It has been proposed that phase-separated transcriptional condensates may facilitate gene activation. However, physical models that clarify the precise role of phase separation in transcriptional regulation remain lacking. Bicoid, a transcription factor that forms an exponential concentration gradient in Drosophila embryos and has been shown to assemble into clusters, offers a unique system to quantitatively connect the transcriptional response to transcription factor concentration. Using fluorescence correlation spectroscopy and single-particle tracking, we have characterized the clusters formed by Bicoid and by Zelda, the global activator of zygotic gene expression, and examined how these clusters relate to the transcriptional responses observed in early embryos. Based on these data, we developed experimentally constrained transcription models that explicitly incorporate liquid-liquid phase-separated condensates. Our analysis shows that the simplest model consistent with current observations requires a three-component phase separation, with Zelda acting as the main driver. In this framework, the principal advantage conferred by phase separation is a sharp local increase in Bicoid concentration at the promoter region.
Fradin et al. (Sun,) studied this question.