Catalysis-based total synthesis of putative mandelalide A

Article abstract of DOI:10.1002/anie.201400605

A concise synthesis of the putative structure assigned to the highly cytotoxic marine macrolide mandelalide A (1) is disclosed. Specifically, an iridium-catalyzed two-directional Krische allylation and a cobalt-catalyzed carbonylative epoxide opening served as convenient entry points for the preparation of the major building blocks. The final stages feature the first implementation of terminal-acetylene metathesis into natural product synthesis, which is remarkable as this class of substrates was beyond reach until very recently; key to success was the use of the highly selective molybdenum alkylidyne complex 42 as the catalyst. Although the constitution and stereochemistry of the synthetic samples are unambiguous, the spectra of 1 as well as of 11-epi-1 deviate from those of the natural product, which implies a subtle but deep-seated error in the original structure assignment. Bitter and sweet: The synthesis of the proposed structure of the cytotoxic macrolide mandelalide A reiterates the notion that structure elucidation of architecturally complex natural products is not always reliable. From the chemical viewpoint, the chosen route attests to the power of (transition) metals as catalysts for stereoselective synthesis. Most notable is the first application of terminal-acetylene metathesis to natural product chemistry.

Full text of DOI:10.1002/anie.201400605

Angewandte
Chemie
DOI: 10.1002/anie.201400605
Total Synthesis
Catalysis-Based Total Synthesis of Putative Mandelalide A**
Jens Willwacher and Alois Fꢀrstner*
Abstract: A concise synthesis of the putative structure assigned
to the highly cytotoxic marine macrolide mandelalide A (1) is
disclosed. Specifically, an iridium-catalyzed two-directional
Krische allylation and a cobalt-catalyzed carbonylative epo-
xide opening served as convenient entry points for the
preparation of the major building blocks. The final stages
feature the first implementation of terminal-acetylene meta-
thesis into natural product synthesis, which is remarkable as
this class of substrates was beyond reach until very recently;
key to success was the use of the highly selective molybdenum
alkylidyne complex 42 as the catalyst. Although the constitu-
tion and stereochemistry of the synthetic samples are unam-
biguous, the spectra of 1 as well as of 11-epi-1 deviate from
those of the natural product, which implies a subtle but deep-
seated error in the original structure assignment.
excellent opportunity to scrutinize catalytic methods that
have been developed in our laboratory (Scheme 1). Previous
T
he supply of meaningful amounts of promising bioactive
experience suggests that a sequence of ring-closing alkyne
metathesis (RCAM)[7,8] followed by semi-reduction of the
resulting cycloalkyne A should allow the macrocyclic frame
compounds that are basically inaccessible from the source
organism is increasingly recognized as a prime task of
contemporary natural product synthesis.[1] Many secondary
metabolites of marine origin fall into this category. This
notion is exemplified by the mandelalides, a small family of
(glycosylated) macrolides that are derived from a rare South
African ascidian of the Lissoclinum genus.[2] Although
preliminary data indicate impressive in vitro cytotoxicity
against human NCI-H460 lung cancer and mouse Neuro-2A
neuroblastoma cell lines,[3] a systematic survey could not be
carried out with the available minute sample (800 mg).[2]
Mandelalide A (1) shows considerable structural homo-
logy to madeirolide A (2), an equally scarce metabolite, which
was isolated from a marine Leiodermatium sponge.[4,5]
Interestingly though, 2 exhibited no appreciable antiprolifer-
ative activity against two pancreatic cancer cell lines.[6] Even if
one takes the differences in the bioassays into account, these
somewhat conflicting results raise questions as to possible
structure–activity relationships and certainly advocate closer
inspection.
of 1 to be forged with excellent selectivity at the Z-configured
[9]
ꢀ
C14 C15 olefin site. Although the latest generation of
catalysts that are based on molybdenum alkylidynes with
triphenylsilanolate ligands is expected to cope with the polar
functionality that is presented by the envisaged cyclization
precursor of type B,[10] the virtues of these catalysts can be
probed in an unorthodox but particularly stringent way by
variation of the substituents on the alkyne termini.
