The Formosalides: Structure Determination by Total Synthesis

Article abstract of DOI:10.1002/anie.202011472

Total synthesis allowed the constitution of the cytotoxic marine macrolides of the formosalide family to be confirmed and their previously unknown stereostructure to be assigned with confidence. The underlying blueprint was inherently modular to ensure that each conceivable isomer could be reached. This flexibility derived from the use of strictly catalyst controlled transformations to set the stereocenters, except for the anomeric position, which is under thermodynamic control; as an extra safety measure, all stereogenic centers were set prior to ring closure to preclude any interference of the conformation adopted by the macrolactone rings of the different diastereomers. Late-stage macrocyclization by ring-closing alkyne metathesis was followed by a platinum-catalyzed transannular 6-exo-dig hydroalkoxylation/ketalization to craft the polycyclic frame. The side chain featuring a very labile unsaturation pattern was finally attached to the core by Stille coupling.

Full text of DOI:10.1002/anie.202011472

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: The Formosalides: Structure Determination by Total Synthesis
Authors: Alois Fürstner, Saskia Schulthoff, James Y. Hamilton, Marc
Heinrich, Yonghoon Kwon, and Conny Wirtz
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.202011472
Link to VoR: https://doi.org/10.1002/anie.202011472

10.1002/anie.202011472
Angewandte Chemie International Edition
The Formosalides: Structure Determination by Total Synthesis
Saskia Schulthoff,[+] James Y. Hamilton,[+] Marc Heinrich, Yonghoon Kwon, Conny Wirtz, and Alois
Fürstner*
Abstract: Total synthesis allowed the question to be answered which of eight possible stereoisomeric
compounds most likely represents the cytotoxic marine macrolides of the formosalide family. The
underlying blueprint was inherently modular to ensure that each conceivable isomer could be reached.
This flexibility derived from the use of strictly catalyst-controlled transformations to set the
stereocenters, except for the anomeric position that is under thermodynamic control; as an extra safety
measure, all chiral centers were set prior to ring closure to preclude any interference of the
conformation adopted by the macrolactone rings of the different diastereomers, which vary from case
to case. Late stage macrocyclization by ring closing alkyne metathesis (RCAM) was followed by a
platinum-catalyzed transannular 6-exo-dig hydroalkoxylation/ketalization to craft the polycyclic frame.
The side chain featuring a very labile unsaturation pattern was finally attached to the core via Stille
coupling. In the end, it proved unnecessary to make the entire panel of isomers; rather, formation of a
single target together with two model compounds allowed the constitution of the formosalides to be
confirmed and their previously unknown stereostructure to be assigned with confidence.
Dinoflagellates are one of the largest groups of marine eukaryotes and, as such, form a major fraction
of plankton.1 In spite of their rather simple morphology, some of these unicellular organisms comprise
disproportionally large genomes and, as a result, are capable of producing secondary metabolites of
exceptional structural complexity and diversity. Examples include the protein phosphatase inhibitor
okadaic acid,2 the family of the ladder polyether toxins (maitotoxin, brevetoxin, ciguatoxin etc.),3 the
complex alkaloid saxitoxin,4 the unusual neurotoxin belizentrin,5,6 and the amphidinolide family of
macrolides;7,8 this selection represents only a tiny fraction of the product spectrum but showcases the
*
S. Schulthoff, Dr. J. Y. Hamilton, Dr. M. Heinrich, Dr. Y. Kwon, C. Wirtz, Prof. A. Fürstner
Max-Planck-Institut für Kohlenforschung
45470 Mülheim/Ruhr (Germany)
Email: fuerstner@kofo.mpg.de
[+]
These authors contributed equally to this work
Supporting Information and ORCID identification numbers of the authors of this article can be found under
##########
1
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
molecular intricacy as well as the potent and diverse biological activities that many dinoflagellate-
derived compounds do possess.
Scheme 1. One of eight possible stereomers that might represent the formosalides: since only the
relative stereochemistry of the color-coded stereoclusters was determined by the isolation team but
their inter-relationships and the absolute configuration could not be established, eight isomers need
to be considered (see the Supporting Information); blueprint of an “integral synthesis”: provided that
the projected transformations proceed under catalyst (reagent) control, the entire panel of possible
isomers comes into reach.
In 2009, Lu and coworkers reported the isolation of two unusual cytotoxic macrolides from a marine
dinoflagellate Prorocentrum sp.;9 this genus is known as a rich source of bioactive secondary
metabolites.2,10 The producing organism (strain PL040104002) was found in the wash-off epiphytes of
seaweeds collected off the southern Taiwanese coastline and was mass-cultured in the laboratory in a
seawater medium. Formosalide A (1) and B (2) differ only in the anomeric substituent at C-8 (Scheme
1); in view of the fact that the cell harvest was extracted with MeOH and the subsequent
chromatographic purification steps used MeOH-containing eluents,9 it is not inconceivable that the
2
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
tertiary methyl ketal in 2 is an artefact of isolation. The 17-membered lactone is reminiscent of the
amphidinolide family, many members of which feature odd-numbered rings as their principle
constituents, which are puzzling from the biosynthetic viewpoint.7,8 Likewise, the 2,5-trans-
disubstituted tetrahydrofuran and the “glycosidic” tetrahydropyran unit inscribed into the macrocyclic
frame of 1 and 2 find close correspondence in the amphidinolide series and other marine natural
products.11 A particularly striking substructure of the formosalides is the skipped, all-cis configured
diene/diene motif within the lipophilic side chain, which is expected to be highly isomerization-prone
for its thermodynamically unfavorable character.
