Insights into assembly and regulation of centromeric chromatin in Saccharomyces cerevisiae

At the core of chromosome segregation is the centromere, which nucleates the assembly of a macromolecular kinetochore (centromere DNA and associated proteins) complex responsible for mediating spindle attachment. Recent advances in centromere research have led to identification of many kinetochore components, such as the centromeric-specific histone H3 variant, CenH3, and its interacting partner, Scm3. Both are essential for chromosome segregation and are evolutionarily conserved from yeast to humans. CenH3 is proposed to be the epigenetic mark that specifies centromeric identity. Molecular mechanisms that regulate the assembly of kinetochores at specific chromosomal sites to mediate chromosome segregation are not fully understood. In this review, we summarize the current literature and discuss results from our laboratory, which show that restricting the localization of budding yeast CenH3, Cse4, to centromeres and balanced stoichiometry between Scm3 and Cse4, contribute to faithful chromosome transmission. We highlight our findings that, similar to other eukaryotic centromeres, budding yeast centromeric histone H4 is hypoacetylated, and we discuss how altered histone acetylation affects chromosome segregation. This article is part of a Special Issue entitled: Chromatin in time and space.     

Insights into assembly and regulation of centromeric chromatin in Saccharomyces cerevisiae

At the core of chromosome segregation is the centromere, which nucleates the assembly of a macromolecular kinetochore (centromere DNA and associated proteins) complex responsible for mediating spindle attachment. Recent advances in centromere research have led to identification of many kinetochore components, such as the centromeric-specific histone H3 variant, CenH3, and its interacting partner, Scm3. Both are essential for chromosome segregation and are evolutionarily conserved from yeast to humans. CenH3 is proposed to be the epigenetic mark that specifies centromeric identity. Molecular mechanisms that regulate the assembly of kinetochores at specific chromosomal sites to mediate chromosome segregation are not fully understood. In this review, we summarize the current literature and discuss results from our laboratory, which show that restricting the localization of budding yeast CenH3, Cse4, to centromeres and balanced stoichiometry between Scm3 and Cse4, contribute to faithful chromosome transmission. We highlight our findings that, similar to other eukaryotic centromeres, budding yeast centromeric histone H4 is hypoacetylated, and we discuss how altered histone acetylation affects chromosome segregation. This article is part of a Special Issue entitled: Chromatin in time and space.     

Nonhistone Scm3 binds to AT-rich DNA to organize atypical centromeric nucleosome of budding yeast

The molecular architecture of centromere-specific nucleosomes containing histone variant CenH3 is controversial. We have biochemically reconstituted two distinct populations of nucleosomes containing Saccharomyces cerevisiae CenH3 (Cse4). Reconstitution of octameric nucleosomes containing histones Cse4/H4/H2A/H2B is robust on noncentromere DNA, but inefficient on AT-rich centromere DNA. However, nonhistone Scm3, which is required for Cse4 deposition in?vivo, facilitates in?vitro reconstitution of Cse4/H4/Scm3 complexes on AT-rich centromere sequences. Scm3 has a nonspecific DNA binding domain that shows preference for AT-rich DNA and a histone chaperone domain that promotes specific loading of Cse4/H4. In live cells, Scm3-GFP is enriched at centromeres in all cell cycle phases. Chromatin immunoprecipitation confirms that Scm3 occupies centromere DNA throughout the cell cycle, even when Cse4 and H4 are temporarily dislodged in S phase. These findings suggest a model in which centromere-bound Scm3 aids recruitment of Cse4/H4 to assemble and maintain an H2A/H2B-deficient centromeric nucleosome.     

Misregulation of Scm3p/HJURP causes chromosome instability in Saccharomyces cerevisiae and human cells

