Methylation of simian virus 40 Hpa II site affects late, but not early, viral gene expression

Regulation of Polyomavirus Transcription by Viral and Cellular Factors

Polyomavirus infection is widespread in the human population. This family of viruses normally maintains latent infection within the host cell but can cause a range of human pathologies, especially in immunocompromised individuals. Among several known pathogenic human polyomaviruses, JC polyomavirus (JCPyV) has the potential to cause the demyelinating disease progressive multifocal leukoencephalopathy (PML); BK polyomavirus (BKPyV) can cause nephropathy in kidney transplant recipients, and Merkel cell polyomavirus (MCPyV) is associated with a highly aggressive form of skin cancer, Merkel cell carcinoma (MCC). While the mechanisms by which these viruses give rise to the relevant diseases are not well understood, it is clear that the control of gene expression in each polyomavirus plays an important role in determining the infectious tropism of the virus as well as their potential to promote disease progression. In this review, we discuss the mechanisms governing the transcriptional regulation of these pathogenic human polyomaviruses in addition to the best-studied simian vacuolating virus 40 (SV40). We highlight the roles of viral cis-acting DNA elements, encoded proteins and miRNAs that control the viral gene expression. We will also underline the cellular transcription factors and epigenetic modifications that regulate the gene expression of these viruses.

DNA methylation controls unmethylated transcription start sites in the genome in trans

AIM

DNA methylation downregulates transcription. However, a large number of genes, which are unmethylated in the promoter region, are inactive. We tested the hypothesis that these genes are regulated by DNA methylation of upstream regulators.

METHODS

We inhibited DNMT1 with 5-aza-2'-deoxycytidine or depleted it with shRNA to map the transcription initiation positions controlled by DNMT1 using ChIPseq with RNApolIIser5 antibody. Ingenuity pathway analysis identified potential methylated upstream regulators. Their functional role in controlling unmethylated promoters was determined by CRISPR/Cas9 gene editing.

RESULTS

We show that a large group of unmethylated promoters is regulated by DNMT1 through DNA methylation dependent silencing of upstream regulators such as transcription factor HNF4A.

CONCLUSION

The landscape of genes regulated by DNA methylation is more wide-ranging than genes downregulated by methylation of their own cis-regulatory sequences; regulation of unmethylated promoters is dependent on the methylation state of upstream trans regulators.

Targeted methylation of CMV and E1A viral promoters

DNA methylation is a gene-silencing and host defense system that can down-regulate viral gene expression in mammalian cells. An established targeted DNA methylation method was used to demonstrate that genome-integrated CMV and adenovirus type 5 E1A promoters were hypermethylated after MCF7 and HEK293 cells were transfected with in vitro methylated viral promoter fragments. In both cases, the targeted methylation-induced gene silencing could be reversed by addition of 5-aza-2'-deoxycytidine, confirming that the CMV and E1A promoters are regulated by DNA methylation. The kinetics of the targeted DNA methylation was determined using a reporter system in live cells. In conclusion, targeted DNA methylation is able to efficiently silence susceptible viral promoters and provides an alternative strategy to study the impact of loci-specific DNA methylation in viral gene expression.

Presence and role of cytosine methylation in DNA viruses of animals

Nucleotide composition varies greatly among DNA viruses of animals, yet the evolutionary pressures and biological mechanisms driving these patterns are unclear. One of the most striking discrepancies lies in the frequency of CpG (the dinucleotide CG, linked by a phosphate group), which is underrepresented in most small DNA viruses (those with genomes below 10 kb) but not in larger DNA viruses. Cytosine methylation might be partially responsible, but research on this topic has focused on a few virus groups. For several viruses that integrate their genome into the host genome, the methylation status during this stage has been studied extensively, and the relationship between methylation and viral-induced tumor formation has been examined carefully. However, for actively replicating viruses—particularly small DNA viruses—the methylation status of CpG motifs is rarely known and the effects on the viral life cycle are obscure. In vertebrate host genomes, most cytosines at CpG sites are methylated, which in vertebrates acts to regulate gene expression and facilitates the recognition of unmethylated, potentially pathogen-associated DNA. Here we briefly introduce cytosine methylation before reviewing what is currently known about CpG methylation in DNA viruses.

A circadian model for viral persistence

Persistently infecting DNA viruses depend heavily on host cell DNA synthesis machinery. Replication of cellular and viral DNA is inhibited by mutagenic stress. It is hypothesized that diurnal regulation of viral DNA replication may occur at the level of cell cycle checkpoints and DNA repair, to protect DNA from exposure to UV light or other mutagens. This highly conserved mechanism is traced back to viruses that persist in prokaryotes and eukaryotes. Inhibition of viral DNA replication and the cell cycle in response to UV light may represent a functional building block in the evolution of circadian-gated DNA replication. Viral DNA replication appears to be closely linked to the circadian clock by interaction of viral promoters, early viral proteins and transcription factors. It is proposed here that under certain conditions viral oncogene expression is phase-shifted relative to that of tumor suppressor and DNA repair genes. The resulting desynchrony of checkpoint controls and DNA repair from diurnal genotoxic exposure produces cyclic periods of suboptimal response to DNA damage. This temporal vulnerability to genotoxic stress produces a "mutator phenotype" with inherent genome instability. The proposed model delineates areas of research with implications for viral pathogenesis and therapeutics.

