PerspectiveMolecular Biology

The Histone Modification Circus

See allHide authors and affiliations

Science  06 Apr 2001:
Vol. 292, Issue 5514, pp. 64-65
DOI: 10.1126/science.1060791

There's an old circus routine that begins with a little car driven into the arena. One clown climbs out, then another emerges and another and another until the audience laughs and wonders how many more can possibly appear. The sequential discoveries of a series of distinct covalent modifications of histone proteins bring this absurd circus skit to mind. Several recent reports are beginning to bring order to the apparent chaos of the many histone modifications that are required for the regulation of gene expression. One of these reports, by Nakayama et al. (1) on page 110 of this issue, examines how histone modifications regulate the silencing of genes. The authors propose a chronological order for a dual histone modification step and elucidate how each modification contributes to gene silencing.

Histone proteins associate with DNA to form nucleosomes, permitting copious amounts of DNA to be neatly packaged into the nucleus. Histones are also direct regulators of gene expression because they can alter the accessibility of gene sequences in the DNA to components of the transcription and replication machinery that must bind to DNA to carry out their work. Addition of acetyl groups (acetylation) to amino acids in the amino terminus of histones provides a valuable model for understanding other histone modifications. The acetylation model is useful because acetylation strongly correlates with gene activation, and because many of the acetyltransferase enzymes that acetylate histones are coactivators of transcription, recruited to promoters by DNA-bound activators. From these observations a model emerges that elaborates the importance of acetylation in promoter-specific alterations of repressed chromatin. This model is substantiated by the finding that certain histone deacetylases—enzymes that remove acetyl groups from histones—are transcriptional corepressors that are recruited to chromatin by DNA-bound repressor proteins.

But why are there so many distinct types of histone modifications, and what is the precise mechanism through which they alter gene transcription? Indeed, the list of well-characterized modifications continues to grow, with the recent addition of phosphorylation (2, 3) and ubiquitination (4, 5). With the Nakayama et al. paper, another histone modification—methylation (the addition of methyl groups to histones)—now takes center stage (1, 6, 7). These investigators (1) describe the relation between histone methylation and deacetylation during gene silencing in the fission yeast. Their study builds upon the recent identification of the histone methyltransferases (enzymes that add methyl groups to histone amino acids) SuV39h1 in mammals, and its homolog Clr4 in yeast (8). Classical genetics had revealed that the genes encoding these enzymes and their relatives in the fly are required for maintaining certain chromosomal regions such as the centromeres in an inert (heterochromatic) state.

Nakayama et al. developed a specific antibody that detects methylation of amino acid lysine 9 (Lys-9) in histone H3, which is a substrate of the Clr4 methyltransferase in fission yeast. They used this antibody in chromatin immunoprecipitation assays to monitor specific DNA sequences associated with H3-Lys-9 methylation. The authors found that histone methylation at centromeres requires Clr4 methyltransferase activity and also depends on a histone deacetylase that specifically removes the acetyl moiety from the nearby Lys-14 of histone H3. Because Lys-14 is a common target of acetyltransferases associated with gene activation, these results suggest that deacetylation of Lys-14 leads to methylation of Lys-9 in an obligatory sequence. A second class of proteins, typified by HP1 (heterochromatin protein 1) in the fly, is also required for the maintenance of gene silencing. In fission yeast, the HP1 homolog Swi6 is localized to the heterochromatin of centromeres. The Nakayama et al. work (1), and a second study by Bannister et al. (6), demonstrate that localization of Swi6 to heterochromatin also requires Clr4 methyltransferase activity. Thus, it appears that deacetylation of H3 leads to its methylation, which then results in the localization of Swi6 to heterochromatin. Previous data have indicated that Swi6 expands across the heterochromatic region, presumably creating a unique silencing chromatin structure. Taken together, this work suggests an obligatory sequence of histone modifications that recruits structural proteins to DNA, resulting in the creation of heterochromatin.

The sequence of histone modifications during gene silencing—first, deacetylation of Lys-14, and then methylation of Lys-9—has a precedent in the dual ordered modifications of histone H3 during gene activation. Histone phosphorylation at serine-10 precedes and promotes acetylation at Lys-14 (9, 10). Thus, together these studies of gene silencing and activation suggest that histone H3 exists in two modification states: an “Off” state characterized by methylation of Lys-9 and deacetylation of Lys-14, and an “On” state wherein Ser-10 is phosphorylated and Lys-14 is acetylated (see the figure). Interestingly, in vitro methylation of Lys-9 inhibits phosphorylation of Ser-10 and phosphorylation of Ser-10 inhibits methylation of Lys-9 (8). This dual arrangement could serve to reinforce gene silencing or activation in vivo.

Modifying histones.

Modifications to amino acid residues in histone protein H3 and the specialized domains of other proteins that bind to modified residues. (A) Dual modification of histone H3 occurs during gene silencing or gene activation. During gene silencing (“Off” state), Lys-9 (K9) is methylated (Me) and Lys-14 (K14) is deacetylated (X). During gene activation (“On” state), Ser-10 (S10) is phosphorylated and Lys-14 is acetylated. Methylation of Lys-9 blocks phosphorylation of Ser-10 and phosphorylation of Ser-10 blocks methylation of Lys-9. (B) As a result of these modifications, the chromodomain (ChrD) of regulatory proteins binds to methylated Lys-9 during gene silencing, and the bromodomain (BrD) of regulatory proteins binds to acetylated Lys-14 during gene activation.

The next compelling question is exactly how these histone modifications alter transcription. A recently advanced idea, the “histone code” hypothesis (11, 12), holds that covalent modifications of histones constitute an intricate pattern that creates a docking surface with which the modules of other proteins can interact. Proteins containing these modules bind to chromatin to alter its structure, provide additional enzymatic activity, or to target other regulatory proteins. Initial support for the docking idea comes from histone acetylation: The bromodomain present in histone acetyltransferases and in other proteins that interact with chromatin binds with higher affinity to peptides bearing acetylated lysine than to unmodified peptides (13, 14). Two recent papers (6, 7) that also examine how histone methylation regulates the silencing of heterochromatin provide strong support for the docking proposal. These studies analyze a domain, called the chromodomain, found in heterochromatin-associated proteins, such as HP1 and Swi6. The chromodomain binds with high affinity to histone H3 peptides bearing methylated Lys-9 both in vitro and in vivo. Thus, a reciprocal docking mechanism may consist of, on the one hand, binding of a bromodomain to acetylated histone during gene activation, and on the other, binding of a chromodomain to methylated histone during gene silencing (see the figure). All chromodomain proteins, however, do not bind to methylated Lys-9 of histone H3 (6), prompting speculation that there may be additional patterns of histone modifications (and even other targeting mechanisms) (15) that specify the binding of other chromodomain proteins. A looming challenge will be to determine the histone modification “code” that specifies the binding to chromatin of numerous other domains in chromatin-associated proteins.

Thus, the silencing and activation of genes may require multiple modifications of histones, which generate unique surfaces for the binding of proteins that carry out further chromatin-related processes. The modification patterns that exist in these histones are just beginning to be decoded, yet it is already abundantly clear that the many distinct covalent modifications of chromatin, although initially confusing, are an important aspect of genomic regulation.


View Abstract

Navigate This Article