Alkyne metathesis, as it stands today, largely relies on the
use of internal acetylene derivatives, because terminal
alkynes are prone to rapid polymerization on contact with
As a quintessential first step, we pursued the total
synthesis of mandelalide A (1) as the most active member
of the series. In the quest for a practical solution, we saw an
[*] Dipl.-Chem. J. Willwacher, Prof. A. Fꢀrstner
Max-Planck-Institut fꢀr Kohlenforschung
45470 Mꢀlheim/Ruhr (Germany)
E-mail: fuerstner@kofo.mpg.de
[**] Generous financial support by the MPG and the Fonds der
Chemischen Industrie (Kekulꢁ stipend to J.W.) is gratefully
acknowledged. We thank Mrs K. Holthusen for assistance in the
preparation of the rhamnosyl donor and the Analytical Departments
of our Institute for excellent support.
Scheme 1. Retrosynthetic analysis of mandelalide A (1) based on ring-
closing alkyne metathesis (RCAM) of a precursor containing one
internal and one terminal acetylene.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201400605.
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
a Schrock alkylidyne.[7] From the synthetic vantage point,
however, their use is desirable as they are often accessible in
fewer steps. The first successful metathesis reactions of
terminal alkynes have been reported only very recently
using catalysts with barely basic ancillary ligands.[11,12] Yet,
attempted ring closure of di-terminal diyne substrates gave
variable outcomes, which might be explained, at least in part,
by the particular constraints of macrocyclization; in contrast,
reactions of substrates that comprise one terminal and one
internal alkyne seem to be more robust.[12] However, all
available information derives from a small set of simple
model compounds, and this tactic definitely needs further
validation. The mandelalide case, invoking an elaborate
cyclization precursor of type B, provides a compelling test.
The required acid sector C was readily prepared from
propane-1,3-diol (Scheme 2). An iridium-catalyzed two-
directional Krische allylation furnished multigram amounts
of the known C2-symmetric diol 4 with exquisite levels of
diastereo- and enantioselectivity and hence served as a prac-
tical entry point.[13] Desymmetrization by iodoetherification
followed by TBS protection afforded alkyl iodide 5,[14] which
was reacted with the lithium enolate of 6. Although somewhat
forcing conditions were needed for this alkylation to proceed,
the methyl-branched chiral center was set with impeccable
selectivity (d.r. = 97:3).[15] Reductive cleavage of the auxiliary
preceded chain extension by cross metathesis with methyl
acrylate.[16] Subsequent oxidation of the primary alcohol
terminus led to the corresponding aldehyde 8, which was
exposed to the lithium anion of 14 to give the desired enyne
10. The yield, however, was modest and variable and could
not be improved, despite considerable experimentation.
Therefore, aldehyde 8 was first subjected to a Takai olefina-
tion[17–19] to give alkenyl iodide 9, which was amenable to
a modified Suzuki propynylation;[20] this reliable sequence
was much better suited for material throughput. Somewhat
unexpectedly though, the saponification of the methyl ester in
10 gave rather complex mixtures; best results were obtained
with Me3SiOK in Et2O, although partial deconjugation of the
double bond could not be avoided even under these mild
conditions (ca. 8%). This outcome, albeit a slight complica-
tion with respect to product characterization, was inconse-
quential and could be rectified prior to macrocyclization (see
below).
The second major building block D was prepared as
shown in Scheme 3. Silylation of commercially available 15
followed by cobalt-catalyzed carbonylative epoxide opening
in the presence of N-(trimethylsilyl)morpholine opened
a convenient entry into amide 17, which reacted with 1-
propenylmagnesium bromide (18) to give enone 19 in good
overall yield.[21,22] In parallel, the terminal alkene 23 was
prepared by a copper/TEMPO-co-catalyzed air oxidation of
20[23] followed by scandium triflate catalyzed crotylation of
the resulting aldehyde 21 with the chlorosilane donor 31. As
expected, this reaction was distinguished by excellent dia-
stereo- and enantioselectivity in favor of the syn-configured
product 22 (d.r. = 98:2, 94% ee).[24] Cross-metathesis of the
derived TES ether 23 with enone 19 worked best using
complex 32 as the catalyst.[25] A SmI2-catalyzed Evans–
Tishchenko reaction of product 24 with an excess of
iPrCHO set the anti-configured 1,3-diol motif while discrim-
inating the two alcohol groups at the same time.[26,27]
Compound 25 thus formed was swiftly elaborated into the
homoallylic alcohol 26 as an adequate precursor for the
signature tetrahydrofuran motif yet to be closed. A brief
screening of several conceivable [I+] or [PhSe+] sources
suggested that a selenoetherification under acidic conditions
was most promising, provided the system was operated in the
presence of a catalytic Lewis base activator.[28] Specifically,
treatment of 26 with a slight excess of N-(phenylselenyl)ph-
thalimide and trifluoroacetic acid in the presence of catalytic
=
amounts of Ph3P S in CH2Cl2 at low temperature furnished
product 27 in high yield and excellent selectivity (d.r. = 14:1).
