3-Oxy-l-azabicyclo(2.2.0]hexa-2,5-dienes ("3-oxy-1-Dewar-pyridines") 6a-d, on cleavage of the enol ether or enol ester function, yield either l-azabicyclo[2.2.0]hex-2-en-5-ones ("1-Dewar-pyridin-3-ones") 9 or 2-azabicyclo[3.1.0] hex-2-en-4-ones 11, depending on the substituents and reaction conditions. Ester cleavage of 2-(benzoyloxy)-l-Dewar-pyridine ?a with methyllithium does not furnish a l-Dewar-2-pyridinone, but a dimer thereof, namely the tricyclic compound 14.Out of the three possible Dewar-pyridinones 1-3, only 1 and derivatives thereof have been synthesized by several routes'-3). The most direct route to the molecular framework of 1, photochemical isomerization of a 2-pyridinone, does not furnish a 1 -Dewar-pyridin-2-one 2 similarly, isomerization of 3-hydroxypyridines to l-Dewar-pyridin-3-ones 3 has not been reported. In a recent communication4), we have reported that cycloaddition of alkynyl esters and ethers 5 to 2,3,4-tri-tertbutylazete (4) yields 3-oxy-1-Dewar-pyridines 6 exclusively when disubstituted alkynes are used, whereas terminal alkynes (5, R = H) lead to a mixture of 6 and the regioisomeric 2-oxy-I -Dewar-pyridine 7.The enol derivatives 6 and 7 appear to be ideal precursors of Dewar-pyridinones of type 2 and 3, resp. In the following, we show that this strategy indeed leads to the molecular framework of 3, but not of 2. Furthermore, we report that under certain conditions intermediarily formed I-Dewarpyridin-3-olates undergo an unexpected and unprecedented rearrangement which ultimately yields 2-azabicyclo[3.1 .O]-hex-2-en-Cones.
ResultsThe results obtained with 6a-d are summarized in Scheme 1 and Table 1. Ester cleavage of alkenyl benzoate 6a either by hydroxide or by methyllithium affords only the 1-Dewar-3-pyridinone derivative 9a. The identity of this novel heterocycle is established by the following spectroscopic data: The P(C = 0) vibration at 1775 cm-I indicates a cyclobutanone substructure. Although higher wave numbers of carbonyl stretching vibrations (1805-1820 cm-') have been reported for monocyclic 3-a~etidinones~,~), the value for 9a agrees quite well with that of the alicyclic analogue (9a, C-CO,tBu instead of N: 1765 cm-' 'I). In the I3C-NMR spectrum, chemical shifts of C-2 and C-3 are nearly the same as those of C-6 and C-5 in the precursor 6a, thus indicating that the azetine moiety has remained intact. The adjacent carbonyl group causes a low-field shift of 18.0 ppm for (2-4 with respect to 6a. The C-6 methylene group, deshielded by two electronegative neighbors, appears in the expected region both in the "C-NMR (6 = 67.7) and in the 'H-NMR spectrum (6, = 3.75, 6B = 3.91, I 'J I =
Hz).Acidic hydrolysis of the (tert-butyldimethylsilyl) enol ether 6 b provides the substituted 1-Dewar-3-pyridinone 9b. Its NMR and IR spectra closely resemble those of 9a, taking into account the influence of the additional 6-phenyl substituent. Most importantly, 6(C-6) is now found at 80.5 and 6(6-H) at 5.18. Furthermore, the 'H-NMR signals of the tBu groups appear at 6 = 0.70, 1.22, 1.26. The ...