Liquid-liquid phase separation (LLPS) is a biophysical phenomenon through which biomolecules condense and accumulate locally. This effect has been linked to a high increase of activity for enzymatic systems in vitro. To form LLPS, biomolecules interact through a mix of specific and/or non-specific interactions, often transient in nature and involving multivalency. In vitro models of biological LLPS often rely on molecular crowders to induce condensation.
Akin to many other phase-separating biological systems, several elements of the DNA double-strand break repair pathway of non-homologous end-joining (NHEJ) have been shown in our lab to undergo phase separation in vitro in presence of crowding agents. The NHEJ process involves the sequential collective action of numerous proteins to successively tether the severed DNA ends, form synapsis, and ligate the DNA. Of interest to our project are the scaffolding protein homodimers XRCC4 and XLF, and the DNA ligase IV. These proteins consist of both folded and very dynamic intrinsically disordered domains, and are involved in in-cis and in-trans interactions. These three proteins are together already sufficient to form condensates in vitro and in presence of crowding agents – which strongly increases ligation activity in presence of blunt-end linear DNA.
However, there is a conundrum: although these three components can condense at low concentrations, slightly below one micromolar, they are present in sub-LLPS concentrations in vivo. How do they get recruited to the NHEJ complex during DNA-repair? How do components issued from the dilute phase enter condensates? What happens at the surface of the condensates?
We aimed to use the molecular insider’s view provided by NMR and especially Relaxation and high-resolution relaxometry (HRR) to study the behaviour of an NHEJ component interacting with condensates at atomic resolution.