Attempted deselenation over Raney Ni with concomitant
cleavage of the terminal benzyl ether gave erratic results.
Therefore, this compound was first subjected to deselenation
under free-radical conditions followed by hydrogenolysis of
the benzyl ether over Pearlmanꢀs catalyst. The resulting
alcohol was oxidized to give the rather sensitive aldehyde 28.
Although treatment with diazophosphonate 33 under stan-
dard conditions (K2CO3, MeOH)[29] provided alkyne 29,
substantial epimerization at the C17 position through
a retro-oxa-Michael/Michael process was noticed. Gratify-
ingly, this stereochemical erosion could be suppressed by
activation of 33 with NaOMe in THF prior to addition of the
aldehyde at low temperature.[30] The terminal alkyne 30 thus
formed was meant to serve in the projected alkyne metathesis
Scheme 2. a) [Ir(cod)Cl]2 (5 mol%), 12 (10 mol%), 3-nitro-4-chloroben-
zoic acid (20 mol%), Cs2CO3, 1,4-dioxane, 908C, 71% (d.r.ꢁ29:1,
99% ee); b) I2, NaHCO3, MeCN, ꢀ408C, 81% (d.r.=5:1); c) TBSOTf,
2,6-lutidine, CH2Cl2, 08C, 96%; d) LDA, LiCl, THF, 08C!458C, 76%
(d.r.=97:3); e) LDA, BH3·NH3, THF, 08C!RT, 96%; f) methyl acrylate,
13 (3 mol%), CH2Cl2, 83% (+ 7% of the Z-isomer); g) Dess–Martin
periodinane, CH2Cl2, 08C!RT, 77%; h) CHI3, CrCl2·THF, THF, ꢀ88C,
aq. serine work-up, 72% (+8% of the Z-isomer); i) propynyl sodium,
B(OMe)3, [Pd(dppf)Cl2]·(CH2Cl2) (10 mol%), THF, 708C, 81%;
j) Me3SiOK, Et2O, 80% (+7% of the b,g-isomer); k) 14, LiHMDS,
THF, ꢀ788C, 41–54% (E/Z=7:1). Aux*=(1S,2S)-2-(methyl-l2-azanyl)-
1-phenylpropan-1-ol, cod=1,5-cyclooctadiene, dppf=1,1’-bis(diphenyl-
phosphino)ferrocene, LDA=lithium diisopropylamide, LiHMDS=
lithium hexamethyldisilazide, Mes=2,4,6-trimethylphenyl, TBS=tert-
butyldimethylsilyl, Tf=trifluoromethanesulfonyl.
2
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
rhamnoside 35 as the 1,2-diacetal 36,[32] methylation
of the remaining alcohol, and exchange of the
bisketal in 37 for two acetyl groups to give product
38 (Scheme 4). Oxidative cleavage of the allyl
moiety[33] followed by installation of an anomeric
trichloroacetimidate afforded product 40 as an
adequate glycosyl donor.[34]
With appreciable amounts of all fragments in
hand, the project entered the crucial assembly
phase in which the macrocyclic perimeter of
mandelalide A had to be forged (Scheme 5). The
seemingly routine esterification of acid 11 with
alcohol 30 was surprisingly challenging. DCC gave
the best results amongst the host of activating
agents investigated, although the reaction was
accompanied by substantial isomerization of the
enoate double bond out of conjugation. This prob-
lem could be fixed upon treatment of the resulting
mixture with catalytic amounts of DBU in MeCN at
508C.[35] In striking contrast, the macrocyclization
of diyne 41 proceeded with remarkable ease even at
ambient temperature when catalyzed by molybde-
num alkylidyne 42 as the arguably most active and
selective catalyst for alkyne metathesis currently
available;[10] the reaction was performed in the
presence of molecular sieves to sequester the
released propyne. In view of the dense decoration
of 41 with polar substituents and the fact that it
comprises a challenging terminal alkyne unit, which
was beyond reach of RCAM until recently, the
outcome attests to the maturity of this method.[7]
Activated Zn(Cu/Ag) in protic medium served
the semi-reduction of 43 to the corresponding E,Z-
diene very well.[36] Slightly acidic conditions then
allowed the TBS ether to be cleaved without
touching the more robust TBDPS groups, which
hence set the stage for the attachment of the
saccharide. As expected, the TESOTf-catalyzed
rhamnosylation of 44 was productive and selective
alike.[37] Finally, global deprotection of the resulting
product 45 was achieved by careful saponification
of the acetate groups followed by cleavage of the
remaining silyl ethers with HF·pyridine in THF/
pyridine. These buffered conditions were the only
way to effect this final step in good yield even
though a variety of other reagents was screened;
however, these either failed to remove the shielded
secondary TBDPS ether and/or led to significant
ring expansion by migration of the lactone acyl
group onto the emerging primary alcohol site.
