Total Synthesis Highlights

近 年 来 完 成 的 天 然 产 物 的 全 合 成 简 介
Total Synthesis Highlights
URL: http://www.organic-chemistry.org/Highlights/totalsynthesis.shtm
1. The Bradshaw/Bonjoch Synthesis of (-)-Anominine
The Hajos-Parrish cyclization was a landmark in the asymmetric construction of polycarbocyclic
natural products. Impressive at the time, the proline-mediated intramolecular aldol condensation
proceeded with an ee that was low by modern standards. Ben Bradshaw and Josep Bonjoch of the
Universitat de Barcelona optimized this protocol, then used it to prepare (J. Am. Chem. Soc. 2010,
132, 5966.) the enone 3 en route to the Aspergillus alkaloid (-)-Anominine (4).
The optimized catalyst for the enantioselective Robinson annulation was the amide 5. With 2.5
mol % of the catalyst, the reaction proceeded in 97% ee. With only 1 mol % of catalyst, the
reaction could be taken to 96% yield, while maintaining the ee at 94%. Conjugate addition
proceeded across the open face of 3 to give, after selective protection, the monoketal 7. After
methylenation and deprotection, oxidation with IBX delivered the enone 9.
With the angular quaternary centers of the natural product in place, the molecule became
increasingly congested. Attempted direct alkylation of 9 led mainly to O-methylation. A solution
to this problem was found in condensation with the Eschenmoser salt, followed by N-oxide
formation and elimination to give the tetraene 10. Selective reduction by the Ganem protocol
followed by equilibration completed the net methylation.

Under anhydrous conditions, the oxide derived from the allylic selenide 12 did not rearrange. On
the addition of water, the rearrangement proceeded smoothly. Protection and hydroboration
converted 13 into 14. The bulk of the folded molecule protected the exo methylene of 14, so
hydrogenation followed by protection and oxidation delivered 15.
Conjugate addition of indole to 15 set the stage for oxidation and bis-methylenation to give 17.
Selective Ru-mediated cross coupling with 18 followed by deprotection then completed the
synthesis of (-)-Anominine (4), which proved to be the enantiomer of the natural product.
D. F. Taber, Org. Chem. Highlights 2011, March 7.
URL: http://www.organic-chemistry.org/Highlights/2011/07March.shtm
2. The Boger Synthesis of (-)-Vindoline
The periwinkle-derived alkaloids vinblastine (2a) and vincristine (2b) are still mainstays of cancer
chemotherapy. The more complex half of these dimeric alkaloids, vindoline (1), presents a
formidable challenge for total synthesis. Building on his previous work (Org. Lett. 2005, 7, 4539.),
Dale L. Boger of Scripps, La Jolla devised (J. Am. Chem. Soc. 2010, 132, 3685.) a strikingly
simple solution to this problem, based on sequential cycloaddition.

The starting point for the synthesis was the ester 3, derived from D-asparagine. This was extended
to 4, condensation of which with 5 gave the enol ether 6. On heating, 7 cyclized to 8, which lost
N 2 to give the zwitterion 9. Addition of the intermediate 9 to the indole then gave 10. In one
reaction, the entire ring system of vindoline, appropriately oxygenated, was assembled, with the
original stereogenic center from D-asparagine directing the relative and absolute configuration of
the final product.
To complete the synthesis, the pendant carbon on 11 had to be incorporated into the pentacyclic
skeleton. After adjusting the relative configuration of the secondary alcohol, the N was rendered
nucleophilic by reduction of the amide to the amine. Oxidation delivered 14, that on activation as
the tosylate smoothly rearranged to the ketone 15. Reduction and regioselective dehydration then
completed the synthesis of vindoline (1).

D. F. Taber, Org. Chem. Highlights 2011, February 7.
URL: http://www.organic-chemistry.org/Highlights/2011/07February.shtm
3. The Tanino Synthesis of (-)-Glycinoeclepin A
(-)-Glycinoeclepin A (3) is effective at picogram/mL concentrations as a hatch-stimulating agent
for the soybean cyst nematode. Approaching the synthesis of 3, Keiji Tanino of Hokkaido
University envisioned (Chem. Lett. 2010, 39, 835. DOI: 10.1246/cl.2010.835) the convergent
coupling of the allylic tosylate 2 with the bridgehead anion 1. The assembly of the fragment 2 was
particularly challenging, as the synthesis would require the establishment not just of the two
adjacent cyclic quaternary centers, but also control of the relative configuration on the side chain.
The preparation of 1 began with the prochiral diketone 3. Enantioselective reduction of the mono
enol ether 4 set the absolute configuration of 5. Iodination followed by cyclization then completed
the assembly of 1.
The construction of the bicyclic tosylate 2 began with m-methyl anisole (7). Following the
Rubottom procedure, Birch reduction followed by mild hydrolysis gave the ketone 8. Epoxidation
followed by β-elimination delivered the racemic 9, which was exposed to lipase to give, after
seven days, the residual alcohol in 40% yield and high ee.

The side chain nitrile was prepared from the diol 12. Homologation gave the nitrile 14, that was
equilibrated to the more stable enol ether 15. The two cyclic quaternary centers of 3 were set in a
single step, by the conjugate addition of the anion of 16 to the crystalline enone 11. Mild
hydrolysis of 17 gave the keto aldehyde, that underwent aldol condensation to give the enone 18.
The hydroboration of 19 followed by coupling of the intermediate organoborane with 20 delivered
21 with 94:6 relative diastereocontrol. Formylation of the enone 22 followed by triflation and
reduction then led to 2.
Although the ketone 1 could be deprotonated with LDA, the only product observed, even at -78°C,
was the derived aldol dimer. The metalated dimethylhydrazone 25, in contrast, coupled smoothly
with 2 to give, after hydrolyis, the desired adduct 26. Pd-mediated carboxylation of the enol
triflate followed by selective oxidative cleavage and hydrolysis then completed the synthesis of
(-)-Glycinoecleptin A (3).
D. F. Taber, Org. Chem. Highlights 2011, January 3.
URL: http://www.organic-chemistry.org/Highlights/2011/03January.shtm
4. The Chen Synthesis of (-)-Nakiterpiosin
(-)-Nakiterpiosin (3), isolated from the thin encrusting sponge Terpios hoshinota, has an IC 50
against murine P388 leukemia cells of 10 ng/mL. Chuo Chen of UT Southwestern Medical Center

developed (J. Am. Chem. Soc. 2010, 132, 371.) a practical synthetic route to 3 based on the
convergent coupling of 1 and 2.
The preparation of 1 was based on the intramolecular [4+2] cyclization of the furan 9, prepared by
Friedel-Crafts acylation of furan (4) with maleic anhydride (5). The absolute configuration of the
secondary alcohol was set by Noyori reduction, using sodium formate as the hydride source.
The cyclization of 9 to 10 proceeded with high diastereocontrol, presumably by way of a chelated
transition state. As expected, cyclization of the silyl ether of 9 delivered the complementary
diastereomer. As the cyclization of 9 was readily reversible, it was taken quickly to the bromide
11. Oxidative cleavage of the diol followed by selective reduction and protection then completed
the synthesis of 1.
The preparation of 2 began with the commercial bromo acid 12. The enantiomerically-enriched
epoxide 13 was constructed in the usual way, by homologation of the aldehyde to the allylic
alcohol followed by Sharpless epoxidation. On exposure to the Yamamoto catalyst, 13 smoothly
rearranged to the aldehyde 14. Condensation of 14 with 15 then gave 16, with only minimal
erosion of enantiomeric excess over the two steps.
Unfortunately, 16 was the incorrect diastereomer, so it had to be inverted. With the aldehyde 17 in
hand, conversion to the dichloride followed by functional group interchange completed the
construction of 2.

Carbonylative coupling of 1 and 2 led to the enone 18. The photochemical Nazarov cyclization of
18 proceeded with the expected high diastereocontrol, to give, after epimerization, the desired
trans-anti-trans product. Deprotection then completed the synthesis of (-)-Nakiterpiosin (3). It is
noteworthy that the full A-ring functionality of 3 was compatible with the conditions of the
photochemical cyclization.
The work of Chen toward the total synthesis of (-)-nakiterpoisin (3) led to a correction of the
relative configurations both of the dichloromethyl substituent and of the secondary bromide. The
availability of 3 by total synthesis is particularly exciting, because it has been shown to interfere
with the Hedgehog signaling pathway. There is the potential, based on this activity, that
derivatives of 3 may prove useful as adjuncts in cancer chemotherapy.
D. F. Taber, Org. Chem. Highlights 2010, December 6.
URL: http://www.organic-chemistry.org/Highlights/2010/06December.shtm
5. The Shair Synthesis of Cephalostatin 1
The cephalostatins and ritterazines, represented by Cephalostatin 1 (3), have the remarkable
property of inducing apoptosis in apoptosis-resistant malignant cell lines. The total synthesis (J.
Am. Chem. Soc. 2010, 132, 275. DOI: 10.1021/ja906996c) of 3 by Matthew D. Shair of Harvard
University required the practical preparation of the complex hexacyclic ketones 1 and 2.
The preparation of 1 started with irradiation of commercial hecogenin acetate 4 to give the known
aldehyde 5. Reaction of 5 with N-phenyltriazolenedione (6) led to the ketal 7. Oxidative cleavage
generated an aldehyde, that on reduction and allylation was converted to 8. Acid-mediated
cyclization led to 9.

The sidechain of 9 was removed, giving 10, that was selectively reduced, leading to 11.
Intramolecular aldol condensation gave 12. The relative configuration of the spiroketal 1 was
established by kinetic bromoetherification of the alcohol 13, followed by free radical reduction of
the resulting tertiary bromide, and acid-catalyzed equilibration.
The synthesis of 2 began with the inexpensive steroid 14. Following the Schönecker protocol, C-H
functionalization led to the ketone 15. Pd-mediated coupling of the derived enol triflate with the
alkyne gave 16, that was oxidized and cyclized to 17. Simmons-Smith conditions converted the
dihydrofuran of 17 into the cyclopropane, that was again opened kinetically with bromine (NBS)
to set the relative configuration of the spiroketal. Free radical reduction followed by protection
and oxidation then completed the preparation of 2.
The coupling of 1 and 2 (not illustrated) to form the central pyrazine of 3 followed the precedent
of Fuchs, combining the 2-azido ketone derived from 1 with the 2-amino methoxime derived from
2. Remarkably, tens of milligrams of 1 and of 2 were prepared, assuring a reasonable supply of 3
for further studies.
D. F. Taber, Org. Chem. Highlights 2010, November 1.
URL: http://www.organic-chemistry.org/Highlights/2010/01November.shtm

6. The Overman Synthesis of Briarellin F
Briarellin F (4) is an elegant representative of the complex polycyclic ethers produced by soft
corals such as Briareum abestinum. Larry E. Overman of the University of California, Irvine
developed (J. Org. Chem. 2009, 74, 5458.) a triply-convergent approach to 4, the central feature of
which was the Prins-pinacol combination of 1 with 2 to give 3.
The aldehyde 2 was assembled by Wittig homologation of the aldehyde 5 with the phosphorane 6,
followed by metalation and formylation. The aldehyde 10 was prepared by opening the
enantiomerically-pure epoxide 8 with the acetylide 9.
Hydroboration of carvone 11 could not be effected with sufficient diastereocontrol. As an
alternative, the mixture of diols was oxidized to the lactone 12. Kinetic quench of the derived silyl
ketene acetal followed by reduction led to the diastereomerically-pure crystalline diol 13. This key
intermediate will have many other applications in target-directed synthesis.
The ketone 14 was converted to the alkenyl iodide 15 by stannylation of the enol triflate, followed
by exposure of the stannane to N-iodosuccinimide. Addition of the alkenyl iodide 15 to the
aldehyde 10 gave the diol 1 as an inconsequential 3:1 mixture of diastereomers. This mixture was
combined with the aldehyde 2 to give, via Lewis acid-mediated rearrangement of the
initially-prepared acetal, the aldehyde 3.

The aldehyde 3 was readily decarbonylated by irradiation in dioxane. Face-selective Al-mediated
epoxidation of the derived homoallylic alcohol proceeded with 10:1 selectivity, and subsequent
MCPBA epoxidation of the cyclohexene was also secured with 10:1 facial control. This set the
stage for the triflic anhydride-mediated closure of the six-membered ring ether. The
Nozaki-Hiyama-Kishi cyclization of 18 proceeded with remarkable selectivity, delivering
Briarellin E (19) as a single diastereomer. Dess-Martin oxidation converted 19 into Briarellin F
(4).
D. F. Taber, Org. Chem. Highlights 2010, October 4.
URL: http://www.organic-chemistry.org/Highlights/2010/04October.shtm
8. The Baran Synthesis of Vinigrol
The diterpene vinigrol (3), isolated from Virgaria nigra F-5408, has eluded total synthesis for
more than twenty years. Attempts to construct the four-carbon bridge on a preformed cis-decalin
have been unavailing. Phil S. Baran of Scripps/LaJolla solved (Angew. Chem. Int. Ed. 2008, 47,
3054, ; J. Am. Chem. Soc. 2009, 131, 17066, ) this problem by adding the extra C-C bond of 1,
that could then be cleaved in course of a Grob fragmentation, leading to 2.
The preparation of 1 started with the dihydroresorcinol derivative 4. Diels-Alder addition of the
ester 5 gave 6, with a modest 2:1 dr. Addition of allyl magnesium chloride to the derived aldehyde
7 proceeded with 6:1 dr. The resulting triene was conformationally sufficiently constrained that
cyclization to 8 proceeded at room temperature over two weeks, or more conveniently at 105°C
for 90 minutes. With 8 in hand, oxidation to the ketone allowed installation of the additional
methyl group of 9. Desilylation followed by OH-directed reduction set the relative configuration
of 1 correctly for the Grob fragmentation to the Z-alkene 2.

There were two remaining problems in the synthesis. The alkene of 2 had to be converted to the
methylated tertiary alcohol, and the ketone had to be elaborated to the ene diol. While seemingly
straightforward, the congested tricyclic skeleton of 2 made many common transformations
difficult. The solution to the former problem was found in the selective dipolar addition of
bromonitrile oxide. Reduction of the ketone then enabled HO-directed hydrogenation of the
alkene, that otherwise was resistant. Dehydration followed by reduction with LiAlH 4 gave the
desired methyl group bearing a primary amine, that was removed by free radical reduction of the
corresponding isonitrile, to give 12.
With 12 in hand, the end of the synthesis appeared to be in sight. In fact, the reduction of a variety
of oxidized intermediates proved difficult. In the end, a sequence that did not require reduction
proved effective. Dihydroxylation of 12 gave a diol, selective oxidation of which delivered the
α-hydroxy ketone 13. Formation of the trisylhydrazone followed by Shapiro reaction gave the
intermediate alkenyl anion, that was trapped with formaldehyde to give the long-sought Vinigrol
(3).
Vinigrol (3) produced by this sequence was racemic. The absolute configuration of the ring
system was set in the course of the Diels-Alder addition of 5 to 4. Use of a chiral catalyst in this
step could set the stage for an enantioselective synthesis of 3.
D. F. Taber, Org. Chem. Highlights 2010, September 6.
URL: http://www.organic-chemistry.org/Highlights/2010/06September.shtm

9. The Nicolaou Synthesis of (+)-Vannusal
The correct assignment of relative configuration for portions of a complex structure that are
remote one from another can present substantial difficulties. This was brought home in the course
of the synthesis of (+)-Vannusal (3) described (Angew. Chem. Int. Ed. 2009, 48, 5642,; 5648, ) by
K. C. Nicolau of Scripps/La Jolla. In fact, they prepared several alternative diastereomers,
including the originally assigned structure, before finally coming to 3, the spectra of which
matched those of the natural product.
Their synthetic strategy was based on the late-stage convergent coupling of the aldehyde 13 with
the iodide 19, leading to 1. The preparation of 13 began with conjugate addition of the 1-propenyl
Grignard reagent 5 to the cyclohexenone 4. Deprotection, oxidation and acetal formation led to 6,
that cyclized with high diastereocontrol to 7. Carbomethoxylation of the ketone followed by
Mn(OAc) 3 cyclization delivered the highly strained norbornane 8 as a single diastereomer.
Condensation of the derived ketone 9 with acetone (10) followed by reduction set the three
remaining ternary stereogenic centers of 13. O-Alkylation of the aldehyde 11 followed by Claisen
rearrangement established the alkylated quaternary center. Functional group manipulation then
converted 12 into 13.
The preparation of the iodide 19 began with the diene 14. Hydroboration followed by acetylation
provided the meso diol. Enzymatic hydrolysis proceeded with high enantioselectivity, to give 15.
Opening of the epoxide 16 with 2-propenyl lithium gave the trans alcohol, that was converted to
the requisite cis alcohol 17 by Mitsunobu esterification followed by hydrolysis. Shapiro iodination
of 18 then delivered 19.

The iodide 19 was enantiomerically pure, but the aldehyde 13 was racemic, so coupling of the two
led to 1 and its diastereomer. The cyclization of 1 with SmI 2 proceeded with remarkable
diastereocontrol, to give the desired 2 directly. Deprotection and oxidation then completed the
synthesis of (+)-Vannusal B (3).
It is noteworthy that throughout this synthesis, the radicals AZADO (20) and 1-Me-AZADO (21),
developed by Professor Iwabuchi ( 2010, March 8), more efficient than the traditional
TEMPO, were used to effect selective catalytic oxidation.
D. F. Taber, Org. Chem. Highlights 2010, August 2.
URL: http://www.organic-chemistry.org/Highlights/2010/02August.shtm
10. The Nicolaou Synthesis of (+)-Hirsutellone B
(+)-Hirsutellone B (3), isolated from the insect pathogenic fungus Hirsutella nivea BCC 2594,
shows good activity (MIC = 0.78 μg/mL) against Mycobacterium tuberculosis. Approaching the
synthesis of 3, K. C. Nicolaou of Scripps/La Jolla envisioned and reduced to practice (Angew.
Chem. Int. Ed. 2009, 49, 6870. ) a spectacular tandem intramolecular epoxide opening - internal
Diels-Alder cyclization (1 → 2) that established all three of the carbocyclic rings of 3 with the
proper relative and absolute configuration.

The construction of 1 began with commercial (R)-(+)-citronellal (4). Wittig homologation
established the (Z)-iodide 5. Selective ozonolysis followed by condensation with the phosphorane
7 set the stage for Jørgensen-Córdova epoxidation (Tetrahedron Lett. 2006, 47, 99.) with H 2 O 2
and a catalytic amount of the Hayashi catalyst 9. Condensation of 10 with the phosphorane 11
followed by Cu-catalyzed coupling of 12 with the organostannane 13 completed the assembly of
1.
This approach underscores the strategic advantages the Jørgensen-Córdova epoxidation has over
the Sharpless protocol. It is not necessary to reduce the aldehyde to the allyic alcohol, then
reoxidize. Further, the Jørgensen-Córdova epoxidation, using catalytic 9, is operationally easier
than the Sharpless procedure, that often uses stoichiometric amounts of tartrate ester.
The cyclization of 1 proceeded by way of 14, with the newly formed stereogenic center having the
diene equatorial on the cyclohexane. Endo cycloaddition catalyzed by the Lewis acid in the
solution then gave 2. The facility with which the cyclization of 14 set both the substituents and the
stereogenic centers of 2 raises the possibility that the biosynthesis might also follow such an
internal [4 + 2] cycloaddition.

To complete the synthesis of 3, it was necessary to construct the strained paracyclophane. The
authors took advantage of the facile cyclization of the thiolate liberated from 19, then installed the
ring-contracted alkene with a Ramburg-Bäcklund rearrangement. They completed the synthesis of
(+)-Hirsutellone B (3) by exposing the ketone 22 to NH 3 in CH 3 OH/H 2 O.
D. F. Taber, Org. Chem. Highlights 2010, July 5.
URL: http://www.organic-chemistry.org/Highlights/2010/05July.shtm
11. The Magnus Synthesis of (±)-Codeine
Although there have been many synthetic approaches to morphine and its methyl ether codeine (3),
the pentacyclic structure of these Papaver alkaloids continues to intrigue organic chemists. Philip
Magnus of the University of Texas devised (J. Am. Chem. Soc. 2009, 131, 16045.) an elegant
route to 3 based on the conversion of 1 to 2 by way of an intramolecular Michael addition.
The starting point for the synthesis was the commercial bromoaldehyde 4. Coupling with 5
delivered the substituted biphenyl 6, that was carried on to the mixed bromo acetal 8. On exposure
to fluoride ion, 8 was desilylated, and the intermediate phenoxide cyclized with impressive facility
to give 1. Exposure of 1 to nitromethane delivered the tetracyclic 2. This reaction apparently was
initiated by Henry addition of the nitromethane to the aldehyde. The intramolecular Michael
addition of the intermediate Henry adduct then proceeded to give the desired cis diastereomer of
the newly formed ring. Finally, loss of water gave 2.
Conjugate reduction of the nitroalkene 2 led to 9 with remarkable diastereocontrol. Exposure of 9
to LiAlH 4 converted the nitro group to the amine, and the enone to the allylic alcohol. On
exposure to acid, the hemiacetal was hydrolyzed. The liberated aldehyde underwent reductive

amination with the free amine, while at the same time ionic cyclization closed the ether ring.
N-Acylation completed the conversion to 10.
The ether 10 had previously been converted to codeine, and then, in a single demethylation step,
to morphine. In that synthesis, the alkene of 10 was directly epoxidized. The resulting “up”
epoxide reacted only sluggishly with phenylselenide anion, and the relative configuration of the
resulting allylic alcohol had to be inverted by oxidation followed by reduction. In the current
synthesis, exposure of the alkene 10 to dibromohydantoin under aqueous conditions, to form the
bromohydrin, effected concomitant arene bromination, to give, after base treatment, the “down”
epoxide 12. Phenylselenide opening of the epoxide was then facile, and the product allylic alcohol
had the correct relative configuration for codeine and morphine. The extra Br was of no
consequence, as it was removed by the final LiAlH 4 reduction.
Except for the stereogenic center of the acetal, the dienone 1 is prochiral. This raises the exciting
possibility that it may be possible to set the absolute configuration of codeine and thus of
morphine by asymmetric catalysis of the Henry reaction of 1.
D. F. Taber, Org. Chem. Highlights 2010, June 7.
URL: http://www.organic-chemistry.org/Highlights/2010/07June.shtm
12. The Dixon Synthesis of (-)-Nakadomarin A
(-)-Nakadomarin A (4), isolated from the sponge Amphimedon sp. off the coast of Okinama,
shows interesting antifungal and antibacterial activity. The key step in the total synthesis of 4
reported (J. Am. Chem. Soc. 2009, 131, 16632. ) by Darren J. Dixon of the University of Oxford
was the diastereoselective addition of the enantiomerically-pure ester 1 to the prochiral nitroalkene
2.

