Nature is a master of using specific tools for specific jobs, so it’s not too surprising that DNA methylation and SETDB1/H3K9 methylation work on different sets of genes, as University of British Columbia researchers report. But occasionally those tools are more like a Swiss Army knife, like SETDB1, which also methylates H3K9s near endogenous retroviruses (ERVs) and seems to protect mouse embryonic stem cells (mESCs) from expressing potentially harmful transcripts.
The crafty Canadian scientists conducted RNA-seq and native ChIP experiments showing that H3K9me3 acts independently of DNA methylation most of the time—a mere 7% of the genes upregulated in a SETDB1 knockout (KO) line also were upregulated in a cell line with DNMT3A, DMNT3B, and DNMT1 knocked out.
Oddly enough, only 13% of the promoters of the upregulated genes in the SETDB1 KO line also were marked with H3K9me3 in the wild-type line. And most of the promoters that lost H3K9 methylation when SETDB1 was knocked out were not upregulated. The story got even more complex when the researchers looked at ERVs. Here’s what they found:
- ERV subfamilies were almost exclusively highly derepressed in the SETDB1 KO versus in the DNMT KO.
- ~40% of SETDB1 binding sites just happen to be really close to ERVs in the mouse genome.
- Most of the class I and class II ERVs that were upregulated in the SETDB1 KO also lost H3K9me3.
- This upregulation of ERVs in the SETDB1 KO led to a lot more strange, “chimeric” transcripts where transcription incorrectly started in the ERV and continued through an exon.
“Our results suggest that transcriptional silencing of ERVs is critical not only for minimizing the deleterious effects associated with increased retrotransposition, but also for maintaining the integrity of the transcriptome, as numerous ERVs have the potential to act as promoters that encode chimeric transcripts, i.e. that splice to genic exons.” Explained Dr. Matthew Lorincz, one of the paper’s authors. The team says that although DNAm may keep ERVs in check in somatic cells, it seems that SET1DB silences them in mESCs. And previous research suggests that incorrect expression of ERVs can lead to cancerous transformation in mice and humans.
Find out more about nature’s tool kit at Cell Stem Cell, June 2011.