Limits...
Globular and disordered-the non-identical twins in protein-protein interactions.

Teilum K, Olsen JG, Kragelund BB - Front Mol Biosci (2015)

Bottom Line: The interactions between intrinsically disordered proteins (IDPs) and other proteins rely on changes in flexibility and this is seen as a strong determinant for their function.This has fostered the notion that IDP's bind with low affinity but high specificity.We find that ordered proteins and the disordered ones act as non-identical twins operating by similar principles but where the disordered proteins complexes are on average less stable by 2.5 kcal mol(-1).

View Article: PubMed Central - PubMed

Affiliation: Structural Biology and NMR Laboratory, Department of Biology, University of Copenhagen Copenhagen, Denmark.

ABSTRACT
In biology proteins from different structural classes interact across and within classes in ways that are optimized to achieve balanced functional outputs. The interactions between intrinsically disordered proteins (IDPs) and other proteins rely on changes in flexibility and this is seen as a strong determinant for their function. This has fostered the notion that IDP's bind with low affinity but high specificity. Here we have analyzed available detailed thermodynamic data for protein-protein interactions to put to the test if the thermodynamic profiles of IDP interactions differ from those of other protein-protein interactions. We find that ordered proteins and the disordered ones act as non-identical twins operating by similar principles but where the disordered proteins complexes are on average less stable by 2.5 kcal mol(-1).

No MeSH data available.


Amino acid composition of protein-protein interfaces extracted from 87 high-resolution structures of protein-protein complexes. (A) Fractional overrepresentation of each amino acid residue type and of the four amino acid residue classes (FWY, CILMV, AGPST, and DEHKNQR) in ORD-IDP complexes relative to ORD-ORD complexes. log2 of the ratios are plotted with positive values indicating overrepresentation in ORD-IDP complexes. (B) Correlation plots of the fractions of the four amino acid residue classes (FWY, CILMV, AGPST, and DEHKNQR) in protein-protein interfaces. Each point represents a protein-protein complex and is colored either red (ORD-ORD) or blue (ORD-IDP).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4496568&req=5

Figure 1: Amino acid composition of protein-protein interfaces extracted from 87 high-resolution structures of protein-protein complexes. (A) Fractional overrepresentation of each amino acid residue type and of the four amino acid residue classes (FWY, CILMV, AGPST, and DEHKNQR) in ORD-IDP complexes relative to ORD-ORD complexes. log2 of the ratios are plotted with positive values indicating overrepresentation in ORD-IDP complexes. (B) Correlation plots of the fractions of the four amino acid residue classes (FWY, CILMV, AGPST, and DEHKNQR) in protein-protein interfaces. Each point represents a protein-protein complex and is colored either red (ORD-ORD) or blue (ORD-IDP).

Mentions: Based on previous collections of data (Stites, 1997; Huang and Liu, 2013) and including several additional data from the literature found by searching PubMed for “ITC protein-protein interactions,” “ITC intrinsically disordered protein,” “thermodynamics protein-protein interactions,” and “thermodynamics intrinsically disordered protein,” we have compiled thermodynamic parameters from close to 200 different protein-protein interaction studies (Supplementary Table 1). The data were standardized to 298 K assuming that ΔCp = 0, as ΔCp has only been estimated for very few of the complexes. We have estimated that the error introduced in ΔG0 is less than 0.2 kcal mol−1 in the most extreme cases where the data were measured at 281 K. For most cases where there is less than 5 K difference the error is less than 0.05 kcal mol−1. We subsequently compared and correlated the parameters for interactions that involve only globular proteins (91 complexes), to the parameters for interactions, where one partner is an IDP (106 complexes). To avoid over-representing a single protein-protein complex we exclusively compared wild-type proteins so that protein specific irregularities will be averaged out. In the cases where a structure of the complex has been determined, we have calculated the interaction surface area using PISA (Krissinel and Henrick, 2007) (Supplementary Table 1), and determined the amino acid composition of the interface using NCONT from the CCP4i suite (Winn et al., 2011). The amino acids were divided into four classes for analysis (FWY, CILMV, AGPST, and DEHKNQR) based on the BLOSSUM50 substitution matrix as defined by Weathers et al. (2004). The interfaces of the all ordered (ORD-ORD) complexes and the ordered-IDP (ORD-IDP) complexes were then compared in this context (Figure 1).


