Biomimetic explorations towards the bisorbicillinoids: Total synthesis of bisorbicillinol, bisorbibutenolide, and trichodimerol

Article abstract of DOI:10.1002/(SICI)1521-3773(19991203)38:23<3555::AID-ANIE3555>3.0.CO;2-Z

Strikingly simple cascade dimerization sequences can be used to assemble the complex frameworks of bisorbicillinoids such as bisorbicillinol (1), bisorbibutenolide (2), and trichodimerol (3). The mechanistic facets of the biomimetic total syntheses of these bioactive natural products were also explored. Inspection of the unique molecular architecture of these compounds reveals that they are likely to be assembled in nature by a dimerization of two oxidized forms of sorbicillin.

Full text of DOI:10.1002/(SICI)1521-3773(19991203)38:23<3555::AID-ANIE3555>3.0.CO;2-Z

COMMUNICATIONS
(1H NMR), the polymer was further purified by dialysis (MWCO
1000) in CHCl3. Polymers were obtained as waxy solids for C16 (Tg ꢀ
408C, Tm ꢀ 408C) and viscous oils for C8 (Tg ꢀ 408C). The NMR
spectra were in accordance with the proposed structure.
and substituents. This encapsulation based only on a hydro-
phobically shielded, hydrophilic solvating microenvironment
represents a different principle in comparison to the topo-
logical entrapment observed for some dendrimer-based
structures.[2] From light scattering measurements it was
tentatively concluded that the solvating species are present
in a unimolecular form in organic solvents, thus, representing
ªinverted unimolecular micellesº. Since there appears to be
no measurable release of encapsulated guests, the term
ªmicelleº is slightly misleading; we favor the term ªmolecular
nanocapsulesº. Release of the encapsulated dyes is achieved
by means of cleaving the ester bond, thus, removing the
hydrophobic molecular shield of the nanocapsules.
[16] In a typical experiment the respective aqueous dye solution (4 mL)
was manually shaken for some seconds with 0.01 wt% solution (4 mL)
of the partially esterified polyglycerol in chloroform. After phase
separation (two experiments after 3 h and 24 h), an aliquot of the
organic layer (2 mL) was transfered into a UV/Vis cuvette and its
absorption spectrum measured. No difference between the 3 h and
24 h measurements was observed. The solutions remained unchanged
for several months.
[17] The preparation of chiral polyglycerol based on the pure enantiomers
of glycidol as a basis for chiral molecular nanocapsules has been
investigated: A. Sunder, R. Mülhaupt, R. Haag, H. Frey, Macro-
molecules, in press.
Molecular nanocapsules and their corresponding host/guest
compounds offer an attractive potential for use in a wide
variety of applications ranging from controlled drug release,
solubilization of inorganic compounds in organic media,
dispersion of polar dyes in hydrophobic polymers, preparation
of inorganic/organic hybrid nanoparticles, to the design of
microreactors and catalysts.[17]
Biomimetic Explorations Towards the
Bisorbicillinoids: Total Synthesis of
Bisorbicillinol, Bisorbibutenolide,
and Trichodimerol**
Received: July 22, 1999 [Z13764IE]
German version: Angew. Chem. 1999, 111, 3758 ± 3761
Keywords: micelles ´ polymers ´ solubilization
K. C. Nicolaou*, Klaus B. Simonsen, Georgios
Vassilikogiannakis, Phil S. Baran, Veroniki P. Vidali,
Emmanuel N. Pitsinos, and Elias A. Couladouros
[1] O. A. Matthias, A. N. Shipway, J. F. Stoddart, Prog. Polym. Sci. 1998,
23, 1.
[2] J. F. G. A. Jansen, E. M. M. de Brabander-van den Berg, E. W. Meijer,
Science 1994, 266, 1226.
Dedicated to Professor Gerasimos J. Karabatsos
on the occasion of his 67th birthday
[3] G. R. Newkome, C. N. Moorefield, G. R. Baker, M. J. Saunders, S. H.
