H, s, indole
NCH3), 3.67–3.84 (2H, m, NCH2Ph), 3.94 (1H, d, J 5.3 Hz,
ArCH2CH), 4.12 (1H, br s, ArCHCH2), 7.11–7.54 (9H, m,
ArH); 13C NMR (75 MHz) δ Ϫ4.0 (q), Ϫ3.9 (q), Ϫ3.6 (q), 18.0
(s), 23.5 (t), 25.3 (q), 25.6 (q), 29.3 (q), 34.4 (t), 48.9 (d), 53.2 (d),
56.4 (t), 93.6 (s), 104.6 (s), 108.8 (d), 118.5 (d), 119.4 (d), 121.6
(s), 126.8 (s), 127.6 (d), 128.6 (d), 128.7 (d), 133.5 (s), 137.2 (s),
162.6 (s); IR νmax/cmϪ1 (CHCl3) 2205, 1635, 1495, 1470, 1415,
1370; MS m/z 469 (Mϩ, 44%), 412 (88), 378 (16), 363 (7), 337
(15), 312 (14), 295 (11), 273 (17), 246 (22), 91 (100), 73 (82);
found m/z 469.2550, calc. for C29H35N3OSi 469.2549.
Attempted Reformatsky reactions with the ketone 8
The ketone 8 (35.5 mg, 0.10 mmol), 2-bromopropionitrile
(19 mg, 0.14 mmol) and an excess of acid washed zinc powder
were suspended in dioxane (3 ml). The flask was suspended in a
sonic bath for 2 h and then left to stand for 24 h. TLC indicated
that no reaction had occurred.
Formation of zinc/silver/graphite matrix. Graphite (300 mg,
25 mmol) was degassed under a stream of dry argon for 10 min
at 150 ЊC and then cooled to room temp. Finely divided, freshly
cut potassium metal (120 mg, 3.06 mmol) was added and
the mixture heated until the metal melted (oil bath ca. 175 ЊC).
Vigorous stirring then afforded a bronze powder of KC8. After
cooling to room temp. dry THF (10 ml) was added, followed
by anhydrous ZnCl2 (200 mg, 1.47 mmol) and AgOAc (20 mg,
0.12 mmol) in small portions. The resulting mixture was
refluxed with stirring to afford a black suspension of zinc/silver/
graphite which was used immediately.
For a second Reformatsky attempt, the ketone 8 (230 mg,
0.65 mmol) and 2-bromopropionitrile (19 mg, 0.14 mmol) were
added to freshly prepared zinc/silver/graphite (1.47 mmol
of zinc) in THF at Ϫ20 ЊC. After stirring for 1 h the reaction
was allowed to warm to room temperature and stirred for 72 h.
TLC indicated that no reaction had occurred.
Large-scale formation of the TBDMS enol ether 14 from a
mixture of the ketone 8 and the alcohol 9
A mixture of ketone 8 and alcohol 9 (1 g) was stirred in dry
DCM (40 ml) and imidazole (250 mg, 3.7 mmol) was added.
When the imidazole had fully dissolved TBDMSCl (426 mg,
2.82 mmol) was added in portions. After 1 h TLC indicated that
the reaction was complete. Work-up as for the small scale pro-
cedure above followed by flash chromatography on silica eluted
with DCM afforded the TBDMS enol ether 14 (894 mg) and
the alcohol 9 (317 mg) both as colourless foams in essentially
quantitative overall yield. The spectral data for this TBDMS
enol ether were identical to those quoted above and evaporation
of a methanolic solution afforded a white solid which was
amenable to elemental analysis.
Attempted cuprate addition to the tetracyclic ketone 8
Mp 149–150 ЊC (MeOH); [α]D Ϫ140 (c 1, MeOH); found
C 74.12, H 7.52; and N 9.00; calc. C 74.16, H 7.51; and N
8.95%.
n-BuLi (0.42 ml of 2.5 M solution in hexanes) was added to
a suspension of CuBrؒMe2S (43.2 mg, 0.21 mmol) in dry THF
(2 ml) maintained at Ϫ78 ЊC. This mixture was allowed to
warm to ca. Ϫ40 ЊC (until homogeneous) and again cooled to
Ϫ78 ЊC, then a solution of the ketone 8 (40 mg, 0.11 mmol) in
dry THF (2 ml) was added dropwise. TLC analysis indicated
that no reaction had occurred after 2 h, and so the mixture
was allowed to warm gradually to room temperature. After a
further 24 h the only process observed was slight decomposition
of the ketone 8.
Attempted DIBAL-H reduction of the TBDMS enol ether 14
The TBDMS enol ether 14 (150 mg, 0.32 mmol) was dissolved
in dry DCM (2 ml) and cooled to Ϫ78 ЊC. DIBAL-H (1.5 ml of
1 M solution in DCM, 1.5 mmol, 5 eq.) was added dropwise
and the solution stirred at this temperature for 30 min. The
reaction was then allowed to warm to room temperature
and stirred for 24 h. No reaction was observed by TLC and
work-up as for the previous DIBAL-H reductions afforded
starting material, contaminated with traces of decomposition
products.
General procedure for the attempted DIBAL-H reduction of the
ketone 8 or the alcohol 9
The tetracyclic material was dissolved in dry DCM and cooled
J. Chem. Soc., Perkin Trans. 1, 2000, 3566–3577
3571

View Article Online
Preparation of (6S,8S,9R,10S)-12-benzyl-9-cyano-8-tert-butyl-
dimethylsilyloxy-5-methyl-6,7,8,9,10,11-hexahydro-5H-6,10-
iminocycloocta[b]indole 15
Hz, NCH2Ph), 3.83–4.03 (6H, m, ArCH2CH and ArCHCH2
and CH(OCH2)2), 5.19 (1H, s, CH(OCH2)2), 5.89 (1H, br d,
J 2.9 Hz, ArCHCH2CH), 7.04–7.51 (9H, m, ArH); 13C NMR
(75 MHz) δ 23.1 (t), 29.1 (q), 29.6 (t), 38.3 (t), 48.2 (d), 51.1
(d), 59.4 (t), 64.8 (t), 104.9 (d), 105.3 (s), 108.6 (d), 118.2 (d),
118.7 (d), 120.8 (d), 124.1 (d), 126.9 (d), 127.1 (s), 128.2 (d),
128.8 (d), 135.4 (s), 136.1 (s), 137.0 (s), 139.0 (s); IR νmax/cmϪ1
(CHCl3) 1675, 1495, 1470, 1370, 1320; MS m/z 386 (Mϩ, 100%),
342 (7), 313 (45), 295 (30), 273 (32), 251 (17), 222 (21), 207
(14), 181 (46), 91 (67); found m/z 386.1989, calc. for C25H26N2O2
386.1994.
The alcohol 9 (300 mg, 0.84 mmol) was dissolved in DCM
(10 ml) and 2,6-dimethylpyridine added (270 mg, 2.52 mmol)
followed by TBDMSOTf (444 mg, 1.68 mmol). After stirring
at room temp. for 24 h TLC indicated the reaction was com-
plete, evaporation and flash chromatography on silica eluted
with DCM then afforded the TBDMS protected alcohol 15 as a
white foam (315 mg, 80%).
1
Rf 0.88 (1:9 ether–CHCl3); H NMR (300 MHz) δ Ϫ0.36
The aldehyde 16 (200 mg, 0.42 mmol) and ethane-1,2-dithiol
(119 mg, 1.26 mmol, 3 eq.) were dissolved in DCM and
BF3ؒOEt2 (59.6 mg, 52 ml, 0.42 mmol, 1 eq.) was added slowly
and the reaction stirred for 24 h at room temp., when TLC
indicated complete consumption of starting material. Aqueous
saturated NaHCO3 was added and the organic layer separated,
washed with saturated NaHCO3 (aq), dried over MgSO4,
filtered and evaporated to afford a smelly yellow oil. Flash
chromatography on silica eluted with DCM afforded the
unsaturated dithioacetal 18 as a pale yellow foam (115 mg,
66%).
