Divergent total synthesis of the antimitotic agent leiodermatolide

Article abstract of DOI:10.1002/anie.201206670

Subtle but distinctive: The stereostructure of the biologically highly promising antimitotic agent leiodermatolide was uncertain. A short, efficient, and flexible total synthesis based on ring-closing alkyne metathesis as the key step has now solved the p

Full text of DOI:10.1002/anie.201206670

Angewandte
Chemie
DOI: 10.1002/anie.201206670
Natural Product Synthesis
Divergent Total Synthesis of the Antimitotic Agent Leiodermatolide**
Jens Willwacher, Nina Kausch-Busies, and Alois Fꢀrstner*
The development of drugs based on microtubule toxins
marked major leaps forward in cancer chemotherapy.[1,2] All
approved spindle poisons exert their primary function by
interfering with tubulin dynamics upon binding to the protein.
Cell-cycle arrest is the immediate biological response, which
ultimately translates into cell death through a complex
signaling cascade. In clinical settings, however, severe side
effects as well as (multidrug) resistance set serious limitations
to the use of such antimitotic agents, and make the search for
more effective and selective drugs indispensable.[3]
It is against this backdrop that the recent discovery of
leiodermatolide has to be seen.[4,5] This complex macrolide of
marine origin exhibits remarkable, yet selective, cytotoxicity
in vitro, with IC50 values for the most sensitive human cancer
cell lines well below 10 nm.[6] This concentration also sufficed
to effect cell-cycle arrest of the highly responsive A549 lung
adenocarcinoma and PANC-1 pancreatic tumor cells at the
G2M phase transition, accompanied by abnormal spindle
formations in both cell types. Surprisingly though, leioder-
matolide exerted no noticeable effect on purified tubulin even
at much higher concentrations.[5] Therefore, this particular
natural product seems to operate through a mechanism
distinct from that of the established antimitotic agents such as
vincristine, colchicine, paclitaxel, epothilone, and discoder-
molide, to mention only the most prominent ones,[1,2] and
hence warrants close chemical, biochemical, and, possibly,
pharmacological inspection.
Leiodermatolide was isolated from a lithistid Leioderma-
tium sponge collected by submersible in the deep waters off
the Florida coastline (0.0011% of the wet weight).[5] Approxi-
mately tenfold lower concentrations were detected in speci-
mens harvested in the Bahamas,[7] whereas an earlier study on
Leiodermatium sponges collected off Palau had only given
structurally unrelated secondary metabolites of mixed poly-
ketide/nonribosomal peptide synthetase origin (“leioto-
lides”).[8,9]
A masterful investigation combining the predictive power
of modern computational tools with advanced NMR spectro-
scopic techniques allowed the constitution and much of the
stereostructure of leiodermatolide to be established.[5] How-
ever, the segregated stereoclusters within the macrolac-
tone[10,11] and the d-lactone terminus could not be correlated
with each other, thus leaving it open as to whether structure
1 or 2 represents the natural product (Scheme 1). The
absolute configuration of leiodermatolide also remained
unknown. Moreover, Maier and co-workers cautiously ques-
tioned whether the unusual axial attachment of the side chain
onto the d-lactone at the tertiary alcohol site C21 had been
correctly assigned by the isolation team.[12]
To resolve these open issues, we pursued an approach to
this rewarding target that was programmed for late-stage
divergence. If necessary for an unambiguous structure assign-
ment, the retrosynthetic analysis outlined in Scheme 1 allows
any conceivable stereoisomer of the terminal d-lactone
(fragment E) to be attached to a truncated macrolide synthon
of type A while forming the E,E diene unit of the side
chain.[10] Provided that A can be forged by ring-closing alkyne
metathesis (RCAM) as envisaged,[13,14] the macrocyclization
and the formation of the Z,Z diene become advantageously
aligned. In this context, it is important to note that the stereo-
and site-selective generation of Z,Z-configured 1,3-dienes by
olefin metathesis seems currently impossible even with the
most advanced catalysts.[15] The proposed strategic substitu-
tion of an alkene for an alkyne, as shown in Scheme 1,
however, allows the orthogonal reactivity of the different
p bonds through a metal alkylidyne catalyst to be harnessed,
and hence the site of ring closure to be reliably determined.
