A novel long-range enhancer regulates postnatal expression of Zeb2: implications for Mowat–Wilson syndrome phenotypes

Muna M. El-Kasti 0 Timothy Wells 0 David A. Carter 0 0 School of Biosciences, Cardiff University , Cardiff, UK The zinc-finger, E-box-binding homeobox-2 (Zeb2) gene encodes a SMAD-interacting transcription factor that has diverse roles in development and disease. Mutations at the hZeb2 locus cause Mowat - Wilson syndrome (MWS), a genetic disorder that is associated with mental retardation and other, case- and sex-dependent clinical features. Recent studies have detailed microRNA-mediated control of Zeb2, but little is known about the genomic context of this gene or of enhancer sequences that may direct its diverse functions. Here, we describe a novel transgenic rodent model in which Zeb2 regulatory sequence has been disrupted, resulting in a postnatal developmental phenotype that is autosomal dominant. The phenotype exhibits a genotypeby-sex interaction and manifests primarily as an acute attenuation of postnatal kidney development in males. Other aspects of embryonic and neonatal development, including neuronal, are unaffected. The transgene insertion site is associated with a 12 kb deletion, 1.2 Mb upstream of Zeb2, within a 4.1 Mb gene desert. A conserved sequence, derived from the deleted region, enhanced Zeb2 promoter activity in transcription assays. Tissue and temporal restriction of this enhancer activity may involve postnatal changes in proteins that bind this sequence. A control human/mouse VISTA enhancer (62 kb upstream of Zeb2) also up-regulated the Zeb2 promoter, providing evidence of a string of conserved distal enhancers. The phenotype arising from deletion of one copy of the extreme long-range enhancer indicates a critical role for this enhancer at one developmental stage. Haploinsufficiency of Zeb2 in this developmental context reflects inheritance of MWS and may underlie some sex-dependent, non-neural characteristics of this human inherited disorder. - ZEB2 (ZFHX1b/SIP1) is a two-handed zinc finger/homeodomain transcription factor (1,2). It is closely related to another protein ZEB1 (ZFHX1a/dEF1); both of these factors regulate gene transcription through similar, E-box-mediated mechanisms (1,3,4). However, ZEB1 and ZEB2 are functionally distinct in many contexts exhibiting different expression patterns (5) and a distinct SMAD interaction domain (2) that mediates differential regulation of TGFb/BMP signalling (6,7). Zeb2 is a complex gene with numerous transcript variants (8,9) that map to 2q22 in the human genome. Zeb2 function has primarily been characterized within mammalian development through gene targeting in mice and clinical analysis of a human Zeb2-associated developmental disorder. A Zeb2 null mutation in mice results in early embryonic death with marked neural tube/crest defects (10). Human Zeb2 mutations in one allele are associated with a complex, multi-organ developmental disorder involving severe mental retardation, termed Mowat Wilson syndrome (MWS; 11; OMIM #235730). These findings are suggestive of both diverse roles for Zeb2 and gene dosage dependency at critical stages of development. Experimental evidence for distinct neural roles of Zeb2 at different developmental stages has come from conditional gene-targeting strategies in mice showing roles in cortical neurogenesis (12), hippocampal formation (13) and myelination (14). The non-neural roles of Zeb2 are poorly understood, but one recent study has described an important role in uterine function (15). Given the extensive characterization of Zeb2 function, surprisingly little is known about the genomic mechanisms that control Zeb2 transcription. Functional promoters have been mapped for mouse Zeb2, but transgenic analysis of promoter/enhancer sequences has not been conducted and functional BACs have not been generated in large-scale regulatory genome projects (GENSAT; www.gensat.org). In the current study, we have obtained evidence for long-range enhancers that have an important role in regulating postnatal expression of Zeb2 but do not appear to participate in embryonic (neuronal) regulation of this gene. This evidence of temporally active enhancer sequences indicates a novel mechanism that allows this gene to have multiple distinct roles in development. Transgenic phenotype In the current study, we have identified Zeb2 enhancers by analysing disruption of these sequences in a novel insertional mutant transgenic line generated in our laboratory. The transgenic rat line (L4) was identified as the single line exhibiting a phenotype of five lines generated with the same transgene. The phenotype was of particular interest because of an abrupt postnatal onset. Transgenic offspring were visually and behaviourally indistinguishable from wild-type littermates at birth, but males failed to develop beyond the postnatal day 6 (PN6) stage culminating in subsequent postnatal lethality. Female transgenic offspring exhibited delayed development but thrived and have been maintained for up to 14 months of age. Following the initial observation of postnatal pathology in males within several independent litters, an ethical strategy was imposed wherein males were killed at PN6. The cause of death is clearly related to a failure of postnatal maturation (presumed renal failuresee below) although because of the litter management, other causes cannot be definitively excluded. It should be noted that post-mortem examination of males on PN6 revealed milk-engorged stomachs, indicating that the postnatal decline in health is not due to maternal rejection. The L4 rat has been maintained as a hemizygous transgenic line though mating with wild-type (WT) Sprague Dawley male rats. The sexually dimorphic phenotype has been stably inherited over .10 generations of breeding, is transmitted to 50% of offspring (male and female) and therefore appears to represent an autosomal dominant trait. Genotyping multiple litters revealed equal proportions of the four types (WT male/female, TG male/female; Fig. 1A). Southern blotting of genomic DNA (Fig. 1B) in multiple generations of offspring revealed a stable banding pattern with no segregation, indicating a single insertion site. The phenotype is apparently unrelated to the transgene sequence because multiple other transgenic rat lines produced with the same transgene exhibit no overt phenotype which is, of course, consistent with the transgene design that is limited to promoter and reporter sequence and does not contain EGR-1 coding sequence (16). Southern blot analysis (Fig. 1B) also revealed divergent integration events and similar transgene copy number in these three lines. Transgene transcript levels in the L4 line are also very similar to another line that we have studied (data not shown, 16,17). The 3.6 kb Southern band in two of the lines including L4 (Fig. 1B) probably represents a head tail concatamer of two copies of the transgene that would be predicted to generate a 3.6 kb fragment when cut with BglII. Integration of only two copies of the transgene is (...truncated)