I Bahar1, R L Jernigan. 1. Molecular Structure Section, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892-5677, USA.
Abstract
BACKGROUND: Two opposite views have been advanced for the packing of sidechains in globular proteins. The first is the jigsaw puzzle model, in which the complementarity of size and shape is essential. The second, the nuts-and-bolts model, suggests that constraints induced by steric complementarity or pairwise specificity have little influence. Here, the angular distributions of sidechains around amino acids of different types are analyzed, in order to capture the preferred (if any) coordination loci in the neighborhood of a given type of amino acid. RESULTS: Some residue pairs select specific coordination states with probabilities about ten times higher than expected for random distributions. This selectivity becomes more pronounced at closer separations leading to an effective free energy of stabilization as large as -2 RT for some sidechain pairs. A list of the most probable coordination sites around each residue type is presented, along with their statistical weights. CONCLUSIONS: These data provide guidance as to how to pack selectively the nonbonded sidechains in the neighborhood of a central residue for computer generation of unknown protein structures.
BACKGROUND: Two opposite views have been advanced for the packing of sidechains in globular proteins. The first is the jigsaw puzzle model, in which the complementarity of size and shape is essential. The second, the nuts-and-bolts model, suggests that constraints induced by steric complementarity or pairwise specificity have little influence. Here, the angular distributions of sidechains around amino acids of different types are analyzed, in order to capture the preferred (if any) coordination loci in the neighborhood of a given type of amino acid. RESULTS: Some residue pairs select specific coordination states with probabilities about ten times higher than expected for random distributions. This selectivity becomes more pronounced at closer separations leading to an effective free energy of stabilization as large as -2 RT for some sidechain pairs. A list of the most probable coordination sites around each residue type is presented, along with their statistical weights. CONCLUSIONS: These data provide guidance as to how to pack selectively the nonbonded sidechains in the neighborhood of a central residue for computer generation of unknown protein structures.