The kinetochore (centromeric DNA and associated proteins) is a key determinant for high fidelity chromosome transmission. Evolutionarily conserved Scm3p is an essential component of centromeric chromatin and is required for assembly and function of kinetochores in humans, fission yeast, and budding yeast. Overexpression of HJURP, the mammalian homolog of budding yeast Scm3p, has been observed in lung and breast cancers and is associated with poor prognosis; however, the physiological relevance of these observations is not well understood. We overexpressed SCM3 and HJURP in Saccharomyces cerevisiae and HJURP in human cells and defined domains within Scm3p that mediate its chromosome loss phenotype. Our results showed that the overexpression of SCM3 (GALSCM3) or HJURP (GALHJURP) caused chromosome loss in a wild-type yeast strain, and overexpression of HJURP led to mitotic defects in human cells. GALSCM3 resulted in reduced viability in kinetochore mutants, premature separation of sister chromatids, and reduction in Cse4p and histone H4 at centromeres. Overexpression of CSE4 or histone H4 suppressed chromosome loss and restored levels of Cse4p at centromeres in GALSCM3 strains. Using mutant alleles of scm3, we identified a domain in the N-terminus of Scm3p that mediates its interaction with CEN DNA and determined that the chromosome loss phenotype of GALSCM3 is due to centromeric association of Scm3p devoid of Cse4p/H4. Furthermore, we determined that similar to other systems the centromeric association of Scm3p is cell cycle regulated. Our results show that altered stoichiometry of Scm3p/HJURP, Cse4p, and histone H4 lead to defects in chromosome segregation. We conclude that stringent regulation of HJURP and SCM3 expression are critical for genome stability.     

The CENP-A chaperone Scm3 becomes enriched at kinetochores in anaphase independently of CENP-A incorporation

Centromeres are specialized chromatin domains where kinetochores assemble. Centromeres contain as a conserved feature nucleosomes that are composed of the canonical histones H2A, H2B and H4 and a centromere-specific histone H3 variant, known as CENP-A in humans and Cse4 in budding yeast. The incorporation of CENP-A homologs into centromeric chromatin is cell cycle regulated and is assisted by related assembly factors named Scm3 in yeast and HJURP in human cells. Here, we describe that the budding yeast Scm3 binds weakly to centromeres during interphase including S phase when Cse4 assembles into centromeres. In anaphase Scm3 then becomes 2.5-fold enriched at kinetochores where it is dynamic with a half recovery time t_½ of 36 sec. In contrast, Cse4 is stably integrated into kinetochores. In addition, ten Scm3 molecules bind to a cluster of 16 kinetochores with 32 Cse4 molecules suggesting a 1:3 ratio at kinetochores between the two proteins. Analysis of conditional lethal scm3–1 mutant cells indicated that Scm3 participates in maintaining Cse4 at centromeres in anaphase. Thus, Scm3 interacts transiently with kinetochores in anaphase where it safeguards Cse4 even after its S phase incorporation into centromeres.     

Scm3 is essential to recruit the histone h3 variant cse4 to centromeres and to maintain a functional kinetochore

The kinetochore is a complex multiprotein structure located at centromeres that is essential for proper chromosome segregation. Budding-yeast Cse4 is an essential evolutionarily conserved histone H3 variant recruited to the centromere by an unknown mechanism. We have identified Scm3, an inner kinetochore protein that immunopurifies with Cse4. Scm3 is essential for viability and localizes to all centromeres. Construction of a conditional SCM3 allele reveals that depletion results in metaphase arrest, with duplicated spindle poles, short spindles, and unequal DNA distribution. The metaphase arrest is mediated by the mitotic spindle checkpoint being dependent on Mad1 and the Aurora kinase B homolog Ipl1. Scm3 interacts with both Ndc10 and Cse4 and is essential to establish centromeric chromatin after DNA replication. In addition, Scm3 is required to maintain kinetochore function throughout the cell cycle. We propose a model in which Ndc10/Scm3 binds to centromeric DNA, which is in turn essential for targeting Cse4 to centromeres.     

Heterochromatin is required for normal distribution of Neurospora crassa CenH3

Centromeres serve as platforms for the assembly of kinetochores and are essential for nuclear division. Here we identified Neurospora crassa centromeric DNA by chromatin immunoprecipitation followed by high-throughput sequencing (ChIP-seq) of DNA associated with tagged versions of the centromere foundation proteins CenH3 (CENP-A) and CEN-C (CENP-C) and the kinetochore protein CEN-T (CENP-T). On each chromosome we found an ～150- to 300-kbp region of enrichment for all three proteins. These regions correspond to intervals predicted to be centromeric DNA by genetic mapping and DNA sequence analyses. By ChIP-seq we found extensive colocalization of CenH3, CEN-C, CEN-T, and histone H3K9 trimethylation (H3K9me3). In contrast, H3K4me2, which has been found at the cores of plant, fission yeast, Drosophila, and mammalian centromeres, was not enriched in Neurospora centromeric DNA. DNA methylation was most pronounced at the periphery of centromeric DNA. Mutation of dim-5, which encodes an H3K9 methyltransferase responsible for nearly all H3K9me3, resulted in altered distribution of CenH3-green fluorescent protein (GFP). Similarly, CenH3-GFP distribution was altered in the absence of HP1, the chromodomain protein that binds to H3K9me3. We conclude that eukaryotes with regional centromeres make use of different strategies for maintenance of CenH3 at centromeres, and we suggest a model in which centromere proteins nucleate at the core but require DIM-5 and HP1 for spreading.     

