Total synthesis of (+)-frondosin A. Application of the RU-catalyzed [5+2] cycloaddition

Article abstract of DOI:10.1021/ja073272b

The first total synthesis of (+)-frondosin A was accomplished in 19 longest linear and 21 total steps from commercially available materials. The key features of the synthesis include a Ru-catalyzed [5+2] cycloaddition, a Claisen rearrangement, and a ring expansion to construct the core of the frondosin A in a diastereoselective and regioselective fashion. This is the first application of a Ru-catalyzed [5+2] cycloaddition in the total synthesis of a natural product. Through this synthesis, the absolute configuration of (+)-frondosin A was established.

Full text of DOI:10.1021/ja073272b

Published on Web 08/31/2007
Total Synthesis of (+)-Frondosin A. Application of the
Ru-Catalyzed [5+2] Cycloaddition
Barry M. Trost,* Yimin Hu, and Daniel B. Horne
Contribution from the Department of Chemistry, Stanford UniVersity,
Stanford, California 94305-5080
Received May 8, 2007; E-mail: bmtrost@stanford.edu
Abstract: The first total synthesis of (+)-frondosin A was accomplished in 19 longest linear and 21 total
steps from commercially available materials. The key features of the synthesis include a Ru-catalyzed
[
5+2] cycloaddition, a Claisen rearrangement, and a ring expansion to construct the core of the frondosin
A in a diastereoselective and regioselective fashion. This is the first application of a Ru-catalyzed [5+2]
cycloaddition in the total synthesis of a natural product. Through this synthesis, the absolute configuration
of (+)-frondosin A was established.
Introduction
potent compound in its class, frondosin A, for which there are
no reported syntheses, was of particular interest to us as a
Frondosin A (1) is a member of a family of five norsesquit-
erpenoid (14-carbon) natural products, frondosins A-E, which
were originally extracted from the marine sponge Dysidea
frondosa, collected in Pohnpei, Federated States of Micronesia.
Researchers at the National Cancer Institute (NCI) later isolated
synthetic target. The proposed structure of frondosin A rests
upon spectroscopic data that were used to assign the relative
stereochemistry, although the absolute configuration remains
unknown. The reported syntheses for other frondosins started
from the assembly of the C and D rings and constructed the A
and B ring systems at a late stage. For frondosin A, we
1
(-)-frondosins A and D from the HIV-inhibitory extract of
2
Euryspongia sp. Interestingly, the frondosins isolated at the
NCI were of optical rotation different from and opposite those
of the frondosins first isolated, so both enantiomers of these
natural products appear to be produced by different species of
sponge, and both are biologically active.
4
5
envisioned that the Rh-catalyzed and Ru-catalyzed [5+2]
cycloaddition would provide an opportunity to quickly access
the bicyclo[5.4.0] structure corresponding to the A-B ring
system of the norsesquiterpene carbon framework present in
frondosin A. This strategy would allow installation of what we
anticipated to be a sensitive exocyclic olefin and the readily
oxidizable hydroquinone at a late stage of the synthesis. In
addition to confirming the structural assignment, an enantiose-
lective synthesis would also establish the absolute stereochem-
istry of frondosin A.
The frondosins inhibit the binding of interleukin-8 (IL-8), a
chemokine which is produced by macrophages, fibroblasts, and
epithelial and endothelial cells, and which has been implicated
in a wide range of acute and chronic inflammatory disorders,
including psoriasis and rheumatoid arthritis. IL-8 has been
identified as a principal factor directing neutrophil (white blood
cell) recruitment to the inflammatory focus, and as such the
IL-8 receptor antagonist is a promising target for development
of novel anti-inflammatory agents. Frondosin A was found to
be the most potent of the frondosins, with IL-8R, IL-8â, and
protein kinase C (PKC) activites all in the low micromolar range.
In addition, the frondosins have been shown to exhibit anti-
HIV activity.
Results and Discussion
The first-generation synthetic plan, based on the Ru-catalyzed
[5+2] cycloaddition as the key step, is summarized in Scheme
1
. The aromatic functionality in frondosin A could be incor-
porated using a cross-coupling reaction, followed by a Claisen
rearrangement. The advanced intermediate 7, in which the
bicyclo[5.4.0] core structure has been constructed, could derive
from a 1,7-enyne (8) utilizing the Ru-catalyzed [5+2] cycload-
dition. The cyclopropyl functionality in compound 8 could be
To date there have been several reported syntheses of
frondosin B and one of frondosin C. As the most biologically
3
(
(
(
1) Patil, A. D.; Freyer, A. J.; Killmer, L.; Offen, P.; Carte, B.; Jurewicz, A.
(4) (a) Wender, P. A.; Takahashi, H.; Witulski, B. J. Am. Chem. Soc. 1995,
112, 4720. (b) Wender, P. A.; Husfeld, C. O.; Langkopf, E.; Love, J. A.;
Pleuss, N.; Tetrahedron 1998, 54, 7203. (c) Wender, P. A.; Dyckman, A.
J.; Husfeld, C. O.; Kadereit, D.; Love, J. A.; Rieck, H. J. Am. Chem. Soc.
1999, 121, 10442. (d) Wender, P. A.; Haustedt, L. O.; Lim, J.; Love, J. A.;
Williams, T. J.; Yoon, J.-Y. J. Am. Chem. Soc. 2006, 128, 6302.
(5) (a) Trost, B. M.; Shen, H. C.; Horne, D. B.; Toste, F. D.; Steinmetz, B. G.;
Koradin, C. Chem.sEur. J. 2005, 11, 2577. (b) Trost, B. M.; Shen, H. C.;
Schulz, T.; Koradin, C.; Schirok, H. Org. Lett. 2003, 5, 4149. (c) Trost, B.
