Concise total synthesis of (-)-frondosin B using a novel palladium-catalyzed cyclization

Article abstract of DOI:10.1002/1521-3773(20020503)41:9<1569::AID-ANIE1569>3.0.CO;2-8

Sidestepping stereochemical inversion: In an expedient, asymmetric total synthesis of (-)-frondosin B a novel palladium-catalyzed cyclization establishes the tetracyclic skeleton (see scheme). The absolute stereochemistry of naturally occurring marine met

Full text of DOI:10.1002/1521-3773(20020503)41:9<1569::AID-ANIE1569>3.0.CO;2-8

COMMUNICATIONS
construction of the tetracyclic core from an alkynyl phenol
precursor of type 5 (Scheme 2). No intramolecular version of
this palladium-catalyzed reaction has been reported to date.
Despite efforts to convert 5 into 6 with a range of palladium
reagents, however, we were unable to effect the desired
cyclization. Instead, extensive decomposition of the starting
material took place.
Concise Total Synthesis of (À)-Frondosin B
Using a Novel Palladium-Catalyzed
Cyclization**
Chambers C. Hughes and Dirk Trauner*
The frondosins (Scheme 1), a family of marine terpenoids
isolated from the sponge Dysidea frondosa, have been found
to inhibit the binding of interleukin-8 (IL-8) to its receptor in
the low micromolar range.[1a] IL-8, which is produced by
macrophages and endothelial cells, recruits T-lymphocytes
MeO
MeO
TfO
OH
O
HO
[Pd]
HO
OTf
Pd0
OH
O
Me
Me
10
O
O
11
8
Me
Me
5
1: frondosin B
2: frondosin A
MeO
O
O
10
11
O
OH
O
OH
Me
X
O
Me
Me
6
Scheme 2. Initial strategy for the key cyclization in the synthesis of
frondosin B (1). Tf ˆ trifluoromethanesulfonyl.
3: frondosin C
4: frondosin D
Scheme 1. The frondosins. The absolute stereochemistry shown is arbi-
trary.
Attempting to gain more insight into the mechanism of the
Arcadi Cacchi reaction, we found that efficient, palladium-
catalyzed bond formation could indeed be achieved provided
the benzofuran moiety was already in place. This observation
greatly simplified and streamlined our overall synthetic plan,
as the formation of the free phenol 5, unlike the correspond-
ing benzofuran 15 (see below), could be realized only through
difficult protective group manipulations.
and neutrophils to an inflammatory site and is implicated in a
wide range of acute and chronic inflammatory disorders. IL-8
receptor antagonists like the frondosins, therefore, interfere
with the inflammatory cascade and might ultimately be used
to prevent autoimmune disorders such as rheumatoid arthritis
and psoriasis. Equally important, members of the frondosin
family have also been shown to exhibit HIV-inhibitory
properties.[1b]
Our synthesis starts with the known R-configured alkyne
7
[4] (obtained in 91% ee by means of a Sharpless epoxidation)
and the known aryl bromide 8[5] (Scheme 3). Sonogashira
coupling[6] of these two components followed by acidic
deprotection[7] gave the primary alcohol 10. Subsequent
saponification of the phenolic acetate with concomitant
cyclization afforded benzofuran 11 in high yield. This material
was then converted to the corresponding iodide 12. Alkyla-
tion of dimethoxylithiocyclohexadiene 13 by using the Piers
protocol[8] followed by hydrolysis afforded cyclohexa-1,3-
dione 14.[9] This surprisingly unstable diketone decomposed
on standing and was therefore immediately converted into
enol triflate 15 by which the stage for the key cyclization was
set.
Recently, (‡)-frondosin B (1) has attracted much attention.
The first total synthesis of this interesting norsesquiterpenoid
by the Danishefsky group confirmed the structure and led to
the assignment of an absolute configuration.[2] Although 1
features only one stereocenter, it represents a considerable
synthetic challenge because of its unusual tetracyclic skeleton,
which incorporates a 2,3-disubstituted benzofuran moiety and
a readily isomerized tetrasubstituted double bond.
