A short and concise asymmetric synthesis of hamigeran B

Article abstract of DOI:10.1002/chem.200400558

The interesting biological properties of the hamigerans wherein hamigeran B is a potent antiviral agent with low cytotoxicity to host cells make these deceptively simple looking structures challenging synthetic targets. A strategy to hamigeran B evolved wherein the three contiguous stereocenters are established ultimately from a Pd catalyzed asymmetric allylic alkylation (AAA). The latter involves an asymmetric allylation of a non-stabilized ketone enolate in 77% yield and 93% ee. By using this process, (S)-5-allyl-2-isopropyl-5- methyl-1-trifluoromethanesulfonyloxycyclopentene becomes available in four steps from 2-methylcyclopentanone. Introduction of the aryl unit by cross-coupling proceeded intermolecularly but failed intramolecularly. On the other hand, reductive removal of the triflate permitted a Heck reaction to effect intramolecular introduction of the aryl ring. The unusual conformational properties of this molecular architecture are revealed by the regioselectivity of the β-hydrogen elimination in the Heck reaction and the diastereoselectivity of the reduction establishing the stereochemistry of the carbon bearing the isopropyl group. The successful route consists of 15 steps from 2-methylcyclopentanone and dimethylorcinol illustrating the efficiency of the route based upon the Pd AAA.

Full text of DOI:10.1002/chem.200400558

FULL PAPER
A Short and Concise Asymmetric Synthesis of Hamigeran B
Barry M. Trost,* Carole Pissot-Soldermann, and Irwin Chen[a]
Abstract: The interesting biological
properties of the hamigerans wherein
hamigeran B is a potent antiviral agent
with low cytotoxicity to host cells make
these deceptively simple looking struc-
tures challenging synthetic targets. A
strategy to hamigeran B evolved
wherein the three contiguous stereo-
centers are established ultimately from
a Pd catalyzed asymmetric allylic alky-
lation (AAA). The latter involves an
asymmetric allylation of a non-stabi-
lized ketone enolate in 77% yield and
93% ee. By using this process, (S)-5-
allyl-2-isopropyl-5-methyl-1-trifluoro-
methanesulfonyloxycyclopentene be-
comes available in four steps from 2-
methylcyclopentanone. Introduction of
the aryl unit by cross-coupling proceed-
ed intermolecularly but failed intramo-
lecularly. On the other hand, reductive
removal of the triflate permitted a
Heck reaction to effect intramolecular
introduction of the aryl ring. The un-
usual conformational properties of this
molecular architecture are revealed by
the regioselectivity of the b-hydrogen
elimination in the Heck reaction and
the diastereoselectivity of the reduction
establishing the stereochemistry of the
carbon bearing the isopropyl group.
The successful route consists of 15
steps from 2-methylcyclopentanone
and dimethylorcinol illustrating the ef-
ficiency of the route based upon the Pd
AAA.
Keywords: alkylation · asymmetric
catalysis
·
Heck reaction
·
palladium · total synthesis
Introduction
The hamigerans are a novel
bioactive family of metabolites
isolated from the poecilosclerid
sponge Hamigera tarangaensis
Berquist and Fromont (family
Anginoidae, syn. Phorbasidae)
from the Hen and Chicken Is-
lands off the coast of New Zea-
Figure 1. Structure of hamigerans.
land.[1] These natural products
contain a unique carbon skele-
ton in which a substituted aromatic nucleus is fused onto a
[4.3.0] or [5.3.0] carbocyclic core that contains three or four
contiguous stereogenic centers one of which features an iso-
propyl group (Figure 1). The biological evaluation of these
compounds ranges from moderate in vitro antitumor activity
against P-388 leukemia cells (hamigeran D) to pronounced
antiviral activity. Indeed, hamigeran B showed 100% virus
inhibition against both the herpes and polio viruses, with
only slight cytotoxicity throughout the host cells.
Despite the relatively small size, the hamigerans offer
both a challenge and an opportunity to apply new synthetic
tools developed in our laboratories. Moreover, the structural
features of hamigerans and the recent development of the
palladium catalyzed asymmetric allylic alkylation (AAA) of
ketone enolates[2] prompted us to devise a straightforward
route to this family. We chose hamigeran B as our target
due to its pronounced antiviral activity. In this paper, we de-
lineate our study culminating in a concise convergent syn-
thesis of hamigeran B.
[a] Prof. B. M. Trost, C. Pissot-Soldermann, I. Chen
Department of Chemistry
Stanford University
Stanford, CA 94305-5080 (USA)
Fax : (+1)650-725-0002
E-mail: bmtrost@stanford.edu
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2005, 11, 951 – 959
DOI: 10.1002/chem.200400558
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
951

Results and Discussion
line[9]) and vinyl triflate 8 in 94% yield in spite of the ex-
treme steric congestion. Carboxylic acid 3 was obtained in
two steps. First, selective oxidation of the terminal olefin in
the presence of the sterically congested tetrasubstituted
olefin by the dihydroxylation/periodate cleavage sequence
afforded aldehyde 13 in 86% yield. Jones oxidation trans-
formed the intermediate aldehyde to the carboxylic acid 3
in 95% yield (Scheme 2).
Under the acidic conditions of the Jones oxidation we
also observed the formation of a by-product in 5% yield.
The NMR analysis of this side product allowed us to assign
its structure as a product of a Friedel–Crafts acylation reac-
tion. Unfortunately, NOE experiments on the compound
provided evidence that the regioselectivity of the acylation
was not ortho as we anticipated, but actually para to the meth-
oxy directing group. Indeed, the aromatic methoxy group
showed NOE correlations to two aromatic protons at d 6.66
and 6.80, the aromatic methyl group showed a NOE correla-
tion to the aromatic proton at d 6.66 and a methyl group of
by the isopropyl moiety showed correlation to the aromatic
proton at d 6.80 (Figure 2).