The assembly of 2 began with the linchpin ketophosphonate 5. Alkylation of the dianion of 5 with
allyl bromide followed by direct condensation of the resulting monoanion with the diacetate 6
gave 7. On exposure to aqueous acid, 7 cyclized to the furan. Oxidation of the liberated primary
alcohol followed by condensation with nitromethane then completed the preparation of 2.
The starting material for the synthesis of 1 was the enantiomerically-pure pyroglutamate
derivative 8. Sulfide displacement followed by N-alkylation with the bromide 10 delivered 11.
Oxidation followed by deprotection then set the stage for the intramolecular Julia-Kocienski
cyclization, that gave 12 with the expected (eight-membered ring) high geometric control.
Addition of the ester 1 to Michael acceptors proceeded across the open face of the lactam, but it
was still necessary to control the face of the nitro alkene 2 to which the lactam anion added.
Catalysis of the addition with the urea 13 delivered 3 with 10:1 diasterocontrol.
Mannich condensation of the nitroalkane 3 with formaldehyde and the amine 14 gave the
bis-lactam 15, conveniently as a single diastereomer. After free radical removal of the nitro group,
it was necessary to achieve selective reduction of the δ-lactam in the presence of the γ-lactam.
Low temperature LiAlH 4 was found to be effective. Direct reduction of the resulting hemiaminal
with formic acid led to the monolactam 16. The hemiaminal from monoreduction of 16 was found
to be unstable and sensitive to over-reduction. Nevertheless, exposure of 16 to Dibal at low
temperature followed by acid-mediated cyclization delivered the diamine 17.

Cyclization of the free base of 17 with the first generation Grubbs catalyst gave (-)-Nakadomarin
A (4) as the minor component of a 40:60 Z/E mixture. Carrying out the cyclization on the
camphorsulfonate salt improved the ratio to 63:37 Z/E.
D. F. Taber, Org. Chem. Highlights 2010, May 3.
URL: http://www.organic-chemistry.org/Highlights/2010/03May.shtm
13. The Corey Synthesis of (+)-Lupeol
The total synthesis of lupeol was one of the crowning achievements of the Robinson
annulation/reductive alkylation approach to stereocontrolled polycarbocyclic construction
developed by Gilbert Stork (J. Am. Chem. Soc. 1971, 93, 4945. ). It is a measure of the progress of
organic synthesis since that time that E. J. Corey of Harvard University could devise (J. Am. Chem.
Soc. 2009, 131, 13928.) an enantioselective synthesis of (+)-lupeol (3) that could be carried out by
a single co-worker. The key step in the synthesis was the Lewis acid-mediated cyclization of 1 to
2.
The preparation of 1 began with the enantioselective epoxidation of farnesol acetate (4). To this
end, asymmetric dihydroxylation delivered the diol 5. Selective mesylation followed by exposure
to dilute methoxide effected ring closure to the epoxide, but also removed the acetate, so this had
to be reapplied.

The synthesis of the aromatic portion of 1 started with the phenol 7. Protection as the very bulky
triisopropylsilyl ether was important for the success of the subsequent cyclization, perhaps
because it discouraged complexation of the Lewis acid with the aryl ether. Metalation followed by
formylation delivered the aldehyde 8, that was reduced and carried on to the bromide 9. The
derived Grignard reagent coupled smoothly with 6 under Li 2 CuCl 4 catalysis.
The cyclization of 1 to 2 proceeded with remarkable efficiency (43%!), for a reaction in which
three new carbon-carbon bonds, four rings and five new stereogenic centers were established. It is
particularly noteworthy that the cyclization cleanly set the trans, anti, trans, anti tetracyclic
backbone of (+)-lupeol (3).
To complete the synthesis of 3, the less substituted alkene of 2 was selectively hydrogenated, then
CH 3 Li was added to give 10. Hydrolysis and dehydration yielded 11, that was reduced and
equilibrated to 12. On brief exposure to MsCl/Et 3 N, 13 cyclized to (+)-lupeol (3).
It is a measure of the remarkable efficiency of this synthesis of (+)-lupeol (3) that it provided
sufficient material to enable studies of the rearrangement of 3 under acidic conditions to other

pentacyclic triterpenes, including, inter alia, germanicol, α-amyrin, β-amyrin and taraxasterol
(14).
D. F. Taber, Org. Chem. Highlights 2010, April 5.
URL: http://www.organic-chemistry.org/Highlights/2010/05April.shtm
14. The Trost Synthesis of (-)-Pseudolaric Acid B
(-)-Pseudolaric Acid B (3), isolated from the bark of the golden larch Pseudolarix kaempferi,
shows potent antifungal activity. A key step in the total synthesis of 3 described (J. Am. Chem.
Soc. 2008, 130, 16424. ) by Barry M. Trost of Stanford University was the free radical cyclization
of 1 that established the angular ester and the trans ring fusion of 2 and thus of 3.
To prepare the bicyclic skeleton of 1, the authors envisioned the Rh-mediated intramolecular
addition of the alkyne of 11 to the alkenyl cyclopropane. The acyclic centers of 11 were
established by Noyori hydrogenation of (equilibrating) racemic 4. One enantiomer reduced much
more quickly than did the other, leading to 5. The absolute configuration of the cyclopropane was
set by Charette cyclopropanation of the monosilyl ether of the inexpensive diol 8. The two
components were then coupled using a Corey-Schlosser protocol. Alkylation of the ylide 10 with
7 gave a new phosphonium salt, that in situ was deprotonated and condensed with the aldehyde 9.
The resulting betaine was deprotonated and quenched, then exposed again to base to give the trans
alkene 11. It is important in this procedure to use PhLi as the base, since the alkyl lithium can
displace the alkyl group on phosphorus.
The product from Ru-catalyzed cyclization was the expected 1,4-diene 12. Fortunately, it was
found that TBAF desilylation led to concomitant alkene migration, to give the more stable
conjugated diene 13. Selective epoxidation of the more electron-rich alkene followed by exposure
to strong base then delivered 14, with the requisite angular oxygenation established.

Pseudolaric Acid B (3) would be derived from cyclization of the selenocarbonate of a tertiary
alcohol. In fact, however, attempted cyclization of such selenocarbonates led only to
decarboxyation and reduction. Even with the selenocarbonate 1 prepared from the secondary
alcohol, the cyclization to 2 required careful optimization, including using not AIBN, but
azobis(dicyclohexylcarbonitrile) as the radical initiator.
Acetylide addition to the ketone 15 could be effected with high diastereocontrol, but lactone
construction proved elusive. Alkaline conditions led quickly to addition of the angular hydroxyl to
the activated alkene in the seven-membered ring. Eventually it was found that the ester exchange
catalyst 16 developed by Otera delivered 17, that could be carried on to (-)-Pseudolaric Acid B
(3).
D. F. Taber, Org. Chem. Highlights 2010, March 1.
URL: http://www.organic-chemistry.org/Highlights/2010/01March.shtm
15. The Nakada Synthesis of (-)-FR182877
The Streptomyces metabolite (-)-FR182877 (3) binds to and stabilizes microtubules, showing the
same potency of anticancer activity as Taxol. Masahisa Nakada of Waseda University assembled
(Angew. Chem. Int. Ed. 2009, 48, 2580.) the hexacyclic ring system of 3 by the tandem
intramolecular Diels-Alder – intramolecular hetero Diels-Alder cyclization of 1, generating seven
new stereogenic centers in a single step.
The construction of the pentaene substrate 1 started with the known aldehyde 4, prepared by
homologation of commercial ethyl 3-methyl-4-oxocrotonate. Addition of the propionyl

oxazolidine anion 5 proceeded with high diastereocontrol, to give 6. The acyl oxazolidinone was
not an efficient acylating agent, so it was converted to the Weinreb amide. Protection and
deprotection then delivered the allylic acetate 7.
The key step in the pentaene assembly was the carefully optimized Negishi-Wipf methylation of 8,
followed by Pd-mediated coupling of the alkenyl organometallic so generated with the allylic
acetate, to give 9. Condensation of the derived keto phosphonate 11 with the known aldehyde 12
then delivered the enone 13.
The Nakada group has worked extensively on the intramolecular Diels-Alder reaction of
substrates such as 1. They have shown that protected anti diols such as 1 cyclize with substantial
diastereocontrol and in the desired sense. In contrast, cyclizations of protected syn diols proceed
with poor diastereocontrol. The enone 13 was therefore reduced to the anti diol and protected,
leading to 14. Oxidation of 14 at room temperature led to a complex mixture, but slow oxidation
at elevated temperature delivered 2. Although the yield of 2 was not much better than if the
reactions were carried out sequentially, first the intramolecular Diels-Alder cyclization, then the
intramolecular hetero Diels-Alder cyclization, with the cascade protocol pure 2 was more readily
separated from the reaction matrix.
With 2 in hand, there was still the challenge of assembling the seven-membered ring. Cyclization
was effected with an intramolecular Heck protocol. The two diastereomers of the allylic alcohol
15 cyclized with comparable efficiency. Ir-catalyzed alkene migration then converted the allylic
alcohols to a mixture of ketones, that was equilibrated to give the more stable diasteromer.

Reduction of the ketone then set the last stereogenic center of 3. Deprotection and subsequent
lactone formation completed the synthesis of (-)-FR182877 (3).
D. F. Taber, Org. Chem. Highlights 2010, February 1.
URL: http://www.organic-chemistry.org/Highlights/2010/01February.shtm
16. The Williams Synthesis of (-)-4-Hydroxydictyolactone
(-)-4-Hydroxydictyolactone (3), representative of the cyclononene xenicanes isolated from the
Dictyotacae algae, readily isomerizes thermally to the more stable (Z)-6,7-isomer. Attempts to
directly form this strained ring system appeared to be fraught with difficulty. David R. Williams
of Indiana University envisoned (J. Am. Chem. Soc. 2009, 131, 9038.) that use of Suzuki coupling
might ameliorate some of the strain, since at the point of commitment to bond formation, the Pd
center would be included in the forming ring. This analysis led specifically to the trans ether 1, as
cyclization of the trans ether appeared likely to be more facile than would cyclization of the
alternative cis diastereomer.
The first challenge was the assembly of the array the four contiguous alkylated stereogenic centers
of 1. To this end, the Z secondary ester 7 was prepared from the acetonide 4, available from
mannitol, and (R)-(+)-citronellic acid, prepared by oxidation of the commercial aldehyde.
Addition of 7 to LDA led to decomposition, but inverse addition of LDA to a mixture of the ester,
TMSCl and Et 3 N smoothly delivered the ketene silyl acetal. On warming, Ireland-Claisen
rearrangement of the ketene silyl acetal led to the acid 8 with remarkable diastereocontrol.
The last alkylated stereogenic center of 1 was installed by reductive cyclization of the formate
ester 9. Again, the cyclization proceeded with remarkable diastererocontrol. While the

intramolecular reaction of in situ prepared allyl metals is well precedented, the addition to a
formate ester had not previously been reported.
Although 11 appears to be ready for the long-awaited Suzuki coupling, in fact the TIPS protecting
group substantially slowed hydroboration. The free alcohol/methyl acetal was the best substrate
for hydroboration, but the free alcohol entered into other side reactions. After extensive
experimentation, a happy medium was found with the methyl acetal/TBS ether 1.
Selenylation of the lactone 12 followed by oxidative elimination of the selenide delivered the
expected Z alkene. Removal of the silyl protecting group had to precede introduction of the second
alkene, as the product 3 deteriorated rapidly on exposure to the alkaline conditions of TBAF
cleavage.
Strained medium rings such as the (E,Z)-cyclononadiene assembled in this study have been
particularly challenging to prepare. It will be interesting to see how generally useful the Suzuki
coupling turns out to be for the construction of such rings.
D. F. Taber, Org. Chem. Highlights 2010, January 4.
URL: http://www.organic-chemistry.org/Highlights/2010/04January.shtm
2009 年
17. The Davies/Williams Synthesis of (-)-5-epi-Vibsanin E
There are currently 61 known vibsane-type diterpenes, as exemplified by (-)-5-epi-Vibsanin E (3).
The first synthesis of 3, described (J. Am. Chem. Soc. 2009, 131, 8329. ) by Huw M. L. Davies of
Emory University and Craig M. Williams of the University of Queensland, was based on the
enantioselective seven-membered ring construction developed by the Davies group and the end
game established by the Williams group. A key step in the synthesis was the intramolecular hetero
Diels-Alder cyclization of 1 to 2.

The absolute configuration of 1 was set by the Rh-mediated cyclopropanation of 4 with the diazo
ester 5. Though closely related to the α-diazo β-keto ester 6, the alkene of 5 donates electron
density to the intermediate Rh carbene, making it more susceptible to the influence of the chiral
ligands. The alkene of the enol ether then participated in the Cope rearrangement, delivering 8.
Routine functional group transformation then converted 8 to 1, that cyclized smoothly to 2.
The enol ether of 2 was reduced with high diastereocontrol to give 10. The ketone was installed by
allylic oxidation, setting the stage for attachment of the two pendant sidechains of 3 by conjugate
addition followed by enolate trapping. Cu-catalyzed addition of the α-oxygenated organolithium
12 proceeded well in the presence of TMS-Cl, to establish the silyl enol ether 13. Allylation of the
regenerated enolate proceeded at oxygen, but the enol ether 14 so prepared rearranged to the
desired C-alkylated product 15 on microwave heating.
The synthesis endgame was based on an unusual transformation, the addition to the keto aldehyde
16 of the phosphonium salt 17, developed (Tetrahedron 2008, 64, 6482.) by the Williams group.
This allowed the introduction of the complete vinyl ester array of (-)-5-epi-Vibsanin E (3).

This synthesis illustrates the power of the elegant enantioselective seven-membered ring
construction developed by the Davies group. The Williams phosphonium salt will also have
general applicability. In a simpler manifestation, conversion of an aldehyde to, e.g., the enol
benzoate, followed by exposure to dilute methoxide, will allow the conversion of an aldehyde to
the aldehyde one carbon longer, without the acidic hydrolysis usually required for such a
transformation.
D. F. Taber, Org. Chem. Highlights 2009, December 7.
URL: http://www.organic-chemistry.org/Highlights/2009/07December.shtm
18.The Kobayashi Synthesis of (-)-Norzoanthamine
The Zoanthus alkaloids, exemplified by (-)-norzoanthamine (3a) and zoanthamine (3b), show
promising activity against osteoporosis. Susumu Kobayashi of the Tokyo University of Science
assembled (Angew. Chem. Int. Ed. 2009, 48, 1400, ; Angew. Chem. Int. Ed. 2009, 48, 1404, ) the
challenging tricyclic core of 3a employing the intramolecular Diels-Alder cyclization of 1 to 2.
The cyclopentane of 1 served as useful scaffolding, even though it was cleaved en route to 3a.
The cyclohexane ring of 1 has five of its six positions substituted, including three that are
alkylated quaternary centers. The starting point for the preparation of 1 was the
enantiomerically-pure Hajos-Parrish ketone 4, containing the first of the those quaternary centers.
Conjugate addition of MeLi established the second quaternary center. The less stable endo alkyl
branch of 1 was installed by conjugate addition to the more reactive α-methylene ketone of the
cross-conjugated 5, followed by kinetic quench. Addition of vinyl cuprate across the open face of
the enone 7 then established the final quaternary center, setting the stage for the intramolecular
Diels-Alder reaction. The silyl enol ether from the cyclization of 1 was not stable, so it was
directly oxidized to the enone 2.

The keto phosphonate 16 for the last two rings of 3a was prepared from the previously-reported
crystalline glutamic acid-derived mesylate 12. Reduction and homologation delivered the ester 14,
that was condensed with the phosphonate anion 15 to give 16.
The congested cyclopentanone 17, derived from 2, was most efficiently deprotonated with n-BuLi.
Exposure of the resulting silyl enol ether to ozone led to the α-hydroxylated product 18.
Unexpectedly but happily, oxidative cleavage of 18 delivered, after deprotection and reprotection,
the more congested aldehyde 19. This cleavage may be proceeding by tautomerization of 18 to the
regioisomeric keto alcohol. The aldehyde 19 was condensed with the keto phosphonate 16, to give,
after hydrogenation, the keto lactone 20. A series of oxidation state adjustments then completed
the synthesis of (-)-norzoanthamine (3a).
The preparation of 3a outlined here underlines the importance of developing new methods for
concise stereocontrolled carbocyclic construction. The utility of an enantiomerically-pure bicyclic
scaffold such as 4 for subsequent relative stereocontrol is particularly apparent
D. F. Taber, Org. Chem. Highlights 2009, November 2.
URL: http://www.organic-chemistry.org/Highlights/2009/02November.shtm

19.The Castle Synthesis of (-)-Acutumine
The complex tetracyclic alkaloid (-)-acutumine 3, isolated from the Asian vine Menispermum
dauricum, shows selective T-cell toxicity. The two adjacent cyclic all-carbon quaternary centers of
3 offered a particular challenge. Steven L. Castle of Brigham Young University solved (J. Am.
Chem. Soc. 2009, 131, 6674.) this problem by effecting net enantioselective conjugate allylation
of the enantiomerically pure substrate 1 to give 2 with high diastereocontrol.
The starting coupling partners (Org. Lett. 2006, 8, 3757,; Org. Lett. 2007, 9, 4033,) for the
synthesis were the Weinreb amide 4, prepared over several steps from 2,3-dimethoxyphenol, and
the diastereomerically- and enantiomerically-pure cyclopentenyl iodide 5, prepared by singlet
oxygenation of cyclopentadiene followed by enzymatic hydrolysis. Transmetalation of 5 by the
Knochel protocol, addition of the resulting organometallic to 4 and enantioselective (and therefore
diastereoselective) reduction of the resulting ketone delivered the alcohol 6. Methods for installing
cyclic halogenated stereogenic centers are not well developed. Exposure of the allylic alcohol to
mesyl chloride gave the chloride 7 with inversion of absolute configuration. Remarkably, this
chlorinated center was carried through the rest of the synthesis without being disturbed.
A central step in the synthesis of 3 was the spirocyclization of 7 to 8. Initially, iodine atom
abstraction generated the aryl radical. The diastereoselectivity of the radical addition to the
cyclopentene was set by the adjacent silyloxy group. The α-keto radical so generated reacted with
the Et 3 Al to give a species that was oxidized by the oxaziridine to the α-keto alcohol, again with
remarkable diastereocontrol.
Conjugate addition to the cyclohexenone 1 failed, so an alternative strategy was developed,
diastereoselective 1,2-allylation of the ketone followed by oxy-Cope rearrangement. The
stereogenic centers of 1 are remote from the cyclohexenone carbonyl, so could not be used to
control the facial selectivity of the addition. Fortunately, the stoichiometric enantiomerically-pure

Nakamura reagent delivered the allyl group preferentially to one face of the ketone 1, to give 9.
The subsequent sigmatropic rearrangement to establish the very congested second quaternary
center of 2 then proceeded with remarkable facility, at 0°C for one hour.
Oxidative cleavage to the aldehyde followed by reductive amination gave 10, that looks as though
it could be poised for intramolecular displacement of the secondary chloride. Nonetheless, Lewis
acid mediated ionization followed by cyclization proceeded smoothly, to establish the fourth ring
of the natural product. Oxidation state adjustment then completed the synthesis of (-)-Acutumine
(3).
The face selective enone allylation followed by oxy-Cope rearrangement (1 → 2), a highlight of
the approach presented here, will have many applications in target-directed synthesis.
20.The Trost Synthesis of (-)-Ushikulide A
(-)-Ushikolide A (4), isolated from a culture broth of Streptomyces sp. IUK-102, showed powerful
activity against murine splenic lymphocyte proliferation (IC 50 = 70 nM). The most important
player in the synthesis of 4 described (J. Am. Chem. Soc. 2008, 130, 16190. ) by Barry M. Trost of
Stanford University was the ProPhenol ligand 1.
The precursor 2 was prepared by coupling the mesylate 7, the alkyne 12, and the aldehyde 13. The
first role of catalyst 1 was in mediating the enantioselective coupling of commercial 5 with 6 to
give, after saponification and CuCl decarboxylation, the mesylate 7. The preparation of 12 began
with the Noyori hydrogenation of the ester 8 to the alcohol 9 in the expected high ee. Note that
although this transformation was carried out at 1800 psi, such reductions proceed well and in

similar ee at 60°C and 60 psi. Brown crotylation of the derived aldehyde 10 delivered 11, that was
homologated to the alkyne 12.
The third fragment 13 was prepared by chiral auxiliary directed aldol condensation. Combination
of 12 with 13 was followed by Au-mediated cyclization, converting the internal alkyne of 14 to
the spiroketal of 15. Pd-catalyzed coupling of 15 with 7 then led to 2 with high diastereocontrol.
The aldol addition of the enolate of 17 to 18 proved elusive under the usual conditions, but with
30 mol % of the Zn catalyst 1 the reaction proceeded smoothly, to deliver 19 with high
diastereocontrol.
To complete the synthesis, hydroboration with 9-BBN was effected on the free carboxylic acid 3,
and Pd-mediated coupling of the derived borane was carried out with the free iodo alcohol 2. As a
result, the product hydroxy acid 20 could be taken directly to the subsequent macrolactonization.

21. The Overman Syntheses of Nankakurines A and B
The tetracyclic alkaloids Nankakurine A and Nankakurine B were isolated from the club moss
Lycopodium hamiltonii. A preliminary study of the biological activity of Nankakurine A
suggested that it could induce secretion of neurotrophic factors and promote neuronal
differentiation. The key step in the first syntheses of Nankakurine A and of Nankakurine B,
reported (J. Am. Chem. Soc. 2008, 130, 11297.) by Larry E. Overman of the University of
California, Irvine was the intriguing intramolecular aza-Prins cyclization of 1 to 2.
The starting material for the synthesis was 5-methyl cyclohexenone 6, prepared from (R)-pulegone.
The diene 5 was prepared from the alkyne 4, following the procedure developed by Diver. There
were two issues in developing the Diels-Alder addition of the enone 4 to the diene 6. The first was
the relative lack of reactivity of 4 as a dienophile. The other issue was the ready epimerization of
the product ketone 9. Both of these problems were solved using the activation method devised by
Gassman. Condensation of 4 with 7 in the presence of the bis-silyl ether 7 and the diene 6 at
cryogenic temperatures led to the ketal 8. It is thought that the active dienophile was the cation 11.
Gentle hydrolysis of the ketal 8 was effected with minimal epimerization. Reductive amination
with the hydrazide 10 proceeded with high diastereocontrol, to give the precursor 1.
The intramolecular aza-Prins cyclization of 1 to 2 proceeded well, though the desired tetracyclic 2
was only observed when base was included in the reaction medium. In the absence of base,
tricyclic alkenes dominated.