Moreover, as the configuration of the targeted alkene can be
determined at will by stereoselective semireduction of the
acetylene linkage, 1,3-dienes of any configuration—including
the generally rather labile Z,Z-configured ones—seem within
reach.[16]
In the leiodermatolide case, we opted to apply the
ꢀ
RCAM/semireduction tactics at the C10 C11 bond rather
ꢀ
than at the neighboring C12 C13 olefin, as this disconnection
arguably leads to more manageable building blocks. The
anticipated bonus of step economy aside, however, this
strategic decision bore considerable risk at the stage of ring
closure: in general, enynes tend to be less reactive than
nonconjugated alkynes, and bulky[17] propargylic substrates
have so far rarely succumbed to RCAM.[13] Both structural
elements are present in the envisaged cyclization precursor B.
At the outset of this project it was also unclear whether an
alkenyl halide is compatible with the available alkyne meta-
thesis catalysts, even though we were optimistic in view of the
excellent tolerance of the latest generation of metal alkyli-
dynes toward many different reactive functional groups.
These reservations notwithstanding, we pursued the syn-
thesis of the required building blocks as shown in Scheme 2.
The cheap malonate derivative 3 served as the point of
departure, which was advanced into 3-iodomethacrolein (6)
by following a literature procedure.[18] A subsequent boron-
mediated Masamune–Abiko reaction with donor 7[19] fur-
[*] Dipl.-Chem. J. Willwacher, Dr. N. Kausch-Busies, Prof. A. Fꢀrstner
Max-Planck-Institut fꢀr Kohlenforschung
45470 Mꢀlheim/Ruhr (Germany)
E-mail: fuerstner@kofo.mpg.de
[**] Generous financial support from the MPG and the Fonds der
Chemischen Industrie (Kekulꢁ stipend to J.W.) is gratefully
acknowledged. We thank the Analytical Departments of our Institute
for the excellent cooperation.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201206670.
Angew. Chem. Int. Ed. 2012, 51, 12041 –12046
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
12041

.
Angewandte
Communications
hyde 9, which was immediately
subjected to a Julia olefination
with the known sulfone derivative
10.[21] The enyne (Z/E > 24:1) thus
obtained was finally deprotected
with TBAF to give 11 in readiness
for esterification with the acid
sector 20.
The preparation of this second
key compound was also straightfor-
ward (Scheme 3). The necessary
syn,anti,syn-configured stereotet-
rad was attained by a Sn(OTf)2-
mediated aldol reaction of the
chiral ketone derivative 13[22] with
an excess of freshly prepared bu-
tynal (14)[23] followed by 1,3-anti-
reduction of the resulting product
15.[24] As expected, the subsequent
O-silylation occurred regioselec-
tively at the propargylic site, leav-
ing the C7-OH group open for
orthogonal protection as a MOM
acetal. This pattern ensured that the
distinctive carbamate group of leio-
Scheme 1. The two possible stereostructures of leiodermatolide and our retrosynthetic analysis for
isomer 1.
nished the anti-aldol product 8 in good yield and respectable
diastereoselectivity.[20] Silylation of the free OH group
followed by routine oxidation-state management gave alde-
Scheme 2. a) CHI3, NaH, Et2O, reflux, 99%; b) KOH, EtOH/H2O,
reflux, 72%; c) LiAlH4, Et2O, 08C!RT, 49%; d) MnO2, CH2Cl2, 98%;
e) 7, Cy2BOTf, Et3N, CH2Cl2, ꢀ788C!RT, 76%; f) TBSOTf, 2,6-dime-
thylpyridine, CH2Cl2, ꢀ108C; g) DIBAl-H, toluene, ꢀ788C, 83% (over
two steps); h) Dess–Martin periodinane, CH2Cl2, 08C; i) 10, KHMDS,
THF, ꢀ558C, 56% (>24:1) (over two steps); j) TBAF, THF, 08C, 99%.
Bn=benzyl, Cy=cyclohexyl, DIBAl-H=diisobutylaluminum hydride,
KHMDS=potassium hexamethyldisilazide, Mes=mesityl, TBAF =
tetra-n-butylammonium fluoride, TBS=tert-butyldimethylsilyl, Tf=tri-
fluoromethanesulfonyl.