Thorough NMR studies allowed only the relative configuration within each of the three color-coded
stereoclusters to be determined but could not establish their mutual relationships nor their absolute
configuration;9 eight possible combinations (four diastereomers and their respective enantiomers)
have hence to be considered. Because of the flexibility of the macrocycle and the lack of
crystallographic data, synthesis is arguably the most definitive way to the answer.12 A priori, such a
project must be capable of producing any conceivable isomer.13 We reasoned that the blueprint
sketched in Scheme 1 qualifies for such an “integral approach”:14 specifically, the late-stage
attachment of the polyene side chain seemed advisable for the sake of convergence and for stability
reasons alike. Convergence is also imperative for the assembly of the eight diastereomeric macrocyclic
cores that are potentially needed. To reach this goal, we opted for a strategy that puts all pertinent
stereocenters in place in a strictly catalyst-controlled fashion,15 capable of overturning, if necessary,
the inherent bias of the individual substrates. Since macrocyclic stereocontrol can be significant,16 it
was planned to do so prior to ring closure in order to prevent any interference of the conformations
adopted by the macrolactones, which will vary across the different diastereomers. As a consequence,
the two critical cyclization reactions necessarily become late-stage events: it was planned to convene
them at the “anomeric” position by taking advantage of the fact that a tert-hemiketal has the same
formal oxidation state as the C-atom of an alkyne.17 When encoded in the form of a triple bond, a
sequence of ring closing alkyne metathesis (RCAM)18,19 (C  B) followed by transannular
functionalization of the resulting cycloalkyne (B  A) should allow the macrocyclic frame and the yet
missing ketal subunit to be formed. The latter maneuver, however, bore considerable risk since 6-exo-
dig hydroalkoxylations of alkynes in general are challenging;20,21,22 in the projected transannular
3
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
format, the conformation of the macrocycle might intervene.23,24 If successful, however, such a merger
of RCAM and -acid catalysis25 streamlines the assembly process and hence minimizes the synthetic
exertion.
Scheme 2. a) [{Ir(cod)Cl}2] (2.5 mol%), (S)-SEGPHOS (5.0 mol%), 14, formic acid, Na2SO4, THF, 60°C, 56%
(7.8 g scale, dr > 20:1, ee > 95%); b) NaH, PMBCl, Bu4NI (10 mol%), DMF, 0°CRT, 91% (9.1 g scale, ee
= 98%); c) TBAF, THF, then NaOH, toluene, reflux, 86% (4.6 g scale); d) (pin)BH, Cy2BH (10 mol%), neat,
82% (4.6 g scale); e) [Pd(PPh3)4] (5 mol%, aq. NaOH (2 M), 1-iodo-1-propyne, THF, 80°C, 69% (2.8 g
scale); f) (DHQ)2PYR (5 mol%), K2OsO42H2O (2.1 mol%), K3Fe(CN)6, K2CO3, tBuOH/H2O, MeSO2NH2, 0°C,
94% (dr = 9:1, 2.9 g scale); g) TESOTf, 2,6-lutidine, CH2Cl2, 0°C, 63% (1.3 g scale); h) acetyl chloride, O-
TMS-quinidine (10 mol%), LiClO4, iPr2NEt, CH2Cl2/Et2O, 78°C, 78% (10, dr = 19:1, 1.1 g scale), 84% (12,
dr = 10:1, 1.0 g scale); i) (i) Me3Al, N,O-dimethylhydroxylamine hydrochloride, CH2Cl2, 0°CRT, then
10 (or 12), 0°C; (ii) allylmagnesium chloride, THF, 78°C0°C, 88% (11, 1 g scale), 95% (13, 0.9 g scale);
cod = 1,5-cyclooctadiene; Cy = cyclohexyl; pin = pinakolato; PMB = p-methoxybenzyl; TBAF = tetra-n-
butylammonium fluoride; TES = triethylsilyl; TIPS = tri(isopropyl)silyl; TMS = trimethylsilyl
4
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
The preparation of the required building blocks in all necessary stereochemical renditions started from
aldehyde 3 derived from cyclopentene in one step (Scheme 2).26 When subjected to an iridium
catalyzed Krische propargylation with 14,27,28 multigram quantities of 4 were secured, basically as a
single isomer (dr > 20:1, ee  95%).29,30 Since the C-C-bond formation is catalyst-controlled, it suffices
to use the enantiomeric SEGPHOS ligand to obtain ent-4 analogously (see the SI). PMB-Protection of
the newly formed alcohol, unmasking of the alkyne, followed by a Cy2BH-entrained hydroboration of
5 in neat HB(pin)31,32 set the stage for chain extension by Suzuki coupling of the resulting alkenyl
boronate 6 with freshly prepared 1-iodo-1-propyne.33 As expected for an E-configured enyne, the
Sharpless asymmetric dihydroxylation of 7 worked well using (DHQ)2PYR as the ligand in the presence
of methanesulfonamide as additive34 to give diol 8 in excellent yield with a dr = 9:1.35 The analogous
reaction of ent-7 under the aegis of the pseudoenantiomeric (DHQD)2PYR furnished ent-8 with similar
efficiency and selectivity.
