Total synthesis of (±)-merrilactone A

Article abstract of DOI:10.1021/ja0761986

The total synthesis of racemic merrilactone A (a neurotrophic agent) is described, featuring simultaneous and stereospecific creation of the C4 and C5 Stereocenters via a notable silyloxyfuran Nazarov cyclization. Full details of the successful synthetic strategy are given, as well as several examples of the interesting reactivity of intermediates that were prepared and studied during the execution of the total synthesis. A detailed investigation of the Lewis acid-catalyzed Nazarov cyclization of silyloxyfurans was conducted, including a systematic study of substrate scope and limitations. In addition, experiments were conducted that suggest the participation of Lewis acidic silicon species in the Nazarov cyclization.

Full text of DOI:10.1021/ja0761986

Published on Web 12/08/2007
Total Synthesis of (()-Merrilactone A
Wei He, Jie Huang, Xiufeng Sun, and Alison J. Frontier*
Contribution from the Department of Chemistry, UniVersity of Rochester,
Rochester, New York 14627
Received August 16, 2007; E-mail: frontier@chem.rochester.edu
Abstract: The total synthesis of racemic merrilactone A (a neurotrophic agent) is described, featuring
simultaneous and stereospecific creation of the C4 and C5 stereocenters via a notable silyloxyfuran Nazarov
cyclization. Full details of the successful synthetic strategy are given, as well as several examples of the
interesting reactivity of intermediates that were prepared and studied during the execution of the total
synthesis. A detailed investigation of the Lewis acid-catalyzed Nazarov cyclization of silyloxyfurans was
conducted, including a systematic study of substrate scope and limitations. In addition, experiments were
conducted that suggest the participation of Lewis acidic silicon species in the Nazarov cyclization.
Introduction
substrates with substituents at both termini of the pentadienyl
cation intermediate will cyclize to generate adjacent stereo-
Merrilactone A (1) was isolated from the pericarps of Illicium
merrillianum in 2000 by Fukuyama and co-workers. The natural
product exhibits unusual neurotrophic activity, promoting growth
of fetal rat neurons at concentrations of 0.1 µmol/L.1 The
interesting structure of this small molecule has attracted the
attention of synthetic groups because of its dense triquinane-
like carbon skeleton, which contains seven contiguous chiral
centers, of which three are quaternary. Danishefsky and Birman
achieved the first total synthesis of (()-merrilactone A, based
on a Diels-Alder cycloaddition.2 A year later, Inoue, Sato, and
Hirama reported a strategy based on the ring contraction of a
1,4-cyclooctenediketone.3 Both of these groups have also
achieved asymmetric syntheses of merrilactone A,4,5 and Inoue
has synthesized the unnatural enantiomer of the natural product.6
Mehta and Singh reported a third (racemic) synthesis based on
the desymmetrization of 1,4-cyclopentenedione,7 and other
groups have also disclosed novel approaches to this unique
carbocyclic system.8,9
centers stereospecifically (see IIfIII, eq 1). However, often a
stoichiometirc amount of a strong Lewis acid is necessary to
promote cyclization, and the final stage of the cyclization
involves the elimination of a proton from an oxyallyl cation
intermediate (e.g., III) that often occurs with poor regioselec-
tivity.12
Recently, catalytic protocols for Nazarov cyclization have
been reported, using a wide range of transition-metal complexes.
Our findings,13 along with those of Trauner,14 Tius,15 Aggar-
wal,16 and Occhiato,17 indicate that polarization of the precursor
divinyl ketone is key to the efficiency of these cyclizations.
These studies, carried out by different research groups, have
shown that compounds that develop high electron density at
one terminus of the pentadienyl cation intermediate are par-
ticularly reactive and cyclize under mild conditions with e20
mol % of catalyst. For example, 1,2-dihydropyran-containing
substrates readily undergo Lewis acid-catalyzed Nazarov
cyclization13-15 as well as regioselective elimination resulting
from the asymmetry of the intermediate oxyallyl cation (eq 2).
The Nazarov cyclization is a 4π electrocyclization typically
involving the conversion of divinyl ketones to cyclopentenones
by activation with a Lewis acid (eq 1).10 Conservation of orbital
symmetry dictates a conrotatory cyclization pathway,11 such that
(11) (a) Woodward, R. B.; Hoffmann, R. Angew. Chem., Int. Ed. Engl. 1969,
8, 781. (b) Woodward, R. B.; Hoffmann, R. The ConserVation of Orbital
Symmetry; Verlag Chemie: Weinheim, Germany, 1970.
(1) Huang, J. M.; Yokoyama, R.; Yang, C. S.; Fukuyama, Y. Tetrahedron Lett.
2000, 41, 6111.
(12) Denmark’s silicon-directed Nazarov cyclization protocol is one solution
to the regioselectivity problem in the elimination step, see: (a) Denmark,
S. E.; Jones, T. K. J. Am. Chem. Soc. 1982, 104, 2642. (b) Jones, T. K.;
Denmark, S. E. HelV. Chim. Acta 1983, 66, 2377. (c) Jones, T. K.; Denmark,
S. E. HelV. Chim. Acta 1983, 66, 2397. (d) Denmark, S. E.; Habermas, K.
L.; Hite, G. A.; Jones, T. K. Tetrahedron 1986, 42, 2821. (e) Denmark, S.
E.; Klix, R. C. Tetrahedron 1988, 44, 4043. (f) Denmark, S. E.; Habermas,
K. L.; Hite, G. A. HelV. Chim. Acta 1988, 71, 168. (g) Denmark, S. E.;
Hite, G. A. HelV. Chim. Acta 1988, 71, 195. (h) Denmark, S. E.; Wallace,
M. A.; Walker, C. B. J. Org. Chem. 1990, 55, 5543.
(13) (a) He, W.; Sun, X.; Frontier, A. J. J. Am. Chem. Soc. 2003, 125, 14278;
addition/ correction J. Am. Chem. Soc. 2004, 126, 10493. (b) Malona, J.
A.; Colbourne, J. M.; Frontier, A. J. Org. Lett. 2006, 8, 5661-5664.
