Chromatin structure of endogenous retroviral genes and activation by an inhibitor of DNA methylation

Methylation of an alpha-foetoprotein gene intragenic site modulates gene activity

By comparing the methylation pattern of Mspl/Hpall sites in the 5' region of the mouse alpha-foetoprotein (AFP) gene of different cells (hepatoma cells, foetal and adult liver, fibroblasts), we found a correlation between gene expression and unmethylation of a site located in the first intron of the gene. Other sites did not show this correlation. In transfection experiments of unmethylated and methylated AFP-CAT chimeric constructions, we then showed that methylation of the intronic site negatively modulates expression of CAT activity. We also found that a DNA segment centered on this site binds nuclear proteins; however methylation did not affect protein binding.

The continued search for evidence of retroviral infection in schizophrenic patients

Retroviral infection has been proposed as an etiologic factor in schizophrenia. In an effort to unmask a latent retrovirus, short term cultures of peripheral lymphocytes from 15 schizophrenic subjects and nine normal controls were exposed to ionizing radiation and co-cultured with indicator cells. Reverse transcriptase activity, a marker of retroviral infection, could not be detected in any of the cultures. Our findings are further evidence against a role for retroviral infection in the etiology of schizophrenia.

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.

Methylation and expression of human pepsinogen genes in normal tissues and their alteration in stomach cancer

In normal human tissues, pepsinogen A mRNA was expressed only in the fundic mucosa of the stomach, whereas pepsinogen C mRNA was expressed in all regions of the stomach mucosa and also in the proximal duodenal mucosa. The distributions of these mRNAs were consistent with those of pepsinogens A and C in the gastroduodenal mucosa. Methylation analysis of DNAs from normal tissues with methylation-sensitive restriction enzymes, HpaII and HhaI, revealed that pepsinogen A and C genes are hypomethylated in tissues producing pepsinogens A and C, suggesting a role of DNA methylation in the regulation of the differential expression of the genes for the two human pepsinogens during normal differentiation. In stomach cancer tissues and cancer cell lines, the expressions of the pepsinogen genes were decreased or lost, in good accordance with their pepsinogen productions. No gross structural changes of the pepsinogen genes were observed in these cancers, but the methylation patterns of the pepsinogen genes were found to be altered in different ways in different cancers. The functional significance of the altered methylation is unknown; however, these results suggest that considerable heterogeneity of the methylation patterns occurs in human stomach cancers.

Epigenetic control of tumor cell morphology

XC cell line derived from a single rat cell transformed by the Prague strain of Rous sarcoma virus produced morphologically different colonies. Among them, two distinct cell types consisting of thick, fusiform cells (L-type), and of flat, polygonal cells (R-type) were apparent. By repeated subclonings, pure cultures, L1 and R1, respectively, were obtained. These clones underwent morphological conversion during prolonged culture; L-type colonies appeared in the R-type clone and vice versa. The kinetic curve suggested that the conversion was multi-stepped. When inoculated into nude mice, L-type cells produced much larger tumors at a higher frequency than R-type cells, and the tumors induced by these two clones were histologically different. The expression of v-src gene was higher in L-type than in R-type cells at both mRNA and protein levels.

Activation of an endogenous retrovirus enhancer by insertion into a heterologous context

An enhancer element is located in the U3 portion of exogenous avian retrovirus long terminal repeats (LTRs). A similar element has not been detected in the LTRs of ev-1 and ev-2, two avian endogenous viruses (evs) that normally are not expressed in vivo. Experiments were initiated to determine whether minor nucleotide differences in the U3 region of a previously untested ev that is ubiquitously expressed in vivo (ev-3) might confer enhancer function on the LTR of this provirus. This question was addressed by inserting U3 regions from ev-3 and from ev-1 and/or ev-2 both upstream of the herpesvirus thymidine kinase gene promoter and in place of the major enhancer domains of the Rous sarcoma virus LTR and determining their relative effects on transcription. U3 regions from all evs tested were unable to enhance transcription from the thymidine kinase gene promoter, indicating that nucleotide differences in the ev U3 regions do not affect their relative enhancer function and therefore are unlikely to play a role in their differential expression in vivo. Unexpectedly, however, all ev U3 regions were able to augment transcription in an orientation-independent manner in the ev-Rous sarcoma virus hybrid LTRs. Further experiments conducted to determine why this enhancer activity is not detectable in intact ev LTRs demonstrated that it was not due to removal of repressor sequences in the ev fragments used that might normally be present in intact ev LTRs. The lack of detectable enhancer activity in intact ev LTRs also was not explained by a defect in ev promoters that makes them unresponsive to enhancers in cis. These experiments therefore identify sequences that, although unable to function detectably as enhancers in their natural context, can function efficiently in a heterologous context. Data are discussed in terms of the modularity of enhancer elements and possible interactions between enhancers and promoter-specific sequences.

