Biomimetic Total Synthesis of Bisorbicillinol, Bisorbibutenolide, Trichodimerol, and Designed Analogues of the Bisorbicillinoids

Article abstract of DOI:10.1021/ja9942843

The bisorbicillinoids are a growing class of novel natural products endowed with unique biological activity and are associated with fascinating hypotheses for their biosynthesis. A full account of our biomimetic explorations toward the bisorbicillinoids including the total syntheses of bisorbicillinol (1), bisorbibutenolide (2), and trichodimerol (4) from sorbicillin (3) is disclosed. Utilizing the novel dimerization reactions discovered and fine-tuned en route to 1 and 4, several analogues of these natural products have been synthesized. Furthermore, studies on the scope of these novel cycloaddition reactions and the isolation of a number of unexpected products along with proposed mechanisms for their formation are reported. These findings add to our knowledge of the largely unexplored chemistry of o-quinols and related aromatic systems.

Full text of DOI:10.1021/ja9942843

J. Am. Chem. Soc. 2000, 122, 3071-3079
3071
Biomimetic Total Synthesis of Bisorbicillinol, Bisorbibutenolide,
Trichodimerol, and Designed Analogues of the Bisorbicillinoids
K. C. Nicolaou,*,† Georgios Vassilikogiannakis,† Klaus B. Simonsen,† Phil S. Baran,†
Yong-Li Zhong,† Veroniki P. Vidali,‡ Emmanuel N. Pitsinos,‡ and Elias A. Couladouros‡
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, Department of
Chemistry and Biochemistry, UniVersity of California, San Diego, 9500 Gilman DriVe, La Jolla, California
92093, and Organic and Bioorganic Chemistry Laboratories, NCRS Demokritos, 15310 Agia ParaskeVi,
Attikis, POB 60228, Athens, Greece
ReceiVed December 7, 1999
Abstract: The bisorbicillinoids are a growing class of novel natural products endowed with unique biological
activity and are associated with fascinating hypotheses for their biosynthesis. A full account of our biomimetic
explorations toward the bisorbicillinoids including the total syntheses of bisorbicillinol (1), bisorbibutenolide
(2), and trichodimerol (4) from sorbicillin (3) is disclosed. Utilizing the novel dimerization reactions discovered
and fine-tuned en route to 1 and 4, several analogues of these natural products have been synthesized.
Furthermore, studies on the scope of these novel cycloaddition reactions and the isolation of a number of
unexpected products along with proposed mechanisms for their formation are reported. These findings add to
our knowledge of the largely unexplored chemistry of o-quinols and related aromatic systems.
Introduction
addition, bisorbicillinol (1),2c bisorbibutenolide (2),5 bisorbicil-
linolide (5),5 and bisorbibetanone,6 all isolated from the
fermentation of Trichoderma sp. USF-2690, exhibit antioxidant
properties.
The dazzling mechanisms by which nature effortlessly knits
together molecules often inspires biomimetic syntheses that not
only lend credence to the proposed biosynthetic mechanisms
but considerably advance synthetic organic chemistry. The
bisorbicillinoids1 (Figure 1) provide an exciting forum for
explorations into such biosynthetic pathways and an opportunity
to discover new chemistry through total synthesis.
Not surprisingly, this diverse class of structurally novel natural
products, isolated from several species of fungi, is endowed
with a broad range of biological activity. For instance, trichod-
imerol (4), isolated from three different sources (Trichoderma
longibrachiatum,2a Penicillium chrysogenun,2b and Trichoderma
sp. USF-26902c), exhibits significant inhibitory activity against
lipopolysaccharide-induced production of tumor necrosis factor
R (TNF-R) in human monocytes and as such represents a new
lead for a potential treatment of septic shock.3 Bisvertinolone
7,4a an oxidized form of bisvertinol (6),4b isolated from
Acremonium strictum, is a novel antifungal agent which
functions via inhibition of â-(1,6)-glucan biosynthesis.4a In
Careful examination of these structurally related dodeca-
ketides reveals that they could all potentially originate from the
natural product sorbicillin (3, Figure 1).7 The first member of
this family, namely, 2′,3′-dihydrobisorbicillinol (1a), isolated
by Dreiding et al. in 1983 from Verticillium intertextum, was
postulated to arise from 3 and 2′,3′-dihydrosorbicillin after an
enantioselective oxidation followed by [4 + 2] dimerization.8
A similar mechanistic proposal was put forth by Abe et al. to
explain the biosynthesis of the newly isolated bisorbicillinol (1,
Figure 2).5 The same group also advanced an elegant proposal
for the biosynthesis of the soon thereafter isolated 2 and 5 via
an anionic cascade rearrangement of 1 (Figure 2).5 Sorbiquinol
(8, Figure 1), isolated by Ayer and co-workers in 1996 from T.
longibrachiatum, was postulated as the [4 + 2] adduct between
the side-chain double bond of sorbicillin (dienophile) and the
enantioselectively oxidized sorbicillin (diene).9
In 1999, we proposed a detailed mechanism for the formation
of trichodimerol (4) from sorbicillin (3) by an oxidation-
* E-mail: kcn@scripps.edu.
† The Scripps Research Institute and University of California, San Diego.
‡ NCRS Demokritos.
(5) (a) Abe, N.; Murata, T.; Hirota, A. Biosci. Biotechnol. Biochem. 1998,
62, 2120-2126. (b) A second fungal metabolite very similar to bisorbi-
butenolide (2) differing only by the relative configuration at C9 was isolated
in 1997 and named trichotetronine: Shirota, O.; Pathak, V.; Hossain, C.
F.; Sekita, S.; Takatori, K.; Satake, M. J. Chem. Soc., Perkin Trans. 1 1997,
2961-2964.
(1) The term bisorbicillinoids to describe all dimeric sorbicillin-derived
natural products was recently introduced by us: Nicolaou, K. C.; Jautelat,
R.; Vassilikogiannakis, G.; Baran, P. S.; Simonsen, K. B. Chem. Eur. J.
1999, 5, 3651-3665.
(2) (a) Andrade, R.; Ayer, W. A.; Mebe, P. P. Can. J. Chem. 1992, 70,
2526-2535. (b) Warr, G. A.; Veitch, J. A.; Walsh, A. W.; Hesler, G. A.;
Pirnik, D. M.; Leet, J. E.; Lin, P.-F. M.; Medina, I. A.; McBrien, K. D.;
Forenza, S.; Clark, J. M.; Lam, K. S. J. Antibiot. 1996, 49, 234-240. (c)
Abe, N.; Murata, T.; Hirota, A. Biosci., Biotechnol., Biochem. 1998, 62,
661-666.
(3) Mazzucco, C. E.; Warr, G. J. Leukocyte Biol. 1996, 60, 271-277.
(4) (a) Kontani, M.; Sakagami, Y.; Marumo, S. Tetrahedron Lett. 1994,
35, 2577-2580. (b) Trifonov, L. S.; Hilpert, H.; Floersheim, P.; Dreiding,
A. S.; Rast, D. M.; Skrivanova, R.; Hoesch, L. Tetrahedron 1986, 42, 3157-
3179.
(6) Abe, N.; Murata, T.; Yamamoto, K.; Hirota, A. Tetrahedron Lett.
1999, 40, 5203-5206.
(7) (a) Cram, D. J.; Tishler, M. J. Am. Chem. Soc. 1948, 70, 4238-
4239. (b) Cram, D. J. J. Am. Chem. Soc. 1948, 70, 4240-4243.
(8) (a) Trifonov, L. S.; Dreiding, A. S.; Hoesch, L.; Rast, D. M. HelV.
Chim. Acta 1981, 64, 1843-1846. (b) Trifonov, L. S.; Bieri, J. H.; Prewo,
R.; Dreiding, A. S.; Hoesch, L.; Rast, D. M. Tetrahedron 1983, 39, 4243-
4256.
(9) Andrade, R.; Ayer, W. A.; Trifonov, L. S. Can. J. Chem. 1996, 74,
371-379.