Grossman, Angew. Chem. 1991, 103, 1207; Angew. Chem. Int. Ed.
Engl. 1991, 30, 1178.
The bisorbicillinoids[1] are a growing family of novel natural
products with interesting and diverse biological activities.
Included within this class are bisorbicillinol (1),[2] bisorbibu-
tenolide (2),[3] trichodimerol (4),[4] bisorbicillinolide (5),[3] and
Â
[4] A. P. H. J. Schenning, C. Elissen-Roman, J.-W. Weener, M. W. P. L.
Baars, S. J. van der Gaast, E. W. Meijer, J. Am. Chem. Soc. 1998, 120,
8199.
[5] K. R. Gopidas, A. R. Leheny, G. Caminati, N. J. Turro, D. A. Tomalia,
J. Am. Chem. Soc. 1991, 113, 7335.
[*] Prof. Dr. K. C. Nicolaou, Dr. K. B. Simonsen,
Dr. G. Vassilikogiannakis, P. S. Baran
Department of Chemistry and
Â
[6] C. J. Hawker, K. L. Wooley, J. M. J. Frechet, J. Chem. Soc. Perkin
Trans. 1 1993, 1287.
[7] S. Stevelmans, J. C. M. van Hest, J. F. G. A. Jansen, D. A. F. J. van -
Boxtel, E. M. M. de Brabander-van den Berg, E. W. Meijer, J. Am.
Chem. Soc. 1996, 118, 7398.
The Skaggs Institute for Chemical Biology
The Scripps Research Institute
10550 North Torrey Pines Road, La Jolla, CA 92037 (USA)
Fax : (‡ 1)858-784-2469
[8] A. I. Cooper, J. D. Londono, G. Wignall, J. B. McClain, E. T. Samulski,
Â
J. S. Lin, A. Dobrynin, M. Rubinstein, A. L. C. Burke, J. M. J. Frechet,
E-mail: kcn@scripps.edu
J. M. DeSimone, Nature 1997, 389, 368.
and
[9] Shape-selective release out of a dendritic box has been reported in:
J. F. G. A. Jansen, E. W. Meijer, E. M. M. de Brabander-van den Berg,
J. Am. Chem. Soc. 1995, 117, 4417.
Department of Chemistry and Biochemistry
University of California San Diego
9500 Gilman Drive, La Jolla, California 92093 (USA)
[10] The solubilization of boron clusters prepared by a reaction within a
dendritic unimolecular micelle has been reported in: G. R. Newkome,
C. N. Moorefield, J. M. Keith, G. R. Baker, G. H. Escamilla, Angew.
Chem. 1994, 106, 701; Angew. Chem. Int. Ed. Engl. 1994, 33, 666.
[11] Y. H. Kim, J. Polym. Sci. Polym. Chem. 1998, 36, 1685.
[12] A. Sunder, R. Hanselmann, H. Frey, R. Mülhaupt, Macromolecules
1999, 32, 4240.
[13] Y. H. Kim, O. W. Webster, J. Am. Chem. Soc. 1990, 112, 4592.
[14] C. J. Hawker, F. Chu, Macromolecules 1996, 29, 4370.
[15] Typical procedure for the preparation of P(G84C160.6): dried hyper-
branched polyglycerol (DPn ˆ 84; 10.1 g) was treated with palmitoyl
chloride (25 mL, 83 mmol) in pyridine in the presence of N-
methylimidazole (0.5 mL) under anhydrous and reflux conditions.
After removal of most of the solvent, K2CO3 was added for work-up
and pyridine residues were removed by azeotropic distillation with
toluene. In cases where residues of free carboxylic acid were detected
V. P. Vidali,[‡] Dr. E. N. Pitsinos, Prof. Dr. E. A. Couladouros[‡]
Organic and Bioorganic Chemistry Laboratories
NCRS Demokritos
153 10 Agia Paraskevi, Attikis, POB 60228, Athens (Greece)
‡
[ ] Further address:
Department of Sciences, Chemistry Laboratory
Agricultural University of Athens
Iera Odos, Athens 11855 (Greece)
[**] We thank Dr. D. H. Huang and Dr. G. Siuzdak for assistance with
NMR spectroscopy and mass spectrometry, respectively. This work
was financially supported by the National Institutes of Health (USA),
The Skaggs Institute for Chemical Biology, a postdoctoral fellowship
from the Alfred Benzons Foundation (K.B.S.), and grants from
Schering Plough, Pfizer, Glaxo, Merck, Hoffmann-La Roche, DuPont,
and Abbott Laboratories.