Sequence conservation and structural organization of the glycerol-3-phosphate dehydrogenase promoter in mice and humans

Cloned segments of the mouse glycerol-3-phosphate dehydrogenase (GPDH) gene, Gdc-1, were used to screen a human library. Human clones obtained spanned 25 kilobases of genomic DNA containing the human GPDH gene, GPD1. The 4 kb of sequence obtained from the 5'-flanking region and first exon of GPD1 was compared with the corresponding mouse sequence. Both sequences share a HindIII site located in what has proven to be the highly conserved 3' untranslated region of an upstream gene of unknown function, D15Kzl. The 3.6-kilobase segment of mouse DNA located between D15Kzl and Gdc-1 was provisionally termed the GPDH promoter. Alignment of the mouse promoter with the corresponding human sequence revealed two conserved domains. An upstream distal promoter region is approximately 900 base pairs in length. A downstream or proximal promoter region consists of approximately 300 base pairs immediately upstream of a TATA-like box and contains the fat-specific elements 1 and 2. Analysis of the chromatin structure of the Gdc-1 promoter revealed four DNase I-hypersensitive sites. They were present in DNA of liver and brown fat, in which GPDH expression is high, but were absent in DNA of spleen, in which GPDH expression is low. Methylation studies of the promoter showed it to be heavily methylated in sperm. However, the DNA from each adult somatic tissue had a unique distribution of nonmethylated sites and could easily be identified by its methylation pattern. These data suggest a structural model of the promoter that explains how Gdc-1 expression is differentially regulated in many types of cells.

Sequence conservation and structural organization of the glycerol-3-phosphate dehydrogenase promoter in mice and humans

Cloned segments of the mouse glycerol-3-phosphate dehydrogenase (GPDH) gene, Gdc-1, were used to screen a human library. Human clones obtained spanned 25 kilobases of genomic DNA containing the human GPDH gene, GPD1. The 4 kb of sequence obtained from the 5'-flanking region and first exon of GPD1 was compared with the corresponding mouse sequence. Both sequences share a HindIII site located in what has proven to be the highly conserved 3' untranslated region of an upstream gene of unknown function, D15Kzl. The 3.6-kilobase segment of mouse DNA located between D15Kzl and Gdc-1 was provisionally termed the GPDH promoter. Alignment of the mouse promoter with the corresponding human sequence revealed two conserved domains. An upstream distal promoter region is approximately 900 base pairs in length. A downstream or proximal promoter region consists of approximately 300 base pairs immediately upstream of a TATA-like box and contains the fat-specific elements 1 and 2. Analysis of the chromatin structure of the Gdc-1 promoter revealed four DNase I-hypersensitive sites. They were present in DNA of liver and brown fat, in which GPDH expression is high, but were absent in DNA of spleen, in which GPDH expression is low. Methylation studies of the promoter showed it to be heavily methylated in sperm. However, the DNA from each adult somatic tissue had a unique distribution of nonmethylated sites and could easily be identified by its methylation pattern. These data suggest a structural model of the promoter that explains how Gdc-1 expression is differentially regulated in many types of cells.

Isolation and characterization of a family of rat endogenous retroviral sequences

The complete nucleotide sequence of one representative rat genomic unit flanked on both sides with RAL elements, which have structural features specific to retroviral LTRs (1), was determined. The total unit was about 7.5 kbp long, and there was a partial homology to known retroviral sequences in gag, pol, and env regions. The sequence also contained minus- and plus-strand primer binding sites, thereby indicating a retroviral nature in replication. Transcription of the sequence was extensive in tumor cells and was strongly correlated with the state of methylation within 5' LTRs, which were highly methylated in the normal but not in the tumor state. In functional assays with bacterial chloramphenicol acetyltransferase constructs containing a series of deleted LTRs, there seemed to be both positive and negative cis-acting effector sequences.

Reactivation of a methylation-silenced gene in adenovirus-transformed cells by 5-azacytidine or by E1A trans activation

In the adenovirus type 2 (Ad2)-transformed hamster cell line HE3, the integrated late E2A promoter of Ad2 DNA is inactive, is methylated at all three 5'-CCGG-3' sequences, and can be reactivated by growing the cells in the presence of 50 microM 5-azacytidine (5-azaC). The three 5'-CCGG-3' sequences then become demethylated. Demethylation and reactivation are stable over 30 passages even after the removal of 5-azaC. The dormant late E2A promoter in cell line HE3 can also be reactivated by transfecting the cells with recombinant plasmids that carry the left terminal E1A and part of the E1B region of Ad2 DNA or the E1A 13S cDNA, but not with plasmids containing the E1A 12S cDNA. The E1A 13S cDNA encodes the 289-amino-acid trans-activating protein of Ad2. The E1A-mediated reactivation of the late E2A promoter is not accompanied by its demethylation in both DNA complements. Cell line HE3 produces constitutively E1A-encoded mRNAs and reactivates the methylated late E2A promoter-chloramphenicol acetyltransferase gene construct after transfection into HE3 cells. Constitutive levels of the endogenous E1A gene products in HE3 cells are detectable but, paradoxically, appear insufficient to reactivate the endogenous, chromosomally integrated E2A gene.

