Broad scope gold(I)-catalysed polyenyne cyclisations for the formation of up to four carbon–carbon bonds†

The polycyclisation of polyenynes catalyzed by gold(I) has been extended for the first time to the simultaneous formation of up to four carbon–carbon bonds, leading to steroid-like molecules with high stereoselectivity in a single step with low catalyst loadings. In addition to terminal alkynes, bromoalkynes can also be used as initiators of polyene cyclisations, giving rise to synthetically useful cyclic bromoalkenes.

We decided to explore the possibility of performing gold(I)catalysed cyclisations analogous to that of pentanene 7 but with a terminal alkyne instead of an alkene to form steroidlike products functionalised with an alkene at the A ring. As a first step towards this goal, we studied a set of 1,5-enynes substituted with different alcohols, phenols, arenes, and heteroarenes as potential nucleophiles. In addition to terminal alkynes, which were not broadly studied previously, 10,18 we decided to employ 1-bromo-1,5-enynes as the initiators of cyclisation. Surprisingly, bromoalkynes have seldom been used in gold-catalysed cyclisations. 19,20 At the outset, we examined the cyclisation of (E)-2,6dimethyldeca-1,5-dien-9-yne (9a) with gold(I) catalysts A-F bearing electronically different bulky groups (Table 1). In all cases, trans-fused hexahydronaphthalene 10a was cleanly obtained as the major product after 1 h by using just 1 mol% catalyst. As we have observed before in other contexts, the best yields were obtained with cationic gold(I) complexes bearing very bulky biphenylphosphine ligands (Buchwald ligands). 1j,21 In this particular instance, cationic dicyclohexylphosphinobiphenyl gold(I) complex C outperforms Johnphos, t-BuXphos, and Xphos complexes A, B, and D ( Table 1, entries 1-4).
Complex C (3 mol%) was used as the catalyst in the cyclisation of aryl substituted 1,5-enynes ( Table 2). The reaction of substrates 9b-f bearing electron-rich aromatic and heteroaromatic rings as cyclisation terminators proceeds to give products 10b-f as single diastereomers in good yields in all cases, with the exception of 10c, which was obtained in 54% yield ( Table 2, entries 1-5). Similar results have been obtained with different metal catalysts using 1,5-enynes analogous to 9b-c with a methyl substituent at the terminal alkyne. 10b,12b,13 The trans-relative configuration was confirmed by X-ray diffraction in the case of indole derivative 10f (Fig. 1), which has a carbon skeleton somewhat related to that of the alkaloids aristomakinine and aristomakine, although for these natural products a cis-hexalin structure has been assigned. 22 The gold(I)catalysed cyclisation of 1,5-enynes 9g-h substituted at C-5, gave spirocyclic derivatives 10g-h in good to excellent yields ( Table 2, entries 6 and 7). The structure of 10h was confirmed by X-ray diffraction (Fig. 1). We also tested several chiral gold(I) catalysts in the cyclisation of enynes 9d and 9f. 23,24 However, despite the excellent cyclisation yields, the enantioselectivities achieved with the dinuclear gold(I) complex of [MeO-DTB-BIPHEP] in the presence of AgNTf 2 were only moderate (54 and 48% ee, respectively) (Scheme 4). 25 The intramolecular addition of alcohols and phenols was briefly studied with substrates 9i-k ( Table 3). As expected considering the precedents, 10 products 10i-k were obtained in good to excellent yields. Spirocyclisations similar to that of 9i to form 10i could be applied for the synthesis of analogues of the natural product filifolinol and other more complex, biologically active compounds with a spirobenzofuran structure. 26 The cyclisation of 1-bromo-1,5-enynes 9l-n and 1-bromo-1,5,9-dienyne 9o with catalyst C took place uneventfully under the usual reaction conditions to give products 10l-o in good yields (Table 4). These results show that bromoalkynes are perfectly suitable initiators of gold(I)-catalysed polycyclisations. The final products are bromoalkenes, which could be further functionalized by metal-catalysed cross-couplings, carbonylations, or by other methods.
The polycyclisation of trienynes 9p-q and tetraenynes 9r-s was similarly performed with catalyst C (1-3 mol%) to give triand tetracyclic compounds 10p-s ( Table 5). Considering that four C-C bonds are formed in a single step, the catalytic transformations of tetraenynes 9r-s into 10r-s are quite remarkable and comparable to that achieved by Gagné in the Pt(II)-cata-lysed polycyclisation of pentanene 7 (Scheme 3). 15,27 However, in our case a lower catalyst loading is required and the final tetracyclic derivatives 10r-s feature two differently substituted double bonds. In 10s, the alkenyl bromide offers a handle for further functionalisation of the A-ring.
Presumably, the cyclisation of 9s to give 10s proceeds by the initial formation of gold(I)-carbene intermediate 5,11 which trig-Scheme 4 Gold(I)-catalysed enantioselective cyclisation of 9d and 9f.   gers a cascade process to form secondary carbocation 12 (Scheme 5). The final tetracyclic compound 10s is then formed by Wagner-Meerwein 1,2 H and Me migrations, 15,28 followed by the proton elimination and protonolysis of the alkenyl-gold(I) bond.
In summary, building upon previous studies, 9,10 we have extended the polycyclisation of polyenynes up to the simultaneous formation of four C-C bonds. These reactions are performed under mild conditions with low catalyst loadings (1-3 mol%) of a cationic dicyclohexylphosphinobiphenyl gold(I)complex with a weakly coordinating acetonitrile ligand. In addition to terminal alkynes, we have also found that bromoalkynes can be used as the initiators of polyene cyclisations, leading to synthetically useful cyclic bromoalkenes. Further work on the development of broad scope and practical solutions of the asymmetric polycyclisation of polyeneynes is underway.