M.; Shen, H. C. Angew. Chem., Int. Ed. 2001, 40, 2313. (d) Trost, B. M.;
Toste, F. D.; Shen, H. J. Am. Chem. Soc. 2000, 122, 2379. (e) Trost, B.
M.; Shen. H. Org. Lett. 2000, 2, 2523.
J.; Johnson, R. K. Tetrahedron 1997, 53, 5047.
2) Hallock, Y. F.; Cardellina, J. H.; Boyd, M. R. Nat. Prod. Lett. 1998, 11,
1
53.
3) (a) Inoue, M.; Carson, M. W.; Frontier, A. J.; Danishefsky, S. J. J. Am.
Chem. Soc. 2001, 123, 1878. (b) Inoue, M.; Frontier, A. J.; Danishefsky,
S. J. Angew. Chem., Int. Ed. 2000, 39, 761. (c) Hughes, C. C.; Trauner, D.
Angew. Chem., Int. Ed. 2002, 41, 1569. (d) Hughes, C. C.; Trauner, D.
Tetrahedron 2004, 60, 9675. (e) Chaplin, J. H.; Flynn, B. L. Chem.
Commun. 2001, 1594. (f) Martinez, I.; Alford, P. E.; Ovaska, T. V. Org.
Lett. 2005, 7, 1133. (g) Li, X.; Kyne, R. E.; Ovaska, T. V. Org. Lett. 2006,
8
, 5153.
10.1021/ja073272b CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 11781-11790
9
11781

A R T I C L E S
Trost et al.
Figure 1. Frondosins A-E.
Scheme 1. Retrosynthesis of Frondosin A by the Ru-Catalyzed [5+2] Cycloaddition
Scheme 2. Synthesis of 1,7-Enyne Model System 17
installed employing an asymmetric cyclopropanation. Dienyne
could be synthesized from compounds 10 and 11 via an allylic
alkylation reaction.
a Pd-catalyzed allylic alkylation to afford dienyne 16. Chemose-
lective cyclopropanation of the more electron rich olefin in
compound 16 gave the desired model substrate 17.
9
One goal of this route was to examine the feasibility of using
,7-enynes in Ru-catalyzed [5+2] cycloadditions. The use of
Unfortunately, when enyne substrate 17 was subjected to the
standard [5+2] cyclization conditions (10% [CpRu(CH3CN)3]-
PF6, acetone, room temperature), none of the desired product
18 could be detected. Attempts to facilitate the reaction by
heating and the examination of different solvents (acetone, DMF,
CH2Cl2) led only to no reaction or to decomposition at higher
temperatures.
1
1
[
,6-enynes in Ru-catalyzed [5+2] reactions to access bicyclo-
5
5.3.0] ring systems was previously described by our group.
Only one example of a 1,7-enyne participating in a [5+2]
cycloaddition exists, an amino-tethered system, wherein a [5+2]
5
d
reaction occurred using either Ru or Rh catalysts (eq 1).
It was at this point that we decided to revise our synthetic
strategy to utilize a 1,6-enyne substrate in the [5+2] cycload-
dition. A retrosynthesis based on the new strategy is shown in
Scheme 3. The key intermediate 19, which bears a bicyclo-
[5.4.0] ring system, could derive from compound 20, which has
a bicyclo[5.3.0] ring system. A 1,6-enyne (21) could be
employed to build the bicyclo[5.3.0] ring system via a Ru-
catalyzed [5+2] cycloaddition. The 1,6-enyne 21 could be
obtained by metalation of vinyl iodide 23 and addition to alkynyl
epoxide 22.
To test whether other 1,7-enynes would participate in a [5+2]
reaction, a simplified model system (17) was prepared as shown
in Scheme 2. Alkyne 14 was reacted with allyl acetate 15 in
6
4c
The synthesis of alkynyl epoxide 22 was achieved as
described in Scheme 4. From commercially available 2-methyl-
(6) Taylor, E. C.; Macor, J. E.; French, L. G. J. Org. Chem. 1991, 56, 1807.
1
1782 J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007
9

Total Synthesis of (
+
)-Frondosin A
A R T I C L E S
Scheme 3. Second-Generation Retrosynthesis for Frondosin A
Scheme 4. Synthesis of Alkynyl Epoxide 22
Scheme 5. Synthesis of Vinyl Iodide 23
3
-butyn-2-ol (24), bromination with phosphorus tribromide
Synthesis of vinyl iodide 23 was accomplished by cyclopro-
panation of allylic alcohol 31 (Scheme 5). The preparation of
allylic alcohol 31 was adopted from the procedure of Charette
afforded a propargyl bromide. An aluminum-mediated addition
to formaldehyde yielded the desired homopropargyl alcohol 25.
Formation of the (dimethylphenylsilyl)alkyne was performed
to reduce the volatility of intermediates in subsequent steps as
well as protection of the alkyne for the subsequent addition
reaction to the epoxide. A Moffatt-Swern oxidation of the
resulting primary alcohol gave the aldehyde 26, which was then
converted to epoxide 22 through the reaction with (chlorom-
ethyl)lithium.7
8
with a similar substrate. The enantioselective cyclopropanation
of alcohol 31 was achieved with chiral oxaborolane ligand 32.9
Alcohol 33 was assigned to be of R,R-configuration on the basis
of Charette’s model for asymmetric cyclopropanation of allylic
alcohols. This assignment was later confirmed by synthesis of
a compound with a known optical rotation (Scheme 16). Alcohol
(
8) Charette, A. B.; Cote, B.; Monroc, S.; Prescott, S. J. Org. Chem. 1995, 60,
888.
(9) Charette, A. B.; Juteau, H. J. Am. Chem. Soc. 1994, 116, 2651.
6
(
7) Sadhu, K. M.; Matteson, D. S. Tetrahedron Lett. 1986, 27, 795.
J. AM. CHEM. SOC.
9
VOL. 129, NO. 38, 2007 11783

A R T I C L E S
Trost et al.
