The nature of protein domain evolution: shaping the interaction network

Abstract

The proteomes that make up the collection of proteins in contemporary organisms evolved through recombination and duplication of a limited set of domains. These protein domains are essentially the main components of globular proteins and are the most principal level at which protein function and protein interactions can be understood. An important aspect of domain evolution is their atomic structure and biochemical function, which are both specified by the information in the amino acid sequence. Changes in this information may bring about new folds, functions and protein architectures. With the present and still increasing wealth of sequences and annotation data brought about by genomics, new evolutionary relationships are constantly being revealed, unknown structures modeled and phylogenies inferred. Such investigations not only help predict the function of newly discovered proteins, but also assist in mapping unforeseen pathways of evolution and reveal crucial, co-evolving inter- and intra-molecular interactions. In turn this will help us describe how protein domains shaped cellular interaction networks and the dynamics with which they are regulated in the cell. Additionally, these studies can be used for the design of new and optimized protein domains for therapy. In this review, we aim to describe the basic concepts of protein domain evolution and illustrate recent developments in molecular evolution that have provided valuable new insights in the field of comparative genomics and protein interaction networks.

Keywords: PDZ domain; Protein domain; interactome.; molecular evolution; superfamily; systems biology.

Fig. (1)

Example of sequence alignment and phylogeny. (A) This figure shows an example alignment of the PDZ domain with different shadings representing the amount of conservation (100, 75 or 50%) at a particular position in the sequence. (B) This tree is the phylogenetic presentation of the alignment in Fig. (1A). It was computed using Bayesian estimation and presents the best-supported topology for the alignment. Numbers indicate % support by the two methods used, while # indicates gene duplication events in the common ancestor and * marks a species-specific duplication event. For computational details, please see [42].

Fig. (2)

Selection on superfamily domain size. (A) Increase in superfamily domain size fitted to a power-law for kinase-like domains (I), Ankyrin-repeats (II), PDZ-like (III), voltage-gated potassium channels (IV), the catalytic domain of metalloproteases (V) and the average increase in superfamily size (VI). R² value for each fit was at least 0.9. (B) Neutral or decreasing family sizes can be found for the MFS general substrate transporters (I), NAD(P)-binding Rossmann folds (II), Ribonucleases H (III), PLP-dependent transferases (IV), periplasmic binding proteins type II (V), ATPase domains of HSP90/topoisomerase II/histidine kinase-like folds (VI) and the average increase in superfamily size (VII) as in 2A.

Fig. (3)

Evolutionary models for protein-protein interactions. The evolution of protein networks is tightly coupled to the addition or deletion of nodes. Additionally, events that introduce mutations in binding interfaces of proteins may result in the addition or loss of links in the network. Node addition may take place through e.g., domain duplication or horizontal gene transfer, while rewiring of the network is mediated by point mutations, alternative splice variants and changes in gene expression patterns.