Reduction of the N-N bond of 2 proceeded smoothly with freshly prepared SmI 2 . After reductive
methylation, hydrogenation removed the benzyl ether, and AlH 3 converted the benzamide to the
benzyl amine. At low temperature, mesylation of the alcohol was apparently faster than
mesylation of the secondary amine, enabling cyclization to 14. Removal of the benzyl protecting
group gave Nankakurine A, which was successfully methylated to give Nankakurine B.
The completion of a total synthesis is an anxious moment, as for the first time it is possible to
compare spectra of the synthetic material with those reported for the natural product. There is
always a concern as to whether or not the spectra are being acquired under precisely the same
conditions employed by those who did the initial isolation. This is particularly true for very polar
molecules such as these diamines. In fact, the spectra in CD 3 OD did not initially match, but on the
addition of small amounts of CF 3 CO 2 H they were brought into congruence.
Although in this synthesis the starting enantiomerically-pure cyclohexenone 4 was derived from
natural sources, one could imagine that enantioselective conjugate methylation of cyclohexenone
or a derivative could get one into the same manifold.
22. The Keck Synthesis of Epothilone B
The total synthesis of Epothilone B (4), the first natural product (with Epothilone A) to show the
same microtubule-stabilizing activity as paclitaxel (Taxol®), has attracted a great deal of attention
since that activity was first reported in 1995. The total synthesis of 4 devised (J. Org. Chem. 2008,
73, 9675.) by Gary E. Keck of the University of Utah was based in large part on the
stereoselective allyl stannane additions (e.g. 1 + 2 → 3) that his group originated.
The allyl stannane 2 was prepared from the acid chloride 5. Exposure of 5 to Et 3 N generated the
ketene, that was homologated with the phosphorane 6 to give the allene ester 7. Cu-mediated
conjugate addition of the stannylmethyl anion 8 then delivered 2.

The benzyloxy aldehyde 1 was prepared from the ester 9 by reduction with Dibal.
Felkin-controlled 1,2-addition of the allyl stannane 2 established the relative configuration of the
secondary alcohol of 3, that was then used to control the relative configuration of the new alcohol
in 10. Addition of the crotyl borane 12 to the derived aldehyde 11 also proceeded with high
diastereocontrol.
The other component of 4 was prepared from the aldehyde 14. Enantioselective allylation, by the
method the authors developed, delivered the alcohol 16. The Z trisubstituted alkene was then
assembled by condensing the aldehyde 17 with the phosphorane 18. Dibal reduction of the product
lactone 19 gave a diol, the allylic alcohol of which was selectively converted to the chloride with
the Corey-Kim reagent. Hydride reduction then delivered the desired homoallylic alcohol, that
was converted to the phosphonium salt 21. Condensation of 21 with 13 gave the diene, that was
carried on to Epothilone B (4).
The synthesis of Epothilone B (4) as originally conceived by the authors depended on ring-closing
metathesis of the triene 22. They prepared 22, but on exposure to the second-generation Grubbs
catalyst it was converted only to 23. The authors concluded that the trans acetonide kept 22 in a
conformation that did not allow the desired macrocyclization.

23. The Johnson Synthesis of Zaragozic Acid C
The zaragozic acids, exemplified by Zaragozic Acid C (3), are picomolar inhibitors of cholesterol
biosynthesis. Jeffrey S. Johnson of the University of North Carolina developed (J. Am. Chem. Soc.
2008, 130, 17281.) an audacious silyl glyoxylate cascade approach to the oxygenated backbone
fragment 1. Intramolecular aldol cyclization converted 1 to 2, setting the stage for the construction
of 3.
The lactone 2 includes five stereogenic centers, two of which are quaternary. The authors were
pleased to observe that exposure of 4 to vinyl magnesium bromide (5) led, via condensation, silyl
transfer, condensation, and again silyl transfer, to a species that was trapped with t-butyl
glyoxylate (6) to give 7 as a single diastereomer. This one step assembled three of the stereogenic
centers of 2, including both of the quaternary centers. The alcohol 7 so prepared was racemic, so
the wrong enantiomer was separated by selective oxidation. Intramolecular aldol condensation of
the derived α-benzyloxy acetate 1 then completed the construction of 2.
Addition of the alkyl lithium 8, again as a single enantiomerically-pure diasteromer, to 2 gave the
hemiketal 9. Exposure of 9 to acid initially gave a mixture of products, but this could be induced
to converge to the tricyclic ester 10. To convert 10 to 11, the diastereomer that was needed for the
synthesis, two of the stereogenic centers had to be inverted. This was accomplished by exposure to
t-BuOK/t-amyl alcohol, followed by re-esterification. The inversion of the secondary hydroxyl
group was thought to proceed by retro-aldol/re-aldol condensation.

Debenzylation of 11 followed by acetylation delivered 12, an intermediate in the Carreira
synthesis of the zaragozic acids. Following that precedent, the ring acetates of 12 were selectively
removed, leaving the acetate on the side chain. Boc protection was selective for the endo ring
secondary hydroxyl, leaving the exo ring secondary hydroxyl available for condensation with the
enantiomerically-pure acid 13. Global deprotection then completed the synthesis of Zaragozic
Acid C (3).
The key to the success of this synthesis of the complex spiroketal 3 was the assembly of 7 in one
step as a single diastereomer from the readily-available building blocks 4, 5, and 6. This process,
reminiscent of group transfer polymerization, will be a useful complement to the cascade
organocatalyzed aldol condensations that have recently been developed.
24. The Carter Synthesis of (-)-Lycopodine
Rich G. Carter of Oregon State University described (J. Am. Chem. Soc. 2008, 130, 9238 ) the first
enantioselective synthesis of the Lycopodium alkaloid (-)-lyopodine (3). A key step in the
assembly of 3 was the diastereoselective intramolecular Michael addition of the keto sulfone of 1
to the enone, leading to the cyclohexanone 2.
The key cyclization substrate 1 bore a single secondary methyl group. While that could have been
derived from a natural product, it was operationally easier to effect chiral auxiliary controlled
conjugate addition to the crotonyl amide 4, leading, after methoxide exchange, to the ester 5. The
authors reported that double deprotonation with LiTMP gave superior results, vs. LDA or BuLi, in
the condensation of 6 with 5 to give 7. Metathesis with pentenone 8 gave the intramolecular
Michael substrate 1.

The authors thought that they would need a chiral catalyst to drive the desired stereocontrol in the
cyclization of 1 to 2. As a control, they tried an achiral base first, and were pleased to observe the
desired diastereomer crystallize from the reaction mixture in 89% yield. The structure of 2 was
confirmed by X-ray crystallography.
To prepare for the intramolecular Mannich condensation, the azide was reduced to give the imine,
and the methyl ketone was converted to the silyl enol ether. Under Lewis acid conditions, the
sulfonyl group underwent an unanticipated 1,3-migration, to give 11. Cyclization of 12 then
delivered the crystalline 14. Reduction converted 14 to the known (in racemic form) ketone 15.
To complete the synthesis, the amine 15 was alkylated with 16 to give the alcohol 17. Oppenauer
oxidation followed by aldol condensation delivered the cyclized enone, that was reduced with the
Stryker reagent to give (-)-Lycopodine (3).
Both the cyclization of 1 to 2 and the cyclization of 9 to 14 are striking. It may be that the steric
demand of the phenylsulfonyl group destabilizes the competing transition state for the cyclization
of 1.
25.

The Hoveyda Synthesis of (-)-Clavirolide C
Conjugate addition-enolate trapping, a strategy originally developed by Gilbert Stork, has become
a powerful method for stereocontrolled ring construction. A key step in the synthesis of
(-)-Clavirolide C (3) reported (J. Am. Chem. Soc. 2008, 130, 12904) by Amir H. Hoveyda of
Boston College occurred early on, with the enantioselective conjugate addition of Me 3 Al to 1 to
give the silyl enol ether 2. Enantioselective conjugate addition to establish a quaternary center β
on a cyclohexanone had been established (2008, August 18), but not yet on cyclopentanones.
Professor Hoveyda found that a modified form of the Ag catalyst that they had published earlier,
in combination with the Lewis acidic AlMe 3 , effected conjugate addition to 1 in 84% ee.
Quenching of the reaction mixture with triethylsilyl triflate led to the enol silyl ether 2.
The assembly of the 11-membered ring of 3 also began with an enantioselective conjugate
methylation, of the lactone 4 with Me 2 Zn, again using a catalyst developed by Professor Hoveyda.
Opening of the lactone 5 followed by Swern oxidation gave the Weinreb amide 6, that was
homologated and reduced to give 7.
Addition of n-BuLi to 2 regenerated the enolate. There were two issues in the addition of that
enolate to the aldehyde 7: syn vs. anti stereocontrol, and control of the configuration of the newly
formed ternary center on the ring relative to the already-established quaternary center. Inclusion of
Et 3 B in the reaction mixture assured anti aldol formation, but there was only a modest preference
for the desired bond formation trans to the slightly more bulky butenyl group, to give 8.
Medium rings are more strained than are larger rings. The diene 8 was reluctant to close with the
second generation Grubbs catalyst, but the catalyst developed by Professor Hoveyda worked well.
The δ-lactone of 3 was then constructed by acylation of 9 with 10 followed by reductive
cyclization with SmI 2 . Conjugate addition to the derived enone 12 on the outside face of the
medium ring alkene gave the desired 13 (9:1 dr). This reaction may be proceeding via the s-cis
conformer, as the more stable s-trans conformer would have been expected to give the other
diastereomer. Dehydration of 13 then delivered (-)-Clavirolide C (3).

This concise synthesis of the dolabellane 3 showcases the power of the catalytic enantioselective
methods for the construction of both ternary and quaternary, including cyclic quaternary, centers
that Professor Hoveyda has developed. Clearly, asymmetric transformation of inexpensive
prochiral ring precursors such as 1 and 4 will make advanced, high ee intermediates such as 2 and
5 much more readily available than they have been in the past.
26. The Zakarian Synthesis of (+)-Pinnatoxin A
(+)-Pinnatoxin A (3), isolated from the shellfish Pinna muricata, is thought to be a calcium
channel activator. A key transformation in the synthesis of 3 reported (J. Am. Chem. Soc. 2008,
130, 3774.) by Armen Zakarian, now at the University of California, Santa Barbara, was the
diastereoselective Claisen rearrangement of 1 to 2.
The alcohol portion of ester 1 was derived from the aldehyde 4, prepared from D-ribose. The
absolute configuration of the secondary allylic alcohol was established by chiral amino alcohol
catalyzed addition of diethyl zinc to the unsaturated aldehyde 5.
The acid portion of the ester 1 was prepared from (S)-citronellic acid, by way of the Evans imide 7.
Methylation proceeded with high diasterocontrol, to give 8. Functional group manipulation
provided the imide 9. Alkylation then led to 10, again with high diastereocontrol. In each case,
care had to be taken in the further processing of the α-chiral acyl oxazolidinones. Direct NaBH 4

eduction of 8 delivered the primary alcohol. To prepare the acid 10, the alkylated acyl
oxazolidinone was hydrolyzed with alkaline hydrogen peroxide.
On exposure of the ester 1 to the enantiomerically-pure base 11, rearrangement proceeded with
high diastereocontrol, to give the acid 2. This outcome suggests that deprotonation proceeded to
give the single geometric form of the enolate, that was then trapped to give specifically the ketene
silyl acetal 12. This elegant approach is dependent on both the ester 1 and the base 11 being
enantiomerically pure.
The carbocyclic ring of pinnatoxin A (3) was assembled by intramolecular aldol condensation of
the dialdehyde 11. This outcome was remarkable, in that 11 is readily epimerizable, and might
also be susceptible to β-elimination. Note that the while the diol corresponding to 11 could be
readily oxidized to 11 under Swern conditions, attempts to oxidize the corresponding hydroxy
aldehyde were not fruitful.
27. The Paquette Synthesis of Fomannosin
The compact sesquiterpene (+)-fomannosin (3), isolated from the pathogenic fungus Fomes
annonsus, presents an interesting set of challenges for the organic synthesis chemist, ranging from
the strained cyclobutene to the easily epimerized cyclopentanone. In the synthesis of 3 developed
(J. Org. Chem. 2008, 73, 4548.) by Leo A. Paquette of Ohio State University, the cyclopentane

was constructed by ring-closing metathesis of 1. The real challenge of the synthesis was the
enantiospecific preparation of 1 from D-glucose.
The starting point for the preparation of 1 was the glucose derivative 4. Selective acetonide
hydrolysis followed by oxidative cleavage gave the ester 5, which on base treatment followed by
hydrogenation delivered the endo ester 6. Condensation of the enolate of 6 with formaldehyde
proceeded with high diastereoselectivity, to give, after protection, the ester 7. Conversion of the
ester to the vinyl group, exposure to methanolic acid and ether formation completed the
preparation of 9.
The construction of the cyclobutane of 1 was effected by an interesting application of the Negishi
reagent (Cp 2 ZrCl 2 /2 x BuLi). Complexation of Cp 2 Zr with the alkene followed by elimination
generated an allylic organometallic 11, which added to the released aldehyde to give the
cyclobutanes 12 and 13 in a 2.4:1 diastereomeric ratio.
Homologation of the aldehyde 13 and subsequent oxidation were straightforward, but subsequent
methylenation of the hindered carbonyl was not. At last, it was found that Peterson olefination
worked well. Metathesis then delivered the cyclopentene 2. The last carbons of the skeleton were
added by intramolecular aldol cyclization of the thioester 16.

The seemingly simple task of converting the alkene of 17 into a ketone proved challenging.
Eventually, dihydroxylation followed by oxidation, and then SmI 2 reduction, completed the
transformation. This still left the challenge of controlling the cyclopentane stereogenic center.
Remarkably, dehydration and epimerization led to (+)-Fomannosin (3) as a single dominant
diastereomer.
28. The Wood Synthesis of Welwitindolinone A Isonitrile
Welwitindolinone A Isonitrile (3) is the first of a family of oxindole natural products isolated from
the cyanobacteria Hapalosiphon welwischii and Westiella intricate on the basis of their activity for
reversing multiple drug resistance (MDR). A key transformation in the total synthesis of 3
reported (J. Am. Chem. Soc. 2008, 130, 2087) by John L. Wood, now at Colorado State University,
was the chlorination of 1, that in one step established both the axial secondary chloro substituent
and the flanking chiral quaternary center.
The starting material for the synthesis of 3 was the diene acetonide 5, readily prepared from the
Birch reduction product 4. Intermolecular ketene cycloaddition proceeded with high regio- and
diastereoselectivity, to give the bicyclooctenone 6.

The triazene-bearing Grignard reagent 7 added to the ketone 6 with the anticipated high
diastereocontrol, to give, after reduction and protection, the cyclic urethane 8. Selective oxidation
of the diol derived from 8 followed by silylation delivered the enone 9. Conjugate addition of
hydride followed by enolate trapping gave the triflate 10. Pd-catalyzed methoxycarbonylation
established the methyl ester 11. Addition of CH 3 MgBr to 11 gave 1, setting the stage for the
establishment of the two key stereogenic centers of 2 and so of 3.
The transformation of 1 to 2 was envisioned as being initiated by formation of a bridging
chloronium ion. Pinacol-like 1,2-methyl migration then proceeded to form the trans diaxial
product, moving the ketone-bearing branch equatorial. In addition to being an elegant solution of
the problem of how to establish the axial chloro substituent of 3, this strategy might have some
generality for the stereocontrolled construction of other alkylated cyclic quaternary centers.
Reduction of the ketone 2 and dehydration of the resulting alcohol led, after deprotection and
oxidation, to the ketone 12. Protection followed by β-elimination gave the enone 13. Direct
reductive amination of 13 failed, but reduction of the methoxime was successful, giving, after
acylation, the formamide 14. Reductive N-O bond cleavage followed by deprotection and
isonitrile formation then set the stage for the planned intramolecular acylation to complete the
synthesis of Welwitindolinone A Isonitrile (3).
The starting diene 5 used in this synthesis was prochiral, leading to racemic 3. Now that the route
to 3 is established, it would be interesting to devise a method for preparing
enantiomerically-enriched 6. Enantiomerically-pure variants of 5 have been prepared, inter alia by

fermentation of halogenated aromatics. Alternatively, an enantioselective version of the [2+2]
cycloaddition to the prochiral 5 could be developed.
2008
27. The Takayama Synthesis of (-)-Cernuine
(-)-Cernuine (3) falls in the subset of the Lycopodium alkaloids that feature a bicyclic aminal core.
There had not been a total synthesis of this class of alkaloids until the recent (Org. Lett. 2008, 10,
1987.) work of Hiromitsu Takayama of Chiba University. The key step in this synthesis was a
diastereoselective intramolecular reductive amination, converting 1 to 2. As is apparent from the
3-D projection, (-)-cernuine (3) has a tricyclic trans-anti-trans aminal core, with an appended
six-membered ring, both branches of which are axial on the core. While the branch that is part of
the aminal could be expected to equilibrate, the other branch had to be deliberately installed.
The synthesis began with (+)-citronellal (4), each enantiomer of which is commercially available
in bulk. After protection and ozonolysis, the first singly-aminated stereogenic center was installed
by enantioselective, and therefore diastereoselective, addition of 5 to the azodicarboxylate 6,
mediated by the organocatalyst 7. Reductive cleavage of the N-N bond followed by acetal
methanolysis converted 8 to 9. Ionization followed by allyl silane addition then delivered 11,
having the requisite axial alkyl branch.
The next two tasks were the assembly of the second of the four rings of 3, and the construction of
the second single-aminated stereogenic center. The ring was assembled by deprotection of 11
followed by acylation with acryloyl chloride, to give 12. Grubbs cyclization followed by
hydrogenation then led to 13. Homologation of 13 to the aldehyde 14 set the stage for
condensation with the camphor-derived tertiary amine 15, following the protocol developed by
Kobayashi. Sequential imine formation, aza-Cope rearrangement, and hydrolysis led to 1 in 94%
de.

One could envision reduction of the lactam carbonyl of 1 to an aldehyde equivalent, that would
then, under acidic conditions, condense to form the desired aminal 2. This approach was, however,
not successful. As an alternative, conditions were developed to convert 1 to the amidine 16.
Reduction then proceeded with the expected high diastereocontrol, to give the cis 1,3-fused aminal
2. This was not isolated, but was directly acylated with acryloyl chloride, to 17.
The synthesis of (-)-cernuine 3 was concluded by Grubbs cyclization of 17 to 3, followed again by
hydrogenation. Note that there was a key difference between this cyclization and the Grubbs
cyclization of 12 that led to 13, in that 17 contained a basic N, while 12 did not. For the
cyclization of 12, the first generation Grubbs catalyst was sufficient, while for the cyclization of
17, the second generation catalyst was required.
This synthesis illustrates the efficacy of the Grubbs cyclization for polycyclic construction. The
approach outlined here also highlights the power of current methods for enantioselective allylation
of imines for the construction of enantiomerically pure, and, in the context of this synthesis,
diastereomerically pure, aminated secondary stereogenic centers.
28. The Roush Synthesis of (+)-Superstolide A
(+)-Superstolide A (3), isolated from the New Caledonian sponge Neosiphonia superstes, shows
interesting cytotoxicity against malignant cell lines at ~ 4 ng/mL concentration. The key
transformation in the synthesis of 3 described (J. Am. Chem. Soc. 2008, 130, 2722. ) by William R.
Roush of Scripps Florida was the transannular Diels-Alder cyclization of 2, which established, in
one step with high diastereocontrol, both the cis decalin and the macrolactone of 3.

The octaene 1 was assembled from four stereodefined fragments. The first, the linchpin 6, was
prepared from the stannyl aldehyde 4. Homologation gave the enyne 5, which on hydroboration
and oxidation gave 6.
Earlier, Professor Roush had optimized the crotylation of the protected alaninal 7. In this case, the
Brown reagent 8 delivered the desired Felkin product 9. Protection followed by ozonolysis gave
the aldehyde 10. Crotylation with the Roush-developed tartrate 11 then gave the alkene 12, setting
the stage for conversion to the iodide 13. Coupling of 13 with 6 completed the preparation of 14.
The third component of (+)-superstolide A (3), the phosphonium salt 21, was assembled by Brown
allylation of the aldehyde 15, to give 17. Protecting group interchange followed by ozonolysis
delivered 18, which via Still-Gennari homologation was carried on to 21. Condensation with the
fourth component, the aldehyde 22, and esterification with 14 then gave 1.

Under high dilution Suzuki conditions 1 was converted to 2. Storage in CDCl 3 for five days, or
brief warming, cyclized 2 to a single diastereomer of the transannular Diels-Alder product, that
was carried on to (+)-superstolide A (3). While acyclic trienes comparable to 2 could be induced
to cyclize, the transannular Diels-Alder reaction proceeded with much higher diastereocontrol
29. The Bergman-Ellman Synthesis of (-)-Incarvillateine
The monoterpene alkaloid (-)-incarvillateine (3) has interesting symmetry properties. The central
cyclobutane diacid core is not itself chiral, but the appended alkaloids are. The key step in the total
synthesis of 3 recently (J. Am. Chem. Soc. 2008, 130, 6316.) described by Robert G. Bergman and
Jonathan A. Ellman of the University of California, Berkeley was the diastereoselective
Rh-catalyzed cyclization of 1 to 2.
The cyclobutane diacid core 5 was assembled from ferulic acid 4 following the procedure of
Kibayashi (J. Am. Chem. Soc. 2004, 126, 16553.).

The starting point for the preparation of 1 was the commercial aldehyde 6. Enantioselective
allylation followed by silylation delivered 7, which on cross metathesis with methacrolein gave
the diene aldehyde 8. Imine formation then completed the construction of 1.
The cyclization of 1 was effected by warming (45°C, 6 h) with 2.5 mol % [RhCl(coe) 2 ] 2 and 5.5
mol % (DMAPh)Pet 2 ligand. While four diastereomers were possible from the cyclization, only
two were observed, with one predominating. Since the product mixture was easily susceptible to
tautomerization, it was carried on directly to reduction and cyclization, to form the lactam 8.
Hydrogenation of 8 to 9 required high temperature and pressure, but delivered 9 as a single
diastereomer. Reduction and desilylation then set the stage for Mitsunobu coupling with 5, to give
11. Dissolving metal conditions removed the tosyl groups from 5 to give (-)-incarvillateine 3.
It will be interesting to see how general this Rh catalyzed cyclization will be. It will also be
interesting to establish the mechanism. The authors described the cyclization of 1 as proceeding
via initial metalation of the alkene C-H bond, followed by insertion of the ester-bearing alkene
into the C-Rh bond to form a new C-Rh bond, and finally reductive elimination. Their previous
observation of metalation of such an unsaturated imine with maintenance of the alkene geometry
suppported this mechanism. The high diastereocontrol also suggested intramolecular C-C bond
formation. Whatever the mechanism, the enantiomerically-pure cyclopentane 2, having four of its
five carbons functionalized, is a versatile intermediate for further transformation.
30. The Toste Synthesis of (+)-Fawcettimine
The tetracyclic Lycopodium alkaloid fawcettimine (3) and its derivatives are of interest as
inhibitors of acetylcholine esterase. F. Dean Toste of the University of California, Berkeley
recently reported (Angew. Chem. Int. Ed. 2007, 46, 7671.) the first enantioselective synthesis of 3.
The key to the synthesis was the rapid assembly of the enantiomerically-enriched hydrindane (2).

