Loss of Ezh2 function remodels the DNA replication initiation landscape.

In metazoan cells, DNA replication initiates from thousands of genomic loci scattered throughout the genome called DNA replication origins. Origins are strongly associated with euchromatin, particularly open genomic regions such as promoters and enhancers. However, over a third of transcriptionally silent genes are associated with DNA replication initiation. Most of these genes are bound and repressed by the Polycomb repressive complex-2 (PRC2) through the repressive H3K27me3 mark. This is the strongest overlap observed for a chromatin regulator with replication origin activity. Here, we asked whether Polycomb-mediated gene repression is functionally involved in recruiting DNA replication origins to transcriptionally silent genes. We show that the absence of EZH2, the catalytic subunit of PRC2, results in increased DNA replication initiation, specifically in the vicinity of EZH2 binding sites. The increase in DNA replication initiation does not correlate with transcriptional de-repression or the acquisition of activating histone marks but does correlate with loss of H3K27me3 from bivalent promoters.

The Efficiency of Global Genome-Nucleotide Excision Repair is Linked to the Fraction of Open rRNA Gene Chromatin, in Yeast.

The yeast rDNA locus is a suitable model to study nucleotide excision repair (NER) in chromatin. A portion of rRNA genes is transcribed and largely depleted of nucleosomes, the remaining genes are not transcribed and folded in nucleosomes. In G1-arrested cells, most rRNA genes do not have nucleosomes. TC-NER removes UV-induced DNA lesions from the transcribed strand of active genes. GG-NER is less efficient and removes DNA lesions from the nontranscribed strand of active genes and from the inactive genome. Different from mammalian cells, in yeast, the rRNA gene-transcribed strand is repaired by RNA polymerase-I-dependent TC-NER. The opposite nontranscribed strand is repaired faster than both strands of inactive rRNA genes. In log-phase cells, RNA polymerase-I are dislodged from the damaged transcribed strand and partially replaced by nucleosomes. Contrary to log-phase cells, in G1-phase cells few, if any, histones are deposited on the open rRNA genes during NER. In this study, we compared GG-NER efficiency in the rRNA gene coding region: without nucleosomes, partially loaded or wholly loaded with nucleosomes. The results indicate that in log-phase cells histones obstruct GG-NER, whereas in G1-phase cells GG-NER is as efficient as TC-NER.

Analyses of rRNA gene chromatin in cell cycle arrested Saccharomyces cerevisiae cells.

The existence of two chromatin structures in the rDNA locus was previously demonstrated for a large variety of organisms, ranging from yeast to human. In yeast there are about 150-200 rRNA genes organized in tandem repeats. Almost half of them are transcribed and largely depleted of nucleosomes (active/open), the other half is not transcribed and is assembled in regular arrays of nucleosomes (inactive/closed). It is proposed that RNA polymerase-I (RNAPI) transcription-elongation removes nucleosomes from closed rRNA genes (opening), and that soon after DNA replication there is deposition of nucleosomes on the open rRNA genes (closing). In G1 arrested cells, nearly all rRNA genes are depleted of nucleosomes, but most of them are not transcribed (inactive/open). In relation to the research article by Charton et al. (Mutat. Res.), the data presented here are on the hydroxyurea concentration-dependent inhibition of yeast culture growth, on cell cycle arrest before completion of genome replication, and on the opening of rRNA gene chromatin. As comparison, data are presented for yeast arrested in the G1-phase of the cell cycle by the pheromone α-factor.

In yeast cells arrested at the early S-phase by hydroxyurea, rRNA gene promoters and chromatin are poised for transcription while rRNA synthesis is compromised.

Hydroxyurea (HU) is an inhibitor of ribonucleotide reductase that is used as a chemotherapeutic agent to treat a number of chronic diseases. Addition of HU to cell cultures causes reduction of the dNTP cellular pool below levels that are required for DNA replication. This trigger dividing cells to arrest in early S-phase of the cell cycle. Cell division hinges on ribosome biogenesis, which is tightly regulated by rRNA synthesis. Remarkably, HU represses the expression of some genes the products of which are required for rRNA maturation. To gain more information on the cellular response to HU, we employed the yeast Saccharomyces cerevisiae as model organism and analyzed the changing aspects of closed to open forms of rRNA gene chromatin during cell cycle arrest, the arrangement of RNA polymerase-I (RNAPI) on the open genes, the presence of RNAPI transcription-factors, transcription and rRNA maturation. The rRNA gene chromatin structure was analyzed by psoralen crosslinking and the distribution of RNAPI was investigated by chromatin endogenous cleavage. In HU arrested cells nearly all rRNA genes were in the open form of chromatin, but only a portion of them was engaged with RNAPI. Analyses by chromatin immuno-precipitation confirmed that the overall formation of transcription pre-initiation complexes remained unchanged, suggesting that the onset of rRNA gene activation was not significantly affected by HU. Moreover, the in vitro transcription run-on assay indicated that RNAPI retained most of its transcription elongation activity. However, in HU treated cells, we found that: (1) RNAPI accumulated next to the 5'-end of rRNA genes; (2) considerably less rRNA filaments were observed in electron micrographs of rDNA transcription units; and (3) rRNA maturation was compromised. It is established that HU inhibition of ribonucleotide reductase holds back DNA replication. This study indicates a hitherto unexplored cellular response to HU, namely altered rRNA synthesis, which could participate to hamper cell division.