Globular and disordered-the non-identical twins in protein-protein interactions.

Teilum K, Olsen JG, Kragelund BB - Front Mol Biosci (2015)

Amino acid composition of protein-protein interfaces extracted from 87 high-resolution structures of protein-protein complexes. (A) Fractional overrepresentation of each amino acid residue type and of the four amino acid residue classes (FWY, CILMV, AGPST, and DEHKNQR) in ORD-IDP complexes relative to ORD-ORD complexes. log2 of the ratios are plotted with positive values indicating overrepresentation in ORD-IDP complexes. (B) Correlation plots of the fractions of the four amino acid residue classes (FWY, CILMV, AGPST, and DEHKNQR) in protein-protein interfaces. Each point represents a protein-protein complex and is colored either red (ORD-ORD) or blue (ORD-IDP).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4496568&req=5

Figure 1: Amino acid composition of protein-protein interfaces extracted from 87 high-resolution structures of protein-protein complexes. (A) Fractional overrepresentation of each amino acid residue type and of the four amino acid residue classes (FWY, CILMV, AGPST, and DEHKNQR) in ORD-IDP complexes relative to ORD-ORD complexes. log2 of the ratios are plotted with positive values indicating overrepresentation in ORD-IDP complexes. (B) Correlation plots of the fractions of the four amino acid residue classes (FWY, CILMV, AGPST, and DEHKNQR) in protein-protein interfaces. Each point represents a protein-protein complex and is colored either red (ORD-ORD) or blue (ORD-IDP).
Mentions: Based on previous collections of data (Stites, 1997; Huang and Liu, 2013) and including several additional data from the literature found by searching PubMed for “ITC protein-protein interactions,” “ITC intrinsically disordered protein,” “thermodynamics protein-protein interactions,” and “thermodynamics intrinsically disordered protein,” we have compiled thermodynamic parameters from close to 200 different protein-protein interaction studies (Supplementary Table 1). The data were standardized to 298 K assuming that ΔCp = 0, as ΔCp has only been estimated for very few of the complexes. We have estimated that the error introduced in ΔG0 is less than 0.2 kcal mol−1 in the most extreme cases where the data were measured at 281 K. For most cases where there is less than 5 K difference the error is less than 0.05 kcal mol−1. We subsequently compared and correlated the parameters for interactions that involve only globular proteins (91 complexes), to the parameters for interactions, where one partner is an IDP (106 complexes). To avoid over-representing a single protein-protein complex we exclusively compared wild-type proteins so that protein specific irregularities will be averaged out. In the cases where a structure of the complex has been determined, we have calculated the interaction surface area using PISA (Krissinel and Henrick, 2007) (Supplementary Table 1), and determined the amino acid composition of the interface using NCONT from the CCP4i suite (Winn et al., 2011). The amino acids were divided into four classes for analysis (FWY, CILMV, AGPST, and DEHKNQR) based on the BLOSSUM50 substitution matrix as defined by Weathers et al. (2004). The interfaces of the all ordered (ORD-ORD) complexes and the ordered-IDP (ORD-IDP) complexes were then compared in this context (Figure 1).

Bottom Line: The interactions between intrinsically disordered proteins (IDPs) and other proteins rely on changes in flexibility and this is seen as a strong determinant for their function.This has fostered the notion that IDP's bind with low affinity but high specificity.We find that ordered proteins and the disordered ones act as non-identical twins operating by similar principles but where the disordered proteins complexes are on average less stable by 2.5 kcal mol(-1).

View Article: PubMed Central - PubMed

Affiliation: Structural Biology and NMR Laboratory, Department of Biology, University of Copenhagen Copenhagen, Denmark.

ABSTRACT
In biology proteins from different structural classes interact across and within classes in ways that are optimized to achieve balanced functional outputs. The interactions between intrinsically disordered proteins (IDPs) and other proteins rely on changes in flexibility and this is seen as a strong determinant for their function. This has fostered the notion that IDP's bind with low affinity but high specificity. Here we have analyzed available detailed thermodynamic data for protein-protein interactions to put to the test if the thermodynamic profiles of IDP interactions differ from those of other protein-protein interactions. We find that ordered proteins and the disordered ones act as non-identical twins operating by similar principles but where the disordered proteins complexes are on average less stable by 2.5 kcal mol(-1).

No MeSH data available.