Concise syntheses of the natural products (+)-sylvaticin and (+)-cis-sylvaticin

Article abstract of DOI:10.1021/ja9049959

Two concise syntheses of the natural products cis-sylvaticin and sylvaticin are reported, using oxidative cyclization methodology as the key step. A sequential solvolysis/hydride shift/intramolecular reduction cascade was used to establish the trans stereochemistry of one of the THF rings of sylvaticin.

Full text of DOI:10.1021/ja9049959

Published on Web 08/12/2009
Concise Syntheses of the Natural Products (+)-Sylvaticin and
(+)-cis-Sylvaticin
Timothy J. Donohoe,*,§ Robert M. Harris,§ Oliver Williams,§ Gra´inne C. Hargaden,§
Jeremy Burrows,† and Jeremy Parker‡
AstraZeneca Pharmaceuticals, Department of Medicinal Chemistry R&D, So¨derta¨lje,
S-151 85, Sweden, AstraZeneca, Process R&D AVlon/Charnwood, AVlon Works, SeVern Road,
Hallen, Bristol, BS10 7ZE, U.K., and Department of Chemistry, UniVersity of Oxford, Chemistry
Research Laboratory, Mansfield Road, Oxford, OX1 3TA, U.K.
Received June 18, 2009; E-mail: timothy.donohoe@chem.ox.ac.uk
Abstract: Two concise syntheses of the natural products cis-sylvaticin and sylvaticin are reported, using
oxidative cyclization methodology as the key step. A sequential solvolysis/hydride shift/intramolecular
reduction cascade was used to establish the trans stereochemistry of one of the THF rings of sylvaticin.
Scheme 1
Introduction
Sylvaticin and its C-12 epimer cis-sylvaticin belong to the
nonadjacent bis-THF subclass of the annonaceous acetogenins.
Sylvaticin was isolated by McLaughlin in 1990 from the dried
fruits of Rollinia sylVatica St. Hil. (Annonaceae).1 Subsequent
isolations were reported in 1993 from Annona purpurea and in
1995 when it was isolated from leaf extracts of Rollinia mucosa
(Jacq.) Baill. together with cis-sylvaticin, bullatalicin, and
muricatetrocin B.2,3 Both cis-sylvaticin and the parent sylvaticin
display potent activity as antitumor agents and exhibit nanomolar
cytotoxicity toward certain human solid tumor cell lines (human
lung carcinoma and human pancreas carcinoma). Their mode
of action is thought to include the inhibition of ATP production
Via the blockage of mitochondrial complex I. The biosynthesis
of these acetogenins is thought to start from a long chain fatty
acid with the terminal γ-lactone functionality initially formed
with the nonadjacent bis-THF core then generated by oxidation
of the unsaturated units present followed by a series of ring-
opening and closing reactions.4
Scheme 2
Both natural products 1 and 2 contain two nonadjacent THF
rings, with sylvaticin bearing the only trans-ring. One THF is
flanked on either side by a hydroxyl group and an alkyl chain,
and the second is flanked by one hydroxyl group and an alkyl
chain ending in a γ-lactone functionality with a hydroxyl group
at the C-4 position (see Scheme 3 for numbering). At the outset
of this work, the structures of 1 and 2 had not been verified by
total synthesis, although the group of McLaughlin had performed
an extremely thorough analysis of both compounds in the
original isolation paper.3 McLaughlin had also shown the precise
positions of the two THF rings (and the hydroxyl groups) by
EIMS of the natural products and their TMS derivatives. The
relative and absolute stereochemistry was determined by various
NMR techniques developed for these acetogenins and included
a full analysis of their di- and tetra-Mosher ester derivatives,
and the structures were finally assigned as shown in Scheme 1.
Previous work in our group culminated in the first synthesis
of (+)-cis-sylvaticin in 2006;5 with Brown and co-workers
subsequently reporting a synthesis of this natural product in
† AstraZeneca Pharmaceuticals.
‡ AstraZeneca, Process R&D.
§ University of Oxford.
