DOI: 10.1002/chem.201403986
Full Paper
& Total Synthesis
Stereoselective Total Synthesis and Structural Elucidation of ()-Indoxamycins A–F Chi He, Chenlong Zhu, Bingnan Wang, and Hanfeng Ding*[a]
Abstract: In this study, a concise and stereoselective approach for the divergent total synthesis of ()-indoxamycins A–F is described. The key steps of the strategy include an Ireland–Claisen rearrangement, an enantioselective 1,6enyne reductive cyclization, and a tandem 1,2-addition/oxa-
Michael/methylenation reaction. The relative and absolute configuration of these natural products has been unambiguously elucidated, and their cytotoxic activities against HT-29 and A-549 tumor cell lines are also reported.
Introduction
()-Indoxamycins A–F (1–6) were tested for antiproliferative activity in the HT-29 tumor cell line. Among this family, ()-indoxamycins A (1) and F (6) were reported to cause significant growth inhibition with IC50 values of 0.59 and 0.31 mm, respectively. The relative and absolute structures of ()-indoxamycins were initially assigned by combined spectral and chemical methods, which revealed the presence of an unprecedented [5,5,6] tricyclic cage-like carbon framework, an a,b-unsaturated carboxylic acid side chain, and a trisubstituted alkene appendage. Three of the six contiguous chiral centers embedded in the congested pentamethylindenyl core are quaternary, of which two are vicinal. The unique chemical structure and potential anticancer activity of ()-indoxamycins prompted us to explore their laboratory synthesis and biological function. In early 2012, Carreira and co-workers reported an exquisite approach for the total synthesis of rac-indoxamycin B.[3] Recently, our group disclosed a divergent route to indoxamycins A, C, and F.[4] Herein, we detail our efforts on the total synthesis of ()-indoxamycins A–F, as well as the structural elucidation and preliminary biological evaluation of these natural products.
Over the past fifty years, the actinomycetes have continuously provided a variety of bioactive secondary metabolites, leading to the formation of more than half of all microbial antibiotics discovered to date.[1] Although this started several decades ago, most efforts in small-molecule discovery have mainly focused on soil-derived actinomycetes. In contrast, microorganisms derived from marine plants and animals have only recently received significant attention from chemists as a rich source of structurally diverse natural products that possess potent biological activities. In 2009, Sato and co-workers isolated a novel class of polyketides from saline cultures of marine-derived actinomycete strain NPS-643, which they subsequently named ()-indoxamycins A–F (Figure 1).[2]
Results and Discussion Total synthesis of ()-indoxamycins A, C, and F Our retrosynthetic analysis is depicted in Scheme 1. By looking into the structures carefully, indoxamycins A, C, and F were conceived to be accessed divergently from a common latestage precursor 7. Construction of 7 might be achieved through a substrate-controlled tandem reaction[5] involving a 1,2-addition/oxa-Michael/methylenation sequence, which would build up the [5,5,6] tricyclic ring system of the molecule. The desired enone-aldehyde intermediate 8 was expected to arise from 1,6-dienyne 9 through a palladium-catalyzed reductive cyclization.[6] Compound 9 could be furnished with the introduction of the two vicinal quaternary centers at C(2a) and C(7b) by an Ireland–Claisen rearrangement[7] from allyl ester
Figure 1. Structures of ()-indoxamycins A–F originally reported by Sato et al.
[a] C. He, C. Zhu, B. Wang, Prof. Dr. H. Ding Department of Chemistry, Zhejiang University 148 Tianmushan Road, Hangzhou 310028 (China) E-mail: [email protected] Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/chem.201403986. Chem. Eur. J. 2014, 20, 1 – 9
These are not the final page numbers! ÞÞ
1
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
Full Paper tained in 90 % yield through esterification of 11 and 12. Under optimized conditions, Ireland–Claisen rearrangement of 10 afforded 13 in 85 % yield as a 1.5:1 mixture of diastereomers. With the smooth construction of the two vicinal quaternary centers at C(2a) and C(7b), subsequent allylic oxidation of 13 followed by DDQ dehydrogenation afforded dienone 14 in 71 % yield over two steps. After protecting group (TMS) removal, dienyne 9 was formed in 85 % yield, which was ready for the next enyne cyclization. In view of the failures in radical approaches,[9] transitionmetal catalyzed reductive cyclization[6] was considered as an alternative solution. Although there were no literature precedents for the reductive cyclizations of cyclohexadienone-containing terminal alkynes, we were delighted to obtain the bicyclic product 15 in 87 % yield as a single diastereomer through simple modification of the standard conditions reported by Trost [Pd2(dba)3, P(o-tol)3, Et3SiH, AcOH, benzene, RT].[6a] The stereochemical outcome of the two methyl groups at C(2a) and C(7b) of bicyclo 15 with syn-relationship was presumably the result of synergistic effects from both the chairlike transi-
Scheme 1. Retrosynthetic analysis of indoxamycins A, C, and F.
10; the latter was easily prepared from the readily available intermediates 11 and 12.[8] Our synthesis endeavor started with the construction of the common precursor 7 (Scheme 2). Allyl ester 10 was first ob-
Scheme 2. Synthesis of the common precursor 7 and ORTEP drawing of tetracyclic compound 26. Ellipsoids set at 40 %. EDC·HCl = N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride, 4-DMAP = 4-(dimethylamino)pyridine, 3,5-DMP = 3,5-dimethylpyrazole, DDQ = 2,3-dichloro-5,6-dicyano-1,4-benzoquinone, Dibal-H = diisobutylaluminium hydride, DMP = Dess–Martin periodinane, Eschenmoser’s salt = N,N-dimethylmethyleneiminium chloride, DME = 1,2-dimethoxyethane.
&
&
Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
2
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Paper tion state (9’) and the formation of thermodynamically more stable cis-5-6-fused ring system. In order to achieve an asymmetric total synthesis, we initially planned to take advantage of the diastereoselective Ireland– Claisen rearrangement that had been developed by Zakarian and co-workers (Scheme 3).[10] Although highly stereoselective enolizations could be achieved by individual treatment of both diastereomers of the enantiopure ester ()-10 a and ()-10 b with chiral Koga-type bases, only slight improvements in diastereoselectivity (up to 2.6:1 d.r.) were obtained, presumably due to the uncontrollability of the chairlike transition state. Compared to the excellent acyclic stereocontrol, moderate results for the reactions of similar esters of cyclic alcohols were also observed by Zakarian et al.[10e]
Table 1. Optimization of stereodivergent 1,6-enyne reductive cyclization.[a]
Entry
[Pd]
Ligand
Time [h]
(+)-15, 2a-epi-15 Yield [%][b] (ee)[c]
1[d] 2[d] 3[d] 4[d] 5[d] 6 7 8 9 10 11 12 13 14[f]
Pd(OAc)2 Pd(dppf)Cl2 Pd(tfa)2 Pd2(dba)3 Pd(tfa)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2
L1 L1 L2 L3 L3 L3 L1 L4 L5 L6 L7 L8 L9 L3
24 24 10 24 10 0.5 1 1 1 2 24 0.5 0.5 0.5
18 (27), n.d.[e] (–) 16 (23), n.d. (–) 45 (29), 36 (21) 15 (65), n.d. (–) 40 (72), 32 (64) 42 (84), 33 (71) 35 (48), 26 (43) 38 (41), 33 (35) 37 (49), 28 (30) 33 (35), 25 (27) n.d. (–), n.d. (–) 32 (25), 24 (22) 39 (43), 34 (41) 46 (93), 43 (80)
[a] Reaction conditions: 9 (0.1 mmol), [Pd] (0.025 equiv), ligand (0.05 equiv), AcOH (2.0 equiv), and Et3SiH (1.5 equiv) in DMSO at 25 8C. [b] Isolated yields. [c] Determined by HPLC analysis. [d] Benzene was used as solvent. [e] Not determined. [f] HCO2H (2.0 equiv) was used in place of AcOH. L1 = (R)-binap, L2 = (R)-Tol-binap, L3 = (R)-segphos, L4 = (R)-3,5xylyl-binap, L5 = (R)-H8-binap, L6 = (R)-MeO-Duphos, L7 = (R)-diop, L8 = (R)dtbm-segphos, L9 = (R)-MeO-dtbm-biphep, tfa = trifluoroacetate.