CpG Methylation of a Plasmid Vector Results in Extended Transgene Product Expression by Circumventing Induction of Immune Responses

Gene therapy has the potential to cure inherited diseases if the delivered genes achieve long-term expression at therapeutic levels in the targeted tissues. Expression is commonly short-lived due to induction of cell-mediated immune responses to the gene therapy vehicle and/or the transgene product, which can be perceived as "foreign" by the host's immune system. Plasmid expression vectors have been used to deliver genes. Bacterial DNA carries immunostimulatory sequences in the form of unmethylated CpG motifs, which induce an inflammatory reaction that in turn promotes activation of transgene product-specific B and T cells. Elimination or methylation of immunostimulatory CpG sequences in plasmid expression vectors prevents the stimulation of transgene product-specific immune responses without necessarily reducing transgene expression. In this study, we tested if a CpG-methylated plasmid expression vector expressing the highly immunogenic glycoprotein of rabies virus can achieve prolonged transgene product expression by circumventing immune recognition. Our data show that mice inoculated with a CpG-methylated plasmid expression vector show delayed clearance of transfected cells and fail to mount a strong immune response to the transgene product. Gene transfer with a CpG-methylated plasmid results in a state of immunological low responsiveness to the transgene product, which may facilitate readministration of the transgene. Nevertheless, mice remain responsive to the transgene product delivered by a viral vector.

Kinetics of accumulation and processing of simian virus 40 RNA in Xenopus laevis oocytes injected with simian virus 40 DNA

We examined the kinetics of accumulation and processing of simian virus 40 (SV40) RNA in stage 6 oocytes of Xenopus laevis microinjected intranuclearly with SV40 DNA. The rates of synthesis and degradation, cellular distribution, size, and sequence specificity of radiolabeled SV40-specific and endogenous oocyte RNA were determined. The kinetics of accumulation of SV40 RNA were biphasic, with greater than 90% of the viral RNA turning over in the nucleus with a half-life of 20 to 40 min. Although most of the primary transcription products were multigenomic in length, some stable polyadenylated SV40-specific RNA similar in size and sequence to late 19S mRNA accumulated in the cytoplasm with time. Differences in strand preference, efficiencies of transcription termination and polyadenylation, and the splice sites used in the synthesis and processing of SV40 RNA in Xenopus oocytes and monkey cells were noted. However, these differences were quantitative, rather than qualitative, in nature. Consequently, they are probably due to regulatory rather than mechanistic differences between the two cell types. We therefore conclude that Xenopus oocytes may be a useful system for studying both mechanistic and cell type-specific regulatory aspects of mRNA biogenesis from cloned DNAs. However, since only a small percentage of the initially synthesized RNA ends up in stable mRNA, it will be important to determine whether mutants of cloned DNAs that produce abnormal amounts of stable mRNAs are altered in promotion and initiation of RNA synthesis, transcription termination, RNA processing, or the stability of the resultant mRNAs.

Template structural requirements for transcription in vivo by RNA polymerase II

Purified simian virus 40 (SV40) DNA is reconstituted into chromatin and transcribed by endogenous RNA polymerase II when microinjected into nuclei of Xenopus laevis oocytes. We have correlated the kinetics of chromatin reconstitution with that of accumulation of virus-specific RNA in this system. A delay of approximately 3 h was found in the appearance of appreciable numbers of both fully supercoiled molecules and transcriptionally active templates. SV40 mini-chromosomes, isolated from virus-infected monkey cells with 0.2 M NaCl, also exhibited this lag in onset of transcriptional activity when microinjected into oocytes. These findings indicate that neither purified SV40 DNA nor SV40 DNA containing a full complement of nucleosomes can function as a template for transcription in vivo before association with appropriate cellular nonhistone chromosomal factors has taken place. In addition, the gradual degradation of linear SV40 DNA in oocytes was not sufficient to account for the fact that it was much less transcriptionally active than circular SV40 DNA. Taken together, these results indicate that the conformational state of the DNA can affect its ability to function as a template for transcription in vivo by RNA polymerase II. In contrast, transcription by RNA polymerase III of purified, circularized cloned DNAs encoding genes for 5S rRNA was detectable long before the injected DNAs had time to reconstitute into chromatin. Therefore, the template structural requirements for transcription in vivo by RNA polymerases II and III are different.

Template structural requirements for transcription in vivo by RNA polymerase II

Purified simian virus 40 (SV40) DNA is reconstituted into chromatin and transcribed by endogenous RNA polymerase II when microinjected into nuclei of Xenopus laevis oocytes. We have correlated the kinetics of chromatin reconstitution with that of accumulation of virus-specific RNA in this system. A delay of approximately 3 h was found in the appearance of appreciable numbers of both fully supercoiled molecules and transcriptionally active templates. SV40 mini-chromosomes, isolated from virus-infected monkey cells with 0.2 M NaCl, also exhibited this lag in onset of transcriptional activity when microinjected into oocytes. These findings indicate that neither purified SV40 DNA nor SV40 DNA containing a full complement of nucleosomes can function as a template for transcription in vivo before association with appropriate cellular nonhistone chromosomal factors has taken place. In addition, the gradual degradation of linear SV40 DNA in oocytes was not sufficient to account for the fact that it was much less transcriptionally active than circular SV40 DNA. Taken together, these results indicate that the conformational state of the DNA can affect its ability to function as a template for transcription in vivo by RNA polymerase II. In contrast, transcription by RNA polymerase III of purified, circularized cloned DNAs encoding genes for 5S rRNA was detectable long before the injected DNAs had time to reconstitute into chromatin. Therefore, the template structural requirements for transcription in vivo by RNA polymerases II and III are different.

Kinetics of accumulation and processing of simian virus 40 RNA in Xenopus laevis oocytes injected with simian virus 40 DNA

We examined the kinetics of accumulation and processing of simian virus 40 (SV40) RNA in stage 6 oocytes of Xenopus laevis microinjected intranuclearly with SV40 DNA. The rates of synthesis and degradation, cellular distribution, size, and sequence specificity of radiolabeled SV40-specific and endogenous oocyte RNA were determined. The kinetics of accumulation of SV40 RNA were biphasic, with greater than 90% of the viral RNA turning over in the nucleus with a half-life of 20 to 40 min. Although most of the primary transcription products were multigenomic in length, some stable polyadenylated SV40-specific RNA similar in size and sequence to late 19S mRNA accumulated in the cytoplasm with time. Differences in strand preference, efficiencies of transcription termination and polyadenylation, and the splice sites used in the synthesis and processing of SV40 RNA in Xenopus oocytes and monkey cells were noted. However, these differences were quantitative, rather than qualitative, in nature. Consequently, they are probably due to regulatory rather than mechanistic differences between the two cell types. We therefore conclude that Xenopus oocytes may be a useful system for studying both mechanistic and cell type-specific regulatory aspects of mRNA biogenesis from cloned DNAs. However, since only a small percentage of the initially synthesized RNA ends up in stable mRNA, it will be important to determine whether mutants of cloned DNAs that produce abnormal amounts of stable mRNAs are altered in promotion and initiation of RNA synthesis, transcription termination, RNA processing, or the stability of the resultant mRNAs.