(1) Mikolajczak, K. J.; Madrigal, R. V.; Rupprecht, J. K.; Hui, Y. H.;
Liu, Y. M.; Smith, D. L.; McLaughlin, J. L. Experentia 1990, 46,
324.
(2) Felicitas Cepleanu, K. O.; Hamburger, M.; Hostettmann, K.; Gupta,
M. P.; Solis, P. HelV. Chim. Acta 1993, 76, 1379.
(3) Shi, G.; Zeng, L.; Gu, Z.; MacDougal, J. M.; McLaughlin, J. L.
Heterocycles 1995, 41, 1785.
(4) For a review see (a) Bermejo, A.; Figade´re, B.; Zafra-Polo, M.-C.;
Barrachina, I.; Estornell, E.; Cortes, D. Nat. Prod. Rep. 2005, 22, 269.
(b) Alali, F. Q.; Liu, X. X.; McLaughlin, J. L. J. Nat. Prod. 1992, 62,
504. (c) Rupprecht, J. K.; Hui, Y. H.; McLaughlin, J. L. J. Nat. Prod.
1990, 53, 237. (d) Saez, J.; Sahpaz, S.; Villaescusa, L.; Hocquemiller,
R.; Cave´, A.; Cortes, D. J. Nat. Prod. 1993, 56, 351. (e) Zafra-Polo,
M. C.; Figadere, B.; Gallardo, T.; Tormo, J. R.; Cortes, D. Phytochem-
istry 1998, 48, 1087. (f) Zeng, L.; Ye, Q.; Oberlies, N. H.; Shi, G.;
Gu, Z. M.; He, K.; McLaughlin, J. L. Nat. Prod. Rep. 1996, 13, 275.
(5) Donohoe, T. J.; Harris, R. M.; Burrows, J.; Parker, J. J. Am. Chem.
Soc. 2006, 128, 13704.
9
12854 J. AM. CHEM. SOC. 2009, 131, 12854–12861
10.1021/ja9049959 CCC: $40.75  2009 American Chemical Society

Syntheses of the Natural Products (+)-Sylvaticin and (+)-cis-Sylvaticin
A R T I C L E S
Scheme 3
2008.6 In both cases, the data from the synthetic sample exactly
matched those of the natural product, thus confirming that the
original assignment was correct. In addition, there are other key
syntheses of related nonadjacent bis-THF annonaceous aceto-
genins reported in the literature.7 Herein, we describe a full
account of our studies which led to the preparation of (+)-cis-
sylvaticin 1, and we also report the first total synthesis of the
parent natural product sylvaticin 2.
Scheme 3). While this approach would be ideal for the
construction of cis-sylvaticin directly, we were also intrigued
by the possibility of using our recently developed hydride-shift/
intramolecular reduction methodology (see DfC-2, Scheme 3)
to construct the trans-THF ring of sylvaticin from a cis-THF
precursor10 and therefore enable the completion of both natural
product syntheses Via a single strategy.
In both cases, we anticipated using the reliable cross-
metathesis (CM) chemistry11 to attach the butenolide portion
B to the bis-THF unit, as had been previously described by
Lee,12 which would allow completion of the syntheses. Analysis
of likely synthetic routes meant that we were required to
disconnect the chain linking the THF rings to the butenolides
in different places during approaches to sylvaticin and cis-
sylvaticin. However, it was decided that an E,E-tetradecatatraene
Results and Discussion
Our strategy for forming the THF rings of these natural
products relied upon the intramolecular oxidative cyclization
reaction of vicinal diols onto alkenes to form cis-THF rings,
using osmium(VI) as a catalyst and an acid as the promoter.8
This methodology has proven to be an extremely reliable route
to THFs, proceeding with complete stereospecificity for syn-
addition across the alkene and also high stereoselectivity for
the formation of cis-2,5-THFs, Scheme 2. Preparation of the
starting diols as single enantiomers (and diastereoisomers) is
readily achieved by a Sharpless AD reaction9 of the corre-
sponding diene. Following this AD sequence with an oxidative
cyclization allows the preparation of cis-THFs with a high
degree of control over all aspects of stereochemistry.
