Comparative Genomic Analysis of the 5′-Upstream Sequences of von Willebrand Factor and 27 Other Endothelial-Specific Genes Identifies Conserved Noncoding Sequences That Are DNase Hypersensitive.

Von Willebrand factor (VWF), a key blood clotting protein, is a standard histologic marker for endothelial origin. Previously published studies characterizing tissue-specific regulatory elements have concentrated on the proximal promoter region (between −2,645 and the end of murine Vwf intron 1). Transgenic mouse models using these sequences have found elements needed for VWF expression in some, but not all, populations of endothelial cells. To identify potential novel VWF distal regulatory elements, as well as sequences contributing to the more general endothelial-specific expression program, we aligned human and mouse orthologous genomic sequences from VWF and 27 other endothelial-specific genes, extending from the start of transcription to the neighboring 5′-upstream gene (up to 100 kb of sequence), excluding the core promoter, and including intron 1. We analyzed the resulting 262 endothelial cell-specific gene conserved noncoding sequences (EC-CNSs) using a specific quantitative PCR assay for DNase I hypersensitive sites, an indicator of an open chromatin state which suggests the presence of a transcription regulatory complex. 33% of the EC-CNSs (including 6 of 9 upstream of VWF) were hypersensitive (HS) in human umbilical vein endothelial cells (HUVECs), while only 16% were HS in control human fibroblast cells. 20% of the EC-CNSs were HS in HUVECs only, suggesting that they may identify novel endothelial-specific transcriptional elements. When the 262 EC-CNSs were compared to CNSs from non-endothelial genes, a single motif, AGGAA(A/G), was identified as overrepresented in EC-CNSs. This association was replicated in an independent set of 28 additional genes with high HUVEC expression. The motif was also overrepresented in the subset of HS EC-CNSs. The AGGAA(A/G) motif includes the core binding sequence [GGA(A/T)] of the Ets family of transcription factors, which contains several members implicated in tissue-specific gene expression. These results suggest that the AGGAA(A/G) motif may represent the target sequence for a specific Ets family member critical for determining the endothelial-specific gene expression program. We are currently testing this hypothesis by analyzing this motif in Vwf BAC transgenic mice containing 177 kb of upstream sequence, including all ten identified EC-CNSs. Additionally, we have expanded our analysis of the VWF genomic region to include over 700 kb of sequence surrounding and including the gene, searching for additional EC-CNSs conserved across eleven mammalian species. This comparative genomic approach, coupled with functional analysis of putative regulatory elements, should provide new insight into the endothelial cell gene expression program and the more general problem of tissue-specific gene regulation.