Eukaryotic transcription factor binding kinetics

Abstract

Transcription represents a key regulatory step in gene expression and transcription factors (TFs) are the key molecular players of this process. TFs inhibit or promote the assembly of the transcriptional machinery by binding DNA and several co-factors. Whereas biochemical studies suggested stable TF-DNA interactions (timescales of hours), fluorescence microscopy showed that TFs bind and dissociate from DNA within few seconds. However, the biological implications of such a dynamic behaviour are still elusive. TF-DNA binding kinetics have been mainly measured in interphase cells, i.e. when active transcription occurs. Though recent findings suggest that even when transcription is silenced in mitosis, some TFs bind to chromatin, a phenomenon known as ‘mitotic bookmarking’. This mechanism was proposed to regulate phenotypic maintenance by ensuring faithful gene reactivation at the mitosis-G1 transition. At present, though, evidence on whether and how binding kinetics determine TFs’ mitotic chromosome binding is missing. In this work, we aimed to determine the role of TFs-DNA binding kinetics in two functional aspects: (i) transcriptional activation of genes involved in the inflammatory response; (ii) differential mitotic chromatin binding of TFs regulating reprogramming and differentiation. We investigated the first aspect by measuring the binding kinetics and transcriptional activity of p65-NFi«B and mutants to its DNA-binding domain (DBD) and transactivation domains (TADs). By performing single-molecule tracking (SMT) of fluorescent p65 molecules and extrapolating p65 transcriptional activation potentials from RNAseq, we demonstrated that p65 binding time (t_b) generally correlates to its capability to induce transcriptional activation. Unexpectedly, deletion of p65 transactivation domains (TADs) did not significantly reduce p65 t_b ( 3-6 s) if compared to wt-p65 (t_b 4 s), although it strongly inhibited p65-mediated transcriptional activation. We also showed 1000-fold reduction of DNA-binding affinity (Y36A/E39D mutant) retained 3% of long-lasting binding events (t_b 12 𝑠) linked to no-activity. This prompted us to speculatively suggest that p65 may bind DNA indirectly. In a second part, we completed the kinetic characterization of Sox2 and Oct4, two TFs of the differentiation/reprogramming network. Preliminary findings showed that Sox2 and Oct4 display different mitotic chromosome binding and comparable dissociation rates from mitotic chromatin. Thus, we tested whether their distinct retention on mitotic chromosomes could arise from different association rates. To test this hypothesis, we performed SMT of Sox2- or Oct4-Halo in ES living cells to compute their bound fraction during different phases of their cell cycle. Sox2 and Oct4 were equally retained on interphase cells and did not show significant differences even during mitosis, suggesting that neither dissociation nor association rates at specific binding sites determine the differential retentions of the two TFs. Our findings indicate that (i) p65 binding kinetics predict functional outcomes as long as protein-protein interactions are preserved; (ii) such interactions do not stabilize p65-DNA binding but critically define binding events as transcriptionally productive; (iii) differential retention of Sox2 and Oct4 on mitotic chromatin is not due to differences in their binding kinetics at specific sites, suggesting a contribution from TFs-DNA non-specific interactions.