Wiz regulates clustered protocadherin genes by restricting CTCF/cohesin loop extrusion in a genomic-distance biased manner

Abstract

Zinc finger (ZF) proteins (ZFPs) constitute the largest family of transcription factors in mammals; however, their regulatory mechanism remains largely elusive. Here we propose COP (C2H2-ZFP occupancy predictor), a deep learning framework that integrates DNA sequence information with protein primary and secondary features to systematically predict ZFP genomic enrichments. Applying COP to the mouse clustered protocadherin (cPcdh) gene locus, we identified dozens of C2H2-ZFPs potentially involved in CTCF-mediated gene regulation with Wiz having the highest number of 12 ZFs. We confirmed Wiz enrichments at all of the CBS (CTCF-binding site) elements across the Pcdh clusters by Myc-tagging the endogenous Wiz gene. Genetic experiments revealed significant increases of expression levels of the cPcdh genes upon Wiz deletion in both neuronal cells in vitro and in mouse brain in vivo. Finally, integrated ChIP-seq, RNA-seq, and 4C-seq analyses demonstrated that Wiz regulates CTCF/cohesin occupancy and long-range enhancer-promoter contacts in a genomic-distance biased manner. Together, these findings have interesting implications regarding molecular mechanism of C2H2-ZFPs in cPcdh gene regulation.