Total Synthesis of Hybocarpone and Analogues Thereof. A Facile Dimerization of Naphthazarins to Pentacyclic Systems

Article abstract of DOI:10.1021/ja030497n

The total synthesis of the lichen-derived antitumor agent hybocarpone (1) and related compounds is described. The successful route to hybocarpone features a novel radical-based dimerization/hydration cascade which generates the bridging hindered carbon-ca

Full text of DOI:10.1021/ja030497n

Published on Web 12/17/2003
Total Synthesis of Hybocarpone and Analogues Thereof. A
Facile Dimerization of Naphthazarins to Pentacyclic Systems
K. C. Nicolaou* and David L. F. Gray
Contribution from the Department of Chemistry and The Skaggs Institute for Chemical Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037, and
Department of Chemistry and Biochemistry, UniVersity of California, San Diego,
9500 Gilman DriVe, La Jolla, California 92093
Received August 18, 2003; E-mail: kcn@scripps.edu
Abstract: The total synthesis of the lichen-derived antitumor agent hybocarpone (1) and related compounds
is described. The successful route to hybocarpone features a novel radical-based dimerization/hydration
cascade which generates the bridging hindered carbon-carbon bond of the molecule in a stereocontrolled
manner, setting the relative configurations of the four contiguous stereocenters in a single step. The
conjecture is made that this process may not be so dissimilar to the biosynthetic pathway leading to the
formation of hybocarpone in nature. The developed sequence to these molecular frameworks also features
the first example of a synthetically useful Diels-Alder trapping of a photochemically generated hydroxy-
o-quinodimethane species with a 1,1-disubstituted olefin to form a quaternary center, and includes an efficient
route to hydroxynaphthoquinone-type structures represented by the monomeric subunit of the natural
product.
Introduction
Hybocarpone (1; Figure 1) is a naturally occurring compound
which was first described in 1999 by J. Elix and his team at
the University of Canberra as part of their program of identifying
lichen metabolites with important biochemical properties.1
Hybocarpone is one of three identifiable compounds produced
by the lichen species Lecanora hybocarpa, which grows in the
Louisiana woodland. When spores are collected from these
lichen and cultured in a specific way, 1 can be isolated from
the mycobiont. This novel pentacyclic compound was shown
to be cytotoxic against the murine mastocytoma P815 trans-
plantable tumor cell line with an IC50 of 150 ng/mL.1 At the
present time, the mechanism by which this natural product
effects cell death in cancer cells is unknown. The flat, aromatic
character of the hybocarpone structure, coupled with its dense
concentration of reactive functionalities such as the 1,2,4-
triphenolic and diketone moieties, however, suggests a possible
DNA intercalation/DNA damage pathway as a viable mode of
action for this compound. A closely related class of natural
products, the naphthazarins, have been studied in detail, leading
some researchers to conclude that they interfere with the cellular
DNA replication machinery (e.g., DNA topoisomerase I and/
or II via formation of a cleavable complex).2 The various
mechanisms proposed to explain the biological activity of the
naphthazarins rely on the quinone moiety as the key functionality
Figure 1. Molecular structures of 1 and related naphthazarin natural
products (2-4).
reacting with biological nucleophiles or electron-transfer re-
agents as a means to induce the observed biological effects.3
Structurally, 1 is composed of two naphthazarin units; however,
it contains no quinone functionality, rendering any comparison
between hybocarpone and the naphthazarins rather speculative
in terms of their mode of action.
Several other lichen species are reported to produce naphtha-
zarin monomers with structures related to that of 1.4-6
(3) (a) Ahn, B.-Z.; Sok, S.-E. Curr. Pharm. Des. 1996, 2, 247-262. (b) You,
Y.-J.; Zheng, X.-G.; Yong, K.; Ahn, B.-Z. Arch. Pharm. Res. 1998, 21,
595-598.
(1) Ernst-Russel, M.; Elix, J.; Chai, C.; Willis, A.; Hamada, N.; Nash, T., III.
Tetrahedron Lett. 1999, 40, 6321-6324.
(4) (a) Yamamoto, Y.; Matsubara, H.; Kinoshita, Y.; Kinoshia, K.; Koyama,
K.; Takahashi, K.; Ahmadjiam, V.; Kurokawa, T.; Yoshimura, I. Phy-
tochemistry 1996, 43, 1239-1242. (b) Chai, C.; Elix, J.; Moore, F.