Scheme 3. a) Bu2BOTf, Et3N, propanal, 97%; b) SO3·pyridine, CH2Cl2,
DMSO, Et3N, ꢀ158C, 88%; c) 14 (5 equiv), Sn(OTf)2, Et3N, CH2Cl2,
ꢀ208C, 55% (88% brsm); d) Me4NBH(OAc)3, HOAc, MeCN, ꢀ508C,
98%; e) TBSOTf, Et3N, CH2Cl2, ꢀ788C!08C, 89%; f) (MeO)NH-
Me·HCl, AlMe3, THF, 08C!RT, 90%; g) MOMCl, (iPr)2NEt, DMF,
=
508C, 89%; h) MeMgCl, Et2O, 08C, 97%; i) CH2 CHMgCl, THF,
ꢀ788C!RT, 87%; j) PBr3, pyridine, Et2O, 08C; k) EtOAc, LDA, CuI,
THF, ꢀ1108C!ꢀ308C, 63% (over two steps); l) Me3SiOK, Et2O,
quant. brms=based on recovered starting material, LDA=lithium
diisopropylamide, MOM=methoxymethyl.
12042 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 12041 –12046

Angewandte
Chemie
dermatolide could later be instal-
led at the correct position, which is
not possible once the macrocycle
has been formed.[25] Successive
=
addition of MeMgCl and H2C
CHMgCl at the Weinreb amide
terminus of 17 furnished the ter-
tiary alcohol 18. Since the pro-
jected Claisen rearrangement of
the derived acetate failed under
a variety of conditions,[26] the rear-
rangement step and the necessary
chain extension were dissected
into two separate operations. To
this end, 18 was first exposed to
PBr3 to give the rather labile allylic
bromide 19, which reacted cleanly
with the lithium enolate of ethyl
acetate in the presence of CuI at
low temperature.[27] The saponifi-
cation of the resulting ester by
using TMSOK was slow but effec-
tive, and provided the required
acid segment 20 in essentially
quantitative yield.
Scheme 4. a) Bu2BOTf (2 equiv), Et3N, propanal, Et2O, ꢀ788C, 74% (d.r.=11:1); b) Ac2O, Et3N,
DMAP cat., CH2Cl2, 08C, 82%; c) LiHMDS, THF, ꢀ788C, 83%; d) (1S)-24, THF, 08C, 86% combined
yield (23/25=1:5.5), 73% (pure 25); e) (1R)-24, THF, ꢀ788C, 88% combined yield (23/25=7.5:1);
f) 26, 27 (5 mol%), CH2Cl2, reflux, 81% (E/Z>19:1). DMAP=4-dimethylaminopyridine, LiHMDS =
lithium hexamethyldisilazide.
The pseudoaxial orientation
of the side chain actually makes
the d-lactone building block E
more synthetically demanding
than its size may suggest.
Attempted
Reformatsky-type
intramolecular
reactions
had previously been shown to
afford exclusively the wrong
array, with the side chain in
the
equatorial
position.[12]
Therefore, we pursued a differ-
ent and pleasingly short
approach (Scheme 4). The
chosen route was inspired by
a literature report on the reduc-
tion of the readily available
cyclic b-ketoester 22 with the
borane adduct [(tBuNH2)·BH3]
in acidic methanol, in which the
incoming hydride was strictly
delivered axially.[28] Attempts to
translate this process to reac-
tions of 22 with a variety of
different carbon nucleophiles,
however, originally met with
limited success. In most cases,
the product yields were zero or
remained disappointingly low
because of the pronounced eno-
lization of the ketone group of
22. Only an allylindium reagent
generated in situ[29] as well as 9-
Scheme 5. a) EDCI·HCl, DMAP, CH2Cl2, 08C, 89%; b) 34 (40 mol%), CH2Cl2/toluene, 1008C, 72%;
c) 1. 28, [Pd(PPh3)4] (20 mol%), Tl(OEt), THF/H2O (3:1); 2. aq HCl (0.5m), tert-butyl methyl ether, 55%;
d) TBAF, THF, 4 ꢂ MS, 08C, 85%; e) Zn(Cu/Ag), THF/H2O/MeOH, 508C, 89%; f) Cl3CC(O)NCO,
CH2Cl2, ꢀ788C, then Al2O3, 84%; g) Me2BBr, CH2Cl2, ꢀ908C!ꢀ788C, 61%. EDCI=N’-(3-dimethyl-
allyl-9-BBN[30] gave good con- aminopropyl)-N-ethylcarbodiimide, MS=molecular sieves.