Overall, this concise route is deemed quite
satisfactory, not least because the majority of the
constructive operations are catalytic in nature (21
steps, longest linear sequence; ca. 4.5% overall
Scheme 3. a) TBDPSCl, imidazole, CH2Cl2, 08C!RT, 94%; b) [Co2(CO)8] (8 mol%),
CO (1 atm), N-(trimethylsilyl)morpholine, EtOAc, 74%; c) 18, THF, ꢀ258C (acidic
quench at ꢀ788C), 83% (E/Z=2:1); d) [Cu(MeCN)4]BF4 (5 mol%), 2,2’-bipyridine
(5 mol%), TEMPO (5 mol%), N-methylimidazole (10 mol%), MeCN, air, 94%;
e) (R,R)-31, Sc(OTf)3 (5 mol%), CH2Cl2, ꢀ788C!08C, 82% (d.r.=98:2, 94% ee);
f) TESCl, Et3N, DMAP, CH2Cl2, 08C!RT, 90%; g) 19, 32 (8 mol%), CH2Cl2, reflux,
79%; h) iPrCHO, SmI2 (35 mol%), THF, ꢀ508C, 78% (d.r.>19:1); i) TBDPSCl,
imidazole, CH2Cl2, 08C!RT, 87%; j) camphorsulfonic acid (30 mol%), CH2Cl2/
MeOH (2:1), 08C, 97%; k) N-(phenylselenyl)phthalimide, trifluoroacetic acid,
=
Ph3P S (12 mol%), CH2Cl2, ꢀ408C!ꢀ208C, 83% (d.r. =14:1); l) Bu3SnH, AIBN,
toluene, 808C, 93%; m) Pd(OH)2/C (cat.), H2 (1 atm), EtOH/EtOAc (9:1), 88%;
n) Dess–Martin periodinane, CH2Cl2, 08C!RT, 91%; o) 33, MeONa, THF, ꢀ788C;
then 28, ꢀ788C!ꢀ508C, 93% (d.r.ꢁ98:2); p) Dibal-H, toluene, ꢀ788C, 97%.
AIBN=azobis(isobutyronitrile), Dibal-H=diisobutylaluminum hydride, DMAP=4-
dimethylaminopyridine, TBDPS=tert-butyldiphenylsilyl, TEMPO=2,2,6,6-tetra-
methyl-1-piperidinyloxy radical, TES=trimethylsilyl, TMS=trimethylsilyl.
reaction.[31] Finally, the isobutylester in 29 was cleaved with
Dibal-H rather than by hydrolysis to avoid scrambling of the
silyl protecting groups.
The yet missing rhamnosyl segment E was prepared by
protection of the two equatorial hydroxy groups of allyl
yield), and more than 20 mg of 1 were made available in this
first round. The spectroscopic signatures of the synthetic
samples leave no doubt about the constitution and stereo-
chemistry of the material; regrettably though, comparison
with the data of the natural product revealed small but
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
assume that this site might be a locus of stereochemical
misassignment. Moreover, handheld models of the compound
with the opposite configuration at the C11 position suggested
that this isomer might also adopt a conformation that matches
the experimentally observed ROESY contacts and J coupling
patterns. To verify this hypothesis, 11-epi-1 was prepared by
following the same assembly route but using the enantiomeric
Myers auxiliary[15] in the alkylation step that sets the chiral
center in question.[38] Much to our dismay, the spectral data of
11-epi-1 deviate from those of the natural product to a similar
extent. Interestingly, however, the recorded coupling constant
(3J11,12 = 9.7 Hz) now fits the one of mandelalide A very well,
which possibly indicates more than one misassignment in the
originally proposed structure.
