This method is designed to follow formation of PRC2-mediated chromatin domains in cell lines, and the method can be adapted to many other systems.
The organization and structure of chromatin domains are unique to individual cell lineages. Their misregulation might lead to a loss in cellular identity and/or disease. Despite tremendous efforts, our understanding of the formation and propagation of chromatin domains is still limited. Chromatin domains have been studied under steady-state conditions, which are not conducive to following the initial events during their establishment. Here, we present a method to inducibly reconstruct chromatin domains and follow their re-formation as a function of time. Although, first applied to the case of PRC2-mediated repressive chromatin domain formation, it could be easily adapted to other chromatin domains. The modification of and/or the combination of this method with genomics and imaging technologies will provide invaluable tools to study the establishment of chromatin domains in great detail. We believe that this method will revolutionize our understanding of how chromatin domains form and interact with each other.
Eukaryotic genomes are highly organized and changes in the chromatin accessibility directly controls gene transcription1. The genome contains distinct types of chromatin domains, which correlate with transcriptional activity and replication timing2,3. These chromatin domains range in size from a few kilobases (kb) to more than 100 kb and are characterized by an enrichment in distinct histone modifications4. The central questions are: how are these domains formed and how are they propagated?
One of the most well-characterized chromatin domains is fostered through the activity of the Polycomb repressive complex 2 (PRC2). PRC2 is a multi-subunit complex composed of a subset of the Polycomb Group (PcG) of proteins5,6, and catalyzes the mono-, di- and trimethylation of lysine 27 of histone H3 (H3K27me1/me2/me3)7,8,9,10. H3K27me2/me3 are associated with a repressive chromatin state, but the function of H3K27me1 is unclear6,11. One of the core components of PRC2, embryonic ectoderm development (EED), binds to the end product of PRC2 catalysis, H3K27me3, through its aromatic cage and this feature results in the allosteric stimulation of PRC212,13. The PRC2 enzymatic activity is crucial for preserving cellular identity during development as the inappropriate expression of certain developmental genes that are contraindicated for a specific lineage, would be detrimental5,6. Hence, unraveling the mechanisms by which PRC2 fosters the formation of repressive chromatin domains in mammals is of fundamental importance to understanding cellular identity.
All of the past experimental systems designed to investigate chromatin domain formation including PRC2-mediated chromatin domains, were performed under steady-state conditions, which are unable to track the unfolding events of chromatin domain formation in cells. Here, we present a detailed protocol to generate an inducible cellular system which monitors the initial recruitment and propagation of chromatin domains. Specifically, we focus on tracking the formation of PRC2-mediated repressive chromatin domains that comprise H3K27me2/3. This system that can capture the mechanistic details of chromatin domain formation, could be adapted to incorporate other chromatin domains, such as the widely studied domains comprising either H2AK119ub or H3K9me. In combination with genomics and imaging technologies, this approach has the potential to successfully address various, key questions in chromatin biology.
Generation of inducible EED rescue mESCs
1. Cell culture
2. Generation of clonal EED knockout (KO) mESCs
3. Engineering the EED knockout mESCs to harbor Cre-ERT2 based inducible EED expression
4. Following nucleation and spreading of PRC2 activity on chromatin
5. Monitoring emergence and growth of the H3K27me3 foci in the mESCs nuclei
A general scheme of the conditional rescue system
Figure 1 shows the targeting scheme to conditionally rescue EED KO cells with either WT or cage-mutant (Y365A) EED that is expressed from the endogenous EED locus. After knocking out EED, a core subunit of PRC2 that is essential for its stability and enzymatic activity, a cassette within the intron following exon 9 of EED is introduced (Figure 1). The cassette consists of the remaining 3' cDNA sequence of EED, in reverse orientation with respect to the endogenous gene sequence. The cassette is flanked by heterologous inverted loxP sites (lox66 and lox71)18. The cells are propagated until the complete loss of H3K27me2/3 is observed. Upon activation of Cre recombinase expression by the addition of 4-OHT, the cassette is inverted such that exon 9 is spliced into the cassette using the splice acceptor sequence (Figure 1). With this system, EED KO cells can now be rescued by either WT or cage-mutant versions of EED both of which have a C-terminal Flag-HA tag. The downstream T2A-GFP provides a marker to select for cells that undergo a successful inversion event. The polyA signal prevents transcription of downstream sequences. With this system, the kinetics of PRC2 recruitment and the formation of H3K27me domains can now be followed.