The three rings and nine stereocenters of cis-sylvaticin 1 and
sylvaticin 2 pose an interesting synthetic challenge with the core
bis-THFs of particular interest. It was proposed that this key
segment C of both natural products could be synthesized in a
very concise manner using a novel bis-oxidative cyclization
reaction to introduce the two cis-THF rings (see EfC-1,
(7) (a) Makabe, H.; Tanaka, A.; Oritani, T. Tetrahedron Lett. 1997, 38,
4247. (b) Marshall, J. A.; Jiang, H. J. Org. Chem. 1998, 63, 7066. (c)
Makabe, H.; Tanaka, A.; Oritani, T. Tetrahedron 1998, 54, 6329. (d)
Zhu, L.; Mootoo, D. R. Org. Lett. 2003, 5, 3475. (e) Zhu, L.; Mootoo,
D. R. J. Org. Chem. 2004, 69, 3154. (f) Crimmins, M. T.; She, J.
J. Am. Chem. Soc. 2004, 126, 12790. (g) Hoye, T. R.; Eklov, B. M.;
Jeon, J.; Khoroosi, M. Org. Lett. 2006, 8, 3383.
(8) (a) Donohoe, T. J.; Butterworth, S. Angew. Chem., Int. Ed. 2003, 42,
948. (b) Donohoe, T. J.; Butterworth, S. Angew. Chem., Int. Ed. 2005,
44, 4766. (c) Donohoe, T. J.; Churchill, G. H.; Wheelhouse (nee´
Gosby), K. M. P.; Glossop, P. A. Angew. Chem., Int. Ed 2006, 45,
8025. (d) Donohoe, T. J.; Wheelhouse (nee´ Gosby), K. M. P.; Lindsay-
Scott, P. J.; Glossop, P. A.; Nash, I. A.; Parker, J. S. Angew. Chem.,
Int. Ed. 2008, 47, 2972.
(9) Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem. ReV.
1994, 94, 2483.
(10) Donohoe, T. J.; Williams, O.; Churchill, G. H. Angew. Chem., Int.
Ed. 2008, 47, 2869.
(6) Brown, L. J.; Spurr, I. B.; Kemp, S. C.; Camp, N. P.; Gibson, K. R.;
(11) Chatterjee, A. K.; Choi, T. L.; Sanders, D. P.; Grubbs, R. H. J. Am.
Chem. Soc. 2003, 125, 11360.
Brown, R. C. D. Org. Lett. 2008, 10, 2489.
9
J. AM. CHEM. SOC. VOL. 131, NO. 35, 2009 12855

A R T I C L E S
Donohoe et al.
Scheme 4
(E,E-4) would serve as a useful starting material for both natural
products syntheses. We would seek to perform regioselective
AD reactions on the two internal alkene units of 4 before
desymmetrizing the system by distinguishing between the two
ends of the molecule.
Asymmetric Dihydroxylation of Tetraenes. The first step in
the synthesis requires the asymmetric dihydroxylation of E,E-
tetradecatetraene 4. 1,5,9,13-Tetradecatetraene is commercially
available but only as an approximately 1:1:1 mixture of the three
possible geometric isomers, Scheme 4. Fortunately the isomers
could be separated by chromatography on silica gel doped with
AgNO3 to yield pure E,E-1,5,9,13-tetradecatetraene, E,E-4. The
isolated tetradecatetraene was confirmed as the E,E-isomer by
the synthesis of authentic E,E-1,5,9,13-tetradecatetraene from
E,E,E-cyclododecatriene 5 in three steps. Both sets of materials
and a mixture of the two had identical NMR data.
Asymmetric dihydroxylation of E,E-4 using (DHQ)2PHAL
as the chiral ligand led to the formation of desired tetraol 8 in
37% yield, >98% ee, >95:5 dr (Scheme 4). The ee and dr were
ascertained by tetra-Mosher ester formation from authentic
standards of both the enantiomer of 8 (made using DHQD
ligand) and the meso-tetraol (made by sequential mono-AD
using DHQD and then with DHQ). The absolute stereochemistry
of tetraol 8 was assigned by using the Sharpless mnemonic.9
Finally, an X-ray crystal structure of tetraol 8 authenticated
its relative stereochemistry, Scheme 4.13 The tetraol 8 was
readily transformed into bis-acetonide 9 by reaction with vinyl-
methyl ether and CSA in DMF.
