Synthesis of the C3-C18 fragment of amphidinolides G and H

Article abstract of DOI:10.1021/ol071024e

A synthesis of an amphidinolides G and H C3-C18 subunits is reported. The C10-C18 segment 4 was prepared by a Negishi cross-coupling, whereas the synthesis of the C3-C9 fragment 5 employed an asymmetric cyanosilylation as the key step. The two segments were coupled by lithiation of iodide 4 and trapping of the anion with amide 5. The allylic epoxide moiety could be synthesized from the protected anti- mesylate 22.

Full text of DOI:10.1021/ol071024e

ORGANIC
LETTERS
2007
Vol. 9, No. 16
3001-3004
Synthesis of the C3−C18 Fragment of
Amphidinolides G and H
Andreas F. Petri,† John S. Schneekloth, Jr.,‡ Amit K. Mandal,† and
Craig M. Crews*,†,‡,§
Departments of Chemistry, Molecular, Cellular, and DeVelopmental Biology, and
Pharmacology, Yale UniVersity, New HaVen, Connecticut 06520-8103
craig.crews@yale.edu
Received May 2, 2007
ABSTRACT
A synthesis of an amphidinolides G and H C3
whereas the synthesis of the C3 C9 fragment 5 employed an asymmetric cyanosilylation as the key step. The two segments were coupled by
lithiation of iodide 4 and trapping of the anion with amide 5. The allylic epoxide moiety could be synthesized from the protected anti- mesylate 22.
−C18 subunits is reported. The C10−C18 segment 4 was prepared by a Negishi cross-coupling,
−
The amphidinolides are a structurally diverse group of
bioactive secondary metabolites isolated from the symbiotic
marine dinoflagellate Amphidinium sp.1 Amphidinolides G
(1) and H (2) are polyketide-based 27- and 26-membered
macrolides, respectively, that were first isolated in 1991 by
Kobayashi et al.2 These two compounds putatively arise from
the same seco acid. The gross structures of 1 and 2 have
been elucidated primarily by means of 2D NMR data, where-
as the absolute stereochemistry was determined on the basis
of X-ray diffraction analysis and degradation3 (Figure 1).
Both amphidinolides G (1) and H (2) were shown to be
among the most potent cytotoxic amphidinolides. Whereas
amphidinolide H (2) exhibits extremely potent cytotoxic
activity against L1210 murine lymphoma and KB epidermoid
carcinoma cells (IC50 ) 0.00048 and 0.00052 µg/mL,
respectively), amphidinolide G (1) showed approximately a
10-fold decrease in activity with IC50 values of 0.0054 and
0.0046 µg/mL, respectively.4 Amphidinolide H (2) has been
implicated in binding to actin subdomain 4 as a potential
mode of action.5
† Department of Molecular, Cellular, and Developmental Biology.
‡ Department of Chemistry.
§ Department of Pharmacology.
Figure 1. Amphidinolides G (1) and H (2).
(1) For reviews, see: (a) Kobayashi, J.; Ishibashi, M. Chem. ReV. 1993,
93, 1753. (b) Chakraborty, T.; Das, S. Curr. Med. Chem.: Anti-Cancer
Agents 2001, 1, 131. (c) Kobayashi, J.; Shimbo, K.; Kubota, T.; Tsuda, M.
Pure Appl. Chem. 2003, 75, 337. (d) Kobayashi, J.; Tsuda, M. Nat. Prod.
Rep. 2004, 21, 77. (e) Kobayashi, J.; Kubota, T. J. Nat. Prod. 2007, 70,
451.
Because of their remarkable biological activity and chal-
lenging structure, amphidinolides G and H represent an
attractive target for synthetic efforts. Although several
(2) Kobayashi, J.; Shigemori, H.; Ishibashi, M.; Yamasu, T.; Hirota, H.;
Sasaki, T. J. Org. Chem. 1991, 56, 5221.
(3) Kobayashi, J.; Shimbo, K.; Sato, M.; Shiro, M.; Tsuda, M. Org. Lett.
2000, 2, 2805.
(4) Kobayashi, J.; Shimbo, K.; Sato, M.; Tsuda, M. J. Org. Chem. 2002,
67, 6585.
10.1021/ol071024e CCC: $37.00
© 2007 American Chemical Society
Published on Web 07/07/2007