The preparation of 2 began with the enantioselective Robinson annulation of the β-keto ester 4
with crotonaldehyde (5), mediated by the organocatalyst 6. In this protocol, originally developed
by Karl Anker Jørgensen, the single stereogenic center was established by conjugate addition,
presumably to the chiral iminium salt generated by the condensation of 5 with 6. Subsequent aldol
(or more likely Mannich) cyclization followed by elimination gave 7. Hydrolysis and
decarboxylation by heating with p-TsOH converted 7 to 1. This procedure was robust enough to
allow preparation of a ten gram batch of 1. This Jørgensen annulation is the current method of
choice for the enantioselective preparation of 2,5-dialkyl cyclohexenones.
Conjugate addition of the propargyl anion equivalent 8 to 1 proceeded with the expected > 95:5
axial diastereoselectivity, to give the silyl enol ether 9. Exposure of the derived iodide 10 to
catalytic [Ph 3 PAu]Cl and AgBF 4 induced smooth cyclization to the cis hydrindane 2.
Before constructing the nine-membered ring amine of fawcettimine (3), it was first necessary to
protect the ketone as the ketal. Pd-mediated coupling of the alkenyl iodide with the organoborane
derived from 11 then proceeded smoothly, as did the subsequent hydroboration of the terminal
alkene.
Neither the mesylate nor the tosylate derived from 12 could be induced to cyclize. In contrast,
intramolecular displacement of the iodide proceeded well, to give 13. Hydroboration followed by
oxidation then gave 15, which on deprotection cyclized to (+)-fawcettimine (3).

Several aspects of this synthesis are attractive. While the stereochemical outcome of the
hydroboration of 14 could not necessarily be predicted with confidence, in fact it did not matter, as
the stereogenic center adjacent to the ketone could be epimerized under the trifluoroacetic acid
deprotection conditions, and only the desired diastereomer would be able to add in an
intramolecular fashion to the cyclohexanone. The construction of 2 from 10 underscores the
importance of the Au-catalyzed cyclizations developed by Professor Toste.
The most important news from this synthesis is the validation in a second research group of the
enantioselective Robinson annulation previously described by Professor Jørgensen. In the
assembly of polycarbocycles, the central challenge is the enantioselective construction of the first
ring. The Jørgensen annulation is a powerful solution to that problem.
31. The Ley Synthesis of Rapamycin
Rapamycin (3) is used clinically as an immunosuppressive agent. The synthesis of 3 (Angew.
Chem. Int. Ed. 2007, 46, 591.) by Steven V. Ley of the University of Cambridge was based on the
assembly and subsequent coupling of the iododiene 1 and the stannyl alkene 2.
The lactone of 1 was prepared by Fe-mediated cyclocarbonylation of the alkenyl epoxide 5,
following the protocol developed in the Ley group.
The cyclohexane of 2 was constructed by SnCl 4 -mediated cyclization of the allyl stannane 9, again
employing a procedure developed in the Ley group. Hydroboration delivered the aldehyde 11,
which was crotylated with 12, following the H. C. Brown method. The alcohol so produced (not
illustrated) was used to direct the diastereoselectivity of epoxidation, then removed, to give 13.
Coupling with 14 then led to 2.

Combination of 1 with 2 led to 15, which was condensed with catechol to give the macrocycle 16.
Exposure of 16 to base effected Dieckmann cyclization, to deliver the ring-contracted
macrolactone 17, which was carried on to (-)-rapamycin (3).
32. The Kozmin Synthesis of Spirofungin A
Often, 6,6-spiroketals such as Spirofungin A (3) have a strong anomeric bias. Spirofungin A does
not, as the epimer favored by double anomeric stabilization suffers from destabilizing steric
interactions. In his synthesis of 3, Sergey A. Kozmin of the University of Chicago took advantage
(Angew. Chem. Int. Ed. 2007, 46, 8854.) of the normally-destablizing spatial proximity of the two
alkyl branches of 3, joining them with a siloxy linker to assure the anomeric preference of the
spiroketal. The assembly of 1 showcased the power of asymmetric crotylation, and of Professor
Kozmin’s linchpin cyclopropenone ketal cross metathesis.

To achieve the syn relative (and absolute) configuration of 6, commercial cis-2-butene was
metalated, then condensed with the Brown (+)-MeOB(Ipc) 2 auxiliary. The accompanying
Supporting Information, accessible via the online HTML version of the journal article, includes a
succinct but detailed procedure for carrying out this homologation. For the anti relative (and
absolute) configuration of 9, it is more convenient to use the tartrate 8 introduced by Roush.
Driven by the release of the ring strain inherent in 10, ring opening cross metathesis with 6
proceeded to give the 1:1 adduct 11 in near quantitative yield. The derived cross-linked silyl ether
12 underwent smooth ring-closing metathesis to the dienone 1.
On hydogenation, the now-flexible ring system could fold into the spiro ketal. With the primary
and secondary alcohols bridged by the linking silyl ether, only one anomeric form, 2, of the spiro
ketal was energetically accessible.
A remaining challenge was the stereocontrolled construction of the trisubstituted alkene. To this
end, the aldehyde 13 was homologated to the dibromide 14. Pd-mediated coupling of the alkenyl
stannane 15 with 14 was selective for the E bromide. The residual Z bromide was then coupled
with Zn(CH 3 ) 2 to give 16. These steps, and the final steps to complete the construction of
spirofungin A (3), could be carried out without exposure to equilibrating acid, so the carefully
established spiro ketal configuration was maintained.

33. The Burke Synthesis of (+)-Didemniserinolipid B
The sulfate (+)-didemniserinolipid B (3), isolated from the tunicate Didemnum sp, has an
intriguing spiroether core. A key step in the synthesis of 3 reported (Org. Lett. 2007, 9, 5357. ) by
Steven D. Burke of the University of Wisconsin was the selective ring-closing metathesis of 1 to
2.
The diol 6 that was used to prepare the ketal 1 was readily prepared from the inexpensive
D-mannitol (4). Many other applications can be envisioned for the enantiomerically-pure diol 6
and for the monoacetate and bis acetate that are precursors to it.
To set up the metathesis, the β,γ-unsaturated ketone 10 was needed. To this end, the keto
phosphonate derived from the addition of the phosphonate anion 8 to the lactone 7 was condensed
with phenyl acetaldehyde 9. The derived enone 10 was a 5:1 mixture of β,γ- and α,βregioisomers.

The diol 6 is C 2 -symmetrical, but formation of the ketal 1 dissolved the symmetry, with one
terminal vinyl group directed toward the styrene double bond, and the other directed away from it.
On exposure to the first generation Grubbs catalyst, ring formation proceeded efficiently, to give 2.
Williamson coupling with the serine-derived alcohol 11 then gave 12.
To establish the secondary alcohol of 13 and so of 3, the more electron rich alkene of 12 was
selectively epoxidized, from the more open face. Diaxial opening with hydride then gave 13.
With 13 in hand, another challenge of selectivity emerged. The plan had been to attach the
ester-bearing sidechain to 13 using alkene metathesis, then hydrogenate. As the sidechain of 3
contained an additional alkene, this had to be present in masked form. To this end, the
α-phenylselenyl ester 14 was prepared. Alkene metathesis with 13 proceeded smoothly, this time
using the second generation Grubbs catalyst. The unwanted alkene was then removed by reduction
with diimide, and the selenide was oxidized to deliver the α,β-unsaturated ester.
34. The Rychnovsky Synthesis of Leucascandrolide A

The macrolactone leucascandrolide A (4), isolated from the calcareous sponge L. caveolata, has
both cytotoxic and antifungal activity. The key step in the synthesis of 4 reported (J. Org. Chem.
2007, 72, 5784.) by Scott D. Rychnovsky of the University of California, Irvine, was the
stereoselective condensation of the aldehyde 1 with the allyl vinyl ether 2 to give 3.
The cyclic ether of 1 was assembled from the crotyl addition product 5. Tandem Ru-catalyzed
metathesis / hydrogenation converted 5 to the lactone 6. Reduction of 6 to the lactol followed by
activation as the acetate gave 7, axial-selective condensation of which with the enol ether 8
delivered the enone 9. Diastereoselective Itsuno-Corey reduction of 9 followed by protecting
group exchange and oxidation then gave 1, containing four of the eight stereogenic centers of
leucascandrolide A (4).
The vinyl ether 2 was readily prepared from the corresponding homoallylic alcohol. Condensation
of 1 with 2 involved Lewis acid activation of the aldehyde, addition of the resulting carbocation to
the vinyl ether, and cyclization with trapping by bromide ion. In this process, the other four of the
eight stereogenic centers were assembled. Three of those centers were formed in the course of the
reaction. While stereocontrol was not perfect, the route is pleasingly succinct, so practical
quantities of diastereomerically pure 3 could be prepared.
To complete the synthesis, the secondary alcohol of 3 was methylated. Selective desilyation of the
primary alcohol followed by oxidation and desilylation then set the stage for the Mitsunobu
macrolactonization. The intermediates in the Mitsunobu reaction are such that the lactonization
can proceed with either inversion of absolute configuration at the secondary center, or retention.
While the usually-employed Ph 3 P gave the lactone with retention of absolute configuration, Bu 3 P
led to clean inversion.

The last challenge was the establishment of the (Z) alkene of the side chain. This was
accomplished using the Toru protocol. Coupling of the secondary bromide with the Cs salt 12
proceeded with inversion of absolute configuration, to give 13. The carboxylates of stronger acids
were not sufficiently nucleophilic to displace the bromide. Aldol condensation of 13 with the
aldehyde 14 gave a mixture of diastereomers, exposure of which to MsCl in pyridine delivered the
requisite (Z) alkene 4.
35. The Smith Synthesis of (+)-Lyconadin A
The pentacyclic alkaloid (+)-lyconadin A (3), isolated from the club moss Lycopodium
complanatum, showed modest in vitro cytotoxicity. A key step in the first reported (J. Am. Chem.
Soc. 2007, 129, 4148.) total synthesis of 3, by Amos B. Smith III of the University of
Pennsylvania, was the cyclization of 1 to 2.
The pentacyclic skeleton of 3 was constructed around a central organizing piperidine ring 9. This
was prepared from the known (and commercial) enantiomerically-pure lactone 4. The akylated
stereogenic center of 9 was assembled by diastereoselective hydroxy methylation of the acyl
oxazolidinone 5 with s-trioxane, followed by protection. Reduction of the imide to the alcohol led
to the mesylate 7, which on reduction of the azide spontaneously cyclized to give, after protection,
the piperidine 8. Selective desilylation of the primary alcohol then enabled the preparation of 9.

The plan was to assemble the first carbocyclic ring of 3 by intramolecular aldol condensation of
the keto aldehyde 15. The enantiomerically-pure secondary methyl substituent of 15 derived from
the commercial monoester 10. Activation as the acid fluoride followed by selective reduction led
to the volatile lactone 11. Opening of the lactone with H 3 CONHCH . 3 HCl gave, after protection,
the Weinreb amide 12. Alkylation of the derived hydrazone 13, selectively on the methyl group,
led, after deprotection, to 15. The intramolecular aldol condensation of 15 did deliver the unstable
cyclohexenone 1. Under the acidic conditions of the aldol condensation, the enol derived from the
piperidone added in a Michael sense, from the axial direction on the newly-formed ring, to give
the trans-fused bicyclic diketone 2.
To move forward, it was necessary to epimerize 2 to the cis ring fusion, and also to differentiate
the two ketones of 2. These two problems were solved simultaneously by deprotection and
epimerization to the cis-fused hemiaminal 16.
Attempts to deoxygenate the tertiary alcohol of 16 failed, so instead, selective reduction followed
by protection delivered 17. Reduction to the axial alcohol followed by dehydration with the
Martin sulfurane then installed the trisubstituted alkene of 18. The C-N bond was re-established
by exposure of 18 to NIS, leading to the crystalline 19. Activation of the derived ketone with
Mander’s reagent followed by reductive deiodination (Et 3 SiH/Pd) gave 20, Michael addition of
which with 21 led to 3.
36. The Betzer and Ardisson Synthesis of (+)-Discodermolide

(+)-Discodermolide (3), a potent anticancer agent that works synergistically with taxol, may yet
prove to be clinically effective. For the synthetic material to be affordable, a highly convergent
synthesis is required. Jean-François Betzer and Janick Ardisson of the Université de
Cergy-Pontoise have described (Angew. Chem. Int. Ed. 2007, 46, 1917.) such a synthesis,
coupling 1 and 2. A central feature of their approach was the repeated application of the inherently
chiral secondary organometallic reagent 5.
The first use of 5 was the addition to the aldehyde 4. The product 6 was ozonized, and the
resulting aldehyde was carried on to the α,β-unsaturated ester. Exposure of the hydroxy ester to
benzaldehyde under basic conditions delivered, by intramolecular Michael addition, the acetal 7.
The next addition of the reagent 5 was to the aldehyde 10. The adduct 11 was deprotonated with
t-BuLi to effect α-elimination, providing, after protection of the alcohol, the alkyne 12. Coupling
of 12 with the amide 7 gave a ketone, enantioselective reduction of which under Itsuno-Corey
conditions led, again after protection of the alcohol, to the alkyne 13. Oxidation followed by
selective hydrogenation and iodine-tin exchange then completed the assembly of 1. Note that PtO 2 ,
not typically used for partial hydrogenation, was the catalyst of choice for this congested alkyne.
The third application of the enantiomerically-pure reagent 5 was addition to the aldehyde that had
been prepared by ozonolysis of 15. Advantage was then taken of another property of the alkenyl
carbamate, Ni-mediated Grignard coupling, to form the next carbon-carbon bond with high
geometric control. Deprotection of the diene 17 so prepared followed by iodination then
completed the synthesis of 2.

The convergent coupling of 1 with 2 was carried out under Suzuki conditions. Reduction of the
iodide of 2 to the corresponding alkyl lithium followed by exchange with B-OMe-9-BBN gave
gave an intermediate organoborane, that smoothly coupled with 1 under Pd catalysis to give 18.
Deprotection and carbamate formation then led to (+)-discodermolide (3).
This synthesis clearly illustrates the power of 5 as an enantiomerically-defined secondary
organometallic reagent, and the synthetic versatility of the product alkenyl carbamates. The ready
availability of the three enantiomerically-pure four-carbon fragments 4, 8, and 14 was also a key
consideration in the design of this synthesis.
37. The Maier Synthesis of Cruentaren A
Cruentaren A (3), isolated from the myxobacterium Byssovorax cruenta, is an inhibitor of
mitochondrial F-ATPase. The synthesis of 3 (Org. Lett. 2007, 9, 655,; Angew. Chem. Int. Ed. 2007,
46, 5209.) by Martin E. Maier of the Universität Tübingen illustrates the power of alkyne
metathesis as a tool for the synthesis of complex natural products. Very recently, Alois Fürstner of
the Max-Planck-Institut, Mülheim, reported (Angew. Chem. Int. Ed. 2007, 46, 9275.) an
alternative synthesis, also based on alkyne metathesis, of cruentaren A (3).
The alcohol portion of 1 was prepared by Marshall homologation of 4 with 5, leading to 6.
Homologation of the derived epoxide 7 then gave 8. Note that the homologation of 7 to 8 required
three steps. This might have been accomplished more directly with the Li salt of 1-propyne, easily
prepared from commercial 1- or 2-bromopropene.

Evans auxiliary-controlled homologation of 9 set the relative and absolute configuration of 10,
which was carried on to 12. To effect coupling, the acid of 12 was activated with carbonyl
diimidazole, then condensed with the bis-alcoholate of 8. This acylation was highly regioselective,
giving 1 as the only observed product.
Cruentaren A (3) has two Z alkenes, so the authors chose a bis-alkyne strategy, with a partial
hydrogenation of both alkynes at the end of the synthesis. To this end, alkyne metathesis was
accomplished with the Schrock tungsten carbine catalyst 13. Homologation to 15 followed by
deprotection and hydrogenation then gave enantiomerically pure cruentaren A (3).
38. The Sammakia Synthesis of the Macrolide RK-397

The polyene macrolide RK-397 (3), isolated from soil bacteria, has antifungal, antibacterial and
anti-tumor activity. Tarek Sammakia of the University of Colorado has described (Angew. Chem.
Int. Ed. 2007, 46, 1066.) the highly convergent coupling of 1 with 2, leading to 3.
The preparation of 1 depended on the powerful methods that have been developed for acyclic
stereocontrol. Beginning with the allylic alcohol 4, Sharpless asymmetric epoxidation established
the absolute configuration of 5. Following the Jung “non-aldol aldol” protocol, exposure of 5 to
TMSOTf delivered the aldehyde 6 in high de. Condensation of 6 with the lithium enolate of
acetone also proceeded with high de. The resulting alcohol was protected as the MOM ether, to
direct the stereoselectivity of the subsequent aldol condensation with 8. Selective β-elimination
followed by reduction and protecting group exchange then gave 1.
The preparation of 2 took advantage of the power of Brown asymmetric allylation. Allylation of
the symmetrical 11 led to the diol 12. This was desymmetrized by selective acetonide formation,
to give 13. Ozonolysis, reductive work-up, and protection of the newly-formed 1,3-diol gave 14,
setting the stage for oxidation and asymmetric allylation to give 15. Reductive deprotection and
oxidation then delivered the acetonide 2.
The tris acetonide 16 was assembled by addition of the enolate derived from 1 to the aldehyde 2,
followed by reduction and protection. Kinetically-controlled metathesis with 17 established the
triene 18. Phosphonate-mediated homologation to the pentaene 19 followed by hydrolysis and
Yamaguchi macrolactonization then completed the synthesis of the macrolide RK-397 (3).

39. The Clark Synthesis of Vigulariol
Vigulariol (3), isolated from the octocoral Vigularia juncea (Pallas), is active against human lung
adenocarcinoma cells (IC 50 = 18 nM). A key step in the synthesis of 3 reported (Angew. Chem. Int.
Ed. 2007, 46, 437.) by J. Stephen Clark of the University of Glasgow was the dipolar
rearrangement of 1 to 2.
The plan for the diastereoselective construction of 1 was based on the known reductive cyclization
of aldehydes such as 7. To this end, the alcohol 6 was prepared and condensed with ethyl
propiolate. As anticipated, the SmI 2 -mediated cyclization of 7 proceeded to give 8 with high
diasterocontrol. The diazo ketone 1 was prepared by reaction of the mixed anhydride (isobutyl
chloroformate) of the acid with diazomethane.
It was anticipated that Cu-catalyzed carbene formation would lead the 1,3-dipole 9, that would
then undergo sigmatropic rearrangement to 2. In the event, the rearrangement was remarkably
efficient, delivering the ten-membered ring carbocycles 10 and the desired 2 in a 1 : 5 ratio.
Exposure of the less stable geometric isomer 10 to catalytic thiyl radical effected equilibration to
2.

Although one might think of medium rings such as that of 2 as being floppy, in fact 2 has a single
preferred conformation, and diastereocontrol for the rest of the synthesis took advantage of that
preferred conformation. The diene 11 derived from 2 has one open face. Addition of methyl vinyl
ketone to that outside face gave a 1 : 2 mixture of the diastereomeric endo and exo cycloadducts.
The mixture was equilibrated to the more stable exo adduct 12. A four-step sequence then
converted 12 into 13.
The completion of the synthesis again depended on the inside-outside bias of the medium ring.
Addition of methyl magnesium chloride to the outside face of the ketone derived from 13
delivered 14 with high diastereocontrol. Epoxidation of 14, again from the outside face, led
directly to vigulariol 3. The intermediate in the MCPBA reaction is the protonated epoxide, and it
may be that the tertiary alcohol poised directly on the opposite face of the alkene opened that
activated intermediate directly.
This synthesis of vigulariol (3) is remarkably concise, with four rings and eight stereogenic
centers being assembled in just twenty steps. All of the stereogenic centers are derived from the
secondary alcohol 6. Use of the known enantiomerically-pure 6 would have delivered vigulariol (3)
in enantiomerically-pure form. This synthesis design once again illustrates the power of
conformational analysis of medium rings.
40. The Trost Synthesis of (-)-Terpestacin
(-)-Terpestacin (3), isolated from Arthrinium sp. FA1744, inhibits the formation of syncytia by
HIV-infected T cells. A key step in the total synthesis of 3 reported (J. Am. Chem. Soc. 2007, 129,
4540. ) by Barry M. Trost of Stanford University was the Ru-catalyzed cyclization of 1 to 2. This
synthesis of (-)-terpestacin (3) elegantly illustrates the power of the Trost enantioselective Pd
catalysts.

The preparation of 1 began with the commercially-available diketone 4. With the enol as the
nucleophile, opening of racemic isoprene monoepoxide 5 using the Trost chiral Pd catalyst led to
the ether 6 in high ee. It is impressive that even though it was directed by a quaternary stereogenic
center, the subsequent Claisen rearrangement to 7 proceeded with complete facial selectivity.
Oxidation to the nonenolizable α-diketone 8 then set the stage for the conjugate addition. Again,
even though the directing stereogenic center was quaternary, the addition proceeded with
substantial diastereocontrol. Further study of the elements that govern the facial selectivity of the
Lewis acid-mediated (Sakurai) addition of allyl silanes to enones would certainly be warranted.
The sulfone 11 was prepared in four steps from commericially-available geranyl bromide, with the
absolute configuration being set by Sharpless asymmetric epoxidation. Subsequent to the
alkylation of 11 with the bromide 10, the now-surplus sulfone was removed reductively with
Pd(OAc) 2 and NaBH 4 . It is noteworthy that the enone of 1 was stable to those conditions.
The cyclization of 1 with the second-generation Grubbs catalyst delivered the desired E alkene 2
in 43% yield. As had been observed before by others, the free allylic alcohol was the preferred
substrate for the Ru metathesis catalyst. No cyclization was observed with tetraenes that had the
alcohol protected.
With 2 in hand, it remained to install the side chain. The relative configuration of the pentenyl side
chain of 14 was set by again using chiral Pd catalysis. The Claisen rearrangement proceeded
smoothly, to give, after protection, the ether 15.

