Regulation of the chromatin environment is an essential process required for proper gene expression. Here, we describe a method for controlling gene expression through the recruitment of chromatin-modifying machinery in a gene-specific and reversible manner.
Regulation of chromatin compaction is an important process that governs gene expression in higher eukaryotes. Although chromatin compaction and gene expression regulation are commonly disrupted in many diseases, a locus-specific, endogenous, and reversible method to study and control these mechanisms of action has been lacking. To address this issue, we have developed and characterized novel gene-regulating bifunctional molecules. One component of the bifunctional molecule binds to a DNA-protein anchor so that it will be recruited to an allele-specific locus. The other component engages endogenous cellular chromatin-modifying machinery, recruiting these proteins to a gene of interest. These small molecules, called chemical epigenetic modifiers (CEMs), are capable of controlling gene expression and the chromatin environment in a dose-dependent and reversible manner. Here, we detail a CEM approach and its application to decrease gene expression and histone tail acetylation at a Green Fluorescent Protein (GFP) reporter located at the Oct4 locus in mouse embryonic stem cells (mESCs). We characterize the lead CEM (CEM23) using fluorescent microscopy, flow cytometry, and chromatin immunoprecipitation (ChIP), followed by a quantitative polymerase chain reaction (qPCR). While the power of this system is demonstrated at the Oct4 locus, conceptually, the CEM technology is modular and can be applied in other cell types and at other genomic loci.
Chromatin consists of DNA wrapped around histone octamer proteins that form the core nucleosome particle. Regulation of chromatin compaction is an essential mechanism for proper DNA repair, replication, and expression1,2,3. One way in which cells control the level of compaction is through the addition or removal of various post-translational histone tail modifications. Two such modifications include (1) lysine acetylation, which is most commonly associated with gene activation, and (2) lysine methylation, which can be associated with either gene activation or repression, depending on the amino acid context. The addition of the acetylation and methylation marks is catalyzed by histone acetyltransferases (HATs) and histone methyltransferases (HMTs), respectively, whereas the removal of the mark is done by histone deacetylases (HDACs) and histone demethylases (HDMs), respectively4,5. Although the existence of these proteins has been known for decades, many mechanisms of how chromatin-modifying machinery works to properly regulate gene expression remain to be defined. Since chromatin regulatory processes are dysregulated in many human diseases, new mechanistic insights could lead to future therapeutic applications.
The chromatin in vivo assay (CiA) is a recently described technique that uses a chemical inducer of proximity (CIP) to control chromatin-modifying machinery recruitment to a specific locus6. This technology has been used to study an expanding list of chromatin dynamics, including histone-modifying proteins, chromatin remodelers, and transcription factors6,7,8,9. CIP-based chromatin tethering of exogenously expressed proteins has also been extended past CiA to modulate non-modified genetic loci by use with a deactivated Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)-associated protein (dCas9)10,11. In the CiA system, mESCs have been modified to express a GFP reporter gene at the Oct4 locus, a highly expressed and strictly regulated area of the mESC genome. Previously, this system has been applied to describe the dose-dependent and reversible recruitment of exogenously expressed heterochromatin protein 1 (HP1), an enzyme that binds H3K9me3 and propagates repressive marks (e.g., H3K9me3) and DNA methylation6. To accomplish this type of recruitment strategy, the mESCs are infected with a plasmid that expresses an FK506-binding protein (FKBP) linked to a Gal4, which is a DNA-protein anchor that binds to a Gal4-binding array upstream of the GFP reporter. The cells are also infected with an FKBP-rapamycin-binding domain (FRB) connected to HP1. When the mESCs are exposed to low nanomolar concentrations of the CIP, rapamycin, both FRB and FKBP, are brought together at the target locus. Within days of rapamycin treatment, expression of the target gene is repressed, as evidenced by fluorescence microscopy and flow cytometry, and histone acetylation is decreased, shown by ChIP and bisulfite sequencing. While the development and validation of this approach have been significant advances in the field of chromatin research and regulation, one drawback is the required exogenous expression of chromatin-modifying machinery.
