Multimodal interactions between a disordered protein and its folded target at single-molecule level.

Abstract

Cap-dependent initiation of translation is regulated by the interaction of the eukaryotic translation initiation factor 4E (eIF4E) with the 120-residue disordered eIF4E binding protein 2 (4E-BP2) in a phosphorylation-dependent manner. In a previous NMR study (Lukhele et al., Structure 2013) we showed that 4E-BP2 interacts with eIF4E via a dynamic bipartite interface consisting of a α-helical primary binding site and a disordered secondary binding site.In our recent paper (Smyth et al., Biophys. J., 2022) we used single-molecule fluorescence resonance energy transfer (smFRET) to show that the conformational changes of 4E-BP2 induced by binding to eIF4E are non-uniform along the sequence; a central region containing both motifs that bind to eIF4E expands and becomes stiffer, the C-terminal region is less affected. In the same paper, segmental chain dynamics and non-local intrachain contacts were measured via fluorescence anisotropy decay and fluorescence correlation spectroscopy, respectively. Surprisingly, we found that the chain dynamics around sites in the C-terminal region, far away from the two known binding motifs, were significantly reduced upon binding to eIF4E, suggesting that this region is also involved in the highly dynamic 4E-BP2:eIF4E complex. Intermolecular smFRET was used to study the interaction between donor-labelled, surface-immobilized eIF4E and acceptor-labelled, freely diffusing 4E-BP2. FRET efficiency-time trajectories from thousands of individual molecules were used to derive the distributions of on- and off-binding times (τON and τOFF) at the sub-ensemble level. We hypothesize that the C-terminal region of 4E-BP2 engages in transient interactions with the disordered N-terminal region of eIF4E. Intermolecular smFRET was measured for different pairs of labelling sites on eIF4E and 4E-BP2. These data provide important new clues about the architecture and the dynamics 4E-BP2:eIF4E complex and can be used as highly informative restraints in computational ensemble modelling approaches.