Tracking the temporal deposition of PRC2-mediated H3K27me2/3
To follow the temporal dynamics of PRC2-mediated chromatin domain establishment, the WT or cage-mutant version of EED is re-expressed in the background of EED KO cells, upon 4-OHT treatment (Figure 2A,B). To follow deposition of H3K27me2/3 marks, ChIP-seq of H3K27me2/me3 are performed after re-expression of WT or cage-mutant EED for the indicated time points. The emergence of H3K27me3 is observed at 12 h after WT EED expression at discrete regions that are denoted as "nucleation sites" (Figure 2C), and then at regions distant from the initial nucleation sites by 24 h13. Eventually, the distribution of H3K27me3 nearly approximates that of the levels seen in WT parental cells at 36 h after WT EED expression. The temporally established downstream sites of H3K27me3 are termed "spreading sites"13. This system can also track the temporal deposition of H3K27me2, which appears to precede H3K27me3 deposition (Figure 2C,D). Lastly, re-expression of cage-mutant EED exhibits different dynamics relative to WT EED (Figure 2C,D). In this case, deposition of H3K27me3 is very inefficient and instead, H3K27me2 becomes apparent and more concentrated at the nucleation sites, indicating that the cage-mutant EED is unable to spread the modification to neighboring regions.
The emergence of H3K27me3 foci and their growth can also be visualized by microscopy (Figure 2E)13. Before the induction of WT EED expression, H3K27me3 staining is not apparent. However, 12 h of WT EED expression shows evidence of H3K27me3 foci formation. These foci increase in number and size by 24 h, and eventually spread to large regions of the nucleus by 36 h of WT EED expression. This system gives evidence of the de novo formation of PRC2-mediated chromatin domains as monitored by ChIP-seq and immunofluorescence (Figure 2C-E)13.
Figure 1: Targeting scheme to conditionally rescue EED KO mESCs either with WT or cage-mt EED (Y365A). Deletion of exon 10 and 11 causes destabilization and degradation of EED and global loss of H3K27me2/me3. A cassette in the intron between exon 9 and 10 is inserted in the reverse direction as indicated. Addition of 4-OHT inverts the cassette such that endogenous exon 9 splices into the cage-mutant or wild type cDNA of the cassette allowing inducible re-expression of WT or cage-mutant EED. This figure has been modified from Oksuz et al., 201813. Please click here to view a larger version of this figure.
Figure 2: Validations and representative applications of inducible EED rescue systems (i-WT-r and i-MT-r). (A). i-WT-r and i-MT-r systems are validated by Western blot using indicated antibodies on whole extracts after 4-OHT treatment to induce expression of WT (i-WT-r) or cage-mutant (i-MT-r) EED, at the time point indicated. (B). Flow cytometry analysis of GFP in i-WT-r cells before (0 h) and after (36 h) 4-OHT treatment for confirming the efficiency of the expression of T2A-GFP, which is indicative of successful flip of the inverted cassette. (C, D). ChIP-seq tracks for H3K27me3 (C) and H3K27me2 (D) near the Emx1 gene in the WT, or in i-WT-r or in i-MT-r cells, for the indicated times of 4-OHT treatment. Early and delayed sites for deposition of the histone marks are indicated. (E). Immunofluorescence using H3K27me3 antibody at 0, 12, 24 and 36 h after rescue of WT EED expression in i-WT-r mESCs. Staining at 36 h is shown at lower exposures. The rightmost panel is a zoomed-in image of the cell labeled with a red arrow. This figure has been modified from Oksuz et al., 201813. Please click here to view a larger version of this figure.
A powerful approach towards understanding the mechanistic details during the formation of a given chromatin domain, is to first disrupt the domain and then track its reconstruction in progress within cells. The process can be paused at any time during the reconstruction to analyze in detail the events in progress. Previous studies on chromatin domains were unable to resolve such events as they were performed under steady-state conditions (e.g., comparing wild-type and gene knockout). Here, we outline a system to assess the dynamic formation of chromatin domains, as highlighted by the recruitment and spreading of PRC2-mediated repressive domains in cells.