To make the technology based on recruitment of only endogenous physiologically-relevant enzymes and make a more modular system, we designed and characterized novel bifunctional molecules, termed CEMs (Figure 1)12. One component of the CEMs includes FK506 which, similar to rapamycin, binds tightly and specifically to FKBP. Thus, the CEMs will still be recruited to the Oct4 locus in the CiA mESCs. The other component of the CEMs is a moiety that binds endogenous chromatin-modifying machinery. In a pilot study, we tested CEMs that contain HDAC inhibitors. While the concept of using an inhibitor to recruit HDACs seems counter-intuitive, the inhibitor is nonetheless able to recruit HDAC activity to the gene of interest. This is accomplished by (1) redirecting the HDAC to the locus, releasing the enzyme, and increasing the density of un-inhibited HDACs in the area and (2) maintaining HDAC inhibition at the locus, but recruiting repressive complexes that bind to the inhibited HDACs, or (3) a combination of both. In a previous study, we showed that the CEMs were able to successfully repress the GFP reporter in a dose- and time-dependent manner, as well as in a manner that was rapidly reversible (i.e., within 24 hours)12. We characterized the ability of the CEM technology presented here to control gene expression using fluorescence microscopy, flow cytometry, and the ability to control the chromatin environment using ChIP-qPCR6. Here, we describe a method for using and characterizing the CEMs, which will facilitate the adaptation of this system to answer additional questions related to chromatin biology.
1) Cell Line Culture for Producing Lentivirus
2) Infection of Mouse Embryonic Stem Cells (mESC) with Lentivirus
3) Chemical Epigenetic Modifiers (CEM)s Preparation and Treatment
4) Analysis of the Expression by Microscopy and Flow Cytometry
5) Analysis of Chromatin by Chromatin Immunoprecipitation (ChIP) Followed by qPCR
We recently developed CEMs and demonstrated that this technology can be applied to regulate gene expression and the chromatin environment at a reporter locus in a dose-dependent and reversible manner. In Figure 1, a model of the lead CEM, CEM23, is shown. HDAC machinery is recruited to the reporter locus by the HDAC inhibitor which, in this case, is the GFP reporter inserted at the Oct4 locus.
We sought to characterize the CEM system at the Oct4 locus in CiA mESCs. Because the cells express GFP, it was possible to quickly visualize the expression of the reporter with fluorescence microscopy. The cells were imaged after 48 hours of treatment with 100 nM of CEM23, along with untreated cells. The phase images show healthy mESCs, which is important because unhealthy or differentiated mESCs would indicate a non-specific GFP repression. The fluorescence images show a bright GFP expression in the control cells and a reduced GFP expression in mESCs treated with CEM23. Representative images are displayed in Figure 2.
To quantify the changes in the GFP expression, flow cytometry was used. Again, the CiA mESCs were treated with 100 nM of CEM23 for 48 hours. Experimental cells treated with CEM23 and control cells were prepared for flow cytometry, using > 100,000 cells per sample. Consistently, we observed a > 30% decrease in GFP-expressing cells among those treated with CEMs. A representative histogram is shown in Figure 3.
Once evidence of CEM-induced GFP gene expression decrease was shown, we tested for changes in the chromatin environment using ChIP. One important factor for preparing samples for ChIP is the extent of chromatin sonication. To have the samples as consistent and properly sheared as possible, 10 million cells were sonicated for 3.5 minutes to obtain chromatin between 200 and 500 bp in size (Figure 4). Because we hypothesized that the repressive CEMs were binding to and recruiting HDAC proteins and repressive complexes to the reporter, we tested for changes in histone tail acetylation. Histone 3 Lysine 27 acetylation (H3K27ac) is commonly found along the transcriptional start site (TSS) of genes. We performed ChIP with an antibody for H3K27ac and then performed a qPCR with primer sets upstream and downstream of the TSS. The results show that a 48-hour treatment with 100 nM of CEM23 decreased the level of H3K27ac at the target locus when compared to control cells not treated with CEMs (p < 0.01, two-tailed Student's t-test with three biological replicates, Figure 5).
Figure 1: CEMs bind to the CiA:Oct4 locus through recruitment to the FKBP and tether endogenous epigenetic machinery. The model of the CEM system shows Gal4-FKBP as the protein-to-DNA anchor. The FK506 portion of the CEM binds to FKBP and the HDAC inhibitor recruits endogenous HDAC proteins to repress GFP. Please click here to view a larger version of this figure.
Figure 2: Fluorescence images of mouse embryonic stem cells (mESCs) with a CiA:Oct4-GFP reporter. Fluorescence microscopy images show a decrease in GFP expression upon CEM treatment. The top panel shows phase and fluorescent images of mESCs grown with untreated media. The bottom panel shows phase and fluorescent of the mESCs treated with 100 nM of CEM23 for 48 hours. The images adapted with permission from ACS Synthetic Biology12. Copyright (2018) American Chemical Society. Please click here to view a larger version of this figure.
