How do low-complexity domain interactions control transcription?

Abstract

Abstract A substantial portion of the proteome consists of intrinsically disordered regions (IDRs) that do not fold into well-defined structures. IDRs often comprise low-complexity domains (LCDs), whose repetitive, compositionally biased sequences drive multivalent interactions that can lead to liquid–liquid phase separation, forming biomolecular assemblies. These LCD-mediated assemblies play an important role in transcriptional control, and productive output depends not on their mere presence but on how strongly and selectively LCDs interact. The present review integrates the optimal-interaction, or ‘Goldilocks,’ principle with RNA polymerase II phosphorylation and RNA-mediated feedback into a unified mechanistic framework spanning the transcription cycle. We connect dysregulation of LCD-mediated multivalent interactions to pathological transcription in cancer, neurodegeneration, and developmental disorders, whether through oncogenic LCD fusions, disease-associated post-translational modifications, or mutations altering phase-separation propensity. We further summarize experimental approaches for studying LCD function, discuss current methodological limitations, and highlight directions for future research.