Scheme 6. Completion of the Bicyclo[5.3.0] Ring
Scheme 7. Third-Generation Retrosynthesis for Frondosin A
Scheme 8. Preparation of [5+2] Cycloaddition Precursor 39
3
3 was then oxidized by the Moffatt-Swern reaction, and the
alcohol 21, for the Ru-catalyzed [5+2] cycloaddition. The
cycloaddition product 38 could be converted to compound 37
through a sequence of steps involving the Claisen rearrangement.
The position of the carbonyl functionality in compound 37 was
anticipated to facilitate a base-mediated double bond migration
into conjugation with the ketone.
resulting aldehyde 34 was transformed into vinyl iodide 23 by
the Takai conditions.10
The coupling of the vinyl iodide 23 and the epoxide 22 was
achieved by a copper-mediated process (Scheme 6). The
deprotection of the alkyne in 35 afforded the cyclization
substrate 21. Finally, the standard ruthenium [5+2] conditions
were applied to the enyne 21 to successfully yield the cyclized
product 20.
It is at this stage that several problems were realized. The
Ru-catalyzed [5+2] cycloaddition reaction was problematic. In
addition to the high catalyst loadings required to achieve modest
conversions, one of the alcohol diastereomers is much less
reactive than the other. More importantly, initial experiments
suggested that several difficulties for the ring expansion and
double bond migration in later stages of the synthesis would
present themselves.
Enyne 39 was prepared through a sequence of steps shown
1
1
in Scheme 8. Aldehyde 40 was prepared in two steps from
commercially available 3-methylbutyne (41). Vinyl iodide 23
was then coupled with aldehyde 40, employing n-butyllithium
in the lithium-halogen exchange and addition reaction. Com-
pound 43 was used without purification in the subsequent
deprotection reaction to afford cyclopropyl enyne 39 as the
precursor for the key [5+2] cycloaddition. Compound 39 is an
approximately 1:1 mixture of two diatereomers with respect to
the newly formed stereogenic center.
To address the problems above, a third strategy was proposed
Scheme 7). Compared to the plan in Scheme 3, the major
change is the use of allylic alcohol 39, instead of homoallylic
(
10) Takai, K.; Nitta, K.; Utimoto, K. J. Am. Chem. Soc. 1986, 108, 7408.
11) Marco-Contelles, J.; Ruiz-Caro, J.; Mainetti, E.; Devin, P.; Fensterbank,
L.; Malacria, M. Tetrahedron 2002, 58, 1147.
(
(
11784 J. AM. CHEM. SOC.
9
VOL. 129, NO. 38, 2007

Total Synthesis of (
+
)-Frondosin A
A R T I C L E S
Scheme 9. Unexpected Diene Formation under the [5+2] Cycloaddition Condition
Scheme 10. Different Reactivities in the [5+2] Cycloaddition Reactions with 39a and 39b
The stage was now set to attempt the [5+2] cycloaddition
reaction. Under standard conditions, no desired products were
observed. Instead, dienyne 44 was isolated as the major product
With strictly dried acetone as the solvent and carefully
purified ruthenium catalyst and by stopping the reaction before
completion by filtering off the catalyst through a silica gel pad,
a small amount of the desired adduct 38 was obtained. Some
recovered starting material 39, with a diastereomeric ratio
different from that of the original starting material, was also
isolated. This demonstrated that one diastereoisomer of the
mixture was consumed faster than the other one in the reaction,
a result similar to that seen in the earlier model system (Scheme
6). Encouraged by this observation, the two diastereoisomers
of 39 were separated by flash chromatography and subjected
to the standard ruthenium-catalyzed [5+2] reaction conditions.
Very different results were found (Scheme 10). A good yield
(Scheme 9). The geometry of the double bonds was determined
by an NOE experiment. It is also interesting to note that
compound 44 seems to be a single enantiomer since derivatizing
this compound with (R)- or (S)-R-methoxyphenylacetic acid
afforded one diastereoisomer in each case. The absolute
stereochemistry of the stereogenic center in compound 44 was
1
determined by comparing the H NMR data of the two
diastereomeric esters.12 The mechanism for the formation of
this diene 44 was proposed as a solvolysis process. The cationic
ruthenium catalyst could complex to the hydroxyl group to
activate the vinylcyclopropane functionality in intermediate 45.
Trace amounts of water present in the solvent may well facilitate
the ring opening to yield diene 44 with an inversion of the
stereochemistry. The same stereochemical outcome was ob-
(80%) of desired product 38a was obtained with compound 39a,
and only a minor amount of the side product 44 was observed.
However, with the other diastereoisomer (39b), the undesired
side product 44 was obtained as the major product (42%). By
changing the solvent to methylene chloride, which is much easier
to dry completely, we were able to minimize the formation of
the undesired elimination product 44. Consistent results were
obtained with both diastereoisomers, resulting in good yields
of the desired cyclization products 38a (88%) and 38b (60%).
13
served in a similar acid-catalyzed solvolysis of cyclopropanes.
(
12) Trost, B. M.; Belletire, J. L.; Godleski, S.; McDougal, P.; Balkovec, J. M.
J. Org. Chem. 1986, 51, 2370.
(
13) (a) Honda, M.; Mita, T.; Nishizawa, T.; Sano, T.; Segi, M.; Nakajima, T.
Tetrahedron Lett. 2006, 47, 5751. (b) Honda, M.; Yamamoto, Y.; Tsuchida,
H.; Segi, M.; Nakajima, T. Tetrahedron Lett. 2005, 46, 6465.
J. AM. CHEM. SOC.
9
VOL. 129, NO. 38, 2007 11785

A R T I C L E S
Trost et al.