syntheses of fragments of amphidinolides G and H and the
structurally related amphidinolide B have been published,1,6,7
to our knowledge no total synthesis of either of these
amphidinolides has been accomplished so far.
Scheme 1. Synthesis of Allenic Acetate 10
Our proposed synthetic route to amphidinolide H (2)
divides the structure into the three major fragments 3, 4, and
5, allowing a convergent assembly of the molecule (Figure
2). To couple these fragments, we could establish the bond
by treatment with ethynylmagnesium bromide. Alcohol 8 was
then homologated under Crabbe´’s conditions9 to form allenol
9 in 68% yield. Acetylation of 9 was accomplished by
treatment with Ac2O and DMAP in pyridine at 45 °C to
provide allenic acetate 10 in 76% yield.
Having established a synthetic route to access the allenic
acetate 10, we turned our attention to the synthesis of the
1,3-diene moiety. Our laboratory has previously demonstrated
that the (E)-1,3-diene moiety of amphidinolide B can be
synthesized from an allenic acetate precursor.7f,10 When the
allenic acetate 10 was treated under the reaction conditions
previously described by Ba¨ckvall11 (LiI, Pd(OAc)2, AcOH,
40 °C), the iodide-mediated SN2′ reaction gave the vinyl
iodide 11 in 80% yield and a low E/Z-selectivity of 2:1. We
were pleased to find that the iodide 11 could be obtained in
76% yield and an E/Z-ratio of 10:1 by performing the
reaction in the absence of Pd catalyst at room temperature
(Scheme 2). Moving forward, we next concentrated our
Figure 2. Retrosynthetic analysis of amphidinolide H (2).
between C9 and C10 by nucleophilic addition of an alkyl-
lithium species derived from fragment 4 into Weinreb amide
5. We propose an aldol coupling to construct the C18-C19
bond and a Wittig olefination with commercially available
ylide 6 across C2 and C3. Finally, a macrolactonization
should complete the carbon skeleton of amphidinolide G (1)
or amphidinolide H (2).
Scheme 2. Synthesis of the C10-C18 Fragment 4
Our synthetic efforts toward the C10-C18 fragment 4
began with ketone 7, which was prepared from (-)-
pseudoephedrine propionamide under the conditions devel-
oped by Myers8 (Scheme 1). Ketone 7 was further elaborated
to propargylic alcohol 8 (as a 1:1 mixture of diastereomers)
(5) (a) Usui, T.; Kazami, S.; Dohmae, N.; Mashimo, Y.; Kondo, H.;
Tsuda, M.; Terasaki, A. G.; Ohashi, K.; Kobayashi, J.; Osada, H. Chem.
Biol. 2004, 11, 1269. (b) Saito, S.-Y.; Feng, J.; Kira, A.; Kobayashi, J.;
Ohizumi, Y. Biochem. Biophys. Res. Commun. 2004, 320, 961.
(6) Amphidinolide H: (a) Chakraborty, T. K.; Suresh, V. R. Tetrahedron
Lett. 1998, 39, 7775. (b) Chakraborty, T. K.; Suresh, V. R. Tetrahedron
Lett. 1998, 39, 9109. (c) Liesener, F. P.; Kalesse, M. Synlett 2005, 2236.
(d) Liesener, F. P.; Jannsen, U.; Kalesse, M. Synthesis 2006, 2590. (e)
Forment´ın, P.; Murga, J.; Carda, M.; Marco, J. A. Tetrahedron: Asymmetry
2006, 17, 2938. (f) Deng, L.; Ma, Z.; Zhang, Y.; Zhao, G. Synlett 2007,
87.
efforts to establish the C12-C13 bond by reacting vinyl
iodide 11 with a suitable coupling partner. Gratifyingly,
iodide 11 could be reacted with the commercially available
organozinc species 12 under Negishi-type cross-coupling
(7) For representative papers regarding synthetic work on amphidinolide
B, see: (a) Chakraborty, T. K.; Suresh, V. S. Chem. Lett. 1997, 6, 565. (b)
Eng, H. M.; Myles, D. C. Tetrahedron Lett. 1999, 40, 2275. (c) Ishiyama,
H.; Takemura, T.; Tsuda, M.; Kobayashi, J. Tetrahedron 1999, 55, 4583.
(d) Cid, M. B.; Pattenden, G. Tetrahedron Lett. 2000, 41, 7373. (e) Zhang,
W.; Carter, R. G. Org. Lett. 2005, 7, 4209. (f) Mandal, A. K.; Schneekloth,
Jr. J. S.; Crews, C. M. Org. Lett. 2005, 7, 3645. (g) Mandal, A. K.;
Schneekloth, J. S., Jr.; Kuramochi, K.; Crews, C. M. Org. Lett. 2006, 8,
427. (h) Gopalarathnam, A.; Nelson, S. G. Org. Lett. 2006, 8, 7.
(8) (a) Myers, A. G.; McKinstry, L. J. Org. Chem. 1996, 61, 2428. (b)
Myers, A. G.; Yang, B. H.; Chen, H.; McKinstry, L.; Kopecky, D. J.;
Gleason, J. L. J. Am. Chem. Soc. 1997, 119, 6496.
(9) Searles, S.; Li, Y.; Nassim, B.; Lopes, M.-T. R.; Tran, P. T.; Crabbe´,
P. J. Chem. Soc., Perkin Trans. 1 1984, 747.
(10) Schneekloth, J. S., Jr.; Pucheault, M.; Crews, C. M. Eur. J. Org.
Chem. 2007, 40.
(11) Horva´th, A.; Ba¨ckvall, J.-E. J. Org. Chem. 2001, 66, 8120.
3002
Org. Lett., Vol. 9, No. 16, 2007

