Hacking the epigenome is no easy task. Each deactivated Cas9 (dCas9) system boasts its own nuances that must be teased apart before we can crack the histone code with its suite of effector domains. When it comes to histone acetylation, we’ve seen the histone acetyltransferase construct dCas9-p300, which writes H3K27ac to activate gene expression. However, every good writer requires an eraser, particularly since the epigenetic regulation of gene expression isn’t so ‘cut and dry’. For example, while the histone deacetylase 3 (HDAC3) erases H3K27ac to repress gene expression, it can also activate gene expression.
In order to develop and optimize an HDAC3 epigenome editing system, the lab of Zhaolan Zhou at the University of Pennsylvania turned to a mouse neuroblastoma cell line: Neuro-2a (N2a). To obtain their genome-wide perspective they employed RNA-seq as well as ChIP-seq, for H3K27ac and HDAC3, with gene-specific follow-ups by qRT-PCR and ChIP-qPCR. As controls for their editing system, they also utilized a catalytically deactivated HDAC3 in addition to a scrambled non-targeting sgRNA for both the catalytically active and deactivated constructs.
Based on their genome-wide scans, Kwon et al. set their sights on Smn1, Mecp2, and Isl1, due to differences in their gene expression levels (high, medium, and low, respectively). Here’s what they discovered:
- The location of endogenous histone acetylation has a big impact
- High enrichment levels prevent deacetylation, but so does being too far from the peaks
- Sites adjacent to the peaks represent the ‘goldilocks zone’ and in the case of Mecp2 a ~1kb deacetylation event occurs
- As observed for other effectors, targeting downstream of the transcription start site (TSS), with either the catalytically active or inactive dCas9-HDAC3 constructs, represses expression due to steric hindrance of CRISPR interference (CRISPRi)
- However, they observed no effect for Smn1, the gene with the highest expression, likely due to high levels of transcriptional machinery and/or the chromatin microenvironment
- More isn’t better, as multiplexing sgRNAs to saturate a locus, downstream of the TSS excluded, prevents their effect on transcription
The team also generated clonal cell lines expressing dCas9-HDAC3 and chose MC3T3-e1 pre-osteoblasts, since these cells express Mecp2 at higher levels than the N2a cells and have a different chromatin environment. They selected for clones containing similar levels of dCas9-HDAC3 and then selected for three levels of sgRNA expression (high, medium, and low) to examine the effects of sgRNA dosage. By using H3K27ac levels as a proxy, they found that there is a certain minimum level of sgRNA expression needed since the low levels of sgRNA didn’t have an effect. Furthermore, high levels of sgRNA had a stronger effect than medium levels. Strikingly, they established that chromatin context was everything, as expression Mecp2 increased in these cells. Overall, this work demonstrates the complexities of both epigenome editing and histone deacetylation.
Learn how to HDAC away at the histone code in Nature Communications, May 2017