Initial forays: Our interest stemmed from the question of
whether setting the stereochemistry of the quaternary center
from which all the remaining stereocenters may evolve
might be achievable by a Pd-catalyzed asymmetric allylic al-
kylation (AAA)[3–6] of a ketone enolate. The methodological
development described previously[7,8] sets a starting point
for our retrosynthetic blueprint depicted in Scheme 1. The
route envisioned was to form the central six-membered ring
at a late stage of the synthesis via an intramolecular Frie-
del–Crafts acylation of acid 3 (see Supporting Information).
The chiral cyclopentanone would stem from the ketone eno-
late derived from 5.
This result led us to assume that the preferred regioselec-
tivity of the Friedel–Crafts acylation is in the para position
to the alkoxy directing group. In order to confirm this as-
sumption, we decided to perform the Friedel–Crafts acyla-
tion on the acyl chloride in the presence of a Lewis acid pro-
moter, assuming that the Lewis acid would simultaneously
coordinate the aromatic alkoxy substituent and the acid
chloride in order to bring the ortho position in close proxim-
ity to the electrophilic center. We treated the in situ formed
acyl chloride of 3 with 3.5 equivalents of aluminium chloride
in benzene. However, isolation and characterization of the
Scheme 1. Synthetic strategy envision for hamigeran B.
Racemic 5 was readily pre-
pared from commercially avail-
able
via a one-pot tandem formyla-
tion/vinylogous etherification
2-methylcyclopentanone
sequence. The palladium-cata-
lyzed AAA of 5 with allyl ace-
tate afforded the allylated prod-
uct 4 in 77% yield and 93% ee,
by using 0.5 mol% of p-allyl-
palladium chloride dimer and
1 mol% (R,R) “standard” Trost
ligand (see Scheme 1).[3b]
Reaction of 4 with two equiv-
alents of lithium dimethylcup-
rate afforded the ketone 7 in
89% yield as a 1:1 mixture of
diastereomers. The vinyl triflate
Scheme 2. Synthesis of carboxylic acid 3. a) NaOMe, HCO2Et, PhH, 08C to rt, then pTsOH, tBuOH, reflux,
61%. b) i) LDA, DME, À788C to 08C. ii) tBuOH, Me3SnCl, DME, 08C. iii) 0.5 mol% [(h3-C3H5)PdCl]2,
1 mol% (R,R)-ligand, allyl acetate, DME, À788C to rt, 77%. c) LiCuMe2, Et2O, À208C to rt, 89%. d) LDA,
PhNTf2, THF, 87%. e) 2 equiv arylboronic acid 9, 2.5 mol% [Pd(PPh3)4], KBr, K3PO4, dioxane, 858C, 94%.
f) 1.3 mol% OsO4, NMO, THF/H2O, 0 8C to rt, then NaIO4, THF/H2O, 86%. g) Jones reagent, acetone, 75%.
8 was then obtained in 87% yield by treatment of ketone 7
with LDA and N-phenyltrifluoromethanesulfonimide. We
crystalline compound formed indicated that under those re-
action conditions, we had obtained the tetracyclic product
12. This product presumably arose from an initial intramo-
lecular Friedel–Crafts acylation and, in the presence of HCl
generated in the course of the reaction, a subsequent Frie-
À
were able to form the C4a C5 bond via a Suzuki cross-cou-
pling reaction between aryl boronic 9 (synthesized in three
steps from commercially available 2-methoxy-6-methylani-
952
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 951 – 959

Hamigeran B
FULL PAPER
Unfortunately, no reaction was observed when phenol 15
was treated with BCl3 and AlCl3 in CH2Cl2. As rare earth
metal trifluoromethanesulfonates were found to be efficient
catalysts for Friedel–Crafts acylation and alkylation,[13] we
also treated phenol 15 with
a catalytic amount of
[Yb(OTf)3], prepared from the corresponding oxide Yb2O3
and trifluoromethanesulfonic acid.[14] However, no reaction
was observed and the starting material could be recovered.
Unfortunately, this route was thwarted by the preferred
regioselectivity para to the alkoxy directing group. Thus, we
decided to revise our strategy for closing the central six-
Figure 2. NOE correlations for 11.
del–Crafts alkylation took place. Under protic conditions,
formation of the tertiary non-benzylic cation is preferred
due to carbonyl group destabilizing the benzylic cation. 1,2-
Hydride shift (or elimination to exocyclic olefin followed
by protonation) produced the tertiary isopropyl cation
which then underwent electrophilic aromatic substitution
(Scheme 3).
À
membered ring considering the construction of C11a C11
À
bond first and then establish the C4a C5 bond via an intra-
molecular palladium-catalyzed, distannane-mediated reduc-
tive coupling of an aryl triflate with
(Scheme 5).
a vinyl triflate
Our new strategy began with intermediate 8 and proceed-
ed along the lines depicted in Scheme 6. The aldehyde 17
derived from the vinyl triflate 8 by an oxidation of the ter-
minal olefin in the presence of the more electron-deficient
and sterically congested tetrasubstituted olefin. Dihydroxy-
lation/periodate cleavage sequence afforded 17 in 94%
yield. Direct reaction with lithiated dimethoxymethylorcinol
18 followed by oxidation with Jones reagent gave the full
carbon skeleton of the target. Under the acidic conditions of
the Jones oxidation, a small percentage of the ketone was
also monodeprotected to afford the phenol 20 (8%). No bis-
deprotected product was ever observed. Deprotection of
one of the two equivalent methoxymethyl protecting groups
to afford 20 was accomplished in 56% yield from 19 by stir-
ring in the presence of 9m H2SO4. To complete the synthesis
Scheme 3. Proposed mechanism for the formation of 12.
