Solvent effect on the syn/anti conformational stability: A comparison between conformational bias Monte Carlo and molecular dynamics methods

The solvent effect on the syn/anti population ratio of the mesityl oxide (MOx) was investigated using a new implementation of conformational bias Monte Carlo (CBMC) and molecular dynamics (MD) methods. It was observed by a previous theoretical work (Theor. Chem. Acc. (2012) 131:1214) that in gas-phase the MOx exists dominantly in syn-form and in aqueous solution in anti-form. The syn/anti free energy difference in the gas phase was used in the intramolecular parametrization and a rotational barrier of approximately 10 kcal mol(-1) was found. Molecular systems with barriers of this order of magnitude have been studied by experimental techniques. However, they have not been discussed yet comparing CBMC and MD simulations. In this work, we show that the intramolecular geometrical information such as bond lengths, angles and torsional angles sampled with CBMC and MD methods are equivalent. Nonetheless, only the CBMC simulations sample appropriately the syn/anti population ratio. With the CBMC configurations in gas phase, it was obtained 95% in syn-form and 5% in anti-form regardless the initial conformation. An inversion of the population was found in water, 25% in syn-form and 75% in anti-form. Comparing the gas phase and in-water CBMC sampling, it was observed that the MOx spends typically approximately 110 successive MC cycles in anti-form and approximately 2300 in syn-form in gas phase. While it was much larger with explicit water, approximately 400 times more for anti-form and approximately 6 times more for syn-form. We argue that this strong stabilization of the anti-form in aqueous solution, does not come from the MOx-water hydrogen bonds interactions, because they are the same for both conformations. Instead, the stabilization comes from the dipole-dipole interaction caused by a larger dipole moment of the MOx in the anti-form, 7.2 D, than in the syn-form, 5.2 D. With the MD sampled configurations in both conditions, we observe that the syn/anti conformational change is a very rare event due to the rotational barrier, which is approximately 17 times larger than the thermal energy. Therefore, the MD sampling of the MOx is not appropriated because it is strongly dependent on the initial conformation even for large simulations with 150 ns up to 400 ns for the isolated solute and for solute-solvent systems.