reduction of 17 was well controlled in the presence of Dibal-H at low temperature to afford diol 19 cleanly, presumably via the chelated alkoxyaluminium intermediate 18. Oxidation of the 19 with freshly prepared Dess–Martin periodinane (DMP) gave 8 in 76 % yield over the two steps. With the enone-aldehyde precursor 8 secured, we turned our attention to the construction of the tricyclic framework of the molecule. After extensive optimization, the expected tandem 1,2-addition/oxa-Michael[13]/methylenation reaction promoted by the Grignard reagent (20) proceeded smoothly to afford the tricyclic enone 24 with no detectable C(2) diastereomers. It is worth noting that oxa-Michael reactions have rarely been considered in the design of domino processes due to the inherent instability of the in situ generated enolate, which would result in elimination or retro-Michael side reactions. The only recent example was reported by Menche and co-workers, who developed a tandem oxa-Michael/Tsuji–Trost reaction leading to the concise synthesis of functionalized tetrahydropyranes.[14] We proposed that the instability of intermediate 23 might be the driving force to facilitate the spontaneous elimination; this gave excellent yield of the product without prior oxidation or addition of any base. Subsequently, isomerization of the double bond of 24 from D4, 9 to D3, 4 under acid catalysis[15] afforded the desired product (1’’E)-7 in 82 % yield (93 % ee), the ee value of which could be further improved to 96 % through recrystallization. To establish the relative and absolute stereochemistry, compound 25 was obtained in 84 % yield by quenching the
Scheme 3. Optimization of diastereoselective Ireland–Claisen rearrangement.
We then moved on to the development of an enantioselective version of the reductive 1,6-enyne cyclization.[11] Inspired by Ito and Mikami’s work regarding the asymmetric Alder-ene reactions, the chiral C2-symmetric bidentate phosphorus ligands were employed. Preliminary screens (Table 1, entries 1– 5) showed Pd(tfa)2/(R)-segphos was an effective catalyst system, with the stereodivergent formation[12] of the enantioenriched bicyclic product (+)-15 (40 % yield, 72 % ee) and 2a-epi-15 (32 % yield, 64 % ee). Furthermore, accelerating the reaction by the cationic PdII catalyst, [Pd(MeCN)4](BF4)2, in DMSO in the presence of (R)-segphos afforded (+)-15 in 42 % yield and 84 % ee (entry 6). However, displacement of segphos with other ligands only gave diminished yields and ee values (Table 1, entries 7–13). Pleasingly, excellent enantioselectivity (93 % ee) and yield (46 %) of (+)-15 were obtained by using the catalyst combination [Pd(MeCN)4](BF4)2/(R)-segphos and formic acid as the additive,[6d, e] accompanied by the formation of 2a-epi-15 in 43 % yield and 80 % ee (entry 14). With successful preparation of (+)-15, a two-step installation of the quaternary center at C(5) delivered 17 in 78 % yield, together with the formation of corresponding methyl enol ether (ca. a 5:1 mixture), which can be recycled through hydrolysis upon treatment with p-TsOH·H2O. Subsequent regioselective Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
These are not the final page numbers! ÞÞ
3
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
Full Paper ment of the configuration at C(6). However, exhaustive trials to invert the configuration, such as Mitsunobu reaction or sequential oxidation and reduction, met with failure. Finally, we opted for the sulfoxide–sulfenate rearrangement.[18] Treatment of 32 with benzenesulfenyl chloride gave allylic sulfoxide 36 in 92 % yield. We were then delighted to find that the Mislow– Evans rearrangement[19] took place unexpectedly during HWE olefination, presumably induced by nucleophilic phosphite reagent, which might act as a thiophile. The desired allylic alcohol 37 was therefore obtained in only one step and 95 % yield. In the event, saponification of 37 afforded (1’’E)-2,6-di-epi-6 in 92 % yield. The synthetic samples display identical spectral and physical properties with those reported for natural ()-indoxamycins A, C and F, respectively;[2] this unambiguously confirmed their relative and absolute configuration.
tandem reaction with aqueous NH4Cl prior to the methylenation. Although solid in racemic form, enantioenriched 25 was unable to afford single crystal suitable for X-ray crystallography. Tetracycle 26 was thus formed in 42 % yield through acidinduced lactonization of 25, the absolute configuration of the former was unambiguously assigned by X-ray crystallographic analysis[16] (see the structure in Scheme 2). This also proved our initial prediction for the relative configuration at C(2), which was based on a Si-face attack of the Grignard reagent 20 on the aldehyde moiety of 8. The late stages of the synthesis are outlined in Scheme 4, which entails a divergent route to the revised structures of ()-indoxamycins A, C, and F. 1,4-Reduction of (1’’E)-7 with LSelectride followed by capture of the resulting enolate with PhNTf2 gave enol-triflate 27 in 79 % yield. Subsequent palladium-catalyzed reductive detriflation provided alkene 28 in 85 % yield. Controlled reduction (Dibal-H, 105 8C) of 28 to aldehyde 29 followed by Horner–Wadsworth–Emmons (HWE) olefination smoothly afforded (1’’E)-2-epi-1 in 77 % yield over the two steps. Starting from the same enone ester (1’’E)-7, sequential reduction gave allylic alcohol 32 in 84 % yield as a single diastereomer. Allylic chloro-dehydroxylation of 32 by treatment with SOCl2 afforded allylic chloride 33 in 97 % yield. Finally, olefination and hydrolysis of the resulting enoic acid 34 in aqueous AgNO3[17] provided (1’’E)-2-epi-3 in 78 % yield over the two steps. To synthesize ()-indoxamycin F, compound 35 was initially constructed from intermediate 32 through HWE olefination in 83 % yield. Unfortunately, the NMR spectra of 35 deviated from those reported for natural indoxamycin F. Careful reexamination of the NMR spectroscopy data revealed the misassign-
Total synthesis of ()-indoxamycin D With sufficient quantity of ()-indoxamycin A in hand, the biomimetic routes to the other indoxamycin family members were investigated. It was found that Riley oxidation[20] of ()-indoxamycin A methyl ester (38) followed by hydrolysis afforded ()-indoxamycin D ((1’’E)-2-epi-4) and (1’’Z)-2-epi-4 in 25 and 49 % yield, respectively (Scheme 5). A similar result (72 % yield, 1:2 d.r.) was also obtained from direct allylic oxidation of ()-indoxamycin A. On the other hand, singlet oxygen ene reaction[21] of 38 gave allylic alcohol 39 and 2’’-epi-39 in 61 % combined yield as a mixture of diastereomers (ca. 2.8:1 d.r.). To our disappointment, despite further experimentation, the other indoxamycin family members could not be accessed since the D1’’, 2’’ double bond seems the most reactive site
Scheme 4. Divergent total synthesis of ()-indoxamycins A, C, and F. Tf = trifluoromethanesulfonyl, L-Selectride = lithium tri-sec-butylborohydride.
&
&
Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
4
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Paper 60 8C] delivered alkene 44 in 80 % yield.[23] Since straightforward reduction of 44 to aldehyde 45 failed, an alternative reduction and oxidation sequence was then executed. HWE olefination of 45 gave 46 in 84 % yield over three steps from 44. After being converted to acetate 48, acid-catalyzed isomerization of the exo-double bond underwent smoothly to provide 49 (70 % yield), the saponification of which afforded acid 50 in 91 % yield. Surprisingly, the spectroscopic data (1H and 13C NMR) of 50 did not match those reported for natural ()-indoxamycin B.[2, 3] Careful inspection of the whole sequence led to a conclusion that the configuration at C(5) had been inverted unexpectedly during the reduction process (44!45), which can be supported by detailed NOESY analysis of 41 and the formation of tetracyclic ether 47 through etherification of 46 in 55 % yield. Faced with this problematic issue, a revised synthetic route was put forward (Scheme 7). Reduction of 44 yielded 51 with complete TBS migration, which was then converted to acetate 52 in 98 % yield. Subsequent TBS deprotection proved challenging, various conditions such as TBAF, p-TsOH·H2O, AcOH, HCl (aq.), pyridinium p-toluenesulfonate gave either a bad ratio or trace amount of the desired product. Fortunately, upon treatment of 52 with HF·pyridine in MeCN at 10 8C, a 5:1 mixture of 53 and 5-epi-53 were obtained in 84 % combined yield. The ratio was sensitive to temperature, a decreased selectivity (1.5:1, 53/5-epi-53) was observed by warming the reaction to 25 8C. With enough 53 in hand, DMP oxidation followed by HWE olefination afforded 55 in 83 % yield over two steps. The following exo-double bond isomerization of 55 only succeeded with excess amount of p-TsOH·H2O (15 equiv) at 100 8C, as prolonged reaction time with catalytic acid resulted in decomposition of the starting material. To our delight, hydrolysis of both the two esters under mild conditions (Me3SnOH, 1,2-dichloroethane, 84 8C)[24] afforded ()-indoxamycin B ((1’’E)-2-epi-2) in 55 % yield over two steps. When treated with aqueous LiOH,
among the four existing double bonds in the skeleton toward allylic oxidation. A variety of conditions[22] such as (CrO3, 3,5DMP), [Pd(OH)2, Cs2CO3, tBuOOH], [PhI(tfa)2, Cs2CO3, tBuOOH], [Mn(OAc)3, tBuOOH], (CuI, tBuOOH), [Pd(OAc)2, p-benzoquinone, anisole, AcOH] and (NBS, benzoyl peroxide) either gave no reaction or caused decomposition of 38.
