Epigenetics in a Dish | The Scientist Magazine®

February TSS Cell Culture

Epigenetics in a Dish



Intro music…

Deanna MacNeil: Welcome to The Scientist Speaks, a podcast produced by The Scientist’s Creative Services Team. Our podcast is by scientists and for scientists. Once a month, we bring you the stories behind news-worthy molecular biology research.


This episode is brought to you by Cytosurge, Molecular Devices, and Eppendorf. Cytosurge is proud to offer the FluidFM Biopsy Solution for live-cell sequencing through single-cell biopsies, where researchers extract cytoplasmic material from a cell without compromising its viability. Because single-cell biopsies provide a snapshot representation of a living cell’s transcriptome at a given point in time, the FluidFM Biopsy Solution enables the correlation between the transcriptome and the cell’s phenotype without removing it from its culture environment. Molecular Devices is one of the world’s leading providers of high-performance bioanalytical measurement systems, software, and consumables for life science research, and pharmaceutical and biotherapeutic development. Included within a broad product portfolio are platforms for high-throughput screening, genomic and cellular analysis, colony selection and microplate detection. These leading-edge products enable scientists to improve productivity and effectiveness, ultimately accelerating research and the discovery of new therapeutics. Eppendorf is a leading life science company that develops and sells instruments, consumables, and services for liquid, sample, and cell handling. Their product range includes pipettes and automated pipetting systems, centrifuges, mixers, spectrometers, thermal cyclers, ultra-low temperature freezers, fermenters, bioreactors, CO2 incubators, shakers, cell manipulation systems, and all accompanying consumables.


Every cell within the human body contains the same DNA, but not all cells look and act alike. The key to cellular diversity lies in which genes the cells express or shut down. Cells convey this information to the appropriate machinery through epigenetic modifications. In this episode, Charlene Lancaster from The Scientist spoke with Jonathan Weissman from the Massachusetts Institute of Technology and Luke Gilbert from the University of California, San Francisco to learn about making epigenetic changes in vitro and the application of these tools in research and the clinic.

Transition music…

Introduction to Epigenetics and Preliminary Editing Tools


Narrator (Charlene Lancaster): Cells often turn gene expression on and off through epigenetic marks. This information is not encoded by the DNA sequence, but instead, epigenetic tags alter a gene’s accessibility to transcriptional machinery. The genes that cells express dictate their phenotypic traits, such as cell morphology, motility, and adhesion properties. Ultimately, this mechanism prevents cells from expressing genes that are not important to them, such as muscle cells expressing genes important for antibody production. While there is some contention within the field as to the true definition of epigenetics, many researchers only consider changes that are heritable, or stably passed down to daughter cells after cell division, as epigenetic modifications. The most studied epigenetic modification is DNA methylation, where DNA methyltransferases deposit covalent marks onto CpG dinucleotides within the genome. Additionally, some scientists consider histone modifications as epigenetic changes, while others question whether these marks are truly heritable or simply instructed by transcription factors and other machinery at each cell division. Despite the importance of epigenetics to all life on earth, scientists including Luke Gilbert have only been able to epigenetically modify the expression of specific genes in cell culture relatively recently.

Luke Gilbert 

People have been epigenetically editing cells in very blunt force ways for a long time. There are very old compounds that were discovered in the 60s and 70s that can globally inhibit DNA methyltransferases and people use these as research tools for a long time to study epigenetic processes. 

The field of precision epigenetic editing only goes back to about 2010, roughly when people started to use programmable DNA binding proteins fused to enzymes that either deposit or remove DNA methylation. 

The Development of CRISPRi and CRISPRa

Narrator: The epigenetic editing toolbox was forever changed through the discovery of the CRISPR-Cas9 molecular scissors. Used by bacteria as a defense mechanism to cleave foreign DNA, researchers including Jennifer Doudna and Stanley Qi reprogrammed the scissors to precisely cut mammalian DNA. During this time, Gilbert was a postdoctoral fellow in Jonathan Weissman’s laboratory at the University of California, San Francisco and collaborations between the University of California, Berkeley and UCSF gave Weissman a front-row seat to learn about this revolutionary finding.

Jonathan Weissman 

I’ve known Jennifer Doudna for many years. Somewhere in late summer of 2012, Jennifer came into my office with Luke and had some evidence that they could edit DNA in mammalian cells. That was exciting but seemed to be already in good hands. 

