Divergent Mechanisms for the Skeletal Rearrangement and [2+2] Cycloaddition of Enynes Catalyzed by Gold

Abstract

Support for the direct route: That cyclobutenes are not necessary intermediates in the skeletal rearrangement of enynes is supported by DFT calculations and kinetic studies. Cyclobutenes may arise from the corresponding syn-cyclopropylgold(I) carbenes (see scheme). Transition-metal-catalyzed reactions of 1,6-enynes proceed via two general pathways (Scheme 1).1, 2 If the metal coordinates selectively to the alkyne 1, cyclopropyl–metal carbenes 2 are initially formed, which can react with alcohols or water to give products of alkoxy- or hydroxycyclization,1, 2 whereas in the absence of nucleophiles, skeletal rearrangement forms dienes 3 (single cleavage) and/or 4 (double cleavage).1, 3 Alternatively, coordination of MXn to the alkyne and the alkene (as in 5) is followed by oxidative cyclometalation to form 6, which usually evolves by β-hydrogen elimination to give Alder–ene-type products.2 Formation of products 3 could also occur by conrotatory ring-opening of cyclobutenes 7,4, 5 which are formed either from 2 or by reductive elimination of 6. A pathway for the formation of 3 via ring-opening of 7 is favored by most authors.4, 6 However, the formation of dienes 4 requires a different mechanistic rationalization. An earlier mechanistic proposal by Oi et al.3 suggested a direct pathway for the skeletal rearrangement via intermediates of type 2. Herein we report experimental and theoretical results that shed new light into this complex mechanistic issue. In particular, this work strongly suggests that cyclobutenes 7 are not necessary intermediates in the skeletal rearrangement of enynes. Alder–ene-type products have not been observed in AuI-catalyzed reactions, which is consistent with the selective coordination of cationic [Au(L)]+ complexes to the alkyne.2c, 7, 8 In the presence of catalysts formed from 8 a–c and AgSbF6,7b0 or new cationic complexes 9 a, b, enyne 10 undergoes a single cleavage rearrangement to form 11 quantitatively at a temperature as low as −63 °C (Scheme 2). On the other hand, enyne 12 undergoes a double cleavage rearrangement with [Au(PPh3)]SbF6 to give exclusively 13.3, 9 These are the skeletal rearrangements occurring at the lowest temperatures. Reaction of enyne 10 with catalyst 9 a (−63 to −26 °C) or 9 b (−43 to −28 °C) was monitored by 1H NMR spectroscopy in CD2Cl2. Under these conditions, smooth and quantitative formation of diene 11 was observed without the build up of any intermediate. The rearrangement is pseudo-first order in 10, which allowed us to determine the thermodynamic parameters shown in Scheme 2. Z=C(CO2Me)2. Δ and ΔH≠ in kcal mol−1; ΔS≠ in cal K−1 mol−1. The large and negative activation entropies suggest that an associative ligand substitution10 (diene 11 by incoming enyne 10) is the rate-determining step of the process. These results establish a very low activation energy (Ea) for the hypothetical conrotatory ring-opening of a cyclobutene of type 7, which therefore should be a fast process at temperatures as low as −63 °C. This is not consistent with the ring-opening of bicycle 14 and its 6,7-dimethyl derivatives,11 for which activation energies of 29.0–32.7 kcal mol−1 and low entropies of activation (1.4–2.2 cal K−1 mol−1) have been determined.