RNA Polymerase-I-Dependent Transcription-coupled Nucleotide Excision Repair of UV-Induced DNA Lesions at Transcription Termination Sites, in Saccharomyces cerevisiae.

If not repaired, ultraviolet light-induced DNA damage can lead to genome instability. Nucleotide excision repair (NER) of UV photoproducts is generally fast in the coding region of genes, where RNA polymerase-II (RNAP2) arrest at damage sites and trigger transcription-coupled NER (TC-NER). In Saccharomyces cerevisiae, there is RNA polymerase-I (RNAP1)-dependent TC-NER, but this process remains elusive. Therefore, we wished to characterize TC-NER efficiency in different regions of the rDNA locus: where RNAP1 are present at high density and start transcription elongation, where the elongation rate is slow, and in the transcription terminator where RNAP1 pause, accumulate and then are released. The Rpa12 subunit of RNAP1 and the Nsi1 protein participate in transcription termination, and NER efficiency was compared between wild type and cells lacking Rpa12 or Nsi1. The presence of RNAP1 was determined by chromatin endogenous cleavage and chromatin immunoprecipitation, and repair was followed at nucleotide precision with an assay that is based on the blockage of Taq polymerase by UV photoproducts. We describe that TC-NER, which is modulated by the RNAP1 level and elongation rate, ends at the 35S rRNA gene transcription termination site.

Nucleosome positioning, nucleotide excision repair and photoreactivation in Saccharomyces cerevisiae.

The position of nucleosomes on DNA participates in gene regulation and DNA replication. Nucleosomes can be repressors by limiting access of factors to regulatory sequences, or activators by facilitating binding of factors to exposed DNA sequences on the surface of the core histones. The formation of UV induced DNA lesions, like cyclobutane pyrimidine dimers (CPDs), is modulated by DNA bending around the core histones. Since CPDs are removed by nucleotide excision repair (NER) and photolyase repair, it is of paramount importance to understand how DNA damage and repair are tempered by the position of nucleosomes. In vitro, nucleosomes inhibit NER and photolyase repair. In vivo, nucleosomes slow down NER and considerably obstruct photoreactivation of CPDs. However, over-expression of photolyase allows repair of nucleosomal DNA in a second time scale. It is proposed that the intrinsic abilities of nucleosomes to move and transiently unwrap could facilitate damage recognition and repair in nucleosomal DNA.

Repair of UV induced DNA lesions in ribosomal gene chromatin and the role of "Odd" RNA polymerases (I and III).

In fast growing eukaryotic cells, a subset of rRNA genes are transcribed at very high rates by RNA polymerase I (RNAPI). Nuclease digestion-assays and psoralen crosslinking have shown that they are open; that is, largely devoid of nucleosomes. In the yeast Saccharomyces cerevisae, nucleotide excision repair (NER) and photolyase remove UV photoproducts faster from open rRNA genes than from closed and nucleosome-loaded inactive rRNA genes. After UV irradiation, rRNA transcription declines because RNAPI halt at UV photoproducts and are then displaced from the transcribed strand. When the DNA lesion is quickly recognized by NER, it is the sub-pathway transcription-coupled TC-NER that removes the UV photoproduct. If dislodged RNAPI are replaced by nucleosomes before NER recognizes the lesion, then it is the sub-pathway global genome GG-NER that removes the UV photoproducts from the transcribed strand. Also, GG-NER maneuvers in the non-transcribed strand of open genes and in both strands of closed rRNA genes. After repair, transcription resumes and elongating RNAPI reopen the rRNA gene. In higher eukaryotes, NER in rRNA genes is inefficient and there is no evidence for TC-NER. Moreover, TC-NER does not occur in RNA polymerase III transcribed genes of both, yeast and human fibroblast.

Expression analysis of ATAD3 isoforms in rodent and human cell lines and tissues.