The oxidative cleavage of the side chain of 15 initially presented some difficulties. Eventually, it
was found that AD-Mix-α (but not β) could effect selective dihydroxylation. Periodate cleavage
followed by reduction then gave 16, which was deprotected to (-)-terpestacin (3).
Note that both 5 and 13 were racemic. Except for the secondary alcohol on the fifteen-membered
ring, the absolute configuration of every stereogenic center in the (-)-terpestacin (3) prepared by
this synthesis derived from the absolute configuration of the enantiomerically-pure Pd complexes
used as first 5 and then 13 were incorporated selectively into the natural product.
41. The Nakada Synthesis of (+)-Digitoxigenin
The total synthesis of steroids such as (+)-digitoxigenin (3) has been studied for more than sixty
years, yet it has never been thought that such studies would lead to a preparative route that would
be competitive with partial synthesis from the abundant plant sterols. The enantioselective
synthesis of 3 recently (Tetrahedron Lett. 2007, 48, 1541.) described by Masahisa Nakada of
Waseda University suggests that the preparative synthesis of even such complex polycarbocycles
may in fact be practical. A key step in the synthesis of 3 was the conjugate addition of the cuprate
derived from 1 to the enone 2.
For this convergent approach to be effective, both 1 and 2 had to first be prepared in
enantiomerically-pure form. The synthesis of 1 started with the bakers’ yeast reduction of the
prochiral diketone 4 to give 5 in high ee. The key problem was the establishment of the ternary
center of 1. This was accomplished by coupling the iodide 8 with the boronic acid 9.
Hydroxyl-directed hydrogenation of 10 then led to the desired 1.

The preparation of 2 began with Cu-catalyzed cyclization of the prochiral diene 12 to give the
crystalline 13. Recrystallization raised the ee of 13 to > 99%. Reductive opening of 13 gave 14,
which was reduced and protected to give 15. Oxidation of 15 with the Ochiai reagent 16 provided
the enone 2. Conjugate addition of 1 proceeded across the more open face of the enone 2, to give
17.
The key step to close the B ring of the steroid was the intramolecular aldol condensation of 18 to
give 19. The β-hydroxy ketone so produced was too sensitive to reduce by direct methods, so the
monoxanthate was prepared from the derived diol, and then reduced under free radical conditions,
leading to the diol 20. Singlet oxygenation of 20 then completed the synthesis of 3.
Clearly, the strength of this synthesis is the elegant construction of the AB enone 2. Even more
important is the demonstration of a general convergent route to the steroids, by addition of an
enantiomerically-pure D ring moiety to an AB fragment. The power of asymmetric synthesis
makes the precursors 1 and 2 readily available in the necessary enantiomerically-pure form.

42. The Pettus Synthesis of (+)-Rishirilide B
(+)-Rishirilide (4), an inhibitor of α 2 -macroglobulin, a tetrameric serum glycoprotein that is an
irreversible protease inhibitor, has a deceptively simple structure. Thomas R. R. Pettus of the
University of California at Santa Barbara has reported (J. Am. Chem. Soc. 2006, 128, 15625.) a
concise route to 4, based on the highly regioselective Diels-Alder addition of the
enantiomerically-pure dienophile 2 to the diene derived from 1.
The preparation of 2 was carried out from the aldehyde 5. Selective protection of the less hindered
phenol followed by reduction gave the benzyl alcohol 6. On exposure to excess Grignard reagent 7,
the primary alcohol of 6 apparently underwent elimination to give the o-quinone methide 8,
conjugate addition to which gave 9. Mitsunobu coupling of 9 with the lactic acid amide 10
proceeded with clean inversion, to give 11. The methoxy methyl amide of 11 was required for the
oxidative dearomatization to 2 to proceed efficiently.
The preparation of the masked diene 1 started with the aldehyde 12. Following the Comins
procedure, metalation of the derived hemiaminal alkoxide and subsequent addition of methyl
iodide gave 13. Addition of SO 2 to the photoenol derived from 13 followed by etherification of the
labile secondary alcohol so formed then gave 1.
The dienophile 2 should undergo cycloaddition with high facial selectivity. That is not relevant in
this application, as addition to the intermediate quinone methide 14 was followed by elimination
and oxidative aromatization. The regioselectivity of the addition, delivering 2, was, however,
critical for the success of the synthesis.

With 3 in hand, what remained was a one-carbon homologation. This was accomplished by
selective O-acylation with dimethyl carbamoyl chloride 16. Subsequent addition of an excess of
anion 17 led, via alkoxide-directed addition to the remaining ketone carbonyl followed by
deacylation, to the desired adduct 18. Careful cleavage, to avoid lactone formation, then delivered
4. Note that this synthesis of (+)-rishirilide B (4) would be classified as enantioselective, since the
intial stereogenic center of 2, that set the absolute configuration of the dearomatized ring of 2, was
not included in the final product.
43. The Gin Synthesis of Nominine
Retrosynthetic analysis of the heptacyclic ring structure of nominine (3), a representative member
of the hetisine family of alkaloids, is not for the faint of heart. David Y. Gin, now at the
Sloan-Kettering Institute, has presented (J. Am. Chem. Soc. 2006, 128, 8734.) an elegant solution
to this problem. One key step in the synthesis was the intramolecular Diels-Alder cyclization of 1
to 2.
The convergent construction of 1 by dipolar cycloaddition presented some interesting challenges.
One half was prepared from 3-methylcyclohexenone (4). Addition of Et 2 AlCN followed by
triflation gave 5 in regiocontrolled fashion. Reduction followed by Pd-catalyzed coupling of the
enol triflate then led to the aldehyde 6. An intriguing question is how one would prepare 6 in
enantiomerically pure form.

The other half of 1 was prepared from the acetal 7. Lateral metalation followed by acylation with
the Weinreb amide 8 gave the ketone 9. Displacement with azide followed by exposure to acid
gave the bridged acetal 10, which was condensed reductively with 6 to give the oxidopyridinium
betaine 12.
Two products, 13 and 14, could arise from the dipolar cycloaddition of 12. It was not clear at the
outset which would be preferred. In the event, it did not matter. Even though the thermal
equilbrium favored the undesired 13, equilibration was efficient, and the two were easily
separated.
With 14 in hand, the stage was set for the intramolecular Diels-Alder cyclization. In fact, the diene
1 was not isolated. The evidence for the intermediacy of 1 was the observation that on exposure to
pyrollidine in methanol at 60°C, the Birch reduction product 17 smoothly cyclized to 2.
Methylenation followed by kinetic, axial-selective SeO 2 oxidation then completed the synthesis of
3.
The concise elegance of this 17-step synthesis of the heptacyclic hetesine alkaloid nominine 3 is
all the more apparent when it is compared with the only previously-reported synthesis, published
just two years earlier, which took 40 steps

44. The Padwa Synthesis of Asphidophytidine
Albert Padwa of Emory University has developed (Org. Lett. 2006, 8, 3275.) a productive
approach to fused indole alkaloids such as asphidophytine (3), based on the dipolar cycloaddition
of the ylide derived by exposure of a diazo ketones such as 1 to a Rh(II) carboxylate catalyst.
The preparation of 1 started with the aniline 4. Ortho iodination followed by N-alkylation with 5
delivered the unsaturated ester 6. Heck cyclization no doubt intially left the alkene still conjugated
with the ester, but traces of acid or base would be expected to easily isomerize this to establish the
aromaticity of the five-membered indole ring. N-methylation followed by saponification then gave
8.
The preparation of 1 continued with the alkylation of 9. Hydrolysis of 10 followed by
homologation and subsequent diazo transfer gave 11, which was coupled with 8 to give 1. The
quaternary center established in the alkylation of 9 was carried through the synthesis, so if
enantiomerically-pure material were desired, an enantioselective route to 10 would have to be
devised.
The push-pull dipole 12 was constructed by exposure of 1 to Rh 2 (OAc) 4 . Loss of N 2 gave the Rh
carbene, which complexed with the nucleophilic amide carbonyl. The dipole 12 was not isolated,
but reacted in situ with the tethered indole to give the hexacyclic adduct 2. Note that two
diastereomers of 2 could have been formed, but only 2, with the bulky t-butyl ester exo, was
observed.

The dipolar cycloadddition established the requisite stereochemical relationship between the three
contiguous quaternary stereogenic centers of 1. It remained to adjust the functional groups around
the newly-formed carbocyclic ring. Exposure to BF 3 .OEt 2 led to ring opening, followed by
trapping of the intermediate carbocation with the t-butyl ester to give the lactone 13, with
concommitant loss of isobutylene. Hydrolysis and decarboxylation gave the alcohol 14, the acetate
of which was removed by reduction with SmI 2 . The derived enol triflate 15 was reduced to the
alkene, which was deoxygenated by way of the thiolactam 16.
The intramolecular dipolar cycloaddition exemplified by the conversion of 12 to 2 is a specific
representative of a general and powerful approach to indole alkaloids, based on cycloaddition of
an intermediate indole to a dipole or a diene. For more recent work along these lines by Professor
Padwa, see Org. Lett. 2007, 9, 279, and Tetrahedron 2007, 63, 5962,
45. The Overman Synthesis of (-)-Sarain A
Larry E. Overman of the University of California, Irvine has recently (Angew. Chem. Int. Ed. 2006,
45, 2912. ) described the first total synthesis of the antibiotic and antitumor alkaloid (-)-sarain A
(3). The structure of 3 is particularly challenging, with two tertiary amines, one of them attached
to a quaternary center. Note that although sarain A is drawn as the amino aldehyde, in fact the
aldehyde and the proximal amine interact to form zwitterionic species that are difficult to purify.
A key transformation in the assembly of 3 was the intramolecular Mannich cyclization of 1 to 2.
This C-C bond forming reaction established the single carbocyclic ring of 3, while at the same
time constructing with high diastereocontrol the tetraalkylated quaternary center.
The enantiospecific preparation of 1 began with diethyl tartrate 4. The enolate of the derived
oxazoline 5 added with high diastereocontrol to the Z-α,β-unsaturated ester 6 to give 7. The diester
7 included the three continguous stereogenic centers of 8. The details of the conversion of 7 to the
hemiaminal 8 had previously been reported by Professor Overman (J. Org. Chem. 1998, 63,
8096, ; Tetrahedron Lett. 1999, 40, 1273, ; Org. Lett. 2005, 7, 933, ).

The aldehyde 1 was in fact not isolated. Rather, the intramolecular Mannich condensation was
developed using the silyl enol ether 8 as the starting material. It is not clear whether the striking
diastereocontrol observed is steric in origin, with the small aldehyde tucked into the more
incumbered space, or if there is an attractive interaction between the aldehyde carbonyl and the
polar carbamate. In principle, the Mannich condensation is reversible, so the reaction may be
under thermodynamic, not kinetic, control.
Desulfonylation of 2 followed by reductive alkylation delivered 9, setting the stage for
Ru-mediated (G1) closure of the first macrocyclic ring. Note that from 9 on, a basic tertiary amine
was carried through the remainder of the synthesis, and from 11 on, each intermediate had two
basic amines.
To close the final ring, the amino diol 11 was protected as the hemiaminal, then oxidized to 13.
Diastereoselective (~3-4 : 1) Grignard addition followed by homologation then gave the precursor
15 for Stille coupling to close the last ring. Reduction of the hemiaminal, oxidation with
bicarbonate-buffered Dess-Martin reagent to the aldehyde (zwitterionic!) and a delicate
deprotection then completed the synthesis of (-)-sarain A (3).
46. The Fukuyama Synthesis of Morphine

Tohru Fukuyama of the University of Tokyo has recently reported (Org. Lett. 2006, 8, 5311. ) an
elegantly conceived total synthesis of (±)-morphine (3), based on the recognition and reduction to
practice of the highly diastereoselective intramolecular Mannich cyclization of 1 to 2.
The stereocontrolled preparation of 1, a considerable challenge, started with the epoxidation of the
diene 4 to give 6. Although racemic 6 was used for this synthesis of morphine, Professor
Fukuyama had previously demonstrated that high ee 6 could be prepared by lipase resolution of
the intermediate bromohydrin 5. Coupling of 6 with 7 led to 8, by net syn S N 2' displacement. The
secondary alcohol was then inverted, to put the α-H syn to what would be the adjacent C-Pd bond,
to undergo β-hydride elimination in the course of the intramolecular Heck cyclization of 10 to 2.
The intramolecular Mannich cyclization of 1 to 2 takes advantage of the native functionality of
morphine 3. The intermediate in the cyclization is apparently the expected 11, which could be
isolated. Note that under the equilibrating conditions of the cyclization, the desired cis ring fused 2
was the only diastereomer observed.
The Ito-Saegusa procedure was used to oxidize the ketone 2 to the enone 12. It was then necessary
to convert the enone into the allylically-inverted alcohol 14. With the morphine skeleton
assembled, the aromatic ring blocks the bottom face of the enone, so both hydrogen peroxide and
hydride attacked from the top face, to give 13. Conversion of the derived thiocarbamate to the free
radical then led to epoxide cleavage, delivering the desired 14. The last three steps to complete the
total synthesis of (±)-morphine (3), oxidation, addition of hydride to the open top face of the
ketone with concomitant reduction of the urethane protecting group to the N-methyl, and
O-demethylation, then followed literature procedures.

47. The Nicolaou Synthesis of Platensimycin
The report last year by scientists at Merck of an antibiotic, platensimycin 3, with a novel
mechanism of action has led to much effort toward the total synthesis of this degraded diterpene.
K. C. Nicolaou of Scripps, La Jolla has now (Angew. Chem. Int. Ed. 2006, 45, 7086.) reported the
first preparation of 3. The key step in the synthesis is the elegantly concise cyclization of 1 to 2.
The preparation of 1 started with iterative Stork-Danheiser alkylation of 4 to give 5. Reduction
followed by hydrolysis unraveled the enol ether to give the enone, which was re-silylated to give 6.
Ru-catalyzed intramolecular ene cyclization of 6 gave the enol ether 7, which was selectively
oxidized to 1. Reductive cyclization of 1 gave 2 as a 2:1 mixture with its diastereomer 8.
The aniline of platensimycin was prepared from 2-nitroresorcinol (9). MOM ether formation
followed by reduction gave 10. Metalation of the protected amine followed by acylation with
Mander’s reagent and thermolysis gave 11.

The cyclization of 2 to 11 proceeded smoothly, as did the alkylation of 11 to 12. The
diastereoselectivity of the alkylation followed from the conformational bias of the ring system.
The conversion of the terminal alkene to the acid 13 initially was troublesome, but a solution was
found in metathesis with vinyl boronate followed by oxidation. Coupling with 11 followed by
deprotection then gave 3.
The absolute configuration of the tricyclic core of 3 was set in the intramolecular ene cyclization
of 6 to 7. There is the possibility that with an enantiomerically-pure catalyst, the product 7 and so
3 could be prepared in high enantiomeric excess.
Total syntheses of platensimycin 3 are underway in several other research groups. The diversity of
the approaches being explored will enrich organic synthesis.
48. The Sorensen Synthesis of (-)-Guanacastepene E
The guanacastepenes, of which guanacastepene E (3) is representative, initially elicited excitement
because of their activity against drug resistant bacterial strains. Although the systemic toxicity of
the guanacastepenes has cooled that enthusiasm, the guanacastepenes remain as architectural
challenges. In particular, a convergent synthesis plan requires the development of a strategy for
constructing the central 7-membered ring with control of relative and absolute configuration,
particularly of the two angularly methylated quaternary centers. Erik J. Sorensen of Princeton
University (J. Am. Chem. Soc. 2006, 128, 7025.) solved this problem by the intramolecular
photochemical dimerization of 1, which delivered 2 with high diastereocontrol. Directed ring
fragmentation then led to 3.

The enantiomercially-pure cyclopentenone of 1 was prepared from the chiral pool starting material
dihydrocarvone (4). Methylation followed by ozonolytic cleavage gave the aldehyde acid 5, which
was converted into the cyanohydrin lactone 6. Intramolecular acylation of the derived nitrile anion
proceeded smoothly, to give, after fragmentation, the 1,2-diketone 7. Stannylation of the nonaflate
led to 8, which was coupled with enantiomerically-pure 9 to give 1.
Construction of the cyclohexene 9 began with the Diels-Alder cycloaddition of the alkyne 11 to
the diene 10, followed by Baeyer-Villiger cleavage to give 13. This established the quaternary
center, and at the same time set the relative configuration of the secondary alcohol of 9, and thus
of 3. The alcohol 15 so prepared was racemic. Screening of esters led to the mandelate acetate 9,
the diastereomers of which could be separated by open column chromatography. The mandelate
ester also proved to be an effective leaving group for the Pd-mediated coupling of 8 and 9 to give
1.
As anticipated, the intramolecular 2+2 photocyclization was directed by the adjacent isopropyl
group, to give 2 with high diastereocontrol. While opening of carbonyl-activated fused
cyclopropanes is amply precedented, far less was known about the opening of carbonyl-activated
fused cyclobutanes. Happily, double one-electron reduction with excess SmI 2 opened the desired
bond, to give, after selenation and oxidation, the dienenone 16. The acetoxy group adjacent to the
ketone was introduced by Rubottom oxidation. On deprotection of 17, the released primary
alcohol spontaneously added to the enone, to give (-)-guanacastepene E (3) .

49. The Crimmins Synthesis of (+)-SCH 351448
The symmetrical macrodiolide (+)-SCH 351448 (4) is the only known selective activator of
transcription from the low density lipoprotein receptor. The highly convergent synthesis of 4 from
1 and 2 reported (Org. Lett. 2006, 8, 2887. ) by Michael T. Crimmins of the University of North
Carolina illustrates the power of the chiral glycidyl anion approach for the preparation of α,α'-bis
chiral ethers.
The upper half of 1 was assembled by Sharpless resolution of the alcohol 5. Allylation of the
enolate derived from 6 then delivered, after reduction and methylenation, the triene 7. The lower
half of 1 was prepared by condensation of the anion derived from 8 with acrolein 9. Methylation
of product 10 led to 11, which was again allylated, to give, after homologation, the aldehyde 12.
Condensation with a chiral acetate equivalent gave 13. The corresponding aldehyde 14 was then
added to the enolate of 15, derived from 7, to arrive at 1.

Although the tandem ring closing metathesis - homologation of 1 with 2 proceeded spectacularly
well, to give 3 in 88% yield, simple dimerization of 3 did not deliver the desired 4. As an
alternative, 15 was cyclized with G1, then acylated with reduced 3 to give 16. Intramolecular
acylation of 17 then worked well, leading to 4.
50. Synthesis of (-)-Colombiasin A and (-)-Elisapterosin B
Historically, there have been three methods for assembling enantiomerically-pure carbocycles:
cyclization of an enantiomerically-pure starting material, enantioselective cyclization of a
prochiral substrate, and enantioselective transformation of a prochiral cyclic starting material.
Recently, a fourth strategy has been developed, enantioselective transformation of a racemic
starting material, with only one enantiomer of the starting material going on to the desired product.
This approach is illustrated by the transformation of racemic 1 to enantiomerically (and
diastereomerically) pure 3. The conversion of 1 to 3 is the key step in a general route established
(J. Am. Chem. Soc. 2006, 128, 2485.) by Huw M. L. Davies of the University at Buffalo to the
diterpenes isolated from the gorgonian coral Pseudoterogoria elisabethae, exemplified by
(-)-colombiasin A (4).

The dihydronaphthalene 1 was prepared by Diels-Alder cycloaddition of the diene 5 to the
benzoquinone 6, the preparation of which had been reported by Nicolaou in the course of a
previous synthesis of (-)-colombiasin A (4). Silylation of the cycloadduct followed by selective
hydrolysis gave the ketone 7. The ketone was converted to the enol triflate 8, which was reduced
with triethylsilane to give racemic 1.
Exposure of 1 to 2 in the presence of the DOSP Rh catalyst, designed by Davies, delivered the two
adducts 8 and 9. These were carried together to the alcohols 10 and 11, which were separated. The
overall yield of diasteromerically- and enantiomerically-pure 11 from racemic 1 was 34%. This
was 68% of theoretical, since was derived from only one of the two enantiomers of 1.
Homologation of 11 to the diene followed by desilylation under oxidative conditions led to the
quinone 12. Followed the precedent of Rychnovsky, on heating 12 was converted to the
Diels-Alder adduct 13. Demethylation then gave (-)-colombiasin A (4).
The tetracyclic triketone (-)-elisapterosin B (14), also a Pseudoterogoria elisabethae diterpene, is
the intramolecular [5 + 2] cycloaddition product from 12. Again following a Rychnovsky

procedure, exposure of the diene 12 to BF 3 etherate led directly to 14, accompanied by minor
amounts of the methyl ether 13.
51. The Ready Synthesis of (-)-Nigellamine A 2
The nigellamine alkaloids, represented by (-)-nigellamine (3), dolabellane diterpenes isolated from
black cumin, apparently have lipid metabolism promoting activity. Joseph M. Ready of UT
Southwestern Medical Center has described (J. Am. Chem. Soc. 2006, 128, 7428.) the first
synthesis of a nigellamine, (-)-nigellamine A 2 (3). Key steps in the synthesis include an
enantioselective cyclization to prepare 1, and the Cr-mediated cyclization of the aldehyde 2.
(-)-Nigellamine A 2 (3) has an angularly substituted trans ring fusion. The key to the synthesis was
the preparation of the angularly-substituted cis-fused lactone 1. The absolute configuration of 1
was set by the Pd-mediated S N 2' cyclization of the malonate 6, which proceeded in 95% ee.
Equally important was the seemingly mundane iodine-mediated cyclization of 7 to 8. This one
transformation differentiated the two esters of 6, secured the relative configuration of one of the
two secondary alcohols of 3, and established the requisite cis ring fusion of the lactone 1.
Nucleophilic displacement of the iodide 8 failed, but two-carbon homologation to the alkyne could
be accomplished by the free radical Fuchs procedure.
There were intial difficulties with the Negishi methylation/iodination of the terminal alkyne, as the
Cp 2 ZrCl 2 reacted preferentially with the lactol. Remarkably, repeating the reaction in the presence

of water led to clean addition to the alkyne. Homologation followed by, again, wet Negishi
methylation/iodination set the stage for oxidation to the lactone aldehyde 2.
Both the bicyclic skeleton of 2 and the geometry of the two alkenes direct the pendant chain
toward the aldehyde. In fact, Ni-catalyzed cyclization proceeded smoothly. The product alcohol
emerged as a single diastereomer, but unfortunately the wrong one, so an oxidation/reduction
cycle was required to correct the secondary alcohol center.
The final challenge was the selective epoxidation of the triene 12. There are two concerns: facial
selectivity, and chemoselectivity. The facial selectivity is inherent, as the geometry of the medium
ring is such that only the desired face is exposed. Chemoselectivity was more challenging.
MCPBA reacted indiscriminantly with each of the three alkenes. Reasoning that a bulkier
epoxidizing agent might be more selective, they found that the Shi dioxirane (13) delivered a 7:1
ratio of the two trisubstituted epoxides. It is interesting that the enantiomer of 13 gave only a 2:1
ratio.
52. The Leighton Synthesis of Dolabelide D
The macrolides dolabelides A-D, isolated from the sea hare Dolabella, are cytotoxic against
HeLa-S3 cells at concentrations of 1.3 - 6.3 μg/mL. The recent (J. Am. Chem. Soc. 2006, 128,
2796. ) synthesis of dolabelide D (3) by James L. Leighton of Columbia University nicely
highlights the powerful reagent-based methods for acyclic stereoselection that they have recently
developed.

