Covering Your Bases

If genetics deals with the information that cells need to operate, epigenetics involves the information packaging. Both of these systems play an essential role in cell behavior—a breakdown in either may tip a cell toward disease. Illuminating how epigenetic chemical decorations are added is one of the most active areas of biomedical research.

One such epigenetic modification is DNA methylation, the addition of methyl groups to cytosine (C) bases along the DNA strand. This chemical mark predominates at Cs that are followed by the base guanine (G). This pattern of methylation is thought to be important for silencing gene expression, which fundamentally affects the character of the cell. Like all epigenetic changes, the pattern can be passed down from cell to cell, influencing the behavior of daughter cells without changing the underlying DNA sequence. But the process by which these patterns are created is still shrouded in mystery. In a study published online

April 16 in Science, led by Anjana Rao, HMS professor of pathology at the Immune Disease Institute and the Program in Cellular and Molecular Medicine at Children’s Hospital Boston, the mystery of DNA methylation begins to be unraveled.

Enzymes responsible for tacking methyl groups onto cytosine bases have been identified previously, but methyl groups also need to be removed. Two waves of global demethylation occur in the early embryo, first in the paternal and maternal genomes right after fertilization and later in the cells of the embryo that are destined to become the egg and sperm. It is thought that the process of creating a new organism requires erasing the existing methyl marks so the DNA methylation pattern can be reset from scratch. Abnormal resetting of DNA methylation patterns might explain, for example, the inefficient reprogramming of induced pluripotent stem (iPS) cells—differentiated cells that were forced to return to a stem cell state and that hold much promise for stem cell therapy. Researchers have been scrambling to find human DNA demethylases, but with little success.

In their search for these genes, Rao and graduate student Mamta Tahiliani joined forces with L. Aravind, a computational biologist at the National Center for Biotechnology. Aravind had noticed that an enzyme in trypanosomes, single-celled parasitic organisms, was able to modify a methyl group on the DNA base thymine (T) at a position corresponding to that of the methyl group on 5-methylcytosine (5mC), the methylated form of cytosine. With the hypothesis that relatives of this enzyme might be able to modify the methyl group of 5mC in mammalian DNA, Aravind searched for related genes in the human genome and pulled out the mammalian TET family of proteins. Interestingly, not only is the TET1 gene fused to the MLL gene in acute myeloid leukemia, but its close family member TET2 has also been implicated in cancer. Little else was known about this family of proteins.

Back in the Rao lab, Tahiliani decided to tackle the problem and to determine if TET1 was able to modify nucleotides. Together with postdoctoral fellow Kian Peng Koh, she overexpressed TET1 in human cells growing in a dish. Intriguingly, they saw a decrease in the level of 5mC using an antibody that recognizes this nucleotide. Excited about the possibility that TET1 might demethylate 5mC, Tahiliani used thin-layer chromatography, a method that segregates molecules into little spots on a plate, to separate out the different nucleotides. She expected to see a decrease in 5mC and a corresponding increase in unmodified cytosine, but she was in for a surprise. Instead, there was an extra spot that did not correspond to either 5mC or cytosine. What was it?

Upon speaking with Yevgeny Brudno, a student in the lab of David Liu, a Howard Hughes investigator and professor of chemistry and chemical biology at Harvard, graduate student and coauthor William Pastor realized that the analytical chemistry techniques applied in Liu’s lab could be used to identify the mystery spot. Together with Liu lab postdoctoral fellow Yinghua Shen, Tahiliani used mass spectrometry to show that the spot was hydroxymethylcytosine (hmC), an oxidized form of 5mC. hmC was previously known to exist in DNA, but was generally thought to be a byproduct of oxidative DNA damage in mammals, not a deliberate DNA modification. The existence of an enzyme that converts 5mC into hmC suggests that hmC has a real biological function and changes how researchers think about this molecule.

To make sure that it was TET1 and not some other enzyme in the cells that was converting 5mC into hmC, Tahiliani purified the catalytic domain of TET1 and put it in a test tube with DNA. Then the researchers used enzymatic assays and mass spectrometry in the Liu lab to analyze the nucleotides. They found that TET1 was, indeed, the responsible party.

But is this all biologically relevant? Koh began to explore the role of TET1 in mouse embryonic stem (ES) cells and observed that these cells had high levels of TET1 that decreased dramatically when he allowed the cells to differentiate. Tahiliani and Koh found that hmC was present in the genome of ES cells and that the amount of hmC decreased when the cells were differentiated. Depleting TET1 in ES cells by RNA interference also decreased the amount of hmC. Moreover, in the same issue of Science, Skirmantas Kriaucionis and Nathaniel Heintz at Rockefeller University report that hmC is present in specific regions of the mouse brain, including Purkinje cells.

The big question is, is TET1 a DNA demethylase? The answer: maybe. It is possible that the enzymes that normally maintain methylation of cytosines cannot recognize hmC, resulting in a passive loss of methylation over time. Alternatively, hmC might be recognized by DNA repair proteins and get excised and replaced with a non-methylated base. Finally, it is also possible that hmC recruits specific hmC-binding proteins while kicking methyl-CG-binding proteins off the DNA. So, TET1 might cause DNA demethylation without actually catalyzing the removal of methyl groups.

Future goals of the Rao lab include developing much-needed tools to study hmC and figuring out its physiological role as well as the role of TET proteins in ES cells and other cell types. According to Tahiliani, “The process of DNA demethylation has always been a black box.” This work begins to pry open that box. Also, as Liu notes, “It’s an example of how collaboration not only between groups but between disciplines and schools can result in real synergy.”

Students may contact Anjana Rao at arao@cbr.med.harvard.edu for more information.

Conflict Disclosure: The authors declare no conflicts of interest.

Funding Sources: National Institutes of Health, Juvenile Diabetes Research Foundation, Harvard Stem Cell Institute, Howard Hughes Medical Institute, American Heart, National Library of Medicine, National Institutes of Health (to L.A. and L.I.), a Lady Tata Memorial Trust, National Science Foundation and the Department of Defense; the content of the work is the responsibility solely of the authors.