(14) Liang, G.; Gradl, S. N.; Trauner, D. Org. Lett. 2003, 5, 4931.
(15) Bee, C.; Leclerc, E.; Tius, M. A. Org. Lett. 2003, 5, 4927.
(16) Aggarwal, V. K.; Beffield, A. J. Org. Lett. 2003, 5, 5075.
(2) Birman, V. B.; Danishefsky, S. J. J. Am. Chem. Soc. 2002, 124, 2080.
(3) Inoue, M; Sato, T; Hirama, M; J. Am. Chem. Soc. 2003, 125, 10772.
(4) Meng, Z.; Danishefsky, S. J. Angew. Chem., Int. Ed. 2005, 44, 1511.
(5) Inoue, M.; Sato, T.; Hirama, M. Angew., Chem. Int. Ed. 2006, 45, 4843.
(6) Inoue, M.; Lee, N.; Kasuya, S.; Sato, T.; Hirama, M.; Moriyama, M.;
Fukuyama, Y. J. Org. Chem. 2007, 72, 3065.
(7) Mehta, G.; Singh, S. R. Angew. Chem., Int. Ed. 2006, 45, 953.
(8) Iriondo-Alberdi, J.; Perea-Buceta, J. E.; Greaney, M. F. Org. Lett. 2005,
7, 3969.
(9) (a) Harada, K.; Kato, H.; Fukuyama, Y. Tetrahedron Lett. 2005, 46, 7407.
(b) Harada, K.; Ito, H.; Hioki, H.; Fukuyama, Y. Tetrahedron Lett. 2007,
48, 6105.
(10) For reviews on Nazarov cyclization, see (a) Habermas, K. L.; Denmark,
S. E.; Jones, T. K. Org. React. (New York) 1994, 45, 1. (b) Harmata, M.
Chemtracts: Org. Chem. 2004, 17, 416. (c) Pellissier, H. Tetrahedron 2005,
61, 6479. (d) Frontier, A. J.; Collison, C. Tetrahedron 2005, 61, 7577. (e)
Tius, M. A. Eur. J. Org. Chem. 2005, 11, 2193.
(17) (a) Occhiato, E. G.; Prandi, C.; Ferrali, A.; Guarna, A.; Venturello, P. J.
Org. Chem. 2003, 68, 9728. (b) Prandi, C.; Ferrali, A.; Guarna, A.;
Venturello, P.; Occhiato, E. G. J. Org. Chem. 2004, 69, 7705.
9
300
J. AM. CHEM. SOC. 2008, 130, 300-308
10.1021/ja0761986 CCC: $40.75 © 2008 American Chemical Society

Total Synthesis of (
±
)-Merrilactone A
A R T I C L E S
Scheme 1. Synthetic Plan Targeting Merrilactone A
Scheme 2. Synthesis of Aldehyde 4aa
Like the dihydropyrans in eq 1, silyloxyfuranyl vinyl ketone
2 (Scheme 1) has high electron density at the 5-position of the
furan. The intermolecular version of this reaction (Michael
addition of trimethylsilyloxyfurans to R,â-unsaturated carbonyl
compounds) is known,18 and the reaction is also closely related
to the catalytic asymmetric Michael reactions of silyl ketene
acetals described by Evans et al.19 These precedents indicate
that a catalytic amount of a Lewis acid should trigger Nazarov
cyclization of 2 to give tricyclic system 3, which could be
elaborated into the ACD ring system of merrilactone A. This
interesting variation on the Nazarov cyclization would represent
a general method to access [5,5]-fused butenolide ring systems
containing adjacent stereocenters.
a Reaction conditions: (a) (COCl)2, DMSO, Et3N, CH2Cl2, -50 °C,
90%; (b) ethyl propiolate, n-BuLi, THF, -78 °C, then 6 88%; (c) (Bu3Sn)-
Cu(CN)Li2, THF, -78 °C, 90%; (d) Br2, CH2Cl2, room temp, 93%; (e)
TIPSOTf, Et3N, CH2Cl2, -78 to 0 °C, quant; (f) (i) t-BuLi, ether, -78 °C;
(ii) DMF, -78 to 10 °C, 79%.
of the alkyne were observed.22 Treatment of 9 with triisopro-
pylsilyl trifluoromethanesulfonate afforded the silyl furan ether
10 in nearly quantitative yield. Lithium-halogen exchange was
achieved using 2 equiv of t-BuLi at -78 °C, and addition of
DMF delivered target aldehyde 4a, which could be stored at
low temperature for months.
Despite a battery of coupling experiments targeting ketone
2, the only productive result was achieved when the lithiofuran
derived from bromide 11 was employed as the nucleophile
(Scheme 3). Thus, treatment of 1123 with t-BuLi and addition
of aldehyde 4a gave 12 in 50% yield upon mild aqueous
workup. The success of this coupling was surprising, and the
selective hydrolysis may result from participation of the adjacent
free hydroxyl group. Oxidation of 12 to the ketone 2 was readily
achieved by Dess-Martin periodinane, although all other
methods examined failed to give any desired product.24 The
ketone 2 was found to be relatively unstable, decomposing
slowly even at low temperature.
Results and Discussion
Initial Studies Targeting ACD Tricycle 3. Assembly of
substrate 2 would require addition of an appropriate nucleophile
to an aldehyde of type 4 (Scheme 1). Synthesis of ring A
precursor 4a started with known alcohol 520 (Scheme 2), which
was easily oxidized to aldehyde 6 in 90% yield. Treatment of
6 with a mixture of n-BuLi and ethyl propiolate at -78 °C gave
alcohol 7 in 80% yield. Stannylcupration of the unprotected
alkynoate proceeded smoothly,21 and in situ cyclization occurred
to afford desired lactone 8. Bromination of 8 was carried out
via titration in dichloromethane with a bromine solution at room
temperature, affording 9 in 93% yield. The analogous iodination
sequence was unselective: byproducts resulting from iodination
(18) Kitajima, H.; Ito, K.; Katsuki, T. Tetrahedron 1997, 53, 17015.
(19) (a) Evans, D. A.; Rovis, T.; Kozlowski, M. C.; Tedrow, J. S. J. Am. Chem.
