Tools are great for building, but sometimes it’s more fun using them to break things down. While we’ve applied deactivated Cas9 (dCas9) tools to build our knowledge of DNA methylation, this time it takes an entire toolbox to break the connection between histone methylation and gene repression.
This constructive deconstruction comes at you from the lab of David Segal at the University of California, Davis. The team bulldozed into the complexities of repressed chromatin by examining two marks enriched for at different features: H3K9me3, a mark of constitutive heterochromatin, and H3K27me3, a mark of facultative heterochromatin.
Designing Novel dCas9 Effectors
First, the team set out to create some mighty histone methylation writers. The group targeted their systems to the HER2 proto-oncogene promoter in HCT116 cells and primarily employed a combo of RT-qPCR and ChIP-qPCR to take tabs on the damage. Here’s what went down when they designed their toolkit:
They fused the catalytic SET domains of H3K9me3 writers, G9A, and SUV39H1 to dCas9 and observed that:
- The constructs repress transcription (~3-fold) only when fused to the N-terminal of dCas9
- Surprisingly, only G9A deposits H3K9me3 (~13-fold), while SUV39H1 produces none, and since both constructs repress expression at similar levels their findings suggest that H3K9me3 is not sufficient for gene inactivation
The team then crafted targetable versions of two effector domains: Ezh2, a writer of H3K27me3, and the N-terminus of Friend of GATA-1 (FOG1), which recruits repressive complexes to remove histone acetylation and write H3K27me3. They discovered that:
- The full length Ezh2 effector deposits H3K27me3 (~9-fold enrichment), while the effector with just the catalytic SET domain does not, but strikingly both repress expression (~2-fold) almost equally, suggesting that the deposition of H3K27me3 is not the key component to gene inactivation
- Unlike other dCas9 effectors, FOG1 prefers the C-terminus for repressing expression (~3-fold). It stacks up as the strongest all of gene repressors (~6-fold) when fused to both the N- and C-termini, leaving a ~6-fold enrichment for H3K27me3
The Sharpest Tools in the Box
With their designer constructs in hand, the talented team then benchmarked their tools, where:
- They expanded the repertoire by including DNMT3A and Krüppel-associated box (KRAB). Then, the team targeted the HER2, MYC, and EPCAM promoters in HCT116 and HEK293T cells, and found that repression can range from 0- to 10-fold
- KRAB and FOG1 are the most potent repressors in most cases
- They also tested out whether deactivated Cpf1 (dCpf1) could replace Cas9 as a scaffold for the effectors, which would open up new target sites for epigenome editing, but, alas, they found that it did not repress gene expression
- Finally, they tested out whether transcriptional repression persists after transient expression of the effectors and observed that a combination of FOG1 and KRAB is best for “transient but strong repression”, whereas the combination of Ezh2 and DNMT3A is suited for “persistent but more modest repression”
Overall, this paper provides some much-needed insight into the emerging field of epigenome editing by revealing that the repression of gene expression by an effector can occur without histone methylation. The team thinks this could be from either steric hindrance or interactions with other complexes. And on top of all that, they give you two new effectors for your toolbox.
First author Henriette O’Geen and senior author David Segal share, “Nature uses epigenetic information to establish persistent patterns of gene expression, and we can too. We have developed a dCas9-based toolbox that can deposit repressive histone marks at single genomic targets. Repressed chromatin states are commonly associated with histone marks H3K9me3 and/or H3K27me3 and/or DNA methylation. It was important to create epigenetic tools that can deposit these epigenetic marks, but also study their interplay. Using our epigenetic toolbox, we found that a combination of dCas9-DNMT3A and dCas9-Ezh2 can persistently silence transcription of the HER2 proto-oncogene.”
The duo concludes, “One of the more surprising findings from our study was that even when the tools correctly write the epigenetic information that is associated with gene silencing, the target genes are often not silenced. These observations led us to conclude that writing these so-called “repressive marks” are not sufficient to cause target gene repression. The reviewers found this to be somewhat of a heretical idea. However, we think it comes down to the fact that we know a lot about “active” and “repressed” epigenetic states, but we clearly need to know more about what is required to transition between these states.”
Go play with the toolbox over at Nucleic Acids Research, July 2017