ORGANIC LETTERS
Synthesis of the A,B,C-Ring System of Hexacyclinic Acid†
2004 Vol. 6, No. 22 3889-3892
Timo Stellfeld,‡ Ulhas Bhatt,§ and Markus Kalesse* Institut fu¨r Organische Chemie, UniVersita¨t HannoVer, Schneiderberg 1 B, 30167 HannoVer, Germany
[email protected] Received July 6, 2004
ABSTRACT
The synthesis of the A,B,C-ring system (2) of hexacyclinic acid (1) is achieved starting from a selective Diels−Alder reaction followed by vinyl cuprate addition. The diastereoselective reduction of the ketone carbonyl at C16 could be achieved with LiAlH4. An intramolecular Michael addition established the ring system stereoselectively, providing access to the selective generation of 9 out of the 14 stereocenters of hexacyclinic acid.
Hexacyclinic acid (1) was isolated through chemical screening by Zeeck, Ho¨fs, and Walker from Streptomyces cellulosae subsp. (stem S 1013) with the aid of their OSMAC approach (OSMAC ) one strain many compounds).1 It reveals a unique hexacyclic carbon skeleton that is paralleled by (-)-FR182877 (3), isolated from Streptomyces sp. No9885 by Fujisawa Pharmaceutical Company in 1998.2 The two compounds exhibit striking similarities in their architecture but have major structural and stereochemical differences. The differences in AB and BC ring fusion sites especially demand different synthetic strategies (Figure 1). The C13-C17 cis fusion in hexacyclinic acid is contrasted by a trans substitution pattern in (-)-FR182877.3 Similarly,
the C10-C18 trans fusion in hexacyclinic acid is contrasted by a cis substitution pattern in (-)-FR182877. Other significant differences are the acetylation at C14 and the carboxylic acid moiety at C25 in hexacyclinic acid; (-)-FR182877 features a free hydroxyl group instead of an acetate at C14 and a C25 methyl group instead of a carboxylic acid moiety, respectively. Probably the most important difference is located at C5 and C20: hexacyclinic acid exhibits a hydroxyl group at C5, whereas (-)-FR182877 has a C5-C20 double bond at this position. It has been proposed that the Michael acceptor serves as the pharma-
†
Dedicated to Prof. A. Zeeck on the occasion of his 65th birthday. Current address: Freie Universita¨t Berlin, Takustr. 3, 14195 Berlin, Germany. § Current address: Albany Molecular Research, Inc., 21 Corporate Circle, P.O. Box 15098, Albany, NY 12212-5098. (1) Ho¨fs, R.; Walker, M.; Zeeck, A. Angew. Chem., Int. Ed. 2000, 39, 3258. (2) (a) Sato, B.; Muramatsu, H.; Miyauchi, M.; Hori, Y.; Takase, S.; Hino, M.; Hashimoto, S.; Terano, H. J. Antibiot. 2000, 53, 123. (b) Sato, B.; Nakaijama, H.; Hori, Y.; Takase, S.; Hino, M.; Hashimoto, S.; Terano, H. J. Antibiot. 2000, 53, 204. (c) Yoshimura, S.; Sato, B.; Kinoshita, B.; Takase, S.; Terano, H. J. Antibiot. 2000, 53, 615. Revised structure: (d) Yoshimura, S.; Sato, B.; Kinoshita, B.; Takase, S.; Terano, H. J. Antibiot. 2002, 55, C1. (3) Hexacyclinic acid numbering is used for both compounds. ‡
10.1021/ol048720o CCC: $27.50 Published on Web 10/01/2004
© 2004 American Chemical Society
Figure 1. Comparison of the retrosynthetic analyses of hexacyclinic acid (1) and FR182877 (3).
Scheme 1.