Tetrahedron Lett. 2001, 42, 8915-8917.
(2) (a) For a review of such naphthazarin natural products including alkannin
and shikonin, see: Papageorgiou, V.; Assimopoulou, A.; Couladouros, E.;
Hepworth, D.; Nicolaou, K. C. Angew. Chem., Int. Ed. 1999, 38, 271-
300. See also: (b) Song, G.-Y.; Kim, Y.; Zheng, X.-G.; You, Y.-J.; Cho,
H.; Chung, J.-H.; Sok, D.-E.; Ahn, B.-Z. Eur. J. Med. Chem. 2000, 35,
291-298.
(5) Stepanenko, L.; Krivoshchekova, O.; Dmitrenok, P.; Maximov, O. Phy-
tochemistry 1997, 46, 565-568.
(6) Berg, A.; Gorls, H.; Dorfelt, H.; Walther, G.; Schlegel, B.; Grafe, U. J.
Antibiot. 2000, 11, 1293.
9
10.1021/ja030497n CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 607-612
607

A R T I C L E S
Nicolaou and Gray
Apparent from Figure 1 is the fact that the monomeric subunit
of hybocarpone is itself a natural product (3). The red
compounds christazarin (2) and 6-methylchristazarin (3) have
been isolated from the cultured mycobiont of Cladonia cris-
tatella.4,5 Shortly after the disclosure of the hybocarpone
structure, the Berg group isolated yet another naturally occurring
hydroxynaphthoquinone possessing interesting biological activ-
ity, which they termed aureoquinone (4; Figure 1).6
The molecular structure of 1 was proposed using a combina-
tion of spectroscopic and molecular modeling data and later
confirmed by single-crystal X-ray crystallographic analysis. It
is characterized by a novel and unprecedented molecular
architecture containing a dinaphtho[2,3-b:2,3-d]furantetraone
skeleton. The hydration of the two carbonyl functionalities, while
not surprising, raises questions of molecular and isomeric
stability. Hybocarpone is actually a hydrated dimer, possessing
a C2 element of symmetry. This structural property is vividly
illustrated by the molecule’s simple 13C NMR spectrum, which
exhibits only 13 carbon signals.1 While examples of the
monomeric hydroxynaphthoquinones abound, it is the dimeric
nature of hybocarpone which makes this naturally occurring
substance especially intriguing. It is reasonable and tempting
to speculate that nature is able to successfully dimerize 3 in a
diastereo- and enantioselective fashion to produce 1.
Prompted by the striking molecular structure and biological
activity of 1, we initiated a program directed toward its total
synthesis, with the specific initial goal of identifying an efficient
means of chemically constructing the unusual central carbon-
carbon bond. In this paper, we describe in detail our results in
this area, which culminated in the total synthesis7 of 1 and a
number of designed analogues (vide infra).
Figure 2. Potential hybocarpone 6 and hydroxynaphthoquinone 7 libraries
from tetralones 5.
Figure 3. Proposed general synthesis of varied tetralones 5 via photo-
chemically induced benzannulation of o-methylbenzaldehydes 8. EWG )
electron-withdrawing group.
odimethanes 9 (Figure 3) from o-alkylbenzaldehydes 8 and their
subsequent trapping in Diels-Alder reactions to afford ben-
zannulated bicycles in varying yields.10-12 We thus reasoned
that if appropriately developed into a synthetically useful
process, such technology could serve as the cornerstone for a
general and flexible tetralone synthesis as projected in Figure
3.
Results and Discussion
As noted above, disconnection of the central carbon-carbon
bond and hemiketals that adjoin the monomeric units of 1 leads
to the known lichen metabolite 3 (Figure 1). This retrosynthetic
disconnection is productive only if one can successfully identify
a synthetic process for constructing this bond in a selective
fashion. At the time this work began, there was no known
method for the direct formation of such a bond. Thus, initial
efforts focused on a dimerization strategy which had a more
solid literature precedent. Thus, oxidative dimerization of ketone
enolates held promise, as it has been shown to be capable of
forging highly congested carbon-carbon bonds to produce 1,4-
diketones.8,9 Mindful of the biological activity exhibited by the
naphthazarins, our initial synthetic designs had, as a secondary
goal, the development of methodology for the construction of
compound libraries with varying substitutions attached to the
rings, including naphthazarins of type 7 (Figure 2) and penta-
cycles of type 6 (Figure 2). It seemed reasonable that tetralones
5 (Figure 2) would be a good platform from which to launch
such an expedition.
