Specificity in transcription factor clustering is encoded in the genome

Transcription factors (TFs) often form clusters in the nucleus. Clusters can facilitate transcription, but it remains unclear how they form. It has been suggested that clusters are seeded by the sequence-specific binding of TFs to DNA, and grow by IDR-IDR interactions that bring in more TFs. This model, however, does not explain how TFs can cluster in specific combinations. Here, we study TF clustering by quantitative imaging of Nanog, Pou5f3, and Sox19b in zebrafish embryos. Using mutant and chimaeric TFs, we show that the formation of a TF cluster requires the DBD as well as at least one of its IDRs. In contrast with the existing model, however, IDRs are not sufficient to join a pre-existing cluster. Instead, both IDR and DBD are needed. Thus, for any TF to join a cluster, motif recognition is required, which explains the specificity in cluster formation. Finally, we show that while IDRs are required to join a cluster, their amino acid sequence is interchangeable, and the DBD can confer specificity to any IDR. Taken together, our work changes the model of cluster formation and explains how specificity is achieved in the organization of transcriptional machinery in the nucleus.