A key connection, also noticed by others,[2] is the formation
of the bond between C10 and C11 to establish the central
seven-membered ring of the target. Our original plan foresaw
the implementation of the Arcadi Cacchi reaction[3] for the
The key bond formation between C10 and C11 was
achieved in good yield by heating 15 with a catalytic amount
of [Pd(PPh3)4] in dimethyl acetamide (DMA) in the presence
of H¸nig×s base.[10] Importantly, no detectable racemization of
the C8 stereocenter occurred under these conditions (see
Experimental Section).[11]
[*] Prof. D. Trauner, C. C. Hughes
Center for New Directions in Organic Synthesis
Department of Chemistry
University of California, Berkeley
Berkeley, CA 94720-1460 (USA)
Fax : (‡1)510-643-9480
E-mail: trauner@cchem.berkeley.edu
The conversion of the carbonyl group into a gem-dimethyl
moiety was achieved by using a protocol developed by Reetz
[**] This work was supported by a generous gift from Merck & Co.
Angew. Chem. Int. Ed. 2002, 41, No. 9
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4109-1569 $ 20.00+.50/0
1569

COMMUNICATIONS
triphenylphosphane or amines was
also excluded. Thus, it appears that
oxidative addition of palladium to the
carbon oxygen bond of the triflate is
essential and that this is a genuine
palladium-catalyzed reaction.
MeO
MeO
MeO
OAc
OAc
Br
O
8
c
R
PMBO
Me
X
Me
X
Me
a
Although isolated examples of Heck
reactions involving furans have been
reported,[13] none have been used in
the total synthesis of a natural prod-
uct.[14] Moreover, no benzofurans have
been employed as substrates. The fact
that the stereocenter at C8 does not
racemize under the reaction condi-
tions raises the question whether a
standard Heck mechanism involving
oxidative addition (15 !17), migrato-
ry insertion (17 !18), and b-hydride
9: X = OPMB
10: X = OH
11: X = OH
12: X = I
7
b
d
OMe
OMe
OMe
Li
OMe
13
e,f
OTf
O
g
h
O
O
O
Me
15
O
Me
14
elimination
(18 !19)
operates
MeO
HO
MeO
(Scheme 4, pathway A). Provided that
the product of migratory insertion, 18,
undergoes syn-elimination of a hydri-
dopalladium species, racemization
would be expected upon rearomatiza-
tion of the resulting enol ether 19.
Since racemization does not occur,
however, an alternative mechanism
may be operating (Scheme 4, path-
ways B and C). In the first, oxidative
addition is followed, instead, by nucle-
ophilic attack of the electron-rich
benzofuran onto the palladium(ii) in-
termediate 17, which is possibly cat-
ionic in nature. Deprotonation of the
cationic intermediate 20 then affords
O
O
O
i,j
k
8
Me
R
Me
Me
O
(–)-1
6
16
Scheme 3. Total synthesis of (À)-frondosin B (À)-1. a) [Pd(PPh3)4], CuI, Et3N, MeCN, reflux, 94%;
b) TFA, CH2Cl2, RT, 83%; c) K2CO3, MeOH, reflux, 92%; d) MsCl, Et3N, THF, 08C, then NaI, Me2CO,
reflux, 93%; e) 13, HMPA, THF, À788C !RT, 85%; f) ion-exchange resin, Me2CO, H2O, reflux;
g) NaHMDS, PhNTf2, THF, Et2O, RT, 75% (2steps); h) [Pd(PPh 3)4], DMA, iPr2NEt, 908C, 70%;
i) MeMgBr, THF, À788C !RT; j) Me2Zn, TiCl4, CH2Cl2, 08C !RT, 82% from 6; k) NaSEt, DMF,
reflux, 90%. PMB ˆ para-methoxybenzyl, TFA ˆ trifluoroacetic acid, Ms ˆ methanesulfonyl, HMPA ˆ
hexamethylphosphoramide, NaHMDS ˆ sodium bis(trimethylsilyl)amide, DMA ˆ N,N-dimethylacet-
amide.
21, which reductively eliminates Pd0 to yield the product (À)-6
with the stereocenter at C8 left intact. In the second, the
highly electrophilic, cationic palladium species 17 undergoes
et al.[12] As a result of the low electrophilicity of the vinyl-
ogous aryl ketone, this procedure had to be carried out in two
separate steps. Reaction of ketone 6 with MeMgBr gave the
corresponding crude, highly acid-sensitive tertiary alcohol,
which was subsequently subjected to the action of [Me2TiCl2]
(formed in situ from Me2Zn and TiCl4). This afforded O-
methyl frondosin B (16) in good overall yield. Finally, O-
demethylation following a previously reported procedure
gave frondosin B (1) in high optical purity.[2]
To our surprise, however, the optical rotation of the (R)-
configured material thus obtained was opposite to that of the
natural product ([a]D25 ˆ À16.8 compared to [a]D ˆ ‡18.6[1a]).