Absent as reaction partners among the literature examples
of photoenolization/Diels-Alder (PEDA) dienophiles were 1,1-
disubstituted olefins, which would lead to the formation of
quaternary carbon centers. Given the quaternary centers present
in our targeted structure, we set out to test whether such partners
could be induced to react in the desired manner. Aldehyde 12
(Scheme 1) is accessible without column chromatography
through a few simple operations starting from the inexpensive
2,6-dimethylphenol by following the procedures of Liebeskind
and Fieser13-15 (with a few significant modifications; see the
Experimental Section in the Supporting Information). For the
initial PEDA reaction attempt, this aldehyde (12) was dissolved
in benzene with 2 equiv of methyl R-ethacrylate (13)16 and
irradiated for 8 h according to the most common procedure used
for such reactions.17 Gratifyingly, the expected union did occur,
delivering the targeted annulation product 14 in ca. 30% yield
as a 2:1 mixture of syn- and anti-diastereoisomers (subsequent
optimization increased the yield of this reaction to 81%).18
Initial Approach to the Hybocarpone Framework. While
searching for a methodology suitable for the construction of
the required tetralone monomeric units, we came across reports
on the photochemical generation of reactive hydroxy-o-quin-
(10) Weedon, A. C. In The Chemistry of Enols; Rappoport, Z., Ed.; Wiley:
Chichester, U.K., 1990; pp 591-638 and references therein.
(11) Charlton, J.; Alauddin, M. Tetrahedron 1987, 43, 2873-2889 and references
therein.
(12) Sammes, P. Tetrahedron 1976, 32, 405 and references therein.
(13) Liotta, D.; Arbiser, J.; Short, J.; Saindane, M. J. Org. Chem. 1983, 48,
2932-2933.
(14) Iyer, S.; Liebeskind, L. J. Am. Chem. Soc. 1987, 109, 2759-2770.
(15) Fieser, L.; Ardao, M.-I. J. Am. Chem. Soc. 1956, 78, 774-781.
(16) Stetter, H.; Kuhlmann, H. Synthesis 1979, 1, 29-30.
(17) Charlton, J.; Plourde, G.; Koh, K. Can. J. Chem. 1989, 67, 574-579.
(7) For a preliminary comunication, see: Nicoalou, K. C.; Gray, D. Angew.
Chem., Int. Ed. 2001, 40, 761-763.
(8) Frazier, R.; Harlow, R. J. Org. Chem. 1980, 45, 5408-5411.
(9) Mazzega, M.; Fabris, F.; Cossu, S.; De Lucchi, O.; Lucchini, V.; Valle, G.
Tetrahedron 1999, 55, 4427-4440.
9
608 J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004

Synthesis of Hybocarpone and Analogues Thereof
A R T I C L E S
Scheme 1. Construction of Dimer 16 and Unsuccessful Attempts
To Elaborate It to 1a
Scheme 2. Construction of a Hindered Carbon-Carbon Bond via
Claisen Rearrangement on a Model Systema
a Reagents and conditions: (a) LDA (1.2 equiv), (TBS)Cl (1.5 equiv),
HMPA, THF, -78 °C, 2 h; then warm to 60 °C, 1 h; (b) CH2N2, ether, 5
min; (c) LAH (6.0 equiv), THF, 4 h, 48% for three steps.
a Reagents and conditions: (a) 13 (6.0 equiv), hν, toluene, 4 h, 81%
(2:1 ratio of diastereoisomers, major isomer shown); (b) CrO3‚2py (6.0
equiv), CH2Cl2, 0-25 °C, 1 h, 86%; (c) 1 M aqueous KOH, EtOH, 90 °C,
6 h, 80%; (d) (i) LDA (1.1 equiv), THF, -78 °C, 1 h; (ii) then FeCl3 (1.1
equiv) in DMF, -78 to 0 °C, 4 h, 52%. LDA ) lithium diisopropylamide.
Benzylic oxidation was accomplished through the use of the
Collins reagent (CrO3‚2py) to give a â-keto ester (86% yield),
which was decarboxylated with KOH in refluxing EtOH to
access the targeted tetralone 15 in 80% yield (Scheme 1).