Soc. 1999, 121, 1994. (b) Evans, D. A.; Scheidt, K. A.; Johnston, J. N.;
Willis, Michael C. J. Am. Chem. Soc. 2001, 123, 4480.
(20) Cruciani, P.; Stammler, R.; Aubert, C.; Malacria, M. J. Org. Chem. 1996,
61, 2699.
(21) (a) Reginato, G.; Capperucci, A.; Degl’Innocenti, A.; Mordini, A.; Pecchi,
S. Tetrahedron, 1995, 51, 2129. See also: (b) Lipshutz, B. H.; Ellsworth,
E. L.; Dimock, S. H.; Reuter, D. C. Tetrahedron Lett. 1989, 30, 2065.
(22) For the bromination/iodination protocol, see: Hollingworth, G. J.; Knight,
J. R.; Sweeney, J. B. Synth. Commun. 1994, 24, 755.
(23) For the synthesis of 11, see the Supporting Information.
9
J. AM. CHEM. SOC. VOL. 130, NO. 1, 2008 301

A R T I C L E S
He et al.
Scheme 3. Synthesis and Attempted Cyclization of 2a
this regard, two types of acyclic substrates were examined: the
simple enones of type 15, and the alkylidene â-ketoesters of
type 16 (Scheme 5). Alkylidene â-ketoester substrates 16 were
easily prepared using a standard Knoevenagel condensation
procedure.28 Initially, it was thought that synthesis of enones
of type 15 could be carried out as shown in Scheme 5 (top).
Indeed, nucleophilic addition of the 4-lithiofuran derived from
10 or 17 to R,â-unsaturated aldehydes 1829 gave alcohol 19 with
high efficiency.30 However, oxidation to ketones 15 required
31
treatment of 19 with a large excess of MnO2 (50 equiv) in
the presence of molecular sieves (4 Å) in dichloromethane at
room temperature for 3 days,32 and yields varied widely from
substrate to substrate. Changes in temperature, solvent, and
MnO2 source did not improve efficiency, and yields dropped
upon scaleup. Therefore, an alternative procedure employing
Weinreb amides 2033 as electrophiles was developed to avoid
the problematic oxidation of the secondary alcohol 19 (Scheme
5, bottom).34
a Reaction conditions: (a) 11/t-BuLi, ether, -78 °C, then 4a; mild
aqueous workup; 50% yield; (b) Dess-Martin periodinane, CH2Cl2, room
temp, 72%.
Scheme 4. Intramolecular Claisen Condensation of Alcohol 12
Initial Nazarov cyclization attempts were carried out with
silyloxyfuran 16a, since earlier experiments showed that
substrates bearing the 2,4,6-trimethoxyphenyl (TMP) substituent
cyclize with great efficiency.13 To our delight, Cu(OTf)2 effected
the cyclization of a 9:1 E/Z mixture of 16a35 to give 90% yield
of [5,5]-fused butenolide 21, containing adjacent stereocenters
(eq 3). The stereochemistry of butenolide 21 corresponded to
conrotatory cyclization of the minor (Z) isomer. It has been
shown that E/Z isomerization of the alkylidene â-ketoester is
facile under the reaction conditions and that the Z isomer
cyclizes while the E isomer does not.36
Unfortunately, attempts to effect Nazarov cyclization of 2
using a variety of Lewis acids, including Cu(OTf)2, TiCl4, SnCl4,
AlCl3, TMSOTf, FeCl3, and ZnCl2, were unsuccessful. Buteno-
lide 13 was the only product isolated.25
During attempts to oxidize the secondary alcohol 12 to the
ketone 2, an unexpected product was formed upon treatment
with NiO2, PDC, or PCC in dichloromethane. This compound
was isolated in 50% yield. Treatment with D2O led to the
disappearance of one proton peak in 1H NMR spectrum,
indicating the presence of a hydroxyl group. The product was
further characterized by 13C, DEPT, and HSQC NMR spec-
troscopy, and the structure was assigned as the bicyclic ketone
14.26 Notably, only one diastereoisomeric product was isolated
from these reactions, although the relative configuration of the
product stereocenters was not determined.
Product 14 is thought to arise from an intramolecular
Mukaiyama-type Claisen condensation, as shown in Scheme 4.
To our knowledge, no other examples of this type of condensa-
tion have been reported previously. Given the high diastereo-
selectivity of the process and the complexity of the ring system
formed, further development of this reaction for synthetic
applications might be warranted.27
Unfortunately, although cyclization of substrate 16a was
successful, the cyclization of other related alkylidene â-ke-
toesters of type 16 did not occur with the same efficiency.
Therefore, it was not surprising to find that copper triflate was
not a competent catalyst for the cyclization of simple ketone
15a (eq 4). However, we were pleased to find that cyclization
was possible when the complex Ir[(dppe)(CO)(DIB)(CH3)]2+
(28) For experimental details, see the Supporting Information and the follow-
ing: Janka, M.; He, W.; Frontier, A. J.; Eisenberg, R.; Flaschenreim, C.
Tetrahedron 2005, 61, 6193.
(29) For synthesis of aldehydes 18, alcohols 19, and enones 15 see Supporting
Information. Initial experiments toward the synthesis of compound 18 were
carried out by Dr. Patrick Caruana.
(30) Ishihara, J.; Yamamoto, Y.; Kanoh, N.; Murai, A. Tetrahedron Lett. 1999,
40, 4387.
Catalytic Nazarov Cyclizations of Silyloxyfurans. When
the cyclization of bicyclic silyloxyfuran 2 failed, alternative
precursors to the merrilactone AC ring system were sought. In
(31) Muehldorf, A. V. Tetrahedron Lett. 1994, 35, 6851.
(32) Dess-Martin oxidation, Swern oxidation, TPAP oxidation, Parikh-Doering
(SO3‚Py) oxidation, pyridinium dichromate (PDC), pyridinium chlorochro-
mate (PCC), BaMnO4, and NiO2 did not achieve the desired oxidation of
19 to 15.