Angew. Chem. Int. Ed. 2012, 51, 12041 –12046
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 12043

.
Angewandte
Communications
versions, but afforded the wrong isomer 23 as the major
product. All attempts to override this inherent bias of the
indium reagent with the aid of chiral modifiers failed to afford
a workable solution. Gratifyingly though, replacement of 9-
allyl-9-BBN by its chiral relative 24 was more successful.[31] In
this case, the addition is reagent-controlled, as evident from
the data shown in Scheme 4. Specifically, (1S)-24 afforded the
required diastereomer 25 with appreciable selectivity (d.r. =
5.5:1) and good yield.[32] The terminal alkene of 25 partici-
pated well in a cross-metathesis reaction with the vinyl MIDA
boronate ester 26 catalyzed by the second-generation Grubbs
carbene 27,[33] which provided the required Suzuki donor 28,
virtually as a single isomer (E/Z > 19:1).
carbamate could be accomplished with Me2BBr at low
temperature to give the targeted product 1 in good yield.[42,43]
To obtain the diastereomeric product 2, the same end
game was executed with similar yields but using the enantio-
meric d-lactone fragment ent-28 (for details, see the Support-
ing Information). Interestingly, the 13C NMR spectra
(150 MHz, CD2Cl2) of 1 and 2 are almost indistinguishable
and did not allow us to decide which of the two isomers
1
represents leiodermatolide.[44] Likewise, their H NMR spec-
tra (600 MHz) are virtually identical, except for the region
between 2.2 and 2.5 ppm, wherein subtle differences can be
noticed that are deemed sufficiently characteristic (Figure 1).
The assembly stage of the project commenced with the
esterification of 11 and 20 to give the requisite cyclization
precursor 29 (Scheme 5). Treatment of this substrate with the
molybdenum alkylidyne complex 35[34] endowed with bulky
triphenylsilanolate ligands resulted only in the fast and clean
formation of an acyclic dimer. This outcome was independent
of whether the reaction was performed at ambient temper-
ature or at 1008C. Gratifyingly though, the catalyst formed
in situ upon activation of complex 34 with CH2Cl2, as
previously described by our research group,[35] furnished the
desired cyclic monomer 30 in 72% yield, with only traces of
the dimer being detectable in the crude mixture.[36] The
transformation required harsh conditions, a high catalyst
loading, and an unusually long reaction time; this fact likely
reflects the challenge derived—collectively—from the strain
of the incipient polyunsaturated ring, the severe steric
hindrance about the reacting propargylic site,[17] and the
electronic toll to be paid for the enyne motif. However, the
chemoselectivity was excellent, with the strict differentiation
of the catalyst between the reactive p systems of the triple
bonds and the inert p bonds of three different types of alkenes
present in the substrate being remarkable and enabling at the
same time.[13]
Since the alkenyl iodide had survived the metathesis step
uncompromised, the stage was set for the attachment of the
side chain by a Suzuki–Miyaura reaction with boronate 28.
Despite the outstanding track record of this venerable
transformation,[37] this particular cross-coupling actually
turned out to be challenging and required considerable
optimization. Best results were obtained using Tl(OEt) in
aqueous THF,[38] followed by an acidic work up to regenerate
the d-lactone ring, which was partly opened in the basic
medium. Attempts to reduce the alkyne in the resulting
polyunsaturated product 31 to the corresponding Z alkene
failed under a variety of conditions as long as the TBS group
on the C9-OH was present.[17] In contrast, the propargylic
alcohol generated by cleavage of the silyl ether, despite being
otherwise very unstable, could be cleanly converted into the
10Z-configured allylic congener 32 on exposure to Zn(Cu/
Ag) as the preferred reducing agent.[39] This compound was
treated with Cl3CC(O)NCO, and the primary adduct hydro-
lyzed upon contact with alumina to release the still missing
carbamate group.[40,41] After several unsuccessful attempts, we
were pleased to find that the delicate final cleavage of the
residual MOM group in the presence of the sensitive allylic
Figure 1. Region of the 1H NMR spectra (600 MHz, CD2Cl2) that shows
subtle, but distinctive, differences for the isomers 1 (red) and 2 (blue),
and allows for a correlation with the published spectrum of the natural
product (black).
