|Home | About | Journals | Submit | Contact Us | Français|
A new torquoselective ring-closure of chiral amide-substituted 1,3,5-hexatrienes and its application in tandem with [4 + 2] cycloaddition are described. The trienes were derived via either a 1,3-H or 1,3-H–1,7-H shift of α-substituted allenamides, and the entire sequence through the [4 + 2] cycloaddition could be in tandem from allenamides.
We recently reported isomerizations of α-substituted allenamides 1 to give amido-dienes 2 via stereoselective 1,3-H shift [Scheme 1].1 In addition, with R = vinyl, the resulting 1,3,5-hexatrienes 2′ were found to be well suited for a 6π-electron electrocyclic ring-closure that could be in tandem with the 1,3-H shift, leading to novel chiral cyclic amido-dienes 32,3 directly from allenamides.4,5 The rapid access of 1,3,5-hexatrienes via a simple isomerization of allenes6–8 allowed us to envision a new torquoselective ring-closure9–13 involving chiral amido-trienes 4. This asymmetric transformation could potentially lead to a remote 1, 6-stereochemical induction while affording cyclic amido-dienes 5, which should be useful for cycloadditions. We communicate here this torquoselective process and its application in tandem with [4 + 2] cycloadditions.
Our intention was quickly met with two unexpected findings. Initially, when heating allenamide 614,15 in ClCH2CH2Cl at 135 °C, instead of isolating the desired amido-diene 8 from ring-closure of the triene 7, we found 915 in almost quantitative yield, thereby implying a 1,5-H shift had taken place [Scheme 2]. Similar results were attained when using triene 7 generated from 6 via an acid-promoted 1,3-H shift using 10 mol % p-TsOH.
We quickly made a minor substrate adjustment to prevent such 1,5-H shift, but that led to the second unexpected finding as shown in Scheme 3. Heating α-prenylated allenamide 10a led to a mixture of two ring-closure products: The desired 2-amido diene 13a and the unexpected 1-amdio-diene 14a in 1:4.5 ratio with 14a being a 10:1 diastereomeric mixture. The latter implied the presence of amido-triene 12a, which could be rationalized through an antarafacial 1,7-H shift9,17 from the initial amido-triene 11a, via the methyl group [in red] syn to the terminal olefin [in blue]. More importantly, stereochemistry for the major isomer of 14a could be assigned using its single crystal X-ray structure [Figure 1]. This unambiguous assignment suggests that a favored disrotatory course could proceed through the transition state as shown for amido-triene 12a.
The tandem 1,3-H–1,7-H shift appears to be general whether commencing from amido-trienes 11b–e [entries 1–4 in Table 1] or directly from α-prenylated allenamides 10b–e [entries 5–8]. It is noteworthy that in the case of allenamide 10b and 10c, cyclic amido-dienes 14b and 14c were the only products resulting from a tandem 1,3-H–1,7-H shift–pericyclic ring-closure. In addition, there was no equilibration between 13 and 14 under the reaction conditions.18
Probing further mechanistically, we found two phenomena associated with this 1,7-H shift process. Firstly, either heating triene 16, derived from γ-substituted allenamide 15, or 15 led to no observable 1,7-H shift with cyclic diene 19 as the only identifiable product. Isolation of 19 implies the ring-closure of triene 16 had taken place to give 18 followed by 1,5-H shift [Scheme 4]. This experiment suggests that there exists a directional preference for the 1,7-H shift in these amido-trienes, and that it is not sufficient simply having a methyl group [in red] at one terminus of the triene syn to the other terminus [in blue].
Secondly, we found a distinct dependence of the 1,7-H shift on the olefinic geometry. As shown in Scheme 5, reactions of allenamides 20-Z and 20-E proceeded through distinctly different tandem pathways. While 1,3-H–1,7-H shift occurred with 20-Z en route to cyclic amido-diene 22 via pericyclic ring-closure of triene 21, the reaction of allenamide 20-E gave 24 with no observable 1,7-H shift. Relative stereochemistry in 22 was assigned based on a disrotatory ring-closure, while the absolute stereochemistry was assessed based on 14a. Cyclic amido-diene 24 was found as a 3:1 isomeric mixture with respect to C5, thereby implying that albeit modest and unassigned at this time,19 a rather impressive 1,6-asymmetric induction took place during the ring-closure.
To both accentuate this dichotomy and render these stereoselective ring-closures synthetically useful, we embarked on tandem processes that would include [4 + 2] cycloadditions. As shown in Scheme 6, reactions of allenamides 25-Z and 26-Z led to tricycles 28a and 28b as a single isomer through a highly stereoselective [4 + 2] cycloaddition of cyclic amido-dienes 27a and 27b, respectively, thereby constituting a quadruple tandem process of 1,3-H-1,7-H shift–6π-electron pericyclic ring-closure–[4 + 2]-cycloaddition. Stereochemical outcome of the cycloaddition was controlled through the C5-stereocenter, which was installed during the torquos elective ring closure.
In contrast, reactions of allenamides 25-E [for a direct comparison with 25-Z] and 29-E led the respective tricycles 31a and 31b [assigned via X-ray: See Figure 2] in excellent yields and high diastereoselectivity proceeding from amido-trienes 30a and 30b or directly from the allenamides in a triple tandem process. In either case, the [4 + 2] cycloaddition presumably went through the diastereomeric pairs 32a/a′ [X = O] and 32b/b′ [X = NTs] given the modest 1,6-induction found for 20-E. It is noteworthy that despite not having assigned these diastereomers with ratios undermined,20 both pairs would actually converge to give 31a and 31b, respectively. These tandem processes provide a rapid assembly of complex tricycles from very simple allenamides, thereby manifesting their tremendous power and synthetic potential.
We have described here a new torquoselective ring-closure of chiral amide-substituted 1,3,5-hexatrienes and its application in tandem with [4 + 2] cycloaddition. The 1,3,5-hexatrienes were derived via either a 1,3-H or 1,3-H–1,7-H shift of α-substituted allenamides, and the entire sequence through the Diels-Alder could be in tandem from allenamides. Applications of these new tandem processes as well as mechanistic understanding and improvement of the observed 1,6-asymmetric induction are underway.
Authors thank NIH [GM066055] for support and Dr. Victor Young [University of Minnesota] for X-ray structural analysis.
Supporting Information Available: Experimental procedures as well as NMR spectra, characterizations, and X-ray structural files are available for all new compounds and free of charge via Internet http://pubs.acs.org.