10.1021/ja9942843 CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/17/2000

3072 J. Am. Chem. Soc., Vol. 122, No. 13, 2000
Nicolaou et al.
Figure 1. Structures of selected bisorbicillinoids and their postulated biosynthetic precursor, sorbicillin (3).
Figure 2. Proposed biosynthetic pathways for bisorbicillinol (1),
bisorbibutenolide (2), and bisorbicillinolide (5).
Michael-ketalization cascade1 and 6 from 3 by a related
pathway (Figure 3).10,11 The first synthesis of enantiomerically
pure 4 from 3 by an oxidation-Michael-ketalization cascade
was reported by Barnes-Seeman and Corey.12
The unusual combination of impressive molecular complexity,
unique and diverse bioactivity, and intriguing biosynthetic
avenues inspired us to embark on a program focusing on the
Figure 3. Proposed biosynthetic pathways from sorbicillin (3) to
trichodimerol (1) and bisvertinol (6).
(10) Nicolaou, K. C.; Simonsen, K. B.; Vassilikogiannakis, G.; Baran,
P. S.; Vidali, V.; Pitsinos, E. N.; Couladouros, E. A. Angew. Chem., Int.
Ed. 1999, 38, 3555-3559.
(11) A previous proposal regarding the biosynthesis of 6 from sorbicillin
postulates a different pathway in which an initial epoxidation of sorbicillin
followed by reduction and then dimerization and ketalization is believed to
be involved: see ref 8.
biomimetic total synthesis of the bisorbicillinoids. Herein, we
present a full account of our biomimetic explorations into this
class of natural products which culminated not only in the total
synthesis of several members of this class, but also in the
(12) Barnes-Seeman, D.; Corey, E. J. Org. Lett. 1999, 1, 1503-1504.

Synthesis of Bisorbicillinoids
J. Am. Chem. Soc., Vol. 122, No. 13, 2000 3073
Scheme 1. Synthesis of Acetate 15 from
2,4-Dimethylresorcinol (12)a
Scheme 2. Biomimetic Total Synthesis of Bisorbicillinol (1)
and Bisorbibutenolide (2) from Acetate 15aa
aConditions: (a) RCOOH (1 equiv), BF3‚Et2O, 120 °C, 1 h; (b)
MeOH/H2O (1:1), 70 °C, 1 h; (c) THF/H2O (1:1), 70 °C, 1 h; (d)
Pb(OAc)4 (1.05 equiv), AcOH/CH2Cl2 (1:1), 0 °C, 30 min.
biomimetic total syntheses of a variety of bisorbicillinoid
analogues.10
Results and Discussion
Our journey to the bisorbicillinoids commenced with the
preparation of a suitably masked form of the expectedly fleeting
quinol 11a/b (Figures 2 and 3). To do this, we first required a
reliable route to gram quantities of sorbicillin (3). Thus the
procedure of McOmie et al.13,14 was satisfactory for producing
the boron complex 13 consistently; however, Michael addition
of methanol occurred frequently when rupturing the boron
complex (MeOH/H2O, reflux) affording compounds of type 17
(see Scheme 1). We reasoned that methanol was present in the
McOmie protocol solely for solubility purposes and, therefore,
employed THF as a cosolvent, thereby eliminating this side
reaction and providing practically pure 3 and analogues 14 after
workup.
After considerable experimentation, it was found that the
requisite R-acetoxy dienone 15a (R ) C5H7, Scheme 1) could
be procured as the major regioisomer together with its regio-
isomer 16a (15a:16a ∼5:1) and in 40% isolated yield upon
treatment with dry lead tetraacetate in degassed acetic acid
(Scheme 1). Several methods for the direct ortho oxidation of
phenols failed when applied to 3, presumably due to the extreme
reactivity of the o-quinol species. For instance, it was known
from the work of Barton et al. that the dimerization of free
o-quinols was instantaneous.15
a Conditions: (a) KOH (10 equiv), THF/H2O (9:1), 0 °C, 2 h; then
1 N HCl (aq), 40%; or (b) 12.1 N HCl, THF/H2O (9:1), 25 °C, 2 h,
43%; (c) KHMDS (1.1 equiv), THF, 25 °C, 24 h; then 1 N HCl (aq),
80%.
using enantiomerically pure 15a (S isomer). The optical rotation
of synthetic 1 ([R]D ) +171.5°, c ) 0.2 in MeOH) was in good
agreement with the literature ([R]D ) +195.2°, c ) 0.5 in
MeOH).2c This remarkable Diels-Alder reaction proceeds with
complete regio- and diastereocontrol (endo selectivity), generat-
ing four stereogenic centers, two of which are fully substituted.
It is notable that the acetate 15a did not give any Diels-Alder
product when heated in benzene or in AcOH for several hours.
As supported by NMR studies (vide infra), the initial stages
of this reaction involve deacetylation, thus revealing a diquino-
late system (bis-deprotonated forms of 11a/b) which rapidly
scrambles to a mixture of diquinolates. Acidification of the
reaction mixture undoubtedly leads to the formation of the
fleeting quinols 11a/b, which rapidly unite in a Diels-Alder
reaction to generate bisorbicillinol (1; see Scheme 2).
Since it is impossible to monitor this transformation by thin-
layer chromatography (TLC), we performed the reaction in
deuterated media (THF-d8/D2O 9:1) and observed the fate of
acetate 15a directly by H NMR spectroscopy. Thus, upon
addition of 10 equiv of KOH, the solution became characteristi-
With acetate 15a in hand, the stage was set to explore the
feasibility of a biomimetic approach to the bisorbicillinoids. To
our delight, treatment of a 0.05 M solution of 15a in THF/H2O
(9:1) with 10 equiv of solid KOH, followed by quenching with
1 N aqueous HCl led to the Diels-Alder adduct, bisorbicillinol
(1) in 40% yield (Scheme 2). The spectral properties of synthetic
1 were identical in all respects to those reported by Abe et al.2c,16
Using a chiral HPLC column, we were able to repeat the reaction
1
cally deep orange and the resonances for acetate 15a dis-
1
appeared. As predicted, the H NMR spectrum indicated the
sole presence of two discrete quinolates (∼3:2 ratio). The
appearance of 1 occurred only after acidification of the reaction
mixture (concentrated HCl). This observation led us to deduce
that the free quinols would be necessary for most dimerization
reactions of this type to occur rather than the corresponding
quinolates. Therefore, a judicious choice of conditions for
neutralization would likely have a profound impact on the nature
(13) Baker, W.; Bondy, H. F.; McOmie, J. F. W.; Tunnicliff, H. R. J.
Chem. Soc. 1949, 2834-2835.
(14) McOmie, J. F. W.; Tute, M. S. J. Chem. Soc. 1958, 3226-3227.
This method was found to be superior on larger scale to the procedure
described by Sartori et al.: Bigi, F.; Casiraghi, G.; Casnati, G.; Marchesi,
S.; Sartori, G.; Vignali, C. Tetrahedron 1984, 40, 4081-4084.
(15) (a) Barton, D. H. R.; Magnus, P. D.; Rosenfeld, M. N. J. Chem.
Soc., Chem. Commun. 1975, 301-301. (b) Barton, D. H. R.; Magnus, P.
D.; Quinney, J. C. J. Chem. Soc., Perkin Trans. 1 1975, 1610-1614. See,
also: Adler, E.; Holmberg, K. Acta Chem. Scan. 1974, B28, 465-472.
(16) We thank Dr. N. Abe of the University of Shizouka for kindly
1
providing us with the H and 13C NMR spectra of 1 and 2.

3074 J. Am. Chem. Soc., Vol. 122, No. 13, 2000
Nicolaou et al.
of the product formation (vide infra). As a corollary to this
observation, it should then be plausible to directly produce 1
via acetate 15a through simple acidic hydrolysis of the acetoxy
function present in 15a giving rise to the transitory quinols 11a/
b, and finally 1 itself. Indeed, treatment of 15a with concentrated
HCl in THF led to 1 in 43% isolated yield (Scheme 2). The
ephemeral nature of quinols 11a/b was once again reflected by
Scheme 3. Synthesis of Sorbicillin Dimer 21 from Acetate
15a and Its Transformation to the Corresponding
Hydrogenated and Acetylated Dimer 22a
1
an H NMR experiment (THF-d8/D2O, concentrated HCl) that
gradually showed the complete consumption of 15a accompa-
nied by proportionate formation of 1 over a 2-h period.