Direct comparison of the published spectrum of leioderma-
tolide (black)[5] with that of compound 1 (red) shows a perfect
match, whereas the pattern of 2 (blue) is slightly dissimilar.[45]
24
The recorded optical rotation of 1 (½aꢁD ¼ꢀ74.3, c = 0.41,
MeOH) implies that the absolute configuration of leioder-
24
matolide (½aꢁD ¼ꢀ84.2, c = 0.34, MeOH)[5] must be as drawn.
Although we have no authentic sample of the natural product
at our disposal, we believe that formula 1 properly depicts the
molecular structure of leiodermatolide.
In summary, we accomplished the first total synthesis of
the structurally challenging and biologically highly promising
12044 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 12041 –12046

Angewandte
Chemie
[17] Literature precedent shows that protecting groups on the C9-
antimitotic agent leiodermatolide by pushing the limits of
alkyne metathesis to new frontiers. The chosen approach is
short, efficient, and flexible, and should hence qualify for
a synthesis-driven mapping of the pharmacophore of this lead
compound. In chemical terms, this case study corroborates
our previous conclusion that RCAM is particularly well-
suited for applications to polyunsaturated targets, where
olefin metathesis often finds its limits.[46] At the same time,
this investigation shows that total synthesis remains an
indispensable tool for structure elucidation, even in the age
of spectroscopy, when dealing with compounds that contain
spatially segregated stereoclusters.[47]
OH group seriously impede reactions at C11, which is the
projected site of ring closure, see Ref. [10].
[18] R. Baker, J. L. Castro, J. Chem. Soc. Perkin Trans. 1 1990, 47 – 65.
[19] a) A. Abiko, J.-F. Liu, S. Masamune, J. Am. Chem. Soc. 1997,
119, 2586 – 2587; b) T. Inoue, J.-F. Liu, D. C. Buske, A. Abiko, J.
Org. Chem. 2002, 67, 5250 – 5256; c) A. Fꢂrstner, J. Ruiz-Caro,
H. Prinz, H. Waldmann, J. Org. Chem. 2004, 69, 459 – 467.
[20] This reaction has precedent, see D. Menche, J. Hassfeld, J. Li, K.
Mayer, S. Rudolph, J. Org. Chem. 2009, 74, 7220 – 7229;
however, we invariably obtained a somewhat lower selectivity
(anti/syn = 13:1 versus > 20:1 in the literature).
[21] a) C. Bonini, L. Chiummiento, V. Videtta, Synlett 2005, 3067 –
3070; b) C. Bonini, L. Chiummiento, V. Videtta, Synlett 2006,
2079 – 2082; c) see also Ref. [16a].
[22] D. A. Evans, H. P. Ng, J. S. Clark, D. L. Rieger, Tetrahedron
1992, 48, 2127 – 2142.
Received: August 17, 2012
Published online: October 18, 2012
[23] Prepared by oxidation of 2-butynol with MnO2.
[24] D. A. Evans, K. T. Chapman, E. M. Carreira, J. Am. Chem. Soc.
1988, 110, 3560 – 3578.
[25] Model studies showed that carbamate formation by monoacy-
lation of macrocyclic C7/C9-diol derivatives results in poor
selectivity and favors the undesired isomer, cf. Ref. [10,11].
[26] Similarly unsuccessful attempts were reported in Ref. [10].
[27] I. Kuwajima, Y. Doi, Tetrahedron Lett. 1972, 13, 1163 – 1166.
[28] K. Hinterding, S. Singhanat, L. Oberer, Tetrahedron Lett. 2001,
42, 8463 – 8465.
[29] Review: J. Augꢁ, N. Lubin-Germain, J. Uziel, Synthesis 2007,
1739 – 1764.
[30] G. W. Kramer, H. C. Brown, J. Org. Chem. 1977, 42, 2292 – 2299.
[31] E. Canales, K. G. Prasad, J. A. Soderquist, J. Am. Chem. Soc.
