H2A.Z maintenance during mitosis reveals nucleosome shifting on mitotically silenced genes.

Profound chromatin changes occur during mitosis to allow for gene silencing and chromosome segregation followed by reactivation of memorized transcription states in daughter cells. Using genome-wide sequencing, we found H2A.Z-containing +1 nucleosomes of active genes shift upstream to occupy TSSs during mitosis, significantly reducing nucleosome-depleted regions. Single-molecule analysis confirmed nucleosome shifting and demonstrated that mitotic shifting is specific to active genes that are silenced during mitosis and, thus, is not seen on promoters, which are silenced by methylation or mitotically expressed genes. Using the GRP78 promoter as a model, we found H3K4 trimethylation is also maintained while other indicators of active chromatin are lost and expression is decreased. These key changes provide a potential mechanism for rapid silencing and reactivation of genes during the cell cycle.

DZNep is a global histone methylation inhibitor that reactivates developmental genes not silenced by DNA methylation.

DNA methylation, histone modifications, and nucleosomal occupancy collaborate to cause silencing of tumor-related genes in cancer. The development of drugs that target these processes is therefore important for cancer therapy. Inhibitors of DNA methylation and histone deacetylation have been approved by the Food and Drug Administration for treatment of hematologic malignancies. However, drugs that target other mechanisms still need to be developed. Recently, 3-deazaneplanocin A (DZNep) was reported to selectively inhibit trimethylation of lysine 27 on histone H3 (H3K27me3) and lysine 20 on histone H4 (H4K20me3) as well as reactivate silenced genes in cancer cells. This finding opens the door to the pharmacologic inhibition of histone methylation. We therefore wanted to further study the mechanism of action of DZNep in cancer cells. Western blot analysis shows that DZNep globally inhibits histone methylation and is not selective. Two other drugs, sinefungin and adenosine dialdehyde, have similar effects as DZNep on H3K27me3. Intriguingly, chromatin immunoprecipitation of various histone modifications and microarray analysis show that DZNep acts through a different pathway than 5-aza-2'-deoxycytidine, a DNA methyltransferase inhibitor. These observations give us interesting insight into how chromatin structure affects gene expression. We also determined the kinetics of gene activation to understand if the induced changes were somatically heritable. We found that upon removal of DZNep, gene expression is reduced to its original state. This suggests that there is a homeostatic mechanism that returns the histone modifications to their "ground state" after DZNep treatment. Our data show the strong need for further development of histone methylation inhibitors.

DNA methylation: the nuts and bolts of repression.

DNA methylation is an epigenetic modification which plays an important role in chromatin organization and gene expression. DNA methylation can silence genes and repetitive elements through a process which leads to the alteration of chromatin structure. The mechanisms which target DNA methylation to specific sites in the genome are not fully understood. In this review, we will discuss the mechanisms which lead to the long-term silencing of genes and will survey the progression that has been made in determining the targeted mechanisms for de novo DNA methylation.

Yeast Hsl7 (histone synthetic lethal 7) catalyses the in vitro formation of omega-N(G)-monomethylarginine in calf thymus histone H2A.