Though the natural product had been assigned the R
configuration in the synthesis by Danishefsky et al., according
to our work, (‡)-frondosin B is S-configured. Having ruled
out trivial mistakes, we were left with the puzzling possibility
that either Danishefsky×s or our material had undergone an
unnoticed inversion of configuration at some stage (see
below).
[15]
À
intramolecular C H insertion to afford the palladium(iv)
intermediate 22, which upon loss of a proton and reductive
elimination again gives (À)-6.
If mechanism A applies, the rearomatization of 19 may
possibly occur in a completely stereoselective fashion and
lead to inversion of the stereocenter at C8.[16] It appears,
however, that the absolute configuration was unintentionally
switched as a result of neighboring group participation in an
early stage of the previous synthesis.[17, 2b] We therefore
suggest that naturally occurring (‡)-frondosin B (1) corre-
sponds to the S isomer.[18]
HO
HO
O
O
For the key cyclization 15 !6, several mechanistic scenarios
can be imagined. Control experiments with various Lewis
acids (e.g. ZnCl2, MgBr2, BF3 ¥ OEt2) gave no product, which
suggests that the reaction does not proceed by means of
conjugate addition/elimination. Nucleophilic catalysis by
Me
Me
(S)-(+)-frondosin B
(R)-(–)-frondosin B
natural
synthetic
1570
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4109-1570 $ 20.00+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 9

COMMUNICATIONS
MeO
MeO
MeO
OTf
[Pd]
O
O
[Pd]OTf
Me
– H[Pd]OTf
O
A
B
C
arom.
H
H
rac-6
Me
H
Me
O
O
O
17
21
19
MeO
MeO
MeO
TfO
O
[Pd]
[Pd]
[Pd]
Pd0
O
O
– Pd0
– H+
O
H
(–)-6
15
Me
Me
Me
O
O
20
21
17
MeO
MeO
H
[Pd]
TfO
H
[Pd]
– Pd0
O
O
– H+
(–)-6
18
Me
Me
O
O
22
17
Scheme 4. Mechanism of the key cyclization.
115.8, 111.7, 111.5, 105.9, 56.3, 37.9, 36.1, 35.1, 30.9, 23.7, 21.4, 20.6; HRMS
(EI) m/z calcd. for C19H20O3, 296.1412; found, 296.1414. The enantiomeric
excess was determined by HPLC [Chiralpak ADTM column, 2% iPrOH in
hexanes at 1.0 mLminÀ1; tR (major enantiomer) ˆ 25.8 min, tR (minor
enantiomer) ˆ 33.8 min] to be 91% (e.r. ˆ 95.5:4.5).
In summary, we have achieved an expedient, asymmetric
total synthesis of (À)-frondosin B (1) from simple starting
materials (22% overall yield from the known alkyne 7). The
salient features of the synthesis include two palladium-
catalyzed reactions: an efficient Sonogashira coupling with
an aryl bromide and a novel cyclization onto a benzofuran.
Furthermore, the usefulness of the Reetz protocol for
installing a gem-dimethyl group in a natural product has been
demonstrated. Studies toward the synthesis of other members
of the frondosin family are underway and will be reported in
due course.
Received: October 18, 2001
Revised: February 11, 2002 [Z18082]
[1] a) A. D. Patil, A. J. Freyer, L. Killmer, P. Offen, B. Carte, A. J.
Jurewicz, R. K. Johnson, Tetrahedron 1997, 53, 5047 5060; b) Y. F.
Hallock, J. H. Cardellina, M. R. Boyd, Nat. Prod. Lett. 1998, 11, 153
160.
[2] a) S. J. Danishefsky, M. Inoue, A. Frontier, Angew. Chem. 2000, 112,
777 780; Angew. Chem. Int. Ed. 2000, 39, 761 764; b) S. J. Dani-
shefsky, M. Inoue, M. W. Carson, A. J. Frontier, J. Am. Chem. Soc.