Narrator: While developing the CRISPR-Cas9 system for use in mammalian cell culture, Doudna and her team generated a dead Cas9 enzyme or dCas9. This version of the protein was still able to form a complex with single guide RNA to bind target genes but was catalytically inactive. Instead, these broken scissors could serve as programmable DNA-binding proteins in vitro and in vivo.

Jonathan Weissman 

This opened up the floodgates for being able to bring in all of these tools we knew for making epigenetic modifications and to do this in a broadly programmable way. The first version of this was CRISPRi, which allowed you to silence nearby genes. That was very quickly followed by CRISPR activation, which let you turn on genes in a programmable way. 



Luke Gilbert

At the core of these technologies are dead Cas9 or deactivated Cas9 fused to additional protein domains via linkers. With CRISPRi or CRISPR interference, the dCas9 protein is fused to a KRAB domain by a long linker. The KRAB domain recruits human proteins, including KAP1, HP1, and SETDB1, to deposit repressive histone marks, such as H3K9 trimethylation, locally around the binding site for dCas9 and that’s correlated with transcription being shut off. 

CRISPRa or CRISPR activation is similar. It’s again a dCas9 fusion protein. There are many more flavors of CRISPR activation than there are of CRISPR inhibition. Most include a protein that’s fused to one or more acidic activation domains. If you space that appropriately to a transcription start site, you’ll recruit mediator, an RNA polymerase, and other proteins and enzymes that locally change epigenetic marks like H3K27 acetylation and those are associated with gene activation.

Narrator: Before the development of CRISPR activation and CRISPR interference, which is also known as CRISPR inhibition, by Weissman, Gilbert, and their team, researchers relied on engineered zinc finger proteins to bind DNA and epigenetically regulate a single gene’s expression in vitro and in vivo. However, developing these proteins was a laborious process. With the advent of CRISPRi and CRISPRa, as well as the relative simplicity of generating single guide RNA libraries, scientists could now easily control all 20,000 protein-coding genes within the human genome. Researchers use both traditional CRISPR-Cas9 and CRISPRi screens to systematically diminish the expression of genes one by one and assess which genes within the genome result in a change to the cell’s phenotype, including its morphology, behavior, and physiology. However, their individual properties give the techniques distinct advantages and disadvantages over one another when it comes to their use in research and their therapeutic potential.

Luke Gilbert 

With Cas9-mediated screens, you’re making breaks in protein-coding exons and then allowing the cell to repair the DNA. You rely on the cell making errors that result in insertions and deletions that render the protein-coding exons out of frame with a stop codon and that’s what we call a knockout. With CRISPR interference, you’re not even producing an RNA. You don’t have any mutations to the genome. 

Jonathan Weissman 

A fundamental difference between CRISPRi and CRISPR cutting when it comes to therapeutics is that CRISPR cutting was once and done. You transiently express the CRISPR cutting and make a permanent change that would then be heritable. CRISPRi as a therapeutic was much less attractive because it required the continuous expression of the dead Cas9. That really limited CRISPRi as a therapeutic. And this is really where CRISPRoff comes in. 




The Development of CRISPRoff and CRISPRon

Narrator: Building on the knowledge they gained from developing the CRISPRi and CRISPRa transcriptional editors, Gilbert, Weissman, and their colleagues developed two new programmable dCas9-based gene editing systems called CRISPRoff and CRISPRon. Unlike CRISPRi and CRISPRa, which only transiently control transcription and do not typically modify DNA methylation, CRISPRoff and CRISPRon stably modulate gene expression by directly editing DNA methylation. 


Luke Gilbert 

CRISPRoff is fused to a de novo DNA methyltransferase, DNMT3A, and CRISPRon is fused to a TET1 enzyme, which is a DNA demethylase. So, these are fusion proteins that locally edit DNA methylation and everybody in the field agrees that DNA methylation is a real epigenetic mark that’s heritable and meets all the criteria for what people consider epigenetics. 


Jonathan Weissman

CRISPRoff marries the once-and-done and the heritability of modifying DNA because we’re making a true epigenetic, heritable change. It does this in a way that doesn’t do any permanent changes to the DNA. If your goal is to prevent a gene from being expressed, I would argue that it’s a more perfect solution. There’s no RNA to instruct the production of the protein. So really, completely silences it as opposed to the CRISPR cutting and base editing approaches, which leave a damaged RNA that can produce a damaged protein.


