Zeb2 DNA-Binding Sites in Neuroprogenitor Cells Reveal Autoregulation and Affirm Neurodevelopmental Defects, Including in Mowat-Wilson Syndrome

Abstract

Functional perturbation and action mechanism studies have shown that the transcription factor Zeb2 controls cell fate decisions, differentiation, and/or maturation in multiple cell lineages in embryos and after birth. In cultured embryonic stem cells (ESCs), Zeb2's mRNA/protein upregulation is necessary for the exit from primed pluripotency and for entering general and neural differentiation. We edited mouse ESCs to produce Flag-V5 epitope-tagged Zeb2 protein from one endogenous allele. Using chromatin immunoprecipitation coupled with sequencing (ChIP-seq), we mapped 2432 DNA-binding sites for this tagged Zeb2 in ESC-derived neuroprogenitor cells (NPCs). A new, major binding site maps promoter-proximal to Zeb2 itself. The homozygous deletion of this site demonstrates that autoregulation of Zeb2 is necessary to elicit the appropriate Zeb2-dependent effects in ESC-to-NPC differentiation. We have also cross-referenced all the mapped Zeb2 binding sites with previously obtained transcriptome data from Zeb2 perturbations in ESC-derived NPCs, GABAergic interneurons from the ventral forebrain of mouse embryos, and stem/progenitor cells from the post-natal ventricular-subventricular zone (V-SVZ) in mouse forebrain, respectively. Despite the different characteristics of each of these neurogenic systems, we found interesting target gene overlaps. In addition, our study also contributes to explaining developmental disorders, including Mowat-Wilson syndrome caused by ZEB2 deficiency, and also other monogenic syndromes.

Keywords: Mowat-Wilson syndrome; Zeb2; chromatin immunoprecipitation sequencing; embryonic stem cells; neural differentiation; neurodevelopmental disorder; syndromes; target genes; transcription factor; transcriptomics.

Figure 1

Characterization of the heterozygous Zeb2-Flag-V5 mESC line and ChIP-qPCR validation. (A) Allele-specific RT-qPCR using two sets of primers located either in exon7 (and therefore able to detect the whole Zeb2 mRNA produced by both alleles; orange bar) or located in exon9 and the V5-tag (thus recognizing specifically the knocked-in tagged allele; light blue bar); (B) Western blot analysis showing V5 epitope containing Zeb2 in ESC-derived NPCs (at D8 of neural differentiation, ND) in nuclear extracts (NE) and cytoplasmic extracts (CE). Membranes were blotted with anti-V5 antibody (left panel, αV5) or anti-Zeb2 antibody (right panel, αZeb2, [20]). As a control, a fraction of Zeb2-rich extract obtained from HeLa cells transfected with a pcDNA3-V5Zeb2 vector was also separated in the same gel; (C) Zeb2 mRNA levels in wild-type (WT, grey bar) and Zeb2-V5 (orange bar, clone 2BE3, indicated as Zeb2-V5) mESCs during ND, as determined by RT-qPCR; (D) RT-qPCR for ND, using marker genes: pluripotency marker genes Nanog, Pou5f1 (Oct4) and Sox2 are downregulated in Zeb2-V5 mESCs similarly to WT. The neuronal marker gene Pax6 is also significantly upregulated during differentiation, such as in WT mESCs; (E) Scheme of the mouse Cdh1 promoter showing the three E-boxes located upstream of the ATG start codon. Zeb2 binds specifically to only two of these, indicated as R1 [25]; (F) ChIP-qPCR showing enrichment for Zeb2-V5 binding to the R1 region of the Cdh1 promoter. Agarose beads were used as negative control (in grey).

Figure 2

Zeb2-V5 protein is recruited at the TSS of transcriptional regulator encoding genes, predominantly those classified in Wnt signaling. (A) 2432 peaks were selected from our ChIP-seq data set (see Section 2). Of these, 94% are associated with protein-coding loci, 5% with miRNAs, and the remaining 1% map to regions without functional annotation (NA); (B) Frequency plot showing the binding of Zeb2-V5 at and around (−10 to +10 kb) the TSS; (C) The 2294 peaks map to 1952 protein-encoding genes, many of which operate in Wnt signaling (Table S1) or are (as shown in panel (D)) transcriptional regulators; (E) Of the 1952 protein-encoding genes, 1244 are differentially expressed during ND when compared to the undifferentiated state (D0). Of these 1244 genes, 335 are differentially expressed at all three time points of ND. We also list a few examples of DEGs uniquely expressed at one time point, as well as these that are shared among three time points; a full list is provided in Table S2; (F) Overlap of the Zeb2-bound regions with H3K27ac, H3K4me1, and H3K4me3 histone marks in the −10/+10 kb from the TSS of up or downregulated genes at D8 of mESCs differentiation; (G) Overlap of the Zeb2-bound regions outside the −10/+10 kb region from the TSS with histone marks.

Figure 3

Schematic representation of the meta-analysis of Zeb2−bound genes versus RNA−seq datasets. (A) 108 genes bound by Zeb2 are also differentially expressed in the three data sets from other studies in mouse models and ESCs (see main text for details, [23,36,37]). Cxcr4 is the only DEG bound by Zeb2 and common among the three data sets (for discussion, see main text); (B) These 108 genes mainly map to signaling pathways regulating stem cell pluripotency, and effects of TGFβ family, FoxO, and Hippo signaling/activity; (C) The 108 genes cluster as important regulators of developmental processes, locomotion, and signaling.

Figure 4

shRNA-mediated KD of Zeb2 discriminates between primary and secondary target genes. (A) Schematic overview of the shRNA transfection targeting Zeb2 (shZeb2) and read-out of the effect. Cellular aggregates at D8 of ND are dissociated and transfected with shZeb2 or against a scrambled, control sequence (shCTRL). Read-out is done two days after the start of shRNA addition. The list of shRNAs is given in Table 3. (B) RT-qPCR measurements: Zeb2 levels after KD were reduced to 40–50% of their normal level (shZeb2, orange bars) compared to shCTRL (blue bars). Bmp7, Cited2, Nanog, Sema3f, Smad1, Smad2, Smad3, and Tgfbr2 were upregulated following Zeb2 KD, whereas genes encoding for neuronal specification and migration (Cxcr4, Lhx5, Ntng2, Pax6, and Tcf4) were downregulated.

Figure 5

Deletion of the Zeb2-binding autoregulatory site (chr2:45109746-45110421) impairs Zeb2 mRNA levels and neuronal markers. (A) Schematic overview of the log (FoldEnrichment) of the peaks identified by ChIP-seq located in the mouse Zeb2 locus, and localization on top of Zeb2 intron/exon structure. The highest peak is located 232 bp upstream of the first translated exon. The grey arrow indicates the TSS of Zeb2; (B) RT-qPCR: Zeb2 mRNA levels are strongly reduced in the Zeb2^ΔP/ΔP clone. For the target genes validated with shRNA (see also Figure 4), most, but not all of the genes found to be affected following Zeb2 KD are also deregulated in Zeb2^ΔP/ΔP mESCs, in particular Id2 and the neuronal markers Cxcr4, Lhx5, Ntng2, Pax6, and Tcf4.