Following these results, we
decided to perform the intra-
molecular Friedel–Crafts acyla-
tion on the free phenol, assum-
ing that it should be a better
ortho-directing group, with a ni-
trile moiety as the electrophilic
partner. Indeed, there is one
precedent in the literature of an
exclusive ortho a-chloroacetyla-
tion of phenols using a combi-
nation of boron trichloride
(BCl3) and aluminium trichlor-
ide (AlCl3).[10] We treated alde-
hyde 13 with hydroxylamine to
give the corresponding oxime[11]
as a mixture of isomers, which
Scheme 4. Formation of nitrile 15. a) NH2OH·HCl, CH3CO2Na, CH3CN/H2O, 1 h, rt, 98%. b) [RuCl2(p-
cymene)]2, MS 4 ꢁ, CH3CN, 808C, 15 min, 76%. c) BBr3, CH2Cl2, À788C to 08C, 85%.
was dehydrated under neutral Scheme 5. Second retrosynthetic analysis.
and mild conditions using a cat-
alyst system of [RuCl2(p-
cymene)]2/molecular sieves[12] to give nitrile 14 in 75% yield
from aldehyde 13. The phenol was then unmasked by treat-
ment with boron tribromide to give 15 in 85% yield
(Scheme 4).
of the cyclization precursor, phenol 20 was treated with trifl-
ic anhydride in the presence of pyridine to afford 16 in ex-
cellent yield (95%). The stage is now set for the final car-
À
bon carbon bond forming reaction to close the central six-
Chem. Eur. J. 2005, 11, 951 – 959
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
953

B. Trost et al.
membered ring and complete the carbon skeleton of hami-
geran B.
The transformation we propose is an intramolecular palla-
dium-catalyzed, distannane-mediated reductive coupling of
an aryl triflate with a vinyl triflate (Figure 3).
halide/aryl halide,[15] aryl triflate/aryl triflate,[15] aryl halide/
vinyl triflate,[16] and vinyl triflate/vinyl halide[17] as well as
allyl ester/allyl ester[18] coupling partners, so the aryl triflate/
vinyl triflate reductive coupling is a reasonable proposition,
even though no specific intramolecular or intermolecular
precedent exists. Our proposed coupling reaction will be es-
pecially challenging due to the extreme steric congestion
surrounding the vinyl triflate.
Several parameters were examined, including solvent, dis-
tannane, additive, palladium source, and exogenous ligands.
Table 1 shows that no conditions were found that yielded
any trace of 2. As a starting point for investigation, the reac-
tion conditions reported for the synthesis of arylstannanes
from aryl halides and distannanes were chosen.[19] However,
catalytically inactive palladium black crashed out of the re-
action mixture quickly upon heating, and only 16 and
phenol 20 were recovered from the reaction mixture, the
latter presumably formed by hydrolysis of the aryl triflate
by adventitious water (entry 1). To stabilize the palladium
species in solution, we switched
Scheme 6. Synthesis of substrate 16. a) cat. OsO4, NMO, THF/H2O then
NaIO4. b) Et2O, À78 to 5, 94%. c) Jones reagent, acetone, 78%. d) aq.
H2SO4, THF, 56%. e) (CF3SO2)2O, pyridine/CH2Cl2, 95%.
to more donating solvents such
as THF and DMF and added
exogenous
triphenylarsine
ligand. Triphenylarsine was
chosen as the initial ligand be-
cause of its known ability to aid
intermolecular Stille couplings
of arylstannanes with vinyl tri-
flates.[20] Indeed, doing both,
precluded precipitation of pal-
ladium black and consequently
led to higher conversions of the
starting material. While none of
the desired cyclized product
was observed, using THF as the
solvent led to the promising iso-
lation of the intermediate aryl-
stannane 22 in 41% yield, as
well as the reduced product 23
in 11% yield (entry 2). This
suggests that the subsequent
coupling reaction with the vinyl
triflate is very slow, as we had
expected from the high degree
of steric congestion.
Therefore, we tried two ideas
Figure 3. Aryl triflate/vinyl triflate coupling.
to facilitate the reaction despite
the steric congestion around the
An intermolecular Stille reaction between the aryl triflate
and the distannane afforded an arylstannnane intermediate
that could immediately cyclize under the reaction conditions
with the vinyl triflate via an intramolecular Stille reaction. It
is also reasonable to propose a vinylstannane intermediate
that cyclizes with the aryl triflate, but the vinyl triflate is far
more sterically congested, as well as less electronically acti-
vated for oxidative addition than the aryl triflate. Such intra-
molecular reductive Stille couplings are known for aryl
vinyl triflate. First, we added stoichiometric amounts of lithi-
um chloride to the reaction mixture because of its known
ability to accelerate the reactivity of triflates in cross-cou-
pling reactions by facilitating transmetalation. Next, we
switched from bis(tributyltin) to the less sterically demand-
ing hexamethylditin. However, switching to hexamethylditin
actually gave worse results (entry 2 vs 3, entry 5 vs 6). Em-
ploying a huge excess of lithium chloride and hexamethyldi-
tin in THF also gave inferior results (entry 4), although it
954
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 951 – 959

Hamigeran B
FULL PAPER
Table 1. Attempts to cyclize: selected conditions.[a]
cessible aryl triflate set the
stage for the carbopalladation
of
a trisubstituted olefin to
form tricycle 24, after b-hydro-
gen elimination. The Heck
strategy involving attack on the
p-system of the double bond re-
duces the steric congestion rela-
tive to the oxidative addition to
the vinyl triflate. Further re-
moval of the vinyl triflate re-
duces the steric congestion for
the carbopalladation. Further-
more, the formation of a six-
membered ring via a 6-exo-trig
mode is less sterically demand-
ing and therefore favored. Fi-
nally, the Heck reaction should
Entry
Solvent
R
LiCl
AsPh3
none
10 mol%
10 mol%
none
Yield
[b]
1
2
3
PhCH3
THF[b]
THF[b]
THF[b]
DMF[c]
DMF[c]
Bu
Bu
Me
Me
Bu
Me
none
none
none
6.0 equiv
3.0 equiv
3.6 equiv
13% 20 and recovered 16
41% 22, 11% 23, 37% 16
no reaction
4[d]
5
mostly 16
no 16, 29% 21
40 mol%
40 mol%
6
complex mixture, including 16, but no 21
[a] All reaction were performed with [Pd(PPh3)4] as the Pd source. [b] Heated to reflux for about 12 h.