Therefore, we conclude that the error made by the
isolation team is subtle but profound, and a considerable
effort will be necessary to solve the riddle.[39] We (and many
others) have experienced such frustrating situations surpris-
ingly often in the recent past.[40,41] Somewhat ironically, they
remind us that contemporary natural product total synthesis
does not only serve the supply management alluded to in the
introduction; all too often it is needed, even in the age of ever
more sophisticated spectroscopy, to decide on structural
issues.[42]
Scheme 4. a) Allyl alcohol, H2SO4 (cat.), reflux, 78%; b) butane-2,3-
dione, MeC(OMe)3, pTsOH·H2O (cat.), MeOH, reflux, 72%; c) NaH,
MeI, DMF, 08C!RT, 64%; d) i) trifluoroacetic acid/H2O (20:1);
ii) Ac2O, DMAP, Et3N, CH2Cl2, 08C!RT, 68%; e) SeO2, HOAc, 1,4-
dioxane, reflux, 86%; f) Cl3CCN, Cs2CO3, CH2Cl2, 98%. pTs =para-
toluenesulfonyl.
distinct differences. The largest deviations in the 13C NMR
spectra were observed for the C11 carbon atom (Dd =+
1.4 ppm) and the attached C25 methyl branch (Dd =
ꢀ1.8 ppm). Furthermore, the recorded coupling constant
(3J11,12 = 7.6 Hz) is not in accordance with that of the natural
product (9.6 Hz). The configuration of this chiral center had
been assigned by the isolation team based on ROESY data
and the analysis of the homo- and heteronuclear coupling
constants.[2] As the macrocycle of mandelalide A is fairly
flexible, and the data hence necessarily represent an average
over more than one conformation, it seemed reasonable to
Received: January 20, 2014
Published online: && &&, &&&&
Keywords: alkyne metathesis · molybdenum · natural products ·
.
structure elucidation · total synthesis
[1] C. A. Kuttruff, M. D. Eastgate,
P. S. Baran, Nat. Prod. Rep. 2014,
DOI: 10.1039/c3np70090a.
[2] J. Sikorska, A. M. Hau, C. Anklin,
S. Parker-Nance, M. T. Davies-
Coleman, J. E. Ishmael, K. L.
McPhail, J. Org. Chem. 2012, 77,
6066 – 6075.
[3] Compound 1, as the most active
member, shows IC50 values of
12 nm and 44 nm for the NCI-
H460 and Neuro-2A cell lines,
respectively; see Ref. [2].
[4] a) P. L. Winder, Ph.D. Thesis,
Florida Atlantic University, 2009;
for a study towards 2 and a struc-
tural discussion, see: b) I. Pater-
son, G. W. Haslett, Org. Lett.
2013, 15, 1338 – 1341.
[5] As 1 derives from an ascidian
whereas
2 was isolated from
a sponge, it is not unreasonable
to assume that the actual produc-
ers of these compounds are (bac-
terial) symbionts.
Scheme 5. a) DCC, DMAP, CH2Cl2, 64% (a,b/b,g isomers=1.5:1); b) DBU (25 mol%), MeCN, 508C,
91%; c) 42 (10 mol%), M.S. (4 ꢂ and 5 ꢂ), toluene, 72%; d) Zn(Cu/Ag), THF/MeOH/H2O (1:1:1),
458C, 88%; e) pTsOH·H2O (cat.), CH2Cl2/MeOH (2:1), 90%; f) 40, TESOTf (30 mol%), CH2Cl2, M.S.
(4 ꢂ), ꢀ508C, 89%; g) K2CO3, MeOH, 08C, 80%; h) HF·pyridine, pyridine, THF, 08C!RT, 80%.
[6] However, madeirolide A was
reported to be strongly fungicidal;
see Ref. [4].
[7] A. Fꢁrstner, Angew. Chem. 2013,
125, 2860 – 2887; Angew. Chem.
Int. Ed. 2013, 52, 2794 – 2819.
=
DCC N,N’-dicyclohexylcarbodiimide, DBU=1,8-diazabicyclo[5.4.0]undec-7-ene, M.S.=molecular
sieves.
4
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
[8] a) A. Fꢁrstner, G. Seidel, Angew. Chem. 1998, 110, 1758 – 1760;
Angew. Chem. Int. Ed. 1998, 37, 1734 – 1736; b) A. Fꢁrstner, O.