Angew. Chem. Int. Ed. 1999, 38, No. 23
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3823-3555 $ 17.50+.50/0
3555

COMMUNICATIONS
program directed towards their biomimetic
total synthesis. Here we report the total
syntheses of three members of this class of
natural products, namely, bisorbicillinol (1),
bisorbibutenolide (2) and trichodimerol (4).
The unifying feature of these structurally
connected dodecaketides resides in the
mechanistically related proposals regarding
their biosynthetic origin from sorbicillin (3;
Figures 1 and 2). Thus, the first member of
this family, bisvertinoquinol (1a), reported
by Dreiding et al. in 1983, was postulated to
arise from a Diels ± Alder reaction between
two different quinols derived from 3 and
2',3'-dihydrosorbicillin by enantioselective
oxidation.[7] A similar hypothesis was ad-
vanced by Abe et al. to explain the biosyn-
thesis of the newly isolated bisorbicillinol (1)
(Figure 1).[2] The same group also proposed
the biosynthesis of the soon thereafter iso-
lated bisorbibutenolide (2) and bisorbicilli-
nolide (5) by anionic rearrangement of 1.[3a]
We have recently proposed a possible bio-
synthetic pathway for trichodimerol (4) in-
volving a two-step Michael addition ± ketal-
ization sequence from an oxidized form of
3.[1] We now suggest that the bisvertinols
(e.g. 6), representing the largest group of the
bisorbicillinoids, are biosynthesized by a
similar mechanism involving a Michael ad-
dition followed by ketalization and finally
reduction (see Figure 2).[8]
OH
O
O
OH
O
O
3'
2'
H
O
O
HO
9
O
HO
OH
O
O
HO
HO
2: bisorbibutenolide
1: bisorbicillinol
1a: 2',3'-dihydrobisorbicillinol
OH
O
O
OH
OH
O
O
HO
H
O
OH
O
3: sorbicillin
4: trichodimerol
OH
OH
O
OH O
O
OH
O
H
H
H
O
O
O
HO
H
O
O
OH
O
H
6: bisvertinol
5: bisorbicillinolide
Figure 1. Structures of selected bisorbicillinoids and their postulated biosynthetic precursor
sorbicillin (3).
bisvertinol (6),[5] all of which are
thought to be biosynthetically de-
rived from sorbicillin (3), itself a
naturally occurring substance (Fig-
ure 1).[6]
Enantioselective
oxidation
OH
O
O
HO
HO
OH
OH
Oxidative dimerization
OH
O
Isolated from various species of
fungi, bisorbicillinol (1), bisorbibute-
nolide (2), and bisorbicillinolide (5)
exhibit antioxidant properties,[2, 3]
while an oxidized form of bisvertinol
(6) represents the first inhibitor of b-
6-glucan biosynthesis, and is, there-
fore, considered as a potential anti-
fungal agent.[5b] Trichodimerol (4),
on the other hand, originally isolated
from the genus Trichoderma, showed
inhibitory activity against lipopoly-
saccharide-induced production of tu-
mor necrosis factor a (TNF-a) in
human monocytes and, consequent-
ly, is considered to be a potential lead
for the treatment of septic shock.[4b]
The novel and complex molecular
architectures of these molecules cou-
pled with their intriguing biological
properties and unusual biosynthetic
pathways prompted us to initiate a
O
HO
OH
O
O
3: sorbicillin
4: trichodimerol
path a
a
a
HO
OH
Ketalization
O
HO
HO
OH
O
OH
a
HO
O
b
O
Reduction
O
OH
O
O
Michael reaction
O
8b
8a
OH
OH
O
O
path b
OH
O
H
H
O
O
H
6: bisvertinol
Figure 2. Proposed biosynthetic pathways from sorbicillin (3) to trichodimerol (4) and bisvertinol (6).
