Convergent total synthesis of the michellamines

1090 J. Org. Chem. 1998, 63, 1090-1097 A Convergent Total Synthesis of the Michellamines| Gerhard Bringmann,*,† Roland Götz,† Paul A. Keller,† Rainer Walter,† Michael R. Boyd,‡ Fengrui Lang,§ Alberto Garcia,§ John J. Walsh,§ Imanol Tellitu,§ K. Vijaya Bhaskar,§ and T. Ross Kelly*,§ Institut für Organische Chemie, Universität Würzburg, Am Hubland, D-97074 Würzburg, Germany, Laboratory of Drug Discovery Research and Development, National Cancer Institute, Building 1052, Room 121, Frederick, Maryland 21702-1201, and Department of Chemistry, Eugene F. Merkert Chemistry Center, Boston College, Chestnut Hill, Massachusetts 02167-3860 Received August 12, 1997 A convergent total synthesis of the anti-HIV michellamines (1) is described. The tetraaryl skeleton of the michellamines was constructed by formation first of the inner (nonstereogenic) biaryl axis and subsequently of the two other (stereogenic) axes in a highly convergent manner. The key transformation features a double Suzuki-type cross-coupling reaction between binaphthalene ditriflate 26 and isoquinolineboronic acid 35. Ditriflate 26 is synthesized in six steps starting from diene 6 and 2,6-dibromobenzoquinone (9) in 21% overall yield. For large scale production of 26, a substantially shortened version of an existing procedure for the preparation of bisnaphthoquinone 13 was also developed, which allows for the preparation of 13 from benzoquinone and diene 6 in five steps and 67% overall yield. Binaphthoquinone 13 was subsequently converted into ditriflate 26 in three steps and 67% overall yield. By the described synthetic strategy, michellamines A (1a) and B (1b) are produced (1a:1b ) 1:2.5) in 24.6% overall yield from diene 6. Curiously, none of the nonnaturally occurring atropoisomer 1c is formed. Introduction The lack of effective drugs for the treatment of AIDS led the United States National Cancer Institute to initiate in the late 1980s a major effort to discover novel HIV (human immunodeficiency virus) inhibitory agents from natural sources.1 In 1991, Boyd et al.2,3 reported the isolation of two anti-HIV alkaloids, michellamines A and B, from the tropical plant Ancistrocladus korupensis native to Cameroon. Michellamine B, the more potent and abundant of the naturally occurring michellamines, aborted viral replication and virus-induced cell killing across an unusually broad range of HIV strains and isolates in diverse human host-cell types.3 By a combination of spectroscopic and degradative studies2-6 michellamines A and B were assigned structures 1a and 1b, respectively, with the relative and absolute configurations shown. Initially, the third pos| Part 102 in the series Acetogenic Naphthylisoquinoline Alkaloids (from the University of Würzburg). For part 101, see: Bringmann, G.; Saeb, W.; Wenzel, M.; François, G.; Schlauer, J. Pharm. Pharmacol. Lett., submitted. Part 47 in the series: HIV-Inhibitory Natural Products (from the National Cancer Institute). For part 46, see: Hallock, Y. F.; Cardellina, J. H., II; Schäffer, M.; Bringmann, G.; Boyd, M. R. BioMed. Chem. Lett., submitted. † Institute für Organische Chemie. ‡ National Cancer Institute. § Boston College. (1) Boyd, M. R. In AIDS Etiology, Diagnosis, Treatment and Prevention; De Vita, V. T., Jr., Hellman, S., Rosenberg, S. A., Eds.; Lippincott: Philadelphia, 1988, p 305. (2) Manfredi, K. P.; Blunt, J. W.; Cardellina, J. H., II; MacMahon, J. B.; Pannell, L. L.; Gordon, M. C.; Boyd, M. R. J. Med. Chem. 1991, 34, 3402. (3) Boyd, M. R.; Hallock, Y. F.; Cardellina, J. H., II; Manfredi, K. P.; Blunt, J. W.; MacMahon, J. B.; Buckheit, R. W.; Bringmann, G.; Schäffer, M.; Cragg, G. M.; Thomas, D. W.; Jato, J. G. J. Med. Chem. 1994, 37, 1740. (4) Bringmann, G. In The Alkaloids; Brossi, A., Ed.; Academic Press: New York, 1986; Vol. 29, pp 141-184. (5) Bringmann, G.; Zagst, R.; Schäffer, M.; Hallock, Y. F.; Cardellina, J. H., II; Boyd, M. R. Angew. Chem., Int. Ed. Engl. 1993, 32, 1190. (6) Bringmann, G.; Gulden, K.-P.; Hallock, Y. F.; Manfredi, K. P.; Cardellina, J. H., II; Boyd, M. R.; Kramer, B.; Fleischhauer, J. Tetrahedron 1994, 50, 7807. sible atropisomer named michellamine C (1c) was also isolated, but was later found to be an artifact formed under too harsh isolation conditions.3,7 The isolation of michellamine C (1c) apparently results from the fact that the michellamines can be interconverted by epimerization under basic conditions.3 At equilibrium the ratio of michellamines A, B, and C is ∼3: 3:1. Although michellamine C (1c) is thermodynamically (7) Bringmann, G.; Harmsen, S.; Holenz, J.; Geuder, T.; Götz, R.; Keller, P. A.; Walter, R.; Hallock, Y. F.; Cardellina, J. H., II; Boyd, M. R. Tetrahedron 1994, 50, 9643. S0022-3263(97)01495-3 CCC: $15.00 © 1998 American Chemical Society Published on Web 01/27/1998 Convergent Total Synthesis of the Michellamines Scheme 1 J. Org. Chem., Vol. 63, No. 4, 1998 1091 Scheme 2 The second synthesis,12 which was completed only a few months later, follows the complementary pathway (B), by first establishing the central bond to result in a dimeric naphthalene unit, which is then connected with the corresponding isoquinoline parts by a double Suzukitype cross-coupling reaction. Since then, additional syntheses have been published from the laboratories of Hoye13 and Dawson14 following pathway A, the central axis again being built up by oxidative dimerization using Ag2O13 or by Suzuki-type cross-couplings.14 These and other15-17 synthetic efforts to build up both mono- and dimeric naphthylisoquinoline alkaloids strongly underline the worldwide interest in this promising field of research. We originally reported the present synthesis in a brief communication in 1994.12 We now provide greater detail and experimental procedures and also describe some of the ancillary studies that facilitated the synthesis’s achievement. Results and Discussion less stable than the other two atropisomers, the reason for the apparent total absence of 1c in nature is unknown. The pronounced biological activity as well as the structural novelty of the michellamines has attracted much interest from synthetic chemists, especially after the U.S. National Cancer Institute published8 an announcement encouraging the research community to pursue synthetic and/or other studies aimed at the production of michellamine B. In principle, the michellamines can be most efficiently dissected retrosynthetically in two manners (Scheme 1). One retrosynthesis leads to a biomimetic pathway (A) that involves the construction of the naphthalene/isoquinoline bonds prior to the formation of the central naphthalene/naphthalene axis. By contrast, the complementary pathway (B) would have the central axis established first, with the two isoquinoline moieties being added subsequently. The first synthesis of the michellamines was achieved in 1994 following the biomimetic concept.7,9,10 In this approach, the monomeric halves of the michellamines, korupensamines A (2, R ) H, X ) H, axis, P-configured) and B (3, R ) H, Y ) H, axis, M-configured), which also occur in A. korupensis,11 were synthesized9,10 and then homo- or cross-coupled by biomimetic oxidative dimerization of appropriately protected 2 and/or 3 using silver(I) oxide.7,10 (8) Anon. J. Nat. Prod. 1992, 55, 1018. (9) Bringmann, G.; Götz, R.; Keller, P. A.; Walter, R.; Henschel, P.; Schäffer, M.; Stäblein, M.; Kelly, T. R.; Boyd, M. R. Heterocycles 1994, 39, 503. (10) Bringmann, G.; Götz, R.; Harmsen, S.; Holenz, J.; Walter, R. Liebigs Ann. Chem. 1996, 2045. The first stage of the synthesis required the preparation of the binaphthalene synthon 5. To that end, we undertook the synthesis of bromonaphthoquinone 11 since dimerization of 11, or derivatives thereof, would afford access to the carbon skeleton of 5. Brassard18 reported that chloronaphthoquinone 8 can be prepared regioselectively in 74% overall yield from the known diene 618,19 and 2,6-dichloro-1,4-benzoquinone (7) by a Diels-Alder reaction followed by aromatization of the adduct on silica gel and final O-methylation (Scheme 2). Since bromides are generally more reactive than chlorides in metal-catalyzed coupling reactions, the bromo analogue 11 was prepared from 9. Using 2,6dibromo-1,4-benzoquinone (9)20 as the dienophile and 618,19 as diene, bromoquinone 10 was synthesized in 70% (11) Hallock, Y. F.; Manfredi, K. P.; Blunt, J. W.; Cardellina, J. H., II; Schäffer, M.; Gulden, K.-P.; Bringmann, G.; Lee, A. Y.; Clardy, J.; François, G.; Boyd, M. R. J. Org. Chem. 1994, 59, 6349. (12) Kelly, T. R.; Garcı́a, A.; Lang, F.; Walsh, J. J.; Bhaskar, K. V.; Boyd, M. R.; Götz, R.; Keller, P. A.; Walter, R.; Bringmann, G. Tetrahedron Lett. 1994, 35, 7621. (13) Hoye, T. R.; Chen, M.; Mi, L.; Priest, O. P. Tetrahedron Lett. 1994, 35, 8747. (14) Hobbs, P. D.; Upender, V.; Liu, J.; Pollart, D. J.; Thomas, D. W.; Dawson, M. I. J. Chem. Soc., Chem. Commun. 1996, 923. Hobbs, P. D.; Upender, V.; Dawson, M. I. Synlett 1997, 965. (15) (a) Hoye, T. R.; Mi, L. Tetrahedron Lett. 1996, 37, 3097. (b) Hoye, T. R.; Chen, M.; Tetrahedron Lett. 1996, 37, 3099. (16) Upender, V.; Pollart, D. J.; Liu, J.; Hobbs, P. D.; Olsen, C.; Chao, W.; Bowden, B.; Crase, J. L.; Thomas, D. W.; Pandey, A.; Lawson, J. A.; Dawson, M. I. J. Heterocyc. Chem. 1996, 33, 1371. (17) (a) Rama Rao, A. V.; Gurjar, M. K.; Ramana, D. V.; Cheda, A. K. Heterocycles 1995, 43, 1. (b) Gable, R. W.; Martin, R. L.; Rizzacasa, M. A. Aust. J. Chem. 1995, 48, 2013. (c) Watanabe, T.; Kamikawa, K.; Uemura, M. Tetrahedron Lett. 1995, 36, 6695. (18) Savard, J.; Brassard, P. Tetrahedron 1984, 40, 3455. (19) Casey, C. P.; Jones, C. R.; Tukeda, H. J. Org. Chem. 1981, 46, 2089. 1092 J. Org. Chem., Vol. 63, No. 4, 1998 Scheme 3 Bringmann et al. Scheme 5 Scheme 4 yield (Scheme 3). Methylation of 10 by refluxing in methyl iodide in the presence of silver (I) oxide gave the desired methyl ether 11 in 97% yield. Bromoquinone 11 was converted into stannane derivative 12 in 71% yield by heating at 100 °C in dioxane with Bu6Sn2 and bis(triphenylphosphine)palladium(II) chloride.21 A second palladium-promoted reaction,21 utilizing the same catalyst, coupled stannane 12 and bromoquinone 11 to afford binaphthoquinone 13. Because 13 is light sensitive, it was usually converted without purification to 14 by reductive acetylation. The yield of 14 from 12 is 65% (Scheme 4). Since the two consecutive coupling reactions leading to 13 use the same reaction conditions, it proved possible to make 13 in a single step. Thus, when half an equivalent of Bu6Sn2 was used in the reaction of 11, after reductive peracetylation, tetraacetate 14 was isolated in 40% yield. We previously reported12 that the dimeriza(20) Prepared in 52% yield using the procedure of Hodgson, H. H.; Nixon, J. J. Chem. Soc. 1930, 1085. The crude product was purified by passing a CH2Cl2 solution of it through a column of silica gel. A number of preparations of 9 have been reported in the literature, but we found this method (the oxidation of 2,4,6-tribromophenol using 90% fuming nitric acid) to be the most reliable one. (21) For recent reviews, see: (a) Farina, V. In Comprehensive Organometallic Chemistry II; Abel, E. W., Stone, F. G. A., Wilkinson, G., Eds.; Pergamon: Oxford, 1995; Vol. 12 (Hegedus, L. S., Ed.), p 161. (b) Knight, D. W. In Comprehensive Organic Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon Press: Oxford, 1991; Vol. 3 (Pattenden, G., Ed.), Chapter 2.3. (c) Bringmann, G.; Walter, R.; Weirich, R. Angew. Chem., Int. Ed. Engl. 1990, 29, 977. tion of 11 to 13 could be realized in a similar yield (43%) by an Ullmann coupling reaction with copper bronze in the presence of tetrakis(triphenylphosphine)palladium(0) in DMF,22 but the current Stille-type23 coupling route is preferred (Scheme 4). The binaphthoquinone 13 had been synthesized before by Laatsch24,25 via an oxidative coupling as shown in Scheme 5. Although our four-step synthesis of 13 is several steps shorter than the one in Scheme 5, we also explored the possibility of improving the latter route. To that end, we first developed an improved synthesis of 19.24-26 The new synthesis starts with the Diels-Alder reaction between diene 618,19 and benzoquinone (16), which affords the adduct 24 cleanly. Desilylation of 24 in methanol with aqueous HCl and subsequent aromatization gave the desired product 19 in 41% overall yield. When excess benzoquinone (16) was used in the DielsAlder reaction, the leftover 16 could serve as the oxidant for the aromatization, which avoided the use of PCC in the original procedure. We were pleased with this modified synthesis because it was shorter than the literature sequence, but it was not a perfect solution. For although the Diels-Alder reaction was fast and clean, the cleavage of ketal 24 only afforded naphthoquinone 19 in a moderate yield. And an O-methylation step was still needed for the preparation of 20. (22) Shimizu, N.; Kitamura, T.; Watanabe, K.; Yamaguchi, T.; Shigyo, H.; Ohta, T. Tetrahedron Lett. 1993, 34, 3421. (23) Stille, J. K. Angew. Chem., Int. Ed. Engl. 1986, 25, 508. (24) Laatsch, H. Liebigs Ann. Chem. 1980, 1321. (25) (a) Krohn, K. Tetrahedron Lett. 1980, 21, 3557. (b) Krohn, K.; Broser, E. J. Org. Chem. 1984, 49, 3766. (26) Rösner, A.; Tolkiehn, K.; Krohn, K. J. Chem. Res. (M) 1978, 3831. Convergent Total Synthesis of the Michellamines Scheme 6 To further improve the synthesis, we sought to optimize the cleavage of ketal 24. Since an acid milder than HCl might give a cleaner reaction, acetic acid was tested and the result was most rewarding. Thus, after the Diels-Alder reaction between diene 6 and benzoquinone (16) in CH2Cl2 was complete, acetic acid was added to the reaction solution. Evaporation of the solvent gave, surprisingly, the desired methyl ether 20 instead of the expected phenol 19 as the product (Scheme 6). The reason why 20 is formed instead of 19 is unclear, but the one-pot reaction provides 20 in 85% overall yield based on diene 6. Compound 20 was then converted to 13 using the previously reported24 procedures (Scheme 5). The synthesis of 13 via 20 still has more steps than the route via 11, but the synthesis of 13/14 via 20 is easier and faster, and better suited for large-scale production. As already mentioned, quinone dimer 13 is light sensitive, so it was usually not purified but instead reductively peracetylated27 directly to afford tetraacetate 14. The transformation is conveniently carried out at room temperature by stirring a mixture of crude 13, zinc powder, acetic anhydride, sodium acetate, and 4-(dimethylamino)pyridine (DMAP) in CH2Cl2 overnight. The conversion of 13 to 14 proceeds cleanly: With a pure sample of 13, 14 was obtained in 95% yield. Tetraacetate 14 contains the entire carbon skeleton of the binaphthalene synthon 5, but it is not properly functionalized for biaryl coupling with the isoquinoline units. Refunctionalization began with a selective bisdeacetylation at the less hindered sites with DBU in methanol28 to give diacetate 25; without purification, 25 was subsequently converted29 into ditriflate 26 using trifluoromethanesulfonic anhydride (Tf2O) and 2,6-lutidine in CH2Cl2. The two-step sequence gave 26 in 70% yield from 14 (Scheme 7). Ditriflate 26 was also converted into the distannanes 27a and 27b and diiodide30 (27) For a leading reference, see: Ulrich, H.; Richter, R. In Methoden de Organische Chemie (Houben-Weyl): Chinone; Georg Thieme Verlag: Stuttgart, 1977; Teil I, p 652. (28) Baptistella, L. H. B.; dos Santos, J. F.; Ballabio, K. C.; Marsaioli, A. J. Synthesis 1989, 436. (29) Stang, P. J.; Hanack, M.; Subramanian, L. R. Synthesis 1982, 85. (30) Ingham, R. K.; Rosenberg, S. D.; Gilman, H. Chem. Rev. (Washington, D.C.) 1960, 60, 459. J. Org. Chem., Vol. 63, No. 4, 1998 1093 Scheme 7 28 (Scheme 7) so that three different types of binaphthalene units were available for coupling with various isoquinoline synthons. We anticipated steric hindrance31 to be the major obstacle in the coupling reaction between the naphthalene and isoquinoline units to give the michellamine skeleton. The isoquinoline unit 4 can be considered a benzene ring with two substituents ortho to the coupling position, while naphthalene 5 () 26-28) can be regarded as a benzene ring with one substituent ortho to the coupling position. It is very congested at the coupling positions. For elaboration of the best coupling conditions, we decided to carry out model coupling studies in order to save the precious enantiomerically pure isoquinoline unit 34. Orcinol derivatives 29, 30, and 31 were used as models because they are similar to the isoquinoline unit in both steric and electronic respects. The stannane derivatives 30 were prepared30 by first lithiation of the corresponding bromide 29 followed by reacting with Bu3SnCl (Scheme 8). The boronic acids32 31 were prepared by lithiation of 29 and subsequent reaction with triisopropyl borate as shown in Scheme 8. After various permutations of potential coupling partners were examined (26-28 with (31) (a) Thompson, W. J.; Gaudino, J. J. Org. Chem. 1984, 49, 5237. (b) Saá, J. M.; Martorell, G.; Garcia-Raso, A. J. Org. Chem. 1992, 57, 678. (c) Fu, J.-m.; Snieckus, V. Tetrahedron Lett. 1990, 31, 1665. (32) For a leading reference, see: Hevesi, L. In Comprehensive Organic Functional Group Transformations; Katritzky, A. R., MethCohn, O., Rees, C. W., Eds.; Pergamon: Oxford, 1995; Vol. 2 (Ley, S. V., Ed.), p 899. 1094 J. Org. Chem., Vol. 63, No. 4, 1998 Scheme 8 Bringmann et al. Scheme 10 Scheme 9 29-31), it was determined that Suzuki-type33,34 crosscoupling reactions between ditriflate 26 and boronic acids 31 gave the best yield for the desired coupling. Coupling reactions between ditriflate 26 and boronic acids 31 were carried out in the presence of tetrakis(triphenylphosphine)palladium(0) and barium hydroxide34 to afford 32a and 32b in 88% and 94% yields, respectively. With a method established for making the final biaryl bonds, we turned to the synthesis of the michellamines themselves. Preparation of the specifically protected key heterocyclic building block 34 (Scheme 9) in a stereochemically homogeneous form, was done as described previously.9,10,35 The subsequent conversion of 34 into the isoquinoline boronic acid 35 was achieved in 89% yield by the same method used above for preparation of the orcinol-derived boronic acids 31 (Scheme 10). Boronic acid 35 was then coupled with ditriflate 26 under the same conditions elaborated for the model compounds to give 36 as a mixture of atropisomers. Deprotection of the benzyl groups was carried out by catalytic hydrogenation; all six benzyl protecting groups were removed by hydrogenation with 10% palladium on charcoal in ethanol at atmo(33) Miyaura, N.; Yanagi, T.; Suzuki, A. Synth. Commun. 1981, 11, 513. (34) (a) Watanabe, T.; Miyaura, N.; Suzuki, A. Synlett 1992, 207. (b) Suzuki, A. Pure Appl. Chem. 1994, 213. (35) Bringmann, G.; Weirich, R.; Reuscher, H.; Jansen, J. R.; Kinzinger, L.; Ortmann, T. Liebigs Ann. Chem. 1993, 877. spheric pressure for 14 h. The two acetates in the naphthalene unit were finally taken off in the last step of the synthesis using methanolic HCl. The resulting mixture of atropisomers was separated by preparative HPLC to give michellamine A (1a) and michellamine B (1b) (1a: 1b ) 1:2.5). The synthetic 1a and 1b were shown to be identical, by direct comparison, to authentic, naturally derived materials. However, we were unable to detect any michellamine C (1c) in the synthetic mixture even though we had a sample of 1c as a TLC/HPLC standard. The question of why 1c is not detectable in either the natural source or our synthetic mixture is intriguing. Atropisomer 1c is a stable compound3 (although slightly underrepresented in a thermodynamically dictated mixture of 1a, 1b, and 1c) and has been synthesized previously.10 While its absence in plant material presumably has to do with the high specificity of the dimerization enzyme, which has recently been isolated,36 at least one of the two coupling steps in our synthesis must proceed with a high diastereoselectivity. Of great interest would be the axial configuration of the intermediate monocoupled product, which would indicate the degree of asymmetric induction by the stereocenters present in 35 and would thus allow an estimation of the additional degree of asymmetric induction exerted by the first-generated biaryl axis on the second coupling step. Regretably, all attempts to isolate or at least detect such a monocoupled intermediate in the reaction of 35 and 26 failed. Biological Evaluation of Synthetic 1a/1b. The antiviral activity of the synthetic michellamines was indistinguishable from that previously reported2,3 for the natural compounds (data not shown). (36) Schlauer, J.; Wiesen, B.; Rückert, M.; AkéAssi, L.; Haller, R. D.; Bär, S.; Fröhlich, K.-U.; Bringmann, G. Arch. Biochem. Biophys. 1998, 350, 87. Convergent Total Synthesis of the Michellamines Conclusion. The convergent synthesis described herein provides the michellamines in an overall yield of 24.6% from diene 6. Experimental Section37 2-Bromo-8-hydroxy-6-methyl-1,4-naphthoquinone (10). To a solution of 2,6-dibromobenzoquinone20 (9) (11.65 g, 43.8 mmol) in 70 mL of dry THF at 0 °C under a nitrogen atmosphere was added dropwise via a syringe pump over 1 h a solution of diene 6 (9.07 g, 48.7 mmol) in 40 mL of dry THF. The resulting solution was stirred at room temperature for 3 h. Then 250 g of 230-400 mesh silica gel was added, and the mixture was shaken until it appeared homogeneous and allowed to stand at room temperature for 48 h. Subsequently, the reaction mixture was put directly onto a silica gel column (4 in. x 10 in.) and purified by eluting with petroleum ether/ethyl acetate (7:3) to give 8.18 g (70%) of bromoquinone 10 as an orange red solid, mp 186-187 °C: IR (KBr) ν 3438, 1638 cm-1; 1H NMR (CDCl3, 400 MHz) δ 2.45 (3H, s), 7.10 (1H, s), 7.45 (1H, s), 7.46 (1H, s), 11.69 (1H, s); 13C NMR (CDCl3, 100 MHz) δ 22.9, 112.6, 122.0, 125.0, 132.1, 140.2, 141.6, 150.0, 163.0, 182.6, 182.9. Anal. Calcd for C11H7O3Br: C, 49.47; H, 2.64. Found: C, 49.40; H, 2.57. In addition, 1.27 g (10%) of methyl ether 11 was obtained as a yellow solid. 2-Bromo-8-methoxy-6-methyl-1,4-naphthoquinone (11). A mixture of 5.38 g of 10 (20 mmol), 6.98 g of Ag2O powder, and 100 mL of iodomethane was refluxed for 1 h. Then the mixture was filtered through Celite and the Celite and the solid were washed with CH2Cl2. The filtrate and the wash were combined and evaporated to give 5.49 g (97%) of bromonaphthoquinone 11 as a yellow solid. The crude product was used in the next reaction without further purification. An analytical sample of 11 was obtained by recrystallization from ethanol as a yellow solid, mp 175-177 °C: IR (CH2Cl2) ν 1667, 1597 cm-1; 1H NMR (CDCl3, 400 MHz) δ 2.48 (3H, s), 4.00 (3H, s), 7.10 (1H, s), 7.41 (1H, s), 7.54 (1H, s); 13 C NMR (CDCl3, 100 MHz) δ 23.0, 57.1, 117.1, 119.2, 121.1, 134.4, 138.8, 143.5, 147.9, 161.3, 176.5, 183.5; HRMS calcd for C12H9BrO3 279.9735, found: 279.9767. Anal. Calcd for C12H9BrO3: C, 51.27; H, 3.23. Found: C, 51.17; H, 3.15. 2-(Tributylstannyl)-8-methoxy-6-methyl-1,4-naphthoquinone (12). A solution of bromonaphthoquinone 11 (370 mg, 1.31 mmol), hexabutylditin (0.730 mL, 1.45 mmol), and bis(triphenylphosphine)palladium(II) chloride (Aldrich, 138 mg, 15 mol %) in 25 mL of anhydrous dioxane was heated at reflux under a nitrogen atmosphere for 1 h. After cooling, the solution was evaporated under vacuum to yield a brown oil which was purified by flash column chromatography on silica gel (petroleum ether/ethyl acetate ) 7:3) to afford 12 as a yellow oil (495 mg, 71%): 1H NMR (CDCl3, 400 MHz) δ 0.85 (9H, t, J ) 7.2 Hz), 1.10 (6H, t, J ) 7.2 Hz), 1.28-1.36 (6H, m), 1.481.51 (6H, m), 2.47 (3H, s), 3.99 (3H, s), 7.07 (1H, s), 7.10 (1H, s), 7.53 (1H, s). Stannane 12 was converted to 13 and 14 without further characterization. 5-Methoxy-7-methyl-1,4-naphthoquinone (20). To a solution of diene 618,19 (2.37 g, 12.7 mmol) in 100 mL of CH2Cl2 under an argon atmosphere at room temperature (37) For general procedures and protocols, see: (a) Kelly, T. R.; Lang, F. J. Org. Chem. 1996, 61, 4633. (b) Reference 10. J. Org. Chem., Vol. 63, No. 4, 1998 1095 was added solid benzoquinone 16 (2.61 g, 24.0 mmol) over 5 min. The resulting dark greenish solution was stirred at room temperature for 18 h. An aliquot of the solution was evaporated to dryness and analyzed by NMR spectroscopy, which showed the formation of the Diels-Alder adduct 24. 24: 1H NMR (CDCl3, 400 MHz) δ 0.20 (9H, s), 1.79 (3H, s), 2.01 (1H, dd, J ) 5.7, 12.8 Hz), 2.85 (1H, d, J ) 12.8 Hz), 3.04 (3H, s), 3.12 (1H, d, J ) 5.7 Hz), 3.23 (1H, m), 5.53 (1H, s), 6.62 (1H, d, J ) 8.4 Hz), 6.74 (1H, d, J ) 8.4 Hz). Then 1.5 mL of acetic acid was added to the reaction mixture and the solvent was evaporated to dryness at room temperature under reduced pressure overnight. A dark, greenish-yellow solid was formed. The solid was mixed with 100 mL of MeOH, and a yellow solid was observed at the bottom of the flask. The solvent was evaporated to dryness and the residue purified by flash column chromatography on silica gel (petroleum ether/ ethyl acetate ) 3:2) to afford 2.18 g (85%) of 20 as a yellow solid, mp 164-166 °C (lit.24 mp 166 °C). 2,2′-Bis(1,4-diacetoxy-8-methoxy-6-methylnaphthalene) (14). (A) From Coupling of 11 and 12. In a sealed tube covered with aluminum foil (13 is light sensitive), a solution of 11 (210 mg, 0.76 mmol), tributylstannane 12 (373 mg, 0.76 mmol), and (PPh3)2PdCl2 (80 mg, 15 mol %) in 10 mL of anhydrous dioxane was heated under argon at 110 °C overnight. After cooling, the mixture was poured onto a stirred mixture of Zn dust (497 mg, 7.60 mmol), 4-(dimethylamino)pyridine (371 mg, 3.04 mmol), NaOAc (623 mg, 7.60 mmol), and Ac2O (2 mL) in 35 mL of CH2Cl2, and the new reaction mixture was stirred in the dark overnight. For the workup, the solution was first filtered through Celite, washed with H2O (3 × 25 mL), dried over Na2SO4, and concentrated under vacuum to afford, after chromatography (petroleum ether/ethyl acetate ) 1:1), tetraacetate 14 as a light brown oil (283 mg, 65%). (B) From Dimerization of 11. In a sealable tube covered with aluminum foil, bromoquinone 11 (438 mg, 1.56 mmol), Sn2Bu6 (0.39 mL, 0.78 mmol), and (PPh3)2PdCl2 (164 mg, 15% mol) were dissolved in 15 mL of anhydrous dioxane under argon. The tube was sealed and heated at 110 °C for 24 h. After being cooled to room temperature, the mixture was poured onto a stirred mixture of Zn dust (1.02 g, 15.6 mmol), 4-(dimethylamino)pyridine (762 mg, 6.24 mmol), NaOAc (1.28 g, 15.6 mmol), and Ac2O (3 mL) in 50 mL of CH2Cl2, and the new reaction mixture was stirred in the dark at room temperature for 15 h. For the workup, the solution was filtered through Celite, washed with H2O, dried over Na2SO4, and concentrated under vacuum. Flash column chromatography on silica gel (petroleum ether/ethyl acetate ) 1:1) afforded 14 (178 mg, 40%) as a light brown oil. (C) From Pure 13. A mixture of binaphthoquinone 13 (2.01 g, 5.00 mmol), dichloromethane (200 mL), zinc dust (10 g), sodium acetate (10 g), acetic anhydride (10 mL), and DMAP (6 g) was stirred in the dark at room temperature for 20 h. Then the mixture was filtered through Celite and the solids were washed with dichloromethane (100 mL). The filtrate and wash were combined and evaporated at reduced pressure and elevated temperature (up to 80 °C) to get rid of the acetic anhydride. The crude product was filtered through a short column of silica gel using dichloromethane as eluent. The filtrate was evaporated to dryness, giving 2.72 g (95%) of 14 as an off-white solid, which was 1096 J. Org. Chem., Vol. 63, No. 4, 1998 sufficiently pure for carrying out the next step without further purification. Crystallization of partially purified 14 from CH2Cl2/ hexanes gave 14 as a white solid, mp 228-230 °C: IR (CH2Cl2) ν 1759 cm-1; 1H NMR (CDCl3, 400 MHz) δ 2.10 (6H, br s), 2.44 (6H, s), 2.49 (6H, s), 3.89 (6H, s), 6.73 (2H, s), 7.11 (2H, br s), 7.22 (2H, s); 13C NMR (CDCl3, 100 MHz) δ 21.3, 21.7, 23.0, 56.8, 110.0, 113.7, 119.3, 121.3, 122.5, 126.7, 130.4, 138.2, 142.2, 143.7, 156.4, 169.9; HRMS calcd for C32H30O10 574.1839, found 574.1820. Anal. Calcd for C32H30O10: C, 66.89; H, 5.26. Found: C, 66.54; H, 5.20. 2,2′-Bis((1-acetoxy-8-methoxy-6-methyl-4-trifluoromethanesulfonyl)oxy)naphthalene) (26). To a solution of tetraacetate 14 (2.80 g, 4.90 mmol) in 65 mL of CH2Cl2 and 65 mL of MeOH was added 1,8-diazabicyclo[5.4.0]undec-7-ene (3.50 mL, 23.4 mmol) at room temperature. After 15 min, the solvent was evaporated at reduced pressure and 50 mL of water was added to the mixture. The mixture was then extracted with CH2Cl2, and the organic layer was separated and dried over Na2SO4. Evaporation of the solvent gave crude diacetate 25 which was used without purification in the next reaction. The crude 25 was dissolved in 80 mL of CH2Cl2 at 0 °C, 2,6-lutidine (1.0 mL, 8.5 mmol) was added, and then trifluoromethanesulfonic anhydride (1.26 mL, 7.50 mmol) was added dropwise over 3 min. The reaction mixture was stirred at room temperature for 10 min. The solvent was evaporated under vacuum, and the resulting oil was purified by flash column chromatography on silica gel (CH2Cl2) to give 2.55 g (70% from 14) of 26 as a white solid. An analytical sample of 26 was obtained by recrystallization from CH2Cl2/hexane as white flakes, mp 190-191 °C: IR (CH2Cl2) ν 1766, 1421, 1205 cm-1; 1H NMR (CDCl3, 400 MHz) δ 2.07 (6H, br s), 2.56 (6H, s), 3.93 (6H, s), 6.83 (2H, s), 7.40 (2H, br s), 7.48 (2H, s); 13C NMR (CDCl3, 100 MHz) δ 21.0, 23.1, 56.9, 110.9, 113.4, 119.4 (q, J ) 320 Hz) 119.6, 120.7, 121.9, 125.9, 130.3, 140.5, 142.7, 144.7, 156.3; HRMS calcd for C30H24S2O12F6 754.0613, found 754.0608. Anal. Calcd for C30H24S2O12F6: C, 47.75; H, 3.21. Found: C, 47.66; H, 3.18. (1R,3R)-N-Benzyl-6,7-bis(benzyloxy)-1,3-dimethyl1,2,3,4-tetrahydroisoquinoline-5-boronic Acid (35). A solution of 120 mg (0.22 mmol) of bromoisoquinoline 349,10,35 in 10 mL of dry THF under argon was cooled to -78 °C. Over the course of 10 min, 0.16 mL (0.24 mmol) of a 1.5 M solution of n-BuLi in hexanes was added and the reaction mixture was stirred 50 min, resulting in an orange solution. Freshly distilled (from sodium) trimethyl borate (0.12 mL, 1.11 mmol) was added, and the reaction mixture was allowed to warm to room temperature overnight. Ten milliliters of water was added, and the mixture was extracted five times with 10-mL portions of CH2Cl2. The combined organic extracts were dried over MgSO4 and evaporated under vacuum. The residue was chromatographed on deactivated (7.5% NH3)37b silica gel using 4:1 petroleum ether/ethyl acetate to give, after recrystallization from ethanol/petroleum ether, 100 mg (89%) of boronic acid 35 as a colorless solid, mp 106108 °C: [R]23D ) +49.3 (c ) 1.5 in MeOH); IR (KBr) ν 3600-3100, 3010, 2920, 2910, 2840 cm-1; 1H NMR (CDCl3, 200.1 MHz) δ 1.30 (3H, d, J ) 6.6 Hz), 1.35 (3H, d, J ) 6.7 Hz), 2.79 (1H, dd, J ) 18.0, 11.3 Hz), 3.13 (1H, dd, J ) 18.0, 4.6 Hz), 3.33 (1H, d, J ) 14.1 Hz), 3.453.56 (1H, m), 3.85 (1H, d, J ) 14.1 Hz), 4.08 (1H, q, J ) 6.7 Hz), 5.02 (2H, s), 5.04 (2H, s), 5.92 (2H, s), 6.42 (1H, Bringmann et al. s), 7.19-7.31 (15H, m). Anal. Calcd for C32H34BNO4 (HBr salt): C, 65.33; H, 6.00; N, 2.38. Found: C, 64.93; H, 6.02; N, 2.36. 5′,5′′-O-Diacetyl-N,N-dibenzyl-6,6′′′,8,8′′′-tetra-Obenzylmichellamine (36). Into a dry Schlenk flask were placed under argon 60.0 mg (0.11 mmol) of isoquinolineboronic acid 35, 37.8 mg (0.050 mmol) of ditriflate 26, 6.0 mg (0.003 mmol) of tetrakistriphenylphosphinepalladium(0), 29.0 mg (0.16 mmol) of barium hydroxide, 3 mL of dimethoxyethane, and 1.5 mL of degassed water. The reaction was heated at 80 °C for 8 h and cooled to room temperature, and volatiles were removed under vacuum. The residue was subjected to preparative thin-layer chromatography on deactivated37b silica gel plates using a 2:1 mixture of petroleum ether/ ethyl acetate as eluent to give 51.0 mg (74%) of 35 in the form of a light brown solid. Evaluation of the 1H NMR spectrum is not worthwhile due to substantial peak broadening (due to hindered rotation) and overlap: IR (KBr, HBr salt) ν 3500-3200, 2940, 2900, 1700, 1650, 1570 cm-1. Anal. Calcd for C92H88N2O10 (HBr salt): C, 79.97; H, 6.42; N, 2.03. Found: C, 79.34; H, 6.44; N, 2.16. Michellamines A (1a) and B (1b). The mixture 36 (50 mg) was dissolved in 2 mL of absolute ethanol and hydrogenated over 5.0 mg of 10% Pd/C for 14 h at room temperature and 1 atm of hydrogen. Catalyst was removed by filtration through a short pad of silica gel, and the filtrate was heated at reflux for 8 h in MeOH that had been saturated in the cold with gaseous HCl. After evaporation of the filtrate, the residue was taken up in MeOH and chromatographed on LH-20 Sephadex, eluting with MeOH. The fractions containing mixtures of 1a and 1b were combined and evaporated. The two atropoisomers were separated on an HPLC equipped with a 254 mm detector, using a 2.1 × 25 cm Rainin Dynamax amine phase column. The crude 1a/1b mixture was dissolved in 7 mL of 87:13 chloroform/methanol, and 0.25mL aliquots were injected and eluted with the same solvent mixture at a flow rate of 12 mL/min to give a total of 6.6 mg (21%) of michellamine A (1a) and 16.5 mg (53%) of michellamine B (1b) which were identical with authentic samples of naturally derived 1a and 1b. 1a: [R]23D ) -8.3 (c ) 0.4 in MeOH) (lit.2 -10.5, c ) 0.83 in MeOH); CD ∆209 -98.3, ∆242 +24.6, ∆258 +17.4; IR (KBr, diacetate) ν 3550-3100, 2960, 2910, 1690, 1600 cm-1; 1H NMR (d4-MeOH, 500.1 MHz) δ 1.21 (6H, d, J ) 6.5 Hz), 1.63 (6H, d, J ) 6.5 Hz), 2.12 (2H, m), 2.34 (6H, s), 2.81 (2H, m) 3.64 (2H, m), 4.10 (6H, s), 4.74 (2H, q, J ) 6.5 Hz), 6.43 (2H, s), 6.75 (2H, s), 6.85 (2H, s), 7.30 (2H, s); 13C NMR (d4-MeOH, 125.0 MHz) δ 18.4, 19.3, 22.1, 33.1, 45.1, 49.4, 57.0, 102.0, 108.0, 113.1, 115.2, 119.1, 120.4, 124.2, 133.1, 134.7, 136.7, 137.5, 152.2, 155.5, 156.9, 159.1. 1b: [R]23D ) -16.2 (c ) 0.72 in MeOH) (lit.2 -14.8, c ) 0.74 in MeOH); CD ∆209 -53.8, ∆214 -53.8; IR (KBr, HBr salt) ν 3600-3150, 2960, 2910, 1600 cm-1; 1H NMR (d4-MeOH, 500.1 MHz) δ 1.16 (3H, d, J ) 6.0 Hz), 1.19 (3H, d, J ) 6.5 Hz), 1.59 (3H, d, J ) 6.5 Hz), 1.63 (6H, d, J ) 6.5 Hz), 2.03 (1H, dd, J ) 18.5, 11.5 Hz), 2.25 (1H, dd, J ) 18.5, 4.5 Hz), 2.33 (3H, s), 2.36 (3H, s), 2.42 (1H, dd, J ) 18.5, 11.3 Hz), 2.69 (1H, dd, J ) 18.5, 4.0 Hz), 3.48-3.55 (2H, m), 4.09 (3H, s), 4.10 (3H, s), 4.62 (1H, q, J ) 6.0 Hz), 4.66 (1H, q, J ) 6.0 Hz), 6.41 (2H, s), 6.75 (1H, s), 6.84-6.85 (3H, m), 7.26 (1H, s), 7.30 (1H, s); 13C NMR (d4-MeOH, 125.0 MHz, diacetate) δ 13.1, 19.1, 20.1, 20.2, 22.1, 22.2, 30.7, 30.7, 33.9, 34.9, 44.5, 44.6, 56.9, Convergent Total Synthesis of the Michellamines 57.0, 101.8, 101.9, 108.3, 114.8, 115.1, 115.2, 118.9, 119.0, 119.2, 119.3, 120.3, 120.4, 124.4, 124.5, 134.6, 134.7, 135.1, 135.2, 136.6, 136.7, 137.4, 137.5, 152.1, 152.2, 155.5, 155.6, 156.3, 156.4, 158.0, 158.1, 180.2, 180.2. Acknowledgment. This work was supported by the National Institutes of Health (Grant CA65093 to T.R.K.), the Deutsche Forschungsgemeinschaft (SFB 251 “Ökologie, Physiologie und Biochemie pflanzlicher und tierischer Leistung unter Stress”), and the Fonds der Chemischen Industrie (financial support and fellowship to R.G.). A.G. and P.A.K. thank NATO and the Alex- J. Org. Chem., Vol. 63, No. 4, 1998 1097 ander von Humboldt Foundation, respectively, for postdoctoral fellowships. Supporting Information Available: Experimental details concerning model studies and other compounds (19, 27a, 27b, 28, 29b, 30a, 30b, 31a, and 31b) not directly related to the final synthetic sequence (6 pages). This material is contained in libraries on microfiche, immediately follows this article in the microfilm version of the journal, and can be ordered from the ACS; see any current masthead page for ordering information. JO971495M