With 1 in hand, we were enticed to explore the hypothesis
put forth by Abe et al. for the biosynthesis of bisorbibutenolide
(2) and bisorbicillinolide (5) from 1 (see Figure 2 and Scheme
2).5 The novel anionic rearrangement that converts 1 to 2
requires deprotonation of the tertiary alcohol leading to 18,
which must then engage the carbonyl group leading to rear-
ranged anion 19 or proceed further to 5 (Figure 2 and Scheme
2). We reasoned that a potassium counterion might best function
in this sequence due to its size and oxophilicity. In the event,
subjection of 1 to 1.1 equiV of KHMDS in THF led to exclusive
formation of 2 in 80% isolated yield. It is notable that 1, which
contains four hydroxyl groups, required only 1 equiv of KHMDS
to complete the transformation (1 f 2). Since 2 is an
intermediate along the presumed anionic cascade pathway to 5
(Figure 2), several conditions commencing from either 1 or 2
to access 5 were evaluated. So far, however, while treatment
of 1 with KHMDS led smoothly to 2, NaHMDS, LiHMDS,
and several other bases failed to convert 1 or 2 to 5.
aConditions: (a) KOH (10 equiv), THF/H2O (10:1), 0 °C, 2 h; then
(b) 1 N HCl (aq) 65% (crude yield); (c) H2, 10%, Pd/C, EtOAc, 25
°C, 2 h, 81%; (d) Ac2O (10 equiv), 4-DMAP (0.2 equiv), EtOAc, 25
°C, 2 h, 80%. NOEs were observed for Mea/Ha (4.5%), Ha/Hb (3.9%),
and Hb/Meb (3.0%) for 22.
Scheme 4. Transformation of Acetate 15a to the Cyclized
Michael Product 24a
After completing syntheses of 1 and 2, we set course towards
other structurally challenging bisorbicillinoids. We reasoned that
an increase in the concentration of the reaction that furnished 1
might also furnish 4 or 6 after acidification (Figure 3). Treatment
of acetate 15a with KOH in a minimum amount of solvent
(THF/H2O 10:1, 0.3 M) led to a major product along with a
small amount of 1. Although it appeared by TLC that the
reaction had led to this new compound in high yield (∼65%
crude), it could only be isolated in a pure form by flash
chromatography, resulting in a much lower isolated yield. The
physical properties of the new compound were suggestive of
the dimeric structure 21 (Scheme 3). Additional proof of
structure was obtained after hydrogenation of the side chains
(H2, 10% Pd/C) and acetylation of the resulting compound
(Ac2O, 4-DMAP) to afford the stable tetraacetate 22, whose
aConditions: KHMDS (3.0 equiv), THF, -78 °C, 2 h, 60%; NOEs
were observed for Mea/Ha (1.8%).
excess KHMDS (3.0 equiv) in anhydrous THF at -78 °C
produced first a red solution and, subsequently, upon acidifica-
tion, the yellow γ-lactone 24 (60% yield). The use of LiHMDS
led only to an orange lithium enolate with complete recovery
of starting material upon acidification. Interestingly, treatment
of 15a with various Lewis acids (e.g., TBSOTf, TMSOTf) led
only to the rearomatization and formation of 3 along with traces
of 1.
1
1
structure was secured using H, 13C, H-1H COSY, 13C APT,
HMQC, and HMBC NMR spectroscopic techniques. Particularly
striking is the fact that dimer 21 is formed as a single
diastereomer, the configuration of which was defined in a 1D
NOE experiment for the following protons: Mea/Ha (4.5%); Ha/
Hb (3.9%); Hb/Meb (3.0%). A mechanism that accounts for the
selective formation of the novel dimer 21 is illustrated in Scheme
3. Thus, it is reasoned that the initially formed dianion (20,
Scheme 3) rearranges by an intramolecular Michael reaction to
an epoxy dianion (not shown), which then engages another
dianion in an intermolecular Michael reaction to stereoselec-
tively generate a quaternary center. Since this is an inter-
molecular reaction between quinolates, a high concentration is
necessary. We attribute the observed diastereospecificity (only
one of four possible diastereomers) to a chelating effect which
governs the attacking face of the incoming quinolate so as to
minimize steric and charged interactions (Scheme 3).
We then realized, as alluded to above, that the key to
generating additional bisorbicillinoids resided in the workup
protocol employed to quench the diquinolate (e.g., 20, Scheme
3) derived from 15a. We also feared that the presence of excess
water (cosolvent) in the reaction would hinder the formation of
certain bisorbicillinoids such as trichodimerol (4) due to
expected ketalization difficulties. Extensive experimentation led
us to identify the use of CsOH‚H2O/MeOH as a reliable method
for the generation of the desired quinolate under protic condi-
tions, but with only stoichiometric quantities of H2O present.
Varying the workup protocol of this reaction was a pivotal step
in the biomimetic total synthesis of 4. Thus, treatment of 15a
with CsOH‚H2O in MeOH for 7 h followed by neutralization
with finely powdered NaH2PO4‚H2O and stirring at 25 °C for
Another group of experiments attempting to convert acetate
15a into other bisorbicillinoids led only to decomposition,
recovery of starting material, or isolation of certain informative
byproducts. For example (Scheme 4), treatment of 15a with

Synthesis of Bisorbicillinoids
J. Am. Chem. Soc., Vol. 122, No. 13, 2000 3075
Scheme 5. Novel Dimerization of the Acetate 15a. Biomimetic Total Synthesis of Trichodimerol (4) and Bisorbicillinol (1)a
aConditions: (a) CsOH‚H2O (10 equiv), MeOH, 25 °C, 7 h; then (b) NaH2PO4‚H2O, 25 °C, 12 h, 16% trichodimerol (4), 22% bisorbicillinol (1).
Table 1. [4 + 2] and [4 + 4] Dimerization of Acetates 15a-g to
Bisorbicillinol (1) and Trichodimerol (4) Analogues
Scheme 6. Synthesis of Acetate 27 from
2,4-Dimethylresorcinol (12) and Failed Attempts To Produce
Dimerization Productsa
aConditions: (a) Zn(CN)2 (2.0 equiv), HCl (g), Et2O, reflux, 1 h,
92%; (b) NaClO2 (2.5 equiv), NaH2PO4 (2.5 equiv), DMSO/H2O
(1:1), 25 °C, 24 h, 69%; (c) CH2N2 (excess), C6H6/MeOH (1:2), 0 °C,
10 min, 78%; (d) Pb(OAc)4 (1.05 equiv), AcOH/CH2Cl2 (1:1), 0 °C,
30 min, 55%.
12 h furnished 4, which was isolated in 16% yield after column
chromatography together with 3 (12%) and 1 (22%, Scheme
5).
