Limits...
On the effect of protein conformation diversity in discriminating among neutral and disease related single amino acid substitutions.

Juritz E, Fornasari MS, Martelli PL, Fariselli P, Casadio R, Parisi G - BMC Genomics (2012)

Bottom Line: Each protein was associated with its corresponding set of available conformers as found in the Protein Conformational Database (PCDB).At the conformer level, we also found that the different conformers correlate in a different way to the corresponding phenotype.Our results suggest that the consideration of conformational diversity can improve the discrimination of neutral and disease related protein SASs based on the evaluation of the corresponding Gibbs free energy change.

View Article: PubMed Central - HTML - PubMed

Affiliation: Departamento de Ciencia y Tecnología, Universidad Nacional de Quilmes, Buenos Aires, Argentina.

ABSTRACT

Background: Non-synonymous coding SNPs (nsSNPs) that are associated to disease can also be related with alterations in protein stability. Computational methods are available to predict the effect of single amino acid substitutions (SASs) on protein stability based on a single folded structure. However, the native state of a protein is not unique and it is better represented by the ensemble of its conformers in dynamic equilibrium. The maintenance of the ensemble is essential for protein function. In this work we investigated how protein conformational diversity can affect the discrimination of neutral and disease related SASs based on protein stability estimations. For this purpose, we used 119 proteins with 803 associated SASs, 60% of which are disease related. Each protein was associated with its corresponding set of available conformers as found in the Protein Conformational Database (PCDB). Our dataset contains proteins with different extensions of conformational diversity summing up a total number of 1023 conformers.

Results: The existence of different conformers for a given protein introduces great variability in the estimation of the protein stability (ΔΔG) after a single amino acid substitution (SAS) as computed with FoldX. Indeed, in 35% of our protein set at least one SAS can be described as stabilizing, destabilizing or neutral when a cutoff value of ±2 kcal/mol is adopted for discriminating neutral from perturbing SASs. However, when the ΔΔG variability among conformers is taken into account, the correlation among the perturbation of protein stability and the corresponding disease or neutral phenotype increases as compared with the same analysis on single protein structures. At the conformer level, we also found that the different conformers correlate in a different way to the corresponding phenotype.

Conclusions: Our results suggest that the consideration of conformational diversity can improve the discrimination of neutral and disease related protein SASs based on the evaluation of the corresponding Gibbs free energy change.

Show MeSH
Distribution of ΔASA (Å2) for substituted positions derived from the analysis of the conformational ensemble for each of the 119 proteins in the dataset.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3303731&req=5

Figure 2: Distribution of ΔASA (Å2) for substituted positions derived from the analysis of the conformational ensemble for each of the 119 proteins in the dataset.

Mentions: The 119 proteins studied in this work were linked to the PCDB database [43]. All conformer coordinates contained in PCDB for each protein were derived from the PDB database (http://www.rcsb.org). The structures were obtained under different conditions, mainly in the presence of different ligands that shift the population equilibrium of the different conformers in the ensemble [37,45]. Our dataset has an average maximum root mean squared deviation (RMSD) between conformers of 2.51 Å and an average number of conformers per protein of 8.6. The distribution of the maximum RMSD (the maximum RMSD displayed between all conformers of a given protein) is shown in Figure 1. Considering that the average RMSD for a protein crystallized under the same condition ranges from 0.1 and 0.4 Å [42] and from the distribution shown in Figure 1, we concluded that our dataset contains proteins with moderated to extreme conformational diversity (for details on conformational diversity per protein see additional file 1). We also computed the relative accessible surface area (ASA) of the positions involved in SASs (neutral and disease related) as described in Methods. We found changes in the maximum Δ(ASA) between conformers, with a maximum value of 98.6 Å2 and an average value of 12.0 Å2 (Figure 2). This distribution reflects the structural changes at the SAS positions between conformers. In fact, using Δ(ASA) values, 33% of the proteins have at least one position that, depending on the chosen conformer, can be classified as buried (ASA<20Å2) or solvent exposed (ASA>20Å2). Previous observations suggested that ΔΔG values upon residue substitution inversely correlate with the corresponding ASA values [46]. We therefore can expect large variations in the ΔΔG estimation upon changes on the different conformers considering the large ASA variation.


On the effect of protein conformation diversity in discriminating among neutral and disease related single amino acid substitutions.

Juritz E, Fornasari MS, Martelli PL, Fariselli P, Casadio R, Parisi G - BMC Genomics (2012)

Distribution of ΔASA (Å2) for substituted positions derived from the analysis of the conformational ensemble for each of the 119 proteins in the dataset.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3303731&req=5

Figure 2: Distribution of ΔASA (Å2) for substituted positions derived from the analysis of the conformational ensemble for each of the 119 proteins in the dataset.
Mentions: The 119 proteins studied in this work were linked to the PCDB database [43]. All conformer coordinates contained in PCDB for each protein were derived from the PDB database (http://www.rcsb.org). The structures were obtained under different conditions, mainly in the presence of different ligands that shift the population equilibrium of the different conformers in the ensemble [37,45]. Our dataset has an average maximum root mean squared deviation (RMSD) between conformers of 2.51 Å and an average number of conformers per protein of 8.6. The distribution of the maximum RMSD (the maximum RMSD displayed between all conformers of a given protein) is shown in Figure 1. Considering that the average RMSD for a protein crystallized under the same condition ranges from 0.1 and 0.4 Å [42] and from the distribution shown in Figure 1, we concluded that our dataset contains proteins with moderated to extreme conformational diversity (for details on conformational diversity per protein see additional file 1). We also computed the relative accessible surface area (ASA) of the positions involved in SASs (neutral and disease related) as described in Methods. We found changes in the maximum Δ(ASA) between conformers, with a maximum value of 98.6 Å2 and an average value of 12.0 Å2 (Figure 2). This distribution reflects the structural changes at the SAS positions between conformers. In fact, using Δ(ASA) values, 33% of the proteins have at least one position that, depending on the chosen conformer, can be classified as buried (ASA<20Å2) or solvent exposed (ASA>20Å2). Previous observations suggested that ΔΔG values upon residue substitution inversely correlate with the corresponding ASA values [46]. We therefore can expect large variations in the ΔΔG estimation upon changes on the different conformers considering the large ASA variation.

Bottom Line: Each protein was associated with its corresponding set of available conformers as found in the Protein Conformational Database (PCDB).At the conformer level, we also found that the different conformers correlate in a different way to the corresponding phenotype.Our results suggest that the consideration of conformational diversity can improve the discrimination of neutral and disease related protein SASs based on the evaluation of the corresponding Gibbs free energy change.

View Article: PubMed Central - HTML - PubMed

Affiliation: Departamento de Ciencia y Tecnología, Universidad Nacional de Quilmes, Buenos Aires, Argentina.

ABSTRACT

Background: Non-synonymous coding SNPs (nsSNPs) that are associated to disease can also be related with alterations in protein stability. Computational methods are available to predict the effect of single amino acid substitutions (SASs) on protein stability based on a single folded structure. However, the native state of a protein is not unique and it is better represented by the ensemble of its conformers in dynamic equilibrium. The maintenance of the ensemble is essential for protein function. In this work we investigated how protein conformational diversity can affect the discrimination of neutral and disease related SASs based on protein stability estimations. For this purpose, we used 119 proteins with 803 associated SASs, 60% of which are disease related. Each protein was associated with its corresponding set of available conformers as found in the Protein Conformational Database (PCDB). Our dataset contains proteins with different extensions of conformational diversity summing up a total number of 1023 conformers.

Results: The existence of different conformers for a given protein introduces great variability in the estimation of the protein stability (ΔΔG) after a single amino acid substitution (SAS) as computed with FoldX. Indeed, in 35% of our protein set at least one SAS can be described as stabilizing, destabilizing or neutral when a cutoff value of ±2 kcal/mol is adopted for discriminating neutral from perturbing SASs. However, when the ΔΔG variability among conformers is taken into account, the correlation among the perturbation of protein stability and the corresponding disease or neutral phenotype increases as compared with the same analysis on single protein structures. At the conformer level, we also found that the different conformers correlate in a different way to the corresponding phenotype.

Conclusions: Our results suggest that the consideration of conformational diversity can improve the discrimination of neutral and disease related protein SASs based on the evaluation of the corresponding Gibbs free energy change.

Show MeSH