3556
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3823-3556 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1999, 38, No. 23

COMMUNICATIONS
Table 1. Selected physical properties of compounds 7a, 12, and 13.
Inspired by biosynthetic considerations, we began our
synthetic studies towards the bisorbicillinoids with sorbicillin
(3) and had acetate 7a (Scheme 1) as the first objective. Thus,
3, obtained by the boron trifluoride catalyzed acylation of 2,4-
dimethylresorcinol[9] with sorbic acid,[10] was subjected to
7a: Rf ˆ 0.41 (silica gel, dichloromethane/acetone 9/1); m.p. 149 ± 1508C
(diethyl ether/n-hexane); IR (film): nÄmax ˆ 2930, 1737, 1650, 1644, 1612,
1
1555, 1215, 1245, 1070, 1018 cm 1; H NMR (500 MHz, CDCl3): d ˆ 11.90
(s, 1H), 7.46 (dd, J ˆ 14.8, 10.8 Hz, 1H), 7.25 (s, 1H), 6.66 (d, J ˆ 14.8 Hz,
1H), 6.38 (m, 1H), 6.31 (m, 1H), 2.15 (s, 3H), 1.93 (d, J ˆ 6.6 Hz, 3H), 1.86
(s, 3H), 1.49 (s, 3H); 13C NMR (125 MHz, CDCl3): d ˆ 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;
‡
3
HR-MS (MALDI): calcd for C16H18O5Na [M ‡ Na ]: 313.1052, found:
OAc
O
O
313.1055
a)
OAc
12: Rf ˆ 0.40 (silica gel, ethyl acetate/hexane 2/3); IR (film): nÄmax ˆ 2931,
[7a:7b ca. 5:1]
1
1769, 1710, 1671, 1370, 1242, 1185, 1071 cm 1; H NMR (500 MHz, C6D6):
OH
OH
d ˆ 5.90 (s, 1H), 4.08 (s, 1H), 3.51 (ddd, J ˆ 13.8, 10.0, 5.5 Hz, 1H), 2.87 (dt,
J ˆ 19.3, 7.3 Hz, 1H), 2.63 (ddd, J ˆ 13.7, 10.1, 5.5 Hz, 1H), 2.19 (dt, J ˆ
19.3, 7.3 Hz, 1H), 1.96 (s, 3H), 1.92 ± 1.86 (m, 1H), 1.86 (s, 3H), 1.78 (s,
3H), 1.73 (s, 3H), 1.72 (s, 3H), 1.69 (s, 3H), 1.68 ± 1.61 (m, 1H), 1.59 (s,
3H), 1.58 ± 1.52 (m, 2H), 1.50 (s, 3H), 1.41 ± 1.34 (m, 2H), 1.33 ± 1.26 (m,
2H), 1.25 ± 1.16 (m, 2H), 1.15 ± 1.07 (m, 2H), 0.88 (t, J ˆ 7.4 Hz, 3H), 0.81
(t, J ˆ 7.3 Hz, 3H); 13C NMR (150 MHz, C6D6): d ˆ 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, 22.9, 21.6,
20.7, 20.0, 19.9, 14.3, 14.1, 10.7, 10.6; HR-MS (MALDI): calcd for
O
O
7b
7a
b)
OH
OH
OH
O
‡
C36H48O12Na [M ‡ Na ]: 695.3038, found: 695.3013
O
OH
13: Rf ˆ 0.38 (silica gel, dichloromethane/methanol 19/1); IR (film): nÄmax
ˆ
O
O
1
2933, 1790, 1619, 1558, 1415, 1349, 1224, 1095, 948 cm
;
1H NMR
8a [diene]
8b [dienophile]
(600 MHz, C6D6): d ˆ 16.95 (s, 1H), 7.48 (dd, J ˆ 14.5, 10.9 Hz, 1H),
6.19 ± 6.11 (m, 1H), 5.74 (d, J ˆ 14.4 Hz, 1H), 5.72 ± 5.64 (m, 1H), 2.77 (dd,
J ˆ 12.2, 8.3 Hz, 1H), 2.08 (dd, J ˆ 17.5, 8.3 Hz, 1H), 2.05 (s, 3H), 1.90 (dd,
J ˆ 17.5, 12.2 Hz, 1H), 1.53 (d, J ˆ 7.0 Hz, 3H), 1.11 (s, 3H); 13C NMR