Synthesis of the Diels-Alder Precursor
Figure 2. Proposed transition states for the Diels-Alder reaction.
cophoric group in (-)-FR182877. The absence of such an electrophilic fragment in hexacyclinic acid may be responsible for the 2 orders of magnitude weaker antitumor activity.4e Furthermore, acetylation at C14 cannot be ruled out as the reason for the observed diminished activity of hexacyclinic acid. Taking these observations into consideration, it becomes obvious that there is a vital interest to provide synthetic access to these compounds not only to deconvolute the biological activity but also to provide compounds with an altered and eventually optimized biological profile. As in the syntheses of (-)-FR182877,4 we envisioned that a Diels-Alder reaction probably acts in the biosynthesis to build up the AB-carbon skeleton of hexacyclinic acid. As a consequence of these retrosynthetic considerations, the functionality at C16 has to be incorporated as a ketone carbonyl group if one considers the transition state shown in Figure 2 as the pivotal control element during the DielsAlder reaction. As a consequence, selective reduction at a later stage in synthesis is a synthetic detour that is not necessary in the synthesis of (-)-FR182877. A second, even more severe obstacle was the fact that precursors having substitution at C18 (10, Scheme 1) would not undergo intramolecular Diels-Alder reaction. This problem could be solved by using the acetylenic precursor 9 and introducing the required side chain through a sequence of cuprate addition and cross metathesis. To close the third ring we considered an intramolecular Michael addition as the method of choice, in contrast to (-)FR182877, where a hetero-Diels-Alder reaction was applied.5 The synthesis began with racemic alcohol 4 that was TBSprotected and cleaved through ozonolysis.6 Aldehyde 5 could (4) (a) Vanderwal, C. D.; Vosburg, D. A.; Weiler, S.; Sorensen, E. J. J. Org. Chem. 1936, 1, 645. (b) Vanderwal, C. D.; Vosburg, D. A.; Sorensen, E. J. Org. Lett. 2001, 3, 407. (c) Vosburg, D. A.; Vanderwal, C. D.; Sorensen, E. J. J. Am. Chem. Soc. 2003, 125, 5393. (d) Evans, D. A., Starr, J. T. Angew. Chem., Int. Ed. 2002, 41, 1787. (e) Evans, D. A., Starr, J. T. J. Am. Chem. Soc. 2003, 125, 13531. (5) Another strategy was reported by Clarke et al. in the synthesis of a model DEF-ring core of hexacyclinic acid: (a) Clarke, P. A.; Grist, M.; Ebden, M. Wilson, C. Chem. Commun. 2003, 1560. Roush et al. reported on a Morita-Baylis-Hillman approach for the formation of the C-ring of FR182877, which is closely related to the Michael aldol reaction: (b) Methot, J. L.; Roush, W. R Org. Lett. 2003, 5, 4223. 3890
be extended with phosphonocrotonate 6 using LDA as a base. DIBAl-H reduction followed by oxidation with manganese dioxide established the double-unsaturated aldehyde 7 in 64% yield over five steps (Scheme 1). An Evans aldol reaction is followed by transformation into the Weinreb amide 8. From here, TBS protection, DIBAl-H reduction, and addition of acetylene Grignard followed by reoxidation with the DessMartin periodinane provided the Diels-Alder precursor 9. In contrast to the C18-substituted triene 10, which did not cyclize under various conditions, acetylenic compound 9 underwent clean cyclization to hexadiene 11 in toluene at 75 °C (Scheme 2).
Scheme 2.
Diels-Alder Reaction and Subsequent Functionalization
The subsequent addition of vinyl cuprate at -78 °C established the desired configuration at C18 through an attack (6) Sugai, T.; Katho, O.; Ohta, H. Tetrahedron 1995, 51, 11987.
Org. Lett., Vol. 6, No. 22, 2004
Scheme 3.
Cross-Metathesis and Michael Addition
Figure 3. Catalysts for the cross-metathesis reaction.