(24) Oxidation conditions examined included manganese dioxide, nickel dioxide,
pyridinium dichromate (PDC), pyridinium chlorochromate (PCC), TPAP/
NMO, and Swern oxidation.
(25) Two related substrates were synthesized, and neither one cyclized. These
results suggest that the alignment of the two olefin moieties for bicyclic
substrates like 2 is poor. See Supporting Information.
(33) For synthesis of the Weinreb amides, see the Supporting Information.
(34) For a single example of the successful coupling of a 4-lithiofuran with a
Weinreb amide, see Kanoh, N.; Ishihara, J.; Yamamoto, Y.; Murai, A.
Synthesis, 2000, 13, 1878. The procedure presented in Scheme 5 was
employed in the synthesis of several Nazarov substrates (see Supporting
Information).
(26) For details, see the Supporting Information.
(27) Many of the reagents that triggered the cyclization were oxidants, suggesting
that an oxidative radical process may be a potential mechanism for the
cyclization. Indeed, when the starting material was treated with tris(4-
bromophenyl) hexachloroantimonate, a cationic radical promoter, the
reaction proceeded smoothly and efficiently even at -78 oC. For another
example of a Michael addition carried out under oxidative conditions, see:
Otera, J.; Fujita, Y.; Sakuta, N.; Fujita, M.; Fukuzumi, S. J. Org. Chem.
1996, 61, 2951.
(35) Assignment of olefin geometry was made by J3 C-H coupling: Kingsbury,
C. A.; Draney, D.; Sopchik, A.; Rissler, W.; Durham, D. J. Org. Chem.
1976, 41, 3863. The ratio of isomers was determined by 1H NMR.
(36) He, W.; Herrick, I. R.; Atesin, T. A.; Caruana, P. A.; Kellenberger, C. A.;
Frontier, A. J. J. Am. Chem. Soc., in press.
9
302 J. AM. CHEM. SOC. VOL. 130, NO. 1, 2008

Total Synthesis of (
±
)-Merrilactone A
A R T I C L E S
Scheme 5. Synthesis of Enone Nazarov Substrates 15a
a Reaction conditions: (a) 10 or 17/t-BuLi, ether, -78 °C, then 18; 40-82% yield; (b) MnO2 (50 equiv), mol. sieves, CH2Cl2, room temp (34-69%
yield); (c) 10 or 17/t-BuLi, ether, then 20; -78 °C to room temp, 68-83% yield.
2 BArf- (22) was the catalyst,37 even though 15a is missing
the TMP group and the ester that is usually involved in two-
point complexation to the catalyst.37,38
In these Nazarov cyclizations, the triisopropylsilyl group is
transferred from the silyloxyfuran to the ketone oxygen during
the reaction, a process analogous to the silicon transfer step
that characterizes the Mukaiyama-Michael addition.40 Since
electrophilic trialkylsilyl species are known to catalyze both
Mukaiyama-Michael reactions41 and Nazarov cyclizations,42
it was necessary to investigate the possibility that a Lewis acidic
silicon species formed in situ could be competing with the Ir-
(III) catalyst.43 A series of experiments indicated that TIPSOTf
did not catalyze the cyclization of any silyloxyfuranyl substrate.
To more closely duplicate the silylium species that might be
formed in situ in the presence of the Ir(III)2+ 2BArf- complex,
NaBArf was mixed with TIPSOTf, a procedure that is expected
to precipitate NaOTf45 and give the ion pair TIPS+‚BArf-.46
The suspension was filtered to remove NaOTf, and the resultant
solution was added to substrate 15a in dichloromethane.
Cyclization was rapid and efficient, which was an interesting
result given the unusual nature of the triisopropylsilylium
cation.46 Furthermore, the results from the “silyilium” cyclization
were identical to the results of cyclizations catalyzed by Ir(III)
complex 22, which calls into question the true identity of the
catalyst for these silyloxyfuran cyclizations.47 While these kinds
of poorly defined Lewis acidic trialkylsilyl-BArf species are not
typically used as catalysts for organic reactions, the related silyl
triflimides R3Si‚NTf2 are known to be highly efficient catalysts
for Mukaiyama aldol48 and Mukaiyama-Michael reactions.49
Indeed, 1 mol % of either TBS‚NTf2 or TIPS‚NTf2 also led to
efficient cyclization of triisopropylsilyloxyfuran 15a.
A series of other ketones of type 15 also underwent smooth
cyclization with catalyst 22 (Table 1). Cyclization of the original
alkylidene â-ketoester 16a could also be effected with Ir(III)
complex 22, with efficiency comparable to the Cu(II) case (entry
1). Tetrasubstituted enones cyclized to give tetrasubstituted enol
silane products bearing a quaternary center (entries 3-4), and
substrates with γ-substitution on the silyloxyfuran also cyclized
efficiently (entries 7-12). As expected, the geometry of the
starting enone was converted directly into the stereochemistry
of the product butenolides 23 via conrotatory cyclization.11,39
In this way, cyclization of pairs of geometric isomers led to
the synthesis of pairs of cyclopentenyl diastereomers (entries 3
vs 4, 5 vs 6, and 7 vs 8). Cyclization occurs even when steric
hindrance is significant: note the congestion of the adjacent
stereocenters formed in entries 10 and 11. However, attempts
to construct a tricyclic structure utilizing this strategy were not
successful (entry 12; see also Scheme 3). It is possible that when
the rigid lactone unit engages in two-point chelation with the
Lewis acid catalyst, the system is unable to achieve the orbital
overlap necessary for cyclization.25 tert-Butyldimethylsilyloxy-
furans also cyclized, albeit in poor yield. These results were
attributed to poor stability of both the substrate and the enol
silane product.
These studies represent the development of an efficient,
catalytic Nazarov cyclization of trialkylsilyloxyfuranyl ketones.
The results suggest that while Ir(III) complex 22 was initially
thought to be the catalyst for the reaction, it appears likely that
a Lewis acidic silicon species is also involved in the process.
(40) (a) Narasaki, K.; Soai, K.; Mukaiyama, T.; Chem. Lett. 1974, 1223. (b)
Saigo, K.; Otsuki, M.; Mukaiyama, T. 1976, 5, 163.