Scheme 11. Determination of the Absolute Configuration for 39a and 39b
Scheme 12. Proposed Mechanism of the Ru-Catalyzed [5+2] Cycloaddition
The absolute configuration of compounds 39a and 39b was
determined by asymmetric reduction of ketone 46 (obtained
from the Dess-Martin oxidation of compound 39) with (R)- or
conformation, as shown in intermediates 47 or 48, after
complexation with the ruthenium catalyst to afford ruthenacy-
clopentanes 49 and 50. To react with ruthenium, the cyclopro-
pane has to adopt a conformation in which the H-C7-C8-
CH2OTIPS dihedral angle is close to 0°. The steric interaction
between the methyl group at the C10 position and the Cp ligand
on the ruthenium in intermediate 50 disfavors this pathway for
the formation of compound 51. Thus, compound 38 was isolated
as a single diastereoisomer.
(
S)-CBS (Corey-Bakshi-Shibata) oxazaborolidine catalysts
1
4
(Scheme 11). This procedure afforded either diastereoisomer
of 39 with a predictable stereochemical outcome. It was shown
that diastereomer 39a has the higher reactivity toward cycload-
dition in the Ru-catalyzed [5+2] process.
The relative stereochemistry of compounds 38a and 38b was
1
assigned on the basis of the coupling constants in H NMR
This mechanism also explains the different reactivities
studies and COSY experiments. The relative stereochemistry
of compound 38a was further confirmed with NOE experiments,
as shown in the Experimental Section. The proposed mechanism
of the Ru-catalyzed [5+2] cycloaddition is anticipated to account
for the observed diastereoselectivity and regioselectivity, as
depicted in Scheme 12. Enyne 39 will react via an extended
observed with diastereomers 39a and 39b. As shown in Scheme
13, the electron-withdrawing hydroxyl group at the pseudoequa-
torial position in intermediate 47b will lower the reactivity of
the double bond toward ruthenium and destabilize the transition
state. This effect can be explained by the overlap of the σ*CO
orbital with the alkene π orbital, which is absent in transition
state 47a. The observation is in agreement with the Stork/Houk-
(
14) (a) Corey, E. J.; Helal, J. C. Angew. Chem., Int. Ed. 1998, 37, 1986. (b)
1
5
Corey, E. J.; Bakshi, R. K.; Shibata, S. J. Am. Chem. Soc. 1987, 109, 5551.
J a¨ ger “inside alkoxy” model.
11786 J. AM. CHEM. SOC.
9
VOL. 129, NO. 38, 2007

Total Synthesis of (
+
)-Frondosin A
A R T I C L E S
Scheme 13. Explanation for Different Reactivities of 39a and 39b in the Cycloaddition
Scheme 14. Preparation of Substrate 37 for the Ring Expansion Reaction
With cycloaddition products 38a and 38b in hand, further
elaboration was necessary to obtain the bicyclo[5.4.0] core
structure of frondosin A. For ease of reaction analysis and
product characterization, compounds 38a and 38b were handled
separately, but under similar reaction conditions. In principle,
the diastereomers of 38 could be employed in the reaction
sequence without prior separation.
Starting from compound 38, removal of the triisopropylsilyl
protecting group was achieved employing TBAF buffered with
acetic acid (Scheme 14). The p-methoxyphenyl group was
selectively installed on the more reactive allylic, and also
primary, hydroxyl group under Mitsunobu conditions to yield
compound 53. Claisen rearrangement under thermal conditions
diazoalkanes being the main strategies. However, due to the
presence of two activated protons at the C10 and C11 positions,
decomposition or formation of undesired products was observed
in many of our ring expansion attempts. Finally, under carefully
controlled conditions, with (TMS)CHN2 and catalytic BF3‚OEt2,
a mixture of compounds 55, 56, and 57 was obtained (Scheme
15). After desilylation with TBAF, the desired ketone 36 was
isolated in 54% yield, along with its regioisomer ketone 58 and
aldehyde 59. Double bond migration was also achieved in the
same transformation. Thus, the bicyclo[5.4.0] core structure of
frondosin A was completed, and efforts to finish the synthesis
were undertaken.
Finishing the synthesis of frondosin A entailed the reduction
of the ketone to a methylene and deprotection of the methyl
ethers. Direct reduction of a carbonyl functionality to the
corresponding methylene group normally involves a Wolff-
Kishner-type reduction. However, in our hands treating ketone
(diethylaniline, 215 °C) proceeded, although it resulted in only
a poor yield of the desired product. It was later found that the
product is sensitive to acid, base, and an oxidative atmosphere.
Thus, a protocol consisting of the Claisen rearrangement and
methylation of the resulting phenol in one operation was
employed to afford compound 54 as a single diastereomer in
good yield. The stereochemistry of compound 51 was confirmed
by NMR studies. The conformation of the bicyclic ring system
contributes to the desired stereoselectivity observed in the
reaction. Dess-Martin oxidation of compound 54 proceeded
smoothly to afford ketone 37.
1
6
36 under standard Wolff-Kishner conditions or Myers’s
1
7
modified conditions resulted in either decomposition or
formation of products lacking the exocyclic double bond. An
alternative strategy involving the reduction of arylsulfonyl
hydrazones with various boron hydrides resulted in complex
1
8
mixtures. The presence of the activated benzylic proton in
ketone 36 may account for the above difficulties.
To finish the bicyclo[5.4.0] core structure of frondosin A, a
ring expansion was required. There were several strategies that
were attempted for the ring expansion, cyclopropanation of the
corresponding silyl enol ether of ketone 37 followed by
fragmentation and Lewis acid-catalyzed ring expansion with
Another alternative for the deoxygenation of ketones involves
making sulfur derivatives, which can then be reduced to give
(
16) (a) Kishner, N. Zh. Russ. Fiz.-Khim. OVa., Chast Khim. 1911, 43, 582. (b)
Wolff, L. Justus Liebigs Ann. Chem. 1912, 394, 86. (c) Huang-Minlon. J.