With the C9-C12 stereotetrad representing the future tetrahydropyran unit of formosalide being
established in both enantiomeric formats, we next sought to set the second stereocluster at C-16/C-
18. Treatment of 8 with TESOTf/2,6-lutidine36 afforded aldehyde 9 in readiness for chain extension via
an acetate aldol equivalent. The asymmetric ketene cycloaddition catalyzed by O-TMS quinidine served
this purpose exceedingly well in that it converted 9 into -lactone 10 (78%, dr = 19:1), but ent-9 into
the diastereomeric -lactone 12 (84%, dr > 20:1 after flash chromatography);37 not only is the reaction
practical and scalable, but double-diastereoselective to the extent needed in the current endeavor.38
With this important aspect rigorously confirmed, it was clear that all stereoisomers of C were within
reach. Therefore, the preparation of the yet missing isomers was postponed until after the other
critical steps of the projected synthesis had been validated. For proof-of-concept, two truncated
formosalide-like macrolides of type A (X = H, R = Me) with simple alkene termini were initially targeted.
To this end, 10 was first converted into the corresponding Weinreb amide,39 which was then
transformed into the ,-unsaturated ketone 11 on treatment with allylmagnesium chloride.40,41 The
diastereomeric compound 13 was obtained analogously, thus setting the stage for fragment coupling
with concomitant 1,3-anti reduction via
a samarium-catalyzed Evans-Tishchenko redox
esterification.42,43,44 The required aldehyde 18 was attained in either enantiomeric form starting from
16 by a cobalt-catalyzed oxidative cyclization which forms the trans-disubstituted tetrahydrofuran ring
5
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
with excellent selectivity (Scheme 3).45 This exact transformation had already served our previous total
synthesis of amphidinolide C and F;46 it illustrates the chemoselectivity of such oxidative Mukaiyama
cyclization reactions in that only the olefinic site of 16 is engaged by the cobalt catalyst,45,47 whereas
the triple bond remains untouched.13a,b,46
Scheme 3. a) (R,R)-27, HOAc (15 mol%), H2O, toluene, 0°C  RT, 36% (>99% ee, 7.5 g scale); b) propyne,
nBuLi, BF3Et2O, THF, 78°C, 87% (6.1 g scale); c) Co(nmp)2 (10 mol%), tBuOOH, O2 (1 atm), iPrOH, 55°C,
84% (dr > 20:1, 3.8 g scale); d) [SO3pyridine], DMSO, iPr2NEt, CH2Cl2, 89% (1 g scale); e) (i) SmI2, PhCHO,
THF; (ii) 18, 11 (or 13) f) TBSOTf, 2,6-lutidine, CH2Cl2, 0°C, 68% (over two steps, 19 + 20 (3:1)), 64% (over
6
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
two steps, 23 + 24 (4:1)); g) 29 (30 mol%), CH2Cl2, toluene, 110°C, 82% (21 + 22), 77% (25 + 26); TBS =
tert-butyldimethylsilyl
With both partners in hand, fragment coupling via an Evans-Tishchenko reaction could be tackled.
Despite the excellent track record of this transformation,43 the specific application to fragments 11 and
18 turned out to be exceptionally taxing (Scheme 3): the catalytic (substoichiometric) variant proved
erratic and was not pursued further; even the reaction using over-stoichiometric amounts of
[(PhCHO)2SmISmI3] as the promoter gave product mixtures, despite considerable experimentation.
Although the 1,3-anti reduction per se proceeded nicely, a mixture of the regioisomeric esters 19 and
20 was formed. This outcome is perplexing in view of the generally accepted mechanism, which
involves a highly ordered chair-like transition state that does not put the site of ester formation at
risk.42,43 The same scrambling also plagued the reaction of the diastereomeric reagent combination 13
+ 18 that furnished the constitutional isomers 23 and 24. We are unaware of any precedent and are
inclined to believe that the observed transesterification reflects an acyl migration occurring post
factum but in situ. The additional O-donor in the THF ring of the aldehyde partner 18, which might
transiently ligate the oxophilic SmIII cation, could play a role.48 In any case, Evans-Tishchenko reactions
using (aliphatic) aldehydes with heteroatom sites in vicinity are rare, and problems have occasionally
been reported.44a,49,50,51
Scheme 4. a) DDQ, CH2Cl2, phosphate buffer (pH 7), 0°C, 47% (30), 96% (33); b) [(C2H4)PtCl2]2 (50 mol%),
2,6-di-tert-butylpyridine, Et2O; c) TBAF, THF; d) TMSCl, MeOH, 0°C, 19% (32, unoptimized, over three
steps (b-d)); 44% (35, unoptimized, over two steps (c,d)); e) [(C2H4)PtCl2]2 (3 mol%), Et2O, 77%; DDQ =
2,3-dichloro-5,6-dicyano-p-benzo-quinone
7
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
As the separation of the regioisomeric esters was tedious, the crude material was O-silylated and the
resulting mixture subjected to ring closing alkyne metathesis (RCAM) using a catalyst generated in situ
by activation of 29 with CH2Cl2, as previously described by our group.52,53 The resulting 17- and 19-
membered macrolides 21 and 22 can be separated by ordinary flash chromatography. The
macrocyclization worked equally well in the diastereomeric series (25/26), thus showing the
robustness of RCAM.18,19
Subsequent oxidative cleavage of the PMB-ether gave alcohols 30 and 33 in readiness for acetal
formation via -acid catalyzed transannular addition of the free –OH group across the triple bond,20 to
be followed by addition of MeOH (or H2O) to the resulting enol ether to form the desired (hemi)ketal
(Scheme 4).20 [(C2H4)PtCl2]2 (Zeise’s salt) in Et2O proved to be the promotor of choice for the 6-exo-dig
hydroalkoxylation step,20,54 which has apparently not yet found any application in transannular
format.20,55,56 For the exceptional sensitivity of the resulting enol ether 31, it proved mandatory to
supplement the PtII catalyst with 2,6-di-tert-butylpyridine to avoid instant decomposition;57 even with
this precaution, loss of material could not be fully suppressed. Gratifyingly, the diasteromeric product
34 proved somewhat more stable. In any case, (crude) 31 and 34 were transformed without delay into
the corresponding methyl ketals on treatment with TBAF followed by addition of MeOH promoted by
admixed TMSCl. It is notable that both transannular addition reactions proceeded with the same
regiochemical outcome, but the macrocyclic conformational diversity impact strongly on the stability
of the products and hence the efficiency of this step.