ATAD3 (ATPase family AAA-Domain containing protein 3) is a mitochondrial inner membrane ATPase with unknown but vital functions. Initial researches have focused essentially on the major p66-ATAD3 isoform, but other proteins and mRNAs are described in the data banks. Using a set of anti-peptide antibodies and by the use of rodent and human cell lines and organs, we tried to detail ATAD3 gene expression profiles and to verify the existence of the various ATAD3 isoforms. In rodent, the single ATAD3 gene is expressed as a major isoform of 67 kDa, (ATAD3l; long), in all cells and organs studied. A second isoform, p57-ATAD3s (small), is expressed specifically throughout brain development and in adult, and overexpressed around the peri-natal period. p57-ATAD3s is also expressed in neuronal and glial rodent cell lines, and during in vitro differentiation of primary cultured rat oligodendrocytes. Other smaller isoforms were also detected in a tissue-specific manner. In human and primates, ATAD3 paralogues are encoded by three genes (ATAD3A, 3B and 3C), each of them presenting several putative variants. Analyzing the expression of ATAD3A and ATAD3B with four specific anti-peptide antibodies, and comparing their expressions with in vitro expressed ATAD3 cDNAs, we were able to observe and define five isoforms. In particular, the previously described p72-ATAD3B is confirmed to be in certain cases a phosphorylated form of ATAD3As. Moreover, we observed that the ATAD3As phosphorylation level is regulated by insulin and serum. Finally, exploring ATAD3 mRNA expression, we confirmed the existence of an alternative splicing in rodent and of several mRNA isoforms in human. Considering these observations, we propose the development of a uniform denomination for ATAD3 isoforms in rodent and human.

UV light-induced DNA lesions cause dissociation of yeast RNA polymerases-I and establishment of a specialized chromatin structure at rRNA genes.

The cytotoxicity of UV light-induced DNA lesions results from their interference with transcription and replication. DNA lesions arrest elongating RNA polymerases, an event that triggers transcription-coupled nucleotide excision repair. Since arrested RNA polymerases reduce the accessibility of repair factors to DNA lesions, they might be displaced. The fate of arrested RNA polymerases-II at DNA lesions has been extensively studied, yielding partially contradictory results. Considerably less is known about RNA polymerases-I that transcribe nucleosomes-depleted rRNA genes at very high rate. To investigate the fate of arrested RNA polymerases-I at DNA lesions, chromatin-immunoprecipitation, electron microscopy, transcription run-on, psoralen-cross-linking and chromatin-endogenous cleavage were employed. We found that RNA polymerases-I density increased at the 5'-end of the gene, likely due to continued transcription initiation followed by elongation and pausing/release at the first DNA lesion. Most RNA polymerases-I dissociated downstream of the first DNA lesion, concomitant with chromatin closing that resulted from deposition of nucleosomes. Although nucleosomes were deposited, the high mobility group-box Hmo1 (component of actively transcribed rRNA genes) remained associated. After repair of DNA lesions, Hmo1 containing chromatin might help to restore transcription elongation and reopening of rRNA genes chromatin.

Chromatin endogenous cleavage and psoralen crosslinking assays to analyze rRNA gene chromatin in vivo.

In eukaryotes, multiple copies of ribosomal RNA (rRNA) genes co-exist in two different chromatin states: actively transcribed (nucleosome depleted) chromatin, and nontranscribed (nucleosomal) chromatin. The presence of two rRNA gene populations compromises the interpretation of analyses obtained by the standard biochemical methods that are used to study chromatin structure (e.g., nuclease digestion and chromatin immunoprecipitation). Here, we provide a protocol to investigate the specific association of proteins with the two rRNA gene chromatin populations in vivo, using Saccharomyces cerevisiae as a model eukaryote.

Topological analysis of ATAD3A insertion in purified human mitochondria.

ATAD3 is a mitochondrial inner membrane-associated protein that has been predicted to be an ATPase but from which no associated function is known. The topology of ATAD3 in mitochondrial membranes is not clear and subject to controversy. A direct interaction of the N-terminal domain (amino-acids 44-247) with the mtDNA has been described, but the same domain has been reported to be sensitive to limited proteolysis in purified mitochondria. Furthermore, ATAD3 has been found in a large purified nucleoid complex but could not be cross-linked to the nucleoid. To resolve these discrepancies we used two immunological approaches to test whether the N-terminal (amino-acids 40-53) and the C-terminal (amino-acids 572-586) regions of ATAD3 are accessible from the cytosol. Using N-terminal and C-terminal specific anti-peptide antibodies, we carried out back-titration ELISA measurements and immuno-fluorescence analysis on freshly purified human mitochondria. Both approaches showed that the N-terminal region of ATAD3A is accessible to antibodies in purified mitochondria. The N-terminal region of ATAD3A is thus probably in the cytoplasm or in an accessible intermembrane space. On the contrary, the C-terminal region is not accessible to the antibody and is probably located within the matrix. These results demonstrate both that the N-terminal part of ATAD3A is outside the inner membrane and that the C-terminal part is inside the matrix.