Scheme 5. Allylic oxidations of ()-indoxamycin A methyl ester (38).
Total synthesis of ()-indoxamycin B With the divergent concept in mind and following the established route described for ()-indoxamycin A, the initial synthetic strategy toward ()-indoxamycin B is shown in Scheme 6. Aldol reaction of 16 with formalin afforded the unstable alcohol 40 in 88 % yield. Immediate silylation of 40 gave 41 in 93 % yield, whereas other methods for the hydroxyl protection under basic conditions led to the regeneration of 16 through a retro-aldol process. By employing previously developed procedures, tricyclic triflate 43 was obtained uneventfully in 50 % yield over four steps from 41. Subsequent detriflation of 43 under optimized conditions [Pd(OAc)2, dppf, Et3SiH, DMF,
Scheme 6. Initial strategy toward total synthesis of ()-indoxamycin B. Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
These are not the final page numbers! ÞÞ
5
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
Full Paper
Scheme 7. Total synthesis of ()-indoxamycin B.
the acetate was removed first, which led to further fragmentation through the vinylogous retro-aldol reaction pathway.[25]
Total synthesis of ()-indoxamycin E Bicycle 8 was initially employed for the tandem 1,2-addition/ oxa-Michael/methylenation reaction promoted by organolithium reagent 57, which delivered an inseparable mixture of 59 and 2-epi-59 in 74 % yield (3:1 d.r.). To achieve a satisfactory diastereoselectivity, bicycle 56 with the endo-double bond at D3, 4 was chosen as our starting point in the synthetic approach to ()-indoxamycin E (Scheme 8). Prepared from 8 in 76 % yield, the tandem reaction of 56 afforded tricyclic enone 58 in 78 % yield as a single diastereomer. In contrast, the corresponding Grignard reagent did not promote the reaction at all. Conjugate reduction of 58 followed by in situ triflation delivered 60 in 82 % yield, which then gave rise to 61 in 95 % yield under palladium-catalyzed reductive conditions. Controlled reduction of ester 61 furnished aldehyde 62 in 80 % yield, the olefination of which led to ()-indoxamycin E ((1’’E)-2-epi-5) in 93 % yield. Scheme 8. Total synthesis of ()-indoxamycin E.
Preliminary biological evaluation of ()-indoxamycins A–F and analogues thereof Synthetic ()-indoxamycins A–F as well as their analogues (1’’Z)-2-epi-4, 35 and 50 were tested for antiproliferative activity against the HT-29 and A-549 tumor cell lines at concentrations of 0.1, 1, 10 and 100 mm, in five independent experiments for each condition by using tamoxifen as a positive control. Surprisingly, none of these compounds caused significant growth inhibition in these assays (Figure 2 and Figure 3). Instead, moderate activity against HT-29 tumor cell line was found for the analogue 35. Our result stands in stark contrast to the work reported by Sato and co-workers (for ()-indoxamycins A and F), however, is consistent with the findings of Carreira and co-workers (for indoxamycins A and B).[25] Further investigations in cytotoxicity assays need to be conducted to achieve a comprehensive biological evaluation of these structures. &
&
Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
Figure 2. Inhibition activity of ()-indoxamycins A–F and analogues against HT-29 colon tumor cell line. The % viability of the cells is shown following 24 h of incubation with the indicated conditions relative to untreated cells. Tam (tamoxifen) served as positive control for the assay.
6
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Paper [6] a) B. M. Trost, F. Rise, J. Am. Chem. Soc. 1987, 109, 3161 – 3163; b) H. Yamada, S. Aoyagi, C. Kibayashi, Tetrahedron Lett. 1997, 38, 3027 – 3030; c) C. H. Oh, C. Y. Rhim, M. Kim, D. I. Park, A. K. Gupta, Synlett 2005, 2694 – 2696; d) C. H. Oh, H. H. Jung, Tetrahedron Lett. 1999, 40, 1535 – 1538; e) C. H. Oh, H. H. Jung, J. S. Kim, S. W. Cho, Angew. Chem. 2000, 112, 768 – 771; Angew. Chem. Int. Ed. 2000, 39, 752 – 755; f) C. H. Oh, J. W. Han, J. S. Kim, S. Y. Um, H. H. Jung, W. H. Jang, H. S. Won, Tetrahedron Lett. 2000, 41, 8365 – 8369. [7] a) R. E. Ireland, D. Habich, D. W. Norbeck, J. Am. Chem. Soc. 1985, 107, 3271 – 3278; b) R. E. Ireland, D. W. Norbeck, J. Am. Chem. Soc. 1985, 107, 3279 – 3285; c) R. E. Ireland, D. W. Norbeck, G. S. Mandel, N. S. Mandel, J. Am. Chem. Soc. 1985, 107, 3285 – 3294; d) J. C. Gilbert, R. D. Selliah, J. Org. Chem. 1993, 58, 6255 – 6265; e) C. P. Dell, K. M. Khan, D. W. Knight, J. Chem. Soc. Perkin Trans. 1 1994, 341 – 349; f) U. Kazmaier, J. Org. Chem. 1996, 61, 3694 – 3699; g) Y. H. Chai, S. P. Hong, H. A. Lindsay, C. McFarland, M. C. McIntosh, Tetrahedron 2002, 58, 2905 – 2928. For a recent review, see: h) K. C. Majumdar, R. K. Nandi, Tetrahedron 2013, 69, 6921 – 6957. [8] a) S. Essig, S. Bretzke, R. Mller, D. Menche, J. Am. Chem. Soc. 2012, 134, 19362 – 19365; b) J. D. Pettigrew, R. P. Freeman, P. D. Wilson, Can. J. Chem. 2004, 82, 1640 – 1648; c) J. D. Pettigrew, J. A. Bexrud, R. P. Freeman, P. D. Wilson, Heterocycles 2004, 62, 445 – 452. [9] a) G. Stork, R. Mook Jr., J. Am. Chem. Soc. 1983, 105, 3720 – 3722; b) G. Stork, R. Mook, Jr., J. Am. Chem. Soc. 1987, 109, 2829 – 2831. [10] a) Y.-C. Qin, C. E. Stivala, A. Zakarian, Angew. Chem. 2007, 119, 7610 – 7613; Angew. Chem. Int. Ed. 2007, 46, 7466 – 7469; b) C. E. Stivala, A. Zakarian, J. Am. Chem. Soc. 2008, 130, 3774 – 3776; c) R. Araoz, D. Servent, J. Molg, B. I. Iorga, C. Fruchart-Gaillard, E. Benoit, Z. Gu, C. Stivala, A. Zakarian, J. Am. Chem. Soc. 2011, 133, 10499 – 10511. For a review on [3,3]-sigmatropic rearrangements in total synthesis, see: d) E. A. Ilardi, C. E. Stivala, A. Zakarian, Chem. Soc. Rev. 2009, 38, 3133 – 3148; e) Z. Gu, A. T. Herrmann, C. E. Stivala, A. Zakarian, Synlett 2010, 1717 – 1722. [11] a) B. M. Trost, D. C. Lee, F. Rise, Tetrahedron Lett. 1989, 30, 651 – 654; b) B. M. Trost, B. A. Czeskis, Tetrahedron Lett. 1994, 35, 211 – 214; c) A. Goeke, M. Sawamura, R. Kuwano, Y. Ito, Angew. Chem. 1996, 108, 686 – 687; Angew. Chem. Int. Ed. Engl. 1996, 35, 662 – 663; d) M. Hatano, M. Terada, K. Mikami, Angew. Chem. 2001, 113, 255 – 259; Angew. Chem. Int. Ed. 2001, 40, 249 – 253; e) M. Hatano, K. Mikami, Org. Biomol. Chem. 2003, 1, 3871 – 3873; f) M. Hatano, K. Mikami, J. Am. Chem. Soc. 2003, 125, 4704 – 4705; g) K. Mikami, M. Hatano, Proc. Natl. Acad. Sci. USA 2004, 101, 5767 – 5769. [12] For a selected review on divergent reactions on racemic mixtures, see: L. C. Millera, R. Sarpong, Chem. Soc. Rev. 2011, 40, 4550 – 4562. [13] For reviews on oxa-Micheal reactions, see: a) C. F. Nising, S. Brse, Chem. Soc. Rev. 2008, 37, 1218 – 1228; b) C. F. Nising, S. Brse, Chem. Soc. Rev. 2012, 41, 988 – 999. [14] L. Wang, P. Li, D. Menche, Angew. Chem. 2010, 122, 9456 – 9460; Angew. Chem. Int. Ed. 2010, 49, 9270 – 9273. [15] a) T. Hudlicky, M. G. Natchus, G. Sinai-Zingde, J. Org. Chem. 1987, 52, 4641 – 4644; b) D.