Posttranslational histone modifications are believed to allow the epigenetic transmission of distinct chromatin states, independently of associated DNA sequences. Histone H3 lysine 9 (H3K9) methylation is essential for heterochromatin formation; however, a demonstration of its epigenetic heritability is lacking. Fission yeast has a single H3K9 methyltransferase, Clr4, that directs all H3K9 methylation and heterochromatin. Using releasable tethered Clr4 reveals that an active process rapidly erases H3K9 methylation from tethering sites in wild-type cells. However, inactivation of the putative histone demethylase Epe1 allows H3K9 methylation and silent chromatin maintenance at the tethering site through many mitotic divisions, and transgenerationally through meiosis, after release of tethered Clr4. Thus, H3K9 methylation is a heritable epigenetic mark whose transmission is usually countered by its active removal, which prevents the unauthorized inheritance of heterochromatin.

Centromere sequences are not conserved between species, and there is compelling evidence for epigenetic regulation of centromere identity, with location being dictated by the presence of chromatin containing the histone H3 variant CENP-A. Paradoxically, in most organisms CENP-A chromatin generally occurs on particular sequences. To investigate the contribution of primary DNA sequence to establishment of CENP-A chromatin in vivo, we utilised the fission yeast Schizosaccharomyces pombe. CENP-ACnp1 chromatin is normally assembled on ∼10 kb of central domain DNA within these regional centromeres. We demonstrate that overproduction of S. pombe CENP-ACnp1 bypasses the usual requirement for adjacent heterochromatin in establishing CENP-ACnp1 chromatin, and show that central domain DNA is a preferred substrate for de novo establishment of CENP-ACnp1 chromatin. When multimerised, a 2 kb sub-region can establish CENP-ACnp1 chromatin and form functional centromeres. Randomization of the 2 kb sequence to generate a sequence that maintains AT content and predicted nucleosome positioning is unable to establish CENP-ACnp1 chromatin. These analyses indicate that central domain DNA from fission yeast centromeres contains specific information that promotes CENP-ACnp1 incorporation into chromatin. Numerous transcriptional start sites were detected on the forward and reverse strands within the functional 2 kb sub-region and active promoters were identified. RNAPII is enriched on central domain DNA in wild-type cells, but only low levels of transcripts are detected, consistent with RNAPII stalling during transcription of centromeric DNA. Cells lacking factors involved in restarting transcription—TFIIS and Ubp3—assemble CENP-ACnp1 on central domain DNA when CENP-ACnp1 is at wild-type levels, suggesting that persistent stalling of RNAPII on centromere DNA triggers chromatin remodelling events that deposit CENP-ACnp1. Thus, sequence-encoded features of centromeric DNA create an environment of pervasive low quality RNAPII transcription that is an important determinant of CENP-ACnp1 assembly. These observations emphasise roles for both genetic and epigenetic processes in centromere establishment.

Specialized chromatin containing CENP-A nucleosomes instead of H3 nucleosomes is found at all centromeres. However, the mechanisms that specify the locations at which CENP-A chromatin is assembled remain elusive in organisms with regional, epigenetically regulated centromeres. It is known that normal centromeric DNA is transcribed in several systems including the fission yeast, Schizosaccharomyces pombe. Here, we show that factors which preserve stable histone H3 chromatin during transcription also play a role in preventing promiscuous CENP-ACnp1 deposition in fission yeast. Mutations in the histone chaperone FACT impair the maintenance of H3 chromatin on transcribed regions and promote widespread CENP-ACnp1 incorporation at non-centromeric sites. FACT has little or no effect on CENP-ACnp1 assembly at endogenous centromeres where CENP-ACnp1 is normally assembled. In contrast, Clr6 complex II (Clr6-CII; equivalent to Rpd3S) histone deacetylase function has a more subtle impact on the stability of transcribed H3 chromatin and acts to prevent the ectopic accumulation of CENP-ACnp1 at specific loci, including subtelomeric regions, where CENP-ACnp1 is preferentially assembled. Moreover, defective Clr6-CII function allows the de novo assembly of CENP-ACnp1 chromatin on centromeric DNA, bypassing the normal requirement for heterochromatin. Thus, our analyses show that alterations in the process of chromatin assembly during transcription can destabilize H3 nucleosomes and thereby allow CENP-ACnp1 to assemble in its place. We propose that normal centromeres provide a specific chromatin context that limits reassembly of H3 chromatin during transcription and thereby promotes the establishment of CENP-ACnp1 chromatin and associated kinetochores. These findings have important implications for genetic and epigenetic processes involved in centromere specification.