(3H, s, one of Si(CH3)2), Ϫ0.14 (3H, s, one of Si(CH3)2),
0.33 (9H, s, SiC(CH3)3), 1.91 (1H, dt, J 14.1, 2.4 Hz, one of
ArCHCH2), 2.13 (1H, dt, J 14.1 4.1 Hz, one of ArCHCH2),
3.08 (1H, d, J 16.4 Hz, one of ArCH2CH), 3.18 (1H, dd,
J 17.0, 7.0 Hz, one of ArCH2CH), 3.33 (1H, dd, J 5.3, 4.1 Hz,
CHCN), 3.54 (3H, s, indole NCH3), 3.54–3.57 (1H, m, ArCH2-
CH), 3.60–3.70 (2H, ABq, J 13.5 Hz, NCH2Ph), 3.87 (1H, dd,
J 4.8, 2.4 Hz, ArCHCH2), 4.22–4.24 (1H, m, CHOTBDMS),
7.05–7.34 (8H, m, ArH), 7.48 (1H, d, J 7.6Hz, ArH); 13C NMR
(75 MHz) δ Ϫ5.5 (q), Ϫ5.2 (q), 17.4 (s), 18.9 (t), 24.9 (q), 28.9
(q), 37.0 (t), 40.9 (d), 47.3 (d), 51.7 (d), 57.6 (t), 64.6 (d), 105.5
(s), 108.5 (d), 118.2 (d), 118.7 (d), 119.8 (s), 120.6 (d), 127.0 (s),
127.3 (d), 128.4 (d), 128.6 (d), 134.0 (s), 137.0 (s), 138.4 (s);
IR νmax/cmϪ1 (CHCl3) 2200, 1495, 1420, 1415, 1370; MS m/z
471 (Mϩ, 100%), 414 (15), 348 (10), 339 (12), 313 (13), 273
(81), 222 (17), 183 (16), 170 (26), 154 (37), 91 (67), 75 (23);
found m/z 471.2706, calc. for C29H37N3OSi 471.2706.
Rf 0.78 (1:9 ether–CHCl3); 1H NMR (300 MHz) δ 1.95 (1H,
dd, J 17.6, 5.3 Hz, one of ArCHCH2), 2.61–2.70 (1H, m, one
of ArCHCH2), 2.83 (1H, d, J 16.9 Hz, one of ArCH2CH),
2.97–3.22 (5H, m, CH(SCH2)2 and one of ArCH2CH), 3.46
(3H, s, indole NCH3), 3.59–3.73 (2H, AB system, J 13.5 Hz,
NCH2Ph), 3.74 (1H, d, J 5.9 Hz, ArCH2CH), 3.95 (1H, d,
J 5.9 Hz, ArCHCH2), 5.02 (1H, s, CH(SCH2)2), 5.95–5.98
(1H, m, ArCHCH2CH), 6.95–7.42 (9H, m, ArH); 13C NMR
(75 MHz) δ 23.6 (t), 29.1 (q), 29.4 (t), 38.3 (t), 39.1 (t), 48.2 (d),
53.8 (d), 56.2 (d), 56.3 (t), 105.1 (s), 108.6 (d), 118.0 (d), 118.7
(d), 120.8 (d), 121.3 (d), 126.9 (s), 126.9 (d), 128.1 (d), 128.8 (d),
135.8 (s), 136.8 (s), 139.9 (s), 138.8 (s); IR νmax/cmϪ1 (CHCl3)
1510, 1470, 1380; MS m/z 418 (Mϩ, 39%), 357 (34), 327 (14),
313 (32), 273 (30), 181 (29), 91 (100); found m/z 418.1536, calc.
for C25H26N2S2 418.1537.
Preparation of (6S,8S,9R,10S)-12-benzyl-8-tert-butyldimethyl-
silyloxy-5-methyl-6,7,8,9,10,11-hexahydro-5H-6,10-imino-
cycloocta[b]indole-9-carbaldehyde 16
The TBDMS ether 15 (144 mg, 0.31 mmol) was dissolved in
DCM (5 ml) and cooled to Ϫ78 ЊC. DIBAL-H (0.9 ml of 1 M
solution, 0.9 mmol, 3 eq.) was added dropwise and the reaction
stirred for 1 h at this temperature, then for 1.5 h at room temp.
Work-up as for the previous attempted DIBAL-H reductions
followed by flash chromatography on silica eluted with DCM
afforded the aldehyde 16 (125 mg, 85%). Since we envisaged this
material being less stable than the related acetal or nitrile, an
exhaustive characterisation was not undertaken.
Acetal 17 formation from the ꢀ,ꢁ-unsaturated aldehyde 6
The α,β-unsaturated aldehyde 6 (303 mg, 0.88 mmol) and excess
ethylene glycol (1 ml) were stirred in dry DCM (20 ml) at room
temp. BF3ؒOEt2 (640 mg, 4.4 mmol, 5 eq.) was added dropwise
and the reaction stirred at ambient temperature for 4 h, when
TLC indicated complete consumption of starting material.
Saturated NaHCO3 (aq) was added and the organic layer
separated, washed sequentially with NaHCO3 (aq) and brine,
then dried over MgSO4. Silica (250 mg) was added to remove
baseline colouration. Filtration and evaporation then afforded
more of the acetal 17 as a white foam (302 mg, 89%).
Rf 0.59 (1:9 ether–CHCl3); 1H NMR (80 MHz) δ Ϫ0.51 (3H,
s, one of Si(CH3)2), Ϫ0.17 (3H, s, one of Si(CH3)2), 0.28 (9H, s,
SiC(CH3)3), 1.78–2.19 (2H, m, ArCHCH2), 2.68–3.37 (4H, m,
ArCH2CH and CHCHO), 3.52 (3H, s, indole NCH3), 3.58–3.89
(3H, m, CH2Ph and ArCH), 4.56–4.81 (1H, m, CHOTBDMS),
6.87–7.54 (9H, m, ArH), 9.65 (1H, s, CHO); MS m/z 474 (Mϩ,
56%), 342 (16), 313 (25), 273 (87), 251 (13), 181 (45), 170 (26),
129 (34), 91 (100), 75 (96); found m/z 474.2689, calc. for
C29H38N2O2Si 474.2703.
Attempted hydroboration of the acetal 17 with BH3ؒTHF
The α,β-unsaturated acetal 17 (27 mg, 0.07 mmol) was dis-
solved in dry THF (2 ml) and cooled to 0 ЊC. BH3ؒTHF (0.5 ml
of 1 M solution in THF, 0.5 mmol) was added dropwise and
the reaction stirred for 24 h. TLC at regular intervals indicated
that a mixture of products was produced simultaneously, and
this reaction was not pursued further.
Formation of (6S,9R,10S)-12-benzyl-5-methyl-6,7,10,11-tetra-
hydro-5H-6,10-iminocycloocta[b]indole-9-carbaldehyde ethylene
acetal 17 and ethylene dithioacetal 18
The aldehyde 16 (30 mg, 0.06 mmol) was dissolved in ethylene
glycol (10 ml) and refluxed for 2 h with a trace of PTSA, by
which time TLC indicated complete consumption of starting
material. The reaction mixture was diluted with benzene and
washed twice with aqueous 1 M NaHCO3 then once with
brine. The organic layer was dried over MgSO4, filtered and
evaporated to afford a brown oil. Flash chromatography on
silica eluted with DCM afforded the unsaturated acetal 17
(15 mg, 65%) as a white foam.
Attempted hydroboration of the acetal 17 with 9-BBN
To the acetal 17 (64 mg, 0.16 mmol) was added 9-BBN (0.5 ml
of 0.5 M solution in THF, 0.25 mmol). The resulting solution
was stirred at ambient temperature for 24 h then aqueous 2 M
NaOH (0.5 ml) and aqueous 35% H2O2 (0.2 ml) were added and
the solution stirred for a further 15 min. The reaction mixture
was then extracted into benzene, washed with water then brine,
dried over MgSO4, filtered and evaporated to afford a yellow oil
(70 mg). Flash chromatography on silica eluted with a solvent
gradient (DCM to 1:9 ether–DCM) afforded the ring-opened
Rf 0.39 (1:9 ether–CHCl3); 1H NMR (300 MHz) δ 2.05 (1H,
dd, J 17.6, 5.3 Hz, one of ArCHCH2), 2.73 (1H, br dd, J 17.6,
5.3 Hz, one of ArCHCH2), 2.93 (1H, d, J 15.8 Hz, one of
ArCH2CH), 3.08 (1H, dd, J 15.8, 5.9 Hz, one of ArCH2CH),
3.54 (3H, s, indole NCH3), 3.67–3.81 (2H, AB system, J 13.5
3572
J. Chem. Soc., Perkin Trans. 1, 2000, 3566–3577

View Article Online
material 19 as a white foam (34 mg, 55%) and also some
starting material 17 (9 mg, 14%).