(150 MHz, C6D6): d ˆ 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; HR-MS (MALDI): calcd for
Diels-Alder reaction
OH
O
O
‡
C16H18O5Na [M ‡ Na ]: 313.1052, found: 313.1063
HO
O
O
HO
diphenylseleninic acid anhydride.[13] The proposed pathway
from 7a to 1 was supported by NMR studies. Thus, when the
above experiment was carried out in an NMR tube using
[D8]THF and D2O, the 1H NMR signals of the rapidly formed
deep orange solution revealed the presence of two distinct
quinolates and the absence of 1 until acidification, whereupon
the resonances corresponding to 1 appeared. These observa-
tions suggested the possibility of converting 7a into 1 directly
by acid hydrolysis of the acetate group. Indeed this was
proven to be the case, whereby treatment of 7a with
concentrated HCl in THF produced 1 in 43% yield. Again a
1H NMR experiment proved the fleeting nature of quinols 8a
and 8b, while it nicely allowed the monitoring of the
appearance of 1 at the expense of 7a during the course of
two hours.
Interestingly, when the saponification of 7a was carried out
with KOH (10 equiv) in a minimum amount of THF/H2O (10/
1), a different product was formed, together with only a small
amount of the previously obtained 1 (Scheme 2). Although
the major product according to thin-layer chromatography
(TLC, ca. 65% crude yield), the new compound proved labile
on silica gel and could only be isolated in pure form by flash
column or preparative thin layer chromatography, and the
yield was much lower. The physical properties of the new
compound were suggestive of the dimeric structure 10.
Further support for this structure was obtained by hydro-
genation of the side chains (H2, 10% Pd/C, EtOAc, 81%) and
acetylation of the resulting compound (11; Ac2O, 4-DMAP,
80%) to furnish tetraacetate 12, whose structure was
fully established by 1H, 13C, 1H ± 1H COSY, HMQC, and
HO
1: bisorbicillinol (racemic)
Scheme 1. Biomimetic total synthesis of bisorbicillinol (1) from sorbicillin
(3). a) Pb(OAc)4 (1.2 equiv), AcOH, 258C, 2 h, 40% 7a and 10% 7b;
b) KOH (10 equiv), THF/H2O (9/1), 08C, 2 h; then 1n aq. HCl, 40%; or
THF/conc. HCl (9/1), 258C, 2 h, 43%.
oxidation with lead tetraacetate in acetic acid to afford the
desired a-hydroxydienone 7a as the major product together
with its regioisomer 7b (ca. 5:1 ratio, Scheme 1). Flash column
chromatography (silica, dichloromethane/acetone 9/1) fol-
lowed by recrystallization furnished 7a in 40% yield
(Table 1).