The effect of CpG sequences on the B cell response to a viral glycoprotein encoded by a plasmid vector

The effect of palindromic CpG sequences on the B cell response to plasmid vectors expressing a highly immunogenic viral glycoprotein was investigated. Methylation of the CpG sequences of bacterial expression vectors abolished their ability to induce an antibody response to the transgene product in mice. The antibody response could be rescued by concomitant injection of oligonucleotides carrying immunostimulatory sequences. The B cell response to two plasmid vectors, both expressing the same viral glycoprotein but containing a different content of the highly stimulatory AACGTT motif, was compared. Comparable B cell responses were induced to the two constructs given at an optimal vaccine dose while the vector containing additional palindromic sequences resulted in higher antibody titers at a suboptimal dose. These data indicate that deletion of CpG motifs or methylation of such sequences in plasmid DNA can abrogate the immune response to the vector encoded antigen and might thus enhance their usefulness as gene therapy vehicles.

Methylation in the preinitiation domain suppresses gene transcription by an indirect mechanism

Although the first observations of the inhibitory effect of methylation on gene activity were made almost a decade ago, the mechanism by which methyl groups affect transcription is still obscure. Here we use engineered promoters methylated in vitro in transient transfections to study the mechanism by which methylation mediates promoter repression. The results clearly show that the location of the methyl groups within the promoter region determines the extent of promoter repression. The most effective suppression was observed when methylation was in the preinitiation domain. The results also support a previous suggestion that a mediator protein is involved in the mechanism of promoter inhibition. The suppressor effect of methylation at sequences flanking the TATA box can be partially overcome in the presence of the simian virus 40 enhancer. In addition, results obtained by transient thymidine labeling of Ltk- cells that were transfected with a methylated thymidine kinase gene from herpes simplex virus, at the level of approximately one template per cell, further support the conclusion that methylation affects primarily transcription preinitiation.

Elevation of large-T antigen production by sodium butyrate treatment of SV40-transformed WI-38 fibroblasts

The effects of sodium butyrate on simian virus 40 early gene expression were determined in SV40-transformed human embryonic lung fibroblasts (SVWI-38). Northern blot analysis and nuclear run-off transcription studies revealed that treatment of cells with millimolar concentrations of sodium butyrate (2.5 to 10 mM) resulted in increased levels of SV40 early gene transcripts, with a concomitant increase in their corresponding proteins (large-T and small-t antigens). Although sodium butyrate treatment enhanced the expression of the early genes, it was associated with a reduction in cell growth and total protein synthesis, as measured by cell number and incorporation of 3H-leucine into macromolecules, respectively. Immunoprecipitation of 35S-labelled cellular proteins with anti-p53 and anti-T antibodies revealed that the level of the cellular protein, p53, declined markedly in the presence of sodium butyrate. Furthermore, in control cells only 30% of the p53 was complexed with large-T antigen, whereas in butyrate-treated cells all the p53 was complexed with large-T antigen. The increased early gene expression was not due to altered methylation patterns, gene amplification, or rearrangement of the integrated SV40 genome. Sodium butyrate treatment did, however, result in the appearance of a new nuclear protein which bound specifically to a SV40 promoter fragment containing large-T antigen binding sites I and II.

DNA methylation and gene expression

A large body of evidence demonstrates that DNA methylation plays a role in gene regulation in animal cells. Not only is there a correlation between gene transcription and undermethylation, but also transfection experiments clearly show that the presence of methyl moieties inhibits gene expression in vivo. Furthermore, gene activation can be induced by treatment of cells with 5-azacytidine, a potent demethylating agent. Methylation appears to influence gene expression by affecting the interactions with DNA of both chromatin proteins and specific transcription factors. Although methylation patterns are very stable in somatic cells, the early embryo is characterized by large alterations in DNA modification. New methodologies are now becoming available for studying methylation at this stage and in the germ line. During development, tissue-specific genes undergo demethylation in their tissue of expression. In tissue culture cells this process is highly specific and appears to involve an active mechanism which takes place in the absence of DNA replication. The X chromosome undergoes inactivation during development; this is accompanied by de novo methylation, which appears necessary to stably maintain its silent state. As opposed to the programmed changes in DNA methylation which occur in vivo, immortalized tissue culture cells demonstrate alterations in DNA modification which take place over a long time scale and which appear to be the result of selective pressures present during the growth of these cells in culture.

Changes in DNA methylation during mouse embryonic development in relation to X-chromosome activity and imprinting

Changing DNA methylation patterns during embryonic development are discussed in relation to differential gene expression, changes in X-chromosome activity and genomic imprinting. Sperm DNA is more methylated than oocyte DNA, both overall and for specific sequences. The methylation difference between the gametes could be one of the mechanisms (along with chromatin structure) regulating initial differences in expression of parental alleles in early development. There is a loss of methylation during development from the morula to the blastocyst and a marked decrease in methylase activity. De novo methylation becomes apparent around the time of implantation and occurs to a lesser extent in extra-embryonic tissue DNA. In embryonic DNA, de novo methylation begins at the time of random X-chromosome inactivation but it continues to occur after X-chromosome inactivation and may be a mechanism that irreversibly fixes specific patterns of gene expression and X-chromosome inactivity in the female. The germ line is probably delineated before extensive de novo methylation and hence escapes this process. The marked undermethylation of the germ line DNA may be a prerequisite for X-chromosome reactivation. The process underlying reactivation and removal of parent-specific patterns of gene expression may be changes in chromatin configuration associated with meiosis and a general reprogramming of the germ line to developmental totipotency.

