Histone H3 variants and modifications on transcribed genes

Histone-modifying complexes may be linked to chromatin assembly pathways through nucleosome assembly machines.

Chromatin structure is disrupted by the process of gene transcription, and the state of the histones plays an important role in gene regulation. Transcribing genes tend to be marked by specific histone modifications that are thought to modulate transcriptional activity. Recently, Ahmad and Henikoff (1) found that transcribed genes in Drosophila are also enriched for histone variant H3.3, which is deposited in chromatin through a DNA replication-independent nucleosome assembly pathway. In an article from the same group in this issue of PNAS, McKittrick et al. (2) demonstrate that a Drosophila cell line contains enough H3.3 to package all of the transcribed genes and that H3.3 is enriched in modifications associated with transcription (e.g., acetylation of select lysines and methylation of lysine 4). Lysine 9 methylation, which is abundant in transcriptionally silent heterochromatin, occurs primarily on the major form of H3 that is assembled into nucleosomes by the replication-dependent pathway. These results led McKittrick et al. to propose that the modification state of H3 associated with a gene is tied to the pathway that assembled the gene into nucleosomes. This hypothesis could also account for observations made by Waterborg (3) 13 years ago, who found that acetylated lysines and lysine 4 methylation are enriched in the alfalfa variant H3.2, whereas lysine 9 methylation was found primarily in the major form, H3.1.

McKittrick et al. extracted histones from Drosophila Kc cells and separated H3.3 from H3 on a reverse-phase HPLC column. These proteins differ in only 4 aa, and antibodies that can distinguish between them are not available. The purified H3 forms were then analyzed for posttranslational modifications by MS and a combination of Western blots and ELISA assays using antibodies specific for H3 modifications. Although no specific modification was restricted to either H3 or H3.3, those most commonly associated with transcribed genes (e.g., di-and trimethylation of lysine 4 and acetylation of lysines 9, 14, 18, and 23) were enriched in the H3.3 variant. By contrast, dimethylated lysine 9, a modification associated with repressed genes, was enriched in H3. H3, like most major histone forms, is synthesized during S phase and is incorporated into newly synthesized chromatin in a pathway linked to DNA replication (4). The finding that this form of H3 carries the majority of lysine 9 methylation (a mark of heterochromatin) suggests that the replication-dependent pathway assembles silent heterochromatin. This modification is conferred by the SuVar3-9 histone methyltransferase and recognized by the heterochromatin-organizing protein HP1 (5).

Replication-independent nucleosome assembly in Drosophila uses histone variant H3.3 instead of H3 in nucleosomes formed between S phases (6). The fact that these nucleosomes accumulate at transcribed genes suggests that nucleosomes are replaced during transcription, a conclusion consistent with recent data (7-9). These studies demonstrated that certain acidic proteins (e.g., Spt6 and Spt16) function both as transcription elongation factors and nucleosome assembly proteins. It appears that elongating RNA polymerase II partially or completely displaces nucleosomes, which then need to be reassembled behind the polymerase to maintain transcriptional specificity. Without efficient nucleosome reassembly during elongation, transcription initiation complexes can form on spurious TATA-like sequences in the body of the gene and lead to inaccurate initiation (7, 8). If Spt6 and/or Spt16 use histone H3.3 in nucleosomes reassembled during transcription in Drosophila, it would lead to the accumulation of H3.3 in transcribed genes as reported by Ahmad and Henikoff (1). It is worth noting that Drosophila Spt16 has been found to track with RNA polymerase on polytene chromosome puffs, sites that are active in transcription (10).

But what of the potential connection between histone H3.3 accumulation on active genes and H3 modifications associated with gene activity? Because both of these are associated with transcription, might it not be a coincidence that transcription-related modifications accumulate on H3.3? Enzyme complexes that carry out transcription-related histone modifications can be recruited to active genes by DNA binding transcription activators, initiation factors, RNA polymerase II, and elongation factors. Gcn5-containing histone acetyltransferase complexes are capable of acetylating the H3 lysines that were monitored by McKittrick et al. and are recruited to genes by transcription activators (11-13). The budding yeast methyltransferase Set1 is a homolog of the Drosophila Trithorax transcriptional regulator and is responsible for methylation of lysine 4 of histone H3 (14, 15). Set1 is recruited to active genes by RNA polymerase via a polymerase-associated complex termed Paf1 (16, 17). These examples illustrate that the transcription machinery can recruit histone-modifying enzymes. Thus, if H3.3 accumulates on a gene while enzymes recruited by the transcriptional machinery continue to modify the histones, H3.3 would then acquire more of these modifications than the H3 it had replaced. McKittrick et al. argue, however, for a more direct link between the replication-independent chromatin assembly pathway that deposits H3.3 on active genes and accumulation of transcription-associated modifications on H3.3. They suggest that a connection between these processes can better explain apparent discrepancies between chromatin immunoprecipitation studies localizing the modified histones to particular sequences and antibody staining of polytene chromosomes.

McKittrick et al. propose that histone-modifying complexes may be linked to chromatin assembly pathways by associating with nucleosome assembly machines (replication dependent or independent). This intriguing possibility is consistent with the finding that specific acetyltransferase complexes associate with nucleosome assembly factors in budding yeast. For example, the histone H3 acetyltransferase complex NuA3 can interact with the Spt16 elongation/nucleosome assembly factor, and the SAS H4 acetyltransferase complex can interact with the Asf1 nucleosome assembly factor (18, 19). Histone-modifying enzymes associated with nucleosome assembly factors could, in principle, bring about modification of the histones bound to those assembly factors. It is not known whether Spt6, Spt16, or other transcription-linked nucleosome assembly factors are associated with modified forms of H3.3. However, there is precedent for the association of modified histones with nucleosome assembly factors. The Drosophila RCAF nucleosome assembly complex, which contains a homolog of Asf1 and mediates nucleosome assembly during DNA replication and repair, contains H3 acetylated on lysine 14 (20). Moreover, the human replication-linked chromatin assembly complex CAC can contain acetylated forms of histone H4 and perhaps H3 (21). Thus, although many details are missing, the types of interactions that could be necessary for nucleosome assembly pathways to participate in histone modification seem feasible. Indeed, one can speculate that the recruitment of histone-modifying complexes by the transcription machinery is primarily required during the initial activation of a gene for modification of nucleosomes formed by the replication-dependent pathway. After the first round of transcription and displacement of several nucleosomes, the replication-independent nucleosome assembly pathway and associated histone-modifying enzymes could maintain the modified state of replacement histones across the gene (Fig. 1). Although such a model is clearly fanciful at present, it suggests further experiments addressing the role of nucleosome assembly pathways in regulating gene expression.

Fig. 1.

Fanciful model of how nucleosome assembly pathways might contribute to the modification of replacement histone variants during transcription. During the first round of transcription, histone-modifying complexes (e.g., acetyltransferases and methyltransferases) associated with promoter-bound transcription factors and elongating RNA polymerase II (Pol II) modify nucleosomes that were assembled during S phase. During transcription polymerase displaces several, but not necessarily all, nucleosomes in the body of the gene. These nucleosomes are replaced by new ones containing replacement histone variants (e.g., H3.3) by a replication-independent nucleosome assembly complex that may be associated with the elongating RNA polymerase. Histone-modifying complexes associated with the nucleosome assembly complex would place transcription-related histone modifications on the replacement histones either before or after their assembly into nucleosomes. Through multiple rounds of transcription, the replacement histones accumulate on the gene and become the primary form of modified histone.