Two such reagents, easily prepared on a large scale, are the amino silanes 4 and 5. These were
used to prepare 8 and 12, which were combined to prepare 1. Homologation of the aldehyde 6
with 4 gave 7. Protection followed by Wacker oxidation then delivered 8. To prepare 12,
methacrolein 9 was homologated with ent-5. Hydroformylation followed by in situ acetal
formation gave 11. Diastereroselective hydroboration followed by oxidation and esterification
then led to 12. Aldol condensation of 12 with 8, taking advantage of the inherent chirality of the
alkoxy enolate, proceeded to give, after protection, a 10:1 ratio of 13 and its diastereomer.
The preparation of alkene 2 depended on a third reagent the Leighton group has developed, the
silane 15. Enantioselective condensation of 14 with 15 set the absolute configuration around Si. In
the next step, intramolecular carbonylative silylation followed by intramolecular crotylation of the
transient aldeyde, the absolute configuration at the Si in 16 directed the new stereogenic centers of
17. To achieve the requisite diastereoselectivity in the aldol condensation of the derived ketone 18
with the aldehyde 19, the enolate of 18 was prepared using the chiral director 20.
To complete the synthesis, it was necessary to form the lactone between 1 and 2, and also to effect
alkene metathesis. The esterification proceeded smoothly, to give, after functional group

deprotection, the linear precursor 20. Alkene metathesis was efficient, but proceeded with little
geometric control. It would have been interesting to know what influence the several protecting
groups might have had on the geometry of the metathesis step.
53. Synthesis and Absolute Stereochemical Assignment of (-)-Galbulimima Alkaloid 13
The galbulimima alkaloids have intriguing physiological activity, but the absolute configuration of
glabulimima alkaloid 13 (3) had not been assigned. Mohammad Movassaghi of the Massachusetts
Institute of Technology has developed (J. Am. Chem. Soc. 2006, 128, 8126.) an elegant route to 3,
based on the convergent coupling of racemic 1 with the enantiomerically-pure 2 (for a racemic
synthesis of 3:).
The enantiomerically-pure iodide 4 was prepared from L-alaninol. Conjugate addition of the
derived radical to methyl vinyl ketone gave 5, which was hydrolyzed to 2.
The Diels-Alder substrate 11 was prepared from the 1,1-dibromo alkene 6. Pd-mediated coupling
of 6 with 7 gave 8. Further coupling with 9 followed by oxidation and enol ether formation gave
10, which underwent smooth cross-coupling with acrolein, yielding 11. The anticipated
intramolecular Diels-Alder cycloaddition led to (racemic) 1 with the expected high
diastereocontrol.

The enantiomerically-pure imine 2 was readily deprotonated, and the resulting anion was
condensed with the aldehyde 1 to give, after dehydration, the imine 12. The enol ether was
selectively brominated. Free radical reduction led to smooth cyclization, to give 13. On unmasking
of the ketone, spontaneous enamine addition ensued, to give the imine 14. Stereocontrolled
reduction with NaBH 4 and protection then gave the pentacylic 15, accompanied by the
diastereomer resulting from condensation of 2 with the other enantiomer of 1. These two
diastereomers were readily separable by flash chromatography.
To complete the synthesis, the enamide 15 was hydrolyzed by aqueous acid. In situ reduction
carried the liberated ketone on to the enone, which was deprotected to give glabulimima alkaloid
13 (3). Both enantiomers of 3 were prepared this way, beginning with each of the enantiomers of 4.
The enantiomer derived from L-alaninol gave the same rotation as the natural product, allowing
the assignment of the absolute configuration.
The convergent coupling employed here allowed the ready preparation of each of the four
enantiomerically-pure diastereomers of the natural product. There are real advantages to such an
approach, because it provides not just one substance, the natural product, but each of the four, for
further evaluation of physiological activity.
54. The Corey Route to the Dolabellanes: Isoedunol and β-Araneosene

A variety of dolabellanes, some of which show substantial physiological activity, have been
isolated from natural sources. E. J. Corey of Harvard University has introduced (J. Am. Chem. Soc.
2005, 127, 13813. ) a unfied approach to the dolabellanes, represented by isoedunol (3), based on
the designed rearrangement of the mesylate 1 to 2.
The key to this approach was the stereocontrolled construction of the cyclobutane 1. The starting
material was the racemic iodo acetonide 4 derived from farnesol. Alkylation of 5 using the
Seebach protocol followed by hydrolysis led the methylthiomethyl ether 7. The ester was
converted to the hydroxy cyclopropane 8 by the Kulinkovic procedure. On activation with Me 3 Al,
8 was smoothly carried on to the enantiomerically-pure cyclobutanone 9. The ring expansion must
not be proceeding by full ionization, as carbocation formation would have led to the racemic
product. The aldehyde derived from 9 was cyclized with SmI 2 to the trans diol 10.
In medium ring derivatives such as 10, one substituent on a ring carbon will be inside, and the
other will be outside. The conformation of 10 is such that formation of the mesylate from the
secondary alcohol led to migration of the more substituted cyclobutane bond, delivering 11. It
follows that the conformation of the diol 12 will be flipped, to keep the OH outside the ring.
Formation of the mesylate from the secondary alcohol of 12 led cleanly to migration of the less
substituted cyclobutane bond, to give the desired cyclopentanone 2.

Addition of 2-propenyl lithium to the cyclopentanone 2 gave the dolebellane isoedunol (3).
Deoxygenation converted 3 to the β-araneosene ( dolabellane13). This strategy for the
construction of the dolabellanes may open a route for the preparation of the cytotoxic dolabellanes
clavulactone (14) and clavirolide (15).
55. Adventures in Complex Indole Synthesis: (-)-Fischerindole I, (+)-Fischerindole G and
(+)-Weltwitindolinone A
Phil S. Baran of Scripps La Jolla has described (J. Am. Chem. Soc. 2005, 127, 15394.) an elegant
entry to the complex indole-derived natural products (-)-fischerindole G (1), (+)-fischerindole I (2)
and (+)-weltwitindolinone A (3).
The starting point for the preparation of 1 and 2 was the epoxide 4 of R-carvone. Enolization
followed by opening with vinyl magnesium bromide 5 delivered the alcohol 6. The yield of this
reaction was modest, but it directly provided 6, with all of the non-indole carbon skeleton of 1,
suitably oxidized, in enantiomerically-pure form.

Remarkably, chlorination of the hindered secondary cyclohexyl alcohol proceeded cleanly, to give
7. Oxidative condensation of the enolate of 7 with indole 8 gave the key intermediate 9. After
some experimentation, it was found that the clay Montmorillonite K-10 would effect the
cyclization to 10. There is not yet a general reductive amination procedure for converting a
cyclohexanone to the equatorial amine, so 10 was reduced to the axial alcohol, which was
converted to the equatorial azide. Selective reduction then delivered the amine 11, the formamide
of which was dehydrated to give 1. The alkaloid 1 so prepared was the enantiomer of the natural
product.
It was not possible to dehydrogenate 1 or the precursor formamide to 2. The ketone 10 was
therefore reduced selectively to the axial amine 12. Exposure of the derived formamide to
t-BuOCl followed by silica gel and Et 3 N installed the desired double bond. Dehydration with the
Burgess reagent then gave 2.
With the absolute configuration of the series established by the preparation of 1, the enantiomer of
the alkaloid 2 was prepared from S-carvone. Both R-carvone and S-carvone are inexpensive and
available in bulk. Exposure of the 2 so prepared to t-BuOCl followed by trifluoroacetic acid
induced oxidative rearrangement to (+)-weltwitindoline A (3), identical, including sign of rotation,
with the natural product.
These three syntheses, which remarkably proceed without resort to functional group protection,
underline the power of the oxidative indole coupling illustrated by the union of 7 and 8 to give 9.
The authors suggest that the facile conversion of 12 to 2, contrasted to the difficulties encountered
in attempting to prepare 2 from 11, could indicate that it is in fact the axial amine 12 or a
derivative that is the biosynthetic precursor to 2.
56. Synthesis of Erythronolide A
Erythronolide A (4), with its array of ten stereogenic centers, is the parent of several classic
antibiotics, including erythromycin. The key step in the total synthesis of 4 recently reported
(Angew. Chem. Int. Ed. 2005, 44, 4036. ) by Erick M. Carreira of the ETH Hönggerberg is the
activation of 1 and subsequent diastereoselective 1,3-dipolar cycloaddition to 2 to give 3.

The synthesis of 1 started with two simple chirons, the alcohol 2, prepared by Noyori reduction of
the acetylenic ketone followed by semi-hydrogenation, and the inexpensive (< one USD/gram)
Roche alcohol (5). Functional group manipulation led to 6, which underwent smooth Mg-mediated
cycloaddition to the enantiomerically-pure alcohol 2, to give 7. Addition of the Grignard reagent 8,
also derived from the Roche ester (5), to the derived methyl ketone also proceeded with high
diastereocontrol, to give the tertiary alcohol. Reduction of 9 and subsequent hydrolysis liberated
the hydroxy ketone, which was reduced selectively to the syn diol. It is a tribute to the stability of
tertiary triethylsilyl ethers that the TES protecting group put on at this stage survived all the way
to the end of the synthesis.
While the diastereoselectivity of 1,3-dipolar cycloaddition was well-precedented in simpler
systems, it was not clear that the diastereoselectivity would be maintained with such a complex
substrate as 1. In fact, addition to 2 proceeded with > 99:1 dr. Wittig homologation of the derived
methyl ketone proceeded with remarkable (33:1) geometric control, setting the stage for
asymmetric dihydroxylation to put in place the last two stereocenters.

Controlled desilylation of 11 followed by selective oxidation delivered the seco acid 12. It had
previously been shown by others that some cyclic protecting groups facilitate macrolactone
formation, while others do not. Fortunately, the two cyclic protecting groups of 12 served well,
and the macrolactonization proceeded efficiently. Protecting group removal and reductive
unmasking of the hydroxy ketone then delivered erythronolide A (4).
Overall, this elegant synthesis is a showcase for the iterative use of the 1,3-dipolar cycloaddition
of enantiomercially-pure allylic alcohols for the preparation of extended arrays of acyclic
stereogenic centers.
57. The Boger Route to (-)-Vindoline
The Vinca-derived vinblastine (2a) and vincristine (2b) are still mainstays of cancer chemotherapy.
The more complex half of these dimeric alkaloids, vindoline (1), has in the past presented a
formidable challenge for total synthesis. Dale L. Boger of Scripps, La Jolla has developed (Org.
Lett. 2005, 7, 4539.) a strikingly simple solution to this problem, based on sequential
cycloaddition.
The starting point for the synthesis was N-methyl 6-methoxytryptamine (3), an improved
preparation of which is described by the authors. This was extended to 4, which was then cyclized
to 5, and acylated with 6 to give 7. On heating, 7 cyclized to 8, which lost N 2 to give the
zwitterion 9. Addition of the intermediate 9 to the indole then gave 10. In one reaction, the entire

ing system of vindoline, appropriately oxygenated, was assembled! The precursor 7 was flat, so it
offered no opportunity for chiral synthesis. Fortunately, 10 and ent-10 proved to be very easy to
resolve by chiral chromatography.
To complete the synthesis, the δ-lactam 11 was oxygenated to 12. Conditions for desulfurization
of the derived thiolactam also effected debenzylation, to give, after acetylation, the ether 13.
Pt-mediated hydrogenolysis gave 14, which was dehydrated to vindoline (1).
A complementary synthetic route, based on the E isomer of 6, and so leading through the ether 15,
is also described. Although slightly longer, this approach was about as efficient as the route via
11.

58. The Stork Synthesis of (-)-Reserpine
The history of reserpine (3), isolated from the Indian shrub Rauwolfia serpentina, is storied. The
anxiolytic activity of reserpine, the first tranquilizer, pointed the way to the development by
Hoffmann-LaRoche of the blockbuster drugs Librium and Valium.
The spectacular first total synthesis of reserpine was achieved by R.B. Woodward in 1956. Several
alternative approaches have been been published since that time. The most recent (J. Am. Chem.
Soc. 2005, 127, 16255. ), by Gilbert Stork of Columbia University, centering on the aldehyde
tosylate 2, illustrates the power of chiral induction for the kinetic establishment of distal
stereocenters. Condensation of 2 with 6-methoxytryptamine (1) led to reserpine (3).
The preparation of 2 started with the previously-reported enantioselective addition of acrylate to
butadiene, to give the acid 6. Iodolactone formation followed by reduction gave the diol 7. Under
the conditions of benzyl ether formation, the iodohydrin cyclized to the epoxide, giving 8. Phenyl
selenide added to 8 to give the expected diaxial product 9, which on oxidation gave 10 in high
enantiomeric excess.
The key reaction in the assembly of 2 was the addition of the kinetic lithium enolate of 10 to the
silyl acrylate 11 to give 12. This reaction is seen as involving two sequential Michael additions,
but the stereochemical outcome is the same as would be expected from a concerted Diels-Alder
cycloaddition. Exposure to TBAF converted the furyl silane to the fluorosilane, which was
debenzylated and carried on to the tosylate 13. On exposure to two equivalents of hydrogen
peroxide, the ketone underwent Baeyer-Villiger oxidation with high regioselectivity. The silane
was also oxidized, delivering 14. Methylation followed by Dibal reduction then gave 2.

An additional stereogenic center is created when 1 and 2 are combined. Initial attempts to carry
out the condensation gave the wrong stereochemical outcome, as Pictet-Spengler condensation
preceded tosylate displacement. To work around this problem, 1 and 2 were combined in the
presence of cyanide ion, to give 15. Heating of 15 gave cyclization, but again to the wrong
diastereomer, perhaps because in the intermediate ion pair from cyanide ionization, the cyanide
ion was blocking one face of the intermediate cation. Fortunately, on stirring at room temperature
in aqueous HCl, 15 did cyclize to the correct diastereomer, providing, after acylation, reserpine
(3).
59. The Overman Route to Gelsemine
Gelsemine 3 has no particular biological activity that recommends it, but its challenging
architecture has been a motivation to a generation of organic synthesis chemists. Larry E.
Overman of the University of California at Irvine has described (J. Am. Chem. Soc. 2005, 127,
18046, 18054,) his total synthesis of gelsemine, the key step of which was the acid-mediated
cyclization of 1 to 2.
The synthesis of 1 started with the preparation of the diene 4 from 3-methyl anisole. Diels-Alder
cycloaddition with methyl acrylate gave the endo adduct 7. The methyl group was converted to the
vinyl group of the natural product by allylic oxidation followed by methylenation. Hydrolysis
followed by inversion and trapping with p-methoxybenzyl alcohol gave the protected amine 7.
This underwent smooth aza-oxy-Cope rearrangement, to give, after protection and bromination,
the ketone 1.

The acid-mediated cyclization of 1 to 2 proceeds by way of the protonated enol 9. It seems likely
that kinetic protonation at the indicated carbon would have led to the other diastereomer of the
bromide. With that diastereomer, however, the Br would be inside in the transition state leading to
cyclization. That may slow down the cyclization sufficiently that epimerization can intervene,
leading to the more easily cyclized diastereomer 9 and thus to 2.
There were several problems to solve in the conversion of 2 to 3. The key step was the
intramolecular Heck cyclization of 11 to 12. There was some concern that such a cyclization
would not proceed with a tetrasubstituted alkene. In fact, Moriarty methoxylation of the ketone 2
followed by generation of the enol triflate delivered 10. Pd-mediated carbonylation followed by
coupling with o-iodoaniline and protection gave 11, which underwent smooth intramolecular Heck
cyclization, to give, after hydrolysis, predominantly the unnatural diastereomer 12.
The completion of the synthesis, including the addition of the last carbon, proved elusive. Early on,
it was found to be important to reduce and protect the 1,3-dicarbonyl system, to give the equatorial
ether 13. Attempted displacement to introduce the last carbon led instead to the aziridine 14.
Taking advantage of this, the aziridine was quaternized, then opened at the less-hindered
secondary site. Deprotection followed by warming with base led, via retro-aldol epimerization of
two stereogenic centers, to the long-sought lactone 16. Deprotection followed by reduction of the
lactone then gave 3.

60. Synthesis of the Potent FBBP12 Ligand Antascomicin B
The FK binding protein ligands that suppress the immune response, such as FK506 and rapamycin,
have been thoroughly studied. FKBP ligands have also been shown to promote the regrowth of
damaged neurons, both peripherally and in the central nervous system. To differentiate these two
disparate activities, it is important to develop potent FKBP ligands that are not immune
suppressive. The antascomicins, represented by antascomicin B (2), are such a class of natural
products. Steven V. Ley of the University of Cambridge has described (Angew. Chem. Int. Ed.
2005, 44, 2732. ) an elegant synthesis of 2, the key step of which was the ring-contracting
transannular Dieckmann cyclization of 1.
The enantiomerically-pure fragments 8, 12 and 17 were coupled to prepare the macrolide 1. The
preparation of 8 started with the commercially-available enantiomerically-pure bromide 3.
Protection and halide exchange set the stage for homologation with allylmagnesium chloride.
Ozonolysis followed by condensation with the acyl oxazolidinone 6 set the last two stereogenic
centers of 8.
The starting point for 12 was the commercially-available enantiomerically-pure Roche ester (9).
Protection, reduction and oxidation gave 10, which was homologated to the ester 11. Reduction to
the allylic alcohol followed by conversion to the chloride and coupling with TMS acetylene led to
12. Addition to 8 of the alkenyl zirconium derived from 12 then gave 13.
Professor Ley used his elegant tartrate method to assemble the carbocyclic fragment 17.
Condensation of two inexpensive components, dimethyl D-tartrate and biacetyl, followed by

eduction and monoprotection delivered the aldehyde 14. Diastereoselective addition of the allyl
stannane 15 led to diene 16, setting the stage for cyclization with the Grubbs second-generation
Ru catalyst.
The sulfone anion derived from 13 added smoothly to 17, to give, after reduction and acylation,
the ester 1. The templating effect of the arene ring of 1 facilitated macrolide formation. The
Dieckmann cyclization then could proceed via a 6-membered ring transition state, leading to the
ring-contracted product 18. In addition to establishing the β-keto amide, this cyclization left
residual oxygenation at the α-carbon, allowing direct elaboration of the 1,2,3-tricarbonyl system of
2.
61. Synthesis of the Potent FBBP12 Ligand Antascomicin B
The FK binding protein ligands that suppress the immune response, such as FK506 and rapamycin,
have been thoroughly studied. FKBP ligands have also been shown to promote the regrowth of
damaged neurons, both peripherally and in the central nervous system. To differentiate these two
disparate activities, it is important to develop potent FKBP ligands that are not immune
suppressive. The antascomicins, represented by antascomicin B (2), are such a class of natural
products. Steven V. Ley of the University of Cambridge has described (Angew. Chem. Int. Ed.
2005, 44, 2732.) an elegant synthesis of 2, the key step of which was the ring-contracting
transannular Dieckmann cyclization of 1.

The enantiomerically-pure fragments 8, 12 and 17 were coupled to prepare the macrolide 1. The
preparation of 8 started with the commercially-available enantiomerically-pure bromide 3.
Protection and halide exchange set the stage for homologation with allylmagnesium chloride.
Ozonolysis followed by condensation with the acyl oxazolidinone 6 set the last two stereogenic
centers of 8.
The starting point for 12 was the commercially-available enantiomerically-pure Roche ester (9).
Protection, reduction and oxidation gave 10, which was homologated to the ester 11. Reduction to
the allylic alcohol followed by conversion to the chloride and coupling with TMS acetylene led to
12. Addition to 8 of the alkenyl zirconium derived from 12 then gave 13.
Professor Ley used his elegant tartrate method to assemble the carbocyclic fragment 17.
Condensation of two inexpensive components, dimethyl D-tartrate and biacetyl, followed by
reduction and monoprotection delivered the aldehyde 14. Diastereoselective addition of the allyl
stannane 15 led to diene 16, setting the stage for cyclization with the Grubbs second-generation
Ru catalyst.
The sulfone anion derived from 13 added smoothly to 17, to give, after reduction and acylation,
the ester 1. The templating effect of the arene ring of 1 facilitated macrolide formation. The
Dieckmann cyclization then could proceed via a 6-membered ring transition state, leading to the
ring-contracted product 18. In addition to establishing the β-keto amide, this cyclization left
residual oxygenation at the α-carbon, allowing direct elaboration of the 1,2,3-tricarbonyl system of
2.

62. Synthesis of (-)-Avrainvillamide and (+)-Stephacidin B
The dimeric alkaloid stephacidin B (1) was recently isolated from a fungus culture. The
“monomer” avrainvillamide (2) had previously been described. Andrew G. Myers of Harvard
University has reported (J. Am. Chem. Soc. 2005, 127, 5342.) the enantioselective total synthesis
of 2, and the dimerization of 2 to 1. The key intermediate in the synthesis was the tetracyclic
amide 3.
The absolute configuration of the target natural products was set by enantioselective reduction of
the enone 6. Usually, catalytic Itsuno-Corey reduction of cyclohexenones without an α-substituent
is not selective. In this case, advantage was taken of that lack of induction from the alkene side,
with the steric bulk on the other side of the ketone directing the reduction. Alkylation of 8 with 9
proceeded to give the expected axial product 10. Cyanation proceeded with remarkable
diastereocontrol, to give, after epimerization and hydrolysis, the amide 11. Conjugate addition of
thiophenol followed by spontaneous cyclization and dehydration led to the amide 12. With the
phenylthio enamide in place, the stage was set for the elegant final cyclization: hydrogen atom
abstraction from the dihydroaromatic followed by fragmentation delivered the formamide radical,
that cyclized efficiently to give the tetracyle 13. Oxidation and iodination of the enone then gave
3.