We were delighted by the serendipitous finding that the NMR data of 32 showed an astounding match
with those of formosalide B, whereas those of isomer 35 were clearly different (see the SI).58 Since the
spectra of the different diastereomers are obviously readily distinguishable59 and based on the
assumption that the presence/absence of the extended side chain is unlikely to alter the spectral
fingerprints of the core, we saw ourselves in the favorable position that the preparation of a single
product isomer rather than the entire ensemble might allow the stereostructure of formosalide to be
ascertained with confidence. To this end, the approach to the model compounds was simply re-routed
by attaching a handle for late-stage introduction of the side chain (Scheme 5). Specifically, 30 was
subjected to a modified Lemieux-Johnson oxidation to cleave the terminal alkene without interference
of the triple bond.60 The resulting crude aldehyde 36 underwent a Stork-Zhao olefination to form the
8
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
required Z-alkenyl iodide 38 as the only double bond isomer, provided that reaction was carried out in
DMPU/THF (4:1).61 Interestingly, however, small amounts of epoxide 37 were also isolated, which
seems to originate from a side reaction reminiscent of the Corey-Chaykovsky epoxidation using sulfur-
(rather than phosphorous) ylides.62 After cleavage of the PMB ether, the resulting compound 39 was
subjected to the Pt-mediated transannular cyclization in buffered medium followed by addition of
MeOH and concomitant cleavage of the TES-groups under slightly acidic conditions to give the
expected methyl ketal 40 in readiness for chain extension via Stille-Migita coupling to complete the
formosalide carbon skeleton. As the configuration of 40 matches that of the much more sensitive
model 32, the transannular addition/ketalization step faced the same issue of passing through an
exceptionally fragile intermediate; each of numerous experiments furnished the desired product, yet
the scatter in the yield was extreme (15-85%), despite the greatest care.
As expected, the required stannane partner 46 is a very delicate compound, the preparation of which
mandated careful optimization in order to prevent isomerization from occurring. The successful route
(Scheme 5) started from commercial 3-butyne-1-ol (41), which was converted via the known alkenyl
iodide 4263 into the phosphonium salt 43 which underwent a Z-selective Wittig reaction with enal 44.
The resulting product 45 had to be handled with greatest care in the dark to avoid instant scrambling.
The same threat overshadowed its subsequent conversion into stannane 46 on treatment with
hexamethylditin and catalytic amounts of [(Ph3P)2PdCl2]:64 the reaction mandates heating but
produces a mixture of isomers as soon as the temperature raises to 40°C. When strictly monitored,
however, the desired alkenyl stannane 46 was secured in good yield and purity.
