enes are controlled in at least two distinct ways. One is the sequence-specific switches that enable regulatory elements to control individual genes by interacting with set stretches of DNA associated with that gene. The other is known as epigenetic regulation, because although it is heritable, it seems to operate independently of the DNA sequence. Two Hot Papers offer an insight into the workings of one part of epigenetic regulation.

      Epigenetic control is exerted through histones, proteins that are intimately associated with DNA, much of which wraps around histones to form the chromatin of chromosomes. In 1993 Bryan Turner, professor of experimental genetics at the University of Birmingham Medical School in the U.K., put forward the idea that chemical modifications to the tails of the histone proteins created a code, somewhat like the DNA code, that regulatory elements could read and act upon. This histone code hypothesis, as it was named in 2000 by C. David Allis, of the University of Virginia School of Medicine, Charlottesville , finds confirmation in the papers at #4 and #5. "Something is writing this code," Allis has said, "and something is reading this code."

     The papers, from groups led by Tony Kouzarides of the University of Cambridge and by Thomas Jenuwein of the Research Institute of Molecular Pathology in Vienna, Austria, focus on methylation. This represents a shift from earlier work, which looked at the more common phenomenon of histone acetylation. Shiv Grewal, of the Cold Spring Harbor Laboratory, New York , had previously shown that in the budding yeast Saccharomyces pombe, two proteins are crucially important to epigenetic control. Swi6 is needed both to silence a stretch of DNA and to maintain the inactivation through cell divisions. Swi6 also needs Clr3, which de-acetylates histones, to keep inactive DNA silent. A related protein, Clr4, methylates histones.

     The mammalian equivalent of Swi6 is known as HP1, while the equivalent of Clr4 is known as SUV39H1. Using almost identical techniques the two groups showed that HP1 proteins, which have long been implicated in gene-silencing and the structure of DNA-histone complexes, would bind to histone H3 if the lysine amino acid at position 9 were methylated, but not if the nearby lysine-4 is methylated, nor if the lysine-9 is unmethylated. To prove the functional significance of these events Kouzarides’s group looked at a mutant of Clr4 in yeast. This mutant lacked the ability to methylate the lysine-9 of H3. As a result, Swi6 did not bind to the H3, and as a result of that a normally silent marker gene was expressed. Jenuwein’s group used mouse cells that lacked SUV39H1 in the same search for proof. These knockout cells did not form the histone-DNA complexes until they were given a functional version of the SUV39H1 mouse equivalent gene.

     Both groups agree on the picture that emerges from their data. SUV39H1 comes along and places a marker methyl group on lysine 9 of histone 3. This marker attracts the attention of HP1. Thus SUV39H1 writes at least this part of the histone code, and HP1 reads it. But HP1 itself contains a domain that binds with SUV39H1, which is then in a position to place a methyl marker on the next H3 histone. Like the structure of the DNA double helix itself, published exactly 50 years ago at this writing, this arrangement is highly suggestive. Says Kouzarides, it offers "a mechanism for epigenetic events to be passed on to the next generation."

     Bryan Turner, originator of the histone code hypothesis, does not begrudge these two papers their moment in the limelight. "I’m absolutely delighted," Turner tells Science Watch. But he points out that  they were "pushing at an open door." By that, Turner means that much of the evidence was already in the literature, albeit in scattered form and not linked together. These results were just what everyone wanted, Turner said, himself included, and "that’s why they are so highly cited." Turner is also convinced that things are going to prove much more complicated than this first, "shiny example" of a complete specific sequence of events underpinning epigenetic control.

     One of the exciting aspects of epigenetic control is that it probably underlies the long-term differentiation of cells. At the start of development, genes are switched on and off according roughly to their position in the developing embryo. As differentiation proceeds, however, a cell’s pattern of activity becomes fixed and is passed on to daughter cells in a stable manner. To begin with, a cell might become either liver or muscle, but thereafter liver cells remain liver cells, while muscle cells remain muscle cells. The ability to manipulate, and possibly even reverse, this type of differentiation could be the key to new approaches to disease.