The outcome of the AD reaction has its origins in the
preferential oxidation of 1,2-trans alkenes over monosubstituted
alkenes. Based upon Sharpless’s rate data, a selectivity of
approximately 7:1 can be expected.12 Calculations can predict
the maximum yield for the dihydroxylation of the two trans
alkenes in E,E-4, assuming this 7:1 rate difference, and also
that the second oxidation proceeded with the same regioselec-
tivity as that of the first. The calculation showed a maximum
theoretical yield of 42% for the formation of 8 which is
close to the 37% achieved in practice.
After the success of the selective double dihydroxylation, the
commercially available mixture of tetradecatetraenes was
subjected to the AD conditions, in the knowledge that cis-
alkenes are dihydroxylated approximately 11 times slower than
trans alkenes.14 In this case the desired tetraol 8 was again
isolated although the compound could not be fully purified until
it was derivatized as bis-acetonide 9 (19% yield over the two
steps, starting with 25 g of tetraene) which was identical to that
prepared from the pure E,E-isomer of 4. In practice, this route
(12) (a) Keum, G.; Hwang, C. H.; Kang, S. B.; Kim, Y.; Lee, E. J. Am.
Chem. Soc. 2005, 127, 10396. (b) White, J. D.; Somers, T. C.; Reddy,
G. N. J. Org. Chem. 1992, 57, 4991. (c) Schaus, S. E.; Brånalt, J.;
Jacobsen, E. N. J. Org. Chem. 1998, 63, 4876. (d) Avediaaian, H.;
Sinha, S. C.; Yazbak, A.; Sinha, A.; Neogi, P.; Sinha, S. C.; Keinan,
E. J. Org. Chem. 2000, 65, 6035.
(13) Diffraction data were collected at 150 K using a Nonius Kappa-CCD
area detector diffractometer (λ ) 0.710 73 Å). Cell parameters and
intensity data were processed using the DENZO-SMN package
(Otwinowski, Z.; Minor, W. Processing of X-ray Diffraction Data
Collected in Oscillation Mode. In Methods in Enzymology; Carter,
C. W., Sweet, R. M., Eds.; Academic Press: 1997; p 276), and
reflection intensities were corrected for absorption effects by the
multiscan method. The structures were solved by direct methods
(Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A.; Burla,
M. C.; Polidori, G.; Camalli, M. J. Appl. Crystallogr. 1994, 27, 435.
Sheldrick, G. M. Acta Crystallogr. 2008, A64, 112–122) and refined
by full-matrix least squares on F using the CRYSTALS suite
(Betteridge, P. W.; Carruthers, J. R.; Cooper, R. I.; Prout, K.; Watkin,
D. J. J. Appl. Crystallogr. 2003, 36, 1487. Tetrgonal, P41212; a )
19.4191(2) Å, c ) 16.2612(2) Å, V ) 6132.12(12) Å3; Z ) 16; Data/
restraints/parameters ) 3090/108/413 (Rint ) 0.043); R1 ) 0.0365,
wR2 ) 0.0468 [I > 3σ(I)]. Full crystallographic results can be found
in the CIF (Supporting Information). CCDC 736594.
(14) Andersson, P. G.; Sharpless, K. B. J. Am. Chem. Soc. 1993, 115, 7047.
9
12856 J. AM. CHEM. SOC. VOL. 131, NO. 35, 2009

Syntheses of the Natural Products (+)-Sylvaticin and (+)-cis-Sylvaticin
A R T I C L E S
Scheme 5
Scheme 6
Hz between the two alkene protons; this confirms the cis
geometry of the newly formed alkene within 12.