9
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
Scheme 5. a) OsO4 (3 mol%), NaIO4, 2,6-lutidine, 1,4-dioxane/H2O, 88%; b) (i) iodomethyl-
triphenylphosphonium iodide, NaHMDS, THF, (ii) 36, DMPU/THF, 78°C, 38 (53%, Z:E > 20:1) + 37 (9%,
dr = 1.8:1); c) DDQ, CH2Cl2, phosphate buffer (pH 7), 0°C, 62%; d) [(C2H4)PtCl2]2 (50 mol%), 2,6-di-tert-
butylpyridine, Et2O; e) TMSCl, MeOH, 0°C, 15-85% (over two steps, see Text); f) 46, Pd(PPh3)4 (20
mol%), CuTC, [Ph2PO2][NBu4], DMF, 0°C  RT, 62%; g) TASF, aq. DMF, 0°C  RT, 41% (1) + 49% (2); h)
(i) Me3Al, Cp2ZrCl2, 1,2-dichloroethane, H2O cat., RT  reflux; (ii) I2, Et2O, 30°C  RT, 61%; i) TsCl,
Et3N, CH2Cl2, 0°C  RT, 78%; j) (i) NaI, acetone, reflux; (ii) PPh3, MeCN, 82%; k) LiHMDS, THF, 78°C,
then 44, DMPU, 78°C, 59%; l) Me3SnSnMe3, [(PPh3)2PdCl2] (5 mol%), THF, 35°C, 79%; Cp =
cyclopentadienyl; CuTC = copper thiophene-2-carboxylate; DMPU = 1,3-dimethyl-3,4,5,6-tetrahydro-
2(1H)-pyrimidinone; HMDS = hexamethyldisilazide; TASF = tris(dimethylamino)sulfur trimethylsilyl
difluoride; Ts = p-toluenesulfonyl
With 40 and 46 in hand, the critical Stille-Migita coupling was performed under conditions previously
developed in this laboratory for fragile polyunsaturated compounds and other demanding cases.65 It
is the combination of Pd(PPh3)4 as catalyst with CuTC as promotor and [Ph2PO2][NBu4] as an effective
yet essentially neutral tin scavenger that allowed the reaction to proceed even at 0°C such that the
10
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
configurational integrity of all stereochemical elements was preserved.13c,66 The remaining TBS ether
in resulting product was instantly deprotected with TASF in aqueous DMF; although these conditions
are known as particularly mild,67 we observed partial hydrolysis of the ketal. In the present context,
this unexpected outcome was deemed a fortuitous coincidence as it furnished synthetic formosalide
A (1) and (2) in a single operation for independent comparison with the spectra of the natural products.
As forecasted by the model compound 32, the recorded NMR data of synthetic 2 were in excellent
agreement with those of the authentic formosalide B reported in the literature. The comparison of
synthetic 1 with formosalide A proved more difficult because this compound shows an interesting
behavior in that the NMR spectrum is not static: a gentle drift of the 13C NMR signals with time and
concentration is observed (see the SI), which is tentatively ascribed to slow conformational changes
that come along with reorganization of the peripheral hydrogen bonding network. This phenomenon
is well documented for polyhydroxylated macrocycles;68 despite the slight fuzziness, the match
between recorded and reported data can rightfully be regarded as excellent too. Moreover, the sign
and magnitude of the []D value of both compounds also correspond well to those of the natural
products (see the SI).9
Therefore, we claim with confidence that the relative and absolute configuration of formosalide A and
B are correctly described by structures 1 and 2, respectively. The project had originally been designed
to cover the entire ensemble of eight conceivable isomers, as required for an ultimate proof. For
favorable circumstances, however, the preparation of three model compounds58 and of a single
product diastereomer sufficed to make a convincing case, allowing us to establish the most likely
stereostructure of these marine natural products. The project bears witness of the excellent level of
contemporary catalyst-controlled asymmetric synthesis as well as of the great strides that macrocycle
formation owes the advent of metathesis in general and RCAM in particular. These methods form a
sound basis for future forays of our laboratory into transannular chemistry, a highly promising field in
which a similar level of confidence has yet to be attained.
11
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
Acknowledgements
Generous financial support by the Swiss National Science Foundation (fellowship to J. Y. H.), the
Alexander-von-Humboldt Foundation (fellowship to Y. K.) and the MPG is gratefully acknowledged. We
thank Alberto Lopez-Rodriguez for valuable assistance and the analytical departments of our Institute
for excellent support.
For the Table of Contents
Identity Card: Total synthesis allowed the question to be answered which of eight possible stereomeric
compounds most likely represents the cytotoxic marine macrolides of the formosalide family. The
successful route was based upon an inherently flexible blueprint that builds all stereogenic centers by
strictly catalyst-controlled transformations prior to formation of the polycyclic framework by late-
stage alkyne metathesis followed by a -acid catalyzed transannular cyclization.
Keywords: Alkyne Metathesis  Natural Products  Platinum  Structure Elucidation  Total Synthesis
12
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
1
The Biology of Dinoflagellates (Ed.: F. J. R. Taylor), Bot. Monogr., Vol. 21, Blackwell, Oxford, 1987.
V. Valdiglesias, M. V. Prego-Faraldo, E. Pásaro, J. Méndez, B. Laffon, Mar. Drugs 2013, 11, 4328-
4349.
2
3
4
5
K. C. Nicolaou, M. O. Frederick, R. J. Aversa, Angew. Chem. Int. Ed. 2008, 47, 7182-7225.
L. E. Llewellyn, Nat. Prod. Rep. 2006, 23, 200-222.
H. J. Domínguez, J. G. Napolitano, M. T. Fernández-Sánchez, D. Cabrera-García, A. Novelli, M.
Norte, J. J. Fernández, A. H. Daranas, Org. Lett. 2014, 16, 4546-4549.
6
F. Anderl, S. Größl, C. Wirtz, A. Fürstner, Angew. Chem. Int. Ed. 2018, 57, 10712-10717.
J. Kobayashi, J. Antibiot. 2008, 61, 271-284.
7
8
A. Fürstner, Isr. J. Chem. 2011, 51, 329-345.
9
C.-K. Lu, Y.-M. Chen, S.-H. Wang, Y.-Y. Wu, Y.-M. Cheng, Tetrahedron Lett. 2009, 50, 1825-1827.
a) C.-K. Lu, Y.-M. Chen, S.-H. Wang, Tetrahedron Lett. 2010, 51, 6911-6914; b) C.-K. Lu, H.-N.
