Interpreting Notch signals: characterisation of dynamic transcription hubs by live imaging

Abstract

The Notch signalling pathway, crucial for development, activates upon cell-to-cell contact. Despite lacking any amplification, activation of target genes is achieved robustly and efficiently after contact-induced cleavage of the Notch transmembrane protein. How relatively few molecules of the activated, cleaved Notch achieve robust activation of genes is not well understood. Additionally, CSL, the Transcription Factor cleaved Notch interacts with, has relatively short DNA residence times, contrasting with a much longer duration of gene transcription. There are gaps in our understanding of how Notch activation results in transcription complex recruitment at gene regulatory regions and how this translates in the activation of target genes. A live imaging approach was taken to understand the mechanisms for robust activation of Notch targets, focusing on the co-activator Mastermind, a key component of the Notch transcription complex. Specifically, its enrichment and behaviour were characterised at a regulated gene locus in Notch active cells, enabling the study of a localised hub or condensate, and how it promotes transcription of target genes. First, analysis of Mastermind recruitment revealed that it forms dynamic transcription hubs in Notch active (Notch ON) cells. The proposed Notch hubs are formed by a combination of DNA-binding and interactions mediated by intrinsically disordered regions. Notably, Mastermind itself was not essential for hub formation as, when depleted, a modified hub persisted although transcription of Notch targets was lost. Factors that were involved in stabilising Mam were investigated, with CBP/p300 and the CDK8 module of the Mediator complex recruited to the transcription hub and required for transcription. However, of the two complexes, only the CDK8 module had an influence on Mam recruitment at the target locus. Further investigation suggested the enrichment of the CDK8 module and Mam in the transcription hub are interdependent. Second, to monitor the proportion of nuclei that were actively transcribing, single-molecule fluorescence in situ hybridization (smFISH) was performed to detect RNA transcripts. Surprisingly, active transcription of Notch targets was detected in only a third of Notch ON nuclei despite the presence of Mastermind transcription hubs in all nuclei. The low proportion of actively transcribing cells correlated with the proportion of nuclei in which RNA Polymerase II (Pol II) and core Mediator (Med1) were recruited into the transcription hub. The proportion of actively transcribing nuclei was significantly increased following treatments with a second signal; the steroid hormone Ecdysone, whose receptor is known to interact with the Notch-regulated locus. This correlated with the increased recruitment of Pol II and Med1 leading to the hypothesis that the second signal increased the valency of interactions at target enhancers. The results have implications for understanding how stochastic signals can become robust. Subsequently, the effects on the transcription hub upon switching Notch signalling OFF were investigated using a temperature-sensitive and an optogenetic strategy. Both revealed that a hub persisted following Notch inactivation, with Mam enrichment lost but the Transcription Factor CSL remaining. The properties were further tested by reactivation of signalling: hub formation and transcription were achieved more efficiently in pre-activated cells compared to naïve cells. This implies that the locus retains a memory of a previous signal, a property which has implications for iterative rounds of signalling in development. Lastly, the expression and role of a shorter, naturally occurring isoform of the Mastermind protein, MamRC were studied. The MamRC isoform contains a small region of the full protein and has previously been described to function as a dominant negative. The endogenous protein was detected in some larval tissues, suggesting it might have a role in development. When expressed at high levels, MamRC forms phase-separated droplets and gives rise to a range of phenotypic defects, some of which resemble those of reduced Notch activity. However, many of these are reversed when cell death is prevented, leading to the conclusion that its inhibitory function is less direct.