When acetate 7a was treated with solid KOH (10 equiv) in
THF/H2O (9/1; 0.05 m) at 08C for 1 h followed by quenching
with 1n aq. HCl, the Diels ± Alder adduct bisorbicillinol (1)
was isolated in 40% yield. The spectral properties of synthetic
1 were identical to those reported by Abe et al.[2, 11] Generat-
ing four stereogenic centers, two of which are quaternary, this
reaction proceeds with remarkable regio- and diastereocon-
trol (endo selectivity).[12]
It is postulated that the first stages of this sequence (7a !1)
involve deacetylation followed by scrambling of the resulting
dianion to a mixture of diquinolates (derived from 8a and 8b,
Scheme 1). Acidification of the reaction mixture presumably
results in the formation of quinols 8a and 8b, which readily
combine in a Diels ± Alder reaction to generate 1. A similar
dimerization was observed by Barton et al. when they
generated ortho-quinols by allowing phenols to react with
Angew. Chem. Int. Ed. 1999, 38, No. 23
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3823-3557 $ 17.50+.50/0
3557

COMMUNICATIONS
(Scheme 3, path a). With less than one equivalent of base,
7a follows a different pathway (path b) initiated by the
formation of its quinolate 14a. Thus, rearrangement of
quinolate 14a followed by acetate migration leads to alkoxy
anion 14b, which undergoes an intramolecular Michael
addition to the enone system to form epoxide 14c. Stereo-
selective epoxide opening by a molecule of water furnishes
diol 15.[14]
O
O
O
O
O
AcO
O
O
a)
OH
O
O
O
9
7a
9
b)
The prospect of generating trichodimerol (4) from acetate
7a was finally realized when we altered the method of
quenching the corresponding monomeric diquinolate
(Scheme 4). Thus, treatment of 7a with CsOH ´ H2O for 7 h
followed by addition of powdered NaH2PO4 ´ H2O and stirring
at 258C for 12 h furnished 4, which was isolated in 16% yield
after column chromatography together with 3 (12%) and 1
(22%). Synthetic 4 (racemic) exhibited identical spectroscop-
ic properties to that reported by Ayer et al.[4a, 15] The very slow
neutralization of this reaction is crucial for the extraordinary
dimerization of acetate 7a.[16]
Finally, the hypothesis put forward by Abe et al.[3] for the
biosynthesis of bisorbibutenolide (2) and bisorbicillinolide (5)
from bissorbicillinol (1) was investigated. Indeed, when a
solution of 1 in THF was treated with one equivalent of
KHMDS at ambient temperature, a more polar compound
was formed (80% yield) whose structure was established to be
that of bisorbibutenolide (2) by spectroscopic means and
RO
O
HO
O
OR
c)
RO
OH
Meb
HO
O
Ha
O
RO
Mea
O
H
O
Hb
HO
H
O
O
d)
10
11: R = H
12: R = Ac
Scheme 2. Synthesis of sorbicillin dimer 10 from acetate 7a. a) KOH
(10 equiv), THF/H2O (10/1), 08C, 2 h; b) 1n aq. HCl, 65% (crude yield);
c) H2, 10% Pd/C, EtOAc, 258C, 2 h, 81%; d) Ac2O (10 equiv), 4-DMAP
(0.1 equiv), EtOAc, 258C, 2 h, 80%. nOes observed for 12: Mea/Ha (4.5%),
Ha/Hb (3.9%), and Hb/Meb (3.0%). 4-DMAP ˆ 4-(dimethylamino)pyri-
dine.
HMBC NMR spectroscopic techniques. Particularly useful in
defining the configuration of 12 were the nOes observed, in a
1D nOe experiment, for the following
protons: Mea/Ha (4.5%); Ha/Hb (3.9%);
Hb/Meb (3.0%). A mechanistic rationale
for the formation of 10 from 7a is pre-
sented in Scheme 2. Thus, it is reasoned
that the initially formed dianion 9 rear-
ranges by an intramolecular Michael reac-
O
O
O
O
O
OAc
O
O
O
O
b)
path b
OH
O
O
O
tion to an epoxy dianion (not shown),
which attacks a second molecule of 9 in an
14a
14b
7a
intermolecular Michael reaction, generat-
ing in the process a quaternary center; the
observed product 10 is formed upon acid-
ification.
