Liquid liquid phase separation of transcription factors Bicoid and Zelda as a mechanism of transcriptional regulation in Drosophila melanogaster embryos.

Abstract

The morphogenic transcription factor (TF) Bicoid (Bcd) helps tissue patterning by activating transcription in a spatially dependent manner along its exponentially de- caying anterior-posterior (AP) concentration gradient. The hunchback gene (hb), one of many Bcd target genes, exhibits a stepwise expression boundary with high anterior activation and low posterior activation, connected by a steep boundary near the mid- point of the embryo. Despite being one of the first and most well-studied examples of morphogenic gene regulation, the hb boundary continues to astound with the remark- able speed, precision, and sharpness with which it forms, evading a complete physical explanation. Recent observations have taken note of the spatial heterogeneity of Bcd molecules within nuclei and have speculated about the potential role of liquid-liquid phase separation (LLPS) on Bcd-hb gene regulation. This work combines theory with experiment, using Airyscan fluorescent microscopy to experimentally probe the clustering properties of Bcd and an associated TF Zelda (Zld), across AP position and time, while using computational models to simulate the potential impact that LLPS of Bcd and Zld may have on hb gene expression. Although theoretical models of a mid-embryo phase transition show an increase in transcription from the non- clustering case and a sharper expression boundary consistent with experimentally measured gene expression, experimental observations of cluster properties contradict iv the assumption of a mid-embryo phase transition. Rather, Airyscan microscopy sug- gests clustering throughout the embryo and a dependence of Bcd clustering on total Bcd nuclear concentration. Extending theoretical models to a three-component LLPS system, gene expression boundary steepness is again improved from the non-clustering case, this time at positions consistent with experimental data from the literature and with better agreement with experimental cluster properties. This data shows that LLPS may play a role in gene regulation in this and similar systems.