Natural Selection of 100% Intrinsically Disordered Biologically Functional Disease Related Pfam Protein Domains

Abstract

An analysis is presented of the distributions of predicted intrinsic disorder in biologically functional Pfam domains in 193024 members of the version 23.0 Pfam seed database representing 12456 unique domains, families, and repeats. 616 mammalian members of this set, representing 315 biological functions, are predicted to have 100% intrinsically disordered Pfam domains. The lower phylogenetic domains have smaller proportions of 100% disordered Pfam domains in the following order: mammals > other eukaryota > viruses > bacteria > archaea. Detailed structure-function maps of sixteen examples of these domains presently known to be directly implicated in human disease show that many of these proteins have multiple Pfam domains, and that efforts to crystallize these proteins have been unsuccessful. We also find that about 2% of Pfam domains in Single cell prokaryotes, and about 6% in multiple cell eukaryotes are predicted to be 100% disordered, with one exception, Pneumocystis (pneumonia), with 28% disordered Pfam domains, suggesting that intrinsic disorder is related to the survival of some single cell organisms infecting mammals. It is proposed that shorter 100% disordered functional domains and regions in proteins confer a selective advantage that has contributed to the evolution of more complex organisms and of single cell organisms that they host. The median length of disordered regions in this 100% disorder group is 59 residues. Distributions of disorder in each of these five phylogenetic domains differ at the 0.999 level with p values less than 1x10∗∗-10. When Pfam domains are separated on the basis of their hydrophobicity and percent disorder, we find at least three linearly independent populations, or flavors, of intrinsic disorder, one of which is the completely disordered state.