Mutational Effects on Protein Structures
Knowledge Gained From Databases, Predictions & Protein Models
Acta Universitatis Tamperensis No. 1498


By Sofia Kahn
June 2010
Tampere University Press
Distributed By Coronet Books
ISBN: 9789514479908
151 pages, Illustrated
$77.50 Paper original

Proteins are the machinery of life. They take care of the functionality of living cells and interactions between cells. Proteins are formed by amino acids whose properties and interactions take part in protein folding and maintain protein stability during the function of the protein. Missense mutations that are one amino acid substitution can have an effect on energies stabilizing the protein and thus have an effect of a different degree on protein structure and/or function. In extreme cases one amino acid change can lead to a severe disease phenotype. In order to investigate structural effects due to mutations structural information is needed.

One aim of the study was to investigate disease-causing mutations in protein secondary structures. For this purpose mutation data from various mutation databases was combined with experimental structural data. The dataset contained 2413 disease-causing missense mutations in 44 proteins. 80% of the mutations appeared in secondary structures. Amino acids are known to have preferences when it comes to different secondary structure types (α-helix, β-sheet, turn, and bend). The study indicated that disease-causing missense mutations do not follow this distribution but that certain amino acids mutate more often depending on the secondary structure type, and that they are often substituted with amino acids with very different properties. In over half of the mutation cases a direct connection to protein stability could be observed. The effects of amino acid substitutions on protein stability can be predicted computationally with stability prediction programs. The performance and accuracy of nine of these programs was assessed using experimental data for 1784 missense mutations as a test set. All the tested stability predictors showed a coherent trend in their predictions which was in line with experimental data. However, the programs tend to fail in predicting neutral mutations. Programs succeed better when considering their ability to predict stability increasing or decreasing mutations separately. As a result programs could be ranked albeit different programs succeeded better than others when considering different parameters.

In the absence of experimental 3D data the alterations in stabilizing interactions can be studied using protein models. Protein models obtained with homology modeling have been widely used to interpret biological phenomena such as alterations in intramolecular energies due to mutations. The aim of the study was to evaluate if models are accurate enough to be used to predict biological phenomena reliably. The 14 homology models used in this study were modeled during the last two decades in-house and each of them had been published along with predictions based on them. The study showed that the biological explanations made based on models with even very low sequence identity (less than 30%) with the template, were in concordance with the interpretations made based on experimental structures and other experimental studies. This analysis indicates that models, when made carefully, can be used to make biological predictions of mutation consequences.

Return to Coronet Books main page