Among several dimerization protocols which were investi-
gated, the most successful was the Harlow procedure, which
employs anhydrous FeCl3 as the oxidant on the pregenerated
lithium enolate of the ketone substrate.8 Thus, exposure of the
enolate of 15 (generated by the action of LDA) to a solution of
FeCl3 in DMF using techniques for the rigorous exclusion of
oxygen allowed for the production of reasonable quantities of
the easily separable meso- and d,l-isomers of dimer 16 (Scheme
1). Unfortunately, this fully substituted tetralone dimer (16)
proved recalcitrant to oxidation, decomposing rapidly upon
exposure to several reagents, and preventing further advance-
ment toward the targeted system.19
A second approach, which looked to the powerful Ireland-
Claisen rearrangement to forge the sterically encumbered bond
at the junction joining the two monomeric units, also suffered
from difficulties encountered in bringing even simple model
systems to the desired oxidation state (Scheme 2). It is, however,
interesting to note the success of this approach, generating the
bond at the quaternary center of polycyclic structures 19 and
20 (Scheme 2).20
With these two approaches failing to deliver the desired
hybocarpone framework, it was finally concluded that a modi-
fied strategy involving the coupling of a highly oxidized
monomer might be more successful, on the basis of the
minimization of both the number and the difficulty of post-
dimerization transformations.
Figure 4. Second-generation retrosynthesis of hybocarpone hexamethyl
ether (21) bases on the oxidative dimerization of hydroxynaphthoquinone
24.
ates. In a redesigned approach toward 1, we aimed to follow a
synthetic strategy which was conceptually much simpler than
the initial attempt discussed above, and perhaps one that was
related to its biosynthesis. The symmetrical structure of 1 and
the existence of the monomeric subunit in nature led to a
conceptual identification of naphthazarin 24 (Figure 4) as a
potential precursor. Requirements for a successful implementa-
tion of this concept include (a) the identification of methodology
suitable for the construction of the extremely hindered carbon-
carbon bond that adjoins the two halves of the dimer and (b)
the successful control of the relative stereochemistry of the four
stereogenic centers of 1 during the assembly process. We
proposed that oxidation of a protected form (24) of 3 via a
single-electron-transfer (SET) process could lead to the highly
reactive radical cation 23, which might have a sufficient lifetime
to undergo intermolecular coupling, leading to intermediate 22
A Second-Generation Approach toward Hybocarpone
Relying on the Dimerization of Highly Oxidized Intermedi-
(18) Nicolaou, K. C.; Gray, D. L. F.; Tae, J. J. Am. Chem. Soc. 2004, 126,
613-627.
(19) Jacob, P. Callery, P. Shulgin, A. Castagnoli N. J. Org. Chem. 1976, 41,
3627-3629.
(20) For the first report of a Claisen rearrangement which creates vicinal
quaternary centers, see: Denmark, S.; Harmata, M. Tetrahedron Lett. 1984,
25, 1543-1546.
9
J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004 609

A R T I C L E S
Nicolaou and Gray
Scheme 3. Synthesis of Naturally Occurring Naphthazarins 3 and
4 via a Series of Selective Oxidation Reactionsa
Figure 5. Calculated relative strain energies of the six possible hybocarpone
diastereoisomers (R ) H) and their corresponding hexamethyl ethers (R )
Me) (see ref 21 for computational parameters).
after loss of two protons. It was anticipated that, at this stage,
a hydration/ketalization event on the hexaketone intermediate
22 would furnish the desired hybocarpone framework 21. This
hydration/cyclization sequence would potentially lead to the
formation of up to six diastereomeric furan systems by reaction
of diketone 22 (Figure 4). Molecular modeling and computa-
tional studies indicated that, among these diastereoisomeric
compounds (see the structures in Figure 5), the natural isomer
21 appeared to be clearly favored in terms of strain energy.21
Since the central dihydroxyfuran systems of these diastereo-
isomeric compounds can exist in equilibrium with their open-
chain counterparts, and because the calculated energy differences
among them are large (>5 kcal/mol), it seemed reasonable to
rely on thermodynamics to set the correct stereochemistry in
this proposed hydration process. Only demethylation separates
compound 21 from 1. Due to the compelling attractiveness of
this simple and straightforward strategy, we proceeded to attempt
its laboratory execution, although there was no literature
precedent for the key coupling reaction.
