Synthesis of the macrolide core of migrastatin

Article abstract of DOI:10.1016/S0040-4039(02)02281-5

A concise and efficient synthesis of the macrolactone core of migrastatin, a new natural product with potent anticancer properties, has been achieved. The key features of our synthetic strategy encompass a Lewis acid catalyzed diene aldehyde condensation

Full text of DOI:10.1016/S0040-4039(02)02281-5

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 9039–9042
Synthesis of the macrolide core of migrastatin†
Christoph Gaula and Samuel J. Danishefskya,b,
*
aLaboratory for Bioorganic Chemistry, Sloan-Kettering Institute for Cancer Research, 1275 York Avenue, New York, NY 10021,
USA
bDepartment of Chemistry, Columbia University, Havemayer Hall, 3000 Broadway, New York, NY 10027, USA
Received 13 October 2002; accepted 14 October 2002
Abstract—A concise and efficient synthesis of the macrolactone core of migrastatin, a new natural product with potent anticancer
properties, has been achieved. The key features of our synthetic strategy encompass a Lewis acid catalyzed diene aldehyde
condensation (LACDAC) to install the three contiguous stereocenters and the trisubstituted (Z)-double bond of migrastatin, and
a (E)-selective ring-closing metathesis (RCM) to construct the macrocycle. © 2002 Published by Elsevier Science Ltd.
Migrastatin (1, Fig. 1) is a novel macrolide natural
product, isolated from a cultured broth of Streptomyces
sp. MK929-43F1 by Imoto and co-workers in 2000.1,2
Recently, it was shown by Kosan Bioscience researchers
that cultures of Streptomyces platensis (strain NRRL
18993) also produce migrastatin.3
These potent anticancer properties render migrastatin
an interesting target for total synthesis. A successful
undertaking of this sort would supply adequate
amounts of material to investigate its biological mode
of action, and even more importantly, could provide
means for preparing structural analogs of 1. From such
a focused collection of congeners might arise agents
with improved biological profiles.
Migrastatin displays a remarkable inhibitory effect on
the migration of human tumor cells.4 The suppression
of tumor cell migration is of great interest, potentially
as a model for a therapeutic approach to the treatment
of tumor metastasis. Furthermore, migrastatin selec-
tively inhibits the anchorage-independent growth of
human small cell lung carcinoma Ms-1 cells.4
In 2002, the relative and absolute configuration of
migrastatin was determined by X-ray crystal structure
analysis of a derivative of 1.5 Migrastatin was revealed
to be a 14-membered lactone with a glutarimide side
chain. The molecule contains five stereogenic centers
and three double bonds. Our general synthetic strategy
toward the total synthesis of migrastatin envisaged the
preparation of key fragment C7ꢀC13 via a Lewis acid
catalyzed diene aldehyde condensation (LACDAC).
For the construction of the macrocycle, we envisioned
ester bond formation, followed by ring-closing metathe-
sis (RCM) to set up the C6ꢀC7 double bond. In this
letter, we describe a rapid and efficient synthesis of the
macrolide core of migrastatin following these ideas.
Happily, the solution for most of the stereochemical
issues takes advantage of chemistry first promulgated in
our laboratories 20 years earlier.
Figure 1. Structure of migrastatin.
The assembly of the C7ꢀC13 subunit began with the
preparation of a-methoxy,b-siloxy aldehyde 4 (Scheme
1). Selective silylation of the primary hydroxyl group of
(S)-3-benzyloxy-1,2-propanediol 2,6 methylation of the
secondary alcohol, and reductive cleavage of the benzyl
ether yielded 3, which was oxidized to aldehyde 4 under
Swern conditions.
Keywords: migrastatin; tumor cell migration; natural product
macrolide; LACDAC; ring-closing metathesis.
* Corresponding author. Tel.: +1-212-639-5501; fax: +1-212-772-8691;
e-mail: s-danishefsky@ski.mskcc.org
† We dedicate this paper on the occasion of his 65th birthday to
Professor Dieter Seebach for his many contributions to science.
0040-4039/02/$ - see front matter © 2002 Published by Elsevier Science Ltd.
PII: S0040-4039(02)02281-5