[c] Heated to 708C for about 12 h. [d] Coreyꢂs conditions for cyclization of a vinyl triflate with a vinyl bromide.
See ref [17].
worked successfully for the Corey group to cyclize a vinyl
bromide with a vinyl triflate. We chose to revert back to
bis(tributyltin) and to continue employing three equivalents
of lithium chloride as an additive in addition to triphenylar-
sine as the exogenous ligand. Choosing DMF as the new sol-
vent gave the most promising result (entry 5). Complete
consumption of the starting material was witnessed, and a
product 21 was isolated in 29% yield that had 1H NMR
spectral attributes that were suggestive of a tricyclic struc-
ture. Coupled with low-resolution mass spectroscopy data
(GC-MS), we have tentatively assigned the structure of 21
to be that depicted above.
A potential mechanism to arrive at 21 from 16 is depicted
in Figure 4. Following oxidative addition of palladium(0),
carbopalladation of the tetrasubstituted olefin forms the tri-
cycle. Expulsion of the triflate leaving group yielded the de-
sired tricycle 2 and palladium triflate. Triflate transfer from
the palladium(ii) species to the tricycle concomitantly re-
duced the palladium(ii) species to a palladium(0) species to
regenerate the catalytic cycle. Further investigations to opti-
mize the formation of 21 and delineate the reaction mecha-
Figure 4. Proposed mechanism for the formation of 21.
nism were severely thwarted by the capriciousness and irre-
producibility of the reaction.
Other parameters such as palladium source and ligand
were also investigated but didnꢂt give any successful results.
Therefore, we decided to reconsider the construction of the
lead to a trisubstituted olefin after syn elimination of the
palladium species which should be easier to hydrogenate in
the following step (Scheme 7).
À
C4a C5 bond via an intramolecular palladium catalyzed
Our new strategy began with intermediate 17 which react-
ed with lithiated dimethylorcinol 26. The change in the pro-
Heck reaction of an aryl triflate with an olefin. This idea
was supported by the isolation
of the by-product 21 in the pre-
vious study.
Heck strategy: The formation
of 21 suggested that a Heck-
type strategy might be viable.
Oxidative addition of palladi-
um(0) to the sterically more ac- Scheme 7. Final retrosynthetic analysis.
Chem. Eur. J. 2005, 11, 951 – 959
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
955

B. Trost et al.
tecting group was only directed towards a higher yielding
synthesis. Therefore, DME was employed as solvent at low
temperature (À558C) to guarantee the reproducibility of the
reaction. The resultant benzylic alcohol was then immedi-
ately oxidized with Dess–Martin periodinane in buffered
medium to avoid elimination of the activated benzylic alco-
hol. With the vision that the
thaw cycles on the solvent to remove any trace of oxygen.
Furthermore, the use of carbonate rather than tertiary
amine bases minimizes the problem of simple hydrogenoly-
sis of the triflate (entry 1 vs 2, Table 2). With either dppe or
dppp as ligands,[22] the desired alkene 24 was isolated in 42–
48% yield which increased to 58% with dppb.[23]
À
last C C bond would be
Table 2. Intramolecular Heck reaction of aryl triflate 25.[a]
formed by an intramolecular
Heck reaction,[21] the vinyl tri-
flate was reductively cleaved to
alkene 29 and the aryl ether
converted to the requisite aryl
triflate 25. (Scheme 8)
Entry
Ligand
Base
T [8C]
Conversion [%][b]
32 [%]
24 [%]
30 [%]
31 [%]
1
2
3
4
5
6
7
dppp
dppp
dppp
dppe
dppb
K2CO3
PMP
107
107
107
107
107
107
107
100
88
100
76
100
86
83
5
29
2
6
5
48
31
37
42
58
8
29
14
18
19
14
38
39
4
5
CsCO3
K2CO3
K2CO3
K2CO3
K2CO3
21
10
15
26
6
We examined the intramolec-
ular Heck reaction of the sub-
stituted cyclopentene 25. Sur-
prisingly, it produced two iso-
meric alkenes 30 and 31 in ad-
dition to the expected alkene
Bdbpp
Bdbpb
8
8
25
[a] All reactions performed with 10 mol% [Pd(OAc)2], 20 mol% ligand, 3 equiv base, with a concentration of
c=0.1m in toluene and cycles of freeze, pump, thaw. [b] Determinated by GC of the crude. [c] Bdbpp=1,3-
bis(dibenzophospholyl)propane, Bdbpb=1,3-Bis(dibenzophospholyl)butane.
Hydrogenation from the least hindered convex face to
give the desired stereochemistry of C-6 seemed to be
straightforward. To avoid reduction of the carbonyl group,
the free phenol was liberated with BBr3 (CH2Cl2, À788C).
Upon hydrogenation over Pd/C, a single diastereomer did
result. X-ray crystallography, however, revealed the product
33 to have exclusively the C-6 epi configuration (Scheme 9,
path a).