Transcriptional regulation and DNA methylation of nuclear genes for photosynthesis in nongreen plant cells

The transcripts of nuclear genes for the small subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase (rbcS), chlorophyll a/b-binding protein (cab), and extrinsic 33-kDa protein involved in photosystem II water oxidation (woxA) were not detectable in the white wild cultured cells of sycamore (Acer pseudoplatanus), in contrast to their high levels in the sibling green mutant cells and the constitutive expression of actin genes (act) in both cell types. We have examined the template activities of nuclear DNAs using the HeLa cell in vitro transcription system. All of the three photosynthesis genes from the green cell line and act from both cell types were well transcribed in vitro, but these photosynthesis genes from the white cell line were not, indicating that the transcriptional regulation is ascribable to DNA templates. Digestion of nuclear DNA with methyl-sensitive and -insensitive isoschizomeric endonucleases and the subsequent Southern hybridization showed that each gene has the identical recognition sites of restriction enzymes in the green and white cell lines, but some of the sites were methylated only in the photosynthesis genes in the white cells. There was observed a clear inverse relationship between the level of expressed transcripts and the extent of DNA methylation. Thus, it is inferred that the selective methylation of DNA is a likely mechanism for suppressing transcription of nuclear genes for photosynthesis in the nonphotosynthetic plant cells.

Sp1 transcription factor binds DNA and activates transcription even when the binding site is CpG methylated

In vertebrates, a negative correlation between gene activity and CpG methylation of DNA, notably in the promoter region, is well established. Therefore, it is conceivable that differential binding of transcription factors to methylated versus unmethylated binding sites is crucial for gene activity. Since the consensus binding site of transcription factor Sp1 contains a central CpG, we have investigated the binding of Sp1 factor to unmethylated and synthetically CpG-methylated DNA. A strong Sp1 binding site was methylated on both strands at two CpG positions, located in the center and at the periphery of the recognition sequence. Our studies show that neither binding in vitro, nor transcription in vivo and in vitro are affected by methylation of the Sp1 binding site. We discuss the possibility that binding of Sp1 factor, which is often associated with promoters of housekeeping genes, prevents CpG methylation.

An excised SV40 intron accumulates and is stable in Xenopus laevis oocytes

Xenopus laevis oocytes injected with simian virus 40 (SV40) DNA synthesize abundant quantities of viral late region RNA. In a previous analysis of the 5' ends of oocyte SV40 late RNAs, it was observed that, in contrast to the majority of the late RNA species, an abundant class of viral late RNAs, whose 5' ends mapped at or near nucleotide 294, was not polyadenylated. The structure of this RNA class has now been characterized further. We have shown that this species consists of a class of small uncapped RNA molecules with heterogeneous 3' ends mapping between nucleotides 417 and 433. This corresponds well with the position of a 139-nucleotide intron within the leader region of late 16S RNA (nucleotides 294-433). The identification of this RNA class as an excised intron was strongly supported by the fact that it displayed anomalous mobilities on different percentage polyacrylamide gels, a property of lariat introns. Furthermore, incubation of oocyte RNA with a HeLa cell extract with lariat debranching activity converted the small RNA to a class that now migrated as less than or equal to 140 nucleotides in length in 8% gels, consistent with the size of the linear intraleader intron. Additional analysis of this RNA showed that it is primarily nuclear in localization and is probably the most stable viral RNA species in the oocyte. These data suggest that oocytes accumulate large quantities of the 16S intraleader intron because of their failure to debranch this RNA efficiently.

Effect of in vitro DNA methylation on beta-globin gene expression

When the human beta-globin gene was methylated at every cytosine residue and was inserted into mouse fibroblasts by DNA-mediated gene transfer, the transcription of the gene was strongly inhibited. This methylation also prevented expression and induction of the gene in mouse erythroleukemia cells. By using partially methylated hybrid molecules, it was shown that methylation-sensitive negative regulatory elements are located in both the 5' and 3' ends of the beta-globin gene but not in the 90-base-pair region usually associated with promoter activity. To further investigate the role of DNA methylation in the regulation of the beta-globin gene, 50-base-pair poly(dG-dC) tracts were introduced into various sites in a mouse-human hybrid gene, and these inserts were methylated by means of the Hha I methylase. Heavy methylation of these artificially added sites had no effect on either transcription initiation or elongation, suggesting that DNA modification operates through fixed endogenous sites in the gene domain.

Fixation of the unmethylated or the 5'-CCGG-3' methylated adenovirus late E2A promoter-cat gene construct in the genome of hamster cells: gene expression and stability of methylation patterns

