|Home | About | Journals | Submit | Contact Us | Français|
The Ficini [2 + 2] cycloaddition using N-sulfonyl substituted ynamides is described, featuring the utility of CuCl2 and AgSbF6 as catalysts. This work represents the first success example of ynamides participating in a thermal [2 + 2] cycloaddition with enones.
More than 40 years ago, Ficini1 disclosed perhaps the most useful carbon-carbon bond forming reaction involving ynamines2: a thermally driven stepwise [2 + 2] cycloaddition3 of ynamine  with cyclic enones, leading to the formation of cyclobutenamine 3 [Scheme 1].4-6 In the last 15 years, ynamides have emerged as a superior synthetic equivalent of ynamines.7,8 Beautiful chemistry in the area of [2 + 2] cycloadditions has followed by way of Tam's Ru-catalyzed ynamide-[2 + 2] cycloaddition of norbornene,9 Danhesier's thermal cycloaddition of ketenes,10 and formal [2 + 2] processes through enyne cycloisomerizations using platinum or gold catalysts developed by Malacria11 and Cossy.12 However, a thermally driven stepwise [2 + 2] cycloaddition in a Ficini manner using ynamides remained elusive.13 Our own efforts in trying to develop this cycloaddition reaction lasted for 13 years. We report here our first success in a Ficini [2 + 2] cycloaddition of ynamides.
Over the last 15 years, we failed numerous attempts at succeeding Ficini [2 + 2] cycloaddition of ynamides using lactam or oxazolidinone substituted ynamides under thermal and/or Lewis-acidic consitions.14 In the current pursuit of this cycloaddition, we chose to employ N-sulfonyl substituted ynamides because the nitrogen pair of the sulfonamido group is more delocalized toward the alkyne.15 Therefore, N-sulfonyl substituted ynamides possess enhanced nucleophilicity over simple amide or urethane substituted ynamides, and they are also less stable than amide or urethane substituted ynamides.
However, to our disappointment, N-sulfonyl substituted ynamides such as 7 and 10 did not undergo any desired thermal cycloaddition [Scheme 2]. Even when we used quinone and adopt the more electron-rich para-methoxy benzensulfonyl group [Mbs] as shown in ynamide 10, no appreciable amount of the desired cycloadduct 9b was observed, thereby further underscoring the superior stability of ynamides over ynamines.
Our next best option would appear to again involve Lewis acids, which had not been successful over the years when using lactam or oxazolidinone substituted ynamides.14 More specifically, our efforts were derailed when using Lewis acids because hydro-halogenations of ynamides, leading to alpha-halogenated enamides, was a serious competing pathway.14,16,17 In addition, when hydro-halogenation is not competing, possible hydrolysis under these suitable Lewis acids represents another challenge associated with ynamides. Consequently, much of ynamide chemistry7a has been limited to halo-substituted Lewis acids that do not involve metals such as Mg, Ti, Sn, Si, B, Al, or In [i.e., CuX2 or ZnX2 is feasible], or Lewis acids with OTf serving as the counter anion. As a result, we screened a small sample of Lewis acids as summarized in Table 1.
Initial failure is quite evident in entries 1-6 when using ynamide 10. However, after observing trace amount of the possible product 11 when using CuCl2 and AgSbF4 [entry 6], we speculated that 10 was polymerizing under these reactions conditions. Therefore, we turned to ynamide 12 with a Me group as the terminal-substitution. Gratifyingly, we found that cycloadduct 1318 could be attained in good yields at three different low temperatures within an hour [entries 7-9]. This result represents the first successful Ficini [2 + 2] cycloaddition using ynamides. Cycloadduct 13 was unambiguously assigned using X-ray [Figure 1]. It is noteworthy that the amido-cyclobutene motif is quite robust. The pericyclic ring-opening does not occur readily since the allowed thermal conrotatory ring-opening would lead to a trans-cycloalkenone.19
The generality of this cycloaddition could be established from examples shown in Figure 2. Several features are: (a) The N-sulfonyl group does not need to be Mbs [entries 1, 2, and 10]; (b) acyclic enones are also suitable [entries 5 and 6];20 (c) the alkyne substitutions [entries 7, 8, 14, and 15] and substitutions on the nitrogen atom [entries 11-15] can be varied, which should significantly enhance the potential applications of these cycloadducts.
Moreover, the [2 + 2] cycloadducts such as 13 could be subjected to hydrolytic conditions and further undergo retro-Claisen via the intermediacy of diketone 29 [Scheme 3], leading to keto-ester 30.21 Intriguingly, while anhydrous conditions led to 30 in 76% yield, when using MeOH-H2O as solvent, keto-imide 3122 was found in addition to 30. Ficini also observed ketoamide formation but only under neutral or basic hydrolytic conditions, and its formation likely proceeded through an aminal intermediate.1,4,23 The modest syn-selectivity was also reported in Ficini's related work,4,23 and the saponified 30-syn was used by Ficini in their synthesis of (±)-juvabione.24
Lastly, a simple and straightforward mechanistic consideration would be that this is step-wise cycloaddition with a nucleophilic 1,4-addition by the ynamide onto the enone activated via the cationic Cu(II) catalyst [see Possibility A in Figure 3]. However, there may be another possibility. That is, the cationic Cu(II) species is activating the alkyne [Possibility B], leading to an intermediate that could participate in a cuprate-like 1,4-addition. While we are not sure of the oxidation state of such copper species, this proposed possibility resonates with our earlier proposal of the intermediacy of C to explain the exclusive syn addition of “H-X” [hydro-halogenation] to ynamides that was observed when using catalysts such as MgX2,14 TiCl4,16 or Rh(I)Cl(Ph3P)3.17 We are currently exploring such a mechanistic possibility.
We have uncovered here the Ficini [2 + 2] cycloaddition using ynamides. These reactions could be catalyzed using CuCl2 and AgSbF6. Efforts are underway to develop synthetic applications of this cycloaddition reaction.
We thank NIH [GM066055] for funding. We thank Dr. Vic Young of the University of Minnesota for providing X-ray structural analysis. We also thank Professor Steve Burke of University of Wisconsin for valuable discussions.