conditions12 to give the ester 13 in 92% yield. No isomer-
ization of the diene moiety was observed under these reaction
conditions. Ester 13 could then be converted into alcohol
14 by reduction with DIBALH (92%). It was determined
that alcohol 14 decomposes to an uncharacterized mixture
of side products upon exposure to silica gel or even
neutralized silica gel. Therefore, the crude alcohol 14 was
carried forward without further purification. Functionalization
of alcohol 14 to the iodide 4 proved to be nontrivial. All
attempts to convert alcohol 14 directly to iodide 4 left the
starting material unchanged or resulted in decomposition of
the substrate.13 Therefore, alcohol 14 was brominated (CBr4,
PPh3, 2,6-lutidine, 99%) and a subsequent Finkelstein
reaction (NaI, acetone, ∆) introduced the iodine in excellent
yield (92%) and furnished the C10-C18 fragment 4. Alkyl
iodide 4 could be stored at -20 °C for 3 months without
any noticeable decomposition.
We proposed to synthesize the C3-C9 fragment 5 via an
asymmetric cyanosilylation of known aldehyde 1514 as key
step. In our initial attemps, we investigated the cyanosily-
lation method developed by Uang15 (which left the aldehyde
15 unchanged) and the method of Pu,16 which gave the
protected cyanohydrin 17 in 30% yield and 69% ee. We were
pleased to find that we could access the TMS-protected cyano-
hydrin 17 in 75% chemical yield and 80% ee17 by employing
the chiral ligand (+)-16 and Ti(OiPr)4 as described by Feng
and co-workers18 (Scheme 3). Exposure of 17 to HCl in
assigned by using the modified Mosher method,19 which
revealed the (S)-configuration of R-hydroxyester 18. Amida-
tion of ester 18 using the Merck procedure20 ((MeO)MeNH‚
HCl, iPrMgCl) afforded Weinreb amide 19 in 84% yield.
The choice of protecting group for the hydroxy function
at C8 was crucial. As it was necessary to utilize a chelating
protecting group, initial studies with a PMB ether or a MOM
ether were carried out (not shown). Unfortunately, later
substrates containing the 1,3-diene unit of amphidinolides
G and H were prone to decomposition upon treatment with
DDQ or under acidic conditions, presumably by way of
isomerization. In order to circumvent this issue, we chose
to protect the C8 hydroxy function of 19 as an MTM
(methylthiomethyl) ether21 (DMSO, AcOH, Ac2O, 60%), thus
completing the synthesis of the C3-C9 fragment 5.
With both fragments 4 and 5 securely in hand, we
envisioned a nucleophilic addition into a Weinreb amide as
the key coupling step to combine the two fragments. The
lithium-halogen exchange to convert the alkyl iodide 4 to
the alkyllithium species 20 proved to be a challenging trans-
formation. In our initial experiments, we performed the
lithium-halogen exchange of iodide 4 using the standard
protocol,22 which involves treatment of 4 with 2.2 equiv of
t-BuLi at -78 °C and subsequent stirring at room temper-
ature for 1 h to decompose excess t-BuLi. Under these condi-
tions, the lithiation was found to be lacking reproducibility,
and the yield of alkyllithium species 20 was usually <40%.23
The low yields of 20 could potentially be caused by an
intramolecular cyclization of the alkyllithium moiety into the
1,3-diene system, as such cyclizations are known to occur
at room temperature.24 Since the cyclization of unsaturated
organolithiums can be suppressed at low temperatures, we
were able to circumvent this problem by modifying the
lithiation protocol.25 When we treated iodide 4 with 1.8 equiv
of t-BuLi at -78 °C and stirred the reaction for 10 min at
-40 °C, lithium halogen exchange to 20 occurred smoothly
and subsequent coupling with Weinreb amide 5 at -78 °C
yielded the ketone 21 in 72% yield26 (Scheme 4). The
chelation-controlled reduction of ketone 21 (LiI/LiAlH4,
Et2O, anti:syn 8:1) provided the anti alcohol in 87% yield
as a single diastereomer after flash chromatography.27
Mesylation of the secondary alcohol (MsCl, Et3N, rt) afforded
Scheme 3. Synthesis of the C3-C9 Fragment 5
(18) Li, Y.; He, B.; Qin, B.; Feng, X.; Zhang, G. J. Org. Chem. 2004,
69, 7910.
(19) (a) Dale, J. A.; Mosher, H. S. J. Am. Chem. Soc. 1973, 95, 512. (b)
Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, H. J. Am. Chem. Soc. 1991,
113, 4092.
(20) Williams, J. M.; Jobson, R. B.; Yasuda, N.; Marchesini, G.; Dolling,
U.-H.; Grabowski, E. J. J. Tetrahedron Lett. 1995, 36, 5461.
(21) Pojer, P. M.; Angyal, S. J. Aust. J. Chem. 1978, 31, 1031.
(22) Bailey, W. F.; Punzalan, E. R. J. Org. Chem. 1990, 55, 5404.
(23) The yield of alkyllithium species 20 was determined by trapping
20 with amide 5 and isolation of the ketone 21 thus formed.
(24) For reviews about the cyclization of usaturated organolithiums,
see: (a) Clayden, J. Organolithiums: SelectiVity for Synthesis; Pergamon
Press: New York, New York, 2002; Vol. 3, pp 293-335. (b) Mealy, M.
J.; Bailey, W. F. J. Organomet. Chem. 2002, 646, 59.
(25) Bailey, W. F.; Jiang, X. Tetrahedron 2005, 61, 3183.
(26) Because the amide 5 had an ee of 80%, the coupling yielded ∼10%
of the (8R)-epimer of ketone 21, which could be separated by column
chromatography at this stage.
ethanol produced the R-hydroxyester 18 in 89% yield. At
this stage, the stereochemistry of the C8-OH group was
(12) For a recent review on Negishi-type couplings, see: Negishi, E.;
Hu, Q.; Huang, Z.; Qian, M.; Wang, G. Aldrichimica Acta 2005, 38, 71.
(13) Other conditions explored were PPh3/I2/imidazole; PPh3/CI4; NaI/
heat; and MsCl/NaI.
(14) Marshall, J. A.; Jiang, H. J. Org. Chem. 1999, 64, 971.
(15) Uang, B. J.; Fu, I. P.; Hwang, C. D.; Chang, C. W.; Yang, C. T.;
Hwang, D. R. Tetrahedron 2004, 60, 10479.
(16) Qin, Y. C.; Liu, L.; Pu, L. Org. Lett. 2005, 7, 2381.
(17) The enantiomeric excess of this reaction was determined by Mosher
ester analysis of the R-hydroxyester 18.
(27) The relative stereochemistry was confirmed after converting the
alcohol into an acetonide by NOESY NMR spectroscopy (see Supporting
Information).
Org. Lett., Vol. 9, No. 16, 2007
3003