Narrator: While CRISPRoff addressed many of the concerns researchers had with CRISPR cutting and CRISPRi, it was unknown if this tool would be capable of silencing every gene. Scientists had examined how DNA methylation controls gene expression for several decades before CRISPRoff’s development and had already determined that they could sort genes into two categories. The first set contains genes with CpG dinucleotide islands in their promoters, which are sites in the genome that are methylated by de novo methyltransferases. Researchers think that cells can silence genes containing these islands, which account for approximately 70% of human genes, through DNA methylation. With 30% of human genes having low levels of CpG dinucleotides around their promoters or transcriptional start sites, scientists do not think that genes within this category are controllable through DNA methylation. But Weissman and Gilberts’s colleague, James Nuñez, challenged this hypothesis.

Luke Gilbert

James did a great experiment where he compared CRISPRoff to a CRISPRoff mutant that had just a single amino acid change in its DNA methyltransferase enzyme core and then targeted all protein-coding genes and just asked the question, which genes can or can’t be turned off by this CRISPRoff construct versus this mutant construct. He saw gratifyingly that we could turn off expression of genes that did have CpG islands at their promoters, but what he saw to our surprise was that you could also turn off expression of genes that did not have CpG islands at their promoters. We followed that up with more detailed, careful experiments to show that methylation was deposited at these genes that don’t have CpG islands, maintained over many cell divisions, and was associated with gene expression being silenced. Those experiments certainly showed that it is possible for DNA methylation to control gene expression for genes that do not have CpG islands.

Jonathan Weissman

There still are sets of genes that are very hard to shut off by methylation. It’s just a minority. Broadly speaking, they do tend to be the ones that have really low CpG levels. But I like to think that the glass is about four-fifths full.


Applications of the Epigenetic Editing Tools


Narrator: With the inclusion of these new programmable genetic editors within the scientific community’s toolbox, researchers are not only able to learn more about epigenetic modifications and their heritability but are now more capable of dissecting molecular pathways important for key biological processes, such as differentiation and development. For instance, scientists have recently performed genome-wide CRISPRi screening on a neutrophil-like cell line. By repressing gene transcription and looking for defects in proliferation, differentiation, and migration, they determined which genes are important for these processes in neutrophils. Researchers are also using these tools to develop better epigenetic editors and potential therapeutics.

Luke Gilbert

My lab and many other labs, including Jonathan’s, have used CRISPRi and [CRISPR]a as a genome-scale screening tool. For example, my lab does a fair amount with CRISPRi and [CRISPR]a genome-scale screens looking for new drug targets, primarily in the context of oncology applications. We also use CRISPRi and [CRISPR]a genome-scale screens to understand new drugs, either after they’re in the clinic or before they go into patients. 

Jonathan Weissman

We’re working with a group at the Broad to be able to epigenetically silence genes in the central nervous system. We developed a new class of epigenetic silencers that recruits the endogenous methyltransferase rather than having to overexpress the fusion. You eliminate any of the toxicity associated with overexpressing a methyltransferase and we’ve gotten very effective and robust silencing in mouse CNS. So, we’re now going after some of the therapeutically interesting disease-causing proteins as a way of treating types of neurodegeneration.

From a therapeutic perspective, the dream of programmable medicine is that we will be able to understand the molecular basis of a disease and in a programmable way immediately make a drug. To do this, you need to have ways of turning on or off or changing genes. Those tools aren’t perfect yet, but they’re really very good between CRISPRi, CRISPRa, CRISPRoff, CRISPR cutting, base editing, and prime editing, among others. We really have a lot of arrows in the quiver of what we can do. The second is we have to deliver this genome editor. So, there’s a lot to be done, but a lot of progress being made. In the end, though, the real limitation is going to be what changes do we need to do to affect the disease and that’s where the functional genomics is going to come in. It’s going to reveal which genes or which combinations of genes we want to turn on and off that will meaningfully impact a broad range of disease processes.



Outro music…


Deanna MacNeil: Thank you for listening to The Scientist Speaks. This episode was produced by the Creative Services Team for The Scientist and narrated by Charlene Lancaster. Thanks again to Cytosurge, Molecular Devices, and Eppendorf for sponsoring this episode. Please join us again in March, as we learn about the role of endogenously produced hallucinogenic compounds in human physiology. To keep up to date with this podcast, follow The Scientist on social media and subscribe to The Scientist Speaks wherever you get your podcasts.

Leave a Reply

Your email address will not be published. Required fields are marked *