The Bergman cyclization of the natural enediyne dynemicin A in its triggered form (2) bound to the minor groove of DNA is compared with the corresponding reaction of its open isomer (4) utilizing QM/MM methodology. The two isomers are typical representatives of 10-membered cyclic (2) and acyclic (4) enediynes, which possess significantly different barriers for the Bergman reaction in the gas phase (2, 20.4 kcal/mol; 4, 31.3 kcal/mol). In the case of the cyclic enediyne (2) the explicit consideration of environmental factors such as the receptor DNA, the solvent water, and charge neutralization by counterions has only minor effects on the energy profile of the cyclization reaction and the corresponding optimized structures when compared with the gas phase. The energetics of the reaction is predominantly determined by QM (electronic) effects. This makes it possible to replace the explicit description of the environment by an implicit one, thus avoiding costly QM/MM calculations and using instead a decoupled QM+MM approach. A conformationally driven hinge mechanism is identified for 2 that makes it possible for the ligand to adjust to the dimensions of the minor groove without significant energy loss. In the case of me acyclic enediyne 4 a QM/MM treatment is necessary to describe the Bergman cyclization in the minor groove. QM/MM corrects the cyclization barrier from 31.3 to 23.7 kcal/mol, which makes the reaction feasible under physiological conditions. The reduction of the barrier is a result of transition-state stabilization, which is caused by an increased dipole moment and hence stronger electrostatic interactions with the environment. In both cases the anionic charge of dynemicin A is largely shielded by water solvation and ion pair formation so that it does not significantly affect the energetics of the Bergman cyclization.
- Bergman cyclization
- decoupled QM+MM approach