9040
C. Gaul, S. J. Danishefsky / Tetrahedron Letters 43 (2002) 9039–9042
Scheme 1. Synthesis of the C7ꢀC13 fragment 8.
It is known that a-chelation control in Lewis acid
mediated reactions of aldehydes of type 4 with carbon
nucleophiles can be achieved, while competing b-chela-
tion is suppressed by the bulky silylether protecting
group.7 Hence, reaction of aldehyde 48 and diene 5
under the influence of TiCl4, followed by cyclization
with camphorsulfonic acid (CSA), yielded the a-chela-
tion controlled dihydropyrone product 6 (Scheme 1).9
The cyclocondensation allowed the construction of the
three contiguous stereocenters of the macrolide and set
the stage for establishing the trisubstituted (Z)-alkene
C11ꢀC12. Treatment of cycloadduct 6 with NaBH4 and
CeCl3·7H2O (Luche reduction)10 led to the correspond-
ing reduced compound, which underwent a Ferrier
rearrangement in aqueous acidic THF to produce lactol
7.11 Reductive opening of lactol 7 with LiBH4 afforded
diol 8 with the desired (Z)-olefin in 55% overall yield
from dihydropyrone 6. We were able to obtain single
crystals of compound 8 (mp 101–102°C) by crystalliza-
tion from ethyl acetate. X-Ray analysis led to a decisive
structural confirmation, revealing the relative configura-
tion of the three stereogenic centers and the geometry
of the double bond to be as proposed on the basis of
our precedents.12 The structure is depicted in Fig. 2.
Protection of the secondary hydroxyl group as a MOM
ether and cleavage of the TBDPS group with HF·py
afforded intermediate 11 in 79% yield. A simple two
step sequence (oxidation of the primary alcohol with
Dess–Martin
periodinane
(DMP)
and
Tebbe
olefination16 of the resulting aldehyde) provided access
to metathesis precursor 12. Anticipating that the risk of
competitive participation of the electron-poor C2ꢀC3
and the sterically hindered C11ꢀC12 double bonds in
the metathesis step is rather low, compound 12 was
subjected to 20 mol% Grubbs Ru–dihydroimidazolyl-
idene catalyst17 in refluxing toluene (0.5 mM).18 Grati-
fyingly, the migrastatin macrolide core 13 was gener-
ated under these conditions19 in 50% yield as a single
olefinic isomer (the desired (E)-congener).20 Notewor-
thy, treatment of 12 with the first generation Grubbs
catalyst (Cy3P)2Cl2RuꢁCHPh in refluxing CH2Cl2 led
exclusively to the dimeric product derived from cross
metathesis of the terminal double bond of the acyl
moiety.
With key intermediate 8 in hand, we explored the
viability of an RCM approach (Scheme 2),13 even
before addressing issues that are associated with the
incorporation of the glutarimide-containing side chain.
Thus, acid chloride 9 was prepared by treatment of the
copper dianion of crotonic acid with allyl bromide,14
followed by the reaction of the resultant acid with
oxalyl chloride.15 Selective acylation of the primary
hydroxyl group of 8 with acid chloride 9 proceeded
smoothly in the presence of DMAP. Interestingly, use
of triethylamine rather than DMAP led to an acylation
product in which the C2ꢀC3 double bond had moved
out of conjugation into the b,g-position (possibly via
vinylketene formation).
Figure 2. ORTEP plot of the X-ray crystal structure of 8.

