Unraveling the Thermodynamics of Ultra-tight Binding of Intrinsically Disordered Proteins

Abstract

Protein interactions mediated by the intrinsically disordered proteins (IDPs) are generally associated with lower affinities compared to those between globular proteins. Here, we characterize the association between the intrinsically disordered HigA2 antitoxin and its globular target HigB2 toxin from Vibrio cholerae using competition ITC experiments. We demonstrate that this interaction reaches one of the highest affinities reported for IDP-target systems (K _D = 3 pM) and can be entirely attributed to a short, 20-residue-long interaction motif that folds into α-helix upon binding. We perform an experimentally based decomposition of the IDP-target association parameters into folding and binding contributions, which allows a direct comparison of the binding contribution with those from globular ultra-high affinity binders. We find that the HigA2-HigB2 interface is energy optimized to a similar extent as the interfaces of globular ultra-high affinity complexes, such as barnase-barstar. Evaluation of other ultra-high affinity IDP-target systems shows that a strategy based on entropy optimization can also achieve comparably high, picomolar affinities. Taken together, these examples show how IDP-target interactions achieve picomolar affinities either through enthalpy optimization (HigA2-HigB2), resembling the ultra-high affinity binding of globular proteins, or via bound-state fuzziness and entropy optimization (CcdA-CcdB, histone H1-prothymosin α).

Keywords: conditional folding; fuzzy interactions; intrinsically disordered proteins; picomolar; protein-protein interactions; toxin-antitoxin; ultra-high affinity.

FIGURE 1

Intrinsically disordered HigA2 binds to its globular target with picomolar affinity. (A) The overall structure of HigA2-HigB2 antitoxin-toxin tetramer (PDB 5JAA). Only one pair of chains (heterodimer) is colored for clarity. Globular toxin HigB2 is shown in dark cyan, while three antitoxin's structural segments are shown in indigo (IDP α-helix motif), violet (IDP β-strand motif) and pink (globular domain). Underneath the sequence of HigA2 antitoxin and its truncated versions is shown in the same color scheme. (B) The ITC binding isotherms corresponding to titrations of different HigA2 segments into HigB2 (direct titration in black) or into the nanobody-HigB2 complex (competition titration in indigo, violet and pink). The inset shows the expected final product of titrations and the lines correspond to global fit of direct and competition titrations.

FIGURE 2

Decomposition of the thermodynamic contributions accompanying HigA2₃₋₂₃-HigB2 association into the folding and binding contributions. The overall thermodynamic parameters for association were determined using the competition ITC experiments and are shown in first set of bars on the left (association exp.). The folding contributions (middle) were estimated from the LR model and were subtracted from the overall values to obtain the binding contributions, shown on the right. Parameters are reported at T = 25°C.

FIGURE 3

Interaction surface optimization and frustration analysis. (A) The interface surface-area normalized binding contributions (ΔF _bind/ASA_int) of ultra-high affinity complexes from IDPs (HigA2_3-23-HigB2 and CcdA_37-72-CcdB₂) and from globular proteins. Thermodynamic parameters for binding (folding is subtracted from the overall values in case of IDPs) were normalized per interface surface area (ASA_int) and are shown as bars and reported at T = 25°C. (B) Frustration index analysis of HigA2_3-23 in complex with the target. Inter-residue interactions are shown on the structure of the HigA2_3-23-HigB2 complex with lines colored according to frustration index (green no frustration, red frustrated interactions). The values of frustration index for HigA2_3-23 are presented on the right panels: upper shows the conformational frustration index, while the bottom one shows the mutational frustration index (Parra et al., 2016).

FIGURE 4

Ultra-high affinity association of IDPs reaching picomolar affinities employ different thermodynamic strategies. Complexes are ordered according to the degree of energy optimization as estimated by the ΔH _bind/ASA_int values. For the HigA2_3-23-HigB2 complex the degree of energy optimization is similar to one observed for the pairs of globular high affinity binders (Supplementary Table S1, Figure 3A). The CcdA_37-72-CcdB₂ complex exhibits fuzziness in the bound-state (Hadži et al., 2017b) and has lower degree of energy optimization (Supplementary Table S1, Figure 3A). The high affinity complex between ProTα and histone H1 is highly dynamic and is entropically stabilized (Borgia et al., 2018).