Activation and repression of transcription in eukaryotes involve changes in the chromatin fiber that can be accomplished by covalent modification of the histone tails or the replacement of the canonical histones with other variants. The histone H2A variant of Drosophila melanogaster, H2Av, localizes to the centromeric heterochromatin and is recruited to an ectopic heterochromatin site formed by a transgene array. His2Av behaves genetically as a PcG gene and mutations in His2Av suppress position effect variegation (PEV), suggesting that this histone variant is required for euchromatic silencing and heterochromatin formation. His2Av mutants show reduced acetylation of histone H4 at Lys 12, decreased methylation of histone H3 at Lys 9, and a reduction in HP1 recruitment to the centromeric region. Neither H2Av accumulation nor histone H4 Lys 12 acetylation is affected by mutations in either Su(var)3-9 or Su(var)2-5. The results suggest an ordered cascade of events leading to the establishment of heterochromatin, requiring the recruitment of the histone H2Av variant followed by H4 Lys 12 acetylation as necessary steps before H3 Lys 9 methylation and HP1 recruitment can take place (Swaminathan, 2005).

The basic unit of chromatin is the nucleosome, which is made up of 146 bp of DNA wrapped around a histone octamer composed of two molecules each of the histones H2A, H2B, H3, and H4. Activation of gene expression requires the transcriptional machinery to overcome the compaction of chromatin: work in recent years has uncovered several strategies to accomplish this goal. ATP-dependent chromatin remodeling has been studied extensively as a mechanism to make the DNA accessible to the transcription apparatus. Covalent modification of the unstructured and solvent-exposed N-terminal tails of histones has also been shown to participate in processes such as activation or repression of transcription and chromosome condensation and segregation (Swaminathan, 2005 and references therein).

Replacement of canonical histones with other variants might provide an alternative mechanism for controlling transcription by inducing altered nucleosomal structures; in fact, recent work has implicated ATP-dependent chromatin remodeling complexes in the process of histone variant replacement (Krogan, 2003; Mizuguchi, 2004). The role of histone variants, and specially those of H3 and H2A, in various nuclear processes has been long appreciated. There are at least three different families of H2A variants present in a variety of organisms from yeast to mammals, and the degree of conservation among members of each family is greater than than to the canonical H2A. H2AX is thought to play a role in DNA double-strand break repair; the serine in the SQEY motif of H2AX is phosphorylated at the site of the DNA damage and serves as a signal for the recruitment of repair proteins. Macro H2A1, another H2A variant, has been shown to have a role in X-chromosome inactivation and dosage compensation in mammals, where it is found to localize to the inactive X after silencing has been established (Swaminathan, 2005).

H2A.Z, the Drosophila version of which is the subject of this overview, is a third histone H2A variant highly conserved across species and, therefore, likely to play an important role in chromatin function (van Daal, 1988; Stargell, 1993). H2A.Z is an essential protein in Drosophila and in mice (van Daal, 1992; Faast, 2001), and has been implicated in both activation and repression of transcription (Dhillon, 2000; Santisteban, 2000; Larochelle, 2003). Gene expression analyses using whole-genome microarrays show that H2A.Z (named Htz1 in yeast) is involved in both activation and silencing of transcription in Saccharomyces cerevisiae (Meneghini, 2003). Recent studies have also shown that H2A.Z is enriched in the pericentric heterochromatin (Rangasamy, 2003) during early mammalian development (Swaminathan, 2005).

Heterochromatin consists of highly compacted DNA that is present around the centromeres and telomeres and is also dispersed at certain sites along the chromosomes. Heterochromatin has been found to be essential for proper chromosome segregation and genomic stability and for maintaining dosage compensation by inactivating one of the X chromosomes in female mammals. Recent studies have helped elucidate several steps in the pathway leading to the formation of heterochromatin. The earliest step established so far requires the RNAi machinery for targeting of small RNAs to heterochromatin by the RNA-induced initiation of transcriptional gene silencing (RITS) complex. Subsequent steps require deacetylation of histone H3 Lys 9 followed by methylation of the same residue by Su(var)3-9. This modified histone then recruits HP1, which in turns recruits the Suv4-20 methyltransferase to trimethylate histone H4 at Lys 20 (Swaminathan, 2005 and references therein).

