ABSTRACT: Osteogenesis is a highly regulated developmental process and continues during the turnover and repair of mature bone. Runx2, the master regulator of osteoblastogenesis, directs a transcription program essential for bone formation through both genetic and epigenetic mechanisms. While individual Runx2 gene targets have been identified, further insights into the broad spectrum of Runx2 functions required for osteogenesis are needed. By performing genome-wide characterization of Runx2 binding at the three major stages of osteoblast differentiation: proliferation, matrix deposition and mineralization, we identified Runx2-dependent regulatory networks driving bone formation. Using chromatin immunoprecipitation followed by high-throughput sequencing (ChIP-Seq) over the course of these stages, we discovered close to 80,000 significantly enriched regions of Runx2 binding throughout the mouse genome. These binding events exhibited distinct patterns during osteogenesis, and were associated with proximal promoters as well as a large percentage of Runx2 occupancy in non-promoter regions: upstream, introns, exons, transcription termination site (TTS) regions, and intergenic regions. These peaks were partitioned into clusters that are associated with genes in complex biological processes that support bone formation. Using Affymetrix expression profiling of differentiating osteoblasts depleted of Runx2, we identified novel Runx2 targets including Ezh2, a critical epigenetic regulator; Crabp2, a retinoic acid signaling component; Adamts4 and Tnfrsf19, two remodelers of extracellular matrix. We demonstrated by luciferase assays that these novel biological targets are regulated by Runx2 occupancy at non-promoter regions. Our data establish that Runx2 interactions withM-BM- chromatin across the genome reveal novel genes, pathways and transcriptional mechanisms that contribute to the regulation of osteoblastogenesis. To identiy the genome-wide occupancy of Runx2, DNA bound by Runx2 at the prolieration, matrix deposition, and mineralization stages were recovered by Runx2 ChIP. Libraries of purified DNA were generated using Illumina SR adapters (Illumina) following manufacturerM-bM-^@M-^Ys manual, and were selected for the inserted fragments of 200 M-BM-1 50 bp, and sequenced 36 bases on an Illumina Genome Analyzer II. Base calls and sequence reads were generated by Illumina CASAVA software (version 1.6, Illumina). Two independent biological repeats of Runx2 ChIP-Seq libraries were prepared for each time point, and two input libraries were prepared with sonicated DNA from day 9 MC3T3-E1 cells. We pooled the reads from two biological replicates for peaking calling using MACS (version 1.4.1) with a stringent p value threshold (p < 1e-10) in contrast to input control, and used these peaks for further bioinformatic analyses. Each sample deposited here contains three files: the sequence file with short reads combined from two biological replicates, a Bed file with peaks called from the pooled short reads, and a Wig file with peak signals.