Scheme 4. Coupling of Fragment 4 with Fragment 5 and Completion of the C3-C18 Subunit 23
22 in 91% yield. MTM deprotection was accomplished under
epoxide, two important characteristic structural features of
amphidinolides G and H have been synthesized successfully.
Efforts toward the eventual total synthesis of 2 are currently
being pursued in our laboratory.
mild alkylating conditions (MeI, NaHCO3, aqueous acetone,
68%), and subsequent treatment of the hydroxy mesylate with
LiHMDS afforded the allylic epoxide 23 in 80% yield.
Epoxide 23 corresponds to the fully functionalized C3-C18
fragment of amphidinolides G (1) and H (2).
Acknowledgment. We gratefully acknowledge financial
support from the NIH (GM062120). A.F.P. thanks the
Deutsche Forschungsgemeinschaft for a postdoctoral fel-
lowship. J.S.S. thanks the American Chemical Society,
Division of Medicinal Chemistry, and Aventis Pharmaceu-
ticals for a predoctoral fellowship.
In summary, we have described an efficient synthetic route
to the C3-C18 subunit 23 of amphidinolides G and H. Our
studies demonstrate that the 1,3-diene moiety of these natural
products can be assembled by an SN2′ reaction of an allenic
acetate 10. We have also demonstrated that a Negishi-type
cross-coupling between a vinyl iodide 11 and an alkylzinc
species 12 can be employed to assemble the C12-C13 bond.
The two advanced fragments 4 and 5 could be combined to
ketone 21 by a lithiation/nucleophilic addition sequence.
Furthermore, we have shown that the allylic epoxide moiety
of amphidinolides G (1) and H (2) can be effectively
synthesized from the protected anti mesylate 22. With the
effective assembly of the 1,3-diene moiety and the allylic
Supporting Information Available: Experimental pro-
cedures and characterization data for all new compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL071024E
3004
Org. Lett., Vol. 9, No. 16, 2007