Synthetic 4 exhibited spectroscopic properties identical to that
reported by Ayer et al., although 4 was unavailable to us for
direct comparison.2a The very slow neutralization of this reaction
is critical for the extraordinary biomimetic dimerization of 15a
to give 4 having no less than eight stereogenic centers, six of
which are quaternary. The reaction could also be carried out
with enantiomerically pure 15a (vide supra) to provide optically
active 4 ([R]D ) -375°, c ) 0.1 in MeOH for synthetic 4 and
[R]D ) -376°, c ) 0.26 in MeOH for naturally occurring 4)2a
as first reported by Barnes-Seeman and Corey while these
studies were still in progress.12
We immediately recognized the exciting potential of this
reaction to furnish designed analogues of 1 and 4. Therefore,
we targeted the acetate 27 as a precursor to analogues of 1 and
4, which would contain a versatile carbomethoxy group poised
for further manipulations. Scheme 6 illustrates our successful
synthesis of acetate 27. Thus, 2,4-dimethylresorcinol (12) was
formylated using the procedure of McOmie et al.13 to afford
aromatic aldehyde 25 in 92% yield. Conversion of aldehyde
25 to the corresponding carboxylic acid using NaClO2/NaH2-
PO4 required the use of a DMSO cosolvent. After treatment of
the resulting carboxylic acid with CH2N2 to provide methyl ester
26, oxidation with Pb(OAc)4 proceeded smoothly to afford
acetate 27 in 55% yield. Despite our repeated attempts, however,
we were unable to coax 27 into a biomimetic dimerization
reaction. We rationalized that the reduced ability of the ester
grouping to act as a Michael acceptor precluded dimerization
b
aIsolated yields. Mixture of three geometrical isomers.
and so we set our sights on modification of the side chain at
the sorbicillin stage.
To probe the effect of differing side chains on the dimeriza-
tion reaction, we synthesized (according to Scheme 1) an array
of sorbicillin mimics which preserved the crucial ketone moiety.
Table 1 illustrates the remarkable tolerance of the dimerization
reaction to changes in the side chain size and electronic
disposition. Conjugation throughout the side chain was found
to be unnecessary as acetates 15d-f led cleanly to the
corresponding [4 + 4] products (4d-f, Table 1), and in the
case of 15e-g provided [4 + 2] products (1e-g, Table 1) as
well. At this time, we are unable to contribute an explanation
for the distribution of products (i.e., [4 + 4] vis-a`-vis [4 + 2])
in this stunning biomimetic cascade dimerization.
Although acetates 15b,c succeeded in furnishing trichodimerol
mimics 4b,c, competing intra- and intermolecular Michael
reactions frequently led to spurious byproducts. For example,
in the case of acetate 15h (Scheme 7), the dimerization reaction
led exclusively to 29 rather than the desired [4 + 4] or [4 + 2]
products. Scheme 7 depicts the unfavorable steric interaction
in 15h which is likely responsible for the efficient formation
of 29 via an intramolecular Michael reaction. Utilizing the pure
E acetate 15b (Table 1), we observed formation of the desired
[4 + 4] product as a mixture of three geometrical isomers

3076 J. Am. Chem. Soc., Vol. 122, No. 13, 2000
Nicolaou et al.
Scheme 9. Abortive at Attempts To Prepare Acetates 36
Scheme 7. Formation of the Intermolecular Michael Adduct
29 from Acetate 15ha
and 37 from Resorcinols 34 and 35
aConditions: CsOH‚H2O (10 equiv), MeOH, 25 °C, 7 h; then
NaH2PO4‚H2O, 25 °C, 12 h, 70%.
Scheme 8. Isolation of the Stable Quinol 30 and Its
Thermal Rearrangement to Quinone 33a
Conclusion
Our explorations into the chemistry of the bisorbicillinoids
have culminated in the total synthesis of several members of
this class, designed analogues, and a wealth of new and
interesting chemistry. In addition, several proposals regarding
the biosynthesis of these compounds have been bolstered by
experimental proof. The biomimetic synthesis of other bisor-
bicillinoids, development of a solid-phase variant of the key
dimerization cascade, and biological evaluation of the novel
bisorbicillinoid mimics reported herein are some of the directions
pointed to by these findings.
Experimental Section
aConditions: (a) CsOH‚H2O (10 equiv), MeOH, 25 °C, 7 h; then
NaH2PO4‚H2O, 25 °C, 12 h, 75%; (b) benzene, reflux, 12 h, 93%; (c)
air, CDCl3, 25 °C, 95%.
General Procedure for the Boron Trifluoride-Catalyzed Acyla-
tion of 2,4-Dimethylresorcinol with Carboxylic Acids. An equimolar
solution of 2,4-dimethylresorcinol13 (12, 500 mg, 3.7 mmol) and the
appropriate carboxylic acid (3.7 mmol) in BF2‚Et2O (5 mL, 0.74 M)
was heated for 1-2 h at 120 °C under an argon atmosphere. The
complete formation of the polar, yellow, stable BF -complex (Scheme
1) was observed by TLC. After cooling, the reaction mixture was
quenched at 0 °C with H2O (5 mL), followed by extraction with EtOAc
(3 × 25 mL) and concentration under vacuum. The resulting material
was refluxed for 30 min to 1 h in a THF/H2O (1:1, 50 mL, 0.07 M),
and the hydrolyzed compound was extracted with EtOAc (3 × 30 mL),
dried (MgSO4), and purified by flash column chromatography using a
mixture of hexane/EtOAc (4:1) as eluent. The yield of the reaction
ranges from 70 to 80%. When the hydrolysis of BF2 complexes 13b
and 13h were carried out in MeOH/H2O (1:1), substantial amounts of
the MeOH Michael adducts 17b and 17h were isolated (Scheme 1).
The spectroscopic data for the desired acylated products 14b-i are
given as Supporting Information.
(because of the symmetry involved, only three isomers are
theoretically possible) apparently due to isomerization processes
of the final [4 + 4] product under the basic reaction conditions.
In several instances in the unsaturated series (not shown), we
also observed conjugate addition of MeOH to the side chain,
leading to a multicomponent mixtures of products.
Not surprisingly, the tert-butyl acetate 15i failed to dimerize
due to extreme steric shielding (Scheme 8). However, the
reaction led cleanly to the free quinol 30, which was stable at
room temperature. The formation of this stable quinol 30 gave
us the rare opportunity to investigate the reactivity of this novel
species when unable to dimerize. Thus, upon heating a solution
of quinol 30 in benzene, we observed a quantitative conversion
to the rearranged trihydroxybenzene 32. A proposed mechanism
for this unique transformation is given in Scheme 8. Further-
more, upon standing in CDCl3 for several hours, a quantitative
conversion of 32 to quinone 33 was observed.
Attempts to modify the central core of 1 and 4 by altering
the presence of methyl groups in the acetate 15a did not succeed.
Although we were able to prepare the demethylsorbicillin
compounds 34 and 35 by a route analogous to that depicted in
Scheme 1, the ensuing oxidation with Pb(OAc)4 resulted only
in complete decomposition (Scheme 9). This difficulty is
attributed to the ease with which re-aromatization must occur
in the expected products along with other competing oxidatively
destructive pathways. The isolation of demethyltrichodimerol
(4a, Figure 1) from natural sources implies that an oxidized
form of demethylsorbicillin 34 should be accessible, and thus,
we are continuing to investigate this transformation (34 f 36,
Scheme 9).
2-Demethylsorbicillin (34). The above-described experimental
procedure was followed in this case, using 4-methylresorcinol instead
of 2,4-dimethylresorcinol: orange semicrystalline solid; Rf 0.60 (silica
gel, hexane/EtOAc 1:1); IR (film) νmax (cm-1) 3441 br w (OH), 1635
1
s (CdO), 1559 m, 1361 m, 1269 m, 1139 m, 998 m; H NMR (600
MHz, CDCl3) δ 13.39 (s, OH), 7.51 (s, 1 H), 7.45 (dd, J1 ) 14.2 Hz,
J2 ) 10.9 Hz, 1 H), 6.94 (s, OH), 6.91 (d, J ) 15.4 Hz, 1 H), 6.38 (s,
1 H), 6.33 (br d, J ) 10.9 Hz, 1 H), 6.28 (m, 1 H), 2.18 (s, 3 H), 1.90
(d, J ) 6.6 Hz, 3 H); 13C NMR (150 MHz, CDCl3) δ 192.4, 164.4,
161.6, 144.9, 141.6, 131.9, 130.5, 121.5, 116.3, 113.8, 103.8, 18.9,
15.3; HRMS (MALDI) calcd for C13H14O3 [M + H+] 219.1021, found
219.1020.