Rf 0.58 (1:9 MeOH–CHCl3); H NMR (300 MHz) δ 2.04
133.1 (d), 135.5 (s), 137.4 (s); IR νmax/cmϪ1 (CHCl3) 1600, 1495,
1470, 1455, 1400, 1385, 1365, 1350; MS m/z 402 (Mϩ, 56%), 386
(13), 313 (7), 311 (16), 295 (100), 251 (8), 242 (54), 223 (19), 208
(15), 194 (14), 181 (24), 91 (68), 73 (49); found m/z 402.1953,
calc. for C25H26N2O3 402.1943.
1
(1H, dd, J 17.3, 4.4 Hz, one of ArCHCH2), 2.71 (1H, d, J 15.8
Hz, one of ArCH2CH), 2.80 (1H, br d, J 17.3 Hz, one of
ArCHCH2), 3.10 (1H, dd, J 15.8, 5.9 Hz, one of ArCH2CH),
3.41–3.57 (2H, m, CH2OCH2CH2OH), 3.58 (3H, s, indole
NCH3), 3.68–3.69 (3H, m, ArCH2CH and CH2OCH2CH2OH),
3.66–3.82 (2H, ABq, J 5.9 Hz, CCH2O), 3.80–3.86 (2H, ABq,
J 13.5 Hz, NCH2Ph), 4.05–4.11 (2H, m, ArCH and OH), 5.64
(1H, br s, ArCHCH2CH), 7.07–7.51 (9H, m, ArH); 13C NMR
(75 MHz) δ 21.6 (t), 29.3 (q), 30.4 (t), 48.4 (d), 52.3 (d), 56.4 (t),
61.9 (t), 70.8 (t), 73.2 (t), 105.0 (s), 108.8 (d), 118.0 (d), 118.9
(d), 121.0 (d), 122.4 (d), 126.9 (s), 127.1 (d), 128.3 (d), 128.8 (d),
135.0 (s), 136.5 (s), 137.0 (s), 138.6 (s); IR νmax/cmϪ1 (CHCl3)
1705, 1655, 1615, 1495, 1470, 1380; MS m/z 388 (Mϩ, 100%),
344 (8), 327 (37), 297 (14), 273 (19), 235 (58), 220 (24), 181 (57),
91 (66); found m/z 388.2153, calc. for C25H28N2O2 388.2151.
Attempted epoxide formation from the N-oxides 20a/b
The N-oxides 20a/b (20 mg, 0.05 mmol) were dissolved in DCM
(2 ml) and stirred over excess NaHCO3 at 0 ЊC. MCPBA (28 mg
of 50% pure, 0.08 mmol) was added and the reaction stirred for
24 h. TLC during this time indicated that a complex mixture
of products was produced. The same work-up procedure as the
previous reaction was employed to afford a yellow oil. None
the desired epoxide could be detected, or isolated from these
products.
Attempted epoxide formation from the ꢀ,ꢁ-unsaturated nitrile 10:
formation of the amide 21
Attempted hydroboration of the dithioacetal 18 with 9-BBN
The nitrile 10 (25 mg, 0.07 mmol) was dissolved in acetone
(1 ml). Aqueous 2 M NaOH (0.18 ml) was added, followed
by dropwise addition of aqueous 35% H2O2 (0.15 ml). After
stirring at ambient temperature for 48 h the solution was
diluted with saturated NaHCO3 (aq), extracted into DCM,
washed with brine and dried over MgSO4. Filtration and
evaporation afforded a yellow oil (30 mg). Flash chroma-
tography on silica eluted with 1:9 MeOH–DCM afforded the
amide 21 as a white foam (21 mg, 79%).
To the dithioacetal 18 (50 mg, 0.12 mmol) was added 9-BBN
(0.5 ml of 0.5 M solution in THF, 0.25 mmol) and the resulting
solution stirred at ambient temperature for 24 h. When TLC
indicated that no reaction had occurred 9-BBN (1 ml of 0.5 M
solution in THF, 0.5 mmol) was added and the reaction mixture
refluxed for 24 h. No reaction, other than slight decomposition,
was observed by TLC.
Rf 0.57 (1:9 MeOH–CHCl3); 1H NMR (90 MHz) δ 2.12
(1H, dd, J 18.5, 5.8 Hz, one of ArCHCH2), 2.70–3.37 (3H, m,
ArCH2CH and one of ArCHCH2), 3.56 (3H, s, indole NCH3),
3.58–4.09 (3H, m, NCH2Ph and ArCH2CH), 4.29 (1H, d, J 5.8
Hz, ArCHCH2), 5.31 (1H, br s, one of C(O)NH2), 6.36 (1H, dd,
J 4.9, 2.2 Hz, ArCHCH2CH), 7.05–7.54 (10H, m, ArH and one
of C(O)NH2); 13C NMR (22.5 MHz) δ 22.6 (t), 29.4 (q), 31.1 (t),
47.4 (d), 51.7 (d), 56.5 (t), 105.3 (s), 108.8 (d), 118.4 (d), 119.1
(d), 121.3 (d), 127.0 (d), 127.2 (d), 128.5 (d), 128.8 (d), 129.6 (d),
134.6 (s), 136.5 (s), 137.1 (s), 138.6 (s), 169.0 (s); IR νmax/cmϪ1
(CHCl3) 3520, 3415, 1675, 1645, 1585, 1495, 1475, 1375; MS
m/z 357 (Mϩ, 100%), 340 (17), 313 (10), 273 (26), 266 (19), 249
(41), 221 (68), 206 (22), 181 (43), 91 (79); found m/z 357.1834,
calc. for C23H23N3O 357.1841.
Attempted epoxide formation from acetal 17: formation of the
N-oxides 20a/b
The acetal 17 (120 mg, 0.31 mmol) was dissolved in DCM
(15 ml) and stirred over excess NaHCO3 at 0 ЊC. Commercially
available MCPBA (201 mg of 50% pure solid, 0.59 mmol) was
added and stirring maintained for 30 min at this temperature.
Saturated NaHCO3 (aq) was added and the organic layer
separated, washed with NaHCO3 (aq) and then twice with
brine, then dried over MgSO4, filtered and evaporated to afford
a yellow oil (175 mg). Flash chromatography on silica eluted
with a careful gradient (1:20 ether–DCM to 1:1:18 ether–
MeOH–DCM) afforded white foams for both isomers of the
N-oxide 20a (71 mg, 57%) and 20b (24 mg, 19%).
Data for 20a (major isomer by ca. 3:1). Rf 0.52 (1:9
MeOH–CHCl3); 1H NMR (90 MHz) δ 2.09 (1H, dd, J 18.3, 4.7
Hz, one of ArCHCH2), 3.13–3.70 (2H, m, one of ArCHCH2
and one of ArCH2CH), 3.55 (3H, s, indole NCH3), 3.86–4.05
(5H, m, CH(OCH2)2 and one of ArCH2CH), 4.24–4.33 (2H,
m, ArCH2CH and ArCHCH2), 4.35–4.67 (2H, ABq, J 10.4 Hz,
NCH2Ph), 5.29 (1H, s, CH(OCH2)2), 5.89–6.00 (1H, m, ArCH-
CH2CH), 7.09–7.64 (9H, m, ArH); 13C NMR (75 MHz)
δ 25.9 (t), 28.9 (q), 29.2 (t), 60.4 (d), 64.5 (2 × t), 67.3 (d), 67.9 (t),
103.0 (s), 104.0 (d), 109.1 (d), 118.4 (d), 119.4 (d), 122.0 (d),
123.9 (d), 125.6 (s), 127.9 (d), 129.1 (d), 129.6 (s), 130.7 (s),
132.6 (d), 132.6 (s), 137.5 (s); IR νmax/cmϪ1 (CHCl3) 1615, 1495,
1470, 1455, 1400, 1385, 1350; MS m/z 402 (Mϩ, 9%), 386 (36),
313 (20), 311 (6), 295 (23), 273 (11), 251 (7), 242 (8), 222 (16),
207 (11), 194 (12), 181 (26), 157 (21), 91 (100), 73 (26); found
m/z 402.1950, calc. for C25H26N2O3 402.1943.