Given the large number of suppressors and enhancers of PEV identified in Drosophila, there must be additional steps involved in the establishment and maintenance of heterochromatin. Although acetylation of histone tails is generally correlated with transcription activation, and establishment of heterochromatin requires deacetylation of H3 Lys 9, acetylation of other residues in H3 or H4 might also play a role in silencing processes. Histone acetyl transferases have been found to be physically associated with Su(var)3-9, and mutations in a MYST domain acetyltransferase behave as suppressors of PEV. These results support the idea that acetylation of specific histone residues might be involved in the formation of heterochromatin. This study demonstrates that the H2A.Z variant of Drosophila, H2Av, plays a role in Pc-mediated silencing and in the establishment of centromeric heterochromatin. In addition, acetylation of H4 Lys 12 is required subsequent to H2Av replacement but before H3 Lys 9 methylation. The results highlight the complexity of the multistep process leading to heterochromatin formation in higher eukaryotes (Swaminathan, 2005).

The establishment of heterochromatin has so far been defined as a four-step process initiated by the RNAi machinery through the production of small RNAs homologous to centromeric DNA repeats that are recruited to prospective heterochromatic regions as part of the RNA-induced initiation of transcriptional gene silencing (RITS) complex. The next step described in this process thus far is the deacetylation and subsequent methylation of histone H3 Lys 9, which serves to recruit HP1. HP1 then recruits the Suv4-20 methyltransferase to trimethylate histone H4 at Lys 20. The work described here suggests that heterochromatin formation is more complex than previously thought, and it involves at least two additional steps. One step requires recruitment of H2Av or replacement of the canonical histone H2A for the H2Av variant. This requirement is highlighted by the observation that mutations in the His2Av gene act as suppressors of position effect variegation by modulating the silencing effect of heterochromatin on the adjacent white gene (Swaminathan, 2005).

The replacement of H2A for H2Av is not specific to heterochromatin, and it may also take place in silenced regions of the euchromatin, since it appears that His2Av behaves genetically as a PcG gene. PcG proteins are responsible for the maintenance of epigenetic silencing of the homeotic genes during Drosophila development. The His2Av gene can be classified as a PcG gene, since mutations in His2Av enhance the phenotype of Pc mutants, suppress the phenotype of mutations in trxG genes, and cause ectopic expression of the Ant gene. The involvement of H2Av in Pc-mediated silencing is not completely unexpected, since H2Av is critical for the establishment of pericentric heterochromatin and both processes share similar strategies. Heterochromatin-induced silencing requires methylation of H3 at Lys 9 by the Su(var)3-9 histone methyltransferase, whereas Pc-induced silencing involves the recruitment of the ESC-E(z) complex to methylate H3 at Lys 27. Although the modified residues are different, in both cases the modification serves as a tag to bind chromo domain-containing proteins, HP1 in the case of pericentric heterochromatin and Pc in euchromatic silencing. Given the parallels between the two processes, it was surprising to find that replacement of H2Av was required for subsequent H3 Lys 9 methylation in heterochromatin but not for H3 Lys 27 methylation in silenced regions of euchromatin. This later conclusion is supported by the observation that neither H3 Lys 27 methylation nor E(z) recruitment is affected by mutations in His2Av (Swaminathan, 2005).

The requirement of H2Av for Pc recruitment but not for H3 Lys 27 methylation points to a slightly different strategy in the establishment of silencing in the euchromatin compared to heterochromatin. Heterochromatic silencing appears to involve a series of events that take place in a linear pathway; each event is dependent on the previous one for proper heterochromatin assembly. In this cascade of events, replacement of H2A by H2Av appears to be an early step in the process, although the results cannot distinguish among the possibilities that H2Av is recruited before, after, or in parallel to the recruitment of the RITS complex by the RNAi machinery. Surprisingly, the observed effects of H2Av on Pc-mediated silencing point to several possible mechanisms, all slightly different from that involved in heterochromatin formation. One formal possibility that would be consistent with the results but that is considered less likely is that H2Av acts downstream of Pc; recruitment of this protein would be required for H2Av replacement, which would then in turn stabilize the binding of the Pc complex. A second possibility is that H2Av replacement is a relatively late event in the process, acting downstream of H3 Lys 27 methylation instead of being required for this modification. In this scenario, H2Av would not be required for the recruitment of the ESC–E(z) complex, but it would be required subsequently to alter chromatin structure and allow Pc recruitment. Alternatively, euchromatic Pc-mediated silencing might be accomplished by two relatively independent parallel pathways that converge at the end to ensure Pc binding to the chromatin. One pathway would alter chromatin structure by recruiting ESC–E(z) and methylating H3 at Lys 27; a second parallel but independent pathway would further alter chromatin structure by replacing H2A for H2Av. Both processes would then be required for the recruitment of the Pc-containing PRC-1 complex (Swaminathan, 2005).

