The art of chromosome dynamics: an interview with Jane Skok

Abstract

In this interview, Professor Jane Skok speaks with Storm Johnson, commissioning editor for Epigenomics, on her work to date in the field of chromosome architecture and regulatory elements.

Jane Skok’s lab uses sophisticated microscopic techniques to visualize recombination in individual cells, tracing the dynamic changes in chromosome architecture and nuclear location at different stages of this complex process. This line of research unites two lifelong passions: science and art. After completing her PhD in immunology and genetics at the Imperial Cancer Research Fund in Lincoln’s Inn Fields, Dr Skok took 12 years off and pursued training in art while caring for her young children. She then returned to science, joining David Gray’s lab at Imperial College London as a postdoctoral fellow to study B cell biology and acquired expertise in Mandy Fisher’s lab to understand how nuclear organization of the antigen receptor genes regulate V(D)J recombination and allelic exclusion. Dr Skok continued to pursue these questions in her own lab at University College London and elucidated the roles of Pax5, locus contraction and nuclear subcompartmentalization in maintaining allelic exclusion. In 2006, Dr Skok was recruited to New York University School of Medicine, where her lab has revealed the activities of several signaling factors in guiding B cell development and they made the surprising discovery that the RAG proteins and the DNA damage response factor ATM help ensure allelic exclusion at the immunoglobulin gene loci. More recently, those at the Skok lab have turned their attention to understanding how localized and long-range chromatin contacts impact gene regulation in health and disease settings.

Your lab combines expertise in both the experimental and analytical aspects of chromosome folding. Why is this ethos important to you?

My lab pioneered the use of DNA FISH to investigate how the localization of genes relative to each other and to different compartments in the nucleus is linked to their activation status. In the early days, DNA FISH was limited in that it was only possible to simultaneously analyze a few genes at a time, and this raised the question about what was happening to the rest of the genome. To investigate, we realized it was going to be important to combine single-cell DNA FISH with genome-wide molecular chromosome conformation capture (3C) approaches, which had been pioneered by other labs. However, we very quickly became frustrated by the analysis of the data sets we obtained. In particular, we were not sure how to determine which interactions were significant within a genome and, even more difficult, how to robustly compare interactions between conditions. It was at this point that I began to appreciate how important the computational component of our future studies was going to be. Initially I tried to get help from the Core at NYU set up for this purpose, but I soon realized this model would not work for me, as I needed people who were invested in the biology and interested enough to develop pipelines to answer questions of interest. So I joined forces with computational biology labs with overlapping interests, and from that time on I have co-mentored a number of computational students and postdocs with PIs like Richard Bonneau (NYU and the Simons Foundation) and Aris Tsirigos (NYU), with whom I maintain long-term collaborations. Having computational people sitting in the lab and having the input from mentors that have the appropriate computational expertise has made a huge difference to our ability to interpret genome-wide data.

You have pioneered several sophisticated imaging techniques. What inspired you to enhance these technologies?

Everyone wants improved technologies to answer their question of interest, but how far you can go in developing these will depend on who is in your lab. For example, custom-made oligonucleotides made it possible to ask about particular aspects of genome organization, such as the relationship between active and inactive genes within a chromosome territory or the spatial organization of chromatin with particular features. At the time that this technology was emerging, I was lucky enough to have Julie Chaumeil in my lab as a postdoc. Trained in Edith Heard’s lab, she is an expert in FISH and imaging, which allowed us to ask a lot of questions that would not have been possible without her expertise [1–3]. Similarly, I have always thought that live imaging was going to be an important aspect of our understanding of gene regulation, but I did not pursue this until a very talented master's student, Yi Fu, who was interested in developing a dual-colored, live imaging system, joined my lab. She succeeded in publishing a first author paper [4], which raised the bar for any master’s students that came after her. If you have people with the motivation, passion and expertise, you can achieve a lot. This is definitely what makes science fun and exciting, as it can take your lab in new directions. I don’t want my lab to be all about my ideas – I enjoy getting input from lab members, and I want the team to develop ideas of their own and have the opportunity to explore these ideas, even if they don’t always work out.