Chou, C.-K. Lee, T.-H. Lee, Org. Lett. 2005, 7, 3893-3896; c) A. R. Yang, S. Lee, Y. D. Yoo, H. S. Kim,
E. J. Jeong, J.-R. Rho, J. Nat. Prod. 2017, 80, 1688-1692; d) S. Lee, A. R. Yang, Y. D. Yoo, E. J. Jeong,
J.-R. Rho, J. Nat. Prod. 2019, 82, 1034-1039.
10
11
12
A. Lorente, J. Lamariano-Merketegi, F. Albericio, M. Alvarez, Chem. Rev. 2013, 113, 4567-4610.
A single synthetic study towards the formosalides has been published, which targeted the C1-
C16 fragment of only one conceivable isomer, see: S. Gajula, A. V. V. Reddy, D. P. Reddy, J. S.
Yadav, D. K. Mohapatra, ACS Omega 2020, 5, 10217-10224.
13
Provided that the constitution had been correctly assigned; for cases in which even an integral
synthesis could not solve the structure, see: a) M. Heinrich, J. J. Murphy, M. K. Ilg, A. Letort, J. T.
Flasz, P. Philipps, A. Fürstner, J. Am. Chem. Soc. 2020, 142, 6409-6422; b) M. Heinrich, J. J.
Murphy, M. K. Ilg, A. Letort, J. T. Flasz, P. Philipps, A. Fürstner, Angew. Chem. Int. Ed. 2018, 57,
13575-13581; c) A. Larivée, J. B. Unger, M. Thomas, C. Wirtz, C. Dubost, S. Handa, A. Fürstner,
Angew. Chem. Int. Ed. 2011, 50, 304-309.
14
15
L. Li, Z. Chen, X. Zhang, Y. Jia, Chem. Rev. 2018, 118, 3752-3832.
This condition does not apply to the “anomeric” position of the ketal, which is under
thermodynamic control.
13
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
16
For a classical treatise, see: W. C. Still, I. Galynker, Tetrahedron 1981, 37, 3981-3996.
17
a) I. V. Alabugin, E. Gonzalez-Rodriguez, R. K. Kawade, A. A. Stepanov, S. F. Vasilevsky, Molecules
2019, 24, 1036; b) Brenzovich, W. E. Angew. Chem. Int. Ed. 2012, 51, 8933-8935.
A. Fürstner, Angew. Chem. Int. Ed. 2013, 52, 2794-2819.
18
19
a) A. Fürstner, Science 2013, 341, 1229713; b) H. Ehrhorn, M. Tamm, Chem. Eur. J. 2019, 25,
3190-3208; c) D. Lee, I. Valchkov, S. Y. Yun, J. K. Cha, Org. React. 2020, 102, 613-931; d) A.
Fürstner, P. W. Davies, Chem. Comun. 2005, 2307-2320.
20
For reviews on -acid catalyzed (spiro)ketal formation, see: a) D. Pflästerer, M. Rudolph, A. S. K.
Hashmi, Isr. J. Chem. 2018, 58, 622-638; b) R. Quach, D. P. Furkert, M. A. Brimble, Org. Biomol.
Chem. 2017, 15, 3098-3104.
21
22
B. Liu, J. K. De Brabander, Org. Lett. 2006, 8, 4907-4910.
6-exo-dig/7-endo-dig Cyclizations are known to compete, see: a) J. A. Goodwin, A. Aponik, Chem.
Commun. 2015, 51, 8730-8741; b) J. P. Costello, E. M. Ferreira, Org. Lett. 2019, 21, 9934-9939.
a) E. Reyes, U. Uria, L. Carrillo, J. L. Vicario, Tetrahedron 2014, 70, 9461-9484; b) P. A. Clarke, A.
T. Reeder, J. Winn, Synthesis 2009, 691-709.
23
24
25
a) Z. Meng, A. Fürstner, J. Am. Chem. Soc. 2019, 141, 805-809; b) K. Gebauer, A. Fürstner, Angew.
Chem. Int. Ed. 2014, 53, 6393-6396.
a) A. Fürstner, P. W. Davies, Angew. Chem. Int. Ed. 2007, 46, 3410-3449; b) A. Fürstner, Chem.
Soc. Rev. 2009, 38, 3208-3221; c) A. Fürstner, Acc. Chem. Res. 2014, 47, 925-938; d) A. Fürstner,
Angew. Chem. Int. Ed. 2018, 57, 4215-4233.
26
27
28
29
30
31
32
R. E. Claus, S. L. Schreiber, Org. Synth. 1986, 64, 150.
L. M. Geary, S. K. Woo, J. C. Leung, M. J. Krische, Angew. Chem. Int. Ed. 2012, 51, 2972-2976.
M. Holmes, L. A. Schwartz, M. J. Krische, Chem. Rev. 2018, 118, 6026-6052.
The stereochemistry was ascertained by Mosher ester analysis, see the SI.
For the sake of reproducibility, it is imperative to run the reaction in the presence of Na2SO4.
A. Arase, M. Hoshi, A. Mijin, S. Nishi, Synthesis 2004, 1814-1820.
For previous use in our laboratory, see: a) A. Fürstner, S. Flügge, O. Larionov, Y. Takjahashi, T.