-S. Hsu, C.-C. Liao, Org. Lett. 2003, 5, 4741 – 4743. [16] CCDC 1006161 (26) 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. [17] S. Chackal-Catoen, Y. Miao, W. D. Wilson, T. Wenzler, R. Brunb, D. W. Boykin, Bioorg. Med. Chem. 2006, 14, 7434 – 7445. [18] For applications in total synthesis, see: a) D. F. Taber, D. Teng, J. Org. Chem. 2002, 67, 1607 – 1612; b) K. M. Engstrom, M. R. Mendoza, M. Navarro-Villalobos, D. Y. Gin, Angew. Chem. 2001, 113, 1162 – 1164; Angew. Chem. Int. Ed. 2001, 40, 1128 – 1130; c) G. Majetich, J. S. Song, C. Ringold, G. A. Nemeth, M. G. Newton, J. Org. Chem. 1991, 56, 3973 – 3988; d) M. J. Pelc, A. Zakarian, Org. Lett. 2005, 7, 1629 – 1631; e) C.-D. Lu, A. Zakarian, Angew. Chem. 2008, 120, 6935 – 6937; Angew. Chem. Int. Ed. 2008, 47, 6829 – 6831; f) N. Hama, T. Matsuda, T. Sato, N. Chida, Org. Lett. 2009, 11, 2687 – 2690; g) E. A. Ilardi, M. J. Isaacman, Y.-C. Qin, S. A. Shelly, A. Zakarian, Tetrahedron 2009, 65, 3261 – 3269. [19] a) P. Bickart, F. W. Carson, J. Jacobus, E. G. Miller, K. Mislow, J. Am. Chem. Soc. 1968, 90, 4869 – 4876; b) R. Tang, K. Mislow, J. Am. Chem. Soc. 1970, 92, 2100 – 2104; c) D. A. Evans, G. C. Andrews, C. L. Sims, J. Am. Chem. Soc. 1971, 93, 4956 – 4957; d) D. A. Evans, G. C. Andrews, Acc. Chem. Res. 1974, 7, 147 – 155.
Figure 3. Inhibition activity of ()-indoxamycins A–F and analogues against A-549 colon tumor cell line. The % viability of the cells is shown following 24 h of incubation with the indicated conditions relative to untreated cells. Tam (tamoxifen) served as positive control for the assay.
Conclusion In summary we have developed a stereoselective approach for the divergent total synthesis of ()-indoxamycins A–F, which were accomplished in 16 steps (4.4 % overall yield), 21 steps (1.8 % overall yield), 17 steps (5.3 % overall yield), 18 steps (1.1 % overall yield), 16 steps (4.3 % overall yield) and 17 steps (5.6 % overall yield), respectively. The salient features of the strategy entail an Ireland–Claisen rearrangement, an enantioselective 1,6-enyne reductive cyclization, and a tandem 1,2-addition/oxa-Michael/methylenation reaction. The described synthesis unambiguously determined the relative and absolute configuration of these natural products, and also enabled their preliminary biological investigations. We believe the methodologies and strategies disclosed herein will find further application in the synthesis of other complex natural skeletons.
Acknowledgements This work was financially supported by NNSFC (21202144, J1210042), the New Teacher’s Fund for Doctor Stations, Ministry of Education (20120101120087), and Zhejiang University. We thank Prof. Guping Tang and Mr. Jun Zhou (Zhejiang University) for assistance with the cytotoxicity assays. Keywords: biological evaluation · enyne cyclization · structure elucidation · tandem reaction · total synthesis [1] W. Fenical, P. R. Jensen, Nat. Chem. Biol. 2006, 2, 666 – 673. [2] S. Sato, F. Iwata, T. Mukai, S. Yamada, J. Takeo, A. Abe, H. Kawahara, J. Org. Chem. 2009, 74, 5502 – 5509. [3] O. F. Jeker, E. M. Carreira, Angew. Chem. 2012, 124, 3531 – 3534; Angew. Chem. Int. Ed. 2012, 51, 3474 – 3477. [4] C. He, C. Zhu, Z. Dai, C.-C. Tseng, H. Ding, Angew. Chem. 2013, 125, 13498 – 13502; Angew. Chem. Int. Ed. 2013, 52, 13256 – 13260. [5] For reviews, see: a) R. A. Bunce, Tetrahedron 1995, 51, 13103 – 13159; b) K. C. Nicolaou, T. Montagnon, S. A. Snyder, Chem. Commun. 2003, 551 – 564; c) A. Padwa, Pure Appl. Chem. 2004, 76, 1933 – 1952. Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
These are not the final page numbers! ÞÞ
7
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
Full Paper [20] a) H. L. Riley, J. F. Morley, N. A. C. Friend, J. Chem. Soc. 1932, 1875 – 1883; b) G. R. Waitkins, C. W. Clark, Chem. Rev. 1945, 36, 235 – 289; c) A. Frstner, T. Gastner, Org. Lett. 2000, 2, 2467 – 2470. [21] G. Blay, L. Cardona, B. Garca, J. R. Pedro, J. Org. Chem. 1993, 58, 7204 – 7208. [22] a) A. Nakamura, M. Nakada, Synthesis 2013, 1421 – 1451; b) N. C. Wilde, M. Isomura, A. Mendoza, P. S. Baran, J. Am. Chem. Soc. 2014, 136, 4909 – 4912. [23] Other conditions such as [Pd(OAc)2, PPh3, Et3N, HCO2H], [Pd(PPh3)4, LiCl, nBu3SnH] and [Pd(dppf)Cl2, dppf, Et3SiH] led to the recovery of 43.
&
&
Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
[24] a) R. L. E. Furln, E. G. Mata, O. A. Mascaretti, J. Chem. Soc. Perkin Trans. 1 1998, 355 – 358; b) R. L. E. Furln, E. G. Mata, O. A. Mascaretti, C. Pena, M. P. Coba, Tetrahedron 1998, 54, 13023 – 13034; c) K. C. Nicolaou, A. A. Estrada, M. Zak, S. H. Lee, B. S. Safina, Angew. Chem. 2005, 117, 1402 – 1406; Angew. Chem. Int. Ed. 2005, 44, 1378 – 1382. [25] O. F. Jeker, Ph.D. Thesis, ETH Zrich (Switzerland), 2013.
Received: June 16, 2014 Published online on && &&, 0000
8
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Paper
FULL PAPER & Total Synthesis
One for all: A stereoselective and divergent total synthesis of ()-indoxamycins A–F has been achieved based on an Ireland–Claisen rearrangement, an enantioselective 1,6-enyne reductive cyclization, and a tandem 1,2-addition/oxa-Michael/ methylenation reaction. The synthesis unambiguously determined the relative and absolute configuration of these natural products, and enabled their preliminary biological investigations.
Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
These are not the final page numbers! ÞÞ
C. He, C. Zhu, B. Wang, H. Ding* && – && Stereoselective Total Synthesis and Structural Elucidation of ()-Indoxamycins A–F
9
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
Full Paper
& Total Synthesis
Stereoselective Total Synthesis and Structural Elucidation of ()-Indoxamycins A–F Chi He, Chenlong Zhu, Bingnan Wang, and Hanfeng Ding*[a]
Abstract: In this study, a concise and stereoselective approach for the divergent total synthesis of ()-indoxamycins A–F is described. The key steps of the strategy include an Ireland–Claisen rearrangement, an enantioselective 1,6enyne reductive cyclization, and a tandem 1,2-addition/oxa-
Michael/methylenation reaction. The relative and absolute configuration of these natural products has been unambiguously elucidated, and their cytotoxic activities against HT-29 and A-549 tumor cell lines are also reported.
Introduction
()-Indoxamycins A–F (1–6) were tested for antiproliferative activity in the HT-29 tumor cell line. Among this family, ()-indoxamycins A (1) and F (6) were reported to cause significant growth inhibition with IC50 values of 0.59 and 0.31 mm, respectively. The relative and absolute structures of ()-indoxamycins were initially assigned by combined spectral and chemical methods, which revealed the presence of an unprecedented [5,5,6] tricyclic cage-like carbon framework, an a,b-unsaturated carboxylic acid side chain, and a trisubstituted alkene appendage. Three of the six contiguous chiral centers embedded in the congested pentamethylindenyl core are quaternary, of which two are vicinal. The unique chemical structure and potential anticancer activity of ()-indoxamycins prompted us to explore their laboratory synthesis and biological function. In early 2012, Carreira and co-workers reported an exquisite approach for the total synthesis of rac-indoxamycin B.[3] Recently, our group disclosed a divergent route to indoxamycins A, C, and F.[4] Herein, we detail our efforts on the total synthesis of ()-indoxamycins A–F, as well as the structural elucidation and preliminary biological evaluation of these natural products.
Over the past fifty years, the actinomycetes have continuously provided a variety of bioactive secondary metabolites, leading to the formation of more than half of all microbial antibiotics discovered to date.[1] Although this started several decades ago, most efforts in small-molecule discovery have mainly focused on soil-derived actinomycetes. In contrast, microorganisms derived from marine plants and animals have only recently received significant attention from chemists as a rich source of structurally diverse natural products that possess potent biological activities. In 2009, Sato and co-workers isolated a novel class of polyketides from saline cultures of marine-derived actinomycete strain NPS-643, which they subsequently named ()-indoxamycins A–F (Figure 1).[2]
Results and Discussion Total synthesis of ()-indoxamycins A, C, and F Our retrosynthetic analysis is depicted in Scheme 1. By looking into the structures carefully, indoxamycins A, C, and F were conceived to be accessed divergently from a common latestage precursor 7. Construction of 7 might be achieved through a substrate-controlled tandem reaction[5] involving a 1,2-addition/oxa-Michael/methylenation sequence, which would build up the [5,5,6] tricyclic ring system of the molecule. The desired enone-aldehyde intermediate 8 was expected to arise from 1,6-dienyne 9 through a palladium-catalyzed reductive cyclization.[6] Compound 9 could be furnished with the introduction of the two vicinal quaternary centers at C(2a) and C(7b) by an Ireland–Claisen rearrangement[7] from allyl ester
Figure 1. Structures of ()-indoxamycins A–F originally reported by Sato et al.
[a] C. He, C. Zhu, B. Wang, Prof. Dr. H. Ding Department of Chemistry, Zhejiang University 148 Tianmushan Road, Hangzhou 310028 (China) E-mail: [email protected] Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/chem.201403986. Chem. Eur. J. 2014, 20, 1 – 9
These are not the final page numbers! ÞÞ
1
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
Full Paper tained in 90 % yield through esterification of 11 and 12. Under optimized conditions, Ireland–Claisen rearrangement of 10 afforded 13 in 85 % yield as a 1.5:1 mixture of diastereomers. With the smooth construction of the two vicinal quaternary centers at C(2a) and C(7b), subsequent allylic oxidation of 13 followed by DDQ dehydrogenation afforded dienone 14 in 71 % yield over two steps. After protecting group (TMS) removal, dienyne 9 was formed in 85 % yield, which was ready for the next enyne cyclization. In view of the failures in radical approaches,[9] transitionmetal catalyzed reductive cyclization[6] was considered as an alternative solution. Although there were no literature precedents for the reductive cyclizations of cyclohexadienone-containing terminal alkynes, we were delighted to obtain the bicyclic product 15 in 87 % yield as a single diastereomer through simple modification of the standard conditions reported by Trost [Pd2(dba)3, P(o-tol)3, Et3SiH, AcOH, benzene, RT].[6a] The stereochemical outcome of the two methyl groups at C(2a) and C(7b) of bicyclo 15 with syn-relationship was presumably the result of synergistic effects from both the chairlike transi-
Scheme 1. Retrosynthetic analysis of indoxamycins A, C, and F.
10; the latter was easily prepared from the readily available intermediates 11 and 12.[8] Our synthesis endeavor started with the construction of the common precursor 7 (Scheme 2). Allyl ester 10 was first ob-
Scheme 2. Synthesis of the common precursor 7 and ORTEP drawing of tetracyclic compound 26. Ellipsoids set at 40 %. EDC·HCl = N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride, 4-DMAP = 4-(dimethylamino)pyridine, 3,5-DMP = 3,5-dimethylpyrazole, DDQ = 2,3-dichloro-5,6-dicyano-1,4-benzoquinone, Dibal-H = diisobutylaluminium hydride, DMP = Dess–Martin periodinane, Eschenmoser’s salt = N,N-dimethylmethyleneiminium chloride, DME = 1,2-dimethoxyethane.
&
&
Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
2
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Paper tion state (9’) and the formation of thermodynamically more stable cis-5-6-fused ring system. In order to achieve an asymmetric total synthesis, we initially planned to take advantage of the diastereoselective Ireland– Claisen rearrangement that had been developed by Zakarian and co-workers (Scheme 3).[10] Although highly stereoselective enolizations could be achieved by individual treatment of both diastereomers of the enantiopure ester ()-10 a and ()-10 b with chiral Koga-type bases, only slight improvements in diastereoselectivity (up to 2.6:1 d.r.) were obtained, presumably due to the uncontrollability of the chairlike transition state. Compared to the excellent acyclic stereocontrol, moderate results for the reactions of similar esters of cyclic alcohols were also observed by Zakarian et al.[10e]
Table 1. Optimization of stereodivergent 1,6-enyne reductive cyclization.[a]
Entry
[Pd]
Ligand
Time [h]
(+)-15, 2a-epi-15 Yield [%][b] (ee)[c]
1[d] 2[d] 3[d] 4[d] 5[d] 6 7 8 9 10 11 12 13 14[f]
Pd(OAc)2 Pd(dppf)Cl2 Pd(tfa)2 Pd2(dba)3 Pd(tfa)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2 [Pd((MeCN)4)](BF4)2
L1 L1 L2 L3 L3 L3 L1 L4 L5 L6 L7 L8 L9 L3
24 24 10 24 10 0.5 1 1 1 2 24 0.5 0.5 0.5
18 (27), n.d.[e] (–) 16 (23), n.d. (–) 45 (29), 36 (21) 15 (65), n.d. (–) 40 (72), 32 (64) 42 (84), 33 (71) 35 (48), 26 (43) 38 (41), 33 (35) 37 (49), 28 (30) 33 (35), 25 (27) n.d. (–), n.d. (–) 32 (25), 24 (22) 39 (43), 34 (41) 46 (93), 43 (80)
[a] Reaction conditions: 9 (0.1 mmol), [Pd] (0.025 equiv), ligand (0.05 equiv), AcOH (2.0 equiv), and Et3SiH (1.5 equiv) in DMSO at 25 8C. [b] Isolated yields. [c] Determined by HPLC analysis. [d] Benzene was used as solvent. [e] Not determined. [f] HCO2H (2.0 equiv) was used in place of AcOH. L1 = (R)-binap, L2 = (R)-Tol-binap, L3 = (R)-segphos, L4 = (R)-3,5xylyl-binap, L5 = (R)-H8-binap, L6 = (R)-MeO-Duphos, L7 = (R)-diop, L8 = (R)dtbm-segphos, L9 = (R)-MeO-dtbm-biphep, tfa = trifluoroacetate.