Ullman coupling of 3 with the aryl iodide 4 to give 5 proved to be more effective than the
alternative coupling with the areneboronic acid. Reduction of 5 with activated zinc powder
converted the nitro group to the N-OH, which spontaneously cyclized to the nitrone 2. While 2 so
preparared gave a 13 C spectrum that was congruent with that of natural avrainvillamide, authentic
material was not available, so a direct comparison could not be made.
In Et 3 N and CH 3 CN, 2 spontaneously dimerized to 1. As 2 is levorotatory (-35°) and 1 is
dextrorotatory (91°), this ready interconversion of the monomer and the dimer will make it
difficult to assign the absolute configuration of either natural product solely by comparison of
rotations.
63. Synthesis of (-)-Littoralisone
The loss of mental function associated with aging is thought to be due at least in part to the
degradation of neurite connections between neurons. Natural products such as (-)-littoralisone (3)
that promote neurite outgrowth in cell culture are therefore interesting lead compounds for
pharmaceutical discovery. The synthesis of 3 recently reported (J. Am. Chem. Soc. 2005, 127,
3696. ) by David W. C. MacMillan of Caltech uses organocatalysis to assemble 1, to control the
relative configuration of the ring system of 2, and to assemble the glucose component of 3.
The preparation of 1 began with commercial enantiomerically-pure citronellol 4. Ozonolysis of
the ester delivered the aldehyde 5, which was hydroxylated with high diastereocontrol

(enantiocontrol), using the proline-catalyzed procedure developed by MacMillan (Org. Chem.
Highlights 2004, January 26.). Protection and homologation of 6 then gave 1.
With most catalysts the dialdehyde 1 cyclized predominantly to the trans-fused product. With the
correct enantiomer of the organocatalyst, however, the cyclization could be directed toward the
desired cis-fused 2. Vilsmeier homologation of the electron-rich alkene followed by oxidation and
lactonization then delivered 8.
A glucose derivative such as 12 would usually be prepared over several protection and
deprotection steps from a commercially-available form of glucose. The authors took a different
approach, using the method for protected glucose synthesis they have developed (Org. Chem.
Highlights 2005, March 21.). Thus, proline-catalyzed dimerization of 9 gave 10, which after
condensation with 11 and benzylation gave 12.
Condensation of 12 with 8 gave the exo glycoside 13. Brief exposure of 13 to 350 nm irradiation
followed by debenzylation then gave 3. It is interesting to note that the photocyclization of 13
proceeded slowly even under ambient laboratory light, suggesting that this step in the biosynthesis
of 3 need not be enzyme-mediated.
64. Total Synthesis of (-)-Sordaricin

Sordaricin (2) is the aglycone of sordarin (3), the parent of a family of clinically-effective
antifungal agents. Lewis N. Mander of the Australian National University has published (J. Org.
Chem. 2005, 70, 1654.) a full account of their enantioconvergent synthesis of 2 (see below), based
on the intramolecular Diels-Alder cyclization of the triene 1. For a complementary total synthesis
by Narasaka of sordaricin (2), see Org. Chem. Highlights 2005, April 4.
To prepare 1, two enantiomerically-pure pieces were needed, 10 and 11. As it developed, 10 was
prepared from one enantiomer of 4, and 11 was prepared from the other enantiomer of 4. In both
the synthesis of 10 and the synthesis of 11, the folded nature of the tricyclic ring system was used
to control relative (and absolute) configuration around the cyclopentane ring. Conjugate addition
to (+)-4 followed by alkylation led to the cis dialkyl 5. The retro Diels-Alder unmasking of the
enone was effected by heating to reflux in 1,2-dichlorobenzene while a stream of nitrogen was
bubbled through the solution. Conjugate addition to 6 proceeded to the more open face of the
enone, to give 7. Removal of the carbonyl under conditions that did not epimerize the adjacent
methyl group followed by oxidation of the isopropenyl group then led to 10.
With the iodide 10 in hand, the preparation of the triene 15 could proceed. The single quaternary
center of 15 was constructed by deprotonation of the nitrile 11, followed by alkylation with 10.
Again, bond formation proceeded on the outside face of the tricyclic ketone, to give 12. Because it
was thought (correctly, as it developed) that the Diels-Alder cyclization would proceed with
higher selectivity at lower temperature, 13 was cracked to give the enone 14.
Methoxycarbonylation with the Mander reagent followed by enol triflate formation and
Cu-mediated coupling then delivered the triene 15.

Deprotection of 15 followed by selective oxidation gave the aldehyde 1. In the event, cyclization
of 1 proceeded smoothly, albeit a little slowly, near room temperature, to give exclusively the
desired regioisomer. Demethylation then gave (-)-sordaricin (2).
It is interesting that on heating 1 or the cycloadduct above 150°C, the predominant product (3~4:1)
from the cycloaddition is the undesired regioisomer 16. So, even though the transition state in the
intramolecular Diels-Alder cyclization is product-like, the more stable transition state for the
cyclization of 1 in fact leads to the thermodynamically less stable regioisomer 2.
65. Total Synthesis of the Tetracyclines
Although the tetracycline antibiotics have been mainstays of antibacterial chemotherapy for
decades, they had eluded efficient total synthesis. In a landmark accomplishment, Andrew G.
Myers of Harvard University recently reported (Science 2005, 308, 395,; J. Am. Chem. Soc. 2005,
127, 8292,) the first such syntheses.
In the Science paper, which was published first, total syntheses of the clinically-important
6-deoxytetracyclines, including doxycycline (9), are described. The starting point for the synthesis
was the enantiomerically-pure ester 2, prepared by fermentation of benzoic acid (1) to the
1,2-dihydrodiol, followed by epoxidation, rearrangement and silylation. Acylation of 3 with 2
gave the ketone 4, which on exposure to LiOTf underwent a very interesting, and
diastereoselective, carbon-carbon bond forming reaction to give, after selective desilylation with
TFA, the alcohol 5. The authors speculate that this reaction is proceeding by initial S N 2´epoxide
opening by the N, followed by ylide formation and 2,3-rearrangement.

The alcohol 5 was the common intermediate for both syntheses. For the deoxy series, 5 was
carried on to the enone 6. Conjugate addition of the anion 7 proceeded with remarkable
diastereoselectivity, to give, after intramolecular acylation and deprotection, doxycycline (9).
The JACS paper describes the total synthesis of the more highly oxygenated (-)-tetracycline (16).
To this end, the alcohol 5 was carried on to the enone 10. Opening of the cyclobutane 11 to the
o-quinone methide followed by Diels-Alder cycloaddition to 10 delivered the endo adduct 12.
Deprotection and oxidation of 12 gave 13, which was further oxidized to the sulfoxide.
Elimination of the sulfoxide gave the naphthalene derivative 14, which underwent spontaneous
oxidation to 15. Reductive deprotection then gave tetracycline (16). The diastereoselectivity of the
air and light-mediated oxidation is remarkable.
66. Synthesis of Erythronolide A
Erythronolide A (4), with its array of ten stereogenic centers, is the parent of several classic
antibiotics, including erythromycin. The key step in the total synthesis of 4 recently reported
(Angew. Chem. Int. Ed. 2005, 44, 4036. ) by Erick M. Carreira of the ETH Hönggerberg is the
activation of 1 and subsequent diastereoselective 1,3-dipolar cycloaddition to 2 to give 3.

The synthesis of 1 started with two simple chirons, the alcohol 2, prepared by Noyori reduction of
the acetylenic ketone followed by semi-hydrogenation, and the inexpensive (< one USD/gram)
Roche alcohol (5). Functional group manipulation led to 6, which underwent smooth Mg-mediated
cycloaddition to the enantiomerically-pure alcohol 2, to give 7. Addition of the Grignard reagent 8,
also derived from the Roche ester (5), to the derived methyl ketone also proceeded with high
diastereocontrol, to give the tertiary alcohol. Reduction of 9 and subsequent hydrolysis liberated
the hydroxy ketone, which was reduced selectively to the syn diol. It is a tribute to the stability of
tertiary triethylsilyl ethers that the TES protecting group put on at this stage survived all the way
to the end of the synthesis.
While the diastereoselectivity of 1,3-dipolar cycloaddition was well-precedented in simpler
systems, it was not clear that the diastereoselectivity would be maintained with such a complex
substrate as 1. In fact, addition to 2 proceeded with > 99:1 dr. Wittig homologation of the derived
methyl ketone proceeded with remarkable (33:1) geometric control, setting the stage for
asymmetric dihydroxylation to put in place the last two stereocenters.
Controlled desilylation of 11 followed by selective oxidation delivered the seco acid 12. It had
previously been shown by others that some cyclic protecting groups facilitate macrolactone
formation, while others do not. Fortunately, the two cyclic protecting groups of 12 served well,
and the macrolactonization proceeded efficiently. Protecting group removal and reductive
unmasking of the hydroxy ketone then delivered erythronolide A (4).

Overall, this elegant synthesis is a showcase for the iterative use of the 1,3-dipolar cycloaddition
of enantiomercially-pure allylic alcohols for the preparation of extended arrays of acyclic
stereogenic centers.
67. Synthesis of (+)-Cyanthiwigin U
The cyanthin diterpenes show physiological activity ranging from cytotoxicity to nerve-growth
factor stimulation. Andrew J. Phillips of the University of Colorado recently described (J. Am.
Chem. Soc. 2005, 127, 5334.) a concise enantioselective synthesis of cyanthiwigin U (3), based on
the metathesis conversion of 1 to 2, using the second generation Grubbs catalyst.
It was clear that 1 would be derived from a Diels-Alder adduct. There has been a great deal of
work in recent years around the development of enantioselective catalysts for the Diels-Alder
reaction, but the catalysts that have been developed to date only work with activated
dienophile-diene combinations. For less reactive dienes, it is still necessary to use chiral auxiliary
control. One of the more effective of those was the known camphor-derived tertiary alcohol, so
that was used in this project. Diels-Alder cycloaddition of the diene 4 with the
enantiomerically-pure enone 5 led to the adduct 6 with high diastereocontrol. Oxidative cleavage
led to the acid 7, which was carried on to the bis-enone 1.
The two-directional tandem metathesis of 1 to 2 proceeded smoothly using 20 mol % of the
second generation Grubbs catalyst under an atmosphere of ethylene. The conversion of 2 to 3 took
advantage of the differing reactivity of the two ketones. Addition of hydride to 2 from the less
hindered face of the less hindered ketone delivered 4. Addition of isopropyl lithium to the
surviving ketone followed by oxidative rearrangement of the resulting tertiary allylic alcohol and
concomitant oxidation of the secondary allylic alcohol gave the diketone 10. Selective addition of
methyl lithium to the less hindered of the two ketones, again from the more open face, then gave
3.

The elegantly concise strategy displayed here for the enantioselective and diastereoselective
construction of the tricyclic enone 3, by two-directional tandem methathesis of the Diels-Alder
derived diketone 1, should have some generality
67. Enantioselective Synthesis of (-)-Epoxomycin
For many classes of physiologically-important natural products, total synthesis requires the
construction of an extended array of acyclic stereogenic centers, with control of relative and
absolute configuration. In the course of a synthesis (J. Am. Chem. Soc. 2004, 126, 15348.) of the
proteasome inhibitor (-)-epoxomycin (3), Lawrence J. Williams of Rutgers University has
developed an elegant and potentially powerful solution to this problem. It was apparent that the
preparation of 3 reduced to the stereocontrolled synthesis of 2, since the balance of the natural
product is made up of readily-available amino acids. The oxygenated quaternary center of 2
appeared to be a particular challenge. The key insight of the synthesis was that both centers could
be established in a single step by selective nucelophilic opening of the enantiomerically-enriched
spiro bis epoxide 1.
The left hand fragment of epoxymycin 3 was assembled from the previously-described amino acid
derivatives 4, 5, and 7, using standard coupling techniques.
The preparation of the allene bis-epoxide 1 started with isovaleraldehyde 9. Addition of the
protected propargyl alcohol 10 under the Carreira conditions led to 11 in > 95% ee. Mesylation
followed by displacement with methyl cuprate provided the allene without loss of enantiomeric
excess. Oxidation of the allene 12 with dimethyldioxirane could have led to any of the four
diastereomers of the spiro bis epoxide. In the event, only two diastereomers were observed, as a

3:1 mixture. That 1 was the major diastereomer followed from its conversion to 3. The
configuration of the minor diastereromer was not noted. Exposure of 1 to nucleophilic azide then
gave the easily-purified 2.
The spirodiepoxide 1 is an intriguing new approach to relative and absolute stereocontrol. It will
be interesting to see what other nucleophiles can be used in the opening. It is possible that a chiral
oxygen transfer reagent, such as the dioxirane prepared by the Shi protocol, would convert 12 to 1
with improved diastereoselectivity (double diastereoselection)
68. Enantioselective Synthesis of the Polyene Antiobiotic Aglycone Rimocidinolide Methyl
Ester
The complex polyene macrolide antibiotics are clinically effective as antifungal agents. Scott
Rychnovsky of the University of California at Irvine has reported (Angew. Chem. Int. Ed. 2004, 43,
2822. ) the first synthesis of rimocidinolide methyl ester (4), the aglycone of rimocidin (1). The
key step in the synthesis is the condensation of the aldehyde 2 with the phosphonate 3, leading to
4.
Preparation of the Aldehyde 2: The absolute configuration of the triene aldehyde 2 was set by
Noyori hydrogenation of ethyl butyrylacetate 5. Silylation and Dibal reduction then gave the
aldehyde 6. Reduction of the homologated ester gave the alcohol, which was oxidized to the

desired aldehyde 7 by the Swern procedure. Condensation of 7 with the Wollenberg stannyl diene
followed by deprotection then gave the unstable aldehyde 2.
The power of convergent synthesis is illustrated by the preparation of the acid 3. The three
components 9, 12, and 17 were each prepared in enantiomerically-pure form using
readily-available chiral reagents, followed by functional group manipulation. One of the more
remarkable transformations was the homologation of the Weinreb amide 15 to give the unstable
allyl ketone, which was then reduced with high diastereoselectivity to give the diol 16.
Convergent coupling of 17 with 9, followed by functional group manipulation, gave the iodide 18,
which was then homologated with 12 to give 19. Although the two monosubstituted alkenes of 19
appear to be similar, dihydroxylation with OsO 4 was remarkably selective, leading to the aldehyde
20.
To complete the synthesis of 1, the acid 3 derived from 20 was converted to the mixed anhydride
(Yamaguchi coupling), then esterified with 2. Exposure to K 2 CO 3 /18-crown-6 gave the all-E
tetraene, which was deprotected to give the aglycone 4
69. Synthesis of (+)-Brasilenyne
(+)-Brasilenyne (3), isolated from the digestive gland of the sea hare Aplysia brasiliana, shows
significant antifeedant activity. Scott Denmark of the University of Illinois has described (J. Am.
Chem. Soc. 2004, 126, 12432.) an elegant synthesis of 3, the key step of which is the Pd-mediated
intramolecular coupling of 1 to give 2.

The enantiomerically-pure intermediate 1 was prepared from the dioxolanone 4, available in three
steps from L-malic acid. Lewis acid-mediated homologation converted 4, a 4:1 mixture of
diastereomers, into 5 as a single diastereomer. After establishment of the alkenyl iodide, it was
necessary to maintain the lactone in its open form. A solution was found in the formation of the
Weinreb amide. The final stereogenic center was established by Brown allylation of the derived
aldehyde. The alkene metathesis to form 1 was carried out with the commercially-available
Schrock Mo catalyst. The authors did not comment on the relative efficacy of alternative alkene
metathesis catalysts.
With 1 in hand, the stage was set for the proposed Pd-mediated coupling. The authors were
pleased to observe that the coupling conditions previously developed in the group worked
efficiently with this much more complex substrate, leading to 2 as a single geometric isomer.
After protection-deprotection and oxidation, homologation using the Corey protocol gave 10.
Formation of the chloride proceeded with the expected clean inversion of absolute configuration,
to give 3.
The least precedented transformation in this synthesis is the homologation of 4 to 5. This appears
to be a general solution to a longstanding challenge, the construction of secondary-secondary
ethers with absolute stereocontrol.
70. Synthesis of (-)-Norzoanthamine

The marine alkaloid (-)-norzoanthamine (3) suppresses bone loss in ovariectomized mice, and so
is of interest as a lead to antiosteoporotic drugs. A substantial challenge in the assembly of 3 is the
stereocontrolled construction of the C ring, with its three all-carbon quaternary centers. In the
synthesis of 3 by Masaaki Mayashita of Hokkaido University (Science 2004, 305, 495. DOI), the
B and C rings were built and two of the three needed quaternary centers were set by the
intramolecular Diels-Alder cyclization of 1 to 2.
The triene 1 was prepared from the enone 4, available in enantiomerically-pure form over several
steps from pulegone. Triply-convergent coupling with the cuprate 5 and the aldehyde 6 led to the
furan 7. Functional group manipulation then gave 1, setting the stage for the intramolecular
Diels-Alder cyclization.
The cyclization of 1 proceeded with 72: 28 diastereoselectivity, leading, after hydrolysis, to the
crystalline diketone 8 as the major product. Reduction of the ketones to the axial alcohols was
followed by spontaneous lactonization, allowing easy differentiation of the several functional
groups. Homologation to 10 followed by condensation with methyl carbonate and subsequent
O-methylation then gave 11. C-Methylation of 11 then set the third quaternary center of the C ring.
The deuteriums were introduced to minimize an unwanted intramolecular hydride transfer in a
later step.
The EFG ring carbons were attached by condensing 13, derived from 12, with the aldehyde 14,
followed by oxidation. Hydrogenation followed by deprotection and oxidation then led to the
triketone 16. Selective enolization of the A-ring ketone allowed facile oxidation to the alkene.

Deprotection and acid-catalyzed cyclization then gave 3. Note that the original stereocenter in 4
was used to set seven of the nine stereocenters of 3.
71. Synthesis of (-)-Tetrodotoxin
Tetrodotoxin 3, the toxic principle of pufferfish poison, is a formidable challenge for synthesis,
with each carbon of the cyclohexane functionalized. Minoru Isobe of Nagoya University recently
reported (Angew. Chem. Int. Ed. 2004, 43, 4782.) a second-generation synthesis of
enantiomerically-pure tetrodotoxin. This synthesis features the rapid construction of the
cyclohexene by Diels-Alder cycloaddition using an enantiomerically-pure dienopile, the early
introduction of the aminated quaternary center, and the use of that center to direct the relative
configuration of further functionalization around the ring.
The synthesis started with levoglucosenone 4, available by the pyrolysis of cellulose, e.g. old
newspapers. Bromination-dehydrobromination gave the enantiomerically-pure Diels-Alder
dienophile 5, which was combined with isoprene to give predominantly the crystalline adduct 1.
Hydrolysis and acetylation led to 6, which was carried on to the geometrically-defined allylic
alcohol 7 via reduction with Zn-Cu couple. Overman rearrangement of 7 proceeded with high
facial control, to give 8.

The next stage of the synthesis was ring oxygenation, to convert 8 into 13. The key to this
transformation was the observation that the amide oxygen of 8 participated in the solvolysis of the
allylic bromide, setting, after hydrolysis, the new secondary stereocenter of 9. Hydroxyl-directed
epoxidation gave 10, which was rearranged with Ti(O-i-Pr) 4 to 11. After some experimentation, it
was found that the derived dione 12 could be reduced to the desired cis diol 13 with LiBr and
LiAlH(O-t-Bu) 3 followed by NaBH 4 /CeCl 3 .
Silylation followed by selenium dioxide oxidation converted 13 into 14. Epoxidation of the
derived TES ether proceeded by addition of oxygen to the more open face of the alkene, leading to
15. Ozonolysis followed by diastereoselective one-carbon homologation provided 17. This set the
stage for intramolecular epoxide opening by the carboxylate, to give 2, in which all of the
stereogenic centers of tetrodotoxin have been established.
Justin Du Bois of Stanford University has put forward (J. Am. Chem. Soc. 2003, 125, 11510. ) a
quite different total synthesis of tetrodotoxin, including an elegant late-stage introduction of the
nitrogen

72. Total Synthesis of (±)-Sordaricin
Carbohydrate derivatives of sordaricin 3 are clinically-effective antifungal agents, but
development efforts were halted when a suffficient supply of 3 could not be established. Koichi
Narasaka of the University of Tokyo recently reported (Chem. Lett. 2004, 33, 942.) a total
synthesis of 3, based on the elegant Pd-mediated cyclization of 1 to 2.
The 5-7-5 skeleton of 1 was assembled from cyclohexenone 4 by conjugate addition followed by
enolate trapping. Simmons-Smith cyclopropanation of 5 led to the cyclopropyl alcohol 6.
Generation of the oxy radical from 6 led to fragmentation to give 7, which further cyclized to give
8. Note that the seven-membered ring is flexible enough that the radical cyclization delivers the
required trans ring fusion. Annulation to 9 followed by kinetically-controlled conjugate addition
then gave 10.
The β-keto ester was protected as the enol acetate, then the ketone 12 was homologated to the
carbonate 1. Despite the strain in the [2.2.1] system being formed, the Pd-mediated cyclization of
1 to 2 proceeded smoothly.

To complete the synthesis, the ketone of 2 was homologated to the alkene 12. Selective oxidative
cleavage of the two vinyl groups followed by reduction provided the diol, the less encumbered
alcohol of which was protected to deliver the fully-differentiated ester 13. Oxidation followed by
ester cleavage then gave 3
73. Synthesis of (-)-Hamigeran B
Of all the ring-forming reactions of organic synthesis, diastereoselective intramolecular
Diels-Alder reactions are among the most powerful. Often, as illustrated by the cyclization of 1 to
2, a single stereogenic center can set the relative and absolute configuration of two rings. The
cyclization of 1 to 2 is the key step in the total synthesis of (-)-hamigeran B recently reported (J.
Am. Chem. Soc. 2004, 126, 613. ) by K.C. Nicolaou of the Scripps Research Institute.
The preparation of 1 started with the addition of lithiated 4 to the enantiomerically-pure epoxide 5,
which was prepared from the racemate using the Jacobsen protocol. Reduction followed by
selective protection of the primary alcohol gave the monosilyl ether, which was further protected
with MOM chloride to give 7. Pd-mediated oxidation to the methyl ketone followed by
condensation with the Horner-Emmons reagent gave the unsaturated ester 8 as an inconsequential
mixture of geometric isomers. Oxidation then set the stage for the crucial cyclization.
On irradiation, the aldehyde 1 underwent photoenolization to give the quinone methide 9.
Intramolecular Diels-Alder cyclization then proceeded with high diastereocontrol to give 2 as a
mixture of epimeric esters.