2001, 123, 1878 1889.
[3] A. Arcadi, S. Cacchi, M. Del Rosario, G. Fabrizi, F. Marinelli, J. Org.
Chem. 1996, 61, 9280 9288.
Experimental Section
To a degassed solution of iPr2NEt (0.78 mL, 4.5 mmol) in dry DMA
(110 mL) was added [Pd(PPh3)4] (65 mg, 0.056 mmol), and the mixture was
heated under Ar to 908C. A solution of triflate 15 (500 mg, 1.12mmol) in
dry DMA (5 mL) was then added with a syringe pump over 3 h. The
reaction mixture was stirred at 908C for an additional 36 h, cooled, diluted
with saturated aqueous bicarbonate solution, and extracted four times with
Et2O (200 mL). The combined Et2O extracts were washed twice with water
(100 mL), once with brine (50 mL), dried over MgSO4, filtered, and
concentrated in vacuo. The crude product was purified by column
chromatography (15% EtOAc in hexanes) to afford 234 mg (70%) of 6
as a viscous oil: [a]2D5 ˆ À36.1 (c ˆ 1.00, CHCl3); IR (film): nÄ ˆ 2934, 1656,
1474 cmÀ1; 1H NMR (400 MHz, CDCl3, 2 58C): d ˆ 7.33 (d, J ˆ 8.8 Hz, 1H),
7.16 (d, J ˆ 2.4 Hz, 1H), 6.87 (dd, J ˆ 8.8 Hz, 2.4 Hz, 1H), 3.85 (s, 3H), 3.28
(m, 1H), 3.06 2.95 (m, 2H), 2.80 (m, 1H), 2.59 2.47 (m, 2H), 2.35 (m,
1H), 2.17 2.04 (m, 3H), 1.55 (m, 1H), 1.38 (d, J ˆ 6.8 Hz, 3H); 13C NMR
(75 MHz, CDCl3, 2 58C): d ˆ 197.6, 164.5, 156.0, 149.3, 149.1, 137.2, 128.3,
[4] A. Murai, T. Oka, Tetrahedron 1998, 54, 1 2 0.
[5] K. M. Johnston, J. Chem. Soc. 1969, 1424 1427. This bromide was
efficiently synthesized by treating 4-methoxyphenol with tetrabutyl-
ammonium tribromide as detailed in D. P. Curran, J. Xu, J. Am. Chem.
Soc. 1996, 118, 3142 3147 (supporting information). The bromophe-
nol was subsequently treated with acetyl chloride in pyridine to
furnish substrate 8 in 77% overall yield.
[6] I. J. Campbell in Organocopper Reagents: A Practical Approach (Ed.:
R. J. K. Taylor), Oxford University Press, Oxford, 1994, p. 217.
[7] D. Kahne, L. Yan, Syn. Lett. 1995, 523 524.
[8] a) E. Piers, J. R. Grierson, J. Org. Chem. 1977, 42, 3755 3757; b) S.
Laschat, F. Narjes, L. E. Overman, Tetrahedron 1994, 50, 347 358.
Angew. Chem. Int. Ed. 2002, 41, No. 9
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4109-1571 $ 20.00+.50/0
1571

COMMUNICATIONS
[9] All attempts to effect a Friedel Crafts type cyclization of the
sensitive diketone 14 with polyphosphoric acid were unsuccessful.
[10] Interestingly, no product was observed for intermolecular versions of
this reaction, see for example Equation (1).
Mechanism and Origin of Stereoselectivity in
Lewis Acid Catalyzed [2 ‡ 2] Cycloadditions of
Ketenes with Aldehydes**
Daniel A. Singleton,* Yingcai Wang, Hong Woon Yang,
and Daniel Romo*
MeO
MeO
[Pd(PPh3)4]
Stereoselectivity as a general phenomenon arises from
differences in free energy associated with diastereomeric
transition states or diastereomeric products. Free-energy
differences between diastereomeric pathways are a require-
ment for stereoselectivity, but they are not sufficient in
themselves. For example, high kinetic enantioselectivity may
be compromised by product racemization. Another common
situation is when stereoisomeric products are formed by
differing mechanisms, so the inhibition of alternative mech-
anisms is often pivotal in the development of stereoselective
methodology. A more complex possibility is that stereo-
isomeric pathways may have different rate-limiting steps.[1]
We describe here evidence for such an event in a cyclo-
addition in which diastereomeric products are formed, and we
discuss the impact of divergent rate-limiting steps on selec-
tivity.
