Transcription factor binding: a sharp view of a fuzzy interaction

From the Hahn lab, Basic Sciences Division

DNA is commonly referred to as the instruction manual of the cell – a compendium of thousands of genes, the particular sequences of which determine the activities of the proteins they encode. But, in informing cellular behavior, gene function is only half the story. What differentiates the identities and behaviors of different cells in our bodies is, as I describe in this month’s other Basic Sciences article, in large part a matter of which genes are expressed and which are not. In this regard, DNA is less an instruction manual and more an instruction encyclopedia: “For cardiomyocytes, see page 375. For macrophages, see page 792.” And the information regulating gene expression is predominantly contained not within genes themselves, but within non-coding DNA, the regions between genes. In a process whose logic is far less well understood than the gene-to-protein paradigm, enhancer and promoter sequences are bound by transcription factors (the proteins responsible for turning genes on or off) to regulate gene expression. The rules that govern how transcription factors interact with each other, and with DNA, and how these interactions impact gene expression, are the focus of Dr. Steve Hahn, a professor in the Basic Sciences Division at Fred Hutch and a member of the UW/Fred Hutch Cancer Consortium. In a new article in Nature Communications, a collaboration between UW/Fred Hutch Cancer Consortium members Dr. Hahn and Dr. Rachel Klevit of the University of Washington, led by research scientist Dr. Lisa Tuttle, identified how transcription factors interact to control this process.

Transcription factors often act not on their own, but in complex with other proteins that can modify their functions. The yeast transcription factor at the center of this study, Gal4, can, depending on context, bind to either the Gal80 protein, which represses Gal4’s ability to activate gene expression, or to the Med15 protein, a subunit of a co-activator complex that promotes Gal4’s ability to activate gene expression. Gal4 binds both Gal80 and Med15 via a short, acidic activation domain (AD). While previous work had suggested that specific sequences in Gal4 and Gal80 interact to bring these proteins together, the mechanism by which Gal4 and Med15 bind each other proved more elusive, and appeared not to rely on any particular sequence in these proteins. In this study, the group sought to work out how the Gal4-Med15 interaction occurs.

“There is some controversy regarding which residues contribute to Gal4 AD function”, the authors noted. They therefore first aimed to clarify the region of the protein responsible for this function. By fusing various versions of the Gal4 AD to a DNA-binding domain and measuring the extent to which these fusion proteins activated gene expression, they identified a 31-residue disordered minimal AD containing two regions, termed Region 1 and Region 2, that each contain large hydrophobic/aromatic residues. They further showed, by replacing these hydrophobic residues, that they are critical to AD function. The authors then used nuclear magnetic resonance spectroscopy (NMR) to characterize the interaction between the Gal4 AD and the Med15 activator-binding domains (ABDs). NMR experiments on the free and ABD-bound Gal4 AD reveal the nature of the interaction. The group observed short-lived binding (“no more than 4ms”) dominated by interactions between Region 2 of the Gal4 AD and the Med15 ABDs. Most importantly, they found that Gal4 binds Med15 in multiple orientations, indicative of a “fuzzy” interaction mediated by dynamic interactions between hydrophobic residues rather than stable interactions between specific amino acids within Gal4 and Med15. Drs. Tuttle and Hahn were struck by the distinct modes that the same region of Gal4 uses to bind different protein partners. “[An] important conclusion is that the Gal4 activator can use either this fuzzy binding mechanism or a conventional stable and sequence-specific binding mechanism depending on the protein partner (Med15 or Gal80). This shows that it is the AD target that primarily dictates whether an interaction is fuzzy or not and that activation domains can be selected to have both fuzzy and specific binding properties.”

Another striking observation that the authors made during their NMR studies was that the binding pattern they saw between Gal4 and Med15 was nearly identical to that observed between Med15 and another transcription factor, Gcn4. “A longstanding question in the transcription field has been to understand how many transcription activators of different sequence interact with a much smaller set of transcription coactivators.  Our new work demonstrates that two activators of completely different sequence (Gal4 and Gcn4) both interact with the Mediator subunit Med15 using the same mechanism”, said Drs. Tuttle and Hahn.

In the future, the Hahn lab is looking to expand their work to examine other types of transcription factor interactions. “We’ve made great progress in understanding the properties and features of acidic ADs. However, proteome analysis and other data suggests that many transcription factors (TFs) use other AD types. Some of these ADs almost certainly work by different mechanisms and are specific for distinct gene classes. Our general strategy will be to identify representative TFs of different classes, pinpoint their ADs and targets and to discover important AD features and mechanisms.”

Gal4-Med15 binding
Model of the interactions of the Gal4 Activation Domain (AD) with the Med15 Activator Binding Domains (ABDs). Image provided by Steve Hahn.

This work was supported by the National Institutes of Health.

Fred Hutch/UW Cancer Consortium members Steve Hahn and Rachel Klevit contributed to this work.

Tuttle, L.M., Pacheco, D., Warfield, L. et al. (2021) Mediator subunit Med15 dictates the conserved “fuzzy” binding mechanism of yeast transcription activators Gal4 and Gcn4. Nat Commun 12, 2220.