2005, 127, 11572 – 11573.
[32] Since the NOESY data for 23 and 25 did not allow for an
unambiguous assignment, the stereochemistry at the branching
site C21 was established by X-ray diffraction; the structures are
contained in the Supporting Information.
[33] B. E. Uno, E. P. Gillis, M. D. Burke, Tetrahedron 2009, 65, 3130 –
3138.
[34] a) J. Heppekausen, R. Stade, R. Goddard, A. Fꢂrstner, J. Am.
Chem. Soc. 2010, 132, 11045 – 11057; b) J. Heppekausen, R.
Stade, A. Kondoh, G. Seidel, R. Goddard, A. Fꢂrstner, Chem.
Eur. J. 2012, 18, 10281 – 10299.
[35] a) A. Fꢂrstner, C. Mathes, C. W. Lehmann, Chem. Eur. J. 2001, 7,
5299 – 5317; b) A. Fꢂrstner, C. Mathes, C. W. Lehmann, J. Am.
Chem. Soc. 1999, 121, 9453 – 9454.
[36] Monitoring of the reaction showed that the formation of 30
proceeds via initial formation of a dimer; catalysts 35 and 34/
CH2Cl2 hence seem to differ in their ability to activate this
dimeric intermediate.
[37] A. Suzuki, Angew. Chem. 2011, 123, 6854 – 6869; Angew. Chem.
Int. Ed. 2011, 50, 6722 – 6737.
[38] a) J. Uenishi, J.-M. Beau, R. W. Armstrong, Y. Kishi, J. Am.
Chem. Soc. 1987, 109, 4756 – 4758; b) S. A. Frank, H. Chen, R. K.
Kunz, M. J. Schnaderbeck, W. R. Roush, Org. Lett. 2000, 2,
2691 – 2694.
[39] W. Boland, N. Schroer, C. Sieler, M. Feigel, Helv. Chim. Acta
1987, 70, 1025 – 1040.
Keywords: alkyne metathesis · antitumor agents ·
natural products · structure eludication · total synthesis
.
[1] Anticancer Agents from Natural Products (Eds.: G. M. Cragg,
D. G. I. Kingston, D. J. Newman), CRC, Boca Raton, FL, 2005.
[2] K.-H. Altmann, J. Gertsch, Nat. Prod. Rep. 2007, 24, 327 – 357.
[3] K. E. Gascoigne, S. S. Taylor, J. Cell Sci. 2009, 122, 2579 – 2585.
[4] A. E. Wright, J. K. Reed, J. Roberts, R. E. Longley, U.S. Pat.
Appl. US 20080033035A1 20080207, 2008.
[5] I. Paterson, S. M. Dalby, J. C. Roberts, G. J. Naylor, E. A.
Guzmꢀn, R. Isbrucker, T. P. Pitts, P. Linley, D. Divlianska, J. K.
Reed, A. E. Wright, Angew. Chem. 2011, 123, 3277 – 3281;
Angew. Chem. Int. Ed. 2011, 50, 3219 – 3223.
[6] The following specific IC50 values were disclosed in Ref. [5]:
human A549 lung adenocarcinoma: 3.3 nm; PANC-1 pancreatic
carcinoma: 5.0 nm; DLD-1 colorectal carcinoma: 8.3 nm; NCI/
ADR-Res. ovarian adenocarcinoma: 233 nm; P388 murine
leukemia: 3.3 nm; Vero monkey kidney cells: 211 nm.
[7] See the Supporting Information of Ref. [5].
[8] J. S. Sandler, P. L. Colin, M. Kelly, W. Fenical, J. Org. Chem.
2006, 71, 7245 – 7251; correction: J. S. Sandler, P. L. Colin, M.
Kelly, W. Fenical, J. Org. Chem. 2006, 71, 8684.
[9] A. Larivꢁe, J. B. Unger, M. Thomas, C. Wirtz, C. Dubost, S.
Handa, A. Fꢂrstner, Angew. Chem. 2011, 123, 318 – 323; Angew.
Chem. Int. Ed. 2011, 50, 304 – 309.
