Amino acid composition and thermal stability of protein structures: the free energy geography of the Protein Data Bank

We study the combined influence of amino acid composition and chain length on the thermal stability of protein structures. A new parameterization of the internal free energy is considered, as the sum of hydrophobic effect, hydrogen-bond and de-hydration energy terms. We divided a non-redundant selection of protein structures from the Protein Data Bank into three groups: i) rich in order-promoting residues (OPR proteins); ii) rich in disorder-promoting residues (DPR proteins); iii) belonging to a twilight zone (TZ proteins). We observe a partition of PDB in several groups with different internal free energies, amino acid compositions and protein lengths. Internal free energy of 96% of the proteins analyzed ranges from -2 to -6.5 kJ/mol/res. We found many DPR and OPR proteins with the same relative thermal stability. Only OPR proteins with internal energy between -4 and -6.5 kJ/mol/res are observed to have chains longer than 200 residues, with a high de-hydration energy compensated by the hydrophobic effect. DPR and TZ proteins are shorter than 200 residues and they have an internal energy above -4 kJ/mol/res, with a few exceptions among TZ proteins. Hydrogen-bonds play an important role in the stabilization of these DPR folds, often higher than contact energy. The new parameterization of internal free energy let emerge a geography of thermal stabilities of PDB structures. Amino acid composition per se is not sufficient to determine the stability of protein folds, since. DPR and TZ proteins generally have a relatively high internal free energy, and they are stabilized by hydrogen-bonds. Long DPR proteins are not observed in the PDB, because their low hydrophobicity cannot compensate the high de-hydration energy necessary to accommodate residues within a highly packed globular fold.