C. Gaul, S. J. Danishefsky / Tetrahedron Letters 43 (2002) 9039–9042
9041
Scheme 2. Synthesis of the macrolide core 13 of migrastatin.
In conclusion, we have developed a direct and efficient
synthesis of the macrolactone core of migrastatin utiliz-
ing a LACDAC/RCM combination as the key feature.
Having demonstrated that macrocyclization can be
achieved via RCM and with the versatile intermediate 8
in hand, we have now begun efforts to adapt our
synthesis to accomodate incorporation of the glutarim-
ide-containing side chain. The results of this ongoing
investigation will be reported in due course.
6. (S)-3-Benzyloxy-1,2-propanediol 2 is commercially avail-
able (Fluka, Aldrich), but only at a high cost. Compound
2 can be easily prepared from inexpensive starting materi-
als. See: (a) Kitaori, K.; Furukawa, Y.; Yoshimoto, H.;
Otera, J. Tetrahedron 1999, 55, 14381–14390; (b) Xiang,
G.; McLaughlin, L. W. Tetrahedron 1998, 54, 375–392.
7. Reetz, M. T.; Kessler, K. J. Org. Chem. 1985, 50, 5434–
5436.
8. The TBDPS group actually serves a dual purpose: It does
not only prevent competing b-chelation, but is also a
protecting group that is sufficiently stable toward acidic
hydrolysis (see the use of CSA in the following
transformations).
Acknowledgements
9. For chelation-controlled cyclocondensations of a-alkoxy
aldehydes with synergistically activated dienes, see: Dan-
ishefsky, S. J.; Pearson, W. H.; Harvey, D. F.; Maring, C.
J.; Springer, J. P. J. Am. Chem. Soc. 1985, 107, 1256–
1268.
Support for this research was provided by the National
Institutes of Health (AI 16943). Postdoctoral Fellow-
ship support is gratefully acknowledged by C.G.
(Deutscher Akademischer Austauschdienst, DAAD).
We thank Drs. Jon Njardarson and Don Coltart for
fruitful discussions. Dr. George Sukenick (NMR Core
Facility, Sloan-Kettering Institute) is acknowledged for
NMR and mass spectral analyses.
10. Luche, J. L.; Gemal, A. L. J. Am. Chem. Soc. 1979, 101,
5848–5849.
11. Ferrier, R. J. J. Chem. Soc. 1964, 5443–5449.
12. Crystallographic data (excluding structural factors) for
compound 8 have been deposited with the Cambridge
Crystallographic Data Centre (CCDC) as deposition no.
CCDC 195014.
References
13. The reaction sequence described in Scheme 2 was actually
conducted starting from compound ent-8. For our early
studies toward the total synthesis of migrastatin it was
more convenient to commence the synthesis with the
1. Nakae, K.; Yoshimoto, Y.; Sawa, T.; Homma, Y.;
Hamada, M.; Takeuchi, T.; Imoto, M. J. Antibiot. 2000,
53, 1130–1136.
known aldehyde ent-4 (prepared from
D-mannitol: Nico-
2. Nakae, K.; Yoshimoto, Y.; Ueda, M.; Sawa, T.; Taka-
hashi, Y.; Naganawa, H.; Takeuchi, T.; Imoto, M. J.
Antibiot. 2000, 53, 1228–1230.
3. Woo, E. J.; Starks, C. M.; Carney, J. R.; Arslanian, R.;
Cadapan, L.; Zavala, S.; Licari, P. J. Antibiot. 2002, 55,
141–146.
laou, K. C.; Piscopio, A. D.; Bertinato, P.; Chakraborty,
T. K.; Minowa, N.; Koide, K. Chem. Eur. J. 1995, 1,
318–333), leading to diol ent-8. Later, we addressed this
problem and developed an efficient synthesis of aldehyde
4 (see Scheme 1).
14. Katzenellenbogen, J. A.; Crumrine, A. L. J. Am. Chem.
4. Takemoto, Y.; Nakae, K.; Kawatani, M.; Takahashi, Y.;
Naganawa, H.; Imoto, M. J. Antibiot. 2001, 54, 1104–
1107.
5. Nakamura, H.; Takahashi, Y.; Naganawa, H.; Nakae,
K.; Imoto, M.; Shiro, M.; Matsumura, K.; Watanabe,
H.; Kitahara, T. J. Antibiot. 2002, 55, 442–444.
Soc. 1976, 98, 4925–4935.
15. Ahmar, M.; Duyck, C.; Fleming, I. J. Chem. Soc., Perkin
Trans. 1 1998, 2721–2732.
16. Tebbe, F. N.; Parshall, G. W.; Reddy, G. S. J. Am.
Chem. Soc. 1978, 100, 3611–3613.

9042
C. Gaul, S. J. Danishefsky / Tetrahedron Letters 43 (2002) 9039–9042
17. Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org.
Lett. 1999, 1, 953–956.
18. For one of the early applications of RCM in complex
natural product synthesis, see: Xu, Z.; Johannes, C. W.;
Salman, S. S.; Hoveyda, A. H. J. Am. Chem. Soc. 1996,
118, 10926–10927.
H), 5.15 (dd, 1H, J=15.4, 8.6, C7-H), 4.90 (d, 1H,
J=6.5, C17-H), 4.75 (d, 1H, J=6.5, C17-H), 4.73 (d,
1H, J=15.9, C13-H), 4.63 (d, 1H, J=15.9, C13-H), 3.55
(app. t, 1H, J=8.6, C8-H), 3.46 (s, 3H, C18-H), 3.34
(dd, 1H, J=8.6, 1.7, C9-H), 3.25 (s, 3H, C16-H), 3.09–
3.02 (m, 1H, C10-H), 2.48–2.15 (m, 4H, C4-H, C5-H),
1.68 (s, 3H, C14-H), 0.89 (d, 3H, J=6.9, C15-H); 13C
NMR (100 MHz, CDCl3): l 165.55, 149.97, 132.82,
130.33, 129.99, 127.49, 122.15, 99.05, 85.99, 83.42, 65.69,
56.51, 56.24, 32.56, 32.22, 30.21, 22.66, 13.67; MS (ESI):
347 ([M+Na]+).
19. Yamamoto, K.; Biswas, K.; Gaul, C.; Danishefsky, S. J.,
manuscript in preparation.
20. Spectroscopic data of 13 (C18H28O5): 1H NMR (400
MHz, CDCl3): l 6.81 (dt, 1H, J=15.6, 7.4, C3-H), 5.75
(d, 1H, J=15.6, C2-H), 5.61–5.53 (m, 2H, C6-H, C11-