[10] A truncated macrocyclic core of type A terminated by an alkenyl
bromide was recently prepared, but no attempts to attach the
side chain through cross-coupling were reported, see I. Paterson,
T. Paquet, S. M. Dalby, Org. Lett. 2011, 13, 4398 – 4401.
[11] For a non-natural diastereomer of the macrocycle, see C. Rink,
V. Navickas, M. E. Maier, Org. Lett. 2011, 13, 2334 – 2337.
[12] V. Navickas, C. Rink, M. E. Maier, Synlett 2011, 191 – 194.
[13] A. Fꢂrstner, Angew. Chem. 2012, DOI: 10.1002/ange.201204513;
Angew. Chem. Int. Ed. 2012, DOI: 10.1002/anie.201204513.
[14] a) A. Fꢂrstner, G. Seidel, Angew. Chem. 1998, 110, 1758 – 1760;
Angew. Chem. Int. Ed. 1998, 37, 1734 – 1736; b) A. Fꢂrstner,
P. W. Davies, Chem. Commun. 2005, 2307 – 2320.
[15] For a discussion, see D. Gallenkamp, A. Fꢂrstner, J. Am. Chem.
Soc. 2011, 133, 9232 – 9235 and the references therein.
ˇ
´
[40] P. Kocovsky, Tetrahedron Lett. 1986, 27, 5521 – 5524.
[41] The tert-OH group at C21 did not interfere with the carbamoy-
lation at C9, although the isolation team had found this site to
easily participate in Mosher ester formations.
[16] For the synthesis of E,Z-configured 1,3-dienes by alkyne meta-
thesis/semireduction, see a) K. Micoine, A. Fꢂrstner, J. Am.
Chem. Soc. 2010, 132, 14064 – 14066; b) A. Fꢂrstner, L. Turet,
Angew. Chem. 2005, 117, 3528 – 3532; Angew. Chem. Int. Ed.
2005, 44, 3462 – 3466; c) A. Fꢂrstner, D. DeSouza, L. Turet,
M. D. B. Fenster, L. Parra-Rapado, C. Wirtz, R. Mynott, C. W.
Lehmann, Chem. Eur. J. 2007, 13, 115 – 134; d) A. Fꢂrstner, M.
Bonnekessel, J. T. Blank, K. Radkowski, G. Seidel, F. Lacombe,
B. Gabor, R. Mynott, Chem. Eur. J. 2007, 13, 8762 – 8783.
[42] a) Y. Quindon, H. E. Morton, C. Yoakim, Tetrahedron Lett.
1983, 24, 3969 – 3972; b) Y. Guindon, C. Yoakim, H. E. Morton,
J. Org. Chem. 1984, 49, 3912 – 3920.
[43] In contrast, a variety of commonly used reagents for the cleavage
of MOM acetals gave complex mixtures, containing significant
amounts of products formed by elimination of the carbamate
Angew. Chem. Int. Ed. 2012, 51, 12041 –12046
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 12045

.
Angewandte
Communications
moiety (anhydr. HCl, TMSCl/(nBu)4NBr, BF3·Et2O, [Ph3C]BF4
(with and without 2,6-di-tert-butylpyridine), ZnBr2/BuSH, (cat-
echol)BX (X = Cl, Br), 9-I-9-BBN).
Int. Ed. 2012, 51, 6929 – 6933; b) V. Hickmann, A. Kondoh, B.
Gabor, M. Alcarazo, A. Fꢂrstner, J. Am. Chem. Soc. 2011, 133,
13471 – 13480; c) K. Lehr, R. Mariz, L. Leseurre, B. Gabor, A.
Fꢂrstner, Angew. Chem. 2011, 123, 11575 – 11579; Angew. Chem.
Int. Ed. 2011, 50, 11373 – 11377.
[44] The NMR spectra of both isomers are contained in the
Supporting Information.
[45] The spectra of synthetic 1 were recorded at different molarities
to make sure that these delicate differences are not caused by
concentration effects.
[47] For another study from our research group illustrating this
notion, see A. Fꢂrstner, M. Albert, J. Mlynarski, M. Matheu, E.
DeClercq, J. Am. Chem. Soc. 2003, 125, 13132 – 13142.
[46] For other instructive cases, see a) W. Chaładaj, M. Corbet, A.
Fꢂrstner, Angew. Chem. 2012, 124, 7035 – 7039; Angew. Chem.
12046 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 12041 –12046