6-Demethylsorbicillin (35). The above-described experimental
procedure was followed in this case, using 2-methylresorcinol instead
of 2,4-dimethylresorcinol: yellow crystals; Rf 0.60 (silica gel, hexane/
EtOAc 1:1); mp 148-149 °C (CHCl3); IR (film) νmax (cm-1) 3171 br
w (OH), 1615 s (CdO), 1564 m, 1487 m, 1366 m, 1266 s, 1219 m,

Synthesis of Bisorbicillinoids
J. Am. Chem. Soc., Vol. 122, No. 13, 2000 3077
1110 m, 988 w, 780 w; 1H NMR (600 MHz, CDCl3) δ 13.35 (s, OH),
7.58 (d, J ) 8.8 Hz, 1 H), 7.47 (dd, J1 ) 14.7 Hz, J2 ) 10.6 Hz, 1 H),
6.92 (d, J ) 14.8 Hz, 1 H), 6.37 (d, J ) 8.8 Hz, 1 H), 6.29 (m, 2 H),
5.37 (br s, OH), 2.14 (s, 3 H), 1.90 (d, J ) 6.2 Hz, 3 H); 13C NMR
(125 MHz, CDCl3) δ 192.7, 164.3, 160.2, 144.9, 141.6, 130.5, 128.7,
121.6, 113.9, 111.4, 106.9, 19.0, 7.8; HRMS (MALDI) calcd for
C13H14O3 [M + H+] 219.1021, found 219.1018.
1 H), 6.40 (d, J ) 15.0 Hz, 1 H), 6.00 (br t, J ) 13.0 Hz, 1 H), 5.81
(br t, J ) 11.8 Hz, 1 H), 5.72 (dq, J1 ) 14.8 Hz, J2 ) 6.7 Hz, 1 H),
5.64 (dq, J1 ) 14.1 Hz, J2 ) 7.0 Hz, 1 H), 3.94 (br s, 1 H), 3.77 (br
s, 1 H), 2.01 (s, 3 H), 1.79 (s, 3 H), 1.47 (s, 3 H), 1.45 (d, J ) 7.5 Hz,
3 H), 1.38 (d, J ) 6.7 Hz, 3 H), 1.30 (s, 3 H); 13C NMR (150 MHz,
C6D6/CD3OD 9:1) δ 208.5, 198.9, 197.2, 175.1, 170.4, 168.8, 146.2,
142.9, 142.3, 139.0, 131.5, 130.7, 124.9, 119.4, 111.5, 109.6, 74.7,
70.3, 68.7, 60.6, 48.4, 42.1, 33.0, 24.7, 18.6, 18.4, 10.5, 8.6; HRMS
(MALDI) calcd for C28H32O8 [M + H+] 497.2175, found 497.2172.
Repeating the reaction with enantiomerically pure (S)-15a gave (+)-
(1); [R]D ) +171.5° (c ) 0.2, MeOH). Reported [R]D ) +195.2°
(c ) 0.5, MeOH).2c
Formation of Dimer 21. When the above hydrolysis of acetate 15a
was carried out in a very concentrated solution of THF/H2O (10:1, 0.3
M) a more polar dimer 21 was formed, together with a small amount
of the previously obtained 1. Dimer 21 was found to be labile on silica
gel and could be isolated in pure form only in 10% yield, by fast flash
column chromatography (silica gel, CH2Cl2/acetone 4:1): yellow
amorphous powder; Rf 0.20 (silica gel, CH2Cl2/acetone 4:1); IR (film)
νmax (cm-1) 3371 br w (OH), 1619 s, 1385 m, 1345 s, 1210 m, 1076
m; 1H NMR (500 MHz, C6D6/CD3OD 9:1) δ 7.38 (m, 2 H), 6.77 (d, J
) 14.6 Hz, 1 H), 6.71 (d, J ) 14.6 Hz, 1 H), 6.32 (dd, J1 ) 14.3 Hz,
J2 ) 12.1 Hz, 1 H), 5.78 (dd, J1 ) 14.0 Hz, J2 ) 12.5 Hz, 1 H), 5.65
(m, 2 H), 4.94 (br s, 1 H), 4.50 (br s, 1 H), 2.03 (s, 3 H), 1.99 (s, 3 H),
1.46 (d, J ) 5.9 Hz, 3 H), 1.38 (s, 3 H), 1.35 (d, J ) 6.9 Hz, 3 H),
1.33 (s, 3 H); 13C NMR (125 MHz, C6D6/CD3OD 9:1) δ 197.9, 192.9,
170.8, 169.6 (br, 3 C), 146.4, 146.1, 142.9, 139.6, 139.4, 136.7, 132.3,
131.4, 125.7, 122.1, 120.6, 88.7, 82.0, 73.1, 59.3, 42.5, 26.0, 25.5, 18.9,
18.8, 8.7, 8.4; HRMS (MALDI) calcd for C28H32O8Na [M + Na+]
519.1995, found 519.2003.
General Procedure for the Lead Tetraacetate Oxidation of 2,4-
Dimethylresorcinols to the Corresponding R-Acetoxydienones. A
solution of the appropriate acylated 2,4-dimethylresorcinol 14 (1.0
mmol) in AcOH/CH2Cl2 (1:1, 25 mL, 0.04 M) was stirred for 20 min
with argon bubbling through the solution, and subsequently, lead
tetraacetate (470 mg, 1.05 mmol, 1.05 eq) was added in one portion at
0 °C. After 30 min stirring at 0 °C, the reaction mixture was quenched
with H2O (25 mL), extracted with EtOAc (50 mL), washed with H2O
(4 × 50 mL), and dried (MgSO4). The solvent was removed in vacuo,
and the remaining traces of AcOH were pumped off under high vacuum.
In most cases, the major byproduct was the undesired, less polar
regioisomer (Rf 0.55-0.60, silica gel, hexane/EtOAc 1:1) (Scheme 1).
The desired, more polar regioisomer (Rf 0.30-0.35, silica gel, hexane/
EtOAc 1:1) was isolated by flash column chromatography using a
mixture of hexane/EtOAc (1:1) as eluent. The spectroscopic data of
the desired regioisomer 15a and the undesired 16a are given below,
while the data for 15b-i are given as Supporting Information.