The Centromere Histone Is Conserved and Associated with Tandem Repeats Sharing a Conserved 19-bp Box in the Holocentromere of Meloidogyne Nematodes

Although centromeres have conserved function, centromere-specific histone H3 (CenH3) and centromeric DNA evolve rapidly. The centromere drive model explains this phenomenon as a consequence of the conflict between fast-evolving DNA and CenH3, suggesting asymmetry in female meiosis as a crucial factor. We characterized evolution of the CenH3 protein in three closely related, polyploid mitotic parthenogenetic species of the Meloidogyne incognita group, and in the distantly related meiotic parthenogen Meloidogyne hapla. We identified duplication of the CenH3 gene in a putative sexual ancestral Meloidogyne. We found that one CenH3 (αCenH3) remained conserved in all extant species, including in distant Meloidogyne hapla, whereas the other evolved rapidly and under positive selection into four different CenH3 variants. This pattern of CenH3 evolution in Meloidogyne species suggests the subspecialization of CenH3s in ancestral sexual species. Immunofluorescence performed on mitotic Meloidogyne incognita revealed a dominant role of αCenH3 on its centromere, whereas the other CenH3s have lost their function in mitosis. The observed αCenH3 chromosome distribution disclosed cluster-like centromeric organization. The ChIP-Seq analysis revealed that in M. incognita αCenH3-associated DNA dominantly comprises tandem repeats, composed of divergent monomers which share a completely conserved 19-bp long box. Conserved αCenH3-associated DNA is also confirmed in the related mitotic Meloidogyne incognita group species suggesting preservation of both centromere protein and DNA constituents. We hypothesize that the absence of centromere drive in mitosis might allow for CenH3 and its associated DNA to achieve an equilibrium in which they can persist for long periods of time.     

The overexpression of a Saccharomyces cerevisiae centromeric histone H3 variant mutant protein leads to a defect in kinetochore biorientation

Chromosomes segregate using their kinetochores, the specialized protein structures that are assembled on centromeric DNA and mediate attachment to the mitotic spindle. Because centromeric sequences are not conserved, centromere identity is propagated by an epigenetic mechanism. All eukaryotes contain an essential histone H3 variant (CenH3) that localizes exclusively to centromeres. Because CenH3 is required for kinetochore assembly and is likely to be the epigenetic mark that specifies centromere identity, it is critical to elucidate the mechanisms that assemble and maintain CenH3 exclusively at centromeres. To learn more about the functions and regulation of CenH3, we isolated mutants in the budding yeast CenH3 that are lethal when overexpressed. These CenH3 mutants fall into three unique classes: (I) those that localize to euchromatin but do not alter kinetochore function, (II) those that localize to the centromere and disrupt kinetochore function, and (III) those that no longer target to the centromere but still disrupt chromosome segregation. We found that a class III mutant is specifically defective in the ability of sister kinetochores to biorient and attach to microtubules from opposite spindle poles, indicating that CenH3 mutants defective in kinetochore biorientation can be obtained.     

Stretching the rules: monocentric chromosomes with multiple centromere domains

The centromere is a functional chromosome domain that is essential for faithful chromosome segregation during cell division and that can be reliably identified by the presence of the centromere-specific histone H3 variant CenH3. In monocentric chromosomes, the centromere is characterized by a single CenH3-containing region within a morphologically distinct primary constriction. This region usually spans up to a few Mbp composed mainly of centromere-specific satellite DNA common to all chromosomes of a given species. In holocentric chromosomes, there is no primary constriction; the centromere is composed of many CenH3 loci distributed along the entire length of a chromosome. Using correlative fluorescence light microscopy and high-resolution electron microscopy, we show that pea (Pisum sativum) chromosomes exhibit remarkably long primary constrictions that contain 3-5 explicit CenH3-containing regions, a novelty in centromere organization. In addition, we estimate that the size of the chromosome segment delimited by two outermost domains varies between 69 Mbp and 107 Mbp, several factors larger than any known centromere length. These domains are almost entirely composed of repetitive DNA sequences belonging to 13 distinct families of satellite DNA and one family of centromeric retrotransposons, all of which are unevenly distributed among pea chromosomes. We present the centromeres of Pisum as novel "meta-polycentric" functional domains. Our results demonstrate that the organization and DNA composition of functional centromere domains can be far more complex than previously thought, do not require single repetitive elements, and do not require single centromere domains in order to segregate properly. Based on these findings, we propose Pisum as a useful model for investigation of centromere architecture and the still poorly understood role of repetitive DNA in centromere evolution, determination, and function.     