Hypothesizing that this product must arise by an equili-
bration in the semihydrogenation intermediate because the
Scheme 8. Synthesis of substrate 25 for Heck reaction. a) cat. OsO4,
NMO, THF/H2O, then add NaIO4. b) i) DME, À558C. ii) Dess–Martin
periodinane, NaHCO3, CH2Cl2, rt, 75% from 8. c) BCl3, CH2Cl2, À208C,
86%. d) 10 mol% Pd(OAc)2, 20 mol% dppf, HCO2H, (C2H5)3N, DMF,
708C, 1 h, 94%. e) (CF3SO2)2O, pyridine/CH2Cl2, 08C, 94%.
final reductive elimination step is too slow as summarized in
Scheme 10, attention turned to iridium since it is known to
minimize such equilibrations.[24,25] Gratifyingly, hydrogena-
tion once again proceeded with complete diastereoselectivi-
ty (Scheme 9, path b). X-ray crystallography confirmed the
correct stereochemistry as in structure 34 for all centers.
It is noteworthy to mention that no hydrogenation condi-
tions were found to make the tetrasubstituted double bond
30a react. Nevertheless, this product could be separated
from 24a by flash chromatogra-
24 [Eq. (1)], surprisingly since it involves a clearly highly
strained tetrasubstituted double bond exocyclic to the ring.
The assessment of the relative stereochemistry of the prod-
phy. However, we focused our
attention on the isomerization
of the quaternary double bond
under acidic conditions. By
using p-toluenesulfonic acid as
catalyst in refluxing toluene led
to the thermodynamically more
stable conjugated double bond
ucts syn to the C-9 methyl group was proved by NOE ex-
35 which proved to be unreactive as well under all hydroge-
nation conditions [Eq. (2)]. By using Amberlyst 15 as a cata-
lyst provided a 1:1 mixture of products arising from the non
regioselective protonation of the double bond [Eq. (3)].
When the isopropylic carbocation is formed, an intramolec-
ular Friedel–Crafts alkylation occurred to irreversibly give
the tetracyclic product 36, as previously encountered
(Scheme 3).
À
periment. Therefore, the stereodiscrimination during the C
C bond formation was fully controlled by the C-9 quaterna-
ry center, installed by the asymmetric allylic alkylation. Re-
action conditions were surveyed to maximize the formation
of 24 and guarantee the reproducibility of the reaction
which turned out to be somewhat capricious. The latter was
rapidly addressed by performing successive freeze-pump-
956
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 951 – 959

Hamigeran B
FULL PAPER
during the Heck reaction and decided to attempt to mini-
mize its formation.
At this stage, only functional group manipulation is re-
quired to arrive at the targeted structure. We first needed to
oxidize the tetralone to the corresponding 1,2-diketone
(Scheme 11). For this purpose, we applied a methodology
developed in our group some years ago.[26] Treatment of the
enolate of compound 34 with phenyl benzenethiosulfonate
afforded the corresponding b-ketosulfides 37 which could be
acetoxylated by using lead tetraacetate in warm benzene.
The b-keto acetoxy sulfide 38 represents a monoprotected
form of the 1,2-dicarbonyl compound 39 which could be un-
masked by saponification of the acetate group under basic
conditions. However, the yield of the overall transformation
proved to be disappointingly low. The free phenol seemed
to be problematic for the acetoxylation step and the condi-
tions (K2CO3 in MeOH) for the saponification required ex-
amination as well.
Scheme 9. Orthogonality between the thermodynamic and kinetic control
in hydrogenation by choice of catalyst.
On the other hand, selenium
dioxide oxidation of 34 effi-
ciently yielded 39. The reaction
proved to be very clean. Finally,
selective
monobromination
ortho to the phenolic group
completed the sequence to the
natural product. This was
smoothly effected by following
the conditions described by
Krohn,[27] which entail the slow
addition via a syringe pump of
stoichiometric amounts of NBS
in the presence of catalytic
iPr2NH in CH2Cl2 at 08C. Com-
parison of the data to that pre-
viously reported[1,28,29] con-
firmed their identify.
Scheme 10. Kinetic versus thermodynamic control of diastereoselectivity of hydrogenation.
Conclusion
Prior to our work, Nicolaou et al. reported the first as well
as asymmetric synthesis utilizing a novel [4+2] photocy-
cloaddition from N-tert-butyl-2-methoxy-p-toluamide.[28,29]
During the course of our studies, Clive reported a racemic[30]
and asymmetric synthesis, the
latter from g-butyrolactone and
2-bromo-6-methoxy-p-tolualde-
hyde.[31] Herein, a short and
concise synthesis of hamiger-
an B has been achieved in 15
linear steps in 10% overall
yield from 2-methylcyclopenta-
none. This synthesis represents
the
shortest
reported
to
Following these results, we did not further envision the
isomerization of the quaternary double bond as an attractive
and efficient option to recycle our product 30a formed
date[23,28–31] and highlights the efficiency of the palladium
catalyzed asymmetric alkylation of ketone enolates. Indeed,
that key step allowed us to introduce the quaternary center
Chem. Eur. J. 2005, 11, 951 – 959
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
957

B. Trost et al.
J=12.8, 6.4, 7.7 Hz, 1H), 1.35 (s, 9H),
1.0 (s, 3H); 13C NMR (CDCl3,
125 MHz):
d = 210.8, 149.0, 134.6,
117.6, 115.2, 79.7, 49.4, 41.2, 32.5, 28.2,
22.0, 21.2; elemental analysis calcd
(%) for C14H22O2: C 75.63, H 9.97;
found C 75.84, H 10.18.
Heck reaction
(3aS,9bR)-1-Isopropyl-6-methoxy-
3a,8-dimethyl-3,3a,4,9b-tetrahydro-cy-
clopenta[a]naphthalen-5-one (24): The
solvent used for the reaction was care-
fully degassed by successive freeze
thaw cycles. To
[Pd(OAc)2] (107 mg,
a
solution of
0.48 mmol,
10 mol%) in toluene (24 mL) in
a
100 mL round bottom flask, was
added 1,3-(diphenylphosphino)butane
(dppb) (409 mg, 0.96 mmol, 20 mol%).