The apparent association of H2Av with silenced regions might appear puzzling in view of findings in other systems. In Tetrahymena, H2A.Z is present in the transcriptionally active macronucleus, but it is not detected in the silenced micronucleus. In S. cerevisiae, the H2A.Z histone variant Htz1 is required for the expression of SWI/SNF-dependent genes such as PHO5 and GAL1; interestingly, although Htz1 is required for activation, it is present at higher levels in the promoters of these genes when they are repressed than when they are transcriptionally active (Santisteban, 2000). Nevertheless, Htz1 has also been shown to be present at silenced loci in yeast, and to be required for HMR and telomere-induced silencing (Dhillon, 2000). A detailed analysis of the role of Htz1 in gene expression in yeast using whole-genome microarrays resulted in the identification of 214 genes that are activated and 107 genes that are repressed by Htz1. Htz1-activated genes are located adjacent to heterochromatin; Htz1 is enriched in the promoters and coding regions of activated genes (Meneghini, 2003), where it appears to block the spreading of heterochromatic silencing (Swaminathan, 2005).

One possible explanation for the functional differences between the yeast and Drosophila H2A.Z homologs is that Drosophila lacks H2A.X, and, therefore, H2Av plays the roles of both histone variants. Nevertheless, the localization of H2A.Z in heterochromatin is not limited to Drosophila. Studies in mice also show an association of this histone variant with pericentric heterochromatin (Rangasamy, 2003). The apparent contradiction between the two conflicting roles for H2A.Z in various organisms could be explained if the role of this histone variant is to assemble a chromatin that is more accessible to other chromatin-remodeling complexes, and that the final result depends on the type of factors recruited to H2A.Z-containing chromatin. This possible role is consistent with the finding that the functionally essential C-terminal domain of H2A.Z (Clarkson, 1999) is exposed on the surface of H2A.Z-containing nucleosomes, and therefore it could serve to recruit other factors (Suto, 2000). Analytical ultracentrifugation experiments suggest that nucleosomes reconstituted with H2A.Z have decreased stability compared to those reconstituted with the canonical H2A histone, and therefore might allow greater accessibility of various factors to H2A.Z-containing chromatin (Abbott, 2001). The decrease in stability is also supported by the crystal structure of H2A.Z-containing nucleosomes, which suggests an altered interface between the H2A.Z-H2B dimer and the H3-H4 tetramer (Suto, 2000). Alternatively, the exposed domain of H2A.Z could mediate internucleosome interactions to give rise to higher-order compacted chromatin structures (Fan, 2002). The H2A.Z-containing chromatin fiber contains regularly spaced nucleosomes (Wallrath, 1995), an arrangement that is characteristic of pericentric heterochromatin (Swaminathan, 2005).

The results discussed here suggest a complex multistep model for heterochromatin assembly. The nature of the first step in the process might involve the RNAi machinery and the recruitment of the RITS complex, although it is also possible that H2Av recruitment takes place prior to this step. The mechanism by which H2Av recruitment is controlled is unclear. Recent studies in yeast have shown that a complex containing swr1, an Snf-2-related ATP-dependent chromatin-remodeling factor, is able to efficiently exchange H2A for H2A.Z (Krogan, 2003; Kobor, 2004; Mizuguchi, 2004). A Drosophila homolog of SWR1 is domino, which has been characterized as a PcG gene and a dominant suppressor of PEV. It is plausible that RITS recruits a Domino-containing chromatin-remodeling complex, which in turn replaces histone H2A for H2Av. This alteration would then facilitate the recruitment of histone-modifying enzymes that would further change chromatin structure. The current results suggest that the next step in the process requires the acetylation of histone H4 Lys 12, but the nature of the enzyme responsible for acetylation of H4 Lys 12 is not known. A candidate for this role could be the product of the chameau gene, which encodes a MYST domain histone acetyltransferase. Mutations in chameau behave as PcG genes and suppress PEV (Grienenberger, 2002), but they fail to affect H4 Lys 12 acetylation, suggesting that the Chameau protein affects a different step in the establishment of heterochromatin. Acetylation of H4 Lys 12 is then followed by methylation of H3 Lys 9 and recruitment of HP1, which in turn tethers Svar4-20 to methylate H4 Lys 20. Additional steps are likely to be required in this complex process. Analysis of the large collection of Su(var) and E(var) mutations identified in Drosophila should make possible the elucidation of additional steps in this pathway, resulting in a comprehensive understanding of the molecular events involved in heterochromatin formation (Swaminathan, 2005).