Am. Chem. Soc. 1946, 68, 2487 (d) Huang-Minlon J. Am. Chem. Soc. 1949,
71, 3301.
(
17) Furrow, M. E.; Myers, A. G. J. Am. Chem. Soc. 2004, 126, 5436.
(
15) (a) Haller, J.; Strassner, T.; Houk, K. N. J. Am. Chem. Soc. 1997, 119,
031. (b) Houk, K. N.; Moses, S. R.; Wu, Y.-D.; Rondan, N. G.; J a¨ ger,
V.; Schohe, R.; Fronczek, F. R. J. Am. Chem. Soc. 1984, 106, 3880.
(18) (a) Caglioti, L.; Magi, M. Tetrahedron 1963, 19, 1127. (b) Hutchins, R.
O.; Milewski, C. A.; Maryanoff, B. E. J. Am. Chem. Soc. 1973, 95, 3662.
(c) Kabalka, G. W.; Baker, J. D. J. Org. Chem. 1975, 40, 1834.
8
J. AM. CHEM. SOC.
9
VOL. 129, NO. 38, 2007 11787

A R T I C L E S
Trost et al.
Scheme 15. Ring Expansion with Ketone 37
Scheme 16. Completion of the Total Synthesis of Frondosin A
Scheme 17. Absolute Configuration of (+)-Frondosin A
the alkane. Attempts to make a thioketal (60) with 1,2-
ethanedithiol, catalyzed by Lewis acids, afforded products
lacking the exocyclic double bond. Evans reported 1,2-bis-
Raney Ni, the desired deoxygenated product 61 could be isolated
in over 80% yield.
Initial attempts for the final deprotection with BBr3 resulted
in decomposition. Treating compound 61 with sodium ethaneth-
iolate in DMF at reflux only afforded monodeprotection
products. The most effective method for the deprotection proved
to be an oxidation-reduction sequence using ceric ammonium
nitrate (CAN) followed by sodium dithionite. Because of the
sensitivity of the system toward oxidation, it was necessary to
perform the reaction only to partial conversion. Thus, an 89%
yield was obtained when the reaction was stopped at 55%
conversion.
[
(trimethylsilyl)thio]ethane and catalytic zinc iodide as mild
19
conditions for thioketal formation. Applying these conditions
with ketone 36 resulted in slow conversion to the thioketal along
with an undesired isomerization of the exocyclic olefin. Real-
izing that zinc iodide might initiate certain radical processes
facilitating the isomerization, different Lewis acids were
screened. It was found subsequently that trimethylsilyl triflate
could catalyze the reaction cleanly to give the desired thioketal
60 (Scheme 16). Upon treatment of the crude product 60 with
The optical rotation of the synthetic (+)-frondosin A matches
the reported value (Scheme 17). Because the absolute stereo-
(
19) Evans, D. A.; Truesdale, L. K.; Grimm, K. G.; Nesbitt, S. L. J. Am. Chem.
1
Soc. 1977, 99, 5009.
11788 J. AM. CHEM. SOC.
9
VOL. 129, NO. 38, 2007

Total Synthesis of (
+
)-Frondosin A
A R T I C L E S
chemistry of the final natural product is derived from alcohol
3, its absolute configuration was verified by conversion to an
ester (63) of known absolute configuration (Scheme 16). Thus,
frondosin A must have an R-stereogenic center at the C8
position. This is in agreement with the original assignment of
solution of diiodomethane (7.2 g, 26.9 mmol) in CH
min. The milky white solution was stirred for 10 min, and allylic
alcohol 31 (1.75 g, 6.77 mmol) in CH Cl (5 mL) was added over 5
min. The reaction was stirred for 10 min, then the ice bath was removed,
and the reaction was warmed to room temperature and stirred for 1 h.
2 2
Cl (5 mL) over
5
3
20
2
2
The reaction was poured into saturated aqueous NH
diluted with CH Cl (50 mL), and extracted. The aqueous layer was
extracted with CH Cl
(2 × 50 mL), and the combined organic layers
were washed with brine, dried (MgSO ), and concentrated. Purification
4
Cl (100 mL),
3
a
(
+)-frondosin B by Danishefsky. Trauner later questioned this
2
2
3d
assignment. Although (+)-frondosin A and (+)-frondosin B
may not necessarily have the same configuration at the C8
position, our results with (+)-frondosin A suggest that Dan-
ishefsky’s original assignment is very likely to be the correct
one, which is also in agreement with recent work by MacMillan
2
2
4
by flash chromatography (5% EtOAc in petroleum ether) yielded 1.40
g of (()-33 (76%) as a clear, colorless oil. R ) 0.38 (20% Et O in
f
2
-
1
petroleum ether). IR (thin film, cm ): 3452, 2944, 2867, 1464, 1108,
2
1
1
based upon his group’s total synthesis of frondosin B.
1063, 1014. H NMR (500 MHz, CDCl
3
): δ 3.90 (d, J ) 11.5 Hz,
1
H), 3.77 (d, J ) 10.0 Hz, 1H), 3.62 (d, J ) 11.5 Hz, 1H), 3.54 (d, J
Conclusion
)
9.8 Hz, 1H), 3.11 (br s, 1H), 1.18 (d, J ) 6.3 Hz, 3H), 1.07 (m,
2
1H), 0.87 (m, 1H), 0.55 (dd, J ) 6.6, 4.9 Hz, 1H), 0.18 (t, J ) 5.2
In summary, the first enantioselective total synthesis of (+)-
frondosin A was accomplished in 7% overall yield, through 19
longest linear and 21 total steps. A Ru-catalyzed [5+2]
cycloaddition of an enantioenriched cyclopropyl enyne was
employed as the key step to construct the bicyclo[5.3.0] ring
system with high regio- and diastereoselectivity. This first
application of the Ru-catalyzed [5+2] cycloaddition in natural
product synthesis highlights the efficiency of this methodology
for generating complex fused ring systems. Subsequent Claisen
rearrangement further elaborated the cycloadduct to give the
frondosin core. This synthesis also serves to determine the
absolute configuration of (+)-frondosin A and shed light on
the configurational assignment of the frondosin family.
