Genes thought to be stuck in the on position can be switched off by a recently discovered family of enzymes. These molecules target a chemical tag on specialized proteins that package DNA and control target genes

One of the off enzymes appears to be crucial for switching off genes whose excess activity might lead to X-linked mental retardation. Excessive activity of another might, by blocking the activity of certain genes, contribute to the loss of differentiation observed in many cancers.

News of these discoveries, published online Feb. 22 in Cell by HMS researchers and their collaborators, comes amid a spate of related papers published in the online Cell and two other scientific journals. The studies report similar findings on this fundamental process in yeast, fruit flies, worms, zebrafish, and mice.

The publishing flurry reflects the intense interest in the protein–DNA assembly known as chromatin and how changes there affect gene expression. Tiny tweaks to the chromatin can influence many biological processes, scientists have discovered. The changes can be inherited without altering the underlying DNA sequence, a phenomenon known as epigenetics.

Scientists are scrambling to learn the rules and players, which may be more complex and dynamic than the genome proper. In the cell nucleus, DNA strands wrap around spools of eight histone proteins. Protruding from these coils of chromatin are histone tails, where much of the epigenetic action happens.

The tails are loaded with tiny chemical marks (including acetyl, phosphate, methyl, and ubiquitin groups). Alone or, more likely, in unknown combinations, these chemical flags can open or close the chromatin to make the DNA more or less accessible and recruit other tags and proteins for gene transcription, repair, or replication. (The histone modifications are distinct from direct DNA methylation, perhaps the best-studied epigenetic state.)

“If you just look at the tails—almost every residue is modified—and do the math, the combination is astronomical,” said Yang Shi, HMS professor of pathology and senior author on one of the Cell papers. Further complicating matters, many marks seem to have specialized crews of personal enzymatic handlers that write, read, and—as the latest studies show in more detail—erase them.

Methylation Reversal

Of all the histone modifications, methylation was considered the most stable until three years ago, when Shi’s group discovered an enzyme that plucked methyl groups off histone tails. The enzyme worked on methyl groups parked at the fourth lysine (K4) and another address on the tails of the third histone (H3). But the demethylase only worked on one or two methyl groups, so it was possible that three methyls conferred the reputedly permanent stability.

Three methyls at H3K4 (H3K4me3) is particularly interesting to scientists because that tag is found at the start of many different active genes. If the histone mark locked certain genes in the on position, people speculated, perhaps it ensured that a dividing skin cell would retain all the active genes that made it a skin cell, for example, and that a dividing liver cell likewise would retain its distinctive properties.

Or maybe, as scientists began to reconsider, a demethylation enzyme for that mark just hadn’t been found yet. The search for the elusive “eraser” heated up with a prediction (and soon, proof of concept) that a certain protein domain, dubbed JmjC, had the theoretical potential to demethylate the lysine on histone tails. Many groups worked to test proteins with the promising domain.

“It was kind of a race,” said postdoctoral fellow Shigeki Iwase, co–first author on the paper from the Shi lab. Iwase and graduate student Fei Lan screened nearly 30 human proteins, a candidate approach that the Shi lab had used previously to identify an enzyme that reverses trimethylation from another H3 lysine residue.

Of the four proteins active in their enzyme assay, Shi and his colleagues focused first on one implicated in human disease. SMCX has been linked to some families with X-linked mental retardation. To explore the potential neurological role of SMCX, Lan teamed up with Peter Bayliss in the Dana–Farber Cancer Institute lab of Thomas Roberts, HMS professor of pathology, for experiments in zebrafish. They found SMCX active almost exclusively in the brain of embryonic zebrafish; those without the protein did not show normal neurological development, and the cells underwent apoptosis.

Next, Iwase worked with graduate student Luis de la Torre-Ubieta in the adjacent lab of Azad Bonni, HMS associate professor of pathology. There, in cultured rat neurons missing the gene for the demethylation enzyme, the axons seemed to grow fine, but the dendrites were stunted. The researchers could rescue the phenotype with RNA that restored the demethylation activity. In further tests, RNA with point mutations found in some human patients and other RNA designed to interfere with the demethylation could not restore proper dendrite growth and length, indicating the importance of the demethylase activity in dendrite growth regulation.

“Ultimately, it’s going to be really interesting to look at this in the neurons that are important for the disease—and not just dendrite growth and length, but at the regions of the dendrite that make synapses,” Bonni said. He speculates that the enzyme is acting on the chromatin controlling a set of genes important for the development of dendrites. “Mutations or abnormalities in chromatin-modifying enzymes are likely to turn up in a variety of neurological disorders,” he said.

Back in Shi’s lab, the researchers cloned a handful of genes with the human mutations. Some versions of the resulting enzyme were unable to erase the mark.

“Human genetic studies identified SMCX as a gene mutated in some families with X-linked mental retardation, but they didn’t know what it did,” Shi said. “Now we show that it is a histone demethylase that regulates histone modification and possibly gene transcription. This study reveals an important link between chromatin biology and human diseases.”

Unleashing Cancer

Meanwhile, researchers in the DFCI lab of Howard Hughes investigator William Kaelin, HMS professor of medicine, had been tracking the various functions of the classic tumor suppressor gene Rb, named for its discovery in retinoblastoma, a relatively rare pediatric eye tumor, and identified as mutated or otherwise incapacitated in many other cancers.

In addition to suppressing a transcription factor that revs up cell proliferation, Rb seemed to independently control cell differentiation by blocking the activity of retinoblastoma binding protein 2 (RBP2), Kaelin’s group found.

Now, it turns out that RBP2 is another member of the family that can demethylate H3K4me3, according to a paper in the online Cell by Kaelin, Howard Hughes investigator Yi Zhang at the University of North Carolina at Chapel Hill, and other collaborators. The findings are consistent with the tendency of cancer cells to dedifferentiate from their tissues of origin.

“If life was simple, RBP2 would demethylate—turn off—certain genes that need to be turned on to differentiate, and Rb would promote differentiation by antagonizing RBP2,” said Kaelin. Unfortunately, it’s probably not that simple. The researchers have evidence that Rb can either antagonize or cooperate with RBP2, depending upon the setting. Another member of the family, Plu1, is overexpressed in breast cancer, he said, suggesting that demethylases play a broader role in cancer than believed.

The round of papers currently or soon to be published characterizes about one half of the JmjC proteins in human cells on the list of potential histone demethylases. “We will most likely see another flurry of papers in a high-impact journal within the next six to eight months,” said Kristian Helin of the University of Copenhagen, senior author on another related Cell paper about the enzyme family. “Taken together, the papers on demethylases show that epigenetic marks may not be as stable as initially thought.”

That’s not necessarily bad news for those who have found naked DNA itself to be intractable to therapeutic intent and the gene activity of complex diseases recalcitrant to scientific explanation. The moveable nature of epigenetic players may make them good candidates for treating many diseases. Another type of histone tag eraser known as histone deacetylases, for example, are in the clinic used as cancer drugs a mere decade after their discovery and likely decades before scientists figure out many aspects about how they work.

“There seems to be a pretty elaborate system for dealing with chromatin fiber in all cells and in very intriguing ways that have yet to be fully figured out,” said David Allis, a chromatin researcher at Rockefeller University who discovered the first histone acetylase enzyme linked to gene activation. “There are marks. Writers put them on. Erasers take them off. And there seem to be readers. The neuronal connection is exciting. This yin–yang balance, or tug of war, of epigenetic gymnastics is going to turn out to be a big aspect of gene regulation.”