Kubota, J. Kobayashi, Chem. Eur. J. 2009, 15, 4011-4029; b) J. Gagnepain, E. Moulin, C. Nevado,
M. Waser, A. Maier, G. Kelter, H.-H. Fiebig, A. Fürstner, Chem. Eur. J. 2011, 17, 6973-6984.
14
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
33
a) J. A. Cabezas, A. C. Oehlschlager, Synthesis 1999, 107-111; b) Y. Sasaki, Y. Horita, C. Zhong, M.
Sawamura, H. Ito, Angew. Chem. Int. Ed. 2011, 50, 2778-2782.
34
35
M. H. Junttila, O. O. E. Hormi, J. Org. Chem. 2009, 74, 3038-3047.
a) H. C. Kolb, M. S. Van Nieuwenhze, K. B. Sharpless, Chem. Rev. 1994, 94, 2483–2547; b) M. C.
Noe, M. A. Letavic, S. L. Snow, Org. React. 2005, 66, 109–625.
36
For the use of TESOTf/2,6-lutidine for acetal cleavage/diol protection, see: H. Fujioka, T. Okitsu,
Y. Sawama, N. Murata, R. Li, Y. Kita, J. Am. Chem. Soc. 2006, 128, 5930-5938.
C. Zhu, X. Shen, S. G. Nelson, J. Am. Chem. Soc. 2004, 126, 5352-5353.
37
38
39
40
S. Masamune, W. Choy, J. S. Petersen, L. R. Sita, Angew. Chem. Int. Ed. 1985, 24, 1-30.
S. Nahm, S. Weinreb, Tetrahedron Lett. 1981, 22, 3815-3818.
Compare: a) T. R. Vargo, J. S. Hale, S. G. Nelson, Angew. Chem. Int. Ed. 2010, 49, 8678-8681; b)
H. Yoshimura, J. Ishihara, S. Hatakeyama, Heterocycles 2018, 97, 702-707; c) P. Karier, F.
Ungeheuer, A. Ahlers, F. Anderl, C. Wille, A. Fürstner, Angew. Chem. Int. Ed. 2019, 48, 248-253.
Attempted direct opening of the -lactone with allylmetal reagents was unrewarding.
D. A. Evans, A. H. Hoveyda, J. Am. Chem. Soc. 1990, 112, 6447-6449.
41
42
43
44
Review: K. J. Ralston, A. N. Hulme, Synthesis 2012, 44, 2310-2324.
For applications from this laboratory, see ref. 40c and the following: a) L. Brewitz, J. Llaveria, A.
Yada, A. Fürstner, Chem. Eur. J. 2013, 19, 4532-4537; b) J. Willwacher, A. Fürstner, Angew. Chem.
Int. Ed. 2014, 53, 4217-4221; c) J. Willwacher, B. Heggen, C. Wirtz, W. Thiel, A. Fürstner, Chem.
Eur. J. 2015, 21, 10416-10430.
45
46
S. Inoki, T. Mukaiyama, Chem. Lett. 1990, 67-70.
a) G. Valot, C. S. Regens, D. P. O’Malley, E. Godineau, H. Takikawa, A. Fürstner, Angew. Chem.
Int. Ed. 2013, 52, 9534-9538; b) G. Valot, D. Mailhol, C. S. Regens, D. P. O’Malley, E. Godineau,
H. Takikawa, P. Philipps, A. Fürstner, Chem. Eur. J. 2015, 21, 2398-2408.
47
48
B. Morra, N. A. Morra, D. G. MacDonald, B. L. Pagenkopf, Synthesis 2020, 52, 847-852.
This notion is supported by the fact that the use of 3-nitrobenzaldehyde furnished the desired
regioisomer exclusively. The resulting product was O-TBS protected before the 3-nitrobenzoate
group was cleaved off with K2CO3 in MeOH. Unfortunately, ester cleavage was accompanied by
complete 1,3-silyl migration, thus thwarting this indirect approach; for details, see the SI.
15
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
49
a) For an advanced application in which recourse to 3-nitrobenzaldehyde had to be made
because indole-, pyrrole- or pyrazolecarbaldehyde failed, see: K. J. Ralston, H. C. Ramstadius, R.
C. Brewster, H. S. Niblock, A. N. Hulme, Angew. Chem. Int. Ed. 2015, 54, 7086-7090; see also: b)
P. D. Dorgon, J. Durrani, M. J. Cases-Thomas, A. N. Hulme, J. Org. Chem. 2010, 75, 7475-7478.
Y. Guo, J. Zhou, R. Gao, M. Zhao, J.-L. Yan, Z. Xu, S. Choi, T. Ye, Org. Lett. 2019, 21, 5471-5474.
D. Romo, S. D. Meyer, D. D. Johnson, S. L. Schreiber, J. Am. Chem. Soc. 1993, 115, 7906-7907.
a) A. Fürstner, C. Mathes, C. W. Lehmann, J. Am. Chem. Soc. 1999, 121, 9453-9454; b) A. Fürstner,
C. Mathes, C. W. Lehmann, Chem. Eur. J. 2001, 7, 5299-5317.
50
51
52
53
The “canopy catalyst” developed in parallel to this project also work well, even though the
reaction is much slower; for this new catalyst family, see: a) J. Hillenbrand, M. Leutzsch, E.