Scheme 11. Formation of diketone 39.
After 15 min, a solution of triflate 25
(2.084 mg, 4.8 mmol) in toluene (24 mL) was added, followed by K2CO3
(1.904 g, 14.4 mmol, 3 equiv). After 5 min, the reaction flask was fitted
with a reflux condenser and heated at 1078C for 8 h (oil bath already at
desired temperature). The product distribution is then determinated by
GC.
in 93% ee, which controls the relative configuration of the
C-5 and C-6 contiguous stereocenters. Thus, one asymmetric
step is necessary to install stereoselectively three stereocen-
ters. Furthermore, the unusual nature of the structure is
highlighted by two abnormal reactivities: first, the formation
of an exocyclic tetrasubstituted double bond during the
Heck studies and second, the high propensity to give net re-
duction of the trisubstituted double bond of 24a from the
more hindered face. The orthogonality between the thermo-
dynamic and kinetic control in hydrogenation by choice of
catalyst is noteworthy.
After cooling down the reaction mixture, the mixture is filtered on silica
gel (5% to 60% ether/petroleum ether) to afford the Heck products
(1.2 g, 91%) containing 58% of the desired endocyclic double bond 24:
Rf =0.25 (50% ether/petroleum ether); [a]2D6:5 =170.1(Æ0.5)8 (c = 0.39,
chloroform, 93% ee); IR (neat): n˜ = 2957, 2869, 1681, 1606, 1567, 1455,
1
1413, 1328, 1298, 1103, 1042 cmÀ1; H NMR (CDCl3, 500 MHz): d = 6.69
(s, 1H), 6.65 (s, 1H), 5.40–5.39 (m, 1H), 3.87 (s, 3H), 3.56 (s, 1H), 2.80
(d, J=14.5 Hz, 1H), 2.37 (s, 3H), 2.31 (d, J=14.5 Hz, 1H), 2.23–2.12 (m,
3H), 1.22 (s, 3H), 1.11 (d, J=6.7 Hz, 3H), 0.77 (d, J=6.7 Hz, 3H);
13C NMR (CDCl3, 125 MHz): d = 198.4, 159.4, 151.5, 145.2, 144.5, 122.9,
121.0, 119.5, 110.6, 57.0, 55.8, 50.6, 45.1, 44.4, 27.2, 25.7, 22.3, 22.2, 21.4;
HR-MS: m/z: calcd for C19H24O2: 284.1776, found 284.1784 [M]+.
Experimental Section
Hydrogenation
(1R,3aS,9bS)-6-Hydroxy-1-isopropyl-3a,8-dimethyl-1,2,3,3a,4,9b-hexahy-
dro-cyclopenta[a]naphthalen-5-one (34): Ir black from Aldrich (60 mg,
0.31 mmol, 15 mol%) was added to a solution of phenol 24a (562 mg,
2.078 mmol) in EtOH (21 mL). The heterogeneous solution was then
charged with 1500 psi of hydrogen pressure. The mixture was stirred at
room temperature for 16 h and filtered over Celite to leave the desired
stereoisomer 34 as a white solid (564 mg, 99%). Rf =0.60 (50% ether/pe-
troleum ether); m.p. 142–1438C (EtOAc/hexane); [a]2D5 = À59.98 (c =
0.78, chloroform, 93% ee); IR (neat): n˜ = 2952, 2872, 2360, 1635, 1566,
Pd AAA reaction
(S)-2-Allyl-5-tert-butoxymethylene-2-methyl-cyclopentanone (4): A 1 L
round bottom flask was charged with 1,2-dimethoxyethane (DME;
165 mL) and diisopropylamine (10.12 g, 14.18 mL, 100 mmol). The so-
lution was cooled to À788C and n-butyllithium (1.6m in hexanes,
62.4 mL, 100 mmol) was added dropwise. The resultant clear solution was
stirred at À788C for 15 min and a solution of cyclopentanone 5 (9.115 g,
50 mmol) in DME (35 mL) was added via cannula. The resultant orange
solution was stirred at 08C for 15 min, charged with anhydrous tert-buta-
nol (25.94 g, 32.83 mL, 350 mmol), followed by a solution of trimethyltin
chloride (9.96 g, 50 mmol) in DME (35 mL). The resultant red orange so-
lution was stirred at 08C for 15 min and then re-cooled to À788C. To this
solution was added a heterogeneous mixture of allyl palladium chloride
dimer (219 mg, 0.6 mmol), (R,R)-Trost standard ligand (863 mg,
1.25 mmol), and allyl acetate (11.01 g, 11.89 mL, 110 mmol) in DME
(35 mL). The resultant red-orange heterogeneous mixture was stirred
overnight at room temperature. The reaction mixture was diluted with
water (200 mL) and the layers were separated. The aqueous layer was re-
extracted twice with Et2O (2ꢃ140 mL), and the combined organic phases
were washed with water and saturated brine (1ꢃ200 mL each) and dried
over magnesium sulfate. Concentration in vacuo, drying of the crude oil
under high vacuum (0.3 Torr) for 90 min and purification of the residue
on silica gel (15% ether/petroleum ether) afforded 4 as yellow oil
(8.56 g, 77%). Rf =0.40 (20% ethyl acetate/petroleum ether); determina-
tion of enantiomeric excess: chiral GC (cyclosil B, isotherm 1208C),
tR(major)=70.162 min, tR(minor)=71.370 min; [a]2D6:9 =À75.28 (c = 1.47,
chloroform, 93% ee); IR (neat): n˜ = 2977, 2868, 1708, 1631, 1456, 1371,
1467, 1450, 1364, 1350, 1307, 1256, 1225, 1190, 804 cmÀ1 1H NMR
;
(CDCl3, 500 MHz): d = 12.48 (s, 1H), 6.60 (s, 1H), 6.52 (s, 1H), 3.00 (d,
J=11.7 Hz, 1H), 2.67 (d, J=17.3 Hz, 1H), 3.53–2.46 (m, 1H), 2.32 (s,
3H), 2.25 (dd, J=17.3, 1.4 Hz, 1H), 1.86–1.81 (m, 1H), 1.70–1.55 (m,
4H), 1.10 (s, 3H), 0.72 (d, J=6.7 Hz, 3H), 0.48 (d, J=6.8 Hz, 3H);
13C NMR (CDCl3, 125 MHz): d = 205.2, 162.3, 147.5, 144.7, 123.2, 115.5,
114.6, 51.8, 49.2, 46.0, 43.8, 39.6, 28.6, 26.9, 25.6, 23.5, 22.2, 18.2; elemen-
tal analysis calcd (%) for C18H24O2: C 79.37, H 8.88; found: C 79.50, H
8.83.
