Genome-wide analysis of NIPBL/cohesin binding in mouse and human cells: Implications for gene regulation and human disease
One of the most powerful tools to arrive in biology in the past decade is high-throughput sequencing, such as Illumina sequencing. These platforms have allowed an unparalleled look at protein-chromatin interactions, mRNA expression, chromatin topology, and a great deal more. Our lab has successfully used these tools to better understand the distribution of cohesin and Nipbl binding in mouse and human cells, with the aim of further clarifying how cohesin and Nipbl regulate gene expression and what genes they regulate. Moreover, we have been able to identify how this binding can change in disease, and correlate these changes with the corresponding gene expression changes taking place in vivo. Careful analysis of ChIP-seq data (chromatin immunoprecipitation coupled with sequencing) has indicated that cohesin binding decreases genome-wide in mouse embryonic fibroblasts (MEFs) derived from a mouse model for Cornelia de Lange Syndrome (CdLS). In fact, cohesin's role in gene activation is most susceptible to Nipbl haploinsufficiency. Moreover, we find that decreased cohesin binding is correlated with the gene expression changes taking place between wildtype and mutant MEFs. Enhancer-promoter interactions, one mechanism by which cohesin can regulate gene expression, are decreased in the mutant MEFs. Our studies have helped characterize how Nipbl haploinsufficiency affects cohesin binding, and suggest how this effect on cohesin binding can affect gene expression in the context of CdLS.
Based on the increased severity of the disease phenotype of CdLS patients with mutations in NIPBL versus mutations in the cohesin subunits, it was postulated that NIPBL might have cohesin-independent functions in the cell. To examine this, we have used ChIP-seq in human cells to identify the global distribution of NIPBL-chromatin interactions. We found that, similar to cohesin, NIPBL is enriched at the promoter region. In contrast to other studies however, we found that NIPBL is present at sites also bound by cohesin and CTCF. While most NIPBL -bound regions are shared with cohesin, about 10% of these sites are free of cohesin and CTCF. Further examination of the cohesin-free sites show 273 genes where NIPBL is bound at the promoter, and could be direct genes targets. Of these, 73 are differentially expressed upon siRNA depletion of NIPBL, two of which were examined in detail to show that NIPBL normally represses the expression of these genes independently of cohesin, which is the first time any ability of NIPBL to repress gene expression has been shown. Taken together, our data indicate that mutations in NIPBL may indeed effect expression of NIPBL target genes, suggesting that this may explain in part the differences in disease severity in CdLS patients.
Over the past decade, our lab has examined the importance of intact heterochromatin on chromosome 4q in FSHD muscular dystrophy. While we have shown that the loss of cohesin and H3K9me3 present at the D4Z4 repeat array on chromosome 4q is characteristic of the disease, and may underlie expression changes seen in patient myoblasts in vivo, the global differences in heterochromatin in FSHD have not been previously studied. Therefore, we have used several sequencing techniques to identify the genome-wide changes to heterochromatin and gene expression in FSHD, with the intent of being able to identify disease-specific signatures and illuminate the interdependence of the two in disease.