path a
a)
Mea
OH
OAc
O
O
OAc
O
HO
HO
In another series of experiments, we
attempted to force acetate 7a into entering
pathways that may lead to other bisorbi-
cillinoids. Thus, while reaction of 7a with
LiHMDS at 78 !258C produced only
the orange lithium enolate, treatment of 7a
with excess KHMDS (3.0 equiv) in anhy-
drous THF at 788C produced first a red
solution and subsequently, upon acidifica-
tion, the yellow g-lactone 13 (60% yield,
O
H3O
H
Ha
H
O
O
O
H
OH
O
H
14c
13
15
Scheme 3. Transformations of acetate 7a. a) KHMDS (3.0 equiv), THF,
788C, 2 h, 60%;
b) KHMDS (0.9 equiv), THF, 78 !258C, 2 h: then 258C, 12 h, 35%. nOes observed for 13: Mea/
Ha (1.8%). KHMDS ˆ potassium salt of 1,1,1,3,3,3-hexamethyldisilazane.
Scheme 3). When the same reaction was performed with
0.9 equiv of KHMDS at 788C, a yellow solution (presum-
ably of the enolate anion) was initially formed, and after
stirring at 258C for 12 h followed by aqueous workup, the
yellow and rather labile trans-diol 15 was isolated. Mechanis-
tically, these events can be rationalized by assuming enolate-
induced cascade reactions. Thus, for the formation of 13,
acetate 7a is required to undergo rearrangement of its quinol
moiety to form its acetate enolate before collapsing, by an
intramolecular Michael addition, on its enone system
OH
OAc
O
O
O
a)
OH
O
OH
HO
OH
O
O
7a
4: trichodimerol
Scheme 4. Biomimetic total synthesis of trichodimerol (4) from acetate 7a.
a) CsOH ´ H2O (10 equiv), MeOH, 258C, 7 h; then NaH2PO4 ´ H2O, 258C,
12 h, 16% 4, 12% sorbicillin (3) and 22% bisorbicillinol (1).
3558
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3823-3558 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1999, 38, No. 23

COMMUNICATIONS
[3] a) N. Abe, T. Murata, A. Hirota, Biosci.
Biotechnol. Biochem. 1998, 62, 2120 ±
2126; b) a second fungal metabolite differ-
ing from bisorbibutenolide (2) only in its
relative configuration at C9 was isolated in
1997 and named trichotetronine: O. Shir-
ota, V. Pathak, C. F. Hossain, S. Sekita, K.
Takatori, M. Satake, J. Chem. Soc. Perkin
Trans. 1 1997, 2961 ± 2964.
OH
OH
O
O
O
O
a)
O
HO
O
O
O
HO
O
HO
HO
HO
16
1: bisorbicillinol
[4] a) R. Andrade, W. A. Ayer, P. P. Mebe,
Can. J. Chem. 1992, 70, 2526 ± 2535;
b) G. A. Warr, J. A. Veitch, A. W. Walsh,
G. A. Hesler, D. M. Pirnik, J. E. Leet, P.-
F. M. Lin, I. A. Medina, K. D. McBrien, S.
Forenza, J. M. Clark, K. S. Lam, J. Anti-
biot. 1996, 49, 234 ± 240.
[5] a) L. S. Trifonov, H. Hilpert, P. Floersheim,
A. S. Dreiding, D. M. Rast, R. Skrivanova,
L. Hoesch, Tetrahedron 1986, 42, 3157 ±
3179; b) M. Kontani, Y. Sakagami, S.
Marumo, Tetrahedron Lett. 1994, 35,
2577 ± 2580.
OH
O
OH
O
O
O
H
O
O
HO
O
O
HO
O
O
OH
OH
2: bisorbibutenolide
17
Scheme 5. Biomimetic synthesis of bisorbibutenolide (2) from bisorbicillinol (1). a) KHMDS
(1.1 equiv), THF, 258C, 24 h; then 1n aq. HCl, 80%.
[6] a) D. J. Cram, M. Tishler, J. Am. Chem.
Soc. 1948, 70, 4238 ± 4239; b) D. J. Cram, J.
1
comparison of its H and 13C NMR spectra with those of the
naturally occurring compound (Scheme 5).[11] Interestingly,
bisorbicillinolide (5) was not observed in this reaction.