b-Lactones are exceptionally versatile intermediates in
organic synthesis,[2] and there has been considerable interest
in their direct synthesis from the [2 ‡ 2] cycloaddition of
ketenes with carbonyl compounds [Eq. (1)] .[3] A growing
number of variants employ achiral[4] or chiral[5] Lewis acids
catalysts to afford optically enriched b-lactones by substrate
or reagent control, respectively.
OTf
O
O
+
(1)
X
iPr2NEt
Me
Me
O
O
PMBO
PMBO
[11] At elevated temperature (1108C) partial racemization was observed
to afford the cyclized product (À)-6 in 86% ee.
[12] M. T. Reetz, J. Westermann, R. Steinbach, J. Chem. Soc. Chem.
Commun. 1981, 237 239.
[13] a) M. Burwood, B. Davies, I. Diaz, R. Grigg, P. Molina, V. Sridharan,
M. Hughes, Tetrahedron Lett. 1995, 36, 9053 9056; b) M. S. McClure,
B. Glover, E. McSorley, A. Millar, M. H. Osterhout, F. Roschangar,
Org. Lett. 2001, 3, 1677 1680.
[14] For an exhaustive review of the Heck reaction, see: I. P. Beletskaya,
A. V. Cheprakov, Chem. Rev. 2000, 100, 3009 3066, and references
therein.
[15] C. Jia, D. Piao, J. Oyamada, W. Lu, Y. Fujiwara, Science 2000, 287,
1992 1995.
[16] Complete inversion would imply a highly stereoselective protonation
of enol ether 16 since a suprafacial [1,3]-hydrogen shift is thermally
not allowed.
[17] The most likely candidate for such a switch is the nucleophilic opening
of an epoxy alcohol with trimethylaluminum, which probably
proceeded with unwanted retention of configuration through partic-
ipation of the ester carbonyl group [Eq. (2)]. See also footnote 42 in
reference [2b]. In fact, a case of retention of configuration is already
documented in the original publication on the ring opening of epoxy
alcohols with organoaluminum compounds: T. Suzuki, H. Saimoto, H.
Tomioka, K. Oshima, H. Nozaki, Tetrahedron Lett. 1982, 23, 3597
3600.
O
O
O
O
(1)
+
•
The synthetic utility and fundamental allure of these
Al
O
formally forbidden cycloadditions make their mechanism of
considerable interest. Experimental studies have been limited
to substituent, solvent, and stereochemical effects in the
uncatalyzed reaction,[6] but several theoretical studies have
O
O
AlMe3
HO
COOMe
O
CH2Cl2
OMe
(2)
[*] Prof. D. A. Singleton, Prof. D. Romo, Y. Wang
Department of Chemistry
OH
OH
Texas A & M University
HO
HO
P.O. Box 30012, College Station, TX 77842-3012 (USA)
Fax : (‡1)979-845-5719
COOMe
+
O
Me
E-mail: singleton@mail.chem.tamu.edu
romo@mail.chem.tamu.edu
MeO
Dr. H. W. Yang
Department of Chemistry
Johns Hopkins University
3400 North Charles Street, Baltimore, MD 21218 (USA)
[18] Attempts to assign the absolute configuration of frondosin B inde-
pendently by X-ray crystallographic analysis of its bromobenzoate and
camphanoyl ester have thus far been unsuccessful.
[**] We thank the NIH (GM-45617, D.A.S.), the NSF (CAREER Award to
D.R., CHE-9624532), and the Robert A. Welch Foundation (A-1145,
D.A.S.; A-1280, D.R.) for support, and the NSF (CHE-9528196) and
the Texas A&M University Supercomputing Facility for computa-
tional resources. D.R. is an Alfred P. Sloan Fellow and a Camille-
Henry Dreyfus Teacher-Scholar. The NSF (CHE-0077917) also
provided funds for purchase of NMR instrumentation.
Supporting information for this article is available on the WWW under
http://www.angewandte.com or from the author.
1572
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4109-1572 $ 20.00+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 9