GENE STRUCTURE

cDNA clone length - 913

Bases in 5' UTR - 149

Exons - 4

Bases in 3' UTR - 338

PROTEIN STRUCTURE

Amino Acids - 141

Structural Domains

The Tetrahymena histone H2A variant designated hv1 is localized exclusively in the transcriptionally active macronucleus and is absent from the quiescent micronucleus. A cDNA clone of the hv1 gene was used to screen a Drosophila cDNA library. A cross-hybridizing clone was recovered and shown by sequence analysis to code for a protein homologous to hv1 as well as to the chicken H2A variant, H2A.F, the sea urchin H2A variant, H2A.F/Z and the mammalian H2A variant H2A.Z. Southern analysis of Drosophila genomic DNA indicates that the H2AvD (H2A variant Drosophila) gene is present in one copy. In situ hybridization places the locus at 97CD on chromosome 3, while the S-phase regulated histone genes are on chromosome 2. Thus the Drosophila H2A variant should be accessible to genetic analysis, which will enable its function to be determined (van Daal, 1988).

The H2AvD sequence contains the nonapeptide, which is conserved in all H2A's, at position 23-31, indicating that this cDNA encodes an H2A-like protein. However, it is clear that the sequence differs markedly from that of the H2A encoded in the Drosophila histone gene cluster. There are 52 amino acid changes in the first 124 amino acids, which indicates a similarity of only 59%. This is approximately the same degree of similarity shown by the H2A variants of chicken (60%), sea urchin (56%-57%), mammals (56%) and Tetrahymena (62%-65%) relative to their respective major S-phase regulated H2A's. The cDNA sequence predicts that H2AvD is a larger molecular weight protein than H2A.1. An antibody to the carboxy terminal portion of H2AvD detects a protein of lower mobility than the major histone H2A on electrophoresis in SDS polyacrylamide gels. An H2A variant (D2) has been previously reported in Drosophila. D2 has a molecular weight of 13,400 daltons and differs from H2A.1 in that it contains no methionine residues and has an increased histidine content. H2AvD shares these properties (van Daal, 1988).

The amino and carboxy terminal ends of the Drosophila and Tetrahymena H2A variants have diverged. However, there is a remarkably high similiarity when comparing the body of the protein in both variants. There are only 16 amino acid changes in the region of amino acids 18-120, which gives a similarity of 84%. The carboxy terminal end of hvl (from amino acid 121) is completely different from that of H2AvD in both amino acid composition and length. The N-terminus shows some sequence similarity (less than 50%), but again the length is different. It should be noted that the major Tetrahymena histone H2A differs significantly from the major H2A's of the other species. The H2A.l's of human, cow, rat and chicken are 95%-99% similar to each other. The Drosophila and sea urchin H2A.l's are 81%-87% similar to those of mammals, chicken and each other. However, the Tetrahymena H2A.1 is only 63-69% similar to those of other species, and so it is perhaps not surprising that H2AvD is quite different from hvl, whereas, it is very similar (97%) to the sea urchin variant (van Daal, 1988).

A comparison of the deduced protein sequences of the H2A variants of chicken (H2A.F), sea urchin (H2A.F/Z) and mammals (H2A.Z) with that of Drosophila (H2AvD) shows clearly that the amino acid sequence of H2AvD is more closely related to the chicken, sea urchin and mammalian H2A variants than it is to the major Drosophila H2A, H2A.l. The conservation in the first 122 amino acids is extremely high. There is 99% similarity between the Drosophila and the sea urchin proteins, with only one amino acid change (ser to asn at position 38). There is 98% similarity between H2AvD and the chicken and mammalian proteins. Both have two amino acid changes (ala to gln at position 21 and ser to thr at position 38 in chicken and ala to thr at position 14 and ala to glu at position 21 in mammals). The carboxyl ends of these four variants all differ. However, the Drosophila variant is the most divergent. The C terminal tail is longer than the others and its sequence is also dissimilar. Southern blot analysis of Drosophila genomic DNA indicates that the H2AvD gene is unique (van Daal, 1988).

date revised: 25 July 2005

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.