Local polarity analysis: a sensitive method that discriminates between native proteins and incorrectly folded models

Abstract
The evaluation of calculated protein structures is an important step in the protein design cycle. Known criteria for this assessment of proteins are the polar and apolar, accessible and buried surface area, electrostatic interactions and other interactions between the protein atoms (e.g. H..O, S..S), atomic packing, analysis of amino acid environment and surface charge distribution. We show that a powerful test of accuracy of protein structure can be derived by analysing the water contact of atoms and additionally taking into account their polarity. On the basis of estimated reference values of the polar fraction of typical globular proteins with known structure (mean, SD and distribution), the evaluation of misfolded structures can be improved significantly. The reference values are derived by moving windows of different length (3-99 amino acid residues) over the amino acid sequence. Model proteins, which are included in the Brookhaven protein structure databank, deliberately misfolded proteins, hypothetical proteins and predicted protein structures are diagnosed as at least partially incorrectly folded. The local fault, mostly observed, is that polar groups are buried too frequently in the interior of the protein. The database-derived quantities are useful in screening the designed proteins prior to experimentation and may also be useful in the assessment of errors in the experimentally determined protein structures.