Role of the conserved carboxy-terminal α-helix of Rad6p in ubiquitination and DNA repair

RAD6 in the yeast Saccharomyces cerevisiae encodes a ubiquitin-conjugating enzyme essential for DNA repair as well as for a number of other biological processes. It is believed that the functions of Rad6p require the ubiquitination of target proteins, but its substrates as well as other interacting proteins are largely unknown. Rad6p homologues of higher eukaryotes have a number of amino acid residues in the C-terminal α-helix, which are conserved from yeast to man but are absent from most other yeast ubiquitin-conjugating enzymes (Ubcs). This specific conservation suggests that the C-terminal a-helix is important for the unique activities of the Rad6p family of Ubcs. We have investigated the effects of mutating this highly conserved region on the ubiquitination of model substrates in vitro and on error-free DNA repair in vivo. C-terminal point and deletion mutants of Rad6p differentially affected its in vitro activity on various substrates, raising the possibility that Rad6p interacts with its substrates in vivo by similar mechanisms. The distal part of the C-terminal u-helix is also essential for error-free DNA repair in vivo. Overexpression of Rad18p, a single-stranded DNA-binding protein that also interacts with Rad6p, alleviates the DNA repair defects of the C-terminal α-helix mutants to different degrees. This indicates that the C-terminal α-helix of Rad6p mediates its interaction with Rad18p, an essential step in DNA repair. Models of Rad6p action propose that its ubiquitination function is followed by proteolysis of unknown ubiquitinated targets. Mutants affecting several functions of the 26S proteasome retain wild-type capacity for error-free DNA repair. This raises the possibility that ubiquitination by Rad6p in DNA repair does not target proteins for proteasomal degradation.     

Scm3 is a centromeric nucleosome assembly factor

The Cse4 nucleosome at each budding yeast centromere must be faithfully assembled each cell cycle to specify the site of kinetochore assembly and microtubule attachment for chromosome segregation. Although Scm3 is required for the localization of the centromeric H3 histone variant Cse4 to centromeres, its role in nucleosome assembly has not been tested. We demonstrate that Scm3 is able to mediate the assembly of Cse4 nucleosomes in vitro, but not H3 nucleosomes, as measured by a supercoiling assay. Localization of Cse4 to centromeres and the assembly activity depend on an evolutionarily conserved core motif in Scm3, but localization of the CBF3 subunit Ndc10 to centromeres does not depend on this motif. The centromere targeting domain of Cse4 is sufficient for Scm3 nucleosome assembly activity. Assembly does not depend on centromeric sequence. We propose that Scm3 plays an active role in centromeric nucleosome assembly.     

Centromere-Like Regions in the Budding Yeast Genome

Accurate chromosome segregation requires centromeres (CENs), the DNA sequences where kinetochores form, to attach chromosomes to microtubules. In contrast to most eukaryotes, which have broad centromeres, Saccharomyces cerevisiae possesses sequence-defined point CENs. Chromatin immunoprecipitation followed by sequencing (ChIP–Seq) reveals colocalization of four kinetochore proteins at novel, discrete, non-centromeric regions, especially when levels of the centromeric histone H3 variant, Cse4 (a.k.a. CENP-A or CenH3), are elevated. These regions of overlapping protein binding enhance the segregation of plasmids and chromosomes and have thus been termed Centromere-Like Regions (CLRs). CLRs form in close proximity to S. cerevisiae CENs and share characteristics typical of both point and regional CENs. CLR sequences are conserved among related budding yeasts. Many genomic features characteristic of CLRs are also associated with these conserved homologous sequences from closely related budding yeasts. These studies provide general and important insights into the origin and evolution of centromeres.     

