'Perfectly' curvilinear profiles for binding as determined by ITC may in fact be multiphasic

TL;DR

This study reanalyzes ITC binding data for proteasome activator PafE, revealing that multiphasic models with discontinuous transitions better fit the data than traditional curvilinear profiles.

Contribution

It demonstrates that binding profiles previously modeled as smooth curves are more accurately described as multiphasic with jumps, challenging conventional analysis methods.

Findings

Multiphasic models fit ITC data significantly better than curvilinear models.

Discontinuous transitions in binding profiles suggest complex binding mechanisms.

Reanalysis impacts interpretation of protein-protein interaction data.

Abstract

In a structural analysis of the proteasome activator PafE in Mycobacterium tuberculosis, the binding of the activator or shorter constructs to the 20S proteasome core particle (20S CP) or derivatives was measured by isothermal titration calorimetry (Bai et al. Proc Natl Acad Sci USA 113: E1983-E1992. 2016). The data were fitted by the authors by nonlinear least squares to give curvilinear profiles that, at least in part, appear to fit the data very well. However, reanalysis of the data shows that the profiles are much better (P < 0.001) represented as multiphasic, i.e. by a series of straight lines separated by discontinuous transitions, often in the form jumps, than by the conventional curvilinear profiles.