Intrinsically disordered proteins and structured proteins with intrinsically disordered regions have different functional roles in the cell

Abstract

Many studies about classification and the functional annotation of intrinsically disordered proteins (IDPs) are based on either the occurrence of long disordered regions or the fraction of disordered residues in the sequence. Taking into account both criteria we separate the human proteome, taken as a case study, into three variants of proteins: i) ordered proteins (ORDPs), ii) structured proteins with intrinsically disordered regions (IDPRs), and iii) intrinsically disordered proteins (IDPs). The focus of this work is on the different functional roles of IDPs and IDPRs, which up until now have been generally considered as a whole. Previous studies assigned a large set of functional roles to the general category of IDPs. We show here that IDPs and IDPRs have non-overlapping functional spectra, play different roles in human diseases, and deserve to be treated as distinct categories of proteins. IDPs enrich only a few classes, functions, and processes: nucleic acid binding proteins, chromatin binding proteins, transcription factors, and developmental processes. In contrast, IDPRs are spread over several functional protein classes and GO annotations which they partly share with ORDPs. As regards to diseases, we observe that IDPs enrich only cancer-related proteins, at variance with previous results reporting that IDPs are widespread also in cardiovascular and neurodegenerative pathologies. Overall, the operational separation of IDPRs from IDPs is relevant towards correct estimates of the occurrence of intrinsically disordered proteins in genome-wide studies and in the understanding of the functional spectra associated to different flavors of protein disorder.

Fig 1. Clusters of the protein variants in the human proteome.

Hierarchical trees of ORDPs, IDPRs, and IDPs, based on different functional profiles (i.e. biological processes, molecular functions, cellular components, and PANTHER protein classes) are here reported. The length of the branches are not identical and reflect the distance matrices in S1 Table. In all the four cases, ORDPs and IDPRs are robustly separated, as indicated by bootstrap scores, the number of times these two groups clustered together divided by the number of bootstrap replicates (1000).

Fig 2. Over- (under-) representation of the protein variants in the PANTHER protein classes.

Bar charts of the normalized differential occurrence of ORDPs, IDPRs, and IDPs in various protein classes, with respect to the human proteome (the reference). Only statistically significant differences are reported (p-value <0.05).

Fig 3. Functional classes are differently enriched in ORDPs, IDPRs, and IDPs.

A functional class is enriched (depleted) in the protein variant that has significantly the highest (lowest) frequency. All the classes shown in this figure passed the enrichment (depletion) test for at least one variant (detailed numerical values can be found in S6 Table).

Fig 4. Over-representation of the variants of disorder in diseases.

Bar charts of the normalized differential occurrence (with respect to the human proteome) of ORDPs, IDPRs, and IDPs in three groups of disease-related proteins.

Fig 5. Enrichment of the protein variants in diseases.

Different groups of disease-related proteins are enriched (depleted) in the protein variant that has significantly the highest (lowest) frequency. Each group is significantly enriched (depleted) in at least one of the variants (detailed numerical values can be found in S11 Table).