Attempted debenzylation of the ꢀ,ꢁ-unsaturated nitrile 10:
formation of the amine 24
The nitrile 10 (25 mg, 0.07 mmol) was dissolved in dry MeOH
and stirred over 10% Pd–C under an atmosphere of hydrogen
whilst being monitored by TLC. After 24 h, filtration and
evaporation afforded a clear oil (21 mg). Flash chromatography
on silica eluted with a solvent gradient (DCM to 1:19 MeOH–
DCM) afforded the free amine 24 as a white foam (16 mg,
80%).
Rf 0.41 (1:9 MeOH–CHCl3); 1H NMR (90 MHz) δ 1.61–2.10
(4H, m, ArCHCH2CH2), 2.99–3.17 (3H, m, ArCH2CH and
CHCN), 3.56 (3H, s, indole NCH3), 3.48–3.73 (1H, m, ArCH2-
CH), 4.14 (1H, br s, ArCHCH2), 7.05–7.58 (4H, m, ArH); 13
C
NMR (22.5 MHz) δ 18.6 (t), 21.0 (t), 29.2 (q), 29.3 (t), 34.2 (d),
49.3 (d), 50.6 (d), 106.8 (s), 109.0 (d), 118.2 (d), 119.3 (d), 121.3
(s), 121.5 (d), 126.2 (s), 133.2 (s), 137.1 (s); IR νmax/cmϪ1
(CHCl3) 2245, 1615, 1470, 1450, 1425, 1380, 1355; MS m/z
251 (Mϩ, 14%), 250 (46), 233 (17), 183 (100), 168 (24), 84 (28);
found m/z 251.1420, calc. for C16H17N3 251.1422.
1
Data for 20b. Rf 0.57 (1:9 MeOH–CHCl3); H NMR (90
MHz) δ 2.45 (1H, dd, J 19.9, 4.7 Hz, one of ArCHCH2), 2.97–
3.22 (3H, m, ArCH2CH and one of ArCHCH2), 3.52 (3H, s,
indole NCH3), 3.84–4.30 (6H, m, CH(OCH2)2 and NCH2Ph),
4.47–4.59 (2H, m, ArCH2CH and ArCHCH2), 5.28 (1H, s,
CH(OCH2)2), 5.90–6.02 (1H, m, ArCHCH2CH), 7.01–7.52
(7H, m, ArH), 7.77–7.89 (2H, m, ArH); 13C NMR (75 MHz)
δ 24.1 (t), 29.4 (q), 31.5 (t), 63.9 (d), 65.2 (t), 65.4 (t), 66.0 (d),
68.7 (t), 104.0 (d), 104.5 (s), 108.8 (d), 118.6 (d), 119.1 (d), 121.2
(d), 123.2 (d), 126.2 (s), 128.1 (d), 129.4 (d), 130.0 (s), 131.8 (s),
Preparation of (6S,8S,9R,10S)-9-cyano-8-hydroxy-5-methyl-
6,7,8,9,10,11-hexahydro-5H-6,10-iminocycloocta[b]indole 25
The alcohol 9 (50 mg, 0.14 mmol) was dissolved in MeOH
(10 ml) and stirred under an atmosphere of H2 with catalytic
10% Pd–C for 48 h. Filtration and evaporation afforded a
colourless oil (36 mg). Flash chromatography on silica eluted
J. Chem. Soc., Perkin Trans. 1, 2000, 3566–3577
3573

View Article Online
with 1:19 MeOH–DCM afforded the amino alcohol 25 as a
white foam (25 mg, 67%).
Si(CH3)2), 0.13 (3H, s, one of Si(CH3)2), 0.88 (9H, s, SiC-
(CH3)3), 0.85–1.08 (3H, m, NCO(CH2)2CH3), 1.44–1.86 (4H, m,
NCO(CH2)2CH3), 2.03–2.57 (2H, m, ArCHCH2), 2.69–3.20
(2H, m, ArCH2CH), 3.69 (3H, s, indole NCH3), {4.95–5.06,
5.19–5.32} (1H, m, ArCH2CH), {5.77, 6.10} (1H, d, J 5.0 Hz,
ArCHCH2), 7.09–7.53 (4H, m, ArH); 13C NMR (22.5 MHz)
δ Ϫ4.0 (q), Ϫ3.8 (q), {13.9, 16.1} (q), 18.0 (s), {18.5, 18.7} (t),
{25.3, 25.7} (q), 27.1 (t), {29.6, 29.7} (q), 35.0 (t), {35.2, 36.2}
(t), {41.6, 45.1} (d), {46.3, 49.7} (d), {92.6, 93.7} (s), {104.7,
106.8} (s), 109.1 (d), 116.7 (s), {118.4, 118.8} (d), 119.7 (d),
{122.0, 122.3} (d), 126.4 (s), {132.1, 133.0} (s), 137.2 (s), {161.5,
164.1} (s), {169.9, 170.4} (s); IR νmax/cmϪ1 (CHCl3) 2210, 1640,
1470, 1430, 1370; MS m/z 449 (Mϩ, 25%), 392 (26), 307 (15),
279 (5), 182 (27), 75 (100), 57 (9); found m/z 449.2491, calc. for
C26H35N3O2Si 449.2499.
Rf 0.35 (1:9 MeOH–CHCl3); 1H NMR (90 MHz) δ 1.83–1.98
(1H, m, one of ArCHCH2), 2.15–2.26 (1H, m, one of ArCH-
CH2), 3.20–3.49 (3H, m, ArCH2CH and CHCN), 3.62 (3H, s,
indole NCH3), 3.90–4.28 (3H, m, ArCH2CH, ArCHCH2 and
CHOH), 7.01–7.54 (4H, m, ArH); 13C NMR (22.5 MHz) δ 23.2
(t), 28.8 (q), 35.3 (t), 38.3 (d), 41.3 (d), 46.2 (d), 62.8 (d), 105.0
(s), 108.9 (d), 117.0 (d), 118.1 (d), 119.9 (d), 120.6 (s), 126.7 (s),
136.2 (s), 139.3 (s); IR νmax/cmϪ1 (CHCl3) 3520, 2231; MS
m/z 267 (Mϩ, 15%), 183 (100); found m/z 267.1368, calc. for
C16H17N3O 267.1372.
Preparation of (6S,10S)-9-cyano-5-methyl-6,7,10,11-tetrahydro-
5H-6,10-iminocycloocta[b]indole 22
The amino alcohol 25 (150 mg, 0.56 mmol), phosphoryl
chloride (343.4 mg, 2.24 mmol) and pyridine (885 mg, 11.2
mmol) were refluxed together in dry benzene (25 ml) for 48 h.
The solution was then cooled, treated with water and stirred
for 30 min. The organic layer was separated, twice washed
with saturated NaHCO3 (aq) then dried over MgSO4, filtered
and evaporated to afford a yellow oil (148 mg). Flash chroma-
tography on silica eluted with 1:19 ether–DCM afforded the
α,β-unsaturated amine 22 as a white foam (35 mg, 25%).