-1 DFT calculations predict an Ea of 25.6 kcal mol−1 for the conrotatory ring-opening of bicyclo[3.2.0]hept-5-ene (15) to 1-vinyl-1-cyclopentene (ΔG298 K=−22.5 kcal mol−1). It is important to note that 15 has a lower olefin strain (OS=16.7 kcal mol−1) than 14 (OS=20.5 kcal mol−1), which is stable up to 118 °C.12 Additional evidence against the ring-opening of a cyclobutene in the low temperature skeletal rearrangement of 10 is provided by the isolation of bicycle 16 as a stable compound.13 DFT calculations14 support pathways for the skeletal rearrangement that do not involve the intermediacy of cyclobutenes 7. Thus, complex 17 a evolves via TS1 to form cation 18, which could furnish dienes 3 by elimination of [Au(L)]+ (Scheme 3). Alternatively, a 1,2-shift gives gold carbene 19 a via TS2 in an almost flat potential surface. Dienes 4 would result from 19 a by β-hydrogen elimination and demetalation. In the case of 20, which is the intermediate in a reaction of an enyne of type 12, a double-cleavage rearrangement was found to give 19 b directly, in agreement with the experimental results. This remarkable process involves a 1,2-shift of a metal carbene with concomitant cleavage of the distal CC bond of the cyclopropane and formation of a double bond. L=PH3. ΔG at 298 K (energies in kcal mol−1). No direct pathway for the formation of a cyclobutene from 17 a was found. In contrast, syn-17′a forms 22 a via TS5, although the anti to syn isomerization from 17 a to 17′a requires a rather high activation energy (Scheme 4).14 This high activation energy of 24.7 kcal mol−1 can be attributed to the loss of conjugation between the gold carbene and the cyclopropane, as shown by the significant shortening of the cyclopropane and CAu bonds and the lengthening of the CC bond connecting the cyclopropane and the gold carbene in TS4. This isomerization process is rather unlikely under the reaction conditions, as the initially formed anti-17 a would undergo a more facile rearrangement via 18 (ΔG≠=9.1 kcal mol−1, Scheme 3). However, an alternative pathway has been found for a more direct formation of complexes 17′a, b by a syn-type attack of the alkene, via TS5, to the (alkyne)gold moiety of 21 a, b (Scheme 4). L=PH3. ΔG at 298 K (energies in kcal mol−1) and selected bond lengths [Å] for 17 a, 17′a, and TS4. Although the anti attack of the alkene is more favorable,7a the syn attack could compete if substitution at the alkene and/or the alkyne disfavors the skeletal rearrangement. In particular, this should be more favorable for the formation of bicyclo[3.2.0]oct-6-enes from 1,7-enynes, in accordance with the calculations (17′b→22 b, Scheme 4) and experiments.4 Significantly, cationic gold complexes catalyze the [2+2] cycloaddition of 1,7-enynes. Thus, enynes 23 and 24 react with complexes [Au(L)]+ at room temperature to give 25 and 26,4b,4d respectively (Scheme 5). Reactions of 23 and 24: a) 9 b (2 mol %), CH2Cl2, room temp., 14 h (80 %); b) 8 c (2 mol %), AgSbF6 (2 mol %), CH2Cl2, room temp., 45 min (67 %); c) PtCl2 (5 mol %), MeCN, 120 °C, 20 h (67 %). Tricycles 25 and 26 do not undergo ring-opening at 120–150 °C to form 1,3-dienes.15 To study the possible effect of transition metals in the ring-opening of the cyclobutene,16 25 was heated in MeCN at 120 °C in the presence of 5 mol % PtCl2 (Scheme 5). Interestingly, under these conditions, PtII,1, 3, 4d,4f,4g which is a known catalyst for the skeletal rearrangement, does not promote the ring-opening of the cyclobutene but rather promotes isomerization to form the less-strained tricycle 27.17 In summary, calculations on the AuI-catalyzed skeletal rearrangement of enynes support the earlier proposals suggested by Oi et al.3 and others,1, 4 although Scheme 3 provides a more rigorous and concise mechanistic picture. An alternative pathway has been found for the formation of cyclobutenes via syn-cyclopropyl–metal carbenes, formed by a syn electrophilic addition of the metal and the alkene to the alkyne. Kinetic experiments indicate that if a conrotatory ring-opening of a cyclobutene intervenes in the skeletal rearrangement, its Ea value would be unreasonably low. Supporting information for this article is available on the WWW under http://www.wiley-vch.de/contents/jc_2002/2005/z501937_s.pdf or from the author. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.