The findings described hereinÐincluding the total synthesis
of bisobicillinol (1), bisorbibutenolide (2), and trichodimerol
(4)Ðshine light on the chemistry of the bisorbicillinoids and
confirm a number of hypothetical proposals regarding their
biosynthesis. Furthermore, the novel synthetic pathways
elucidated during this study add to the repertoire of cascade
reactions as an enabling technology for complex molecule
construction, while the unusual molecular frameworks pro-
duced enrich our pool of advanced synthetic intermediates
and building blocks.
Am. Chem. Soc. 1948, 70, 4240 ± 4243.
[7] a) L. S. Trifonov, A. S. Dreiding, L. Hoesch, D. M. Rast, Helv. Chim.
Acta 1981, 64, 1843 ± 1846; b) L. S. Trifonov, J. H. Bieri, R. Prewo, A. S.
Dreiding, L. Hoesch, D. M. Rast, Tetrahedron 1983, 39, 4243 ± 4256.
[8] A previous proposal regarding the biosynthesis of bisvertinol (6) from
sorbicillin postulates a different pathway, in which an initial epox-
idation of sorbicillin followed by reduction and then dimerization and
ketalization is believed to be involved.[7]
[9] W. Baker, H. F. Bondy, J. F. W. McOmie, H. R. Tunnicliff, J. Chem.
Soc. 1949, 2834 ± 2835.
[10] J. F. W. McOmie, M. S. Tute, J. Chem. Soc. 1958, 3226 ± 3227. This
method was found to be superior on larger scale to the procedure
described by the Sartori group; see F. Bigi, G. Casiraghi, G. Casnati, S.
Marchesi, G. Sartori, C. Vignali, Tetrahedron 1984, 40, 4081 ± 4084.
[11] We thank Dr. N. Abe of the University of Shizuoka for kindly
providing us with the 1H and 13C NMR spectra of bisorbicillinol (1)
and bisorbibutenolide (2).
[12] The acetate 7a did not give any Diels ± Alder product 1 when heated
in benzene or in AcOH for several hours.
Received: September 6, 1999
Revised version: October 11, 1999 [Z13976IE]
German version: Angew. Chem. 1999, 111, 3762 ± 3766
[13] a) D. H. R. Barton, P. D. Magnus, M. N. Rosenfeld, J. Chem. Soc.
Chem. Commun. 1975, 301 ± 301; b) D. H. R. Barton, P. D. Magnus,
J. C. Quinney, J. Chem. Soc. Perkin Trans. 1 1975, 1610 ± 1614; see also:
E. Adler, K. Holmberg, Acta. Chem. Scand. Ser. B 1974, 28, 465 ± 472.
[14] Attempts to induce 7a to react under Lewis acid catalysis conditions
were also made. Interestingly, when 7a was exposed to TBSOTf or
TMSOTf, 3 was the major product formed together with small
amounts of 1. TBS ˆ tert-butyldimethylsilyl, Tf ˆ trifluormethanesul-
fonyl, TMS ˆ trimethylsilyl.
Keywords: biomimetic synthesis ´ bisorbicillinols ´ dimeri-
zations ´ natural products ´ trichodimerol
[1] The term ªbisorbicillinoidsº was recently introduced by us to describe
all dimeric natural products derived from sorbicillin: K. C. Nicolaou,
R. Jautelat, G. Vassilikogiannakis, P. S. Baran, K. B. Simonsen, Chem.
Eur. J. 1999, 5, 3651 ± 3665.
[2] N. Abe, T. Murata, A. Hirota, Biosci. Biotechnol. Biochem. 1998, 62,
661 ± 66.
[15] Trichodimerol (4) was unavailable to us for direct comparison.
[16] For the first synthesis of enantiomerically pure trichodimerol by
dimerization of resolved acetate 7a, see D. Barnes-Seeman, E. J.
Corey, Org. Lett. 1999, 1, 1503 ± 1504.
Angew. Chem. Int. Ed. 1999, 38, No. 23
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3823-3559 $ 17.50+.50/0
3559