The late E2A promoter of adenovirus type 2 (Ad2) DNA can be inactivated by in vitro methylation of three 5'-CCGG-3' sequences at positions +23, +5, and -215 relative to the cap site in this promoter. This inactivation has been documented in transient expression experiments both in Xenopus laevis oocytes and in mammalian cells (K.-D. Langner, L. Vardimon, D. Renz, and W. Doerfler, Proc. Natl. Acad. Sci. USA 81:2950-2954, 1984; K.-D. Langner, U. Weyer, and W. Doerfler, Proc. Natl. Acad. Sci. USA 83:1598-1602, 1986). In the present study, in vitro-methylated or unmethylated promoter-gene assemblies were permanently fixed by integration in the hamster genome. In individually established cell lines, the degree of promoter methylation was correlated to gene activity. The pAd2E2AL-CAT construct, in which the late E2A promoter controls expression of the procaryotic chloramphenicol acetyltransferase (cat) gene, was fixed in BHK21 hamster cells by cotransfection with and selection for the pSV2-neo construct (P. J. Southern and P. Berg, J. Mol. Appl. Genet. 1:327-341, 1982) in which the early simian virus 40 promoter controls the gene for neomycin phosphotransferase. The pAd2E2AL-CAT construct was transfected in the unmethylated or in the 5'-CCGG-3' methylated form. The pSV2-neo plasmid was cotransfected in the unmethylated form. The stability of in vitro-imposed methylation patterns and cat gene expression were followed and correlated in a number of established cell lines which contained the constructs integrated in a non-rearranged configuration. The foreign DNA did not persist in the episomal state but was integrated, frequently in multiple tandems of the plasmid DNA. Among 19 cell lines established after transfecting the unmethylated pAd2E2AL-CAT construct, the late E2A promoter remained unmethylated (examined in 10 cell lines), and the cat gene was expressed in 18 cell lines. On the other hand, among 14 cell lines which were generated by transfection with the methylated construct, 7 cell lines did not express the cat gene, and the three 5'-CCGG-3' sequences in the late E2A promoter remained almost completely methylated. In five cell lines, the E2A promoter sequences were partly demethylated and the cat gene was expressed at low levels. Last, in two cell lines, demethylations were found to be extensive and strong cat expression was observed. It remained a question of considerable interest what factors determined the stability of methylation patterns that had been preimposed by in vitro methylation on specific sequences in a promoter, after this promoter was fixed by integration in the mammalian genome.(ABSTRACT TRUNCATED AT 400 WORDS)

Methylation of specific cytosine residues enhances simian virus 40 T-antigen binding to origin region DNA

Specific binding of simian virus 40 large T antigen to origin region DNA requires the interaction of T antigen with multiples of a consensus recognition pentanucleotide sequence (5'-G[T]-A[G]-G-G-C-3'). To assess the interaction of T antigen with cytosine residues in the recognition sequences, bacterial methylases were used to methylate simian virus 40 form I DNA in vitro at specific cytosine residues. Methylation of a subset of the cytosine residues in the pentanucleotide sequences resulted in enhanced binding of T antigen to origin region DNA. Enhanced binding to the methylated pentanucleotides indicates that the methyl groups introduced on this subset of pentanucleotide cytosine residues could not have sterically interfered with the interaction of T antigen with the recognition sequences. This lack of steric interference suggests that T antigen does not make close contact in the major groove with these particular cytosine residues during normal binding.

Effects of DNA methylation on specific transcription by RNA polymerase II in vitro

In vitro transcription in a HeLa cell lysate by RNA polymerase II directed by a chicken feather keratin gene promotor has been studied using unmethylated template DNA and DNA methylated in vitro by HpaII methylase. The efficiency of specific gene transcription from methylated DNA was dependent on topology of the input DNA, the most significant effect being complete inhibition of transcription from one template which contained three methylation sites, one just 5' and two greater than 500 bases 3' to the site of transcription initiation. The inhibition of transcription depends on a factor(s) which is variably present in lysate preparations and is labile on storage at -70 degrees.

Patterns of methylation of polyomavirus DNA in polyoma-transformed rat cells

The patterns of methylation of integrated polyomavirus (Py) DNA and flanking cellular sequences were determined in an inducible line of Py-transformed rat cells, designated LPT, by an analysis of cleavage patterns of LPT DNA generated by the restriction enzymes MspI, HpaII, and HhaI. The Py DNA in LPT cells is integrated into a single chromosomal site and includes whole viral genomes arranged in a head-to-tail configuration. Amplification of the viral DNA and synthesis of infectious virus can be induced in these cells by treatment with carcinogens. The experiments reported here show that in uninduced LPT cells only the late Py genes, which encode the Py capsid proteins, and sequences flanking one of the two viral DNA-cell DNA junctions are methylated. The early genes, encoding the Py T antigens, and sequences flanking the second junction are unmethylated. Since only the early genes are transcribed and translated in uninduced LPT cells, it is apparent that an inverse correlation exists between transcription and methylation of the integrated Py genes in LPT cells, as reported in other cellular and viral systems. The patterns of methylation of integrated Py genomes were also examined in cells of a subclonal derivative of the LPT line in which the virus cannot be activated. In these cells, not only the late Py genes were found to be methylated, but also the 3' portion of the early gene which encodes the Py large T antigen. These findings may have implications for understanding Py induction in LPT cells.

One role for DNA methylation in vertebrate cells is strand discrimination in mismatch repair

Although the occurrence of 5-methylcytosine (m5C) in DNA is widespread, the function of this modified base remains unclear. At some specific sites it apparently has an effect in controlling gene expression, but many sites do not appear to be involved in this regulation. Balanced against its regulatory usefulness at some sites is the mutational risk it imposes upon the cell. Deamination of m5C can lead to its replacement by thymine (T). One possible role for excess methylation is strand discrimination in the repair of mismatches. We constructed the complementary hemimethylated single-base-pair mismatches, G T and A C, at a CG site in simian virus 40 DNA, transfected these into the host African green monkey kidney cells (CV-1), and examined DNA of the progeny for repair at this site. Hemimethylation at two Hha I sites (Gm5CGC) bracketing the mismatch directed repair to occur only on the unmethylated strand. Methylation at the multiple Cm5CATGG and Gm6ATC sites, a pattern normally seen in bacteria, also instructed repair to proceed on the unmethylated strand, although less efficiently. Hemimethylation at only one site, adjacent to the mispaired bases (Hpa II, Cm5CGG) produced repaired molecules in a ratio that may represent random repair of the A C mismatch and strand-directed repair in the complementary G T mismatch. The -mCG- -GT- mismatch could result from deamination of m5C in the most commonly methylated dinucleotide in vertebrates, CpG. Methylation may be able to compensate for the errors it causes by serving as a mechanism for strand discrimination in correcting those errors. In addition, single-strand nicks were also shown to direct repair.

