It is (reasonably) common knowledge that DNA is written in a simple four-letter code: ACTG. Too simple, actually. A fifth base—5-methylcytosine 5mC—plays a crucial role in determining which bits of the DNA are actually open for business. Methylation is the basis of the epigenetic marks that control gene activation and with it the differentiation of embryonic stem cells into different tissues and the normal development and functioning of an organism. Differentiated cells can be restored to a state rather like stem cells, so-called induced pluripotent stem cells (iPSCs), suggesting that the epigenetic marks have been removed. And while epigenetic marks may have an impact on future generations, as a result of the way they affect gene functioning, in each generation the marks seem to be completely erased from the primordial germ cells, which go on to form sperm or eggs. What erases the marks?

The answer, according to two papers that currently rank among biology’s hottest as tracked by Thomson Reuters Web of Science, is “ten-eleven-translocation” proteins, designated Tet1, Tet2 and Tet3 (see adjoining table, papers #1 and #2). This newly discovered family of proteins catalyzes the oxidation of 5mC to 5-hydroxymethylcytosine, 5hmC. This (6th) nucleotide base was originally discovered in 1972 but dismissed as the uninteresting result of damage to the DNA. The laboratories of Nat Heintz at Rockefeller University and Anjana Rao at Harvard Medical School re-discovered it in 2009, when Rao’s group showed that Tet1converted 5mC to 5hmC. Then it turned out that mutations in Tet2 were associated with cancers of many kinds of cells in blood, and tumors in general have far less 5mC, which squares with the idea that cancer genes are not being properly regulated.

A 2011 paper from Rao’s lab (paper #3) showed that Tet1 and Tet2, but not Tet3, were expressed in mouse embryonic stem cells. Tet1 and Tet2 seemed to operate independently and to control different differentiation pathways, but both had binding sites for the products of Oct4 and Sox2, two genes intimately associated with pluripotency and iPSCs. This established a link between epigenetic erasure in the differentiation of embryonic stem cells and iPSCs, and suggested a role for Tet1 and Tet2 in both. (Tet3 is probably more important in the maintenance of more differentiated cells.)

The papers at #1 and #2 clarify Tet’s likely mechanism. Beyond converting 5mC to 5hmC, Tet can further convert 5hmC to 5-formylcytosine (5fC, a 7th base) and 5fC to 5-carboxylcytosine (5caC, an 8th). In the final step, thymine-DNA glycosylase (TDG), previously known mostly as a proofreading enzyme that corrects copying mistakes, removes the carboxyl group and releases cytosine for recycling. In paper #2, a group under Guo-Liang Xu, of the Institute of Biochemistry and Cell Biology of the Chinese Academy of Sciences in Shanghai, demonstrate clearly that cells depleted of TDG accumulate 5caC. In paper #1, Yi Zhang, one of the original discoverers of Tet’s activity, now with the Howard Hughes Medical Institute and Children’s Hospital, Boston, and his group detect both 5fC and 5caC in the genomic DNA of mouse embryo stem cells and organs, and show that the amount of both is affected as expected by overexpression or depletion of Tet proteins. The two papers, published back to back in Science, differ in some particulars. Zhang’s group finds that conversion of 5fC can accumulate relative to 5caC, suggesting a bottleneck in the process, while for Xu’s group conversion is efficient with no accumulation of 5fC.

Both papers, as noted above, are currently among the dozen most-cited biology reports published in the last two years (aside from reviews), according to citations tallied during a recent two-month period by Essential Science Indicators Hot Papers, a subset of the Web of Science.

Last year Zhang’s lab added another piece to the puzzle, demonstrating a crucial role for Tet1 in meiosis (paper #4). This is exciting because it links Tet1 directly to epigenetic erasure. Meiosis halves the number of chromosomes and shuffles the genome as a result of crossing over and DNA exchange between paired chromosomes, and is a process unique to the germ line cells that produce sperm and eggs. Before meiosis starts, DNA methylation is erased across the genome. Tet1 has no great role in that process in primordial germ cells, but is crucial later on in meiosis. Female mice deficient in Tet1 produce fewer eggs cells, and those egg cells are likely to contain unpaired and damaged chromosomes.

Zhang’s latest contribution to the story is hot off the presses at Nature (paper #5) and looks specifically at the role of Tet1 in genomic imprinting. This is a more specific process whereby one allele of a pair is silenced by methylation, so that the copy from only the mother or only the father functions in the offspring. Although almost all methylation is erased during the reprogramming that starts embryonic development, the imprinting pattern survives and is copied over during the formation of gametes, so that genes in the sperm are imprinted with the paternal pattern, while genes in the egg receive the maternal pattern. Male mice deficient in Tet1 give rise to offspring that, despite being genetically identical, show diverse patterns of growth, with some pups much larger than average (and growing more rapidly) and some much smaller or entirely resorbed by the female. Pups are also much more likely to die within three days of birth than controls.

Growth is important because theories of the evolution of genomic imprinting tie it to a conflict between male and female in the allocation of resources to offspring. Average placenta size is smaller for offspring of Tet1-deficient males, and so is the average size of the litter, suggesting that the female is in some sense fighting back to re-exert control over the resources going to the pups. Overall, this most recent paper demonstrates that Tet1 plays a crucial role in erasing maternal epigenetic marks in the male line and paternal marks in the female. The failure to erase those marks results in genomic imprinting defects in the next generation.

Where now? Tet proteins have extended the DNA code to 8 letters—or perhaps four letters and four diacriticals. More importantly, in addition to prospects for cancer treatment and better induced stem cells, abnormalities in genetic imprinting cause human diseases such as Prader-Willi syndrome and Angelman syndrome. Tet proteins may have a role to play in these, too.

Dr. Jeremy Cherfas is a science writer based in Rome, Italy.