(1R,3aR,9bR)-6-Hydroxy-1-isopropyl-3a,8-dimethyl-2,3,3a,9b-tetrahydro-
1H-cyclopenta[a]naphthalene-4,5-dione (39): SeO2 (132 mg, 1.18 mmol)
was added to a solution of phenol 34 (322mg, 1.18 mmol) in dioxane
(7 mL), H2O (225 mL) and 1 drop AcOH. The resultant suspension was
stirred for 24 h at 1008C. The yellow solution was then cooled down to
room temperature and filtered through a silica gel pad to remove inor-
ganic products. After evaporation, the crude mixture was purified by
chromatography on silica gel (5% to 10% Et2O/PE) to give 39 as a
yellow solid (248 mg, 73%) together with starting material 34 (60 mg,
90% based on reovered starting material). 39: Rf =0.23 (10% ether/pe-
troleum ether); m.p. 92–938C; [a]2D3:7 = À189.08 (c = 0.29, dichlorome-
thane); IR (neat): n˜ = 2959, 1725, 1634, 1567, 1453, 1368, 1338, 1303,
1264, 1206, 1156, 980, 945 cmÀ1; H NMR (CDCl3, 500 MHz): d = 7.51 (t,
J=2.3 Hz, 1H), 5.76–5.68 (m, 1H), 5.05–5.00 (m, 2H), 2.49–2.38 (m,
2H), 2.19–2.11 (m, 2H), 1.84 (ddd, J=12.8, 7.1, 8.2 Hz, 1H), 1.58 (ddd,
1
958
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 951 – 959

Hamigeran B
FULL PAPER
1217, 1197, 1154, 1098, 1035, 936, 851, 743 cmÀ1
;
1H NMR (CDCl3,
[8] B. M. Trost, G. M. Schroeder, J. Am. Chem. Soc. 2004, in press.
[9] J. H. Chan, J. S. Hong, R. N. Hunter III, G. F. Orr, J. R. Cowan,
D. B. Sherman, S. M. Sparks, B. E. Reitter, C. W. Andrews III, R. J.
Hazen, M. St. Clair, L. R. Boone, R. G. Ferris, K. L. Creech, G. B.
Roberts, S. A. Short, K. Weaver, R. J. Ott, J. Ren, A. Hopkins, D. I.
Stuart, D. K. Stammers, J. Med. Chem. 2001, 44, 1866–1882.
[10] T. Toyoda, K. Sasakura, T. Sugasawa, J. Org. Chem. 1981, 46, 189–
191.
[11] H. A. Dondas, M. Frederickson, R. Grigg, J. Markandu, M. Thorn-
ton-Pett, Tetrahedron 1997, 53, 14339–14354.
[12] S. H. Yang, S. Chang, Org. Lett. 2001, 3, 4209–4211.
[13] A. Kawada, S. Mitamura, J.-I. Matsuo, T. Tsuchiya, S. Kobayashi,
Bull. Chem. Soc. Jpn. 2000, 73, 2325–2333.
[14] J. H. Fosberg, V. T. Spaziano, T. M. Balasubramanian, G. K. Liu,
S. A. Kinsley, C. A. Duckworth, J. J. Poteruca, P. S. Brown, J. L.
Miller, J. Org. Chem. 1987, 52, 1017–1021.
[15] T. R. Kelly, Q. Li, V. Bhushan, Tetrahedron Lett. 1990, 31, 161–164.
[16] M. Mori, N. Kaneta, M. Shibasaki, J. Org. Chem. 1991, 56, 3486–
3493.
[17] A. X. Huang, Z. Xiong, E. J. Corey, J. Am. Chem. Soc. 1999, 121,
9999–10003.
[18] B. M. Trost, K. M. Pietrusiewicz, Tetrahedron Lett. 1985, 26, 4039.
[19] H. Azizian, C. Eaborn, A. Pidcock, J. Organomet. Chem. 1981, 215,
49–58.
500 MHz): d = 11.92 (s, 1H), 6.72 (s, 1H), 6.69 (s, 1H), 3.39 (d, J=
9.1 Hz, 1H), 2.61 (ddd, J=13.3, 7.7, 5.5 Hz, 1H), 2.38 (s, 3H), 2.30–2.24
(m, 1H), 1.83–1.76 (m, 1H), 1.70–1.64 (m, 1H), 1.57–1.51 (m, 1H), 1.29
(s, 3H), 1.24–1.16 (m, 1H), 0.54 (d, J=6.6 Hz, 3H), 0.42 (d, J=6.5 Hz,
3H); 13C NMR (CDCl3, 125 MHz): d = 200.1, 184.3, 164.6, 150.8, 144.1,
123.4, 116.7, 116.2, 56.8, 56.4, 51.4, 33.8, 28.1, 26.9, 24.4, 23.2, 22.5, 19.8;
elemental analysis calcd (%) for C18H22O3: C 75.50, H 7.74; found: C
75.71, H 7.49.