Differential activation of the hprt gene on the inactive X chromosome in primary and transformed Chinese hamster cells

We have investigated the genetic activation of the hprt (hypoxanthine-guanine phosphoribosyltransferase) gene located on the inactive X chromosome in primary and transformed female diploid Chinese hamster cells after treatment with the DNA methylation inhibitor 5-azacytidine (5azaCR). Mutants deficient in HPRT were first selected by growth in 6-thioguanine from two primary fibroblast cell lines and from transformed lines derived from them. These HPRT- mutants were then treated with 5azaCR and plated in HAT (hypoxanthine-methotrexate-thymidine) medium to select for cells that had reexpressed the hprt gene on the inactive X chromosome. Contrary to previous results with primary human cells, 5azaCR was effective in activating the hprt gene in primary Chinese hamster fibroblasts at a low but reproducible frequency of 2 x 10(-6) to 7 x 10(-6). In comparison, the frequency in independently derived transformed lines varied from 1 x 10(-5) to 5 x 10(-3), consistently higher than in the nontransformed cells. This increase remained significant when the difference in growth rates between the primary and transformed lines was taken into account. Treatment with 5azaCR was also found to induce transformation in the primary cell lines but at a low frequency of 4 x 10(-7) to 8 x 10(-7), inconsistent with a two-step model of transformation followed by gene activation to explain the derepression of hprt in primary cells. Thus, these results indicate that upon transformation, the hprt gene on the inactive Chinese hamster X chromosome is rendered more susceptible to action by 5azaCR, consistent with a generalized DNA demethylation associated with the transformation event or with an increase in the instability of an underlying primary mechanism of X inactivation.

Activation and demethylation of the intracisternal A particle genes by 5-azacytidine

Treatment of C3H 10T1/2 mouse embryo fibroblasts with the cytidine analogues 5-azacytidine and 5-aza-2-deoxycytidine causes altered gene expression and results in the manifestation of phenotypic changes and altered cell morphology. This includes the conversion of these cells to adipocytes, chondrocytes and myotubes. The effects of these analogues on intracisternal A particle (IAP) gene expression in mouse C3H 10T1/2 cells have been examined. Treatment with either 3 microM 5-azacytidine or 0.3 microM 5-aza-2-deoxycytidine for 24 h was associated with an immediate increase in IAP gene transcription, and with the subsequent appearance of IAPs in the cisternae of the cells 24 h after removal of the drug. Control cells contained no, or very few, IAPs and IAP mRNA. Analysis of the methylation status of the IAP genes, using the restriction endonucleases HpaII, MspI and HhaI, showed that these genes were already demethylated at the end of the 24-h treatment period. IAP gene transcripts were detectable even after a 16-h drug treatment period, at which stage the genes were not yet fully demethylated. After further growth in fresh medium for 90 h, the levels of IAP RNA started to decline, but the demethylated CpG sites were not yet remethylated. These results suggest the involvement of other factors, in addition to methylation, in the regulation of IAP gene expression. These drugs were found to have no stimulatory effect on several oncogenes examined in this study.

Gene structure and transcription in mouse cells with extensively demethylated DNA

In previous work, three clonal cell lines with extremely low DNA methylation levels were derived by multiple consecutive treatments of C3H 10T1/2 C18 (10T1/2) cells with 5-aza-2'-deoxycytidine (5-aza-CdR). In this study we examined the methylation status of genes in these three methyl-deficient clones to assess the specificity of the induced hypomethylation. Complete demethylation of virtually all 5'-CCGG-3' sites was observed in four genes examined, but some sites common to all three clones were persistently methylated even after further exhaustive 5-aza-CdR treatment. Thus, there is a subset of methylation sites within these cells which can never be stably demethylated. The extensive demethylation was not always associated with changes in the level of RNA expression of the genes examined but was strongly correlated with an altered chromatin structure of the unexpressed alpha 1-globin gene and the muscle determination gene MyoD1. These results provide a direct correlation between hypomethylation and the induction of a transcriptionally competent chromatin state.

Physical linkage of mouse lambda genes by pulsed-field gel electrophoresis suggests that the rearrangement process favors proximate target sequences

The first complete map of a mammalian immunoglobulin gene locus is presented. Mouse lambda genes were mapped by pulsed-field gel electrophoresis. The gene order is V2-Vx-C2-C4-V1-C3-C1. The distance between V2 or Vx and the C2-C4 cluster is 74 or 55 kilobases (kb), respectively, whereas that between V1 and C3-C1 is only 19 kb; V2 and C3-C1 are at least 190 kb apart. