Angewandte

DOI: 10.1002/anie.201004047

Chemie

Natural Product Synthesis

Total Synthesis of Nominal Banyaside B: Structural Revision of the Glycosylation Site**

Corinna S. Schindler, Louis Bertschi, and Erick M. Carreira*
The natural products that comprise the aeruginosin family of serine protease inhibitors[1] represent one of the seven main classes of cyanobacterial peptides.[2] They include the four glycopeptides, banyasides A[3a] (1) and B[3a] (2), suomilide,[3b] and spumigin HKVV,[3c] which incorporate a common, densely functionalized tricyclic azabicyclononane (Abn) core. The biological activity and their complexity render these as notable synthesis targets. These structures present two key challenges: access to the Abn core and installation of the side-chains adorning the periphery, which include lipids, cationic and anionic peptides, as well as glycoside appendages. We recently reported a nine-step synthesis of the Abn core.[4, 5] Herein, we disclose studies focused on the side-chains. These studies have led to an approach for the installation of a hindered peptide along with O-sulfation, and glycosylation reactions. A key outcome of this work is the proposed revised structure for natural banyaside B (3), in which the glycoside is linked to the Abn core at the axial-C-9 OH (Figure 1). With a high-yielding route to the Abn core secured,[4] completion of the synthesis necessitated incorporation of peptides at the hindered secondary amine and C-3 carboxy group along with glycosylation at the C-7 position. Efforts to couple 4 (Scheme 1) and carboxybenzyl-d -leucine employing common reagents (e.g. EDC, DCC, BOPCl, BOP, DEPBT,

Scheme 1. Successful functionalization of the Abn N-2-amine. THF = tetrahydrofuran.

Figure 1. Azabicyclononanes of the aeruginosin family.

[*] C. S. Schindler, L. Bertschi, Prof. Dr. E. M. Carreira ETH Zrich, HCI H335, 8093 Zurich (Switzerland) Fax: (+ 41) 44-632-1328 E-mail: [email protected] [**] This work was supported by a grant from the Swiss National Science Foundation. C.S.S. thanks the Roche Research Foundation for a predoctoral fellowship. We are grateful to Mr. S. Diethelm for the purification and detailed MS-studies of 2 and sulfated-27, as well as to Dr. W. B. Schweizer for the X-ray crystallographic analyses. We gratefully acknowledge Prof. Dr. S. Carmeli for providing 1H and 13 C NMR data for banyaside B. Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/anie.201004047.
Angew. Chem. Int. Ed. 2010, 49, 9229 9232

PyBOP, HATU)[6] did not yield any product, and the starting secondary amine was either recovered or the corresponding C-7 or C-9 esters formed. Attempts at coupling under more harsh conditions were unsuccessful.[7] Traces of amide (< 5 %) were observed (LCMS) following the procedure reported by Vedejs and Driver[8] with the symmetrical anhydride. Collectively, these results make it clear that the N-2 amine is sterically congested and rendered inaccessible. The steric effect is exacerbated by the substituents of the Abn core, which attenuate the reactivity of the embedded amine. An approach was pursued in which simple carboxamides would be installed at the N-2 position under forcing conditions, and the adducts were subsequently elaborated to the desired d -leucine derivative. Amide formation at the N-2 position was effected by treatment of 6[9] with 10.0 equivalents of 4-methylpent-2-enoic anhydride in tetrahydrofuran at 110 8C (Scheme 1).[10] Subsequent treatment of 7 with LiOOH in tetrahydrofuran led to selective oxazolidinone deprotection, forming 8 in 92 % yield. The functionalization of 8 was subsequently envisioned.[11, 12] Treatment of 8 with catalyst [Mn(dpm)3]/ PhSiH3/O2 gave a 1:1 mixture of lactones 9 and 13 (79 % yield). After their separation, configurational assignment was effected with 2D NOESY.[13] Lactones 9 and 13 were each subjected to lactone opening with amine 10,[14] whereupon alcohols 11 and 12 were obtained in 85 % and 81 % yield, respectively.

 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

9229

Communications
The key point in the synthesis required conversion of alcohol 11 into azide 12 (Scheme 2). Neither the use of DEAD/dppa[15] nor DEAD/HN3 provided desired 12, instead (Figure 2). Thus, the 13C shifts for the anomeric carbon atom as well as the C-7 carbon atom on the Abn core of synthetic 2 differed by d % 2 ppm. The 13C shifts for the C-8 carbon atom

Scheme 2. a) [Mn(dpm)3], PhSiH3, O2, EtOH, 23 8C, 79 % (9/13, 1:1); b) 10 (1.0 equiv), CH2Cl2, 23 8C, 85 %; c) 1. MsCl (1.1 equiv), Et3N, CH2Cl2, 23 8C; 2. NaN3 (1.5 equiv), DMF, 23 8C, 81 % d) 1. MsCl (1.1 equiv), Et3N, CH2Cl2, 23 8C; 2. LiCl (1.1 equiv), DMF, 23 8C, 78 %; e) NaN3 (1.5 equiv), DMF, 23 8C, 89 %. Mn(dpm)3 = tris(2,2,6,6-tetramethyl-3,5-heptanedionato)manganese(III), Ms = mesyl, Ac = acetyl, Boc = tert-butoxycarbonyl, DMF = N,N-dimethylformamide. Figure 2. Comparison of nominal banyaside B (2) and C-9 axial a-glycoside (27) with the NMR spectroscopic data published for natural banyaside B.

leading to recovery of 11. Mesylation of 11 followed by exposure to NaN3 gave 12 in 81 % yield (2 steps). The diastereomeric alcohol was converted into 12 by a doubleinversion procedure that proceeds through the chloride and of synthetic 2 were off by d = + 3.3 ppm, while d(C-7) were off provides a stereoconvergent approach to 12. by d = 2.2 ppm. Despite these discrepancies, the dehydroAzide reduction with aqueous PPh3 led to formation of pyrrole fragment as well as the d -leucine subunit compared well to the natural product. diketopiperazine 15 and amine 14 as a 1:1 mixture. Treatment Because the differences described above were localized of 12 with Pd/C also gave 15 as the sole product. The use of about the C-7 glycosylation site in synthetic 2 (Figure 2), we PMe3 was found to be critical in furnishing 14. However, when speculated that the natural compound might more properly the amine product was subjected to peptide coupling, 15 was be described with the glycosidic linkage at C-9-OH (3). again isolated. Attention was then focused on methods that enabled rapid coupling. With anhydride 16, adduct 17 was formed in 70 % yield. Tripeptide 17 underwent deprotection (LiOHH2O/aq. THF/MeOH) to afford the diol 18 in 95 % yield (Scheme 3). Glycosyl donor 19 was prepared and exposed to acceptor 18 in the presence of Me3SiOTf (0.1 equiv), leading to the formation of a diastereomeric mixture of glycopeptide 20 in 75 % yield (a/b ratio 2:1). These diastereomers were separated following deprotection (tris(dimethylamino)sulfonium difluorotrimethylsilicate, TASF) and thus afforded 21 (92 % yield, Scheme 4). The pyrrolidine was deprotected with trifluoroacetic acid (TFA) and then allowed to react with 22 to install the protected guanidine and furnish 23 in 69 % yield. The primary alcohol was subjected to O-sulfation (SO3py) to afford 24, which in turn was deprotected to give 2. Careful comparison and analysis of the NMR spectroscopic data for synthetic 2 to that reported for Scheme 3. a) H , Pd/C (10 mol %), MeOH, 95 %; b) 1. PMe , aq. THF, 23 8C; 2 3 the natural banyaside B revealed significant differ- 2. 16 (5.0 equiv), CH2Cl2, 23 8C, 70 %; c) LiOH (3.0 equiv), THF, H2O, MeOH, ences of up to 3 ppm in the 13C NMR chemical shifts 23 8C, 95 %. TBDPS = tert-butyldiphenylsilyl.

9230

www.angewandte.org

 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Angew. Chem. Int. Ed. 2010, 49, 9229 9232

Angewandte
between the free primary OH and its O-sulfate (Figure 2, circles). Thus, upon O-sulfation of 23, the 13C shifts for the side-chain bearing the sulfate underwent a displacement from d = 83.0 to 79.8 ppm for the CHOMe carbon atom and d = 61.6 ppm to 66.3 ppm for the HO3SOCH2 carbon atom. Consequently, we proposed that natural banyaside B was the glycosylated aanomer at axial-C-9 OH instead of the equatorial-C-7 OH as shown for 3 (Figure 1). In order to provide additional corroborative evidence for the structural revision, glycoside 27 was subjected to the sulfation conditions described for 23. Sulfated 27 displayed an identical retention factor by LCMS to that described by Pluotno and Carmeli[3a] for the natural product (YMC-Pack, ODS-A, isocratic MeOH/H2O Scheme 4. a) 19 (1.0 equiv), Me3SiOTf (0.1 equiv), 4 M.S., THF, 23 8C, 75 % (a/b, 2:1); b) TASF (5.0 equiv), DMF, 23 8C, 92 %; c) 1. F3CCO2H (10.0 equiv), CH2Cl2, (1:1), EIC m/z 950.40, UV at 210 nm): k = 1.94 23 8C; 2. 22 (3.0 equiv), Et3N, CHCl3, 23 8C, 69 %; d) SO3py (1.0 equiv), CH2Cl2, 23 8C, (natural banyaside B), and k = 1.98 (synthetic 25 %; e) F3CCO2H/CH2Cl2 (9:1), 23 8C, 95 %. M.S. = molecular sieves, TES = triethylbanyaside 3). By contrast, the nominal banyaside silyl, TASF = tris(dimethylamino)sulfonium difluorotrimethylsilicate, Tf = trifluorome2 elicited k = 1.73 (for details, see the Supporting thanesulfonyl. Information).[3a] In conclusion, the total synthesis of nominal banyaside B (2) has been completed in 23 steps, During the course of optimizing the glycosylation, we relying on the initially reported synthesis of the Abn core observed the equatorial-OH (C-7 OH) to be more-reactive common to banyasides A (1) and nominal banyaside B (2), than the axial-OH (C-9 OH). Thus, glycosylation of diol 4 suomilide, and spumigin HKVV. The introduction of the with 1.0 equivalent of 19 furnished an 8:1 product mixture, three peptide side-chains proved to be challenging and in fact assigned as the C-7 (26) and C-9 (25) monoglycosides, could not be realized following standard peptide coupling respectively (Scheme 5). Noteworthy, and diagnostic for the procedures, thereby clearly demarking the limitations of current methods for the installation of hindered peptide bonds. The total synthesis of nominal banyaside B (2) was realized by relying on a late-stage glycosylation and introduction of the guanidine subunit on the sensitive 2,5dehydropyrrole. Finally, a key outcome of the investigations is the revision of the structure for banyaside B with the glycolipid appended at C-9.
Received: July 2, 2010 Revised: August 17, 2010 Published online: October 22, 2010

Chemie

Scheme 5. Attempted glycosylations of the banyaside core (4).

C-7 isomer, was HMBC correlations between the anomeric proton and Abn C-7 as well as a COSY cross-peak between the H-9 proton and the Abn 9-OH (Figure 2). The minor isomer (25) lacks a HMBC signal involving the anomeric proton and the Abn C-9 carbon atom, a feature also noted in the spectra for natural banyaside B. These results prompted us in turn to prepare the corresponding regioisomer of 20, which was converted into 27, following procedures described above. The 13C NMR shift for the anomeric carbon atom in 27 was found to be d = 98.9 ppm, which compares well with d = 98.7 ppm for natural banyaside B. Indeed, 27 shows strong similarities and chemical shifts in the 13C NMR spectra with the isolated natural product (Figure 2, triangles with Dd < 1 ppm). The only major differences are in the d -leucine sidechain (circles, Figure 2, 27), in line with the difference
Angew. Chem. Int. Ed. 2010, 49, 9229 9232

Keywords: aeruginosins banyaside b glycopeptides natural products total synthesis

[1] For a recent review on the chemistry of the aeruginosins, see: K. Ersmark, J. R. Del Valle, S. Hanessian, Angew. Chem. 2008, 120, 1220; Angew. Chem. Int. Ed. 2008, 47, 1202. [2] M. Welker, H. von Doehren, FEMS Microbiol. Rev. 2006, 30, 530. [3] a) A. Pluotno, S. Carmeli, Tetrahedron 2005, 61, 575; b) K. Fujii, K. Sivonen, K. Adachi, N. Kazuyoshi, Y. Shimizu, H. Sano, K. Hirayama, M. Suzuki, K. Harada, Tetrahedron Lett. 1997, 38, 5529; c) K. Fuji, K. Harada, M. Suzu, K. Adachi, H. Sano, K. Noguchi, K. Hirayama, K. Sivonen, Tennen Yuki Kagobutsu Toronkai Koen Yoshishu 1996, 38, 277. [4] C. S. Schindler, C. R. J. Stephenson, E. M. Carreira, Angew. Chem. 2008, 120, 8984; Angew. Chem. Int. Ed. 2008, 47, 8852. [5] C. S. Schindler, E. M. Carreira, Chem. Soc. Rev. 2009, 38, 3222.

 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.angewandte.org

9231

Communications
[6] EDC = 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochlo-ride, DCC = dicyclohexylcarbodiimide, BOPCl = bis(2oxo-3-oxazolidinyl)phosphinic chloride, BOP = benzotriazole1-yl-oxy-tris-(dimethylamino)-phosphonium hexafluorophosphate, DEPBT = 3-(diethoxyphosphoryloxy)-1,2,3-benzotriazin-4(3H)-one, PyBOP = benzotriazol-1-yl-oxytripyrrolidinophosphonium hexafluorophosphate, HATU = 2-(7-aza-1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate, HOBt = 1-hydroxybenzotriazole. [7] Details may be found in the Supporting Information. The use of Leuchs anhydride was also examined; see: H. R. Kricheldorf, Angew. Chem. 2006, 118, 5884; Angew. Chem. Int. Ed. 2006, 45, 5752. [8] E. Vedejs, S. T. Driver, J. Am. Chem. Soc. 1993, 115, 3358. [9] Crystallographic data for 6 : C18H24N2O9, M = 412.395, monoclinic, space group P21/n, a = 10.3693(2), b = 10.3316(3), c = 18.4211(4); V = 1971.70(8) 3, 1calcd = 1.389 Mg m3, T = 173 K, reflections collected: 8319, independent reflections 4506 (R(int) = 0.038), R(all) = 0.0635, wR(gt) = 0.0505. CCDC 779698 (6) contains the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac. uk/data_request/cif. [10] Complete removal of the DCC-derived byproduct was found to be crucial for the successful preparation of 8. [11] a) T. Mukaiyama, S. Isayama, S. Inoki, K. Kato, T. Yamada, T. Takai, Chem. Lett. 1989, 449; b) S. Inoki, K. Kato, T. Takai, S. Isayama, T. Yamada, T. Mukaiyama, Chem. Lett. 1989, 515; c) K. Kato, T. Yamada, T. Takai, S. Inoki, S. Isayama, Bull. Chem. Soc. Jpn. 1990, 63, 179. [12] a) J. Waser, E. M. Carreira, Angew. Chem. 2004, 116, 4191; Angew. Chem. Int. Ed. 2004, 43, 4099; b) J. Waser, E. M. Carreira, J. Am. Chem. Soc. 2004, 126, 5676; c) J. Waser, H. Nambu, E. M. Carreira, J. Am. Chem. Soc. 2005, 127, 8294; d) J. Waser, B. Gaspar, H. Nambu, E. M. Carreira, J. Am. Chem. Soc. 2006, 128, 11693; e) B. Gaspar, E. M. Carreira, Angew. Chem. 2007, 119, 4603; Angew. Chem. Int. Ed. 2007, 46, 4519; f) B. Gaspar, E. M. Carreira, Angew. Chem. 2008, 120, 5842; Angew. Chem. Int. Ed. 2008, 47, 5758. [13] For details, see the Supporting Information. [14] For the preparation of 10, see: S. Hanessian, R. Margarita, A. Hall, S. Johnstone, M. Tremblay, L. Parlanti, J. Am. Chem. Soc. 2002, 124, 13342. [15] DEAD = diethyl azodicarboxylate, dppa = diphenylphosphoryl azide.

9232

www.angewandte.org

 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Angew. Chem. Int. Ed. 2010, 49, 9229 9232