The ester 2 has a trans 6-5 ring fusion, whereas in the desired 3 the ring fusion is cis. This was
easily corrected, as the 6-5 ring fusion is more stable cis. Acid-mediated dehydration to give the
alkene proceeded with concomitant removal of the MOM protection. Osmylation followed by
acetonide formation and subsequent oxidation gave the ketone, which was readily epimerized to
10. The acetonide was designed to block H addition to the bottom side, so hydrogenation of 11
would give the isopropyl group endo. In fact, hydrogenation led to the undesired exo isopropyl,
but hydroboration proceeded from the exo face, leading to the desired 12.
Hydrolysis of the acetonide followed by oxidation and bromination provided the ketone 13, which
is itself a natural product, hamigeran A. Hydrolysis under aerobic conditions led first to
decarboxylation, then to autooxidation, to give(-)-hamigeran B 3
74. Synthesis of the Dendrobatid Alkaloid 251F
The Dendrobatid “poison arrow” frogs of Central and South America exude a potent mixture of
alkaloids from their skins. It was originally thought that the frogs biosynthesized these alkaloids,
but it has since been shown that they are of dietary origin. The skin exudate of the Colombian frog
Minyobates bombetes causes severe locomotor difficulties, muscle spasms and convulsions upon
injection in mice. The major component of the alkaloid mixture is 251F 3. Jeff Aubé of the
University of Kansas recently described (J. Am. Chem. Soc. 2004, 126, 5475.) the enantioselective
total synthesis of 3. The key step in the synthesis was the cyclization of the keto azide 2.
The ketone 1, with four ternary stereogenic centers in one cyclopentane ring, was a significant
challenge for synthesis. A clever solution flowed from the idea of coupling a chiral Diels-Alder
reaction with alkene metathesis. The relative and absolute configuration of 1 were set by the
Diels-Alder cycloaddition of the acyl oxazolidine 4 to cyclopentadiene, to give 5. Homologation
of 5 to the enone 6 set the stage for alkene metathesis, mediated by the Grubbs catalyst 7, to give
1.

Conjugate addition to 1 proceeded across the open face of the bicyclic system to give an enolate,
condensation of which with the enantiomerically-pure aldehyde 8 gave the enone 9. Conjugate
reduction of the enone also removed the benzyl ether, to give the alcohol. Conversion of the
alcohol to the azide gave 10. Ozonolysis followed by selective reduction then gave 2.
The acid-mediated intramolecular addition of the azide to the ketone proceeds by way of addition
followed by a pinacol-type rearrangement, to give the amide 12. The regioselectivity of the bond
migration is remarkable. Reduction of the amide then gave 3.
75. Enantioselective Synthesis of (+)-Tricycloclavulone
Triclavulone 3, recently isolated from Clavularia vulgaris, is one of the most complex of the
family of about forty structurally-related fatty acid-derived marine prostanoids described from this
species. Hisanaka Ito and Kazuo Iguchi of the Tokyo University of Pharmacy and Life Science
recently reported (J. Am. Chem. Soc. 2004, 126, 4520. ) the total synthesis of 3, starting with the
preparation of the enantiomerically-enriched bicyclic ketone 1.

The enantioselective Cu-catalyzed cycloaddition of the reactive enone 4 to the alkyne 5 proceeded
efficiently, but with only modest enantiomeric excess. This was improved at a later point in the
synthesis, by recrystallization of 2.
The folded geometry of 1 directed the Grignard addition, to give, after protection, the ester 7.
Homologation to the allylic alcohol 8 set the stage for Grubbs ring closure, to give, after oxidation,
the tricyclic sulfone 9, having the skeleton of triclavulone 3.
There were two more stereocenters to set. It was expected that cuprates would add to the open face
of the strained cyclobutene. The control of the other stereocenter was more problematic. One
solution was to prepare an α-sulfonyl lactone. To this end, the ketone was converted to the
secondary carbonate. As hoped, conjugate addition was followed by intramolecular acylation, but
the reaction continued to full acyl transfer, to give 10. Fortunately, desilylation of 10 proceeded
with concomitant lactonization. Desulfonylation then gave 2, which could be brought to high ee
by recrystallization.
With 2 in hand, the rest of the synthesis proceeded smoothly. Reprotection followed by reduction
and oxidation gave the keto aldehyde 11. Condensation of the keto phosphonate 12 with 11 gave
the enone 13. Enantioselective transfer hydrogenation of 13 gave the allylic alcohol 14 with 11:1
diastereoselectivity. Protecting group interchange then gave triclavulone 3, identical in every
respect with natural material.

76. Synthesis of Amphidinolide T1
The amphidinolides are a class of structurally diverse and physiologically potent natural products.
The key step in the total synthesis of enantiomerically-pure amphidinolide T1 3 recently reported
(J. Am. Chem. Soc. 2004, 126, 998. ) by Timothy Jamison of MIT, the Ni-mediated cyclization of
1 to 2, clearly illustrates the power of organometallic C-C bond formation in organic synthesis.
The alcohol fragment of 1 was prepared by Evans alkylation of 4 to give, after reduction and
protection, the alkyne 5. Ni-mediated coupling of the alkyne 5 with the enantiomerically-pure
epoxide 6, following the procedures developed by the Jamison group, led to the alcohol 7 with
high regio- and geometric control.
The acid portion of 1 was assembled by enantio- and diastereocontrolled addition of Z-crotyl
borane to the aldehyde 8, following the Brown protocol. Hydroboration and oxidation led to 9,
which was condensed with the allenyl silane 10 to give 11 with high diastereocontrol. Conversion
of the alcohol to the iodide followed by three-carbon homologation by the Myers procedure then
led to 1, which was cyclized with > 10:1 regio- and diasterocontrol to give 12. Ozonolysis and
methylenation of the less hindered ketone then delivered 3.
In both of the Ni-mediated steps in this synthesis, the Ni-alkyne complex is acting as an acyl anion,
in one case opening an epoxide and in the other case adding to the aldehyde in an intramolecular
sense. Such Ni-reduced phenylalkynes are among the easiest to prepare and least expensive of
acyl anion equivalents

77. Catalytic Asymmetric Synthesis of Quinine and Quinidine
The tetracyclic alkaloid quinine 1 and the diastereomeric alkaloid quinidine 2 share a storied
history. Eric Jacobsen of Harvard recently completed (J. Am. Chem. Soc. 2004, 126, 706.)
syntheses of enantiomerically-pure 1 and of 2. For each synthesis, the key reaction for establishing
the asymmetry of the target molecule was the enantioselective conjugate addition developed by
the Jacobsen group.
For both 1 and 2, the synthesis started with the alkenyl amide 3. Salen-mediated conjugate
addition proceeded with remarkable induction, to give 5 in 92% ee as a mixture of diastereomers.
Reduction and cyclization followed by deprotonation and kinetic quench delivered the
enantiomerically-enriched cis dialkyl piperidine 6. Homologation of the two sidechains then gave
the alkenyl boronic ester 8.
The quinoline portion of the target alkaloids was prepared by condensing p-anisidine 9 with ethyl
propiolate, followed by bromination. Coupling of 10 with the boronic ester 8 proceeded to give 11,
the intermediate for the synthesis of both 1 and 2. Selective direct epoxidation of 11 using the
usual reagents failed, but Sharpless asymmetric dihydroxylation was successful, providing the diol
in > 96:4 diastereoselectivity, with only traces of the tetraol and of the product from
dihydroxylation of the terminal vinyl group. The diol could be converted cleanly to the desired
epoxide 11. Deprotection followed by cyclization led to quinine 1. Preparation of the
diastereomeric epoxide, using AD-mix-α, followed by cyclization gave quinidine 2. The brevity of
the preparation of these two classical alkaloids is a testament to the power of reagent-controlled
synthesis.

78. Synthesis of Deacetoxyalcyonin Acetate
Deacetoxyalcyonin acetate 1 and euncellin 2 are representative members of the eunicellin class of
diterpenes. The synthesis of deacetoxyalcyonin acetate 2 by Gary Molander of the University of
Pennsylvania (J. Am. Chem. Soc. 2004, 126, 1642.) illustrates the power of intramolecular
organometallic carbonyl addition for ring construction.
The six-membered ring of 1 was commerically available in enantiomerically-pure form as
α-phellandrene 3. The challenge was to stitch the highly-substituted ten-membered ring of 1 onto
the disubstituted alkene of 3. The strategy that was conceived was to first construct the
seven-membered ring of 8, then effect three-carbon ring expansion to give 11.
The plan for seven-membered ring construction was to effect stepwise 4 + 3 cycloaddition of 6 to
the protected dialdehyde 5. The preparation of 5 began began with 2+2 cycloaddition between 3
and methoxy ketene, to give 4 with high regio- and diastereocontrol. Photochemical cleavage then
gave 5. The acid-mediated 4 + 3 proceeded via initial addition to the less congested ionized
aldehyde, to give the β-keto ester 7. Alkylation of the dianion of 7 followed by ester hydrolysis
and selenation /oxidation established the enone 8.

Three-carbon ring expansion was carried out in two stages. First, two-carbon homologation of the
exo methylene ketone 8 followed by trapping of the intermediate enolate as the triflate led to 9.
Nozaki-Hiyama-Kishi coupling followed by acetylation smoothly converted 9 into 10.
The trisubstituted alkene of 10 was more readily oxidized than was the congested tetrasubstituted
alkene, so the more reactive alkene was temporarily epoxidized. After ozonolysis, the epoxide was
reduced off using the Sharpless protocol. It is a tribute to the specificity of this reagent that the
easily-reduced α-acetoxy ketone is not affected. Selective silylation of the more accessible ketone
followed by methylenation, hydrolysis and addition of methyl lithium to the outside face of the
previously protected carbonyl then delivered 1.
79. Synthesis of (-)-Podophyllotoxin
(-)-Podophyllotoxin 1 and its derivative etoposide 2, derived from natural sources, are in current
clinical use. Michael Sherburn of Australian National University reports (J. Am. Chem. Soc. 2003,
125, 12108. DOI: 10.1021/ja0376588) complementary total syntheses of 1 and of its enantiomer.
The key step in each of these syntheses is a spectacular intramolecular alkene arylation,
exemplified by the conversion of 4 to 5.

The absolute configuration of 1 was set by conjugate addition to the oxazoline 3 followed by
trapping of the product anion, following the precedent of Meyers. Reduction of the oxazoline to
the alcohol followed by thionocarbonate formation then set the stage for the key aryl transfer
reaction.
The aryl transfer is thought to be initiated by addition of the silyl radical to the thiocarbonyl. The
radical so formed adds to the alkene to generate the benzylic radical 7. This radical adds to the
arene to give 8, which fragments to 5.
Oxidation of the silyl group of 5 to the ketone followed by Pd-mediated decarboxylation led to the
ketone 9. The conversion of 9 to podophyllotoxin 1 followed the literature precedent.

80. Synthesis of (-)-Strychnine
The total synthesis of (-)-strychnine 3 reported (J. Am. Chem. Soc. 2003, 125, 9801.) by Miwako
Mori of Hokkaido University is a tour de force of selective organopalladium couplings.
The absolute configuration of the final product was established at the outset, by Pd-catalyzed
de-racemizing coupling of 4 with the o-bromoaniline derivative 5. Using the inexpensive
Binol-derived ligand (S)-BINAPO 6 , a model coupling was carried out on several cyclohexenol
derivatives having different one-carbon substituents at C-2. The best ee's were observed with the
silyloxymethyl group. Several alcohol derivatives were then tried, and it was found that the allylic
phosphate gave the best rates and ee's. Using the optimized 4, the coupling with 5 to give 7
proceeded in 84% ee.
The sidechain of 7 was extended by one carbon, to give the nitrile 8. A second organopalladium
step then was used to cyclize 8 to 2. Using Ag 2 CO 3 as the base suppressed unwanted alkene
migration. The reaction ran more slowly in DMSO than in DMF, but byproduct formation was
suppressed.
The crystalline nitrile 2 (99% ee from EtOH) was reduced and protected to give the carbamate 9,
setting the stage for another Pd-catalyzed ring-forming step. Allylic oxidation of 9 gave the enone
only in unacceptably low yield. Pd-mediated cyclization, by contrast, proceeded efficiently to give
the alkene 10. Hydroboration followed by oxidation then gave the ketone 11, a useful intermediate
for the construction of a variety of Strychnos alkaloids.

For strychnine 3, the ketone 11 was converted to the alkene 12 by reduction of the enol triflate
derived from the more stable enolate. Deprotection and acylation gave 13, which was cyclized
with Pd to give, after equilibration, the diene 14. Alkylation, to give 15, followed by Pd-mediated
cyclization then gave 16, which was reduced and cyclized to (-)-strychnine 3.
81. Enantioselective Total Synthesis of (+)-Amphidinolide T1
Amphdinolide T1 1 is representative of a family of macrolides, isolated from the Amphidinium
marine dinoflagellates, that show significant antitumor properties. Arun Ghosh of the University
of Illinois at Chicago recently completed (J. Am. Chem. Soc. 2003, 125, 2374.) a total synthesis of
1, based conceptually on the convergent coupling of the enantiomerically-pure fragments 2 and 3.
For each of the two fragments, a key component was assembled by the syn selective aldol
condensation developed by Ghosh. For 2, addition to 3-benzyloxypropionaldehyde gave 4, which
was carried on to the protected lactol 6. Homologation to 7 allowed Grubbs coupling with the
fragment 8, leading to 9. Activation of the lactol by condensation with benzenesulfinic acid then
gave 2.

The enantiomerically-pure aldehyde 14was prepared by adding dithiane to the
commercially-available glycidyl tosylate 10. For the other half of 3, another syn-selective aldol
condensation gave 12, which was carried on to the iodide 13. Reduction with t-butyl lithium,
addition of the resulting organolithium to 11 and oxidation then gave the coupled ketone, which
was homologated using the Petasis procedure to give 14.
In fact, the sensitive disubstituted alkene of 14 turned out to not be stable to the subsequent AlCl 3
coupling conditions, so the alkene and the secondary alcohol were protected together as the
bromoether 15. Condensation of the derived enol ether 16 with the sulfone 2 in the presence of
DTBMP (2,6-di-t-butyl-4-methylpyridine) then gave 17. Yamaguchi lactonization followed by
regeneration of the alkene by zinc reduction completed the synthesis of 1.
82. Synthesis of (+)-4,5-Deoxyneodolabelline
The dolabellanes, represented by 3-hydroxydolabella-4(16), 7, 11(12)-triene-3,13-dione 1 and the
neodolabellanes, represented by (+)-4,5-deoxyneodolabelline 2, are isolated from both terrestrial
and marine sources. They show cytotoxic, antibiotic and antiviral activity. The recent synthesis of
(+)-4,5-deoxyneodolabelline 2 by David Williams of Indiana University (J. Am. Chem. Soc., 2003,
125, 1843.) highlights both the strengths and the challenges of the current state of the art in
asymmetric synthesis.

The synthetic plan was to assemble both the dihydropyran 3 and the cyclopentane 4 in
enantiomerically-pure form, then to effect Lewis acid-mediated coupling of the allyl silane of 4
with the anomeric ether of 3 to form a new stereogenic center on the heterocyclic ring. A critical
question was not just the efficiency of this step, but whether or not the desired stereocontrol could
be achieved at C-3.
The construction of the heterocycle 3 started with enantiomerically-pure ethyl lactate. Protection,
reduction and oxidation led to the known aldehyde 6. Chelation-controlled allylation gave the
monoprotected-diol 7. Formation of the mixed acetal with methacrolein followed by
intramolecular Grubbs condensation then gave 3. The dihydropyran 3 so prepared was a 1:1
mixture at the anomeric center.
The preparation of the cyclopentane 4 proved to be more of a challenge. Rather than attempt an
enantioselective synthesis, racemic 11 was prepared in straightforward fashion from
commercially-available 2-methylcyclopentenone, by conjugate addition followed by alkylation of
the regenerated ketone enolate. Ozonolysis followed by selective reduction then led to 11.
Resolution was accomplished by enantioselective reduction of the racemic ketone, to give a 1:1
mixture of separable diastereomers. Reoxidation of one of the diastereromers gave ketone 11,
which was determined to be a 96:4 mixture of enantiomers. Homologation followed by allylic
silylation then gave 4 as an inconsequential mixture of diastereomers.
Condensation of the allyl silane 4 with 3 proceeded to give exclusively the desired trans
dihydropyran 5. McMurry coupling of the derived keto aldehyde gave the diol 13 as a mixture of
diastereomers. Oxidation of the mixture gave 2 and its C-8 diastereomer in a ratio of 8:1.

83. Synthesis of (+)-Phomactin A
The diterpene (+)-Phomactin A 4 is an antagonist of platelet activating factor. The preparation of 4
recently reported (J. Am. Chem. Soc. 2003, 125, 1712.) by Randall Halcomb of the University of
Colorado elegantly illustrates the use of readily-available natural products as starting materials for
natural product synthesis.
The synthetic plan called for a late-stage intramolecular reductive coupling of the iododiene 3 to
establish the macrocyclic ring of 4. The iododiene 3 was to be assembled by condensation of the
highly-substituted cyclohexene 1 with the aldehyde 2.
The aldehyde 2 was prepared from the inexpensive geraniol ether 5. Selective ozonolysis followed
by Wittig homologation gave the bromodiene, which was converted via dehydrobromination and
alkylation to the alkyne 6. Regioselective hydridozirconation followed by iodination of the C-Zr
bond gave the alkenyl iodide 7 with high geometric control. The two stereogenic centers of 2 were
then established by Sharpless asymmetric epoxidation.
The preparation of the cyclohexene 1 began with pulegone 8, available commercially in high
enantiomeric purity. Methylation followed by retro aldol condensation to remove the unwanted
isopropylidene group gave 2,3-dimethylcyclohexanone, which on
bromination-dehydrobromination gave 9. Vinylation followed by alkylative enone transposition

gave 11, which was brominated over several steps to give 12. Conditions to reduce the ketone 11
directly to the axial alcohol were unavailing, so the dominant pseudoequatorial alcohol from
NaBH 4 reduction was inverted, to give 1.
Condensation of 1 with 2 led to 3, setting the stage for the key macro ring closure. Happily,
conditions could be developed to effect this important transformation, a B-alkyl Suzuki coupling.
The ligand dppf is 1,1'-bisdiphenylphosphinoferrocene. The use of AsPh 3 , rather than a phosphine,
as the supporting ligand was important, as was the use of the thallium base.
84. Synthesis of the Mesotricyclic Diterpenoids Jatrophatrione and Citlalitrione
Leo Paquette of Ohio State recently reported (J. Am. Chem. Soc. 2003, 125, 1567.) the total
synthesis of jatrophatrione 1 and citlalitrione 2. These diterpenes, which share a central
highly-subsituted 5-9-5 core, show remarkable tumor-inhibitory activity.
The 5-9-5 skeleton was assembled by the addition of the alkenyl cerate derived from 6 with the
ketone 4, to give 7. Oxy-Cope rearrangement then gave the 5-9-5 enolate, which was quenched
with methyl iodide to give 8. The ketone 8 underwent spontaneous intramolecular ene cyclization,
to give 9.

The transient 5-9-5 ketone 8 has two cis-fused rings. To invert the ring stereochemistry, the alkene
9 was oxidized to the enone 10. After some experimentation, it was found that a CuH preparation
would reduce the enone to give predominantly the trans-fused ketone. Monomesylation of the
derived diol set the stage for Grob fragmentation to reopen the nine-membered ring, providing,
after reduction, the alcohol 12.
At this point, there were two problems in selective alkene functionalization to be addressed.
Although all attempts at oxidation of the cyclopentene failed, intramolecular hydrosilylation
proceeded smoothly, to give 13. On exposure of the derived cyclic carbonate to ,
the cyclononene then underwent allylic oxidation, to give 14.
Attempts to functionalize the homoallylic alcohol 15 quickly revealed that this product of an
intramolecular aldol condensation was sensitive to base. Fortunately, heating with
thiocarbonyldiimidazole effected clean dehydration to give predominantly the desired regioisomer
of the diene. Methanolysis followed by oxidation then gave the triketone 1, which on epoxidation
with MCPBA gave 2 as the minor component of a 3:1 mixture.
Further Information:
J. Yang, Ph.D. Thesis, Ohio State University, 2003.

85. Total Synthesis of Ingenol
The total synthesis of the tetracyclic Euphorbia tetraol ingenol 3 reported by Keiji Tanino of
Hokkaido University (J. Am. Chem. Soc. 2003, 125, 1498.) illustrates the power of
diastereoselective carbocationic rearrangements, as exemplified by the conversion of 1 to 2.
The construction of the tricyclic epoxide depended on several highly diastereoselective
transformations. The addition of lithio t-butyl acetate to ketone 4 proceeded to give 5 as a single
diastereomer, even though the ketone is flanked by a quaternary center. The authors speculate that
lithium chelation with the methyl ether directed addition. Even more spectacular was the
cyclization of the propargylic acetate 7 to 10. The Co complex activated the acetate for ionization,
while at the same time establishing the proper geometric relationship for bond formation.
Dissolving metal reduction of the Co complex then gave the alkene.
The elegant pinacol rearrangement of 1 to 2, mediated by (ArO) 2 AlCH 3 , exposed a ketone that
might usually need to be protected. In this case, however, the ketone is so buried in the
inside-outside ingenol skeleton that it is unreactive. After several further manipulations, a
spectacular osmylation of the diene 12 led to ingenol 3, in an overall 45-step sequence.
The ingenol 3 prepared by this route was racemic. It is interesting to speculate how one might
efficiently prepare 4 or its precursors in enantiomerically-pure form.

86. Total Synthesis of the Galbulimima Alkaloid GB 13
Lew Mander of the Australian National University recently reported (J. Am. Chem. Soc. 2003, 125,
2400. ) the total synthesis of the pentacyclic alkaloid GB 13 3, which had been isolated from the
bark of the rain forest tree Galbulimima belgraveana. In the course of the synthesis, he took full
advantage of benzene precursors, while at the same time carefully establishing each of the eight
stereogenic centers of 3.
The core tricyclic ketone 1 was assembled by Birch reduction of 2,5-dimethoxybenzoic acid,
followed by alkylation with 3-methoxybenzyl bromide, to give 4. Acid-catalyzed electrophilic
cyclization of 4 gave the tricyclic ketone 5, which on decarboxylation and protection gave 1.
Diazo transfer to 1 followed by irradiation in the presence of bis-(trimethylsilyl)amide led to ring
contraction with concomitant carbonyl extrusion, to give 7. Dehydration to the nitrile followed by
selenation then set the stage for a highly diastereoselective ytterbium-catalyzed Diels-Alder
reaction, to give, after reduction and protection, the pentacyclic intermediate 2.
Intermediate 2 appears to have two extraneous carbons, the red nitrile and the blue carbon in the
aromatic ring. In fact, the blue carbon was carried all the way through, to appear as the α-methyl
group on the piperidine ring. Birch reduction of 2 deleted the now-superfluous nitrile, and reduced
the aromatic ring, to give, after hydrolysis, the enone 10. Eschenmoser fragmentation of the
intermediate epoxy ketone then gave the keto alkyne 11. The subsequent condensation with
hydroxylamine followed by reduction proceeded with spectacular (but anticipated) stereocontrol,
to establish the three stereogenic centers of the trisubstituted piperidine ring. Oxidation of 12 then
gave the enone 3.