Hamigeran B (1): A solution of NBS (50 mg, 0.03 mmol) in CH2Cl2
(20 mL) was added dropwise over ꢀ4 h (syringe pump) to a stirred and
cooled (08C) solution of diketone 39 (80 mg, 0.028 mmol) and iPr2NH
(ca. 0.2 mL, 0.14 mmol) in CH2Cl2 (40 mL). The ice bath was removed
and the mixture was stirred for 3 h. Evaporation of the solvent and flash
chromatography of the residue over silica gel, using 7:3 CH2Cl2/toluene
gave hamigeran B (86 mg, 85%) as a yellow solid. Rf =0.36 (30% tolu-
ene/CH2Cl2); m.p: 162–1638C (lit.[1] m.p. 163–1658C;) [a]2D5:8 =À211.18 (c
= 0.15, dichloromethane); IR (neat): n˜ = 2958, 2361, 1996, 1725, 1634,
1540, 1436, 1397, 1280, 1169, 1030, 778, 743 cmÀ1 1H NMR (CDCl3,
;
500 MHz): d = 12.6 (s, 1H), 6.82 (s, 1H), 3.39 (d, J=9.2 Hz, 1H), 2.64
(ddd, J=13.1, 7.6, 5.2 Hz, 1H), 2.51 (s, 3H), 2.34–2.27 (m, 1H), 1.84–1.77
(m, 1H), 1.72–1.65 (m, 1H), 1.58–1.52 (m, 1H), 1.29 (s, 3H), 1.22–1.16
(m, 1H), 0.51 (d, J=6.7 Hz, 3H), 0.45 (d, J=6.5 Hz, 3H); 13C NMR
(CDCl3, 125 MHz): d = 199.0, 184.4, 160.8, 150.2, 142.7, 124.2, 117.2,
111.5, 56.9, 56.2, 51.2, 33.7, 28.1, 26.7, 24.4, 24.3, 23.3, 19.7.
[20] V. Farina, B. Krishnan, D. R. Marshall, G. P. Roth, J. Org. Chem.
1993, 58, 5434–5444.
[21] I. P. Beletskaya, A. V. Cheprakov, Chem. Rev. 2000, 100, 3009–3066.
[22] B. M. Trost, W. Tang, Angew. Chem. 2002, 114, 2919–2921; Angew.
Chem. Int. Ed. 2002, 41, 2795–2797.
[23] After completion of our work, Mehta reported a strategy similar to
ours that resulted in a synthesis of 6-epi-hamigeran B. In his intra-
molecular Heck reaction he reports a 2:1 mixture of 24 and the
simple hydrogenolysis product 32 in a total 55–60% yield but does
not report the tetrasubstituted olefin; see G. Metha, H. M. Shinde,
Tetrahedron Lett. 2003, 44, 7049–7053.
Acknowledgement
We thank the National Institutes of Health (GM-13598) and the National
Science Foundation for their generous support of our programs. Dr.
Carole Pissot-Soldermann thanks the Swiss National Science Foundation
and La Sociꢄtꢄ de Secours des Amis des Sciences and Mr. Irwin Chen
the US National Science Foundation for fellowships. We thank Prof. D.
Clive for experimental details regarding the bromination step.
[24] S. Nishimura, F. Mochizuki, S. Kobayakawa, Bull. Chem. Soc. Jpn.
1970, 43, 1919.
[25] S. Nishimura, H. Sakamoto, T. Ozawa, Chem. Lett. 1973, 855–858.
[26] B. M. Trost, J. Am. Chem. Soc. 1977, 99, 4405–4412.
[27] K. Krohn, S. Bernhard, U. Flçrke, N. Hayat, J. Org. Chem. 2000, 65,
3218–3222.
[28] K. C. Nicolaou, D. Gray, J. Tae, Angew. Chem. 2001, 113, 3791–
3795; Angew. Chem. Int. Ed. 2001, 40, 3679–3683.
[29] K. C. Nicolaou, D. L. F. Gray, J. Tae, J. Am. Chem. Soc. 2004, 126,
613–627.
[30] D. L. J. Clive, J. Wang, Tetrahedron Lett. 2003,44, 7731–7733.
[31] D. L. J. Clive, J. Wang, Angew. Chem. 2003, 115, 3528–3531; Angew.
Chem. Int. Ed. 2003, 42, 3406–3409.
[1] K. D. Wellington, R. C. Cambie, P. S. Rutledge, P. R. Bergquist, J.
Nat. Prod. 2000, 63, 79–85.
[2] B. M. Trost, G. M. Schroeder, J. Am. Chem. Soc. 1999, 121, 6759–
6760.
[3] a) B. M Trost, M. L. Crawley, Chem. Rev. 2003, 103, 2921–2944;
b) B. M. Trost, D. L. Van Vranken, C. Bingel, J. Am. Chem. Soc.
1993, 115, 441–458.
[4] B. M. Trost, Chem. Pharm. Bull. 2002, 50, 1–14.
[5] B. M. Trost, D. L. Van Vranken, Chem. Rev. 1996, 96, 395–422.
[6] B. M. Trost, C. B. Lee, Asymmetric Allylic Alkylation Reactions, in
Catalytic Asymmetric Synthesis (Ed.: I. Ojima), Wiley-VCH, New
York, 2000, pp. 593–650.
[7] B. M. Trost, C. Pissot-Soldermann, I. Chen, G. M. Schroeder, J. Am.
Chem. Soc. 2004, 126, 4480–4481.
Received: June 4, 2004
Published online: December 20, 2004
Chem. Eur. J. 2005, 11, 951 – 959
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
959