In 2019, your group published the paper ‘EpiMethylTag: simultaneous detection of ATAC-seq or ChIP-seq signals with DNA methylation’ [5]. How has this method advanced epigenetic research?

This was another technique that stemmed from an idea of a former postdoc of mine, Priscillia Lhoumaud. I really liked the concept of being able to determine whether an ATAC-seq signal or a CHIP-seq signal comes from the same DNA molecule that is enriched for DNA methylation. If you do genome-wide studies, most of the data is representative of a population of cells and you never know what this means in terms of signal. For example, does a signal come from all cells in the population or a subset? This is a problem because even cultured cells can be heterogeneous and diseases like cancer even more so. The issue becomes even more complicated if you try to correlate enrichment of two factors when the data is derived from two separate experiments. And to get even more granular, you have to think of what is happening at the individual allele level. EpiMethylTag solves part of the problem, but the drawback is that it relies on a locus to be methylated in order to get the second signal. Inspired by the ideas of a senior research associate in my lab, Catherine Do, we are beginning to study the functional effects of genetic variants by analyzing allele specific chromatin interactions. Phasing of the data is greatly helped by long read sequencing, but this hasn’t yet gotten to prime time, so there is currently not much data available.

In your 2018 Review ‘Enhancer talk’, you provide several recommendations for future study on individual enhancer research [6]. Since then, has there been any progress in this field?

Identifying regulatory elements and connecting them to their target genes is a problem that is of great interest to me and many others in the community. As pointed out in our review, deciphering the contributions of multiple regulatory components to stage- and cell type-specific gene regulation is not an easy puzzle to solve, and there is no one rule for all. Different labs are tackling this in different ways. For example, the Mundlos lab models genetic alterations in patients in a mouse setting to identify the different elements that contribute to gene regulation. This is a powerful approach to understanding normal limb development and for teasing apart how genetic alterations and rearrangements contribute to different disease outcomes.

One particularly novel approach is synthetic regulatory dissection, a method pioneered by Jef Boeke (NYU) (ref on biorxiv). This involves the de novo synthesis of variant loci extending around 150 kilobases in length that encode different combinations of the regulatory elements that potentially control a particular gene or cluster of genes. This approach enables the rapid analysis of multiple different constructs in which regulatory elements can be deleted, moved or mutated – a process that would be far more drawn out with CRISPR targeting.

In our lab, we have taken a more global approach to match regulatory elements with their target genes. A computational postdoc, Theo Sakellaropoulos, who is co-mentored by myself and Aris Tsirigos, developed a computational pipeline called METHc (methyl connections) that couples DNA methylation changes to alterations in target gene expression in tumor and normal samples across multiple cancers, linked to patient outcome using TCGA data. METHc allows us to identify putative cis regulatory elements controlling one or more target gene and vice versa in the context of their topological interactions, chromatin accessibility and contribution of transcription factor (TF) binding, We hope to be able to publish this year.

I am also particularly intrigued by the role of transposeable elements (TEs) in gene regulation, as these harbor binding sites for transcription factors and chromatin architectural proteins such as CTCF. These are generally ignored in genome-wide analyses because of the repetitive nature of their sequences. However, TEs comprise around half our genome and have been important in spreading TF binding sites across the genome, and as a result they play an important role in gene regulation and transcriptional networks. A computational PhD student, Ramya Raviram, co-supervised by myself and Rich Bonneau, became interested in this issue and in particular whether TEs participate in chromatin interactions that could be important in gene regulation [7]. This inspired her to adapt existing 3C methodologies to the analysis of chromatin contacts involving TEs.

It is clear that 3D chromatin organization contributes to gene regulation. In this regard, the architectural proteins CTCF and cohesin play an important role. In the lab, we are now studying the contribution of cancer-associated CTCF point mutations to tumorigenesis, analyzing their impact on the interaction of the protein with DNA, binding stability, binding profiles, cohesin, nuclear organization and gene regulation. Since CTCF is a zinc finger protein, analyzing mutations in functionally distinct residues of the zinc finger provides general insight into zinc finger proteins as a whole, and these are important because they make up about 50% of TFs. I really became intrigued with zinc finger proteins through a collaboration with a zinc finger expert, Marcus Noyes, who is at NYU. He gave me a lot of insight into the importance of functionally distinct residues and how they might influence DNA binding.