(41) RajanBabu, T. V. J. Org. Chem. 1984, 49, 2083.
(42) Andrews, J. F. P.; Regan, A. C. Tetrahedron Lett. 1991, 32, 7731.
(43) (a) Carreira, E. M.; Singer, R. A. Tetrahedron Lett. 1994, 35, 4323. (b)
Denmark, S. E.; Chen, C.-T. Tetrahedron Lett. 1994, 35, 4327. (c) Hollis,
T. K.; Bosnich, B. J. Am. Chem. Soc. 1995, 117, 4570. (d) Hiraiwa, Y.;
Ishihara, K.; Yamamoto, H. Eur. J. Org. Chem. 2006, 1827.
(44) Ref deleted in proof.
(45) Arndtsen, B. A.; Bergman, R. G. Science 1995, 270, 1970.
(46) Lambert, J. B.; Zhang, S.; Stern, C. L.; Huffman, J. C. Science 1993, 260,
1917.
(47) No reaction occurred in the presence of NaBArf alone, indicating that the
silicon species was responsible for the catalytic cyclization. See Supporting
Information.
(37) Janka, M.; He, W.; Frontier, A. J.; Eisenberg, R. J. Am. Chem. Soc. 2004,
126, 6864.
(38) See reference 28.
(48) (a) Boxer, M. B.; Yamamoto, H. Org. Lett. 2005, 7, 3127. (b) Boxer, M.
B.; Yamamoto, H. J. Am. Chem. Soc. 2006, 128, 48.
(49) Inanaga, K.; Takasu, K.; Ihara, M. J. Am. Chem. Soc. 2005, 127, 3668.
(39) Stereochemistry of products was determined by 1D nOe experiments.
9
J. AM. CHEM. SOC. VOL. 130, NO. 1, 2008 303

A R T I C L E S
He et al.
Table 1. Nazarov Cyclization of Type 15 and 16 Silyloxyfurans
with Ir(III) Catalyst 22a
the necessary cis-[5,5] ring fusion. Unfortunately, compound
23i decomposed rapidly upon exposure to several different
fluoride-containing reagents. It was remarkably stable under
acidic conditions but, despite extensive experimentation, selec-
tive deprotection could not be achieved.
Construction of the ABC Tricycle 26 from Nazarov
Cyclization Product 23i. Because assembly of the ACD system
was problematic, a strategy involving construction of ring B
prior to deprotection of 23i was pursued (Scheme 7). Depro-
tection of the TMS group of 23i was best achieved by treatment
of AgNO3 and KCN, providing terminal alkyne 24.50 Radical
cyclization proceeded smoothly with azobis(isobutyronitrile)
(AIBN) and tributyltin hydride (Bu3SnH)51 to give vinylstannane
25 in high yield (88%) along with a trace amount (∼ 4%) of
the target exo-olefin 26. Addition of p-toluenesulfonic acid
monohydrate (p-TsOH‚H2O) in benzene to the reaction effected
conversion of remaining vinylstannane 25 to 26 in situ. A trace
amount of the free alcohol 27, arising from the deprotection of
the tert-butyldimethylsilyl (TBS) group, was also isolated.
Tricycle 26 contained all the carbon atoms of the merrilactone
A carbon skeleton except C-12. To explore intramolecular
acylation procedures, it was necessary to selectively deprotect
the primary alcohol. Since p-TsOH‚H2O was observed to effect
the partial deprotection of the tert-butyldimethyl (TBS) group
during the cyclization sequence (Scheme 7), we chose first to
explore this avenue. After extensive optimization, it was found
that treatment of 26 with 4 equiv of p-TsOH‚H2O in ethanol
for 4 h furnished the desired primary alcohol 27 in 40% yield,
along with recovered 26 (40%) and the acetal 28 (15%) (Scheme
8). Longer reaction times led exclusively to acetal 28.
Consistent with the observations noted above, primary alcohol
27 was found to be unstable under both aqueous workup
conditions and on silica, forming acetal 28 as a single diaste-
reomer in less than 10 min in both cases. Global deprotection
attempts resulted in efficient formation of acetal 28, which
proved inert to hydrolysis with p-TsOH‚H2O even after long
reaction times at elevated temperatures. However, subjection
of compound 26 to similar forcing conditions triggered oxycy-
clization, affording ether 30 in low yield. The fully deprotected
hydoxyketone was not observed in either case (Scheme 8).
1
The structure of ether 30 was assigned by H, 13C, and 2D-
COSY NMR spectroscopy (see Supporting Information). The
structure and stereochemistry of acetal 28 were unambiguously
established by X-ray crystallography (Figure 1).52 The C-3 (C-7
in the X-ray structure) side chain and the C-11 (C-12 in the
X-ray structure) side chain had the expected syn relationship
predicted from the Nazarov cyclization and required for
synthesis of merrilactone A.
Fortunately, 1 equiv of TBAF effected the selective depro-
tection of the triisopropylsilyl (TIPS) group at -78 °C to give
the tert-butyldimethylsilyl (TBS) protected hydroxy ketone 31
(Scheme 9). If excess TBAF was used, global deprotection was
achieved to yield hydroxyketone 29. As was observed in the
a Reaction conditions: Ir[(dppe)(CO)(DIB)(CH3)]2+ 2 BArf- (2 mol %),
b
CH2Cl2, room temperature. R) CH2CH2CCTMS. c Inseparable mixture of
olefin isomers.35 d Obtained in a 1:3.2 mixture of diastereomers about the
H/ Ph stereocenter; the major diastereomer observed was 23g. e Obtained
in a 1:3 mixture of diastereomers about the H/ n-propyl stereocenter; major
diastereomer observed was 23h. f Other Lewis acids such as Cu(OTf)2 and
AlCl3 and Ir(III) gave comparable results.
Further experimentation is necessary to determine whether Ir-
(III) complex 22 is able to both initiate and catalyze cyclization
or that it simply initiates catalysis by a Lewis acidic silicon
species. The methodology provides rapid, stereospecific access
to [5,5]-fused butenolide-containing systems with adjacent
stereocenters.
