The regulation of endothelial cell gene expression plays a central role in maintaining hemostatic balance. However, little is known about the key sequence elements responsible for the control of the endothelial gene expression program. We previously identified the cause of low levels of von Willebrand factor (VWF) in the inbred mouse strain RIIIS/J as a regulatory mutation in the gene encoding an N-acetylgalactosaminyltransferase, GALGT2. This low VWF allele, which we termed Mvwf1, results from a tissue-specific switch in expression program from intestinal epithelium to vascular endothelium. The ectopic expression of Galgt2 in vascular endothelial cells results in aberrant post-translational modification of VWF, leading to accelerated clearance. The specific regulatory DNA sequences responsible for this remarkable lineage-specific transcriptional switch were mapped genetically to an ~300kb region surrounding the Galgt2 gene. To further define the critical regulatory sequences, we have generated transgenic mice using three overlapping C57BL/6J bacterial artificial chromosomes (BACs) encompassing the Galgt2 gene. All three transgenes directed wild-type (C57BL/6J) GI epithelial expression of Galgt2, mapping the GI specific regulatory elements to an 84kb interval shared by these three BACs. Direct sequencing of RIIIS/J over >80kb flanking the transcriptional start site identified a region of 2–3% sequence divergence from C57BL/6J beginning ~25kb upstream of exon 1 and extending ~10kb into intron 1. This region contains several insertions and deletions ranging from 120bp to 6.4kb. We are currently testing chimeric RIIIS/J:C57BL/6J BACs to fine map the critical regulatory elements responsible for the switch. PCR genotyping of 49 mouse strains identified a total of 9 strains that have RIIIS/J polymorphisms within the region of high sequence divergence. Extensive sequencing confirmed a common haplotype block of at least 55kb in length that is shared by all 9 Mvwf1 strains that also exhibit the GI epithelial to endothelial switch in Galgt2 expression. These data identify an ancient founder allele present in a number of mouse strains and retained at a low level within the wild mouse population. In a survey of the Jackson Phenome Database, Mvwf1 strains accounted for 5 of the 6 highest aPTTs, establishing this single allele as the most common cause of prolonged aPTTs in laboratory mice. The prevalence of this allele also suggests that it has been maintained through positive selective pressure. In summary, we have identified a common allele leading to low VWF levels in the mouse. We have characterized a haplotype block shared by all Mvwf1 mice that begins at least 35kb upstream of the Galgt2 transcriptional start site and overlaps the 84kb interval that confers wild-type Galgt2 expression. Continued analysis of these sequences should provide important new insights into endothelial cell-specific gene expression and the more general problem of tissue specific gene regulation.

Author notes

Corresponding author