The HSL7 (histone synthetic lethal 7) gene in the yeast Saccharomyces cerevisiae encodes a protein with close sequence similarity to the mammalian PRMT5 protein, a member of the class of protein arginine methyltransferases that catalyses the formation of omega-N(G)-monomethylarginine and symmetric omega-N(G),N'(G)-dimethylarginine residues in a number of methyl-accepting species. A full-length HSL7 construct was expressed as a FLAG-tagged protein in Saccharomyces cerevisiae. We found that FLAG-tagged Hsl7 effectively catalyses the transfer of methyl groups from S-adenosyl-[methyl-3H]-L-methionine to calf thymus histone H2A. When the acid-hydrolysed radiolabelled protein products were separated by high-resolution cation-exchange chromatography, we were able to detect one tritiated species that co-migrated with an omega-N(G)-monomethylarginine standard. No radioactivity was observed that co-migrated with either the asymmetric or symmetric dimethylated derivatives. In control experiments, no methylation of histone H2A was found with two mutant constructs of Hsl7. Surprisingly, FLAG-Hsl7 does not appear to effectively catalyse the in vitro methylation of a GST (glutathione S-transferase)-GAR [glycine- and arginine-rich human fibrillarin-(1-148) peptide] fusion protein or bovine brain myelin basic protein, both good methyl-accepting substrates for the human homologue PRMT5. Additionally, FLAG-Hsl7 demonstrates no activity on purified calf thymus histones H1, H2B, H3 or H4. GST-Rmt1, the GST-fusion protein of the major yeast protein arginine methyltransferase, was also found to methylate calf thymus histone H2A. Although we detected Rmt1-dependent arginine methylation in vivo in purified yeast histones H2A, H2B, H3 and H4, we found no evidence for Hsl7-dependent methylation of endogenous yeast histones. The physiological substrates of the Hsl7 enzyme remain to be identified.

Protein arginine methyltransferase 6 specifically methylates the nonhistone chromatin protein HMGA1a.

The HMGA family proteins HMGA1a and HMGA1b are nuclear nonhistone species implicated in a wide range of cellular processes including inducible gene transcription, modulation of chromosome structure through nucleosome and chromosome remodeling, and neoplastic transformation. HMGA proteins are highly modified, and changes in their phosphorylation states have been correlated with the phase of the cell cycle and changes in their transcriptional activity. HMGA1a is also methylated in the first DNA-binding AT-hook at Arg25 and other sites, although the enzyme or enzymes responsible have not been identified. We demonstrate here that a GST fusion of protein arginine methyltransferase 6 (PRMT6) specifically methylates full-length recombinant HMGA1a protein in vitro. Although GST fusions of PRMT1 and PRMT3 were also capable of methylating the full-length HMGA1a polypeptide, they recognize its proteolytic degradation products much better. GST fusions of PRMT4 or PRMT7 were unable to methylate the full-length protein or its degradation products. We conclude that PRMT6 is a good candidate for the endogenous enzyme responsible for HGMA1a methylation.

A novel SET domain methyltransferase modifies ribosomal protein Rpl23ab in yeast.

In vivo studies have shown that the ribosomal large subunit protein L23a (Rpl23ab) in Saccharomyces cerevisiae is methylated at lysine residues. However, the gene encoding the methyltransferase responsible for the modification has not been identified. We show here that the yeast YPL208w gene product, a member of the SET domain family of methyltransferases, catalyzes the reaction, and we have now designated it Rkm1 (ribosomal lysine (K) methyltransferase 1). Yeast strains with deletion mutations in candidate SET domain-containing genes were in vivo labeled with S-adenosyl-l-[methyl-(3)H]methionine. [(3)H]Methyl radioactivity was determined after lysates were fractionated by SDS gel electrophoresis. When compared with the parent strain or other candidate deletion strains, a loss of a radiolabeled 15-kDa species was observed in the rkm1 (Deltaypl208w) knock-out strain. Treatment of wild-type cell extracts with RNase or proteinase K demonstrated that the methyl-accepting substrate is a protein. Cellular lysates from parent and knockout strains were fractionated using high salt sucrose gradients. Analysis of the gradient fractions by SDS gel electrophoresis demonstrated that the 15-kDa methyl-accepting substrate elutes with the large ribosomal subunit. In vitro methylation experiments using purified ribosomes confirmed that the methyl-accepting substrate is a ribosomal protein. Amino acid analysis of the in vivo labeled 15 kDa polypeptide showed that it contains epsilon-[(3)H]dimethyllysine residues. Mass spectrometry of tryptic peptides of the 15 kDa polypeptide identified it as Rpl23ab. Analysis of the intact masses of the large ribosomal subunit proteins by electrospray mass spectrometry confirmed that the substrate is Rpl23ab and that it is specifically dimethylated at two distinct sites by Rkm1. These results show that SET domain methyltransferases can be involved in translational roles as well as in the previously described transcriptional roles.