13
Hz, 1H) ppm. CNMR (125 MHz, CDCl
1
3
): δ 72.8, 66.2, 27.4, 17.9,
6.6, 15.8, 13.8, 11.7 ppm. HRMS (EI+): m/z calcd for C15
31 2
H O
Si
+
[
M - H] 271.2093, found 271.2092.
1R,2R)-2-Methyl-1-(hydroxymethyl)-1-[(triisopropylsilanyl)oxy]-
cyclopropane (33). To a solution of Et Zn (43 mL, 1.0 M in hexanes,
freshly prepared from neat Et Zn, 43 mmol) in DME (4.4 mL, 42.3
mmol) and CH Cl (90 mL) at -30 °C was added dropwise a solution
of diiodomethane (6.8 mL, 84.4 mmol) in CH Cl (20 mL) over 10
(
2
2
2
2
2
2
min. The clear solution was stirred for 30 min, then butylboronic acid
N,N,N′,N′-tetramethyl-L-tartaric acid diamide ester 32 (7.0 g, 25.9 mmol)
in CH
Cl
2 2
20 mL) was added over 10 min. After 15 min, allylic alcohol
Cl (20 mL) was added. The reaction
3
1 (5.5 g, 21.3 mmol) in CH
2
2
was allowed to warm to room temperature and stirred for 24 h. The
reaction was quenched with 1 N HCl (100 mL) and saturated aqueous
Experimental Section
NH
layer was extracted with diethyl ether (3 × 100 mL). The combined
organic layers were washed with water, saturated aqueous NaHCO
and brine, dried (MgSO ), and concentrated. Purification by flash
chromatography (silica gel, 4% EtOAc in petroleum ether) yielded
product 33 (6.0 g, quantitative yield) as a clear oil. The enantiomeric
excess of the product was determined to be 94.6% ee by conversion
4
Cl (100 mL). The organic layer was separated, and the aqueous
Selected experimental procedures for the preparation of 42, 33, 38a,
3
,
5
4a, 36, and 1 (frondosin A) appear below. Full experimental details
for all compounds are given in the Supporting Information.
-(Trimethylsilyl)-3,3-dimethyl-4-pentyn-1-ol (42). To a solution
4
5
of 3-methylbutyne (41; 8.06 g, 118 mmol) in diethyl ether (100 mL)
at -20 °C was added n-BuLi (110 mL, 2.19 M solution in hexanes,
into the trifluoroacetate ester and chiral GC, described below. R )
f
2
37 mmol) followed by tetramethylethylenediamine (TMEDA; 17.7
24
0
.38 (5:1 petroleum ether/Et
2
2 2
O). [R]D ) -27.4° (c ) 1.10, CH Cl ).
mL, 118 mmol). The reaction was warmed to room temperature and
heated to a gentle reflux for 10 h until a clear orange solution formed.
The reaction was recooled to -78 °C, and a solution of ethylene oxide
The spectral data matched those of the racemic compound (()-33
described above.
(1R,2R)-Trifluoroacetic Acid [2-Methyl-1-[[[(triisopropyl)silanyl]-
(
6 mL, condensed from cylinder, 118 mmol) in THF (20 mL) was
oxy]methyl]cyclopropyl]methyl Ester. The enantiomeric excess of the
above product was determined by conversion to the trifluoroacetate
ester of alcohol (-)-33: To a solution of (-)-33 (27 mg, 0.10 mmol)
in CH Cl (0.6 mL) under Ar at 0 °C was added pyridine (16 µL, 0.2
2 2
mmol) followed by trifluoroacetic anhydride (28 µL, 42 mg, 0.2 mmol),
cannulated into the solution of the above dianion over 5 min. The
reaction was stirred at -78 °C for 2 h, and then trimethylsilyl chloride
(30 mL, 240 mmol) was added dropwise. The reaction was allowed to
warm to room temperature and stirred for 5 h. The solution was
concentrated to about 200 mL under reduced pressure, 1 N HCl (200
mL) was added, and the mixture was stirred vigorously for 1 h. The
organic layer was separated, and the aqueous layer was extracted with
diethyl ether (2 × 100 mL). The combined organic layers were washed
and the solution was stirred for 1 h. The reaction was quenched with
H
2
O (5 mL) and diluted with diethyl ether (5 mL) and the organic
layer separated, washed with saturated aqueous CuSO , saturated
aqueous NaHCO , and saturated aqueous NaCl, dried (MgSO ), and
4
3
4
with saturated aqueous NaHCO
3
and NaCl, dried (MgSO
4
), and
O in
petroleum ether) yielded product 42 (11.70 g, 63.6 mmol, 53% yield)
concentrated. Purification by flash chromatography (5% EtOAc in
petroleum ether) yielded the trifluoroacetate ester of (-)-33 (28.7 mg,
concentrated. Purification by flash chromatography (20% Et
2
78%) as a clear, colorless, oil. The enantiomeric excess was determined
1
11
as a colorless liquid. H NMR and IR match the reported data. R )
f
to be 94.6% ee by chiral GC (Cyclosil B, 120 °C, 15 mL/min, minor
-
1
0
2
.17 (20% Et
162, 1451, 1409, 1364, 1251, 1062, 1029. 1H NMR (500 MHz,
): δ 3.83 (t, J ) 6.2 Hz, 2H), 1.68 (t, J ) 6.2 Hz, 2H), 1.21 (s,
H), 0.59 (s, 9H) ppm.