The most critical step for this inducible system is the accurate design of DNA constructs to re-express the desired proteins (or RNA) after their respective deletion from the cells. Various mutations, tags, and fluorescent markers could be introduced within these constructs, depending on the downstream applications. For example, instead of using self-cleaving T2A peptide immediately before GFP to disconnect it from the protein of interest, various fluorescent fusion proteins could be generated for high resolution imaging and/or single-particle tracking experiments to monitor the initial events of chromatin domain formation in living cells.
Although this method could be modified in many ways to provide insights into the formation of chromatin domains, it does have some limitations. First, it requires a cell line that harbors the Cre-ERT2 gene. Second, optimizations are necessary to determine the time points after 4-OHT treatment. Third, this system is not reversible so as to monitor the events during deconstruction of a chromatin domain. Degron-based methods could be used as an alternative to the method described here26,27. These systems enable reversible and rapid degradation of proteins, but usually require costly molecules for protein destabilization, such as auxin or shield26,27. Furthermore, these degron-based methods have limitations. They could only be used to re-express the WT version of the protein after its degradation. In contrast, the system described herein allow conditional expression of the mutant version of the protein of interest in addition to its WT counterpart. Combining such a degron system with the method described here would be a powerful means to reversibly modulate formation of chromatin domains. An alternative method for following the formation of chromatin domains entails the use of specific inhibitors to deplete a defined chromatin modification from the chromatin and then washout the inhibitor to follow de novo deposition of the modification. For example, treatment with EZH2 inhibitor and subsequent washout is used for tracking the re-formation of H3K27me domains28. However, EZH2 inhibitor is not effective in completely depleting H3K27me, even 7 days after treatment. In this case, the presence of pre-existing H3K27me might recruit and activate PRC212, thereby complicating interpretation of the results. As well, the use of the inhibitor and washout strategy is limited due to the absence of specific and potent inhibitors for many chromatin domains.
In addition to the applicability of this method to cellular systems, it can be used to generate inducible mutations/tagging on desired proteins in animals. In some cases, mutating a protein or inserting a tag could be lethal during development. This method bypasses this lethality and provides a temporal and spatial control of the mutation/tagging of proteins by coupling with tissue specific Cre-ERT2 mouse strains. These types of experiments would be valuable to determine the effect of a mutation in a specific tissue and/or at a specific stage of development. Importantly, it allows for the isolation of cells that undergo a successful recombination via the reporter within the cassette. This facilitates biochemical analyses, such as affinity purification from specific tissues, to isolate the tissue-specific interactome for a given protein. This system could be geared to monitor dynamic changes in any given cellular process, and hence is not limited to tracking chromatin domain formation.
The authors have nothing to disclose.
We thank Drs. L. Vales, D. Ozata and H. Mou for revision of the manuscript. The D.R. Lab is supported by the Howard Hughes Medical Institute and the National Institutes of Health (R01CA199652 and R01NS100897).