A new type of protein methylation activated by tyrphostin A25 and vanadate.

It has been reported that S-adenosylmethionine-dependent protein methylation in rat kidney extracts can be greatly stimulated by tyrphostin A25, a tyrosine kinase inhibitor. We have investigated the nature of this stimulation. We find that addition of tyrphostin A25, in combination with the protein phosphatase inhibitor vanadate, leads to the stimulation of methylation of polypeptides of 64, 42, 40, 36, 31, and 15 kDa in cytosolic extracts of mouse kidney. The effect of tyrphostin appears to be relatively specific for the A25 species. The enhanced methylation does not represent the activity of the families of protein histidine, lysine or arginine methyltransferases, nor that of the l-isoaspartyl/d-aspartyl methyltransferase, enzymes responsible for the bulk of protein methylation in most cell types. Chemical and enzymatic analyses of the methylated polypeptides suggest that the methyl group is in an ester linkage to the protein. In heart extracts, we find a similar situation but here the stimulation of methylation is not dependent upon vanadate and an additional 18 kDa methylated species is found. In contrast, little or no stimulation of methylation is found in brain or testis extracts. This work provides evidence for a novel type of protein carboxyl methylation reaction that may play a role in signaling reactions in certain mammalian tissues.

Spliceosome Sm proteins D1, D3, and B/B' are asymmetrically dimethylated at arginine residues in the nucleus.

We report a novel modification of spliceosome proteins Sm D1, Sm D3, and Sm B/B'. L292 mouse fibroblasts were labeled in vivo with [3H]methionine. Sm D1, Sm D3, and Sm B/B' were purified from either nuclear extracts, cytosolic extracts or a cytosolic 6S complex by immunoprecipitation of the Sm protein-containing complexes and then separation by electrophoresis on a polyacrylamide gel containing urea. The isolated Sm D1, Sm D3 or Sm B/B' proteins were hydrolyzed to amino acids and the products were analyzed by high-resolution cation exchange chromatography. Sm D1, Sm D3, and Sm B/B' isolated from nuclear fractions were all found to contain omega-NG-monomethylarginine and symmetric omega-NG,NG'-dimethylarginine, modifications that have been previously described. In addition, Sm D1, Sm D3, and Sm B/B' were also found to contain asymmetric omega-NG,NG-dimethylarginine in these nuclear fractions. Analysis of Sm B/B' from cytosolic fractions and Sm B/B' and Sm D1 from cytosolic 6S complexes showed only the presence of omega-NG-monomethylarginine and symmetric omega-NG,NG'-dimethylarginine. These results indicate that Sm D1, Sm D3, and Sm B/B' are asymmetrically dimethylated and that these modified proteins are located in the nucleus. In reactions in which Sm D1 or Sm D3 was methylated in vitro with a hemagglutinin-tagged PRMT5 purified from HeLa cells, we detected both symmetric omega-NG,NG'-dimethylarginine and asymmetric omega-NG,NG-dimethylarginine when reactions were done in a Tris/HCl buffer, but only detected symmetric omega-NG,NG'-dimethylarginine when a sodium phosphate buffer was used. These results suggest that the activity responsible for the formation of asymmetric dimethylated arginine residues in Sm proteins is either PRMT5 or a protein associated with it in the immunoprecipitated complex.

DAL-1/4.1B tumor suppressor interacts with protein arginine N-methyltransferase 3 (PRMT3) and inhibits its ability to methylate substrates in vitro and in vivo.