a Reagents and conditions: (a) Pb(OAc)4 (1.4 equiv), hν, AcOH, 2 h,
71% (3:1 ratio of diastereoisomers); (b) HCl, AcOH, 90 °C, 4 h, 72%; (c)
OsO4 (0.1 equiv), NMO (3.0 equiv), THF-tBuOH-H2O-py (20:20:4:1),
12 h, 92%; (d) IBX (3.0 equiv), DMSO, 25 °C, 0.5 h, 92%; (e) 1.5 M
aqueous KOH-THF (1:3.5), air, 1 h, 83%; (f) BBr3 (10.0 equiv), CH2Cl2,
-78 to +25 °C, 3 h, 40-50%. NMO ) N-methylmorpholine N-oxide, and
IBX ) 1-hydroxy-1,2-benziodoxol-3(1H)-one 1-oxide.
The new synthetic journey commenced from â-keto ester 35
(Scheme 3), which was already in hand from the previous
attempt as described above (Scheme 1). To progress from this
intermediate, conditions for the selective oxidation of the
secondary benzylic position in the presence of a benzylic methyl
group needed to be identified. Common protocols (e.g., NBS,
DDQ, PDC) were found to be nonselective, attacking indis-
criminately both potential oxidation sites within 35. However,
photochemically activated Pb(OAc)4 was observed to favorably
discriminate among the two positions (ratio ca. 25:1), making
this a highly useful transformation.22 The precise origin of the
observed selectivity and the generality and scope of this
transformation leading to acetate 37 in 71% yield (3:1 mixture
of acetate diastereoisomers) remain to be explored. Exposure
of this oxidation product to HCl in AcOH at 90 °C gave the
olefin 39 (72% yield), which was subjected to dihydroxylation
with NMO-OsO4(cat.) to afford vicinal diol 41 (92% yield,
ca. 10:1 ratio of diastereoisomers, stereochemistry not deter-
mined). While a number of common oxidation protocols failed
to yield the next targeted intermediate, IBX oxidation of 41
was selective and high-yielding, leading to the corresponding
hydroxyketone 43 (92% yield), which was directly dissolved
in THF and treated with aqueous KOH in the presence of air to
afford, via a decarboxylation/oxidation sequence, the desired
naphthazarin 24 in 83% yield (two steps). The unusual rapidity
of the latter decarboxylation and the temperature (ambient) at
(21) The hybocarpone diastereoisomeric structures were modeled with Accelrys
Insight II, 1998. Molecular dynamics calculations were performed with
Accelrys Discover v. 2.98 on a cluster of SGI Origin servers running Irix
6.5. The relative energies of the diastereoisomeric structures were computed
with the class II pairwise force field cff91 v. 2.0 (see: Dinur, U., Hagler,
A. In ReViews in Computational Chemistry; Lipkowitz, K., Boyd, D., Eds.;
VCH: New York, 1991; Vol. 2, p 527), with library parameters for bond,
valence, torsion, and out-of-plane terms. Typically, 500 structures were
generated within 250 ps at 800 K. They were annealed to 300 K in 5 ps.
The structures were further minimized to convergence using a conjugate
gradient algorithm, and the lowest energy structure cluster was selected as
the preferred structure. We thank Dr. I. Ioannou for performing these
calculations.
(22) We are aware of one other report of benzylic oxidation employing
photochemically activated Pb(OAc)4; see: Geiwiz, J.; Haslinger, E. HelV.
Chim. Acta 1995, 78, 818-832.
9
610 J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004

Synthesis of Hybocarpone and Analogues Thereof
A R T I C L E S
Scheme 4. Synthesis of 1 via Successful Application of the
brief event, exposure of 24 to 1 equiv of CAN in MeCN at
-35 °C under sonication conditions led to the isolation of two
isomeric hybocarpone derivatives, 21 and 25 (21:25 ratio ca.
3:2), which were separated on deactivated silica gel and
spectroscopically characterized. A key to success in this
transformation was the quench, which must be strongly basic
or no product is isolated. Although the yield of the pentacycles
25 and 21 was rather low (combined 19% in the best run), the
majority of the starting monomer 24 was recovered unchanged,
and in high purity, by acid/base extraction. This recovered
material (24) could be resubjected to dimerization without
additional purification. After three such reiterations, 25 and 21
were obtained in a combined 36% yield.