Guth, A. Rumbo, G. Seidel, J. Am. Chem. Soc. 1999, 121, 11108 –
11113.
[29] S. Mꢁller, B. Liepold, G. J. Roth, H. J. Bestmann, Synlett 1996,
521 – 522.
[30] K. C. Nicolaou, K. C. Fylaktakidou, H. Monenschein, Y. Li, B.
Weyershausen, H. J. Mitchell, H. Wie, P. Guntupalli, D. Hep-
worth, K. Sugita, J. Am. Chem. Soc. 2003, 125, 15433 – 15442.
[31] C-Methylation of 30 is possible but tricky because of the
presence of the CH-acidic ester group and requires addition of
hexamethylphosphoramide (HMPA) for optimal results. The
successful engagement of the terminal alkyne in the RCAM-
based macrocyclization is therefore not only advantageous in
terms of step count, but also from a practical perspective.
[32] S. V. Ley, D. K. Baeschlin, D. J. Dixon, A. C. Foster, S. J. Ince,
H. W. M. Priepke, D. J. Reynolds, Chem. Rev. 2001, 101, 53 – 80.
[33] K. Kariyone, H. Yazawa, Tetrahedron Lett. 1970, 11, 2885 – 2888.
[34] R. R. Schmidt, W. Kinzy, Adv. Carbohydr. Chem. Biochem. 1994,
50, 21 – 123.
[9] A. Fꢁrstner, Science 2013, 341, 1229713.
[10] a) J. Heppekausen, R. Stade, A. Kondoh, G. Seidel, R. Goddard,
A. Fꢁrstner, Chem. Eur. J. 2012, 18, 10281 – 10299; b) J.
Heppekausen, R. Stade, R. Goddard, A. Fꢁrstner, J. Am.
Chem. Soc. 2010, 132, 11045 – 11057.
[11] B. Haberlag, M. Freytag, C. G. Daniliuc, P. G. Jones, M. Tamm,
Angew. Chem. 2012, 124, 13195 – 13199; Angew. Chem. Int. Ed.
2012, 51, 13019 – 13022.
[12] P. Persich, J. Llaveria, R. Lhermet, T. de Haro, R. Stade, A.
Kondoh, A. Fꢁrstner, Chem. Eur. J. 2013, 19, 13047 – 13058.
[13] a) X. Gao, H. Han, M. J. Krische, J. Am. Chem. Soc. 2011, 133,
12795 – 12800; b) Y. Lu, I. S. Kim, A. Hassan, D. J. Del Valle,
M. J. Krische, Angew. Chem. 2009, 121, 5118 – 5121; Angew.
Chem. Int. Ed. 2009, 48, 5018 – 5021..
[35] R. Ballini, G. Bosica, D. Fiorini, Tetrahedron Lett. 2001, 42,
8471 – 8473.
[14] The minor isomer was recycled to diol 4 on treatment with
activated zinc in DMA (80%).
[36] W. Boland, N. Schroer, C. Sieler, M. Feigel, Helv. Chim. Acta
1987, 70, 1025 – 1040.
[15] A. G. Myers, B. H. Yang, H. Chen, L. McKinstry, D. J. Kopecky,
J. L. Gleason, J. Am. Chem. Soc. 1997, 119, 6496 – 6511.
[16] A. K. Chatterjee, T.-L. Choi, D. P. Sanders, R. H. Grubbs, J. Am.
Chem. Soc. 2003, 125, 11360 – 11370.
[17] T. Okazoe, K. Takai, K. Utimoto, J. Am. Chem. Soc. 1987, 109,
951 – 953.
[18] A. Fꢁrstner, Chem. Rev. 1999, 99, 991 – 1045.
[19] Optimal results were obtained upon work-up with a basic serine
solution; see: D. P. Stamos, X. C. Sheng, S. S. Chen, Y. Kishi,
Tetrahedron Lett. 1997, 38, 6355 – 6358.
[20] G. Seidel, A. Fꢁrstner, Chem. Commun. 2012, 48, 2055 – 2070.
[21] S. N. Goodman, E. N. Jacobsen, Angew. Chem. 2002, 114, 4897 –
4899; Angew. Chem. Int. Ed. 2002, 41, 4703 – 4705.