The key reaction in this synthesis of cis-sylvaticin is the
formation of two THF rings and three stereogenic centers using
a novel double oxidative cyclization reaction to access the bis-
THF core of the natural product. Pleasingly, initial attempts at
the oxidative cyclization of 12 afforded bis-THF 13 in a
reasonable yield of 60%. The loss of alkene peaks between 5.84
1
and 4.99 ppm in the H NMR showed consumption of the
starting material 12, and the peaks at 83.1, 82.9, 82.1, and 80.0
ppm in the 13C NMR confirmed the formation of two THF rings
1
in the product. However, careful inspection of the H and 13C
NMR data showed the presence of a byproduct of cyclization
which proved very difficult to separate from the desired bis-
THF 13. Mass spectral evidence suggested the byproduct to be
an overoxidized form of 13, and the presence of a carbonyl
was confirmed by a peak at 213 ppm in the 13C NMR and an
absorption at 1716 cm-1 in the IR spectrum. We have assigned
the structure of the undesired product as a ketone derived from
overoxidation of product 13 at C-24. Extensive optimization of
the oxidative cyclization conditions was then attempted, con-
centrating on promoting oxidative cyclization at the expense
of overoxidation. We found that reducing the amount of TFA
maximized the tetraol/ketone ratio (40:1) with the optimum
conditions involving addition of the osmium in three batches
over 48 h. Using this protocol, a yield of 77% of compound 13
could be obtained.
To prepare the bis-THF core for the Wittig reaction, oxidation
of the primary alcohol to the aldehyde was required. Unfortu-
nately, attempts to carry out this transformation directly from
13 failed. However, this problem could be overcome by simple
protecting group manipulation; tetra-TBS protection of 13
followed by monodeprotection of the least sterically hindered
primary TBS group with TBAF led to the formation of alcohol
14 in 59% yield over these two steps, Scheme 5. Oxidation of
14 with TPAP proceeded smoothly to give aldehyde 15 in 86%
yield.15 We then followed literature precedent from Lee12 who
used cross metathesis (CM) as an efficient tactic for joining
related compounds in the synthesis of Annonaceous acetogenins.
In preparation, aldehyde 15 was methylenated under Wittig
conditions to afford alkene 16.
starting from the mixture of geometric isomers was used to
provide material for the ensuing syntheses.
Desymmetrization and Oxidative Cyclization. The C2 sym-
metrical bis-acetonide 9 was then desymmetrized by oxidation
of one of its two alkene units; this was achieved by subjecting
9 to dihydroxylation conditions. However, due to the homotopic
nature of the two alkenes, standard dihydroxylation conditions
led to a statistical mixture of the desired compound 10, starting
material 9, and the doubly dihydroxylated compound 11. It was
found that the yield of 10 could be increased to 60% under
asymmetric dihydroxylation conditions at 0 °C. This increase
in yield is likely to be due to the differential product solubility
in the biphasic solvent system of the AD reaction rather than
by asymmetric induction. The tetraol byproduct 11 could be
recycled to 9 in 65% yield over two steps by oxidative cleavage
of the two diol units to yield a bis-aldehyde followed by bis-
methylenation (Scheme 5).
The oxidative cyclization precursor 12 was then prepared by
oxidative cleavage of 10 to give an aldehyde, and a Wittig
reaction followed using the nonstabilized ylid formed from
deprotonation of commercially available undecyltriphenylphos-
phonium bromide with KHMDS (81% of 12 over two steps).
The presence of four alkene peaks between 138.0 and 114.9
ppm in the 13C NMR confirmed the formation of a second alkene
unit. The complex alkene 2 H multiplet at 5.44-5.31 ppm
corresponding to the disubstituted alkene could be simplified
by performing a homonuclear decoupling experiment, irradiating
the 2 H quartet at 2.04 ppm (CH3(CH2)8CH2CHdCH) and
allowing the determination of a coupling constant, J, of 10.8
(15) (a) Ley, S. V.; Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis
1994, 639. (b) Lenz, R.; Ley, S. V. J. Chem. Soc., Perkin Trans. 1
1997, 3291.
9
J. AM. CHEM. SOC. VOL. 131, NO. 35, 2009 12857

A R T I C L E S
Donohoe et al.
Scheme 7
Desymmetrization and Oxidative Cyclization for the Syn-
thesis of Sylvaticin. The synthesis of sylvaticin also required
desymmetrization of bis-alkene 9 so that the two alkenes could
be modified in sequence; this was achieved from diol 10,
Scheme 6. Instead of periodate cleavage, as before, protection
of the diol as its bis-acetate allowed us to selectively cleave
the remaining alkene using ozonolysis followed by a Wittig
reaction with a nonstabilized ylid, to yield cis-alkene 17 as the
sole alkene isomer from the reaction. We then turned our
attention to the other end of the molecule and proceeded to
cleave the diol derived from hydrolysis of bis-acetate 17 with
basic sodium periodate. The resultant aldehyde was subjected
to olefination using the Kocienski variant of the Julia reaction,16
which gave trans alkene 18 in good yield and with 14:1
Scheme 8
1
selectivity for the E-isomer. A series of H NMR experiments
on compound 18 (and precursors to it) allowed us to assign the
two 3J values across the alkene units as 11 and 15 Hz,
respectively, which confirms the stereochemcial assignment as
shown.
(16) Blakemore, P. R.; Cole, W. J.; Kocienski, P. J.; Morley, A. Synlett
1998, 26.
9
12858 J. AM. CHEM. SOC. VOL. 131, NO. 35, 2009

Syntheses of the Natural Products (+)-Sylvaticin and (+)-cis-Sylvaticin
A R T I C L E S
Scheme 7.18 This tactic left the remaining three hydroxyl groups
at C-24, C-19, and C-16 untouched, and these were derivatized
as a silicon-hydride based silyl ether under standard conditions
(20f21). Next, reaction with mild acid, generated by the action
of iodine in methanol,19 promoted selective removal of the
acetal, while leaving the silicon-based protecting groups intact.
This protecting group manipulation enabled us to activate each
of the newly unmasked alcohols 22 as a mesylate derivative,
23. The fate of each of the two OMs functionalities within 23
is different: the primary OMs group is destined to be used to
chain extend the carbon skeleton so that it can ultimately be
attached to the butenolide fragment by cross metathesis; the
secondary OMs group is intended to intitiate a hydride shift/
directed ionic hydrogenation sequence that is designed to remove
oxygen functionality at C-11 and form a trans 2,5-disubstituted
THF ring at the same time. Preliminary studies clearly showed
that it was easiest to chain extend at the primary center before
attempting hydride shift/directed reduction. Therefore, bis-
mesylate 23 was reacted with allylmagnesium bromide in the
presence of catalytic Cu(I), in diethylether at -20 °C, Scheme
7. The conditions for this transformation had to be carefully
optimized so as to prevent double displacement by the orga-
nocuprate with 24 being isolated in 71% yield. Finally, the
solvolysis of the remaining mesylate within 24 in hexafluoro-
isopropanol (HFIPA) was attempted in the presence of zinc
acetate. Based on our previous methodology, we have evidence
that this reaction sequence is initiated by a hydride shift/
expulsion of the OMs group, to generate oxa-carbenium ion A.
This is subjected to an intramolecular ionic hydrogenation,
whereby the Si-H bond from the appended ROSiHtBu2 group
attacks the oxa-carbenium ion from the lower face as drawn in
an intramolecular fashion.10 Pleasingly, the reaction worked as
planned to generate the trans-THF ring 25 in 88% yield. Despite
extensive experimentation the conditions required to effect this
transformation always gave an amount of desilylated material,
and eventually we opted to extend the reaction time (to 48 h)
so that all three of the silyl “protecting groups” were removed.
As a direct consequence of this, three OTBS groups were then
added at this stage (26) so that the synthesis could be continued
unimpeded.
Scheme 9
Synthesis of the Butenolide Fragment and Completion of
the Syntheses. The butenolide fragment was prepared from two
commercially available building blocks, namely epichlorohydrin
and propylene oxide, following the general route described by
Lee.12 The commercial nature of these two starting materials
means that we can readily prepare any stereochemical isomer
of the CM precursor.