Figure 3: Flow cytometry analysis quantitates a decreased expression of GFP in mESCs upon CEM23 treatment. mESCs without CEM23 (blue) were compared with mESCs treated with 100 nM of CEM23 for 48 hours (red). Please click here to view a larger version of this figure.
Figure 4: Sonication of chromatin is uniform, and a smear is visible between 200 and 500 bp. To perform ChIP-qPCR, the chromatin from mESCs was sonicated. Each sample consisted of approximately 10 million cells, which were sonicated for 3.5 minutes, to produce similarly-sheared chromatin samples between 200 and 500 bp in length. Please click here to view a larger version of this figure.
Figure 5: CEM23 treatment causes a decrease in H3K27ac at the CiA locus. ChIP-qPCR was performed to test for changes in the chromatin environment. After a 48-hour treatment with 100 nM of CEM23, a decrease in H3K27ac was observed at CiA:Oct4 (**p < 0.01, two-tailed Student's t-test with three biological replicates. The error bars represent the standard deviation). Please click here to view a larger version of this figure.
Here, we described the recently developed CEM system being applied to regulate gene expression and chromatin environment at a specific gene in a dose-dependent manner. We provide an accurate method to study the dynamics involved in regulating gene expression through the selective recruitment of specific endogenous chromatin regulatory proteins. This is a highly modular technology that can be applied to investigate how different protein- and chromatin-modifying complexes work in concert to properly regulate the chromatin environment, as well as to study how these processes are dysregulated, with the benefit of achieving gene specificity.
Here, three ways were demonstrated to visualize changes in the chromatin structure and gene expression with new technology. After perturbation of specific targeted loci with CEMs, real-time changes to gene expression could be analyzed by fluorescence microscopy. Then, changes in gene expression levels could be measured more quantitatively with flow cytometry at defined endpoints. Finally, changes to posttranslational epigenetic marks on chromatin were examined by ChIP; specifically, in this case, H3K27ac. We did not test HDAC recruitment with ChIP-qPCR because of the diversity of HDACs that act in concert at a single locus. It is likely that a heterogenous population of HDAC enzymes was recruited and it would, therefore, be technically difficult to test them all by ChIP. In a previously published work, we have demonstrated that the direct recruitment of HDAC3 by a GAL4 fusion caused similar activity on gene expression and chromatin modification by ChIP12. If a non-fluorescent reporter is being modulated, changes in gene expression can also be measured by (1) an RNA expression analysis with reverse transcriptase qPCR or (2) protein expression with western blotting, or imaging and conducting flow cytometry with fluorescent secondary antibodies.
In relation to current CIP-based techniques like the CiA system12, this CEM technology has the advantage of recruiting endogenous epigenetic modulators to the target gene. Redirecting the cell's own machinery creates a more physiologically relevant means of modulating expression. Exogenously expressing master enzymes, like HATs and transcription factors, might result in side effects from their increased expression, especially while not being actively recruited to the gene locus.
While this article shows the functionality of this novel system with CEMs composed of HDAC inhibitors, several other inhibitors or protein recruiters can be synthesized in place of the HDAC inhibitor. To achieve gene activation, HAT inhibitor- or HDMT inhibitor-based CEMs can be synthesized. To repress and then overexpress the same gene, it is possible to treat the cells with a repressive CEM, wash them with regular media, and then add an activating CEM. This would allow for a clean system to study the dynamics of recruiting repressive and activating complexes to the same genomic region. When designing future CEMs, several characteristics need to be considered, including cell permeability, inhibitor potency, structure, and inhibitory kinetics. The linker length between FK506 and the recruiter moiety will influence the permeability of the CEMs, as well as the position of the recruited protein. When comparing two CEMs that differed only in linker length, the CEM with the shorter linker was more effective12. While it has not been tested directly here, the higher effectiveness is a result of greater cell permeability. Choosing inhibitors with chemical structures amenable to modification without disrupting the portion that binds the active site of the target is also important. The potency and specificity were also considered by selecting inhibitors with a high potency and specificity for a given class of proteins. The kinetics of inhibition may influence the effectiveness of the CEMs. CEMs comprised of inhibitors with slow on-off rates tend to be less effective.