Low expression of human histocompatibility leukocyte antigen-DR is associated with hypermethylation of human histocompatibility leukocyte antigen-DR alpha gene regions in B cells from patients with systemic lupus erythematosus

The relationship between the expression of HLA-DR antigens and the HLA-DR alpha gene methylation was examined in systemic lupus erythematosus (SLE). Using permanent B cell lines, we found reduced DR expression in SLE. The low DR expression was correlated with high anti-DNA antibody titers in patients' sera. The amounts of DR alpha message were lower in SLE cells than in normal controls, suggesting that the low expression of DR antigens is associated with gene functions. The extent of DNA methylation was examined at five CCGG sites in the HLA-DR alpha locus. DNA from both SLE and normal cells showed variable methylation patterns. Since the DR alpha gene is a single-copy gene, such a variability is the result of assaying a mixture of transformed clones containing methylated DR alpha gene, with other clones containing unmethylated DR alpha gene. A distinctive feature of normal cells was a consistent methylation pattern: 12 normal cell lines showed exactly the same pattern. In contrast, 28 SLE cell lines showed a cell-line-specific methylation, and hypermethylation at the DR alpha locus. The hypermethylation is often associated with transcriptionally inactive genes. Thus, our results suggest that (a) B cells with hypermethylated DR genes might express no or few DR antigens; (b) the ratio of cells with differently methylated DR genes is consistent in normal individuals, while, in SLE patients, cells with hypermethylated DR genes predominate, resulting in apparently reduced DR antigen expression; and (c) the aberrant DR expression could be associated directly with immunoregulatory dysfunctions in SLE disease.

DNA methylation diminishes bleomycin-mediated strand scission

Three DNA duplexes differing substantially in sequence were derived from pBR322 plasmid DNA and supercoiled SV40 DNA by digestion with appropriate restriction endonucleases. Following treatment with the restriction methylase HhaI (recognition sequence: GCGC) or HhaI and HpaII (CCGG), the unmethylated and methylated DNAs were compared as substrates for the antitumor agent bleomycin. Bleomycin-mediated strand scission was shown to diminish substantially at a number of sites in proximity to the methylated cytidine moieties, especially where multiple sites had been methylated within a DNA segment of limited size. Detailed analysis of the DNA substrates revealed that both strands of DNA within a methylated region became more refractory to cleavage by bleomycin and that the protective effect could extend as many as 14 base pairs in proximity to the 5-methylcytidine moieties. Among the methylated DNA segments that became more resistant to bleomycin cleavage was a HpaII site of SV40 DNA, methylation of which has previously been shown to diminish the synthesis of the major late viral capsid protein following microinjection into Xenopus laevis oocytes. Study of the cleavage reaction at varying salt levels suggested that diminished bleomycin strand scission may be due, at least in part, to local conformational changes of the DNA to Z form (or other non-B-form structures). The results are generally consistent with the hypothesis that one mechanism for the expression of selective therapeutic action by certain DNA damaging agents could involve the recognition of specific methylation patterns.

Inhibition of herpes simplex thymidine kinase gene expression by DNA methylation is an indirect effect

The biological activity of in vitro methylated HSV-TK DNA was analysed after microinjection into thymidine kinase negative rat 2 cells. It was found that the fully methylated DNA (Hpa II methylase) was as active as the non methylated control DNA for about 48 hours after injection. DNA reextraction experiments and blot analysis showed that DNA demethylation was not the reason for the observed TK activity. With prolonged cultivation time the methylated DNA becomes rapidly inactive and 100 hrs after injection thymidine incorporation was no longer detectable in the recipient cells. In transformed cells, obtained after coinjection with SV40 DNA, the HSV-DNA was partially demethylated and inactive. Addition of 5-azacytidine to the culture medium induced further demethylation and reactivation of the thymidine kinase gene.

A domain of methylation change at the albumin locus in rat hepatoma cell variants

A well-differentiated rat hepatoma cell line, Fu5-5, yields variant clones whose rate of secretion of serum albumin ranges from 40 to less than 0.08 micrograms of albumin/mg of cell protein per 48 h. Clones were classified as high producers (10 to 40 micrograms/mg per 48 h), intermediate producers (1 to 10 micrograms/mg per 48 h), low producers (0.1 to 1.0 micrograms/mg per 48 h), and null variants (less than 0.1 micrograms/mg per 48 h). Albumin synthetic rates are proportional to secretion rates and range from 0.9 to less than 0.002% of total protein synthesis as measured by pulse-labeling. Steady-state albumin mRNA levels were measured by filter hybridization of fragmented, end-labeled mRNA and by Northern blotting. Message levels are proportional to albumin synthetic rates except for a high producer in which albumin mRNA is less elevated than the synthetic rate. The extent of methylation was quantitated at each of 24 CpG-containing sites or site clusters at the albumin locus. These sites span a region that contains the albumin gene as well as 10 kilobases of the 5' flank and 1 kilobase of the 3' flank. An 8-kilobase region is described, with boundaries in the 5' flank and in the middle of the gene, within which all 11 sites examined showed a correlation of undermethylation with the high-producer phenotype. In contrast, 12 of 13 sites outside of this region showed no phenotype correlation. Null variants derived from a high producer underwent de novo methylation of this domain. Six independent hybrid clones derived from the cross of a high producer with a null variant showed extinction of albumin production and hypermethylation of the domain. Apparently these cells retain the capacity for the de novo methylation of these specific sites.