reduction of 17 was well controlled in the presence of Dibal-H at low temperature to afford diol 19 cleanly, presumably via the chelated alkoxyaluminium intermediate 18. Oxidation of the 19 with freshly prepared Dess–Martin periodinane (DMP) gave 8 in 76 % yield over the two steps. With the enone-aldehyde precursor 8 secured, we turned our attention to the construction of the tricyclic framework of the molecule. After extensive optimization, the expected tandem 1,2-addition/oxa-Michael[13]/methylenation reaction promoted by the Grignard reagent (20) proceeded smoothly to afford the tricyclic enone 24 with no detectable C(2) diastereomers. It is worth noting that oxa-Michael reactions have rarely been considered in the design of domino processes due to the inherent instability of the in situ generated enolate, which would result in elimination or retro-Michael side reactions. The only recent example was reported by Menche and co-workers, who developed a tandem oxa-Michael/Tsuji–Trost reaction leading to the concise synthesis of functionalized tetrahydropyranes.[14] We proposed that the instability of intermediate 23 might be the driving force to facilitate the spontaneous elimination; this gave excellent yield of the product without prior oxidation or addition of any base. Subsequently, isomerization of the double bond of 24 from D4, 9 to D3, 4 under acid catalysis[15] afforded the desired product (1’’E)-7 in 82 % yield (93 % ee), the ee value of which could be further improved to 96 % through recrystallization. To establish the relative and absolute stereochemistry, compound 25 was obtained in 84 % yield by quenching the
Scheme 3. Optimization of diastereoselective Ireland–Claisen rearrangement.
We then moved on to the development of an enantioselective version of the reductive 1,6-enyne cyclization.[11] Inspired by Ito and Mikami’s work regarding the asymmetric Alder-ene reactions, the chiral C2-symmetric bidentate phosphorus ligands were employed. Preliminary screens (Table 1, entries 1– 5) showed Pd(tfa)2/(R)-segphos was an effective catalyst system, with the stereodivergent formation[12] of the enantioenriched bicyclic product (+)-15 (40 % yield, 72 % ee) and 2a-epi-15 (32 % yield, 64 % ee). Furthermore, accelerating the reaction by the cationic PdII catalyst, [Pd(MeCN)4](BF4)2, in DMSO in the presence of (R)-segphos afforded (+)-15 in 42 % yield and 84 % ee (entry 6). However, displacement of segphos with other ligands only gave diminished yields and ee values (Table 1, entries 7–13). Pleasingly, excellent enantioselectivity (93 % ee) and yield (46 %) of (+)-15 were obtained by using the catalyst combination [Pd(MeCN)4](BF4)2/(R)-segphos and formic acid as the additive,[6d, e] accompanied by the formation of 2a-epi-15 in 43 % yield and 80 % ee (entry 14). With successful preparation of (+)-15, a two-step installation of the quaternary center at C(5) delivered 17 in 78 % yield, together with the formation of corresponding methyl enol ether (ca. a 5:1 mixture), which can be recycled through hydrolysis upon treatment with p-TsOH·H2O. Subsequent regioselective Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
These are not the final page numbers! ÞÞ
3
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
Full Paper ment of the configuration at C(6). However, exhaustive trials to invert the configuration, such as Mitsunobu reaction or sequential oxidation and reduction, met with failure. Finally, we opted for the sulfoxide–sulfenate rearrangement.[18] Treatment of 32 with benzenesulfenyl chloride gave allylic sulfoxide 36 in 92 % yield. We were then delighted to find that the Mislow– Evans rearrangement[19] took place unexpectedly during HWE olefination, presumably induced by nucleophilic phosphite reagent, which might act as a thiophile. The desired allylic alcohol 37 was therefore obtained in only one step and 95 % yield. In the event, saponification of 37 afforded (1’’E)-2,6-di-epi-6 in 92 % yield. The synthetic samples display identical spectral and physical properties with those reported for natural ()-indoxamycins A, C and F, respectively;[2] this unambiguously confirmed their relative and absolute configuration.
tandem reaction with aqueous NH4Cl prior to the methylenation. Although solid in racemic form, enantioenriched 25 was unable to afford single crystal suitable for X-ray crystallography. Tetracycle 26 was thus formed in 42 % yield through acidinduced lactonization of 25, the absolute configuration of the former was unambiguously assigned by X-ray crystallographic analysis[16] (see the structure in Scheme 2). This also proved our initial prediction for the relative configuration at C(2), which was based on a Si-face attack of the Grignard reagent 20 on the aldehyde moiety of 8. The late stages of the synthesis are outlined in Scheme 4, which entails a divergent route to the revised structures of ()-indoxamycins A, C, and F. 1,4-Reduction of (1’’E)-7 with LSelectride followed by capture of the resulting enolate with PhNTf2 gave enol-triflate 27 in 79 % yield. Subsequent palladium-catalyzed reductive detriflation provided alkene 28 in 85 % yield. Controlled reduction (Dibal-H, 105 8C) of 28 to aldehyde 29 followed by Horner–Wadsworth–Emmons (HWE) olefination smoothly afforded (1’’E)-2-epi-1 in 77 % yield over the two steps. Starting from the same enone ester (1’’E)-7, sequential reduction gave allylic alcohol 32 in 84 % yield as a single diastereomer. Allylic chloro-dehydroxylation of 32 by treatment with SOCl2 afforded allylic chloride 33 in 97 % yield. Finally, olefination and hydrolysis of the resulting enoic acid 34 in aqueous AgNO3[17] provided (1’’E)-2-epi-3 in 78 % yield over the two steps. To synthesize ()-indoxamycin F, compound 35 was initially constructed from intermediate 32 through HWE olefination in 83 % yield. Unfortunately, the NMR spectra of 35 deviated from those reported for natural indoxamycin F. Careful reexamination of the NMR spectroscopy data revealed the misassign-
Total synthesis of ()-indoxamycin D With sufficient quantity of ()-indoxamycin A in hand, the biomimetic routes to the other indoxamycin family members were investigated. It was found that Riley oxidation[20] of ()-indoxamycin A methyl ester (38) followed by hydrolysis afforded ()-indoxamycin D ((1’’E)-2-epi-4) and (1’’Z)-2-epi-4 in 25 and 49 % yield, respectively (Scheme 5). A similar result (72 % yield, 1:2 d.r.) was also obtained from direct allylic oxidation of ()-indoxamycin A. On the other hand, singlet oxygen ene reaction[21] of 38 gave allylic alcohol 39 and 2’’-epi-39 in 61 % combined yield as a mixture of diastereomers (ca. 2.8:1 d.r.). To our disappointment, despite further experimentation, the other indoxamycin family members could not be accessed since the D1’’, 2’’ double bond seems the most reactive site
Scheme 4. Divergent total synthesis of ()-indoxamycins A, C, and F. Tf = trifluoromethanesulfonyl, L-Selectride = lithium tri-sec-butylborohydride.
&
&
Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
4
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Paper 60 8C] delivered alkene 44 in 80 % yield.[23] Since straightforward reduction of 44 to aldehyde 45 failed, an alternative reduction and oxidation sequence was then executed. HWE olefination of 45 gave 46 in 84 % yield over three steps from 44. After being converted to acetate 48, acid-catalyzed isomerization of the exo-double bond underwent smoothly to provide 49 (70 % yield), the saponification of which afforded acid 50 in 91 % yield. Surprisingly, the spectroscopic data (1H and 13C NMR) of 50 did not match those reported for natural ()-indoxamycin B.[2, 3] Careful inspection of the whole sequence led to a conclusion that the configuration at C(5) had been inverted unexpectedly during the reduction process (44!45), which can be supported by detailed NOESY analysis of 41 and the formation of tetracyclic ether 47 through etherification of 46 in 55 % yield. Faced with this problematic issue, a revised synthetic route was put forward (Scheme 7). Reduction of 44 yielded 51 with complete TBS migration, which was then converted to acetate 52 in 98 % yield. Subsequent TBS deprotection proved challenging, various conditions such as TBAF, p-TsOH·H2O, AcOH, HCl (aq.), pyridinium p-toluenesulfonate gave either a bad ratio or trace amount of the desired product. Fortunately, upon treatment of 52 with HF·pyridine in MeCN at 10 8C, a 5:1 mixture of 53 and 5-epi-53 were obtained in 84 % combined yield. The ratio was sensitive to temperature, a decreased selectivity (1.5:1, 53/5-epi-53) was observed by warming the reaction to 25 8C. With enough 53 in hand, DMP oxidation followed by HWE olefination afforded 55 in 83 % yield over two steps. The following exo-double bond isomerization of 55 only succeeded with excess amount of p-TsOH·H2O (15 equiv) at 100 8C, as prolonged reaction time with catalytic acid resulted in decomposition of the starting material. To our delight, hydrolysis of both the two esters under mild conditions (Me3SnOH, 1,2-dichloroethane, 84 8C)[24] afforded ()-indoxamycin B ((1’’E)-2-epi-2) in 55 % yield over two steps. When treated with aqueous LiOH,
among the four existing double bonds in the skeleton toward allylic oxidation. A variety of conditions[22] such as (CrO3, 3,5DMP), [Pd(OH)2, Cs2CO3, tBuOOH], [PhI(tfa)2, Cs2CO3, tBuOOH], [Mn(OAc)3, tBuOOH], (CuI, tBuOOH), [Pd(OAc)2, p-benzoquinone, anisole, AcOH] and (NBS, benzoyl peroxide) either gave no reaction or caused decomposition of 38.