Four hydrophobic amino acid residues in the C-terminal effector domain of the yeast Mig1p repressor are important for its in vivo activity

The Mig1 repressor is a zinc finger protein that mediates glucose repression in yeast. Previous work in Saccharomyces cerevisiae has shown that two domains in Mig1p are required for repression: the N-terminal zinc finger region and a C-terminal effector domain. Both domains are also conserved in Mig1p homologs from the distantly related yeasts Kluyveromyces lactis and K. marxianus, and these Mig1 proteins can fully replace the endogenous Mig1p in S. cerevisiae. We have now made a detailed analysis of the conserved C-terminal effector domain in Mig1p from K. marxianus, using expression in S. cerevisiae to monitor its function. First, a series of small deletions were made within the effector domain. Second, an alanine scan mutagenesis was carried out across the effector domain. Third, double, triple and quadruple mutants were made that affect certain residues within the effector domain. Our results show that four conserved residues within the effector domain, three leucines and one isoleucine, are particularly important for its function in vivo. The analysis further revealed that while the C-terminal effector domain of KmMig1p mediates a seven- to nine-fold repression of the reporter gene, a five- to sixfold residual effect also exists that is independent of the C-terminal effector domain. Similar results were obtained when the corresponding mutations were made in ScMig1p. Moreover, we found that mutations in these residues affect the interaction between Mig1p and the general corepressor subunit Cyc8p (Ssn6p). Modeling of the C-terminal effector domain using a protein of known structure suggests that it may be folded into an α-helix. Received: 30 March 1998 / Accepted: 18 August 1998     

Probing the Architecture of a Simple Kinetochore Using DNA–Protein Crosslinking

In budding yeast, accurate chromosome segregation requires that one and only one kinetochore assemble per chromosome. In this paper, we report the use of DNA–protein crosslinking and nondenaturing gel analysis to study the structure of CBF3, a four-protein complex that binds to the essential CDEIII region of Saccharomyces cerevisiae centromeres. We find that three subunits of CBF3 are in direct contact with CDEIII over a region of DNA that spans 80 bp. A highly asymmetric core complex containing p58^CTF13 p64^CEP3 and p110^NDC10 in direct contact with DNA forms at the genetically defined center of CDEIII. This core complex spans ~56 bp of CEN3. An extended complex comprising the core complex and additional DNA-bound p110^NDC10 also forms. It spans ~80 bp of DNA. CBF3 makes sequence-specific and -nonspecific contacts with DNA. Both contribute significantly to the energy of CBF3–DNA interaction. Moreover, important sequence-specific contacts are made with bases that are not conserved among yeast centromeres. These findings provide a foundation for understanding the organization of the CBF3–centromere complex, a structure that appears to initiate the formation of microtubule attachment sites at yeast kinetochores. These results also have implications for understanding centromere-binding proteins in higher cells.     

Dynamics of a novel centromeric histone variant CenH3 reveals the evolutionary ancestral timing of centromere biogenesis

The centromeric histone H3 variant (CenH3) serves to target the kinetochore to the centromeres and thus ensures correct chromosome segregation during mitosis and meiosis. The Dictyostelium H3-like variant H3v1 was identified as the CenH3 ortholog. Dictyostelium CenH3 has an extended N-terminal domain with no similarity to any other known proteins and a histone fold domain at its C-terminus. Within the histone fold, α-helix 2 (α2) and an extended loop 1 (L1) have been shown to be required for targeting CenH3 to centromeres. Compared to other known and putative CenH3 histones, Dictyostelium CenH3 has a shorter L1, suggesting that the extension is not an obligatory feature. Through ChIP analysis and fluorescence microscopy of live and fixed cells, we provide here the first survey of centromere structure in amoebozoa. The six telocentric centromeres were found to mostly consist of all the DIRS-1 elements and to associate with H3K9me3. During interphase, the centromeres remain attached to the centrosome forming a single CenH3-containing cluster. Loading of Dictyostelium CenH3 onto centromeres occurs at the G2/prophase transition, in contrast to the anaphase/telophase loading of CenH3 observed in metazoans. This suggests that loading during G2/prophase is the ancestral eukaryotic mechanism and that anaphase/telophase loading of CenH3 has evolved more recently after the amoebozoa diverged from the animal linage.     

