Published on Web 03/04/2006
Total Synthesis of (()-Haouamine A
Phil S. Baran* and Noah Z. Burns
Department of Chemistry, The Scripps Research Institute, 10550 North Torrey Pines Road,
La Jolla, California 92037
Received January 15, 2006; E-mail: pbaran@scripps.edu
Structurally unique natural products have historically served as
conduits to mend gaps in the science of synthesis.1 Recently, Zub´ıa
and co-workers discovered the structurally unprecedented alkaloids
haouamine A and B (1 and 2, Figure 1) from a tunicate (Aplidium
haouarianum) residing off the southern coast of Spain.2 Their
heptacyclic framework of mysterious biosynthetic origin exists as
an inseparable mixture of two interconverting isomers that accom-
modate a congested indeno-tetrahydropyridine ring system, an
unusual oxygenation pattern, and a highly strained aza-paracyclo-
phane with a bent aromatic ring. These architectural features
combined with exquisitely selective anticancer activity in human
colon carcinoma cells (IC50 ) 0.1 µg/mL) make 1 a particularly
attractive target for total synthesis.3 Herein, we describe a short
and efficient total synthesis of 1 facilitated by the invention of a
powerfully simplifying annulation method to access the indeno-
tetrahydropyridine ring system and a new method for macrocy-
clization leading to the bent aromatic ring.
Initial forays into the haouamine problem were guided by our
postulated biosynthesis put forth in Figure 1 wherein the key C-C
and C-N bonds would be formed through simple oxidation and
condensation events of ammonia and four meta-hydroxylated
phenylacetaldehydes. The final retrosynthetic analysis (Figure 1)
was derived from an amalgamation of empirical intelligence gained
from “failures” in attempting to reduce this plan to practice. Thus,
the indeno-tetrahydropyridine 3 was chosen as an ideal intermediate
from which to evaluate strategies for forging the unusual aza-
paracyclophane, which in turn was envisioned as arising from oxime
4 Via a cascade annulation sequence designed exclusively for this
ring system.
The synthesis commenced with alkylation of the thermodynamic
potassium enolate of readily available 54 with allylic iodide 64 in
54% yield followed by oxime formation to furnish 4 in 75% yield
(Scheme 1). On the basis of the precedent set by Grigg and co-
workers,5 exposure of 4 to an electrophilic halogen source was
expected to elicit a 5-exo-trig cyclization to nitrone 7. It was
reasoned that a subsequent reduction of the nitrone would occur
stereoselectively to produce 8, which, in principle, could cyclize
to the unprecedented N-hydroxyaziridinium species 9. Assuming
that this new chemical entity can even be formed, orbital consid-
erations6 allow one to predict a subsequent position selective
fragmentation to the desired N-hydroxypiperidine 10 in preference
to the alternate N-hydroxyl enamine. Chemoselective reduction of
the N-O bond (in the presence of the aryl bromide and olefin)
could then lead to the key intermediate 3. This designed cascade
could be reduced to practice by executing the following sequence:
(1) treatment of 4 with 2,4,4,6-tetrabromo-2,5-cyclohexadienone
at 0 °C to furnish the intermediate nitrone 7 as an inconsequential
mixture of two diastereomers, (2) immediate reduction with NaBH4
at 50 °C to the cis-fused hydroxy pyrrolidine 8, (3) continued heat-
ing for 1 h to elicit hydroxyaziridinium formation and subsequent
ring expansion to 10, and (4) selective reduction7 of the crude
N-hydroxylpiperidine using In0 to furnish 3 in 57% overall yield
(corresponding to an average yield of 90% per transformation). The
structures of the stable intermediates 7 and 10 were secured spec-
troscopically by intercepting the cascade and by X-ray crystal-
lographic analysis (Figure 2) of related intermediates 7′ (prepared
Figure 1. Retrosynthetic analysis of haouamine A (1).
from debromo-4 using t-BuOCl) and 10′ (debromo-10).4 Although
the N-hydroxyaziridinium species 9 was not spectroscopically ob-
servable, it is logical to invoke such an intermediate by analogy to
the chemistry of related pyrrolidine systems lacking an N-hydroxyl.8
With gram-scale access to the core indeno-tetrahydropyridine
ring system in only three steps, attention turned to construction of
the hallmark aza-paracyclophane sector. The pseudo-boat config-
uration adopted by one of the phenols clearly posed a challenge to
the current state of the art in macrocyclization methodology. Not
surprisingly, several standard approaches such as transition metal-
based biaryl synthesis, Witkop photocyclization,9 and intramolecular
alkylation all failed. A new strategy for this type of cyclization was
conceiVed of on the premise that a nonaromatic conformational
mimic of the bent aromatic ring might serVe as a Viable precursor
if it were able to undergo subsequent aromatization. The pyrone-
alkyne Diels-Alder reaction10 fits these criteria since it leads to a
cyclohexadiene having a boat configuration and an embedded
leaving group (CO2). To the best of our knowledge, such a strategy
for macrocyclization is unknown.
Implementation of this approach commenced with a Stille
coupling of pyrone 124,11 with Boc-protected piperidine 11 to afford
the pyrone-piperidine conjugate 13. Subsequent Boc removal and
alkylation delivered 14 in 70% yield. Global demethylation using
BBr3 and peracetylation afforded 15 in 67% yield and set the stage
for the key macrocyclization event. After extensive experimentation,
15 was converted into 1 upon microwave irradiation of 15 in di-
chlorobenzene at 250 °C for 10 h followed by basic (K2CO3,
MeOH) acetate hydrolysis. The use of fully acetylated precursor 15
was essential to minimize thermal decomposion of 1. Remarkably,
the macrocyclization event proceeded with high atropselectivity in
favor of 1 (10:1, separable by HPLC).12 Synthetic (()-haouamine
A (1), isolated in 21% yield along with 30% of recovered 15
(separated by PTLC prior to deacetylation), was spectroscopically
identical to a natural sample kindly provided by Professor Zub´ıa.
The synthetic challenge posed by the ornate molecular archi-
tecture of haouamine has been met in eight steps from readily
available indanone 5 (prepared in two steps from commercially
9
3908
J. AM. CHEM. SOC. 2006, 128, 3908-3909
10.1021/ja0602997 CCC: $33.50 © 2006 American Chemical Society