Physicochemical basis of pairwise interactions between amino acid residues in Brookhaven data bank

We have analyzed the contacts between nonhomologous multisubunit proteins and complexes from Brookhaven data bank. We have analyzed distribution of amino acids in the contact area for 122 protein sequences and calculated frequencies of amino acid residues contacts in ternary structure. The pair of residues is considered as a contact if the minimum distance between their atoms is less than 2.8 A. We have compared the calculated frequencies of external contacts in protein complexes with contacts inside the tertiary structure (internal contacts) extracted from Protein contacts atlas available via Internet. The contacts CC, VV, WW, II, LL, FF occur more frequently for both samples than expected on the random basis, but their frequency in tertiary structure is higher. The pairs RN, RE, AV, CC, EK, IL, IF, IV, LL, LF, LV are statistically more frequent comparing to the average values in both samples.

The most/less frequent contacts are the same in tertiary and ternary structure for 16 cases from 32 possible. There are no cases where statistically frequent contact for one sample is rare for another. The most preferable contacts in both samples are contacts between hydrophobic amino acids or between polar amino acids. Pairings between individual amino acid residues are also very frequent. The size of amino acid residue plays an important role in forming the sample. The role of large aromatic residues in protein contacts is discussed.