Scheme 5. Allylic oxidations of ()-indoxamycin A methyl ester (38).
Total synthesis of ()-indoxamycin B With the divergent concept in mind and following the established route described for ()-indoxamycin A, the initial synthetic strategy toward ()-indoxamycin B is shown in Scheme 6. Aldol reaction of 16 with formalin afforded the unstable alcohol 40 in 88 % yield. Immediate silylation of 40 gave 41 in 93 % yield, whereas other methods for the hydroxyl protection under basic conditions led to the regeneration of 16 through a retro-aldol process. By employing previously developed procedures, tricyclic triflate 43 was obtained uneventfully in 50 % yield over four steps from 41. Subsequent detriflation of 43 under optimized conditions [Pd(OAc)2, dppf, Et3SiH, DMF,
Scheme 6. Initial strategy toward total synthesis of ()-indoxamycin B. Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
These are not the final page numbers! ÞÞ
5
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
Full Paper
Scheme 7. Total synthesis of ()-indoxamycin B.
the acetate was removed first, which led to further fragmentation through the vinylogous retro-aldol reaction pathway.[25]
Total synthesis of ()-indoxamycin E Bicycle 8 was initially employed for the tandem 1,2-addition/ oxa-Michael/methylenation reaction promoted by organolithium reagent 57, which delivered an inseparable mixture of 59 and 2-epi-59 in 74 % yield (3:1 d.r.). To achieve a satisfactory diastereoselectivity, bicycle 56 with the endo-double bond at D3, 4 was chosen as our starting point in the synthetic approach to ()-indoxamycin E (Scheme 8). Prepared from 8 in 76 % yield, the tandem reaction of 56 afforded tricyclic enone 58 in 78 % yield as a single diastereomer. In contrast, the corresponding Grignard reagent did not promote the reaction at all. Conjugate reduction of 58 followed by in situ triflation delivered 60 in 82 % yield, which then gave rise to 61 in 95 % yield under palladium-catalyzed reductive conditions. Controlled reduction of ester 61 furnished aldehyde 62 in 80 % yield, the olefination of which led to ()-indoxamycin E ((1’’E)-2-epi-5) in 93 % yield. Scheme 8. Total synthesis of ()-indoxamycin E.
Preliminary biological evaluation of ()-indoxamycins A–F and analogues thereof Synthetic ()-indoxamycins A–F as well as their analogues (1’’Z)-2-epi-4, 35 and 50 were tested for antiproliferative activity against the HT-29 and A-549 tumor cell lines at concentrations of 0.1, 1, 10 and 100 mm, in five independent experiments for each condition by using tamoxifen as a positive control. Surprisingly, none of these compounds caused significant growth inhibition in these assays (Figure 2 and Figure 3). Instead, moderate activity against HT-29 tumor cell line was found for the analogue 35. Our result stands in stark contrast to the work reported by Sato and co-workers (for ()-indoxamycins A and F), however, is consistent with the findings of Carreira and co-workers (for indoxamycins A and B).[25] Further investigations in cytotoxicity assays need to be conducted to achieve a comprehensive biological evaluation of these structures. &
&
Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
Figure 2. Inhibition activity of ()-indoxamycins A–F and analogues against HT-29 colon tumor cell line. The % viability of the cells is shown following 24 h of incubation with the indicated conditions relative to untreated cells. Tam (tamoxifen) served as positive control for the assay.
6
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Paper [6] a) B. M. Trost, F. Rise, J. Am. Chem. Soc. 1987, 109, 3161 – 3163; b) H. Yamada, S. Aoyagi, C. Kibayashi, Tetrahedron Lett. 1997, 38, 3027 – 3030; c) C. H. Oh, C. Y. Rhim, M. Kim, D. I. Park, A. K. Gupta, Synlett 2005, 2694 – 2696; d) C. H. Oh, H. H. Jung, Tetrahedron Lett. 1999, 40, 1535 – 1538; e) C. H. Oh, H. H. Jung, J. S. Kim, S. W. Cho, Angew. Chem. 2000, 112, 768 – 771; Angew. Chem. Int. Ed. 2000, 39, 752 – 755; f) C. H. Oh, J. W. Han, J. S. Kim, S. Y. Um, H. H. Jung, W. H. Jang, H. S. Won, Tetrahedron Lett. 2000, 41, 8365 – 8369. [7] a) R. E. Ireland, D. Habich, D. W. Norbeck, J. Am. Chem. Soc. 1985, 107, 3271 – 3278; b) R. E. Ireland, D. W. Norbeck, J. Am. Chem. Soc. 1985, 107, 3279 – 3285; c) R. E. Ireland, D. W. Norbeck, G. S. Mandel, N. S. Mandel, J. Am. Chem. Soc. 1985, 107, 3285 – 3294; d) J. C. Gilbert, R. D. Selliah, J. Org. Chem. 1993, 58, 6255 – 6265; e) C. P. Dell, K. M. Khan, D. W. Knight, J. Chem. Soc. Perkin Trans. 1 1994, 341 – 349; f) U. Kazmaier, J. Org. Chem. 1996, 61, 3694 – 3699; g) Y. H. Chai, S. P. Hong, H. A. Lindsay, C. McFarland, M. C. McIntosh, Tetrahedron 2002, 58, 2905 – 2928. For a recent review, see: h) K. C. Majumdar, R. K. Nandi, Tetrahedron 2013, 69, 6921 – 6957. [8] a) S. Essig, S. Bretzke, R. Mller, D. Menche, J. Am. Chem. Soc. 2012, 134, 19362 – 19365; b) J. D. Pettigrew, R. P. Freeman, P. D. Wilson, Can. J. Chem. 2004, 82, 1640 – 1648; c) J. D. Pettigrew, J. A. Bexrud, R. P. Freeman, P. D. Wilson, Heterocycles 2004, 62, 445 – 452. [9] a) G. Stork, R. Mook Jr., J. Am. Chem. Soc. 1983, 105, 3720 – 3722; b) G. Stork, R. Mook, Jr., J. Am. Chem. Soc. 1987, 109, 2829 – 2831. [10] a) Y.-C. Qin, C. E. Stivala, A. Zakarian, Angew. Chem. 2007, 119, 7610 – 7613; Angew. Chem. Int. Ed. 2007, 46, 7466 – 7469; b) C. E. Stivala, A. Zakarian, J. Am. Chem. Soc. 2008, 130, 3774 – 3776; c) R. Araoz, D. Servent, J. Molg, B. I. Iorga, C. Fruchart-Gaillard, E. Benoit, Z. Gu, C. Stivala, A. Zakarian, J. Am. Chem. Soc. 2011, 133, 10499 – 10511. For a review on [3,3]-sigmatropic rearrangements in total synthesis, see: d) E. A. Ilardi, C. E. Stivala, A. Zakarian, Chem. Soc. Rev. 2009, 38, 3133 – 3148; e) Z. Gu, A. T. Herrmann, C. E. Stivala, A. Zakarian, Synlett 2010, 1717 – 1722. [11] a) B. M. Trost, D. C. Lee, F. Rise, Tetrahedron Lett. 1989, 30, 651 – 654; b) B. M. Trost, B. A. Czeskis, Tetrahedron Lett. 1994, 35, 211 – 214; c) A. Goeke, M. Sawamura, R. Kuwano, Y. Ito, Angew. Chem. 1996, 108, 686 – 687; Angew. Chem. Int. Ed. Engl. 1996, 35, 662 – 663; d) M. Hatano, M. Terada, K. Mikami, Angew. Chem. 2001, 113, 255 – 259; Angew. Chem. Int. Ed. 2001, 40, 249 – 253; e) M. Hatano, K. Mikami, Org. Biomol. Chem. 2003, 1, 3871 – 3873; f) M. Hatano, K. Mikami, J. Am. Chem. Soc. 2003, 125, 4704 – 4705; g) K. Mikami, M. Hatano, Proc. Natl. Acad. Sci. USA 2004, 101, 5767 – 5769. [12] For a selected review on divergent reactions on racemic mixtures, see: L. C. Millera, R. Sarpong, Chem. Soc. Rev. 2011, 40, 4550 – 4562. [13] For reviews on oxa-Micheal reactions, see: a) C. F. Nising, S. Brse, Chem. Soc. Rev. 2008, 37, 1218 – 1228; b) C. F. Nising, S. Brse, Chem. Soc. Rev. 2012, 41, 988 – 999. [14] L. Wang, P. Li, D. Menche, Angew. Chem. 2010, 122, 9456 – 9460; Angew. Chem. Int. Ed. 2010, 49, 9270 – 9273. [15] a) T. Hudlicky, M. G. Natchus, G. Sinai-Zingde, J. Org. Chem. 1987, 52, 4641 – 4644; b) D.