1
Data for the ester 28. Rf 0.56 (1:9 ether–CHCl3); H NMR
(90 MHz) δ 0.84–1.15 (6H, m, NCO(CH2)2CH3 and OCO-
(CH2)2CH3), 1.45–1.82 (4H, m, NCOCH2CH2CH3 and OCO-
CH2CH2CH3), 2.22–2.51 (5H, m, NCOCH2CH2CH3, OCO-
CH2CH2CH3 and one of ArCHCH2), 3.07–3.39 (3H, m,
ArCH2CH and one of ArCHCH2), 3.69 (3H, s, indole NCH3),
{5.06–5.15, 5.27–5.38} (1H, m, ArCH2CH), {5.88, 6.18} (1H,
d, J 5.9 Hz, ArCHCH2), 7.04–7.55 (4H, m, ArH); 13C NMR
(22.5 MHz) δ 13.5 (q), 13.9 (q), {18.2, 18.3} (t), 18.6 (t), {24.7,
26.7} (t), 29.6 (q), {33.1, 34.1} (t), 35.2 (t), 35.9 (t), {41.4, 45.1}
(d), {46.0, 49.9} (d), {101.8, 102.8} (s), {104.3, 106.8} (s), 109.2
(d), 114.0 (s), {118.4, 118.8} (d), 119.8 (d), {122.2, 122.4} (d),
126.2 (s), 132.8 (s), 137.3 (s), {158.5, 160.8} (s), 169.7 (s), 170.4
(s); IR νmax/cmϪ1 (CHCl3) 2230, 1765, 1715, 1650, 1470, 1420,
1380, 1350; MS m/z 405 (Mϩ, 100%), 335 (89), 264 (29), 247
(70), 222 (17), 183 (31), 170 (37), 144 (13), 71 (70); found
m/z 405.2046, calc. for C24H27N3O3 405.2052.
1
Rf 0.70 (1:9 MeOH–CHCl3); H NMR (300 MHz) δ 2.39
(1H, dd, J 19.1, 5.5 Hz, one of ArCHCH2), 2.97–3.06 (1H, m,
one of ArCHCH2), 3.08 (1H, d, J 15.9 Hz, one of ArCH2CH),
3.35 (1H, dd, J 15.9, 5.9 Hz, one of ArCH2CH), 3.70 (3H, s,
indole NCH3), 4.87 (1H, dd, J 11.4, 5.3 Hz, ArCH2CH), 5.36
(1H, dd, J 10.6, 5.5 Hz, ArCHCH2), 6.66 (1H, br d, J 5.3 Hz,
ArCHCH2CH), 7.14–7.53 (4H, m, ArH); 13C NMR (75 MHz)
δ 25.7 (t), 29.7 (q), 30.9 (t), 45.2 (d), 49.6 (d), 104.9 (s), 109.2 (d),
114.1 (s), 116.5 (s), 118.5 (d), 120.0 (d), 122.6 (d), 126.0 (s),
131.7 (s), 137.0 (s), 141.4 (d); IR νmax/cmϪ1 (CHCl3) 2205, 1635,
1470, 1430; MS m/z 249 (Mϩ, 100%), 234 (13), 183 (82), 168
(25), 79 (56); found m/z 249.1266, calc. for C16H15N3 249.1266.
Fluoride ion mediated desilylation of the TBDMS enol ether 27
To the amide 27 (130 mg, 0.29 mmol) stirred in MeOH (10 ml)
at room temp. was added TBAF (113 mg, 0.44 mmol). After 30
min, evaporation and flash chromatography on silica eluted
with 3:97 MeOH–DCM afforded the ketone 29 as a white foam
(100 mg, 100%). This material displayed very broad NMR
resonances and was not fully characterised.
Debenzylation of the TBDMS enol ether 14
The TBDMS enol ether 14 (250 mg, 0.53 mmol) was dissolved
in dry MeOH (30 ml) and then evaporated to dryness to
ensure that traces of halogenated solvent were removed. The
amorphous white solid obtained was redissolved in dry
MeOH (40 ml) and stirred over catalytic 10% Pd–C for 45 min
under an atmosphere of hydrogen. Filtration and evaporation
afforded the free amine 26 (202 mg, 100%) which was found to
be unstable and was therefore used immediately, without full
characterisation.
Rf 0.46 (1:9 MeOH–CHCl3); IR νmax/cmϪ1 (CHCl3) 2210,
1735, 1650, 1470, 1415, 1380; MS m/z 335 (Mϩ, 65%), 265 (28),
248 (11), 222 (10), 183 (100), 89 (19), 71 (96), 57 (32), 43 (83).
Sodium borohydride reduction of the ketone 29
The ketone 29 (120 mg, 0.36 mmol) was dissolved in dry MeOH
(10 ml) and NaBH4 (136 mg, 3.6 mmol) was added in portions.
After stirring at ambient temperature for 72 h the solvent was
removed by evaporation and the residue partitioned between
DCM and water. The organic layer was separated and the
aqueous layer was repeatedly extracted with DCM (until the
extract exhibited little or no UV chromophore: ca. 6–9 extracts
required). The combined organic extracts were dried over
MgSO4, filtered and evaporated to afford a pale yellow oil (100
mg). TLC indicated that at least 3 products were present and
our attempts to purify this material by flash chromatography on
silica eluted with a variety of systems met with little success.
This mixture was therefore used directly in the subsequent step.
Rf 0.13 (1:9 ether–CHCl3); 1H NMR (80 MHz) δ 0.11 (3H, s,
one of Si(CH3)2), 0.13 (3H, s, one of Si(CH3)2), 0.88 (9H, s,
SiC(CH3)3), 1.90–2.05 (1H, m, one of ArCHCH2), 2.65–2.80
(1H, m, one of ArCHCH2), 3.00–3.07 (2H, m, ArCH2CH),
3.64 (3H, s, indole NCH3), 4.18–4.29 (1H, m, ArCH2CH), 4.49
(1H, br d, J 5.7 Hz, ArCHCH2), 7.09–7.47 (4H, m, ArH);
IR νmax/cmϪ1 (CHCl3) 3320, 2210, 1730, 1630, 1470, 1415, 1365.
Formation of (6S,10S)-12-butyryl-9-cyano-8-tert-butyldimethyl-
silyloxy-5-methyl-6,7,10,11-tetrahydro-5H-6,10-iminocyclo-
octa[b]indole 27 and the ester 28
The free amine 26 (202 mg, 0.53 mmol) was stirred over Cs2CO3
(326 mg, 1 mmol) in DCM (20 ml). Butyryl chloride (85 mg,
83 µl, 0.8 mmol) was added dropwise and the reaction stirred
at ambient temperature for 30 min. Filtration and evaporation
then afforded a yellow oil (264 mg). Flash chromatography on
silica eluted with a solvent gradient (DCM to 1:19 MeOH–
DCM) afforded the amide 27 (145 mg, 61%) and the ester
28 both as white foams (28 mg, 13%).
Formation of (6S,10S)-12-butyryl-9-cyano-5-methyl-6,7,10,11-
tetrahydro-5H-6,10-iminocycloocta[b]indole 23
The mixture of products obtained above (100 mg), phosphoryl
chloride (184 mg, 1.2 mmol) and pyridine (475 mg, 6 mmol)
were refluxed together in dry benzene for 24 h. The solution was
then cooled, treated with water and stirred for 30 min. The
organic layer was separated, twice washed with saturated
NaHCO3 (aq) then dried over MgSO4, filtered and evaporated
Data for the amide 27 (minor rotamer in italics). Rf 0.77 (1:9
ether–CHCl3); 1H NMR (90 MHz) δ 0.10 (3H, s, one of
3574
J. Chem. Soc., Perkin Trans. 1, 2000, 3566–3577

View Article Online
to afford a yellow oil (115 mg). TLC indicated that this material
was composed of at least 3 products. However, careful flash
chromatography on silica eluted with a solvent gradient (1:19
ether–DCM to 1:1:18 MeOH–ether–DCM) afforded some
of the α,β-unsaturated product 23 as a white foam (27 mg,
24%) and some ketone 29 was recovered as a yellow foam
(10 mg, 8%). Other lower running components could not be
isolated cleanly.
91 (60); found m/z 357.1844, calc. for C23H23N3O 357.1841;
[α]D Ϫ102.9 (c 1.7, DCM).
Larger-scale LAH reductions of the TBDMS enol ether 14
The TBDMS enol ether 14 (250 mg, 0.53 mmol) was added
in portions to a suspension of LAH (30 mg, 0.79 mmol) in
dry THF (15 ml) at 0 ЊC. The reaction was then allowed to
warm to room temp. and stirred for 72 h. Work-up and
flash chromatography as for the previous reaction afforded the
vinylogous amide 31 as a pale yellow foam (38 mg, 20%) and a
higher running component 32 as a white foam (100 mg, 55%)
which could only be partially characterised before dimerising
to an unknown structure.