What advice would you give to women in science, and how can organizations better support the motherhood of researchers?

I do not think of myself as a woman in science but rather as a person in science. I never wanted to be given any special treatment ‘just because I am a woman’ – that would be demeaning. For me it is all about being the best person for the job. Women don’t need to be positively discriminated – we just need to make sure we are not negatively discriminated.

I may have been lucky in that I never felt I was unfairly treated by any organization I was part of and was given the same opportunities as everyone else. However, working alongside men is not always easy, as even those who think they are enlightened have a tendency to treat you like their assistants and you get handed jobs they find tedious and at the end they often take credit for your work. It is also not uncommon to be treated like you are invisible or have men talk over you. In these circumstances, it is important to try to make yourself heard.

My advice to women in science is – don’t give up, and never take no for an answer. This advice could equally well be given to any person in science, but it is my perception that women are generally not as self-confident as men, and this can be a handicap to their careers. I would also encourage women not to shy away from leadership positions. I agreed to be chair on my NIH study section – not because this was a job I particularly aspired to but because I had never sat on a study section where a woman was chair. So I thought it was time for me to do my bit to level the playing field.

That said, there is no level playing field with men, starting with the fact that we are the ones who give birth – that means it is already much easier for a man to have a family than a woman. Men don’t have to spend 9 months incubating their offspring, dealing with morning sickness and all the other niceties of pregnancy, including having to take time out of their work day for regular health checks. And after the children are born, women are generally the primary caretakers, which means women have to compete with men and take care of their children at the same time. Sometimes child care interferes with work to such a degree that there is no alternative other than to take a career break – as in my case, where I had a sick child who needed to be taken care of for a prolonged period of time. To succeed, you have to try even harder once you get back in the race, but it can be done.

Organizations can do much to support women with young families, first by providing affordable day care for those at every stage in their career. They can also promote more eligible women into leadership positions. This is definitely the best way to alter existing patterns of behavioral bias that women are frequently subjected to.

What are the future directions for your work?

The lab, as always, is focused on chromatin organization and its impact on gene regulation. As you might have gathered from the things I have talked about, we have switched from studying programmed recombination to cancer. Our interest in the architectural protein CTCF extends to its paralog CTCFL, also known as BORIS. CTCFL is another interesting protein that in normal circumstances is only transiently expressed in premeiotic male germ cells, where it plays a unique role in spermatogenesis by regulating expression of pluripotency and testis-specific genes. CTCFL is also implicated in numerous cancers, where its aberrant expression promotes cancer cell survival through mechanisms that are poorly understood, and it has been linked with advanced stage and poor prognosis in several cancers.

CTCFL has a very similar zinc finger domain to CTCF, but its N and C terminal regions are very different from those of CTCF, and we recently showed that as a result, CTCFL cannot interact with cohesin and thus does not have an insulator function. However, given the similarity in the DNA-binding zinc finger domain, CTCFL competes with CTCF for a third of its binding sites, and thus through binding at these sites and eviction of CTCF it can alter chromatin structure and gene regulation [8].

In contrast to mice, which express only one full-length isoform of CTCFL, the human gene is transcribed from three gene promoters and the transcribed RNA is alternatively spliced into 23 isoforms, resulting in the expression of 17 unique proteins. Although CTCFL has been studied in cancer, no one has paid attention to which isoforms are expressed in different cancer types and what functional impact each of these isoforms has. This is something we are currently investigating.

Aside from the areas I have already talked about, I have recently become interested in the mechanisms underlying the emergence of drug-tolerant persisters (DTPs), and Ben Neel and I just got awarded an R01 grant to study this in HER2+ breast cancer. This is a new and exciting area of research for the lab.