Effect of regional DNA methylation on gene expression

The effect of DNA methylation on the transcriptional activity of the hamster adenine phosphoribosyltransferase (aprt) and the herpes thymidine kinase (tk) genes has been investigated. By using M13 constructs containing these gene sequences, specific segments of each gene were methylated in vitro by restriction fragment primer-directed second-strand synthesis using the substrate 2'-deoxy-5-methyl-cytidine triphosphate (dmCTP). These hybrid-methylated molecules were inserted into mouse Ltk- cells by DNA-mediated cotransfer. In all cases, the integrated sequences retained the in vitro-directed methylation pattern. The aprt gene was inhibited by CpG methylation in the 5' region but was unaffected by methylation at the 3' end or in adjacent M13 sequences. In contrast to this, DNA methylation in both the 5' promoter region and the 3' structural region of the tk gene had a strong inhibitory effect. This suggests that this modification may affect transcription by mechanisms that do not involve the direct alteration of recognition sequences for RNA polymerase.

Methylation pattern of fish lymphocystis disease virus DNA

The content and distribution of 5-methylcytosine in DNA from fish lymphocystis disease virus was analyzed by high-pressure liquid chromatography, nearest-neighbor analysis, and with restriction endonucleases. We found that 22% of all C residues were methylated, including methylation of the following dinucleotide sequences: CpG to 75%, CpC to ca. 1%, and CpA to 2 to 5%. Comparison of relative digestion of viral DNA with MspI and HpaII indicated that CCGG sequences were almost completely methylated at the inner C. The degree of methylation of GCGC was much lower. The methylation pattern of fish lymphocystis disease virus DNA differed from that of the host cell DNA.

A domain of methylation change at the albumin locus in rat hepatoma cell variants

A well-differentiated rat hepatoma cell line, Fu5-5, yields variant clones whose rate of secretion of serum albumin ranges from 40 to less than 0.08 micrograms of albumin/mg of cell protein per 48 h. Clones were classified as high producers (10 to 40 micrograms/mg per 48 h), intermediate producers (1 to 10 micrograms/mg per 48 h), low producers (0.1 to 1.0 micrograms/mg per 48 h), and null variants (less than 0.1 micrograms/mg per 48 h). Albumin synthetic rates are proportional to secretion rates and range from 0.9 to less than 0.002% of total protein synthesis as measured by pulse-labeling. Steady-state albumin mRNA levels were measured by filter hybridization of fragmented, end-labeled mRNA and by Northern blotting. Message levels are proportional to albumin synthetic rates except for a high producer in which albumin mRNA is less elevated than the synthetic rate. The extent of methylation was quantitated at each of 24 CpG-containing sites or site clusters at the albumin locus. These sites span a region that contains the albumin gene as well as 10 kilobases of the 5' flank and 1 kilobase of the 3' flank. An 8-kilobase region is described, with boundaries in the 5' flank and in the middle of the gene, within which all 11 sites examined showed a correlation of undermethylation with the high-producer phenotype. In contrast, 12 of 13 sites outside of this region showed no phenotype correlation. Null variants derived from a high producer underwent de novo methylation of this domain. Six independent hybrid clones derived from the cross of a high producer with a null variant showed extinction of albumin production and hypermethylation of the domain. Apparently these cells retain the capacity for the de novo methylation of these specific sites.

The cardiac myosin light chain (MLC-2A) gene in chicken is methylated in both expressing and nonexpressing tissues

Certain specific methylation sites of the chicken cardiac myosin light chain (MLC-2A) gene in DNA from embryonic heart tissue as well as from embryonic livers and brains were studied. MLC-2A specific mRNA was present in the heart only, indicating that the gene is transcriptionally active in this tissue and inactive in liver and brain. Using the methylation-sensitive restriction enzymes Hpa II and Hha I, the gene was found to be identically methylated in DNA from all three examined tissues, irrespective of its expression. Moreover, the gene was extensively methylated throughout, including the 5' end and 3' flanking regions, with the exception of 3 Hpa II sites, one located in the middle part of the gene, one in the last intron and one in the 3' flanking sequence. These results suggest that specific genes can be actively expressed even when they are highly modified, and that changes in the methylation pattern from one tissue to another are not necessarily associated with the gene activation.

Methylation of the SV40 HpaII site does not affect late viral gene expression in microinjected tissue culture cells

Intranuclear coinjection of the late SV40 KpnI/BclI DNA fragment and the early promotor/enhancer HpaII/BglI DNA segment into permissive monkey and non-permissive mouse cells allows late SV40 gene expression without T-antigen synthesis and DNA replication. These conditions were chosen to analyse the effect of DNA methylation on V-antigen synthesis detached from the process of DNA replication. We found that in vitro methylation of a single cytosine nucleotide proximal to the major late mRNA cap site by the HpaII methylase does not block capsid protein synthesis. This result is in contrast to reported data obtained in Xenopus laevis oocyte injection experiments [(1982) Proc. Natl. Acad. Sci. USA 79, 5142-5146].

Activation of a chicken embryonic globin gene in adult erythroid cells by 5-azacytidine and sodium butyrate

Adult White Leghorn chickens were rendered anemic by injection with 1-acetyl-2-phenylhydrazine and then treated with parenteral 5-azacytidine, and levels of embryonic globin RNA in circulating reticulocytes were measured. A very small but detectable amount of correctly initiated embryonic p-type globin RNA was detected in reticulocytes from birds treated with 5-azacytidine, while none was detected in reticulocytes from those receiving only phenylhydrazine or phenylhydrazine plus 1-beta-D-arabinofuranosylcytosine (cytosine arabinonucleoside). An attempt to increase embryonic globin RNA induction by treatment with parenteral sodium butyrate after 7 days of 5-azacytidine administration resulted in a 5- to 10-fold increase in the level of embryonic globin RNA. However, sodium butyrate did not induce embryonic gene expression when given alone or after treatment with cytosine arabinonucleoside. Sodium butyrate treatment also caused a DNase I-hypersensitive site to be exposed at the 5' end of the rho-globin gene only after 5-azacytidine induced demethylation of several CpG sites in and around the gene. The implications of this model of gene activation in vivo are discussed in the context of multistep gene regulation.