Rf 0.44 (1:9 ether–CHCl3); 1H NMR (90 MHz) δ 1.00 (3H, br
t, J 6.9 Hz, NCO(CH2)2CH3), 1.52–1.85 (2H, m, NCOCH2-
CH2CH3), 2.13–2.51 (3H, m, NCOCH2CH2CH3 and one of
ArCHCH2), 2.74–3.22 (3H, m, ArCH2CH and one of ArCH-
CH2), 3.70 (3H, s, indole NCH3), {4.89–4.99, 5.20–5.31} (1H,
m, ArCH2CH), {5.74, 6.10} (1H, d, J 5.7 Hz, ArCHCH2),
6.58–6.69 (1H, m, ArCHCH2CH), 7.09–7.54 (4H, m, ArH); 13
C
Data for 32. Rf 0.77 (1:9 ether–CHCl3); 1H NMR (80 MHz)
δ 2.44–3.40 (4H, m, ArCH2CH and ArCHCH2), 3.53 (3H, s,
indole NCH3), 3.86 (2H, br s, NCH2Ph), 4.14–4.30 (2H, m,
ArCHCH2 and ArCH2CH), 5.32 (1H, br s, one of CCH2), 6.09
(1H, br s, one of CCH2), 7.03–7.50 (9H, m, ArH); IR νmax/cmϪ1
(CHCl3) 1690, 1610, 1490, 1465.
NMR (22.5 MHz) δ 14.0 (q), 18.8 (t), 26.5 (t), 29.7 (q), 30.4 (t),
35.3 (t), 40.8 (d), {49.8, 50.6} (d), 104.5 (s), {109.2, 109.9} (d),
114.7 (s), 117.2 (s), 118.3 (d), 119.8 (d), 122.2 (d), 126.2 (s),
133.7 (s), 137.3 (s), 142.8 (d), 170.4 (s); IR νmax/cmϪ1 (CHCl3)
2225, 1650, 1470, 1430, 1420, 1385, 1350; MS m/z 319 (Mϩ,
100%), 279 (11), 249 (46), 248 (38), 232 (74), 221 (16), 206 (17),
183 (75), 167 (23), 157 (18), 149 (37), 71 (33), 57 (24), 43 (95);
found m/z 319.1669, calc. for C20H21N3O 319.1685.
Selected data for the dimer of 32. Rf 0.57 (1:9 ether–CHCl3);
1H NMR (300 MHz) δ 1.20 (1H, t, J 7.0 Hz), 1.58–1.64 (1H, m),
1.96–2.00 (1H, m), 2.10 (2H, br d, J 15.9 Hz), 2.24 (1H, dd,
J 12.9, 2.3 Hz), 2.56 (1H, d, J 16.4 Hz), 2.69 (1H, d, J 15.3 Hz),
2.78–2.86 (1H, m), 2.94 (1H, d, J 15.3 Hz), 3.01–3.08 (1H, m),
3.26 (1H, d, J 6.5 Hz), 3.40 (3H, s, one indole NCH3), 3.41 (3H,
s, one indole NCH3), 3.44–3.56 (3H, m), 3.68 (1H, d, J 13.5 Hz,
one CH2Ph), 3. 84 (1H, d, J 13.5 Hz, one CH2Ph), 4.08 (2H, d,
J 4.7 Hz), 6.90 (1H, dd, J 5.9, 2.9 Hz), 7.02–7.49 (18H, m,
ArH); 13C NMR (22.5 MHz) >50 resonances; IR νmax/cmϪ1
(CHCl3) 1720, 1600, 1490, 1465, 1415, 1375; FAB MS 685
([M ϩ H]ϩ, 14%), 663 (44), 647 (27), 341 (54), 338 (27), 281
(39), 207 (50), 154 (100); MS m/z 684 (Mϩ, 1%), 342 (100), 314
(12), 273 (48), 251 (21), 197 (24), 182 (31), 170 (58), 91 (70);
found m/z 684.3446, calc. for C46H44N4O2 684.3464.
For a second attempt, the TBDMS enol ether 14 (423 mg,
0.90 mmol) was dissolved in dry THF (15 ml) at 0 ЊC. LAH
(60 mg, 1.58 mmol) was added in portions over 5 h until TLC
of the reaction mixture indicated complete consumption of the
starting material. Work-up and flash chromatography as for the
previous reaction afforded the vinylogous amide 31 as a pale
yellow foam (85 mg, 26%) and the desilylated ketone 8 as a
white foam (216 mg, 67%).
General procedure for attempted ring closure of 23, 27 and 29
The amides 23, 27 and 29 (10 mg) were each dissolved in dry
THF (1 ml) and cooled to Ϫ78 ЊC. LDA (1.1 eq. of 1.5 M
solution in THF) was added and the reaction monitored by
TLC. When no reaction was observed after several hours at
Ϫ78 ЊC the reaction was allowed to warm to Ϫ20 ЊC and later
to ambient temperature. When TLC indicated the presence
of significant quantities of starting material the reaction was
cooled to Ϫ78 ЊC, and more LDA added dropwise. This process
was repeated until most of the starting material was consumed.
Work-up was achieved by cooling the reaction back to Ϫ78 ЊC,
quenching with saturated NH4Cl (aq), extracting into EtOAc,
washing with brine and drying over MgSO4. Filtration and
evaporation invariably gave yellow oils which were composed
of several products (by TLC). Investigation of the crude NMR
data indicated that little or no cyclisation had occurred, and the
isolation of a single product which could be fully characterised
proved impossible.
Small-scale LAH reduction of the TBDMS enol ether 14:
formation of the vinylogous amide 31
Attempted enamine reaction with the vinylogous amide 31
The TBDMS enol ether 14 (50 mg, 0.11 mmol) was added to
a suspension of LAH (7 mg, 0.18 mmol) in dry THF (4 ml) at
0 ЊC. The reaction was then allowed to warm to room temp. and
stirred for 24 h. Water (50 µl), aqueous 2 M NaOH (150 µl)
and water (50 µl) were added sequentially, then the mixture was
partitioned between EtOAc and water. The organic phase was
separated, dried over MgSO4, filtered and evaporated to afford
a yellow oil (40 mg). Flash chromatography on silica eluted
with a solvent gradient (DCM to 1:19 MeOH–DCM) afforded
the vinylogous amide 31 as a pale yellow foam (17 mg, 45%).
The vinylogous amide 31 (15 mg, 0.04 mmol), butyraldehyde
(28.8 mg, 0.40 mmol) and a trace of PTSA were stirred
over activated molecular sieves in CHCl3 (5 ml). After a 5 day
reaction time no reaction was visible by TLC. Under reflux or
with more acid present decomposition of the starting material
was the only process observed.
Attempted hydrogenation of the vinylogous amide 31: formation
of the ketone 33
1
Rf 0.51 (1:9 MeOH–CHCl3); H NMR (300 MHz) δ 2.39
The vinylogous amide 31 (33 mg, 0.09 mmol) was dissolved
in MeOH (10 ml) and stirred over an excess of 10% Pd–C
under an atmosphere of hydrogen. After 24 h the catalyst was
removed by filtration and the filtrate evaporated to dryness to
afford a mixture of the ketone 33 and inorganic contaminants.
Extraction of the residue into DCM followed by evaporation
afforded the ketone 33 as a white foam (16 mg, 66%).