The C-Terminal Domain of CENP-C Displays Multiple and Critical Functions for Mammalian Centromere Formation

CENP-C is a fundamental component of functional centromeres. The elucidation of its structure-function relationship with centromeric DNA and other kinetochore proteins is critical to the understanding of centromere assembly. CENP-C carries two regions, the central and the C-terminal domains, both of which are important for the ability of CENP-C to associate with the centromeric DNA. However, while the central region is largely divergent in CENP-C homologues, the C-terminal moiety contains two regions that are highly conserved from yeast to humans, named Mif2p homology domains (blocks II and III). The activity of these two domains in human CENP-C is not well defined. In this study we performed a functional dissection of C-terminal CENP-C region analyzing the role of single Mif2p homology domains through in vivo and in vitro assays. By immunofluorescence and Chromatin immunoprecipitation assay (ChIP) we were able to elucidate the ability of the Mif2p homology domain II to target centromere and contact alpha satellite DNA. We also investigate the interactions with other conserved inner kinetochore proteins by means of coimmunoprecipitation and bimolecular fluorescence complementation on cell nuclei. We found that the C-terminal region of CENP-C (Mif2p homology domain III) displays multiple activities ranging from the ability to form higher order structures like homo-dimers and homo-oligomers, to mediate interaction with CENP-A and histone H3. Overall, our findings support a model in which the Mif2p homology domains of CENP-C, by virtue of their ability to establish multiple contacts with DNA and centromere proteins, play a critical role in the structuring of kinethocore chromatin.     

The H3 histone chaperone NASPSIM3 escorts CenH3 in Arabidopsis

Centromeres define the chromosomal position where kinetochores form to link the chromosome to microtubules during mitosis and meiosis. Centromere identity is determined by incorporation of a specific histone H3 variant termed CenH3. As for other histones, escort and deposition of CenH3 must be ensured by histone chaperones, which handle the non‐nucleosomal CenH3 pool and replenish CenH3 chromatin in dividing cells. Here, we show that the Arabidopsis orthologue of the mammalian NUCLEAR AUTOANTIGENIC SPERM PROTEIN (NASP) and Schizosaccharomyces pombe histone chaperone Sim3 is a soluble nuclear protein that binds the histone variant CenH3 and affects its abundance at the centromeres. NASP^SIM3 is co‐expressed with Arabidopsis CenH3 in dividing cells and binds directly to both the N‐terminal tail and the histone fold domain of non‐nucleosomal CenH3. Reduced NASP^SIM3 expression negatively affects CenH3 deposition, identifying NASP^SIM3 as a CenH3 histone chaperone.     

Nonhistone Scm3 and histones CenH3-H4 assemble the core of centromere-specific nucleosomes

The budding yeast histone H3 variant, Cse4, replaces conventional histone H3 in centromeric chromatin and, together with centromere-specific DNA-binding factors, directs assembly of the kinetochore, a multiprotein complex mediating chromosome segregation. We have identified Scm3, a nonhistone protein that colocalizes with Cse4 and is required for its centromeric association. Bacterially expressed Scm3 binds directly to and reconstitutes a stoichiometric complex with Cse4 and histone H4 but not with conventional histone H3 and H4. A conserved acidic domain of Scm3 is responsible for directing the Cse4-specific interaction. Strikingly, binding of Scm3 can replace histones H2A-H2B from preassembled Cse4-containing histone octamers. This incompatibility between Scm3 and histones H2A-H2B is correlated with diminished in vivo occupancy of histone H2B, H2A, and H2AZ at centromeres. Our findings indicate that nonhistone Scm3 serves to assemble and maintain Cse4-H4 at centromeres and may replace histone H2A-H2B dimers in a centromere-specific nucleosome core.     

DNA deletion as a mechanism for developmentally programmed centromere loss

A hallmark of active centromeres is the presence of the histone H3 variant CenH3 in the centromeric chromatin, which ensures faithful genome distribution at each cell division. A functional centromere can be inactivated, but the molecular mechanisms underlying the process of centromere inactivation remain largely unknown. Here, we describe the loss of CenH3 protein as part of a developmental program leading to the formation of the somatic nucleus in the eukaryote Paramecium. We identify two proteins whose depletion prevents developmental loss of CenH3: the domesticated transposase Pgm involved in the formation of DNA double strand cleavages and the Polycomb-like lysine methyltransferase Ezl1 necessary for trimethylation of histone H3 on lysine 9 and lysine 27. Taken together, our data support a model in which developmentally programmed centromere loss is caused by the elimination of DNA sequences associated with CenH3.