With AC ring system 23i in hand, attempts were again made
to synthesize the tricyclic ACD system 3 (Scheme 6). Selective
deprotection of the primary silyloxy group of 23i and subsequent
conversion to the carbonate derivative would set the stage for
intramolecular acylation, in order to install lactone ring D with
(50) Kobayashi, Y.; Shimazaki, T.; Taguchi, H.; Sato, F. J. Org. Chem. 1990,
55, 5324. A variety of fluoride-based deprotection reagents such as TBAF,
HF, and HF‚Py were unselective, and no reaction occurred using K2CO3
in methanol.
(51) Shanmugam, P.; Srinivasan, R.; Rajagopalan, K. Tetrahedron, 1997, 53,
6085.
(52) With the exception of the numbering that appears in Figures 1 and 2
(assigned for the purposes of X-ray analysis only), the carbon numbering
in the manuscript will correspond to the numbering of the natural product
merrilactone A (see ref 1).
9
304 J. AM. CHEM. SOC. VOL. 130, NO. 1, 2008

Total Synthesis of (
±
)-Merrilactone A
A R T I C L E S
Scheme 6. Proposed Construction of ACD Tricycle 3 via Intramolecular Acylation
Scheme 7. Radical Cyclization of Terminal Alkyne 24a
Scheme 9. Selective Deprotection of 26 with
Tetrabutylammonium Fluoridea
a Reaction Conditions: (a) AgNO3, then KCN, THF/H2O/EtOH, 93%;
(b) (i) AIBN, Bu3SnH, PhH, reflux, then (ii) p-TsOH‚H2O (one pot), 88%
(26), 4% (27).
a Reaction conditions: (a) TBAF (1 equiv), THF, -78 °C; (b) TBAF (3
equiv), THF, room temp.
Scheme 8. Treatment of 26 with p-TsOH‚H2Oa
Scheme 10. Attempted Intramolecular Acylation of Carbonates
32a
a Reaction conditions: (a) LDA, ClCO2R, 90%; (b) TMSOTf (72%),
TiCl4, or BF3‚Et2O.
a Reaction conditions: (a) p-TsOH‚H2O, EtOH, room temp; (b)
p-TsOH‚H2O, benzene, reflux, 2 h.
via acylation of ketone enolates with pendant carbonates,53 it
seemed possible that the TIPS enol ether might be nucleophilic
enough to allow direct acylation with carbonates of type 32 to
give desired tricycle 33 (Scheme 10). Carbonate substrates 32a
and 32b were prepared by acylation of primary alcohol 27. LDA
was found to be the ideal base for these reactions, while
nucleophilic bases such as 4-dimethylaminopyridine (DMAP)
or n-BuLi yielded primarily TIPS acetal 28. Since this trans-
formation is similar to the Mukaiyama aldol reaction, in the
sense that an enol ether attacks a Lewis acid activated carbonyl
electrophile to form a new carbon-carbon bond,54 Lewis acidic
conditions were explored first. In a surprising twist, attempted
acylation of phenyl carbonate 32a using TMSOTf as the Lewis
acid promoter led to exclusive formation of cyclopropane 34.
No lactone product 33 was detected. The same result was
observed with ethyl carbonate 32b, and with alternative Lewis
acids TiCl4 and BF3‚Et2O.
formation of acetal 28, both 29 and 31 were obtained as single
diastereomers (rather than as pairs of C-6 epimers), although
it was not possible to assign the C-6 stereochemistry with
confidence.
Thus, after extensive experimentation, conditions were identi-
fied that allowed selective deprotection of either silyl group of
26 or its global deprotection (Schemes 8 and 9).
Efforts To Form the ABCD Tetracycle from Derivatives
of 26. Encouraged by literature examples of lactone formation
When carbonates 32 underwent SN2-type displacement instead
of acylation, alternative C-12 carbon donors were sought.
Chloroformate 35 and orthoester 36 were appealing because both
were expected to collapse to reactive oxonium ions upon
treatment with a Lewis acid (Scheme 11).55 The acylation could
(53) (a) Goldsmith, D. J.; John, T. K.; Van Middlesworth, F. Synth. Commun.
1980, 110, 551. (b) Molander, G. A.; Quirmbach, M. S.; Silva, L. F.;
Spencer, K. C.; Balsells, J. Org. Lett. 2001, 15, 2257. (c) Cho, S. Y.;
Carcache, D. A.; Tian, Y.; Li, Y.; Danishefsky S. J. J. Am. Chem. Soc.
2004, 126, 14358.
(54) However, the reaction between a silyl enol ether and carbonate, either in
an intermolecular fashion or in an intramolecular fashion, had not been
reported.
(55) Activation of orthoformates with Lewis acids has been reported to give
electrophilic species that are reactive toward CdC double bonds: Li, T.
T.; Lesko, P.; Ellison, R. H.; Subramanian, N.; Fried, J. H. J. Org. Chem.
1981, 46, 111.
Figure 1. X-ray structure of TIPS acetal 28.
9
J. AM. CHEM. SOC. VOL. 130, NO. 1, 2008 305

A R T I C L E S
He et al.
Scheme 11. Cyclization Experiments with Chloroformate 35 and
Orthoformate 36a
Scheme 13. Intermolecular Acylation of Ketone 31a
a Reaction Conditions: (a) LDA (1.8 equiv), ClCO2Et (1.5 equiv) gave
38 (∼15% yield) and diastereomers 39 (2.3:1 ratio, 67% yield); (b) TBAF,
THF, 0 °C to room temp; gave 40-trans (57% yield) and 41 (26% yield).
a Reaction conditions: (a) phosgene, CH2Cl2 90%; (b) HC(OMe)3,
MgCl2, CH2Cl2; (c) TMSOTf or AgOTf, 88%; (d) TiCl4, 68% (from 27).
Scheme 12. Attempted Intramolecular Acylation under Basic
Conditionsa
a Bases examined: n-BuLi, LDA, NaH, NaOMe, NaH/KH, KH, KOtBu.
then proceed via a 5-exo reaction pathway, with less steric
congestion at the site of the C-6/C-12 bond formation. As shown
in Scheme 11, both precursors could be prepared in one step
from alcohol 27.56,57
The results of these cyclization attempts were disappointing,
however. Treatment of chloroformate 35 with promoters TMS‚
OTf and AgOTf only triggered hydrolysis of the triisopropylsilyl
(TIPS) enol ether. No cyclization was observed. Treatment of
unstable orthoformate 36 with TiCl4 gave cyclopropane 34 in
good yield as the exclusive product of the reaction.
These experiments provided strong evidence that the Lewis
acid-mediated acylation of triisopropylsilyl (TIPS) enol ether
was not a viable strategy for the synthesis of lactone 33.
Acylation of the corresponding enolates, expected to be smaller
and more reactive, was pursued next. The first experiment,
treatment of carbonate 32b with fluoride ion under basic
conditions, gave ketone 37 at -78 °C (90% yield) and formation
of cyclopropane 34 at 0 °C (66% yield; Scheme 12). Formation
of the enolate from ketone 37 in the presence of metals that
might be able to coordinate both the ketone oxygen and the
carbonyl oxygen of the carbonate was also explored to encour-
age proper alignment of the nucleophile and the electrophile.
However, deprotonation of 37 always gave cyclopropane 34,
sometimes along with the hydroxyketone 29.
In a final effort to build ring D onto an intermediate derived
from intermediate 26, intermolecular acylation of R-methyl
ketone 31 was investigated. Although the stereoselectivity of
acylation at C-6 was expected to be poor, any opportunity to
access desired lactone 33 had become attractive. Upon initiation
of these experiments, it was found that excess LDA was required
to drive the acylation of 31 with ethylchloroformate to comple-
tion. However, a byproduct tentatively assigned as C-10 acylated
lactone 38 was always formed under these conditions (Scheme
Figure 2. X-ray crystal structure of compound 41.
13). Fortunately, it was possible to isolate a mixture of
diastereomeric ketoesters 39 (trans and cis) in 67% combined
1
yield. H NMR revealed a 2.3:1 ratio of epimers, but relative
stereochemistry could not be assigned from the spectrum.
Exposing the mixture of ketoesters 39 to TBAF led to the
formation of two new products. One of them was confidently
assigned as the hydroxy ketoester 40-trans, which was expected
to be inert under lactonization conditions owing to the trans
relationship of the alcohol and the ester. The other product was
assigned as γ-butyro-δ-valerolactone 41, via X-ray crystal-
lographic analysis (Figure 2).52
Isolation of 40-trans from the reaction mixture suggests that
only 39-cis undergoes the rearrangement/fragmentation. An
oxidation event must also occur to deliver product 41, probably
via the action of adventitious oxygen in the reaction mixture.58
A possible mechanistic pathway for the transformation is shown
in Scheme 14. TBAF should effect deprotection of the silyl
ethers 39 to afford primary alcohols 40. Alcohol 40-cis could
then undergo lactonization as planned to give the elusive product
33, generating a molecule of ethanol. It is then possible that
(56) Chloroformate: Lucas, M. A.; Schiesser, C. H. J. Org. Chem. 1996, 61,
(58) (a) Bailey, E. J.; Barton, D. H. R.; Elks, J.; Templeton, J. F. J. Chem. Soc.
1962, 1578. (b) Gardner, J. N.; Carlon, F. E.; Gnoj, O. J. Org. Chem. 1968,
33, 3294. (c) Gardner, J. N.; Popper, T. L.; Carlon, F. E.; Gnoj, O.; Herzog,
H. L. J. Org. Chem. 1968, 33, 3695-3699.
5754.
(57) Orthoester: Perron, F.; Gahman, T. C.; Albizati, K. F. Tetrahedron Lett.
1988, 29. 2023.
9
306 J. AM. CHEM. SOC. VOL. 130, NO. 1, 2008

Total Synthesis of (
±
)-Merrilactone A
A R T I C L E S
Scheme 14. Proposed Mechanism of the Formation of the
δ-Valerolactone 41
Scheme 16. Completion of the Total Synthesisa
a Reaction Conditions: (a) NaH, MeI, gave 33 (84% yield) and 49 (14%
yield); 6:1 ratio; (b) NaH, MeI, and then HMPA, gave 33 only (97% yield);
(c) NaBH4, MeOH gave 50a (42% yield) and 50b (51% yield); (d) Dess-
Martin periodinane (99% yield);( e) p-TsOH‚H2O, benzene, reflux, 4 h,
92% yield; (f) m-CPBA; (g) p-TsOH‚H2O, 68% yield over two steps.
Scheme 15. Formation of Ring D via Intramolecular Acylationa
These experiments indicate that while cyclization of carbonate
46 gives exclusively lactone 48, the analogous cyclization of
C-6 methyl substituted analogue 37 gave only cyclopropane 34
(see Scheme 12). Evidently, the steric influence of the C-6
methyl in 37 was significant enough to completely change the
course of the reaction.60
The final steps of the total synthesis were carried out as shown
in Scheme 16. Installation of the C-6 methyl was first attempted
by treatment of 48 with NaH and MeI. The reaction was sluggish
and gave a 1:6 ratio of the O-methylation product 49 to the
desired C-methylation product 33. Fortunately, addition of
HMPA both accelerated the reaction and delivered 33 in 97%
yield.61 Reduction of the C-7 ketone was unselective, which
was not surprising given that the two hydride-accepting faces
appeared to be equally accessible. Of the reducing agents
examined, DIBAL-H and L-selectride appeared to overreduce
33, while NaBH4 effected exclusive albeit unselective reduction
of the ketone. Fortunately, the reduction was achieved in high
yield and the two alcohols 50a and 50b could be separated by
chromatography. Undesired alcohol 50a could then be oxidized
back to the starting ketone 33 in quantitative yield using Dess-
Martin periodinane. This recycling process enabled us to convert
50a into the desired alcohol 50b in 73% yield after two cycles.
