Most biochemical reactions have evolved in crowded intracellular environments. However, the complexity of intracellular environments is often neglected in structural or functional studies of proteins. In these cases, reactions involving proteins are deliberately separated from the perturbations of co-solutes in order to simplify data acquisition and interpretation. Having acquired an enormous body of knowledge under these simplified dilute-buffer conditions, methodological progress of the past two decades has made the study of proteins inside living cells increasingly accessible and concomitantly kindled an interest to investigate proteins in their native habitat. Naturally, major questions that arose were to what extent ubiquitous transient interactions alter protein structure, function and thermodynamics and, not least, what role protein surfaces and their physicochemical properties play in determining the frequency and duration of these diffusive encounters.

By looking at the rotational-tumbling rates of three structurally well-characterized proteins in live cells with nuclear magnetic resonance (NMR) relaxation, we expand on previous research performed in the bacterium Escherichia coli and establish the physicochemical principles that determine diffusive interactions in the mammalian cytosol of the human ovarian cancer cell line A2780. Just as in E. coli, net charge is the dominating factor in regulating protein interactivity, albeit with the impact on rotational retardation greatly diminished. We ascribe this to the generally lower macromolecular concentrations in the eukaryotic cytosol, and put forward a hypothesis in which less stringent rules regarding protein surface decoration in eukaryotes could have facilitated the development of multi-cellular organisms. Furthermore, by developing a model where a distribution of differently sized interaction partners is taken into account when examining rotational retardation, we reconcile transverse and longitudinal in-cell relaxation with theory, and are able to estimate the populations of the bound and free form of a set of reporter proteins. Looking at the populations of bound protein instead of a mean-field rotational retardation finally allows us to re-assess the guiding rules behind diffusive cytosolic interactions. Last, we outline a putative mechanism behind the in-cell destabilization of a variant of Superoxide dismutase 1 (SOD1^barrel). By mimicking generic poly-anionic intracellular co-solutes with poly-acetic acid (NaPAc1200), we identify the positively charged N-terminal portion of the unfolded form of the protein as the interaction site with the highest affinity. Further examining the unfolded ensemble of SOD1^barrel with a mutationally destabilized variant reveals a compact state, that remains almost unchanged upon binding to NaPAc1200. This suggests that NaPAc1200-mediated destabilization occurs mainly through mass action, in full accord with the postulated mechanism for in-cell protein destabilization.