[22] It is essential that the EtOAc that is used in the carbonylation is
rigorously dried and acid-free; likewise, the Grignard addition
must be quenched at ꢀ788C to avoid Michael addition of the
released morpholine; see: A. Gomtsyan, R. J. Koenig, C.-H. Lee,
J. Org. Chem. 2001, 66, 3613 – 3616; for details, see the
Supporting Information.
[23] J. M. Hoover, S. S. Stahl, J. Am. Chem. Soc. 2011, 133, 16901 –
16910.
[24] H. Kim, S. Ho, J. L. Leighton, J. Am. Chem. Soc. 2011, 133,
6517 – 6520.
[25] Z.-Y. Zhan, PCT WO 2007/003135A1, 2007.
[26] D. A. Evans, A. H. Hoveyda, J. Am. Chem. Soc. 1990, 112, 6447 –
6449.
[27] a) K. J. Ralston, A. N. Hulme, Synthesis 2012, 2310 – 2324; see
also: b) L. Brewitz, J. Llaveria, A. Yada, A. Fꢁrstner, Chem. Eur.
J. 2013, 19, 4532 – 4537.
[37] The minor anomer (d.r. ꢂ 16:1) could be removed after cleavage
of the acetates.
[38] A Scheme showing the entire route is contained in the
Supporting Information; note, however, that not all steps leading
to 11-epi-1 were fully optimized.
[39] A comparison of 1 and 2 shows that their “northern” stereo-
clusters, which comprise the tetrahydrofuran ring, are enantio-
meric; one may be tempted to speculate that this entire domain
has been misassigned in 1.
[40] For reviews, see: a) K. C. Nicolaou, S. A. Snyder, Angew. Chem.
2005, 117, 1036 – 1069; Angew. Chem. Int. Ed. 2005, 44, 1012 –
1044; b) M. E. Maier, Nat. Prod. Rep. 2009, 26, 1105 – 1124.
[41] a) A. Larivꢂe, J. B. Unger, M. Thomas, C. Wirtz, C. Dubost, S.
Handa, A. Fꢁrstner, Angew. Chem. 2011, 123, 318 – 323; Angew.
Chem. Int. Ed. 2011, 50, 304 – 309; b) P. Buchgraber, T. N.
Snaddon, C. Wirtz, R. Mynott, R. Goddard, A. Fꢁrstner, Angew.
Chem. 2008, 120, 8578 – 8582; Angew. Chem. Int. Ed. 2008, 47,
8450 – 8454; c) A. Kondoh, A. Arlt, B. Gabor, A. Fꢁrstner,
Chem. Eur. J. 2013, 19, 7731 – 7738; d) A. Fꢁrstner, L. C.
Bouchez, L. Morency, J.-A. Funel, V. Liepins, F.-H. Porꢂe, R.
Gilmour, D. Laurich, F. Beaufils, M. Tamiya, Chem. Eur. J. 2009,
15, 3983 – 4010; e) A. Fꢁrstner, K. Radkowski, C. Wirtz, R.
Goddard, C. W. Lehmann, R. Mynott, J. Am. Chem. Soc. 2002,
124, 7061 – 7069; for a correction of a putative bioactivity with
the help of synthetic material, see: f) K. Micoine, P. Persich, J.
Llaveria, M.-H. Lam, A. Maderna, F. Loganzo, A. Fꢁrstner,
Chem. Eur. J. 2013, 19, 7370 – 7383.
[42] For an instructive case, see: J. Willwacher, N. Kausch-Busies, A.
Fꢁrstner, Angew. Chem. 2012, 124, 12207 – 12212; Angew. Chem.
Int. Ed. 2012, 51, 12041 – 12046.
[28] a) S. E. Denmark, D. Kalyani, W. R. Collins, J. Am. Chem. Soc.
2010, 132, 15752 – 15765; b) S. E. Denmark, W. R. Collins, Org.
Lett. 2007, 9, 3801 – 3804.
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

.
Angewandte
Communications
Communications
Total Synthesis
Bitter and sweet: The synthesis of the
proposed structure of the cytotoxic mac-
rolide mandelalide A reiterates the notion
that structure elucidation of architectur-
ally complex natural products is not
always reliable. From the chemical view-
point, the chosen route attests to the
power of (transition) metals as catalysts
for stereoselective synthesis. Most nota-
ble is the first application of terminal-
acetylene metathesis to natural product
chemistry.
J. Willwacher,
A. Fꢁrstner*
&&&& — &&&&
Catalysis-Based Total Synthesis of
Putative Mandelalide A
6
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
These are not the final page numbers!