Periodic Reporting for period 2 - bInDR (DNA binding specificity in-vivo)
Période du rapport: 2022-12-01 au 2024-05-31
Our project addresses a fundamental question in gene regulation: how transcription factors (TFs) locate their target sites in large genomes specifically and rapidly. Both properties are not easily explained in prevailing models. We have proposed a new paradigm, which attributes both the specificity and speed of the TF-target process to intrinsically disordered regions (IDRs) within the TF sequence. IDRs are enriched in TF sequences but have remained largely uncharacterized. The bInDR project aims to define this role of IDRs in directing TF binding across genomes, focusing on budding yeast as a model. To establish this new paradigm, we are addressing its generality amongst TFs, the mechanistic basis through which IDR directs DNA binding, the new molecular sequence grammar that encodes this specificity, and the dynamic consequences for the TF-target search process
I the past 2.5 years since we began working on this project, we have made significant advance in all three aims of the project:
1. General use of IDR in directing TF binding: we confirmed the generality of the IDR-based TF binding. For this, we considered >50 budding yeast TFs, which we analyzed trough systematic sequence truncations and genome-wide mapping. Through this analysis we were able to classify TFs based on the relative contribution of the IDR and the DNA binding domain (DBD) to binding preference. A paper describing these results was published in Molecular Cell (Kumar et al. 2023)
2. The IDR sequence grammar associated with its role in directing binding preference: We began to investigate the IDR sequence grammar using the budding yeast TF Msn2 as a model. This TF has a long (>500AA) IDR. We found That this IDR directs Msn2 binding preferences using multiple specificity determinants that spread across this full sequence. To reveal the sequence bases of these determinants we designed and analyzed >120 IDR mutants, each containing 12-120 residue changes spread across the sequence. These mutants were designed to test different hypothesis and were analyzed using genome-wide profiling. This analysis defined the key IDR design features to be a spread of hydrophobic (mostly aliphatic) residues within an otherwise hydrophilic sequence. A paper describing these results was published in Nucleic Acid Research (Jonas et al. 2023)
3. Towards the molecular basis of IDR-based binding preferences: In thinking of the mechanistic basis of IDR-directed binding, we considered interaction with other co-binding TFs as our most favored hypothesis. To test this hypothesis, we again focused on the Msn2 model. Through different projects in the group, we mapped binding profiles of >95% of budding yeast TFs, all under the same conditions, which enabled to define all TFs co-localizing to Msn2-bound promoters. We identified 14 such TFs, and tested for their potential binding interactions with Msn2 through systematic analysis in which we deleted these TFs individually and in combination. Through this analysis, we identified some unidirectional dependencies between these Msn2-related TFs. However, none of these complied with the prevailing model of co-binding dependencies. Furthermore, in the context of Msn2, this analysis refuted the possibility that the IDR direct binding preference through interactions with other TFs. A paper describing these results was published in Cell Systems (Lupo et al. 2023)
4. IDRs encode multiple functions using interwoven sequence grammars: IDRs of TFs also contain activation domains (ADs). While these can be short (~dozens of AA), their design appears to resemble the one we identified as associated with IDR role in binding preferences. This suggested to us that these two properties – activation and binding – might be guided by the same molecular mechanisms, for example interaction with co-activators. This hypothesis was further raised by reviewers of our paper, which made it more completing for us to investigate at some depth. We therefore extended the analysis of our 161 designed IDR mutants by analyzing their effect on transcription induction and co-factor recruitment recruiting. Similar to the binding determinants, multiple ADs were spread across IDRs, in locations that overlapped regions contributing to binding preferences. However, those determinants of gene induction and binding preferences differ in two key aspects as binding determinants were highly redundant and invariant to specific sequence motifs, while transition activity required the existence of multiple determinants and was dependent on specific sequence motifs. Based on this data, we propose a model in which TF IDRs encode multiple functions using an interwoven grammar of hierarchical complexity. A paper describing these results was published in Nucleic Acid Research (Mendel et al. 2023)
1. General use of IDR in directing TF binding: we confirmed the generality of the IDR-based TF binding. For this, we considered >50 budding yeast TFs, which we analyzed trough systematic sequence truncations and genome-wide mapping. Through this analysis we were able to classify TFs based on the relative contribution of the IDR and the DNA binding domain (DBD) to binding preference. A paper describing these results was published in Molecular Cell (Kumar et al. 2023)
2. The IDR sequence grammar associated with its role in directing binding preference: We began to investigate the IDR sequence grammar using the budding yeast TF Msn2 as a model. This TF has a long (>500AA) IDR. We found That this IDR directs Msn2 binding preferences using multiple specificity determinants that spread across this full sequence. To reveal the sequence bases of these determinants we designed and analyzed >120 IDR mutants, each containing 12-120 residue changes spread across the sequence. These mutants were designed to test different hypothesis and were analyzed using genome-wide profiling. This analysis defined the key IDR design features to be a spread of hydrophobic (mostly aliphatic) residues within an otherwise hydrophilic sequence. A paper describing these results was published in Nucleic Acid Research (Jonas et al. 2023)
3. Towards the molecular basis of IDR-based binding preferences: In thinking of the mechanistic basis of IDR-directed binding, we considered interaction with other co-binding TFs as our most favored hypothesis. To test this hypothesis, we again focused on the Msn2 model. Through different projects in the group, we mapped binding profiles of >95% of budding yeast TFs, all under the same conditions, which enabled to define all TFs co-localizing to Msn2-bound promoters. We identified 14 such TFs, and tested for their potential binding interactions with Msn2 through systematic analysis in which we deleted these TFs individually and in combination. Through this analysis, we identified some unidirectional dependencies between these Msn2-related TFs. However, none of these complied with the prevailing model of co-binding dependencies. Furthermore, in the context of Msn2, this analysis refuted the possibility that the IDR direct binding preference through interactions with other TFs. A paper describing these results was published in Cell Systems (Lupo et al. 2023)
4. IDRs encode multiple functions using interwoven sequence grammars: IDRs of TFs also contain activation domains (ADs). While these can be short (~dozens of AA), their design appears to resemble the one we identified as associated with IDR role in binding preferences. This suggested to us that these two properties – activation and binding – might be guided by the same molecular mechanisms, for example interaction with co-activators. This hypothesis was further raised by reviewers of our paper, which made it more completing for us to investigate at some depth. We therefore extended the analysis of our 161 designed IDR mutants by analyzing their effect on transcription induction and co-factor recruitment recruiting. Similar to the binding determinants, multiple ADs were spread across IDRs, in locations that overlapped regions contributing to binding preferences. However, those determinants of gene induction and binding preferences differ in two key aspects as binding determinants were highly redundant and invariant to specific sequence motifs, while transition activity required the existence of multiple determinants and was dependent on specific sequence motifs. Based on this data, we propose a model in which TF IDRs encode multiple functions using an interwoven grammar of hierarchical complexity. A paper describing these results was published in Nucleic Acid Research (Mendel et al. 2023)
We expect to establish IDRs as the key missing piece in our understanding of the fundamental process of TF-target search across genomes, explaining key mysteries of this process.