(Z)-4-Hydroxytamoxifen (5 mg) | Sigma | H7904-5MG | For induction of EED expression |
16% Paraformaldehyde aqueous solution (10×10 ml) | Electron Microscopy Sciences | 15710 | For immunofluorescence |
2-mercaptoethanol | LifeTechnologies | 21985-023 | For mESCs culture |
2% Gelatin Solution | Sigma | G1393-100ml | For mESCs culture |
Accutase 500 ML | Innovative Cell Tech/FISHER | AT 104-500 | For mESCs culture |
Alexa Fluor 594 AffiniPure Donkey Anti-Rabbit IgG (H+L) | Jackson immunoresaerch | 711-585-152 | For immunofluorescence |
Aqua-Mount Mounting Medium | FISHER/VWR | 41799-008 | For immunofluorescence |
CHAMBER SLD TC PRMA 8-CHM 16 PK | Fisher Sci | 177445PK | For immunofluorescence |
DAPI (4',6-Diamidino-2-Phenylindole, Dihydrochloride) – 10 mg | Life Tech | D1306 | For immunofluorescence |
ERK inhibitor, PD0325901 | Stemgent | 04-0006 | For mESCs culture |
ESGRO Recombinant Mouse LIF Protein | Millipore/Fisher | ESG1107 | For mESCs culture |
FBS Stem Cell Qualified | Atlanta | S10250 | For mESCs culture |
Gibson Assembly Master Mix | NEB | E2611L | For Donor template cloning |
GSK3 inhibitor, CHIR99021 | Stemgent | 04-0004 | For mESCs culture |
H3K27me2 (D18C8) rabbit mAB | Cell Signaling | 9728S | Antibody for ChIP-seq |
H3K27me3 | Cell Signaling | 9733S | Antibody for ChIP-seq |
Histone H2Av antibody (pAb) | Active motif | 39715 | Spike-in control for ChIP-seq |
Knockout DMEM | Invitrogen | 10829-018 | For mESCs culture |
L-glutamine | Sigma | G7513 | For mESCs culture |
Lipofectamine 2000 | LifeTech | 11668019 | For transfection |
MangoTaq DNA Polymerase | Bioline | BIO-21079 | For Genotyping PCR |
Normal donkey serum (10 mL) | Jackson ImmunoResearch | 017-000-121 | For immunofluorescence |
Penicillin-Streptomycin | Sigma/Roche | P0781 | For mESCs culture |
pSpCas9(BB)-2A-GFP (PX458) | Addgene | 48138 | For gRNA cloning |
QuickExtract DNA Extraction Solution | Lucigen | QE0905T | For Genotyping PCR |
Triton X-100 | Sigma | T8787-250ML | |
Zero Blunt PCR Cloning Kit | Thermo Fisher | K270020 | For Donor template cloning |
Primers/gBlocks | |||
EED-KO-gRNA-1 | Sequence: ctctggctactgtcaactag. gRNAs pairs to knockout EED in C57BL/6 ESCs for i-WT-r and i-MT-r systems. | ||
EED-KO-gRNA-2 | Sequence: TAGGCTATGACGCAGCTCAG. gRNAs pairs to knockout EED in C57BL/6 ESCs for i-WT-r and i-MT-r systems. | ||
EED-gRNA-inducible | Sequence: atggcaccccgaaattagaa. gRNA and Donor to generate i-WT-r system in the EED-KO background. | ||
i-WT-r Donor | https://benchling.com/s/seq-l2LLlWNEnLrfGXcbdCxI. gRNA and Donor to generate i-WT-r system in the EED-KO background. | ||
EED-gRNA-inducible | Sequence: atggcaccccgaaattagaa. gRNA and Donor to generate i-WT-r system in the EED-KO background. | ||
i-MT-r Donor | https://benchling.com/s/seq-n8eiZCB2XAkOuzzpv6qM. gRNA and Donor to generate i-MT-r system in the EED-KO background. | ||
Genotyping Primers | |||
Gnt-EED-KO-FW-1 | Sequence: ctgtaggctgccatctgtga. Wild type allele will produce a product of 1.9 kb. Knockout allele will produce a product of 200 bp. | ||
Gnt-EED-KO-REV-1 | Sequence: agccagggctacacagagaa. Wild type allele will produce a product of 1.9 kb. Knockout allele will produce a product of 200 bp. | ||
Inducible_Genotype-FW-1 | Sequence: tgcagtgaaacaaatttggaa. When the casette is inserted, the primers will produce 1863 bp. The wild type allele will produce a product of ~200 bp. | ||
Inducible_Genotype-REV-1 | Sequence: gagaggggtggcactgtaaa. When the casette is inserted, the primers will produce 1863 bp. The wild type allele will produce a product of ~200 bp. | ||
Inducible_Genotype-FW-2 | Sequence: ccccctctttctccttttct. When the casette is inserted, the primers will produce 3200 bp. The wild type allele will produce a product of 1560 bp. | ||
Inducible_Genotype-REV-2 | Sequence: atgcctgggtgaatgaaaaa. When the casette is inserted, the primers will produce 3200 bp. The wild type allele will produce a product of 1560 bp. |