2,6-Dimethyl-6-acetoxy-4-(2,4-hexadienoyl)-3-hydroxy-2,4-cyclo-
hexadien-1-one (15a, desired acetate). Yellow crystals; Rf 0.35 (silica
gel, hexane/EtOAc 1:1); mp 149-150 °C (ether/hexane 1:1); IR (film)
ν
max (cm-1) 2930 br w, 1737 m (CdO, acetate), 1650 s (CdO, enone),
1
1555 s, 1371 m, 1270 m, 1245 s, 1070 m, 1018 m, 768 m; H NMR
(500 MHz, CDCl3) δ 11.90 (s, OH), 7.46 (dd, J1 ) 14.8 Hz, J2 ) 10.8
Hz, 1 H), 7.25 (s, 1 H), 6.66 (d, J ) 14.8 Hz, 1 H), 6.38 (m, 1 H), 6.31
(m, 1 H), 2.15 (s, 3 H), 1.93 (d, J ) 6.6 Hz, 3 H), 1.86 (s, 3 H), 1.49
(s, 3 H); 13C NMR (125 MHz, CDCl3) δ 195.4, 193.7, 170.4, 162.9,
152.3, 148.7, 145.3, 130.5, 125.9, 120.6, 112.1, 78.6, 24.5, 21.0, 19.6,
7.6; HRMS (MALDI) calcd for C16H18O5Na [M + Na+] 313.1052,
found 313.1055. Separation of the two enantiomers of acetate 15a was
accomplished by semipreparative HPLC (Daicel Chiracel AD, hexane/
i-PrOH/MeOH 80:15:5 containing 0.1% TFA, flow rate 5.5 mL/min,
tr ) 15.3 min (S isomer) and tr ) 18.1 min (R isomer); (S)-15a: [R]D
) -580° (c ) 0.1, MeOH). Reported [R]D ) -606° (c ) 0.9, MeOH).12
(R)-15a: [R]D ) +595 ° (c ) 0.2, MeOH). Reported [R]D ) +615° (c
) 1.0, MeOH).12
Hydrogenated Tetraacetate 22. The crude reaction mixture from
the above reaction (85 mg, 0.17 mmol) was dissolved in EtOAc (5
mL) and 10% Pd/C (15 mol %) was added: the yellow reaction mixture
was stirred under H2 until the solution turned colorless (2 h). The
solution was filtered through Celite and concentrated. Flash column
chromatography (silica gel, CH2Cl2/acetone 4:1) gave 70 mg (81%) of
the hydrogenated dimer as a clear oil, which was converted directly to
the tetraacetate 22. To a solution of the hydrogenated dimer (26 mg,
0.05 mmol) in EtOAc (2 mL, 0.025 M) were added Ac2O (50 µL, 0.5
mmol, 10 equiv) and 4-DMAP (1 mg, 0.01 mmol, 0.2 equiv), and the
reaction mixture was stirred for 2 h, concentrated, and purified by flash
column chromatography (silica gel, hexane:EtOAc 2:1) to give tet-
raacetate 22 (28 mg, 80%) as a stable clear oil: Rf 0.25 (silica gel,
hexane/EtOAc 2:1); IR (film) νmax (cm-1) 1769 s (CdO, acetate), 1710
s (CdO, enone), 1671 s (CdO, enone), 1370 m, 1242 m, 1185 m,
2,6-Dimethyl-6-acetoxy-4-(2,4-hexadienoyl)-5-hydroxy-2,4-cyclo-
hexadien-1-one (16a, undesired acetate). Yellow oil; Rf 0.60 (silica
gel, hexane/EtOAc 1:1); IR (film) νmax (cm-1) 1738 m (CdO, acetate),
1681 m (CdO, enone), 1606 s (CdO, enone), 1537 m, 1372 m, 1247
1
1071 m; H NMR (500 MHz, C6D6) δ 5.90 (s, 1 H), 4.08 (s, 1 H),
1
3.51 (ddd, J1 ) 13.8 Hz, J2 ) 10.0 Hz, J3 ) 5.5 Hz, 1 H), 2.87 (dt, J1
) 19.3 Hz, J2 ) 7.3 Hz, 1 H), 2.63 (ddd, J1 ) 13.7 Hz, J2 ) 10.1 Hz,
J3 ) 5.5 Hz, 1 H), 2.19 (dt, J1 ) 19.3 Hz, J2 ) 7.3 Hz, 1 H), 1.96 (s,
3 H), 1.88 (m, 1 H), 1.86 (s, 3 H), 1.78 (s, 3 H), 1.73 (s, 3 H), 1.72 (s,
3 H), 1.69 (s, 3 H), 1.65 (m, 1 H), 1.59 (s, 3 H), 1.55 (m, 2 H), 1.50
(s, 3 H), 1.38 (m, 2 H), 1.30 (m, 2 H), 1.20 (m, 2 H), 1.11 (m, 2 H),
0.88 (t, J ) 7.4 Hz, 3 H), 0.81 (t, J ) 7.3 Hz, 3 H); 13C NMR (125
MHz, C6D6) δ 205.2, 194.5, 184.9, 168.6, 167.7, 166.2, 165.4, 163.4,
157.8, 157.7, 127.3, 127.0, 119.7, 82.7, 79.6, 79.2, 71.5, 59.7, 44.0,
33.0, 32.0, 31.2, 27.7, 25.4, 23.4, 23.0, 22.9 (2 CH2), 21.6, 20.7, 20.0,
19.9, 14.3, 14.1, 10.7, 10.6; HRMS (MALDI) calcd for C36H48O12Na
[M + Na+] 695.3038, found 695.3013. Assignment of 1H and 13C
m, 997 w, 732 w; H NMR (600 MHz, CDCl3) δ 11.75 (s, OH), 7.39
(dd, J1 ) 14.8 Hz, J2 ) 10.7 Hz, 1 H), 7.35 (d, allylic coupling J )
1.0 Hz, 1 H), 6.44 (d, J ) 14.8 Hz, 1 H), 6.33 (dd, J1 ) 15.0 Hz, J2
) 10.5 Hz, 1 H), 6.27 (m, 1 H), 2.16 (s, 3 H), 1.95 (d, allylic coupling,
J ) 1.0 Hz, 3 H), 1.91 (d, J ) 6.6 Hz, 3 H), 1.52 (s, 3 H); 13C NMR
(150 MHz, CDCl3) δ 200.3, 194.4, 173.5, 169.9, 144.0, 141.7, 136.1,
130.7, 124.6, 117.1, 105.4, 82.8, 23.3, 20.0, 19.0, 15.9; ESIMS
(C16H18O5): m/z (%) negative 289 ([M - H], 75).
Dimerization of Acetate 15a under Basic (KOH) Conditions.
Bisorbicillinol (1). To a solution of acetate 15a (29 mg, 0.1 mmol) in
THF/H2O (9:1, 2 mL, 0.05 M) was added KOH (56 mg, 1.0 mmol, 10
equiv) dissolved in a minimun amount of H2O (one drop) at 0 °C. The
reaction mixture turned orange immediately, indicating the formation
of the corresponding diquinolate. After 1.5 h stirring at 0 °C, the reaction
mixture was treated with 1 N HCl (2 mL) followed by extraction with
CH2Cl2 (3 × 5 mL). The combined organic phases were washed with
H2O (2 × 10 mL) and brine (10 mL) and dried (MgSO4). Removal of
the solvent followed by flash column chromatography (silica gel,
CH2Cl2/acetone 5:1) gave the Diels-Alder adduct 1 (10 mg, 40%) as
a pale yellow amorphous powder. The spectral properties of synthetic
1 were identical to those reported by Abe et al.:2c Rf 0.30 (silica gel,
CH2Cl2/acetone 4:1); IR (film) νmax (cm-1) 3415 br w (OH), 1739 s
1
signals were aided by H-1H COSY, 13C APT, HMQC and HMBC
experiments. NOE experiments (C6D6): irradiation of signal at δ 5.90
(Hb, Scheme 3) effects signals at δ 4.08 (3.9%, Ha) and 1.96 (3.0%,
Meb); irradiation of signal at δ 4.08 (Ha, Scheme 3) effects signals at
δ 5.90 (3.8%, Hb) and 1.55 (4.5%, Mea).
[4 + 2] Dimerization of Acetate 15a Under Acidic (HCl)
Conditions. To a solution of acetate 15a (29 mg, 0.1 mmol) in THF
(2 mL, 0.05 M) was added concentrated HCl (12.1 N, 206 µL, 2.5
mmol, 25 equiv) followed by stirring for 2 h at 25 °C. The reaction
mixture was diluted with EtOAc (5 mL), washed with H2O (5 mL)
and brine (5 mL), and dried with MgSO4. Evaporation of the solvent
followed by flash column chromatography (silica gel, CH2Cl2/acetone
5:1) furnished 1 (10.5 mg, 43%). The spectroscopic data of synthetic
1
(CdO), 1632 m, 1390 m, 1244 m, 1210 m, 1012 m; H NMR (600
MHz, C6D6/CD3OD 9:1) δ 7.37 (dd, J1 ) 15.0 Hz, J2 ) 11.0 Hz, 1
H), 7.34 (dd, J1 ) 14.9 Hz, J2 ) 11.0 Hz, 1 H), 6.74 (d, J ) 14.8 Hz,

3078 J. Am. Chem. Soc., Vol. 122, No. 13, 2000
Nicolaou et al.
