Preview

PDF, English Download (12MB) | Terms of use

Abstract

Transcription is a multistep process that is tightly regulated by transcription factors (TFs). TFs typically comprise two subdomains - a DNA-binding domain (DBD) and an activation domain (AD). Properties of TF target site binding are attributed to the DBD. The AD is thought to determine interactions with the transcriptional machinery, the acquisition of co-activating chromatin modifications and, potentially, the formation of phase-separated nuclear compartments via multivalent interactions. However, it remains unclear if DBD and AD really act independently in determining crucial parameters that govern the induction of transcription. In particular, it is unknown if TF binding kinetics and binding site residence times regulate transcription independent of equilibrium binding parameters and whether phase separation caused by multivalent TF interactions is functionally relevant for activation. In this thesis, experimental and analytical approaches were developed and applied that provide mechanistic insights from the highly informative analysis of TF binding and transcription kinetics. Techniques were introduced to measure TF binding kinetics and to follow transcription using light dependent TF recruitment. These approaches were automated and software packages for the analysis of the resulting data were developed to test a large number of conditions in single cells with high time resolution. The wide applicability of light-induced transcription time courses was demonstrated by two proof-of-concept applications: the detection of transcriptional memory and the discrimination of stochastic models using heterogeneous single-cell trajectories. The framework was then applied to reveal a functional link between binding properties of the DBD, multivalent interactions of the AD and the dynamics of transcriptional (co-)activation. Specifically, the following conclusions could be reached: (1) Reduced TF residence time decreased transcription, even for identical binding site occupancy. (2) Multivalent interactions of the AD stabilized chromatin binding of weakly bound TFs and led to the recruitment of an indirectly bound fraction of molecules. (3) ADs with strong multivalent interactions activated faster and more strongly. (4) Phase-separation into macroscopic droplets did not enhance transcription and could in some conditions even have a suppressive effect. (5) Acetylation of histone at lysine residue 27 (H3K27ac) and the binding of BRD4, which interacts with H3K27ac, were induced by indirectly and transiently bound activators under conditions that were not sufficient to induce RNA production. (6) H3K27ac and BRD4 were not strictly necessary for transcription, but had an enhancing effect. Based on these findings the thesis provides an integrated view of TF activity, in which multiple, interdependent properties of DBD and AD increase transcriptional output. These include long TF residence time, high binding site occupancy, complex stabilization by multivalent interactions and interactions with co-activators, but not phase-separation into macroscopic compartments. These findings provide insights into the different TF features that govern their ability to activate transcription and for the design of synthetic TFs.