-S. Hsu, C.-C. Liao, Org. Lett. 2003, 5, 4741 – 4743. [16] CCDC 1006161 (26) 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. [17] S. Chackal-Catoen, Y. Miao, W. D. Wilson, T. Wenzler, R. Brunb, D. W. Boykin, Bioorg. Med. Chem. 2006, 14, 7434 – 7445. [18] For applications in total synthesis, see: a) D. F. Taber, D. Teng, J. Org. Chem. 2002, 67, 1607 – 1612; b) K. M. Engstrom, M. R. Mendoza, M. Navarro-Villalobos, D. Y. Gin, Angew. Chem. 2001, 113, 1162 – 1164; Angew. Chem. Int. Ed. 2001, 40, 1128 – 1130; c) G. Majetich, J. S. Song, C. Ringold, G. A. Nemeth, M. G. Newton, J. Org. Chem. 1991, 56, 3973 – 3988; d) M. J. Pelc, A. Zakarian, Org. Lett. 2005, 7, 1629 – 1631; e) C.-D. Lu, A. Zakarian, Angew. Chem. 2008, 120, 6935 – 6937; Angew. Chem. Int. Ed. 2008, 47, 6829 – 6831; f) N. Hama, T. Matsuda, T. Sato, N. Chida, Org. Lett. 2009, 11, 2687 – 2690; g) E. A. Ilardi, M. J. Isaacman, Y.-C. Qin, S. A. Shelly, A. Zakarian, Tetrahedron 2009, 65, 3261 – 3269. [19] a) P. Bickart, F. W. Carson, J. Jacobus, E. G. Miller, K. Mislow, J. Am. Chem. Soc. 1968, 90, 4869 – 4876; b) R. Tang, K. Mislow, J. Am. Chem. Soc. 1970, 92, 2100 – 2104; c) D. A. Evans, G. C. Andrews, C. L. Sims, J. Am. Chem. Soc. 1971, 93, 4956 – 4957; d) D. A. Evans, G. C. Andrews, Acc. Chem. Res. 1974, 7, 147 – 155.
Figure 3. Inhibition activity of ()-indoxamycins A–F and analogues against A-549 colon tumor cell line. The % viability of the cells is shown following 24 h of incubation with the indicated conditions relative to untreated cells. Tam (tamoxifen) served as positive control for the assay.
Conclusion In summary we have developed a stereoselective approach for the divergent total synthesis of ()-indoxamycins A–F, which were accomplished in 16 steps (4.4 % overall yield), 21 steps (1.8 % overall yield), 17 steps (5.3 % overall yield), 18 steps (1.1 % overall yield), 16 steps (4.3 % overall yield) and 17 steps (5.6 % overall yield), respectively. The salient features of the strategy entail an Ireland–Claisen rearrangement, an enantioselective 1,6-enyne reductive cyclization, and a tandem 1,2-addition/oxa-Michael/methylenation reaction. The described synthesis unambiguously determined the relative and absolute configuration of these natural products, and also enabled their preliminary biological investigations. We believe the methodologies and strategies disclosed herein will find further application in the synthesis of other complex natural skeletons.
Acknowledgements This work was financially supported by NNSFC (21202144, J1210042), the New Teacher’s Fund for Doctor Stations, Ministry of Education (20120101120087), and Zhejiang University. We thank Prof. Guping Tang and Mr. Jun Zhou (Zhejiang University) for assistance with the cytotoxicity assays. Keywords: biological evaluation · enyne cyclization · structure elucidation · tandem reaction · total synthesis [1] W. Fenical, P. R. Jensen, Nat. Chem. Biol. 2006, 2, 666 – 673. [2] S. Sato, F. Iwata, T. Mukai, S. Yamada, J. Takeo, A. Abe, H. Kawahara, J. Org. Chem. 2009, 74, 5502 – 5509. [3] O. F. Jeker, E. M. Carreira, Angew. Chem. 2012, 124, 3531 – 3534; Angew. Chem. Int. Ed. 2012, 51, 3474 – 3477. [4] C. He, C. Zhu, Z. Dai, C.-C. Tseng, H. Ding, Angew. Chem. 2013, 125, 13498 – 13502; Angew. Chem. Int. Ed. 2013, 52, 13256 – 13260. [5] For reviews, see: a) R. A. Bunce, Tetrahedron 1995, 51, 13103 – 13159; b) K. C. Nicolaou, T. Montagnon, S. A. Snyder, Chem. Commun. 2003, 551 – 564; c) A. Padwa, Pure Appl. Chem. 2004, 76, 1933 – 1952. Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
These are not the final page numbers! ÞÞ
7
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
Full Paper [20] a) H. L. Riley, J. F. Morley, N. A. C. Friend, J. Chem. Soc. 1932, 1875 – 1883; b) G. R. Waitkins, C. W. Clark, Chem. Rev. 1945, 36, 235 – 289; c) A. Frstner, T. Gastner, Org. Lett. 2000, 2, 2467 – 2470. [21] G. Blay, L. Cardona, B. Garca, J. R. Pedro, J. Org. Chem. 1993, 58, 7204 – 7208. [22] a) A. Nakamura, M. Nakada, Synthesis 2013, 1421 – 1451; b) N. C. Wilde, M. Isomura, A. Mendoza, P. S. Baran, J. Am. Chem. Soc. 2014, 136, 4909 – 4912. [23] Other conditions such as [Pd(OAc)2, PPh3, Et3N, HCO2H], [Pd(PPh3)4, LiCl, nBu3SnH] and [Pd(dppf)Cl2, dppf, Et3SiH] led to the recovery of 43.
&
&
Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
[24] a) R. L. E. Furln, E. G. Mata, O. A. Mascaretti, J. Chem. Soc. Perkin Trans. 1 1998, 355 – 358; b) R. L. E. Furln, E. G. Mata, O. A. Mascaretti, C. Pena, M. P. Coba, Tetrahedron 1998, 54, 13023 – 13034; c) K. C. Nicolaou, A. A. Estrada, M. Zak, S. H. Lee, B. S. Safina, Angew. Chem. 2005, 117, 1402 – 1406; Angew. Chem. Int. Ed. 2005, 44, 1378 – 1382. [25] O. F. Jeker, Ph.D. Thesis, ETH Zrich (Switzerland), 2013.
Received: June 16, 2014 Published online on && &&, 0000
8
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Paper
FULL PAPER & Total Synthesis
One for all: A stereoselective and divergent total synthesis of ()-indoxamycins A–F has been achieved based on an Ireland–Claisen rearrangement, an enantioselective 1,6-enyne reductive cyclization, and a tandem 1,2-addition/oxa-Michael/ methylenation reaction. The synthesis unambiguously determined the relative and absolute configuration of these natural products, and enabled their preliminary biological investigations.
Chem. Eur. J. 2014, 20, 1 – 9
www.chemeurj.org
These are not the final page numbers! ÞÞ
C. He, C. Zhu, B. Wang, H. Ding* && – && Stereoselective Total Synthesis and Structural Elucidation of ()-Indoxamycins A–F
9
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