(1H, d, J 17.0 Hz, one of ArCHCH2), 2.48 (1H, d, J 15.3 Hz,
one of ArCH2CH), 2.93 (1H, dd, J 17.0, 5.9 Hz, one of ArCH-
CH2), 3.34 (1H, dd, J 15.3, 5.3 Hz, one of ArCH2CH), 3.57
(3H, s, indole NCH3), 3.78–3.98 (2H, ABq, J 12.9 Hz,
NCH2Ph), 3.93 (1H, d, J 5.3 Hz, ArCH2CH), 4.11 (1H, d, J 5.9
Hz, ArCHCH2), 5.16 (1H, br s, one of CHNH2), 6.79–6.86
(1H, m, CHNH2), 7.06–7.47 (9H, m, ArH), 9.30 (1H, br s, one
of CHNH2); 13C NMR (22.5 MHz) δ 27.4 (t), 29.3 (q), 42.3 (t),
48.9 (d), 54.4 (d), 56.8 (t), 103.6 (s), 107.2 (s), 108.8 (d), 118.1
(d), 118.9 (d), 121.2 (d), 127.1 (s), 127.1 (d), 128.4 (d), 128.6 (d),
135.6 (s), 137.3 (s), 138.9 (s), 149.2 (d), 195.7 (s); IR νmax/cmϪ1
(CHCl3) 3500, 1650, 1585, 1470; MS m/z 357 (Mϩ, 100%), 340
(16), 314 (7), 273 (21), 266 (17), 250 (19), 187 (22), 170 (52),
1
Rf 0.46 (1:9 MeOH–CHCl3); H NMR (300 MHz) δ 1.04
(3H, d, J 7.0 Hz, CHCH3), 2.41 (1H, dd, J 14.1, 2.4 Hz, one of
ArCH2CH), 2.64 (3H, s, NbCH3), 2.72 (1H, d, J 16.8 Hz, one
of ArCHCH2), 2.92 (1H, q, J 7.0 Hz, CHCH3), 2.99 (1H, dd,
J 16.8, 5.9, one of Ar CHCH2), 3.09 (1H, dd, J 14.1, 4.7 Hz, one
of Ar CH2CH), 3.60 (3H, s, indole NCH3), 3.58–3.62 (1H,
m, ArCH2CH), 4.26 (1H, br s, ArCHCH2), 7.04–7.26 (3H, m,
J. Chem. Soc., Perkin Trans. 1, 2000, 3566–3577
3575

View Article Online
ArH), 7.44 (1H, d, J 8.2 Hz, ArH); 13C NMR (22.5 MHz)
δ 11.3 (q), 16.6 (t), 29.2 (q), {41.0, 41.3} (q), 46.0 (t), {47.4,
48.0} (d), {53.9, 54.9} (d), 61.2 (d), 103.2 (s), 108.8 (d), 118.3
(d), {119.0, 119.2} (d), {121.4, 121.8} (d), 126.6 (s), 133.9 (s),
137.4 (s), 208.8 (s); IR νmax/cmϪ1 (CHCl3) 1710, 1520, 1470,
1450, 1420, 1380; MS m/z 268 (Mϩ, 39%), 197 (100), 182 (10),
170 (24), 168 (10), 151 (13); found m/z 268.1578, calc. for
C17H20N2O 268.1576.
Data for the ketone 35. Rf 0.77 (1:9 Et2O–CHCl3); 1H NMR
(400 MHz) δ 1.44 (9H, s, (CH3)3C), 2.39 (1H, br d, J 13.7 Hz,
one of ArCHCH2), 2.55–2.64 (1H, br d, J 16.9 Hz, one of
ArCH2CH), 2.96–3.06 (2H, m, one of ArCHCH2 and CH-
CH2NHBOC), 3.12 (1H, dd, J 16.9, 6.3 Hz, one of ArCH2CH),
3.16–3.35 (2H, m, CH2NHBOC), 3.51 (3H, s, indole NCH3),
3.68–3.95 (3H, m, CH2Ph and ArCH2CH), 4.19–4.28 (1H, m,
ArCHCH2), 5.01 (1H, br s, NHBOC), 7.07–7.48 (9H, m, ArH);
13C NMR (50 MHz) δ 18.2 (t), 28.4 (q), 29.2 (q), 38.2 (t), 46.7
(t), 51.4 (d), 54.1 (d), 56.5 (t), 57.8 (d), 79.2 (s), 103.2 (s),
108.9 (d), 118.3 (d), 119.1 (d), 121.6 (d), 126.5 (s), 127.4 (d),
128.6 (d), 134.0 (s), 137.4 (s), 138.5 (s), 155.9 (s), 209.7 (s);
MS m/z 459 (Mϩ, 26%), 386 (9), 342 (12), 273 (94), 91 (86);
found m/z 459.2520, calc. for C28H33N3O3 459.2522.
LAH reduction and Boc protection of the nitrile alcohol 9:
formation of the Boc-protected amine 34
The alcohol 9 (1.88 g, 5.26 mmol) was dissolved in dry THF
(40 ml) and stirred at 0 ЊC. LAH (15 ml of 1 M solution, 15
mmol) was added dropwise and then the reaction was allowed
to warm to room temperature and stirred overnight, then
refluxed for 5 h. After cooling back to 0 ЊC, water (0.5 ml), 2 M
NaOH (0.5 ml) and water (0.5 ml) were added sequentially,
then the mixture was stirred and allowed to warm to room
temperature then filtered to remove inorganic waste and
evaporated. The organic residue was then suspended in EtOAc
and washed three times with brine, dried over MgSO4, filtered
and evaporated to afford a pale yellow oil, homogeneous by
TLC (1.8 g). This primary amine (1.8 g, 4.99 mmol) was dis-
solved in CHCl3 (50 ml) and stirred over NaHCO3 (1 g) at room
temperature. Boc2O (1.091 g, 5.0 mmol) dissolved in CHCl3
(25 ml) was added dropwise over 30 min and then the reaction
stirred for 1 h. Water (25 ml) was added and the mixture
vigorously stirred for 5 min, then the organic phase was
separated, washed with brine, dried over MgSO4, filtered and
evaporated to afford a yellow oil (2.39 g). Flash chroma-
tography on silica eluted with a solvent gradient (1:2 EtOAc–
hexane to 1:1 EtOAc–hexane) afforded the Boc-protected
amine 34 as a white foam (1.55 g, 64% from the amine 34).
Rf 0.25 (1:9 Et2O–CHCl3); 1H NMR (400 MHz) δ 1.52 (9H,
s, (CH3)3C), 2.13 (1H, d, J 13.9 Hz, one of ArCHCH2), 2.27
(1H, dt, J 13.9, 4.4 Hz, one of ArCHCH2), 2.43–2.53 (1H, m,
CHCH2NHBOC), 2.78 (1H, d, J 17.0 Hz, one of ArCH2CH),
3.13 (1H, dd, J 17.0, 7.1 Hz, one of ArCH2CH), 3.26–3.47 (3H,
m, ArCH2CH and CH2NHBOC), 3.63 (3H, s, indole NCH3),
3.65 (1H, d, J 13.5 Hz, one of CH2Ph), 3.77 (1H, d, J 13.5 Hz,
one of CH2Ph), 3.89–3.97 (2H, m, ArCHCH2 and CHOH),
4.85 (1H, br s, NHBOC), 7.14–7.43 (8H, m, ArH), 7.57 (1H, d,
J 7.8 Hz, ArH); 13C NMR (50 MHz) δ 17.3 (t), 28.4 (q), 29.2
(q), 37.8 (t), 40.4 (t), 45.0 (d), 47.6 (d), 52.8 (d), 57.3 (t), 65.8 (d),
79.2 (s), 103.7 (s), 109.1 (d), 118.3 (d), 119.4 (d), 121.5 (d), 126.7
(s), 127.0 (d), 128.3 (d), 128.5 (d), 137.2 (s), 137.3 (s), 139.0 (s),
156.2 (s); IR νmax/cmϪ1 (CHCl3) 3506, 3455, 1707, 1508, 1468,
1368; MS m/z 461 (Mϩ, 5%), 273 (10), 199 (10), 84 (80), 49
(100); found m/z 461.2684, calc. for C28H35N3O3 461.2678.
PCC oxidation of the alcohol 34
The alcohol 34 (1.33 g, 2.88 mmol) was dissolved in DCM
(2 ml) and added to a suspension of PCC–Florosil (50:50,
2.5 g, 6 mmol of PCC) in DCM (10 ml) and stirred at
room temperature overnight. The suspension was then filtered
through a pad of Florosil and evaporated to give a yellow oil
(1.19 g). This material was pre-sorbed onto silica (2 g) then
subjected to flash chromatography on silica eluted with 1:2
EtOAc–hexane to afford starting material alcohol 34 (174 mg)
and a 1:1.5 mixture of the Boc-protected ketone 35 and the
pentacycle 36 as a white foam (539 mg, 41 or 47% based on
recovered starting material).