Oxidative Quinone Coupling Strategya
To our delight, the more complex 1H NMR spectrum (CDCl3,
500 MHz) of 25 (Scheme 4) was observed to undergo
simplification as this, apparently unfavorable, isomer was
transformed to the thermodynamically most stable hybocarpone
isomer (21). This acid-catalyzed transformation was strong
evidence that both of the isolated pentacycles bear the trans
relationship of the two ethyl groups at the junction joining the
two monomeric units of both isomers. This selectivity might
be predictable on the basis of severe steric congestion within
the cis isomer (see Figure 5). The generation of 1 from its
hexamethylated derivative (21) was effected by exposure to
AlBr3-EtSH in methylene chloride at 0 °C (60% yield).
Synthetic 1 was purified by reversed-phase (RP)-HPLC and fully
characterized by spectroscopic means. It exhibited spectral data
(1H and 13C NMR, mass spectrometry) identical to those
reported for natural hybocarpone.1,23
a Reagents and conditions: (a) 24 in MeCN; then add CAN (1.0 equiv)
in degassed MeCN, -35 to 0 °C, sonication, 2 min; then 5 M aqueous
KOH, 0-25 °C, 10 min, 19% plus 60% recovered starting material, 21:25
ratio ca. 3:2; (b) AcOH, CHCl3, 5 min, >95%; (c) AlBr3 (1 M in CH2Br2,
25 equiv), EtSH-CHCl2 (1:2), 0 °C, 1 h, 60%. CAN ) cerium ammonium
nitrate.
At this juncture, we wondered whether there was something
unique about 1 which facilitates its synthesis via an oxidative
dimerization approach. We pondered the notion that the
developed dimerization protocol could be applied to other
hydroxynaphthoquinones, leading to the corresponding penta-
cyclic products. To address this question, we prepared two
related compounds for dimerization studies. The first such
candidate was the hydroxynaphthoquinone 45, which differs
from 24 by the incorporation of a methyl, rather than an ethyl,
group on the reacting carbon atom. The synthesis of this
compound has already been mentioned above (Scheme 3). The
second substrate (48; Scheme 5) was prepared by hydrogenation
(PtO2, H2, EtOAc) of the naturally occurring (and commercially
available) lapachol (47). This molecule (48) retains a high degree
of steric bulk around the reacting carbon, but is unsubstituted
on the aryl ring, changing the electronics of the hydroxynaph-
thoquinone. In each case, the expected pentacyclic compound
(46, 49) was isolated upon exposure of the monomeric unit to
CAN in acetonitrile. For monomer 45, the reactivity appeared
to be similar to that of 24, although it was noted that more
starting material (70-80%) was recovered and a correspond-
ingly lower amount of product 46 was formed (6-9%) in each
of the several trials attempted. For the simpler precursor 48,
the dimerization process is noticeably less efficient (ca. 3% yield
of 49). However, in several runs, 93% of the starting material
was recovered unchanged and at a high level of purity. Also of
note is the different behavior of the dimeric pentacycles 46 and
49 as compared to 1. The lapachol-derived compound 49 was
isolated as a single isomer, and no other dimeric products were
which it proceeds suggest that the neighboring hydroxyl group
plays a facilitating role in this conversion (43 f 24, Scheme
3). Indeed, a similar substrate which bore a protecting group
on that hydroxyl group required elevated temperatures and
prolonged reaction times to effect decarboxylation.
The efficient arrival at 24 represented an important milestone
in the synthetic effort toward hybocarpone. Cleavage of the three
aryl ethers was accomplished in modest yield by exposure of
this intermediate (24; Scheme 3) to BBr3 (-78 to +25 °C) for
several hours. Thus, upon workup of the reaction mixture, the
deep-red-colored 3 was obtained in 49% yield. The advancement
of the latter compound (3) toward 1 was attempted under various
conditions, albeit without success. In a second foray and with
an eye on 4, the aforementioned synthetic sequence leading to
24 (Scheme 3) was revisited, beginning with methyl methacryl-
ate in the initial PEDA reaction to afford the new starting
material 36. This sequence led from the latter compound (36)
to the trimethyl ether-protected 45, which was deprotected to
afford 4 in yields similar to those for 3. Although the synthetic
sequences to 24 and 45 involved several steps, they are
operationally simple and generally rapid. Furthermore, the entire
route requires only a few chromatographic separations. Thus,
this synthetic strategy may provide practical access to a number
of substituted naphthazarins and tetralones such as 15, 3, and 4
(Schemes 1 and 3).