One constraint of the current CEM technology is the requirement of a Gal4-binding array at the target gene locus. To recruit CEMs to a region other than the Oct4 CiA locus, any of the hundreds of engineered Gal4 upstream activation sequence (UAS) lines available to the scientific community can be used. One way the system will be made more modular is by incorporating a deactivated Cas9 (dCas9) with chemical epigenetic enzyme recruitment10,11,15. This nuclease-dead protein from the CRISPR-Cas9 system will still be recruited to any region of the genome to which a guide RNA (gRNA) is designed, but it will not cut the DNA. Using a dCas9-FKBP fusion, the FKBP component will recruit the CEMs where the dCas9 is recruited, greatly increasing the ease and flexibility of potential target genes. Imagine using this technology to target disease-relevant genes as a potential therapeutic or a means by which to reversibly and temporally study disease mechanisms at the chromatin level.
The authors have nothing to disclose.
The authors would like to thank the members of the Hathaway and Jin laboratories for their helpful discussions. The authors also thank Dan Crona and Ian MacDonald for their critical reading of the manuscript. This work was supported in part by Grant R01GM118653 from the U.S. National Institutes of Health (to N.A.H.); and by Grants R01GM122749, R01CA218600, and R01HD088626 from the U.S. National Institutes of Health (to J.J.). This work was also supported by a tier 3 and a student grant from the UNC Eshelman Institute for Innovation (to N.A.H and A.M.C, respectively). Additional funding from a T-32 GM007092 (to A.M.C) supported this work. Flow cytometry data was obtained at the UNC Flow Cytometry Core Facility funded by a P30 CA016086 Cancer Center Core Support Grant to the UNC Lineberger Comprehensive Cancer Center.
DMEM | Corning | 10-013CV | |
NEAA | Gibco | 11140-050 | |
HEPES (for tissue culture) | Corning | 25-060-Cl | |
BME (2-Mercaptoethanol) | Gibco | 21985-023 | |
FBS | Atlantic Biologicals | S11550 | Individual lots tested for quality |
Covaris tubes | ThermoScientific | C4008-632R | |
Covaris caps | ThermoScientific | C4008-2A | |
PEI (polyethylenimine) | Polysciences | 23966-1 | |
Transfection media (Opti-mem) | Gibco | 31985070 | |
Virus centrifuge tubes | Beckman Coulter | 344058 | |
Virus filter membrane | Corning | 431220 | |
Polybrene | Santa Cruz Biotechnology | SC134220 | |
0.25 % Trypsin | Gibco | 25200-056 | with EDTA |
0.05 % Trypsin | Gibco | 25400-054 | with EDTA |
Puromycin | InvivoGen | ant-pr | |
Blasticidin | InvivoGen | ant-bl | |
SYBR Green Master Mix (asymmetrical cyanine dye) | Roche | 4913914001 | |
H3K27ac antibody | Abcam | ab4729 | |
(ChIP kit) ChIP-IT High Sensitivity Kit | Active Motif | 53040 | |
Sonicator | Covaris | E110 | |
Ultracentrifuge | Optima | XPN-80 | |
SW-32 centrifuge rotor | Beckman Coulter | 14U4354 | |
SW-32 Ti centrifuge buckets | Beckman Coulter | 130.2 | |
LentiX 293 Human Embryonic Kidney | Clontech | 632180 | |
PBS | Corning | 46-013-CM | |
BSA | Fisher | BP9700 | |
EGTA | Sigma | E3889 | |
EDTA | Fisher | S311-100 | |
NaCl | Fisher | S271-1 | |
Glycerol | Fisher | G33-1 | |
Tris | Fisher | BP152 | |
NP40 | Roche | 11754599001 | |
Triton | MP Biomedicals | 807426 | |
Glycine | Fisher | BP381-500 | |
TE buffer | Fisher | BP2474-1 | |
HEPES (for ChIP) | Fisher | BP310-100 | |
Gelatin | Sigma | G1890 | |
Flow cytometer, Attune NxT | Thermo Fisher | A24858 | |
FKBP-Gal4 plasmid | Addgene | 44245 | |
psPAX2 plasmid | Addgene | 12260 | |
pMD2.G plasmid | Addgene | 12259 | |
Nanodroplets | MegaShear | N/A | |
DNA Nanodrop | Thermo Fisher | ND1000 | |
Protease Inhibitors (Leupeptin) | Calbiochem/EMD | 108975 | |
Protease Inhibitors (Chymostatin) | Calbiochem/EMD | 230790 | |
Protease Inhibitors (Pepstatin) | Calbiochem/EMD | 516481 | |
CiA:Oct4 mESC | The kind gift of G. Crabtree | ||
Lif-1C-alpha – producing Cos cells | The kind gift of J. Wysocka |