A COMPARATIVE SEMIEMPIRICAL AND ABINITIO STUDY OF TAUTOMERIZATION ENERGIES - 7-METHYL-2,3,4,7-TETRAHYDROISOTHIAZOLO [5,4B]PYRIDINE-3,4-DIONE AND OTHER MODELS OF ANTIBACTERIAL QUINOLONE ANALOGS

The tautomerization energies of 2,3-dihydroisothiazole-3-one (1), 7-methyl-2,3.4,7-tetrahydroisothiazole[5,4b]pyridine-3,4-dione (2), 7-methyl-2,3,4,7-tetrahydroisoxazolopyridine-3,4-dione (3) and 2,3-dihydrobenzoisothiazole-3-one (4) were computed using the MNDO method with the original and PM3 parameterization, and in the ab initio scheme with the MINI-1, STO-3G, 3-21G*, 4-31G* and 6-31G** basis sets. Second-order Moller-Plesset (MP2) and zero-point, energy (ZPE) corrections were applied to the smallest tautomeric pair. The effect of the surrounding medium was mimicked by using the self-consistent solvent cavity method of Tomasi and coworkers. The calculated lactim-lactam tautomerization energies are sensitive to the basis set used. For the isolated molecules the MINI-1, STO-3G and 6-31G** basis sets correctly predict the lactim form to be the more stable one; 4 is exceptional in that the lactam form is the more stable one if computed at the Hartree-Fock (HF) level but the lactim form is more stable at the MP2/6-31G** level. MNDO results are in good qualitative agreement with large-basis-set calculations, whereas the PM3 parameterization was found to be less reliable. In a medium having a high dielectric constant the lactam form is strongly favoured.