Reuse of structural domain-domain interactions in protein networks

Abstract

Background: Protein interactions are thought to be largely mediated by interactions between structural domains. Databases such as iPfam relate interactions in protein structures to known domain families. Here, we investigate how the domain interactions from the iPfam database are distributed in protein interactions taken from the HPRD, MPact, BioGRID, DIP and IntAct databases.

Results: We find that known structural domain interactions can only explain a subset of 4-19% of the available protein interactions, nevertheless this fraction is still significantly bigger than expected by chance. There is a correlation between the frequency of a domain interaction and the connectivity of the proteins it occurs in. Furthermore, a large proportion of protein interactions can be attributed to a small number of domain interactions. We conclude that many, but not all, domain interactions constitute reusable modules of molecular recognition. A substantial proportion of domain interactions are conserved between E. coli, S. cerevisiae and H. sapiens. These domains are related to essential cellular functions, suggesting that many domain interactions were already present in the last universal common ancestor.

Conclusion: Our results support the concept of domain interactions as reusable, conserved building blocks of protein interactions, but also highlight the limitations currently imposed by the small number of available protein structures.

Figure 1

Comparison of coverage of iPfam domain pairs on protein interactions. For each species, the height of the column reflects the number of known protein–protein interactions in the data set. The columns are split according to the proportion of interactions that contain an iPfam domain pair (blue), that contain any other Pfam domains on both proteins (red), and those that contain no Pfam domain pair (yellow).

Figure 2

Frequencies of iPfam domain pairs in E. coli, S. cerevisiae and H. sapiens protein interactions. Each point in this graph represents a set of protein interactions. The abscissa reflects the number of interactions in each set that contain the same iPfam domain pair. The ordinate shows the number of distinct such sets, each defined by a different iPfam domain pair. In both H. sapiens (blue) and S. cerevisiae (green) a small number of iPfam domain pairs covers a large fraction of the interactome, whereas in E. coli, no iPfam domain occurs in more than 4 experimental interactions at a time. Dotted lines denote fitted monomial functions, showing that the distributions follow a power law.

Figure 3

Average frequency of iPfam domain pairs relative to degree of node. Each point represents a protein in the interaction networks of H. sapiens (blue) and S. cerevisiae (green). For each protein, we calculate the degree, defined as the number of interactions the protein is involved in. On the y-axis, we show the average number of iPfam domain pairs in edges adjacent to proteins of degree x. We calculated a Spearman correlation of 0.68 and 0.71, for H. sapiens and S. cerevisiae. The correlation is outlined by dotted lines.

Figure 4

Venn diagramm showing the fractions of iPfam domain pairs found in the E. coli, S. cerevisiae and H. sapiens binary protein interaction sets. The three circles represent the iPfam domain pairs observed in the respective species. The overlaps denote co-observed iPfam domain pairs. The grey set in the background represents iPfam domain pairs not found in the three species.

Figure 5

Species distribution of iPfam domain pairs. This pie chart shows how many iPfam domain pairs were found in PDB structures from each species. The total number is larger than the 4030 unique iPfam pairs in the database because an iPfam pair can be found in structures from several species.