Sometimes an indirect approach can give a direct answer. For example, two recent papers use dCas9 based epigenome editing to cut through gene regulatory mechanisms. No cutting required. These epigenomic editing toolkits use single-cell readouts to provide a clearer story of how our favorite chromatin modifications and enhancers do their jobs.
Mixing and Matching Modifications with Modules
Determining whether modifications act directly or indirectly to affect transcription has been complicated. To get a better handle on cause-and-effect, as well as quantitative data, Jamie Hackett’s creative group at EMBL (Italy) developed a modular epigenome editing toolkit that lets you mix and match nine key modifications targeted to specific sequences in mouse embryonic stem cells. Because it uses just the catalytic domains of histone- or DNA-modifying enzymes, it avoids complex side effects and gets the real scoop on modifications in physiological context. Here’s a little bit about the platform:
- Effectors with the catalytic domains of modifying enzymes are linked to a superfolder GFP and a fragment that binds GCN4 sites
- dCas9 with a tail of GCN4 motifs binds the effectors, and the doxycycline-inducible system includes an enhanced guide RNA for targeting
With a single-cell readout system, modifications made by the constructs could directly and quantitatively turn transcription on or off. Next, the team turned their investigative reporting skills toward H3K4me3, and here’s what they found:
- Depletion of H3K4me3 at a set of promoters decreases expression at those locations, and also leads to the removal of H3K27ac and gain of H3K27me3
- Restoring H3K4me3 activity reactivates some of the genes, and targeting the mod to silenced genes can activate some of them
- H3K4me3 helps promoters become acetylated and removes silencing factors, like Polycomb
Genomic context is important for some modifications. For example, H3K36me3 was the strongest gene silencer in a reporter with CTCF motifs and was on constitutively. Programming H3K36me3 with the platform decreased H3K4me3 and increased DNA methylation, but it had no effect on a couple of other marks. Although the H3K36me3 effector could silence the endogenous Xist gene, it couldn’t silence it when the CTCF motif was removed.
Finally, the team combined various pairs of effectors and found some synergies, such as the H3K27me3 and H2AK119ub pair that significantly increases transcription more than either could alone.
Overall, the intrepid team reports that data resulting from their new toolkit suggests that context is important, and modifications can directly affect transcription.
Bi-directional Interactions
In another up-to-the-minute editing development, Howard Chang’s talented team (Howard Hughes Medical Institute, Stanford University) used dCas9-based tools to bidirectionally change expression levels at two loci simultaneously in opposite directions to investigate gene and enhancer interactions. They crafted CRISPRai, which allows activation or inhibition of genetic elements with a single-cell readout to be applied to human cells. Here’s some info on the new system:
- Doxycycline-inducible CRISPRa contains VPR fused to a dCas9 from S. aureus, whereas the CRISPRi version includes KRAB fused to an S. pyogenes dCas9
- Each construct pairs with different guide RNAs
- In addition, the team went beyond Perturb-Seq, which only analyzes single perturbations, to develop CRISPRai Perturb-seq that handles bidirectional changes
The team then applied the toolkit to living cells, such as K562, Jurkat cells primary T cells, and CAR-T cells. Some genes, like SPI1, respond more to activating constructs than constructs that repress transcription, so it becomes activated more easily. GATA1, however, is the opposite and responds more to repressing constructs, so it becomes repressed more easily. Here are a few other newsflashes:
- When perturbing the SPI1 and GATA1 transcription factors, some genes become synergistically regulated
– These genes have less SPI1 at promoters and enhancers but the same GATA1 at promoters compared to those genes with just additive effects - Other genes are much less expressed, or buffered, and have less of both factors at the promoter
- Testing perturbations of cis elements—transcription start sites (TSSs) and enhancers—showed that for IL2, a repressed promoter trumps activated enhancers
– But two very strong IL2 enhancers (E4 and E6) can actually overcome TSS perturbations; these are called “gatekeepers”
The system also showed how enhancers interact with each other, and further experiments demonstrated that some gatekeeper enhancers can affect chromatin accessibility and are in more contact with the transcription start site, but others are not.
Wrapping up, the soundbite here is that CRISPRai is useful for studying coding and noncoding elements in the context of endogenous loci.
Making the Most of Modifications
Both toolkits make the most of targeting specific loci to tweak and edit modifications and to determine their effects in endogenous or near-endogenous contexts. Senior author Jamie Hackett shares that it “unlocks the ability to program desired gene expression levels in a highly tunable manner.” In this way, the editing systems could someday help decipher how the epigenome contributes to health and disease states.
Get the full scoop on the modular toolkit at Nature Genetics, May 2024, and the CRISPRai system at Nature Biotechnology, May 2024.