1 were identical to those given above and in accordance to those
reported by Abe et al.2c
10.5 mmol, 2.5 eq) in H2O (3 mL). The slurry was stirred vigorously
at ambient temperature for 24 h, at which time it was diluted with
EtOAc (25 mL) and 1 N HCl (10 mL) was added. Extraction with
EtOAc (3 × 25 mL) followed by drying with MgSO4 and removal of
the solvent under vacumn gave the crude carboxylic acid, which was
directly dissolved in benzene/MeOH (1:2, 10 mL) and treated with
excess CH2N2 (etheral solution) at 0 °C. After 10 min, the solvent was
removed and the crude methyl ester was flash chromatographed (silica
gel, hexane/EtOAc 5:1) to give pure 26 (440 mg, 54%) as off-white
crystals: Rf 0.30 (silica gel, hexane/EtOAc 5:1); mp 136-138 °C
(CHCl3); IR (film) νmax (cm-1) 3435 br w (OH), 1646 (CdO, methyl
ester) 1606 m, 1441 m, 1294 m, 1206 s, 1146 m, 1109 m, 1019 m,
Anionic Rearrangement of Bisorbicillinol (1) to Bisorbibutenolide
(2). To a 0 °C cold solution of 1 (10 mg, 0.02 mmol) in dry THF (0.5
mL, 0.04 M) was added KHMDS (0.5 M solution in toluene, 44 µL,
0.022 mmol, 1.1 equiv), and the mixture turned orange. The reaction
was allowed to warm to room temperature, and the mixture stirred for
24 h. Addition of 1 N HCl solution (1 mL) was followed by extraction
with EtOAc (2 × 5 mL). The combined extracts were washed with
brine (5 mL) and dried (MgSO4), followed by flash column chroma-
tography of the resulting residue (silica gel, CH2Cl2/acetone 4:1). The
more polar natural product 2 was isolated as a pale yellow amorphous
powder (8 mg, 80%). The spectral data of synthetic 2 were identical to
those reported by Abe et al.:5a Rf 0.18 (silica gel, CH2Cl2/acetone 4:1);
IR (film) νmax (cm-1) 3425 br w (OH), 1740 s (CdO), 1667 m, 1632
1
786 m; H NMR (500 MHz, CDCl3) δ 11.05 (s, OH); 7.47 (s, 1 H),
5.17 (s OH), 3.89 (s, 3 H), 2.17 (s, 3 H), 2.13 (s, 3 H); 13C NMR (125
MHz, CDCl3) δ 194.6, 159.8, 158.0, 128.5, 114.6, 110.0, 104.7, 51.9,
15.4, 7.7; HRMS (MALDI) calcd for C10H12O4 [M + H+] 197.0814,
found 197.0807.
1
m, 1385 m, 1202 m; H NMR (500 MHz, CDCl3) δ 7.33 (dd, J1 )
14.7 Hz, J2 ) 11.0 Hz, 1 H), 7.23 (dd, J1 ) 15.4 Hz, J2 ) 11.0 Hz, 1
H), 6.38 (dq, J1 ) 15.0 Hz, J2 ) 7.0 Hz, 1 H), 6.28 (m, 3 H), 6.14 (d,
J ) 15.0 Hz, 1 H), 6.10 (d, J ) 14.6 Hz, 1 H), 3.37 (d, J ) 6.2 Hz,
1 H), 3.33 (d, J ) 2.2 Hz, 1 H), 3.19 (dd, J1 ) 6.2 Hz, J2 ) 2.2 Hz,
1 H), 1.92 (d, J ) 7.8 Hz, 3 H), 1.90 (d, J ) 7.7 Hz, 3 H), 1.55 (s, 3
H), 1.47 (s, 3 H), 1.27 (s, 3 H), 1.10 (s, 3 H); 13C NMR (125 MHz,
CDCl3) δ 208.1, 202.6, 195.0, 175.4, 173.6, 169.8, 148.0, 145.6, 143.9,
141.0, 130.9, 130.1, 126.8, 117.7, 108.4, 98.5, 82.5, 74.9, 62.5, 51.3,
43.5, 42.3, 23.5, 23.1, 19.2, 19.0, 11.0, 6.2; HRMS (MALDI) calcd
for C28H32O8Na [M + Na+] 519.1995, found 519.2001.
2,6-Dimethyl-6-acetoxy-4-carbomethoxy-3-hydroxy-2,4-cyclo-
hexadien-1-one (27). The above-described general procedure for the
lead tetraacetate oxidation of 2,4-dimethylresorcinols 14 to the corre-
sponding R-acetoxydienones 15 was followed for the preparation of
27: pale yellow glass; Rf 0.30 (silica gel, hexane/EtOAc 1:1); IR (film)
νmax (cm-1) 1737 s (CdO, acetate), 1704 s (methyl ester), 1657 m
1
(CdO, enone), 1441 m, 1372 m, 1241 s, 1104 w, 1020 w 732 w; H
NMR (500 MHz, CDCl3) δ 10.68 (s, 1 H), 7.36 (s, 1 H), 3.90 s (3 H),
2.11 s (3 H), 1.95 s (3 H), 1.54 (s, 3 H); 13C NMR (125 MHz, CDCl3)
δ 195.4, 182.5, 170.0, 160.1, 152.1, 118.5, 111.8, 78.0, 53.2, 23.8, 20.4,
7.3; ESIMS (C12H14O6) m/z (%) negative 253 ([M - H], 100).
General Procedure for the Dimerization of the Acetates Using
CsOH‚H2O as Base. To a solution of the appropriate acetate 15 (0.1
mmol) in MeOH (6 mL, 0.016 M) was added CsOH‚H2O (167 mg,
1.0 mmol, 10 equiv) at 25 °C, and the reaction mixture was stirred for
7 to 8 h. The resulting yellow-orange monomeric diquinolate was
quenched with powdered NaH2PO4‚H2O (550 mg, 4.0 mmol, 40 equiv).
After 12 h stirring at 25 °C, the remaining solid NaH2PO4‚H2O was
removed by filtration and the reaction mixture was concentrated and
then purified by careful gradient flash column chromatography (silica
gel, hexane/EtOAc 1:1 f 1:2). The Rf value for the less polar [4 + 4]
adducts ranged from 0.45 to 0.55, while for the more polar [4 + 2]
adducts ranged from 0.30 to 0.40 (silica gel, hexane/EtOAc 1:2). The
yields of these reactions are listed in Table 1. In the case of acetate
15h, the intramolecular Michael adduct 29 (21 mg, 70%, Scheme 7)
was isolated. The stable quinol 30 was isolated in high yield (22 mg,
75%) in the case of acetate 15i (Scheme 8). Spectroscopic data for
compounds 4b-f, 1e-g, 29, and 30 are given as Supporting Informa-
tion.
Trichodimerol (4). The spectral data of synthetic 4 were identical
to those reported by Ayer et al.:2a pale yellow amorphous powder; Rf
0.55 (silica gel, hexane/EtOAc 1:2); IR (film) νmax (cm-1) 3430 br w
(OH), 1610 s (CdO), 1610 m, 1405 m, 1285 m, 1148 s, 1000 m, 930
m; 1H NMR (500 MHz, C6D6) δ 17.20 (s, 2 OH), 7.36 (dd, J1 ) 14.7
Hz, J2 ) 11.1 Hz, 2 H), 6.24 (d, J ) 14.8 Hz, 2 H), 5.86 (dd, J1 )
13.3 Hz, J2 ) 11.4 Hz, 2 H), 5.54 (dq, J1 ) 14.8 Hz, J2 ) 6.9 Hz, 2
H), 3.20 (br s, 2 OH), 3.10 (s, 2 H), 1.55 (s, 6 H), 1.43 (s, 6 H), 1.53
(dd, J1 ) 6.4 Hz, J2 ) 1.8 Hz, 6 H); 13C NMR (125 MHz, C6D6) δ
197.9 (2 C), 175.9 (2 C), 143.6 (2 CH), 140.4 (2 CH), 130.9 (2 CH),
118.5 (2 CH), 104.0 (2 C), 102.8 (2 C), 78.7 (2 C), 58.8 (2 C), 57.5 (2
CH), 21.2 (2 CH3), 19.0 (2 CH3), 18.7 (2 CH3); HRMS (MALDI) calcd
for C28H32O8Na [M + Na+] 519.1995, found 519.1999. Repeating the
reaction with enantiomerically pure (S)-15a gave (-)-(4); [R]D ) -375°
(c ) 0.1, MeOH). Reported [R]D ) -376° (c ) 0.3, MeOH).2a
3,6-Dimethyl-5-(trimethylacetyl)-1,2,4-trihydroxybenzene (32).