DAL-1 (differentially expressed in adenocarcinoma of the lung)/4.1B is a tumor suppressor gene on human chromosome 18p11.3 whose expression is lost in >50% of primary non-small-cell lung carcinomas. Based on sequence similarity, DAL-1/4.1B has been assigned to the Protein 4.1 superfamily whose members interact with plasma membrane proteins through their N-terminal FERM (4.1/Ezrin/Radixin/Moesin) domain, and cytoskeletal components via their C-terminal SAB (spectrin-actin binding) region. Using the DAL-1/4.1B FERM domain as bait for yeast two-hybrid interaction cloning, we identified protein arginine N-methyltransferase 3 (PRMT3) as a specific DAL-1/4.1B-interacting protein. PRMT3 catalyses the post-translational transfer of methyl groups from S-adenosyl-L-methionine to arginine residues of proteins. Coimmunoprecipitation experiments using lung and breast cancer cell lines confirmed this interaction in mammalian cells in vivo. In vitro binding assays demonstrated that this was an interaction occurring via the C-terminal catalytic core domain of PRMT3. DAL-1/4.1B was determined not to be a substrate for PRMT3-mediated methylation but its presence inhibits the in vitro methylation of a glycine-rich and arginine-rich methyl-accepting protein, GST (glutathione-S-transferase-GAR (glycine- and arginine-rich), which contains 14 'RGG' consensus methylation sites. In addition, induced expression of DAL-1/4.1B in MCF-7 breast cancer cells showed that the DAL-1/4.1B protein significantly inhibits PRMT3 methylation of cellular substrates. These findings suggest that modulation of post-translational methylation may be an important mechanism through which DAL-1/4.1B affects tumor cell growth.

PRMT7 is a member of the protein arginine methyltransferase family with a distinct substrate specificity.

We have identified a mammalian arginine N-methyltransferase, PRMT7, that can catalyze the formation of omega-NG-monomethylarginine in peptides. This protein is encoded by a gene on human chromosome 16q22.1 (human locus AK001502). We expressed a full-length human cDNA construct in Escherichia coli as a glutathione S-transferase (GST) fusion protein. We found that GST-tagged PRMT7 catalyzes the S-adenosyl-[methyl-3H]-l-methionine-dependent methylation of the synthetic peptide GGPGGRGGPGG-NH2 (R1). The radiolabeled peptide was purified by high-pressure liquid chromatography and acid hydrolyzed to free amino acids. When the hydrolyzed products were separated by high-resolution cation-exchange chromatography, we were able to detect one tritiated species which co-migrated with an omega-NG-monomethylarginine standard. Surprisingly, GST-PRMT7 was not able to catalyze the in vitro methylation of a GST-fibrillarin (amino acids 1-148) fusion protein (GST-GAR), a methyl-accepting substrate for the previously characterized PRMT1, PRMT3, PRMT4, PRMT5, and PRMT6 enzymes. Nor was it able to methylate myelin basic protein or histone H2A, in vitro substrates of PRMT5. This specificity distinguishes PRMT7 from all of the other known arginine methyltransferases. An additional unique feature of PRMT7 is that it seems to have arisen from a gene duplication event and contains two putative AdoMet-binding motifs. To see if both motifs were necessary for activity, each putative domain was expressed as a GST-fusion and tested for activity with peptides R1 and R2 (acetyl-GGRGG-NH2). These truncated proteins were enzymatically inactive, suggesting that both domains are required for functionality.

Characterization of the Drosophila protein arginine methyltransferases DART1 and DART4.

The role of arginine methylation in Drosophila melanogaster is unknown. We identified a family of nine PRMTs (protein arginine methyltransferases) by sequence homology with mammalian arginine methyltransferases, which we have named DART1 to DART9 ( Drosophila arginine methyltransferases 1-9). In keeping with the mammalian PRMT nomenclature, DART1, DART4, DART5 and DART7 are the putative homologues of PRMT1, PRMT4, PRMT5 and PRMT7. Other DART family members have a closer resemblance to PRMT1, but do not have identifiable homologues. All nine genes are expressed in Drosophila at various developmental stages. DART1 and DART4 have arginine methyltransferase activity towards substrates, including histones and RNA-binding proteins. Amino acid analysis of the methylated arginine residues confirmed that both DART1 and DART4 catalyse the formation of asymmetrical dimethylated arginine residues and they are type I arginine methyltransferases. The presence of PRMTs in D. melanogaster suggest that flies are a suitable genetic system to study arginine methylation.