1,2-trans)-2-Methyl-1-(hydroxymethyl)-1-[(triisopropylsilanyl)-
oxy]cyclopropane [(()-33]. To a solution of Et Zn (13.5 mL, 1.0 M
in hexanes, 13.5.0 mmol) in CH Cl (10 mL) at 0 °C was added a
2
O in petroleum ether). IR (thin film, cm ): 3345, 2967,
product (1S,2S) t
R
) 47.838 min and major product (1R,2R) t ) 48.558
R
min).
CDCl
6
3
(
1R,3aE,6R,7E,8aS)-3,3,6-Trimethyl-7-[[(triisopropylsilanyl)oxy]-
methyl]-1,2,3,5,6,8a-hexahydroazulen-1-ol (38a). To a solution of 1,6-
enyne 39a (2.95 g, 7.8 mmol) in CH Cl (150 mL) at 0 °C was added
a solution of [CpRu(CH CN) ]PF (0.34 g, 0.78 mmol) in CH Cl (50
(
2
2
2
3
3
6
2
2
2
2
mL). The reaction was stirred for 3 h as the solution warmed to room
temperature, then the solution was filtered through a silica gel pad to
remove the catalyst, and the filtrates were concentrated. Flash chro-
(
20) di Lugano, F. R.; Monteiro, J.; Fliche, C.; Braun, J.; Le Goffic, F. Synth.
Commun. 1992, 22, 1155.
(
21) A recent asymmetric synthesis by MacMillan et al. also suggests that (+)-
frondosin B has an R-chiral center at the C8 position: D. W. C. MacMillan,
Princeton University, personal communication, 2007.
matography (5-10% Et
g, 6.86 mmol, 88% yield) as a colorless oil. R
2
O in petroleum ether) yielded product 38a (2.59
) 0.17 (10% Et O in
f
2
J. AM. CHEM. SOC.
9
VOL. 129, NO. 38, 2007 11789

A R T I C L E S
Trost et al.
petroleum ether). IR (thin film, cm-1): 3375, 2945, 2867, 1464, 1383,
4
organic layers were washed with brine, dried (MgSO ), and concentrated
to give the crude product as a pale yellow liquid.
1
1
368, 1248, 1064. H NMR (500 MHz, CDCl
3
): δ 5.70 (d, J ) 1.5
Hz, 1H), 5.50 (dt, J ) 7.0, 3.4 Hz, 1H), 4.18 (s, 1H), 4.08 (m, 2H),
The crude product was then dissolved in acetonitrile (10 mL), and
TBAF trihydrate (0.63 g, 2.0 mmol) was added at room temperature.
The reaction was stirred for 4 h and then diluted with ethyl acetate (30
mL) and water (30 mL). The mixture was extracted with ethyl acetate
3
1
.54 (s, 1H), 2.45 (m, 1H), 2.26 (m, 1H), 2.10 (m, 1H), 1.72 (d, J )
.5 Hz, 1H) 1.70 (d, J ) 4.0 Hz, 1H), 1.18 (s, 3H), 1.15 (d, J ) 3.2
1
3
Hz, 3H), 1.08 (s, 3H), 1.04 (m, 21H) ppm. C NMR (125 MHz,
CDCl
3
): δ 149.2, 147.7, 120.1, 119.3, 75.7, 67.7, 66.0, 49.2, 48.3, 41.3,
(2 × 30 mL), washed with brine, dried (MgSO
4
), and concentrated.
+
31.1, 18.2, 17.5, 15.4, 12.1 ppm. HRMS (EI ): m/z calcd for C23
42 2
H O -
Flash chromatography (5-10% ethyl acetate in petroleum ether) yielded
the desired product 36 (185 mg, 0.52 mmol, 54% yield) along with
ketone 58 (78 mg, 0.22 mmol, 23%) and aldehyde 59 (48 mg, 0.136
+
24
Si [M] 378.2954, found 378.2954. [R]D ) 29.3° (c ) 0.820, CH
Cl ). NOE:
2
-
2
mmol, 14%). R ) 0.25 (10% ethyl acetate in petroleum ether). IR
f
-
1
(
thin film, cm ): 2919, 2850, 1672, 1496, 1458, 1415, 1277, 1218,
1
1
6
2
3
105, 1050. H NMR (500 MHz, CDCl ): δ 6.82 (d, J ) 8.8 Hz, 1H),
.67 (dd, J ) 3.2, 8.8 Hz, 1H), 6.50 (d, J ) 3.0 Hz, 1H), 5.24 (t, J )
.7 Hz, 1H), 4.69 (d, J ) 3.0 Hz, 1H), 4.23(m, 1H), 3.82 (s, 3H), 3.68
(
(
(
1
5
s, 3H), 2.83 (m, 1H), 2.32 (m, 3H), 2.19 (m, 1H), 2.01 (s, 1H), 1.91-
m, 1H), 1.78 (m, 1H), 1.41 (m, 1H), 1.24 (s, 3H), 1.20 (s, 3H), 1.08
d, J ) 6.9 Hz, 3H) ppm. 13C NMR (125 MHz, CDCl
): δ 197.8,
63.5, 154.7, 153.7, 152.4, 134.5, 132.3, 114.4, 112.9, 111.3, 105.9,
7.2, 55.4, 45.0, 38.1, 37.4, 36.7, 34.9, 27.2, 26.8, 25.8, 19.3 ppm.
