Mechanisms of transcription control by intrinsically disordered region interactions
Among numerous proteins involved in transcription, we currently focus on DNA-binding transcription factors, which are key regulators of mammalian transcription. They contain sequence-specific DNA-binding domains and transactivation domains that participate in specific protein-protein interactions to direct gene activation. Whereas numerous atomic structures of DNA-binding domains have provided us a concrete understanding of how transcription factors interact with DNA, transactivation domains ubiquitously contain intrinsically disordered low-complexity sequence domains (LCDs) that are not amenable to conventional structure-function relationship analysis. Mutations in these LCDs not only disrupt transcription but also have been implicated in numerous human diseases. However, how LCDs successfully execute specific transactivation functions has been a long-standing puzzle due to a lack of tools for probing LCD behaviors.
Recent studies have demonstrated the power of quantitative single-cell and single-molecule imaging in tackling this fundamental question. Direct visualization of the dynamic behavior of transcription factor LCDs in vivo has revealed that dynamic, multivalent, and selective interactions occur among the LCDs, which mediate the formation of local high-concentration “hubs” of transcription factors at target genomic loci. Under certain conditions, such as overexpression of LCDs, these hubs can develop into macroscopic liquid-liquid phase separated droplets. Multivalent LCD-LCD interactions stabilize transcription factors binding to DNA, recruit RNA polymerase II, and play an essential role in transactivation.
The discovery and characterization of LCD-LCD interactions represents one step further toward resolving LCD-mediated transcriptional regulation. However, there remains a big black box between an LCD interaction hub and the transcription output. Besides transcription factors, a massive crowd of other regulatory proteins are involved in transcription, including the basal transcription machinery, co-activator complexes, and epigenetic regulators. All these players form an extremely complex network involving thousands of specific protein-protein interactions to enable precise transcriptional regulation. It remains unclear how the assembly of LCD hubs mediates these critical interactions and how these interactions lead to transcription.
We aim to elucidate the detailed mechanisms by which a transcription factor LCD interaction hub directs gene activation. We want to visualize the dynamic interplay between LCD hub assembly and transcription, map the complex molecular interaction network underlying the interplay, and compare the regulatory behaviors of different LCDs. To this end, we leverage our expertise in single-molecule imaging and combine it with genome editing and biochemical approaches to build novel quantitative imaging platforms. In addition, we employ a host of different approaches in genetics, molecular and cell biology, as well as functional genomics to address our research questions. We are interested in both normal transcription factors and aberrant transcription factors that cause different forms of cancer and diabetes. We are currently investigating oncogenic transcription factors that are produced by chromosome translocation and contain LCDs from the FET family proteins (FUS/EWS/TAF15), such as the chimeric transcription factor EWS::FLI1, which drives oncogenesis of Ewing’s sarcoma, the second most common pediatric bone cancer.