3.5. Acid-Base Complexes
For organic acid-base pairs, theoretical studies at satisfactorily high level convincingly prove that a neutral O–H…N bond rather than the ionic O−…+HN bridge is formed generally for complexes in the gas phase. Such calculations have been performed with acid components such as formic and acetic acids with various bases [115,116,240]. It has been found at the MP2/aug-cc-pvdz level, however, that the ionic acetic acid…methylguanidine complex is also stable on the gas-phase potential energy surface [115]. Investigations for further acid-base complexes is necessary whether a geometry optimization starting from an ion-pair structure could find a local energy minimum for this complex or the proton on the base moves onto the –COO− group.
For an aqueous solution the question is, whether the intramolecular H-bond (from the perspective of the complex) saves its neutral-form character as in the gas phase or the ion-pair tautomer is stabilized. The answer could depend on the theoretical approach that has been applied for estimating the solvent effects. Liljefors and Orrby [240] concluded that consideration of an explicit water molecule would stabilize the ionic complex in continuum solvents with ε = 4–6. Without the explicitly considered water molecule, the ion-pair form is stabilized in continuum solvents with ε > 9.
Accordingly, Nagy and Erhardt [115] studied the possible proton jump within the complex using the IEF-PCM approach up to the CCSD(T)/CBS level and using MP2/aug-cc-pvdz and MP2/aug-cc-pvdz optimized geometries in solutions characterized by ε = 5 and 15. It was found that the acid …methylamine complexes are much more stable in the neutral than the ion-pair form. The complex is almost exclusively ionic, however, when the base is methylguanidine. This study aimed to explore the neutral or ionic character of the so-called salt-bridge formation when an Asp/Glu side chain can be close to the side chain of a Lys/Arg residue within a protein. The prediction was that the H-bond with lysine would be neutral in low-polarity environment, whereas the bridge is ionic with arginine in any environment. In partial disagreement with the above results, Silva et al., [241] advocated the ionic character of any salt-bridge if an Asp side chain, as the acid partner is involved. These authors emphasized the need for considering more than three explicit water molecules and accounting for vibrational contributions to the total relative free energy even when a continuum solvent model is being applied. Preferably, consideration of large solvation shells was advised.
Nagy and Erhardt [116] studied the acetic acid interaction with methyl, dimethyl and trimethyl amines. The relative internal free energies were estimated using the IEF-PCM method at the ab initio CCSD(T)/CBS//MP2/aug-cc-pvdz and the DFT/B97D/aug-cc-pvtz levels. Relative solvation free energies were calculated through MC/FEP processes. The predicted total relative free energies varied considerably depending on the simulation conditions. When a solution of finite, about 0.1 molar concentration was modeled in the presence of a dissolved Na+Cl− ion pair, the CH3COO−…+HNHx(CH3)3−x (x = 0, 1, 2) ionic structure was found to predominate in aqueous solution. In contrast, when the low-dielectric-constant region of a protein was simulated by a solution model with chloroform solvent, the neutral form was found to be much more stable than the ion-pair. This is a remarkable conclusion regarding the type of the intermolecular H-bond for the complex, because the two modeling scenarios were supposed to account for the conditions of the drug-receptor interaction in an aqueous phase (e.g., on the surface of an enzyme) in comparison to a tightly binding environment within a cavity deeply buried in a protein.