The pattern of histone modification in the chromatin surrounding a gene is important for its transcriptional activity. Methylation, acetylation and phosphorylation are among the different covalent modifications that contribute to the combinatorial potential of histone patterns. But a new study published in Nature by the Kouzarides group hints at the existence of yet another level of complexity that determines the activation state — the precise number of methylation events.

Tony Kouzarides and colleagues focused their attention on a group of proteins that contain so-called SET domains. SET domains catalyse methylation of specific lysine (K) residues in the amino-terminal tails of histones. The authors identified Set1 and Set2 as the lysine methylases responsible for methylation of histone H3, and found that Set1 is specifically responsible for methylation of H3 at residue K4.

Given that lysine residues can be mono-, di- and even tri-methylated, Kouzarides and co-workers raised antibodies that can distinguish between di- and tri-methylated K4 of H3. Using these antibodies, they showed that Set1 is responsible for both di- and tri-methylation of K4 of H3.

Set1 is thought to be a transcriptional repressor. However, the detection of di-methylated K4 H3 at euchromatic loci — which generally represent transcriptionally active loci — indicated that Set1 might also function as an activator. So, Kouzarides and colleagues carried out gene-expression-profiling analysis, which revealed 480 genes whose activity was significantly reduced in the absence of Set1. Reduction of the top-20 genes varied between 53% and 38%, which indicates that Set1 is required, but not responsible, for gene activation.

Using chromatin immunoprecipitation assays, Kouzarides and colleagues showed that Set1-activated genes are methylated at K4 of H3, and that constitutively active genes contain both di- and tri-methylated K4 of H3. What came as a surprise was that the methionine-regulated MET16 gene, as well as several other inducible genes, is both di- and tri-methylated at K4 of H3 — but only when the gene is active. In the repressed state, however, K4 of H3 is di- but not tri-methylated.

The authors propose that “the role of dimethylated K4 H3 may be to determine a transcriptionally 'permissive' chromatin environment, whereas the trimethylated state may allow for an 'active' chromatin conformation”. Finally, these data indicate that an — as yet unknown — mechanism must exist that prevents Set1 from adding a third methyl group.