1
Data for the pentacycle 36. Rf 0.77 (1:9 Et2O–CHCl3); H
NMR (400 MHz) δ 1.57 (9H, s, (CH3)3C), 2.57 (1H, dd, J 15.1,
2.6 Hz, one of ArCHCH2), 3.01–3.10 (2H, m, one of ArCH-
CH2 and CHCH2NHBOC), 3.58 (3H, s, indole NCH3), 3.80
(1H, dd, J 11.6, 1.8 Hz, one of CH2NBOC), 3.88 (1H, d, J 13.3
Hz, one of CH2Ph), 3.94–4.02 (2H, m, one of CH2Ph and one
of CH2NBOC), 4.08–4.14 (1H, m, ArCHCHN), 4.28–4.33
(1H, m, ArCHCH2), 5.51 (1H, d, J 6.4 Hz, ArCHCHN),
7.14–7.46 (8H, m, ArH), 8.16 (1H, d, J 8.0 Hz, ArH); 13C NMR
(50 MHz) δ 28.5 (q), 28.7 (q), 43.3 (t), 44.8 (t), 48.7 (d), 49.5 (d),
50.1 (d), 57.7 (t), 60.5 (d), 79.5 (s), 107.2 (s), 108.6 (d), 119.9 (d),
121.5 (d), 121.9 (d), 126.2 (s), 127.6 (d), 128.6 (d), 134.5 (s),
137.9 (2 × s), 155.0 (s), 206.7 (s); IR νmax/cmϪ1 (CHCl3) 1710,
1683, 1506, 1430, 1368; MS m/z 457 (Mϩ, 27%), 401 (12), 342
(15), 273 (16), 91 (37); found m/z 457.2359 calc. for C28H31N3O3
457.2365.
Formation of Nb-benzylsuaveoline 2c
The mixture of ketone 35 and pentacycle 36 (75 mg) was dis-
solved in 1:9 water–TFA and stirred at room temperature for
1 h. The solution was then evaporated to dryness, and then
three times redissolved in toluene and evaporated. The residue
was used directly in the next step and was dissolved in dry
DCM (10 ml) and stirred over molecular sieves at room
temperature while PrCHO (36 mg, 0.48 mmol) was added in
DCM (1 ml). After 72 h the solution was filtered through
Florosil to remove sieve dust, then evaporated to afford a
yellow oil composed of at least 5 products by TLC. Flash
chromatography on silica eluted with 5:95 MeOH–CHCl3
afforded 6 small analytical samples, one of which had an
appropriate Rf and displayed the EI mass ion and primary
fragmentation expected for Nb-benzylsuaveoline 2c: MS (EI,
70 eV) m/z 393 (Mϩ, 9%), 273 (30), 170 (54), 91 (100); lit.14a
MS (EI, 15 eV) m/z 393 (Mϩ, 100%), 273 (67).
Swern oxidation of the alcohol 34
Oxalyl chloride (70 mg, 48 µl, 0.55 mmol) was stirred in DCM
at Ϫ60 ЊC (CO2–CHCl3 bath). DMSO (86 mg, 78 µl, 1.1 mmol)
in DCM (1 ml) was added slowly and the mixture stirred for
2 min. The alcohol (46 mg, 0.1 mmol) was dissolved in DCM
(1 ml) and added dropwise and the reaction stirred for 30 min
at Ϫ60 ЊC. TEA (250 mg, 350 µl, 2.5 mmol) was then added
and the reaction allowed to warm to room temperature. Water
(10 ml) was added and the mixture stirred for 10 min, then the
organic layer separated and the aqueous phase extracted with
DCM. The combined organic extracts were washed with brine 5
times then dried over MgSO4, filtered and evaporated to afford
a pale yellow oil. Flash chromatography on silica eluted with
5:95 Et2O–CHCl3 gave the ketone 35 as a pale yellow foam
(17 mg, 37%), along with a complex mixture (33 mg, homo-
geneous by TLC) which was composed of at least 3 non-indolic
products (>50 resonances in 13C NMR, but the characteristic
quaternary resonance normally observed around 103 ppm
was not present).
Acknowledgements
We thank Dr G. K. Barlow, Mrs B. Chamberlain, Dr
T. A. Dransfield, Mr B. R. Glennie and Dr N. R. McLay
at York and Dr A. S. F. Boyd and Dr R. Fergusson at Heriot-
Watt for NMR and mass spectra, and EPSRC and Sanofi-
Synthelabo for funding.
3576
J. Chem. Soc., Perkin Trans. 1, 2000, 3566–3577

View Article Online
12 P. D. Bailey, I. D. Collier, S. P. Hollinshead, M. H. Moore,
References
K. M. Morgan, D. I. Smith and J. M. Vernon, J. Chem. Soc., Perkin
Trans. 1, 1997, 1209.
13 S. P. Hollinshead, DPhil Thesis, York, 1987.
1 S. F. Martin, C. W. Clark and J. W. Corbett, J. Org. Chem., 1995, 60,
3236.
2 A. Jossang, P. Fodor and B. Bodo, J. Org. Chem., 1998, 63, 7162 and
14 (a) X. Fu and J. M. Cook, J. Org. Chem., 1993, 58, 661;
(b) D. Soerens, PhD Thesis, University of Wisconsin–Milwaukee,
1978; (c) R. W. Weber, PhD Thesis, University of Wisconsin–
Milwaukee, 1985.
15 Y. Horiguchi, S. Matsuzawa, E. Nakamura and I. Kuwajima,
Tetrahedron Lett., 1986, 27, 4025.
16 P. D. Bailey, S. P. Hollinshead, M. H. Moore, K. M. Morgan,
D. I. Smith and J. M. Vernon, Tetrahedron Lett., 1994, 35, 3585.
17 (a) B. H. Han and P. Boudjouk, J. Org. Chem., 1982, 47, 5030;
(b) A. K. Bose, K. Gupta and M. S. Manhas, J. Chem. Soc., Chem.
Commun., 1984, 86; (c) T. Kitazume, Synthesis, 1985, 855.
18 (a) E. Erdik, Tetrahedron, 1987, 43, 2203; (b) C. Rene, A. Furstner
and H. Weidmann, J. Chem. Soc., Chem. Commun., 1986, 775.
19 W. Zwanenburg, Tetrahedron Lett., 1970, 935.
references therein.
3 J. Li and J. M. Cook, J. Org. Chem., 1998, 63, 4166.
4 S. P. Majumbar, P. Potier and J. Poissen, Tetrahedron Lett., 1972,
1563.
5 (a) M. A. Amer and W. E. Court, Phytochemistry, 1981, 20, 2569;
(b) A. M. A. G. Nasser and W. E. Court, Phytochemistry, 1983,
22, 2297.
6 (a) S. Enders, H. Takayama, S. Suda, M. Kitajima, N. Aimi, S. Sakai
and J. Stockigt, Phytochemistry, 1993, 32, 725; (b) L. Polz,
J. Stockigt, H. Takayama, N. Uchida, N. Aimi and S. Sakai,
Tetrahedron Lett., 1990, 31, 6693.
7 M. L. Trudell and J. M. Cook, J. Am. Chem. Soc., 1989, 111,
7504.
8 P. Yu and J. M. Cook, J. Org. Chem., 1998, 63, 9160.
9 X. Fu and J. M. Cook, J. Am. Chem. Soc., 1992, 114, 6910.
10 P. D. Bailey and N. R. McLay, J. Chem. Soc., Perkin Trans. 1, 1993,
441.
11 P. D. Bailey, S. P. Hollinshead, N. R. McLay, K. M. Morgan,
S. J. Palmer, S. N. Prince, C. D. Reynolds and S. D. Wood, J. Chem.
Soc., Perkin Trans. 1, 1993, 431.
20 E. N. Zil’berman, Russ. Chem. Rev., 1984, 53, 900.
21 R. Grigg, T. R. B. Mitchell and S. Sutthivaiyakit, Tetrahedron
Lett., 1979, 1067.
22 P. D. Bailey, P. J. Cochrane, F. Irvine, K. M. Morgan, D. J. Pearson
and K. T. Veal, Tetrahedron Lett., 1999, 40, 4593.
23 P. D. Bailey, K. M. Morgan, G. M. Rosair and R. Ll. Thomas,
Tetrahedron Lett., 1999, 40, 8255.
J. Chem. Soc., Perkin Trans. 1, 2000, 3566–3577
3577