Thermal Rearrangement of Quinol 30. A solution of quinol 30 (15
mg, 0.06 mmol) in dry benzene (5.0 mL, 0.01 M) was refluxed for 12
h. The solvent was removed by concentration, and the resulting residue
was purified by flash column chromatography (silica gel, hexane/EtOAc
1:1) to give 32 (14 mg, 93%) as a yellow glass. Compound 32 was
stored under argon to prevent its oxidation to 33: Rf 0.40 (silica gel,
hexane/EtOAc 1:1); IR (film) νmax (cm-1) 3413 br w (OH), 1666 s
(CdO), 1631 m, 1462 m, 1391 m, 1361 m, 1285 s, 1129 m, 1075 m;
1H NMR (500 MHz, C6D6) δ 5.71 (s, OH), 5.42 (s, OH), 4.58 (s, OH),
Formation of Intramolecular Michael Adduct 24. To a -78 °C
cold solution of acetate 15a (9.5 mg, 0.03 mmol) in dry THF (1.0 mL,
0.03 M) was added dropwise KHMDS (0.5 M solution in toluene, 196
µL, 0.09 mmol, 3 equiv), and the mixture turned red. The reaction
mixture was stirred for an additional 30 min and then it was quenched
at -78 °C with 1 N HCl (2.0 mL). After being warmed to 25 °C, the
mixture was extracted with CH2Cl2 (2 × 3 mL), washed with brine (5
mL), and dried (MgSO4). The residue obtained after evaporation was
purified by flash column chromatography (silica gel, CH2Cl2/MeOH
19:1) to give 24 (6 mg, 60%) as a yellow glass: Rf 0.38 (silica gel,
CH2Cl2/MeOH 19:1); IR (film) νmax (cm-1) 1790 s (CdO, lactone),
1619 s (CdO, enone), 1558 m, 1415 m, 1349 m, 1224 m, 1095 m,
1
948 w; H NMR (600 MHz, C6D6) δ 16.95 (s, OH), 7.48 (dd, J1 )
14.5 Hz, J2 ) 10.9 Hz, 1 H), 6.15 (m, 1 H), 5.74 (d, J ) 14.4 Hz, 1
H), 5.68 (m, 1 H), 2.77 (dd, J1 ) 12.2 Hz, J2 ) 8.3 Hz, 1 H), 2.08 (dd,
J1 ) 17.5 Hz, J2 ) 8.3 Hz, 1 H), 2.05 (s, 3 H), 1.90 (dd, J1 ) 17.5 Hz,
J2 ) 12.2 Hz, 1 H), 1.53 (d, J ) 7.0 Hz, 3 H), 1.11 (s, 3 H); 13C NMR
(150 MHz, C6D6) δ 189.9, 175.2, 168.7, 163.9, 140.2, 137.7, 131.1,
119.4, 110.9, 101.7, 84.0, 41.5, 36.4, 23.9, 18.6, 7.8; HRMS (MALDI)
calcd for C16H18O5Na [M + Na+] 313.1052, found 313.1063; NOE
experiments (C6D6): irradiation of signal at δ 1.11(Mea, Scheme 4)
effects signal at δ 2.77 (1.8%, Ha).
6-Formyl-2,4-dimethylresorcinol (25). A vigorously stirred solution
of 2,4-dimethylresorcinol (12, 3.94 g, 0.028 mol) and Zn(CN)2 (6.8 g,
0.056 mol, 2.0 equiv) in dry ether (50 mL, 0.6 M) was charged with
a rapid stream of HCl to allow gentle reflux. After 1 h the resulting
aldimine hydrochloride precipitated out, and the solution was cooled
to 25 °C. The ether was decanted off and the solid was washed with
dry ether (2 × 20 mL). The resulting white solid was dissolved in
H2O (150 mL) and the solution was refluxed for 1 h, whereupon the
product precipitated out. Ethanol (50 mL) was added, and the reaction
mixture was heated to redissolve the crystals and finally cooled to 0
°C. The resulting crystals were filtered off, washed with cold H2O/
EtOH (1:1, 50 mL), and dried under vacuum over P2O5 to give 25
(4.40 g, 92%) as greenish crystals: Rf 0.45 (silica gel, CH2Cl2/MeOH
19:1); mp 178-179 °C (H2O/EtOH 1:1); IR (film) νmax (cm-1) 3410
br w (OH), 1634 s (CdO), 1595 m, 1480 m, 1434 m, 1290 m, 1253
1
m, 1147 m, 951 w; H NMR (600 MHz, DMSO-d6) δ 11.44 (s, OH),
9.71 (br s, OH), 9.65 (s, 1 H), 7.27 (s, 1 H), 2.11 (s, 3 H), 2.01 (s, 3
H); 13C NMR (150 MHz, DMSO-d6) δ 195.1, 161.5, 159.5, 132.8,
117.0, 113.9, 110.4, 16.0, 8.0; ESIMS (C9H10O3) m/z (%) negative 165
([M - H], 100).
6-Carboxymethyl-2,4-dimethylresorcinol (26). To a solution of 25
(690 mg, 4.2 mmol) in DMSO (12 mL, 0.35 M) were added NaClO2
(950 mg, 10.5 mmol, 2.5 equiv) in H2O (12 mL) and NaH2PO4 (1.3 g,

Synthesis of Bisorbicillinoids
J. Am. Chem. Soc., Vol. 122, No. 13, 2000 3079
1.88 (s, 3 H), 1.78 (s, 3 H), 1.21 (s, 9 H); 13C NMR (125 MHz, CDCl3)
δ 219.2, 145.1, 144.4, 137.1, 124.1, 118.9, 111.2, 45.1, 27.7 (3 CH3),
13.6, 8.9; ESIMS (C13H18O4) m/z (%) negative 235 ([M - 3 H], 100).
This peak corresponds to the oxidized form 33.
Acknowledgment. We thank Drs. D. H. Huang and G.
Siuzdak for NMR spectroscopic and mass spectroscopic as-
sistance, respectively. This work was financially supported by
the National Institutes of Health, The Skaggs Institute for
Chemical Biology, a postdoctoral fellowship from the Alfred
Benzons Foundation (to K.B.S.), a predoctoral fellowship from
the National Science Foundation (to P.S.B.), and grants from
Schering Plough, Pfizer, Glaxo, Merck, Hoffmann-La Roche,
DuPont, and Abbott Laboratories.
3,6-Dimethyl-5-(trimethylacetyl)-2-hydroxy-p-quinone (33). Air
Oxidation of 32. Compound 32 (7 mg, 0.03 mmol) was completely
oxidized to 33 upon standing in a NMR tube in either CDCl3 (0.6 mL,
0.05 M, 4 h) or C6D6 (0.6 mL, 0.05 M, 24 h). Removal of the solvent
followed by flash column chromatography (silica gel, hexane/EtOAc
2:1) gave 33 (6.5 mg, 95%) as a yellow semicrystalline solid: Rf 0.60
(silica gel, hexane/EtOAc 1:1); IR (film) νmax (cm-1) 3384 s (OH),
1702 s (CdO, enone), 1654 s (CdO), 1631 s, 1462 w, 1391 m, 1362
m, 1304 s, 1161 m, 1047 m; 1H NMR (600 MHz, CDCl3) δ 1.94 (s, 3
H), 1.93 (s, 3 H), 1.20 (s, 9 H); 13C NMR (150 MHz, CDCl3) δ 211.2,
186.5, 183.4, 151.3, 146.3, 135.0, 116.8, 44.5, 26.9 (3 CH3), 12.9, 7.8;
ESIMS (C13H16O4) m/z (%) negative 235 ([M - H], 100).
Supporting Information Available: Spectral data (Rf, mp,
IR, 1H NMR, 13C NMR, HRMS) for 14b-i, 15b-i, 4b-f, 1e-
g, 29, and 30 (print/PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
JA9942843