Isomerization of 50b by treatment with p-TsOH‚H2O in
refluxing benzene for 4 h furnished Danishefsky’s intermediate
51. Finally, following the procedures of Danishefsky and
a Reaction Conditions: (a) AgNO3, then KCN, THF/H2O/EtOH, 87%;
(b) (i) AIBN, Bu3SnH, PhH, reflux, then (ii) p-TsOH‚H2O (one pot), 91%;
(c) 3 equiv TBAF, 0 °C, 99%; (d) DMAP, Py, ClCO2Et, 95%; (e) 20 equiv
NaH, THF, 47:48 ) 1:1; (f) p-TsOH‚H2O (90% from 46).
R,R-disubstituted â-ketoester 33 undergoes ring opening via
attack by either ethanol or water present in the TBAF solution
to give intermediate 42. Autoxidation of 4259 would give
R-hydroperoxide 43, which could then fragment and lactonize
as shown to deliver 41.
These experiments demonstrated that intermolecular acylation
of R-methyl ketone 31 is also not a viable approach to the
synthesis of merrilactone A, so again, alternative routes were
sought.
Completion of the Synthesis from Nazarov Cyclization
Product 23j. Given the disappointments experienced during
attempts to form ring D via acylation of derivatives of
tetrasubsituted silyl enol ether 26, an alternative strategy
involving acylation of the less hindered C-6 carbon of trisub-
situted silyl enol ether 44 was considered (Scheme 15). This
approach was promising because the two examples of intramo-
lecular acylation previously reported involved R-unsubstituted
ketones.53 To this end, the chemistry developed for ring ABC
formation from 23i (cf. Scheme 7) was carried out on 23j (Table
1, entry 11) delivering 44 without complication. Deprotection
with 3 equiv of TBAF delivered ketone 45 in good yield, and
the DMAP/Py/ClCO2Et protocol reported by Molander gave
target carbonate 46.53b Finally, treatment of 46 with excess NaH
in THF triggered intramolecular nucleophilic lactonization53c
to give a 1:1 mixture of the desired lactone 48 along with a
single diastereomer of â-ketoester 47 (relative stereochemistry
at C-6 was not determined). Subsequent treatment of this mixture
with p-TsOH‚H2O converted 47 cleanly into 48, furnishing the
desired bislactone 48 in 90% yield from carbonate 46. Incred-
ibly, no cyclopropanation was observed in this reaction.
(60) A reversible Dieckmann condensation process could also account for the
observed results (see below, eq 1). After the initial acylation to generate
intermediates 46 and 37, R,R-disubstituted â-ketoester 37 could be
susceptible to ring opening under the reaction conditions, potentially leading
to irreversible formation of cyclopropane 34. In contrast, deprotonation of
â-ketoester 46 under the reaction conditions could discourage the retro-
Dieckmann process, preserving the lactone ring system 53. Experiments
conducted on a related ring system (eq 2) confirm the retro-Dieckmann
type of bond cleavage in presence of NaOEt or NaH (see Supporting
Information). Although the reaction analogous to the transformation of 37
to 34 was not observed in the model system, we cannot rule out the
possibility that the reaction does occur in the merrilactone ring system.
The authors thank Professor David Hart (The Ohio State University) and
Professor Michael Harmata (The University of Missouri-Columbia) for
suggesting this reaction mechanism to explain the contrasting results of
the acylation reactions of 37 and 46.
(59) (a) Bailey, E. J.; Barton, D. H. R.; Elks, J.; Templeton, J. F. J. Chem. Soc.
1962, 1578. (b) Gardner, J. N.; Carlon, F. E.; Gnoj, O. J. Org. Chem. 1968,
33, 3294. (c) Gardner, J. N.; Popper, T. L.; Carlon, F. E.; Gnoj, O.; Herzog,
H. L. J. Org. Chem. 1968, 33, 3695.
(61) Jacobi, P. A.; Selnick, H. G. J. Org. Chem. 1990, 55, 202.
9
J. AM. CHEM. SOC. VOL. 130, NO. 1, 2008 307

A R T I C L E S
He et al.
Birman, alcohol 51 was smoothly converted into merrilactone
A (1) in 68% yield. The spectroscopic data of both the
intermediate epoxide and the final product 1 were identical to
those reported in literature for (()-merrilactone A.
developed represent an unusual version of the Nazarov cycliza-
tion that should find additional applications in organic synthesis.
Acknowledgment. We are grateful to the Research Corpora-
tion, the Petroleum Research Fund, NSF (CAREER: CHE-
0349045) and the University of Rochester for generous financial
support. W.H. and A.J.F. thank Prof. Robert K. Boeckman, Jr.
(University of Rochester) and Prof. Scott Denmark (University
of Illinois) for helpful discussions. Prof. Richard Eisenberg and
Dr. Mesfin Janka (University of Rochester) are acknowledged
for providing iridium catalyst 22. We are also grateful to Dr.
William Brennessel (University of Rochester) for solving the
structures of acetal 28 and lactone 41 by X-ray crystallography,
Dr. Sandip Sur (University of Rochester) for assistance with
NMR spectroscopy, and Dr. Alice Bergmann (University of
Buffalo) for carrying out high-resolution mass spectroscopy.
Conclusion
An efficient stereoselective total synthesis of (()-merrilactone
A was achieved via the newly developed catalytic Nazarov
cyclization of silyloxyfurans. The primary roadblock in early
efforts was in the construction of lactone ring D: undesired
cyclopropanation formation predominating during all attempts
to effect intramolecular acylation of tetrasubstituted derivatives
derived from Nazarov product 23i. This difficulty was finally
overcome when derivatives of trisubstituted Nazarov product
23j were used instead of the tetrasubstituted derivatives 23i.
During the synthetic studies, an interesting intramolecular
Claisen condensation of a silyloxyfuran was observed, and
Lewis acidic silicon species were implicated as participants in
the catalytic cyclization reactions. This total synthesis of (()-
merrilactone A demonstrates that catalytic Nazarov cyclization
methodology can be efficiently applied to problems in natural
product synthesis. In particular, the silyloxyfuran cyclizations
Supporting Information Available: Experimental procedure,
characterization data, X-ray crystal structure coordinates and
files for compounds 28 and 41 in CIF format. This material is
available free of charge via the Internet at http://pubs.acs.org.
JA0761986
9
308 J. AM. CHEM. SOC. VOL. 130, NO. 1, 2008