With quantities of naphthazarin 24 at hand, we proceeded to
investigate its dimerization via the proposed (Figure 4) radical-
based pathway to 1. Upon screening a number of reagents,
success was finally found in the use of cerium ammonium nitrate
(CAN) as an oxidant, which proved equal to the task of initiating
the crucial coupling reaction (Scheme 4). Thus, in a remarkably
(23) We thank Professor J. A. Elix, University of Canberra, for a generous gift
of natural 1.
9
J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004 611

A R T I C L E S
Nicolaou and Gray
Scheme 5. Oxidative Dimerization of Hydroxynaphthoquinones 45
and 48a
was fully characterized) which appeared to arise from de-
methylation of the aryl ethers. This is not surprising, as CAN
is known to cleave aryl methyl ethers at low temperature,
particularly when they are made more labile by a carbonyl group
located at the ortho position.19 This hypothesis is further
supported by the large percentage of starting material cleanly
recovered in the reaction of the aryl-unsubstituted naphtho-
quinone 48 (Scheme 5), which does not suffer from such side
reactions under the same reaction conditions.
Conclusion
With its novel and imposing structure, the naturally occurring
hybocarpone molecule provided a unique opportunity for
strategy and process development as well as analogue construc-
tion. Direct access to the hybocarpone skeleton was possible
by oxidative dimerization of the corresponding naphthazarin
monomer employing CAN as the oxidant. The success of this
process may be taken as support of the notion that nature is
also using this pathway in its biosynthesis of this bioactive
molecule, although no direct evidence for such a mechanism
exists at present. The synthetic technology developed en route
to 1 was also applied for the construction of two novel
hybocarpone analogues, namely, compounds 46 and 49, as a
demonstration of the potential power of this strategy to deliver
libraries of this natural product for biological screening purposes.
Among the novel synthetic objectives achieved in the described
synthesis are the selective benzylic acetoxylation reaction, the
first example of applying the photoenolization/Diels-Alder
reaction protocol to construct quaternary centers, and a general
method for the preparation of highly substituted tetralones and
hydroxynaphthoquinones. The highlight of this study, however,
must be the novel cascade sequence involving the oxidative
dimerization/hydration process of the hydroxynaphthoquinones
to the pentacyclic systems represented by 1. The exploration
of the generality and scope of the photoenolization/Diels-Alder
reaction and its application to the synthesis of the hamigerans
are described in the following paper in this issue.18
a Reagents and conditions: (a) 45 or 48 in MeCN; then add CAN (1.0
equiv) in degassed MeCN, -35 to 0 °C, sonication, 2 min; then 5 M aqueous
KOH, 0-25 °C, 10 min, 9% 46 plus 71% recovered starting material; 4%
49 plus 93% recovered starting material; (b) PtO2(cat.), H2 (1 atm), EtOAc,
4 h, filter catalyst, then bubble in air, 2 h, 83%.
evident. With the aureoquinone-derived dimer 46, we were able
to isolate two compounds in a ca. 1:1 ratio. The isomer we
1
assigned structure 46 (Scheme 5) exhibited a simple H NMR
spectrum reminiscient of that of 1. The other isomer exhibits
the correct mass, by mass spectrometry, but its 1H NMR
spectrum supports a nonsymmetrical structure. However, unlike
the hybocarpone isomer, it does not convert to the other isomer
(i.e., 46) under the conditions employed in the hybocarpone case
(25 f 21). Furthermore, this second isomer appears to be prone
to decomposition and has eluded full characterization.
The dimerization reaction itself is visually informative; a jet
black color is formed when CAN is added to the hydroxynaph-
thoquinone. This color fades over several minutes at -35 °C
or over 20 s at room temperature to a yellow hue very similar
to that of the starting material. Allowing the reaction to proceed
overnight did not increase the conversion. When the conditions
were modified such that the black color was not observed (i.e.,
running the reaction in 4:1 MeCN-H2O), no dimerization
occurred. A number of highly polar and colored compounds
were noticed as side products in these dimerization reactions
in the hybocarpone series (although no individual side product
Acknowledgment. This work was financially supported by
the National Institutes of Health, The Skaggs Institute for
Chemical Biology, a Louis R. Jabinson fellowship (to D.L.F.G.),
and grants from Amgen, Merck, and Pfizer.
Supporting Information Available: Experimental procedures
and compound characterization. This material is available free
of charge via the Internet at http://pubs.acs.org.
JA030497N
9
612 J. AM. CHEM. SOC. VOL. 126, NO. 2, 2004