from the convex face and furnished the final syn addition product 12. The selective reduction at C16 thus became the pivotal transformation. Here, the hydride source had to approach the carbonyl carbon from the most hindered side, the interior of the concave face. A variety of reagents were tested for diastereoselective reduction, but most reducing reagents, and in particular bulky regents such as L-selectride, gave the undesired isomer with good selectivity, inversion of which under Mitsunobu conditions could not be achieved. Only reduction with LiAlH4 provided the required configuration at C16 with >20:1 selectivity. Next, a cross metathesis is required to install the conjugated ester functionality for an intramolecular Michael addition. Several metathesis catalysts have been used in order to induce this transformation, and it turned out that the Hoveyda-Grubbs second-generation catalyst 187 and Grela’s p-nitro-substituted analogue 198 (Figure 3) could facilitate this transformation in equal yields (80%, Scheme 3). TBDPS-protection and hydrolysis of one TBS group was followed by TPAP oxidation in 79% yield over three steps to yield ketoester 15. With compound 15 in our hands, we were in a position to perform the crucial step to close the C-ring via a Michael reaction. Several reagents and conditions were tried for this transformation without success.9 Eventually, thermodynamic TMS-enol ether formation was achieved with TMSI and (TMS)2NH in 1,2-dichloroethane as the solvent10 and instantly provided the cyclobutane adduct 16 through a subsequent aldol reaction in 68% yield.11 Finally, cyclobutane 16 was opened with TBAF in THF to establish the A,B,C-ring system (2) of hexacyclinic acid with nine contiguous chiral centers in 20 steps. Cyclization (7) Kingsbury, J. S.; Harrity, J. P. A.; Bonitatebus, P. J.; Hoveyda, A. H. J. Am. Chem. Soc. 1999, 121, 791. (8) Grela, K.; Harutyunyan, S.; Michrowska, A. Angew. Chem., Int. Ed. 2002, 41, 4038. (9) For example, treatment of ketone 15 with secondary amines like proline, pyrrolidine, or diethylamine did not give any reaction. (10) Ihara, M.; Taniguchi, T.; Makita, K.; Takano, M.; Ohnishi, M.; Taniguchu, N.; Fukumoto, K.; Kabuto, C. J. Am. Chem. Soc. 1993, 115, 8107. (11) 1H NMR indicates the formation of four stereoisomers of the cyclobutane intermediate 16. The geometry of the intermediate enol ether is suspected to be a mixture, with the (E)-isomer as the major isomer. For the formation of a thermodynamic enol ether with TMSI/(TMS)2NH, see also: Sedrani, R.; Krallen, J.; Cabrejas, L. M. M.; Papageorgiou, C. D.; Senia, F.; Rohrbach, S.; Wagner, D.; Thai, B.; Eme, A.-M. J.; France, J.; Oberer, L.; Rihs, G.; Zenke, G.; Wagner, J. J. Am. Chem. Soc. 2003, 125, 3849.
Org. Lett., Vol. 6, No. 22, 2004
product 2 was obtained as an inseparable mixture (4:1) of two diastereomers. NOESY experiments and coupling constants give strong evidence that the major isomer exhibits the desired stereochemistry, matching the title compound 2 (Figure 4). The NOESY cross-peak between H20 and H16
Figure 4. Assignment of the stereochemistry confirms the formation of the desired diastereomer 2 (the numbering corresponds to the natural product).
supports the correct stereochemistry at C19. Assuming that the configuration between the four- and five-membered rings in 16 is cis since the trans configuration would produce high ring strain, the new stereocenters in 2 must be cis-configured. For C8, we have no direct evidence, but the coupling constants of H9b also indicate H9b-H8 trans configuration. Interestingly, the intramolecular Michael addition parallels recent results on the biosynthesis of hexacyclinic acid from the Zeeck group. Through feeding experiments, it could be shown that the tetrahydropyran oxygen O6 is not derived from acetate or propionate but from external water (Figure (12) Meyer, S. W. Dissertation, Georg-August-Universita¨t zu Go¨ttingen, Go¨ttingen, Germany, 2003. 3891
Further work toward the completion of the total synthesis of hexacyclinic acid continues and will be reported in due course. Acknowledgment. This work was supported by the DFG and the Fonds der Chemischen Industrie. We thank Prof. Zeeck for insight into the biosynthesis and A. Stelmakh for helpful discussion. A sample of the ruthenium catalyst 19, provided by Dr. Karol Grela, is gratefully acknowledged. Figure 5. Formation of the C-ring seems to take place through a Prins-like reaction in the biosynthesis.
5).12 A vinylogous Prins reaction is therefore proposed as the biosynthetic route, in contrast to the hetero Diels-Alder reaction in (-)-FR182877.
3892
Supporting Information Available: Experimental procedures and spectroscopic data for 11-13, 15, 16, 2, and intermediates, as well as a copy of the NOESY spectrum of 2. This material is available free of charge via the Internet at http://pubs.acs.org. OL048720O
Org. Lett., Vol. 6, No. 22, 2004