Thus, opening of (S)-(-)-propylene oxide with the dianion
of (phenylthio)acetic acid gave lactone 27 in 73% yield as a
1:1 mixture of enantiopure diastereomers, Scheme 8. Lactone
27 was then deprotonated with LDA and quenched with iodides
28 and 29 (themselves derived from (R)-epichlorohydrin in three
steps) as the electrophile to afford 30 and 31, each as a 3:1
mixture of diastereomers. Alkenes 32 and 34 were then prepared
by oxidation/elimination of the thiophenyl group, and in
accordance with literature precedent, the elimination process
was completely regioselective. For operational reasons, it was
The pivotal double oxidative cyclization was then attempted
on diene 18, using our recently developed conditions which
avoid the use of osmium(VIII) by utilizing pyridine-N-oxide
as a mild reoxidant.6 In addition, we have also found that
previous conditions that required a low pH for cyclization
(usually with the addition of TFA) can be replaced by Lewis
acidic conditions and operate at a much higher pH. These new
mild conditions were essential for the successful cyclization of
18 whereby the pH (ca. 4) was low enough to allow in situ
deprotection of the two acetals and subsequent double oxidative
cyclization to 19, while retaining the PMB group.17
The next sequence of reactions was aimed at installing the
tBu2SiH protecting group on the C-16 hydroxyl so that it can
participate in hydride transfer to an oxo-carbenium ion at a later
stage, Vide infra. To do this, the PMB group was oxidized with
DDQ and the resulting oxo-carbenium ion was captured by the
proximal alcohol to generate the six-membered acetal 20,
(18) (a) Oikaw, Y.; Yoshioka, T.; Yonemitsu, O. Tetrahedron Lett. 1982,
23, 889. (b) Kim, C. Y.; An, H. J.; Shin, W. K.; Yu, W.; Woo, S. K.;
Jung, S. K.; Lee, E. Angew. Chem., Int. Ed. 2006, 45, 8019.
(19) (a) Szarek, W. A.; Zamojski, A.; Tiwari, K. N.; Ison, E. R. Tetrahedron
Lett. 1986, 27, 3827. (b) Wahlstrom, J. L.; Ronald, R. C. J. Org. Chem.
1998, 63, 6021.
(17) Donohoe, T. J.; Winship, P. C. M.; Walter, Daryl, S. J. Org. Chem.
2009, in press (DOI: 10.1021/jo901243y).
9
J. AM. CHEM. SOC. VOL. 131, NO. 35, 2009 12859

A R T I C L E S
Donohoe et al.
Scheme 10
beneficial to remove the TBS group from compound 32, and
this was accomplished with acid in methanol giving 33 in 77%
overall yield from 30.
the procedure reported in the isolation paper. Formation of a
cyclic formaldehyde acetal 37 was accomplished with DMSO
and TMSCl to give a derivative that was also shown to match
the data in the literature. Finally, formation of the di-(R) and
di-(S)-Mosher ester derivatives of compound 37 provided
material with NMR data that matched the original data.
In each case, two fragments 33 and 34 could then be joined
to the bis-THF unit using CM methodology, Scheme 9. As
suggested by Lee, 4 equiv of alkenes 33 and 34 were used in
each metathesis cross-coupling reaction to ensure that alkenes
16 and 26 did not homodimerize (all are terminal alkenes of
type I and therefore prone to homodimerization).11,12 Pleasingly,
the metathesis reaction of 16 with 33 proceeded to give 35 and
homodimerized 33 which could be recycled into the CM reaction
to good effect. The synthesis of (+)-cis-sylvaticin 1 was then
completed in two final steps, which comprised diimide reduction
of the more symmetrical and electron-rich double bond,20
followed by acid deprotection of the three TBS groups.
Likewise, the synthesis of (+)-sylvaticin was accomplished
using a related protocol, Via 36, which furnished the natural
product 2 in good overall yield. This work represents a short
route to these two natural products, and it is noteworthy that
cis-sylvaticin was completed in just 13 steps from commercially
available tetraene 4 and that sylvaticin required 19 steps to
complete.