Transcription of methylated eukaryotic viral genes in a soluble in vitro system

SV40 and adenovirus-2 (Ad2) recombinant plasmids containing long segments of poly(dC-dG) cloned adjacent to transcription control regions were methylated in vitro with Hbal methylase and transcribed in a soluble in vitro system. The addition of up to 40 or more base pairs of poly(m5dC-dG) immediately upstream or downstream of promoter regions was shown to have no effect on the accuracy or efficiency of specific transcription from these promoters in vitro. Methylation at various naturally occurring C-G sequences within or near these promoters also had no effect on transcription in vitro. The significance of these results with respect to possible mechanisms whereby DNA methylation might regulate eukaryotic gene expression is discussed.

DNA methylation of three 5' C-C-G-G 3' sites in the promoter and 5' region inactivate the E2a gene of adenovirus type 2

The E2a gene of human adenovirus type 2 (Ad2) encodes the 72-kilodalton DNA-binding protein. We previously described perfect inverse correlations between the methylation of all 5' C-C-G-G 3' sequences in the Ad2 E2a gene in virus-transformed hamster cells containing viral DNA sequences in an integrated state and the extent to which this gene is expressed. We subsequently showed that in vitro methylation of all 14 5' C-C-G-G 3' sequences in the cloned E2a gene by prokaryotic Hpa II DNA methyltransferase leads to transcriptional inactivation after microinjection into Xenopus laevis oocytes. The unmethylated cloned E2a gene is expressed in these cells. We report here the construction of partly methylated clones of the E2a gene. In the promoter (5')-methylated construct, three 5' C-C-G-G 3' sequences at the 5' end of the subclone were methylated. One of these sites is located 215 base pairs (bp) upstream (bp 26,169 of Ad2 DNA), and two sites are located 5 and 23 bp downstream from the cap site (bp 25,931 and 25,949 of Ad2 DNA) of the E2a gene. This construct was transcriptionally inactive upon microinjection into nuclei of X. laevis oocytes. In the gene (3')-methylated construct, 11 5' C-C-G-G 3' sequences in the main part of the transcribed gene region were methylated in vitro. This construct was transcribed in X. laevis oocytes, and at least some of the Ad2-specific RNA synthesized was initiated at the same sites as in Ad2-infected human KB cells. Both mock-methylated constructs were transcribed into Ad2-specific RNA in X. laevis oocytes. These results demonstrate that DNA methylations at or close to the promoter and 5' end of the E2a gene cause transcriptional inactivation. Perhaps only one methyl group would be adequate for inactivation; in vivo methylation of more than one cytosine may be a form of safeguard or redundancy.

Methylation of human T-cell leukemia virus proviral DNA and viral RNA expression in short- and long-term cultures of infected cells

Leukemic peripheral blood lymphocytes from individuals infected with the human T-cell leukemia/lymphoma virus (HTLV) were found to express little or no viral RNA before being put into tissue culture. Within 24-48 hr, viral RNA expression increased at least four- to eightfold. Established HTLV-infected cell lines constitutively express viral RNA. Southern blots of DNA from HTLV-infected cells digested with the methylation-sensitive restriction enzyme HpaII showed that the proviral DNA was methylated in all of the uncultured peripheral blood cells tested. In contrast, no proviral methylation was detected in any of the cell lines examined, suggesting a functional correlation between methylation and viral RNA expression. However, DNA from HTLV-infected lymphocytes cultured for 48 hr (by which time increases in viral RNA expression are evident) did not differ detectably with respect to proviral DNA methylation from uncultured cells, suggesting that the increase in viral RNA expression after short-term culture is mediated by mechanisms independent of changes in DNA methylation.

Simian virus 40 late promoter region able to initiate simian virus 40 early gene transcription in the absence of the simian virus 40 origin sequence

To improve our knowledge of the simian virus 40 (SV40) late promoter control region, we took advantage of the fact that T antigen can be expressed with a heterologous promoter. We constructed three chimeric plasmids (pEMP-273, pEMP-LCAP-162, and pEMP-LCAP-113) each with a putative late promoter sequence positioned immediately upstream from the SV40 early gene coding region but in an orientation opposite to its natural orientation in the SV40 genome. After transfection of the recombinant DNA into HeLa or CV1 cells, T antigen accumulation, as scored by indirect immunofluorescence, was used as a functional test for promoter activity. We found that the sequence mapping from nucleotides 332 to 273 is not sufficient for promoting transcription of SV40 early gene but does potentiate the promoter activity of the 72-base-pair repeats in initiating the transcription at natural late cap sites. Considering that both plasmids pEMP-LCAP-162 and pEMP-LCAP-113 lack all of the sequence of the SV40 replication origin, we conclude that SV40 transcription can be mediated through a putative late promoter in the absence of the sequence for the SV40 replication origin.

5-Methyldeoxycytidine in the Physarum minichromosome containing the ribosomal RNA genes

5-Methyldeoxycytidine (5MC) was analyzed by high pressure liquid chromatography (HPLC) and by restriction enzyme digestion in rDNA isolated from Physarum polycephalum. rDNA from Physarum M3C strain microplasmodia has a significant 5MC content (about half that of the whole genomic DNA). This rDNA contains many C5MCGG sites because it is clearly digested further by Msp I than by Hpa II. However, most 5MC is in other sites. In particular, alternating CG sequences appear to be highly methylated. HPLC of deoxyribonucleosides shows tha most of the transcribed regions contain little or no 5MC. Restriction digestion indicates that there is little or no 5MC in any of the transcribed regions including the transcription origin and adjacent sequences. Over 90% of the total 5MC is in or near the central nontranscribed spacer and most methylated restriction sites are in inverted repeats of this spacer. rDNA is very heterogeneous with respect to 5MC. The 5MC pattern doesn't appear to change with inactivation of the rRNA genes during reversible differentiation from microplasmodia (growing) to microsclerotia (dormant), showing that inactivation is due to changes in other chromatin variables. The 5MC pattern is different between Physarum strains. The possible involvement of this 5MC in rDNA chromatin structure and in cruciform and Z-DNA formation is discussed.