Yiannakas, C. P. Gordon, C. Wille, N. Nöthling, C. Copéret, A. Fürstner, J. Am. Chem. Soc. 2020,
142, 11279-11294; b) J. Hillenbrand, M. Leutzsch, A. Fürstner, Angew. Chem. Int. Ed. 2019, 58,
15690-15696.
54
As shown by the formation of 34, fairly low catalyst loadings can suffice; however, small-scale
reactions were best performed using 0.5 - 1 equiv. of Zeise’s salt and no further optimization
was attempted.
55
56
-Acid catalyzed transannular reactions in general are rare; for leading references see ref. 56
and the following: B. M. Trost, G. Dong, J. Am. Chem. Soc. 2010, 132, 16403-16416.
For examples from this laboratory, see ref. 44a, 46 and the following: a) S. Benson, M.-P. Collin,
A. Arlt, B. Gabor, R. Goddard, A. Fürstner, Angew. Chem. Int. Ed. 2011, 50, 8739-8744; b) A.
Ahlers, T. de Haro, B. Gabor, A. Fürstner, Angew. Chem. Int. Ed. 2016, 55, 1406-1411.
A. Zhdanko, M. E. Maier, Chem. Eur. J. 2014, 20, 1918-1930.
57
58
The same is true for a third model comprising the inverted tetrahydrofuran ring derived from
ent-18; the strong deviations at the inverted positions C2 an C5 are particularly indicative, see
the SI.
59
This cannot be taken for granted; for a case in which two diastereoisomeric macrolides gave
spectra that were almost indistinguishable at 600 MHz and the final decision as to which one
represents the natural product required biological profiling, see: a) J. Willwacher, N. Kausch-
Busies, A. Fürstner, Angew. Chem. Int. Ed. 2012, 51, 12041-12046; b) D. Mailhol, J. Willwacher,
16
This article is protected by copyright. All rights reserved.

10.1002/anie.202011472
Angewandte Chemie International Edition
N. Kausch-Busies, E. E. Rubitski, Z. Sobol, M. Schuler, M.-H. Lam, S. Musto, F. Loganzo, A.
Maderna, A. Fürstner, J. Am. Chem. Soc. 2014, 136, 15719-15729.
60
61
62
63
64
65
W. Yu, Y. Mei, Y. Kang, Z. Hua, Z. Jin, Org. Lett. 2004, 6, 3212-3214.
G. Stork, K. Zhao, Tetrahedron Lett. 1989, 30, 2173-2174.
E. J. Corey, M. Chaykovsky, J. Am. Chem. Soc. 1965, 87, 1353-1364.
G. Wang, E.-I. Negishi, Eur. J. Org. Chem. 2009, 1679-1682.
H. Azizian, C. Eaborn, A. Pidcock, J. Organomet. Chem. 1981, 215, 49-58.
A. Fürstner, J.-A. Funel, M. Trembley, L. C. Bouchez, C. Nevado, M. Waser, J. Ackerstaff, C. C.
Stimson, Chem. Commun. 2008, 2873-2875.
66
a) A. Fürstner, L. C. Bouchez, J.-A. Funel, V. Liepins, F.-H. Porée, R. Gilmour, F. Beaufils, D. Laurich,
M. Tamiya, Angew. Chem. Int. Ed. 2007, 46, 9265-9270; b) A. Fürstner, C. Nevado, M. Waser, M.
Tremblay, C. Chevrier, F. Teplý, C. Aissa, E. Moulin, O. Müller, J. Am. Chem. Soc. 2007, 129, 9150-
9161; c) A. Fürstner, J. Ackerstaff, Chem. Commun. 2008, 2870-2872; d) J. Gagnepain, E. Moulin,
A. Fürstner, A. Chem. Eur. J. 2011, 17, 6964-6972; e) J. Preindl, S. Schulthoff, C. Wirtz, J. Lingnau,
A. Fürstner, Angew. Chem. Int. Ed. 2017, 56, 7525-7530; f) C.-X. Zhuo, A. Fürstner, J. Am. Chem.
Soc. 2018, 140, 10514-10523.
67
68
a) K. A. Scheidt, H. Chen, B. C. Follows, S. R. Chemler, D. S. Coffey, W. R. Roush, J. Org. Chem.
1998, 63, 6436-6437; b) C. Aïssa, R. Riveiros, J. Ragot, A. Fürstner, J. Am. Chem. Soc. 2003, 125,
15512-15520.
For striking cases from our laboratory, see: a) S. Benson, M.-P. Collin, G. W. O’Neil, J. Ceccon, B.
Fasching, M. D. B. Fenster, C. Godbout, K. Radkowski, R. Goddard, A. Fürstner, Angew. Chem.
Int. Ed. 2009, 48, 9946-9950; b) A. Arlt, S. Benson, S. Schulthoff, B. Gabor, A. Fürstner, Chem.
Eur. J. 2013, 19, 3596-3608; c) T. Nagano, A. Fürstner, J. Am. Chem. Soc. 2007, 129, 1906-1907;
d) T. Nagano, J. Pospisil, G. Chollet, S. Schulthoff, V. Hickmann, E. Moulin, J. Herrmann, R. Müller,
A. Fürstner, Chem. Eur. J. 2009, 15, 9697-9706.
17
This article is protected by copyright. All rights reserved.