However, Curran has recently shown that following the
synthesis of a library of stereoisomers of the mono-THF
acetogenin murisolin, a number of the stereoisomers possessed
identical 1H and 13C NMR data to those of the natural product,21
a problem that is typical of the acetogenin family and has also
been observed by Brown during studies on cis-solamin.22 As a
result of these reports, thorough proof that both of our synthetic
compounds were actually the natural products was deemed
necessary.
Of particular concern to us was the relationship between the
stereogenic centers of the butenolide fragment (C-4 and C-36)
and those of the corresponding bis-THF unit. Fortunately, our
synthetic approach enabled the formation of the butenolide
metathesis precursors with any of the four possible configura-
tions. These could then be coupled by CM and the final spectra
compared to those of the natural material.
Verification of the Structures of cis-Sylvaticin and Sylva-
ticin. It was essential that we were able to match the data from
synthetic 1 and 2 with those reported by McLaughlin and so
confirm the structures of these two natural products. Our first
Preliminary results in the cis-sylvaticin series showed that
the C-4,36 bis-epidiastereosiomer of cis-sylvaticin 38 had
identical NMR data to those of cis-sylvaticin itself (although
its [R]D was different, see Scheme 10). Derivatives which
differed at only C-4 or C-36 were distinguishable and different
from the natural product and were therefore discounted. Finally,
preparation of the tetra-Mosher ester derivatives of both cis-
1
observations were that the H and 13C NMR spectra and [R]D
of the synthetic material were indeed a very good match with
the literature data (see Supporting Information). Moreover, the
tetra-(R) and tetra-(S)-Mosher esters of cis-sylvaticin 1 were
also in accord with the data reported by McLaughlin. The
derivatization of sylvaticin was also undertaken according to
(21) (a) Hoye, T. R.; Hanson, P. R.; Hasenwinkel, L. E.; Ramirez, E. A.;
Zhuang, Z. Tetrahedron Lett. 1994, 35, 8529. (b) Curran, D. P.; Zhang,
Q.; Lu, H.; Gudipati, V. J. Am. Chem. Soc. 128, 9943.
(20) Nelson, D. J.; Henley, R. L.; Yao, Z.; Smith, T. D. Tetrahedron Lett.
1993, 34, 5885.
(22) Bhunnoo, R. A.; Hobbs, H.; Laine´, D. I.; Light, M. E.; Brown, R. C. D.
Org. Biomol. Chem. 2009, 7, 1017.
9
12860 J. AM. CHEM. SOC. VOL. 131, NO. 35, 2009

Syntheses of the Natural Products (+)-Sylvaticin and (+)-cis-Sylvaticin
A R T I C L E S
sylvaticin 1 and the C-4,36 bis-epimer 38 produced a set of
compounds that clearly showed the original assignment to be
correct.
was prepared by utilizing a one-pot hydride shift/intramolecular
oxo-carbenium ion reduction protocol.
Extensive derivatization of the products and stereoisomers
thereof has shown clearly that the original assignment of
stereochemistry, as reported by McLaughlin, was indeed
correct.
A similar sequence on the C-4,36 bis-epimer of sylvaticin
39 (which had NMR but not specific rotation data that matched
the literature) gave a derivative 40 that also had a poor match
of its di-(R) and di-(S)-Mosher esters. We can conclude from
these studies that the relative and absolute stereochemistry of
cis-sylvaticin and sylvaticin are as shown in Scheme 1 and were
correctly assigned by McLaughlin in the original isolation
paper.3
To conclude, we have reported efficient and concise routes
to two natural products and described here the first total
synthesis of sylvaticin. The key step in our route is the double
oxidative cyclization of a protected tetraol onto a diene unit
that forms two rings in one reaction with a high degree of
stereoselectivity. Moreover, the trans-THF ring of sylvaticin
Acknowledgment. We thank the EPSRC and AstraZeneca for
funding this project. Professor J. McLaughlin is thanked for
providing extensive spectra of both natural products. We thank A.
Cowley, A. Thompson and the Oxford Chemical Crystallography
Service for assistance and use of instrumentation.
Supporting Information Available: Experimental procedures
and spectral data for new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
JA9049959
9
J. AM. CHEM. SOC. VOL. 131, NO. 35, 2009 12861