Biomolecular condensation of human IDRs initiates endogenous transcription via intrachromosomal looping or high-density promoter localization.

Abstract

Protein intrinsically disordered regions (IDRs) are critical gene-regulatory components and aberrant fusions between IDRs and DNA-binding/chromatin-associating domains cause diverse human cancers. Despite this importance, how IDRs influence gene expression, and how aberrant IDR fusion proteins provoke oncogenesis, remains incompletely understood. Here we develop a series of synthetic dCas9-IDR fusions to establish that locus-specific recruitment of IDRs can be sufficient to stimulate endogenous gene expression. Using dCas9 fused to the paradigmatic leukemogenic NUP98 IDR, we also demonstrate that IDRs can activate transcription via localized biomolecular condensation and in a manner that is dependent upon overall IDR concentration, local binding density, and amino acid composition. To better clarify the oncogenic role of IDRs, we construct clinically observed NUP98 IDR fusions and show that, while generally non-overlapping, oncogenic NUP98-IDR fusions convergently drive a core leukemogenic gene expression program in donor-derived human hematopoietic stem cells. Interestingly, we find that this leukemic program arises through differing mechanistic routes based upon IDR fusion partner; either distributed intragenic binding and intrachromosomal looping, or dense binding at promoters. Altogether, our studies clarify the gene-regulatory roles of IDRs and, for the NUP98 IDR, connect this capacity to pathological cellular programs, creating potential opportunities for generalized and mechanistically tailored therapies.