Exploring Relationships between the Density of Charged Tracts within Disordered Regions and Phase Separation

Abstract

Biomolecular condensates form through a process termed phase separation and play diverse roles throughout the cell. Proteins that undergo phase separation often have disordered regions that can engage in weak, multivalent interactions; however, our understanding of the sequence grammar that defines which proteins phase separate is far from complete. Here, we show that proteins that display a high density of charged tracts within intrinsically disordered regions are likely to be constituents of electrostatically organized biomolecular condensates. We scored the human proteome using an algorithm termed ABTdensity that quantifies the density of charged tracts and observed that proteins with more charged tracts are enriched in particular Gene Ontology annotations and, based upon analysis of interaction networks, cluster into distinct biomolecular condensates. These results suggest that electrostatically-driven, multivalent interactions involving charged tracts within disordered regions serve to organize certain biomolecular condensates through phase separation.

Figure 1.

Net charge per residue plots of NPM1 (a) and another nuclear protein, SURF6 (b). Regions of predicted disorder highlighted in purple. Highlighted charged tracts have area greater than 1 Process diagram for the calculation of ABTdensity (c).

Figure 2.

Scatter plot showing correlation between ABTscore values (blue data points) and ABTdensity values (orange data points), and the number of disordered residues in each protein with one or more disordered regions (a). Enhanced box plots showing ABTscore (b) and ABTdensity (c) distributions for the whole proteome, nucleolar proteome, stress granule (SG) proteome, and NUP98 interactors. P values are reported when a protein set’s mean is different from the whole proteome’s at a p<0.05.

Figure 3.

Heatmap of the Gene Ontology enrichment analysis for processes annotations of proteins in Groups 1 through 4. Asterisks indicate significance at p≤0.05.

Figure 4.

Interaction network for Group 1. Nodes represent proteins and edges represent physical or genetic interactions. Orphan proteins are not shown. Each color represents clusters described in Table 3. Yellow, nucleolar proteins; green, proteins associated with nucleosomes and heterochromatin; cyan, proteins associated with transcription bodies; and blue, proteins associated with protein degradation.