3
(
1R,6R,8S,8aS,E)-8-(2,5-Dimethoxyphenyl)-3,3,6-trimethyl-7-
methylene-1,2,3,5,6,7,8,8a-octahydroazulen-1-ol (54a). A solution of
3a (0.75 g, 2.28 mmol) in diethylaniline (8 mL) in a sealed tube was
5
+
+
30 3
HRMS (EI ): m/z calcd for C23H O [M] 354.2195, found 354.2190.
heated in a 220 °C oil bath. After 16 h, the reaction was cooled, and
the solvent was removed under reduced pressure at 100 °C. The residue
was dissolved in acetone (10 mL), and K
CH I (0.62 mL, 10 mmol) were added. The reaction was heated at
reflux for 3 h and then diluted with diethyl ether (50 mL). The organic
layer was washed with water and brine, dried (MgSO ), and concen-
trated. Flash chromatography (5-10% EtOAc in petroleum ether)
yielded product 54a (0.44 g, 1.28 mmol, 56% yield, 69% brsm) as a
pale yellow liquid and the starting material 53a (0.144 g, 0.44 mmol).
2
D
4
[R] ) 21.0° (c ) 0.59, CH
2
Cl
2
).
Frondosin A (1). Compound 61 (4.4 mg, 0.013 mmol) in 1:1 H
2
O/
2 3
CO (0.69 g, 5.0 mmol) and
CH
3
CN (0.13 mL, degassed by argon sparging for 30 min) at 0 °C
3
was added dropwise to a solution of CAN (14 mg, 0.026 mmol) in 1:1
O/CH CN (0.13 mL, also degassed). After 10 min, NaHCO (43
mg, 0.52 mmol) was added, followed by Na (23 mg, 0.13 mmol).
H
2
3
3
4
2 2 4
S O
The reaction was stirred vigorously at room temperature for 1 h. The
mixture was filtered through a silica gel pad and concentrated under
an argon atmosphere. Purification by achiral HPLC (2% isopropyl
alcohol in heptane) yielded frondosin A (2.0 mg, 0.0064 mmol, 49%
yield, 89% yield brsm), along with starting material 61 (2.0 mg, 0.0059
1
R
f
) 0.28 (10% EtOAc in petroleum ether). H NMR (500 MHz,
CDCl
3
): δ 6.85 (d, J ) 2.9 Hz, 1H), 6.82 (d, J ) 8.9 Hz, 1H), 6.73
dd, J ) 3.1, 8.8 Hz, 1H), 5.61 (m, 1H) 4.82 (s, 1H), 4.43 (s, 1H),
(
4
(
1
1
20
1
mmol, 45%). H NMR, IR, and [R]D match the reported data. R )
f
.04 (d, J ) 11.5 Hz, 1H), 3.82 (t, J ) 3.7 Hz, 1H), 3.76 (s, 3H), 3.74
s, 3H), 2.89 (d, J ) 11.5 Hz, 1H) 2.51 (m, 2H), 1.95 (m, 1H), 1.70-
.56 (m, 4H), 1.22 (s, 2H) 1.17 (d, J ) 5.5 Hz, 3H) 1.05 (s, 3H) ppm.
C NMR (125 MHz, CDCl
28.4, 120.0, 116.6, 112.5, 111.2, 109.2, 74.9, 56.6, 55.7, 55.1, 47.1,
2.7, 41.2, 37.1, 33.1, 32.0, 21.8 ppm. IR (neat, cm ): 3564, 2954,
-1
0
2
1
.33 (25% ethyl acetate in petroleum ether). IR (thin film, cm ): 3384,
956, 2932, 2871, 1636, 1497, 1452, 1410, 1361, 1279, 1212, 1190,
1
147. H NMR (400 MHz, CDCl
3
): δ 6.77 (d, J ) 8.5 Hz, 1H), 6.67
13
3
): δ 158.2, 153.6, 153.2, 150.9, 132.8,
(
(
dd, J ) 2.9, 8.5 Hz, 1H), 6.59 (d, J ) 3.0 Hz, 1H), 4.85 (s, 1H), 4.52
s, 1H), 3.95 (s, 1H), 2.49, (m, 2H), 2.04 (m, 1H), 1.87 (m, 3H), 1.56-
1
4
1
-
1
1
.51 (m, 4H), 1.29 (m, 1H), 1.10 (s, 3H), 1.09 (s, 3H), 1.06 (d, J ) 7.0
Hz, 3H) ppm. HRMS (MALDI-TOF): m/z calcd for C21 , [M +
H] 313.2089, found [M + H] 313.2081. [R]D ) 37.6° (c ) 0.20,
+
495, 1464, 1281, 1227, 1049. HRMS (EI ): m/z calcd for C22
H
30
O
3
+
24
28 2
H O
[
M] 342.2195, found 342.2210. [R]D ) -52.0° (c ) 0.31, CH
2
Cl
2
).
+
+
20
NOE:
1
20
MeOH), lit. [R]D ) 31.5° (c ) 0.25, MeOH).
Acknowledgment. We thank the National Science Foundation
and the National Institutes of Health, General Medical Sciences
(Grant GM13598), for their generous support of our research.
We also thank Johnson Matthey Co. for their kind donations of
ruthenium chloride catalyst. Mass spectra were provided by the
Mass Spectrometry Regional Center of the University of
CaliforniasSan Francisco supported by the NIH Division of
Research Resources.
(
7R,9S)-9-(2,5-Dimethoxyphenyl)-4,4,7-trimethyl-8-methylene-
,3,4,5,6,7,8,9-octahydrobenzo-cyclohepten-1-one (36). To a solution
of ketone 37 (0.33 g, 0.97 mmol) in CH Cl (10 mL) at -25 °C was
added BF ‚Et O (0.15 mL, 1.22 mmol) followed by (TMS)CHN (0.61
mL, 2 M in ether, 1.22 mmol). The reaction was stirred at -25 °C for
h and then quenched with saturated aqueous NaHCO (20 mL). The
mixture was extracted with diethyl ether (3 × 20 mL), and the combined
2
2
2
Supporting Information Available: Experimental procedures
and spectroscopic data for all new compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
3
2
2
8
3
JA073272B
11790 J. AM. CHEM. SOC.
9
VOL. 129, NO. 38, 2007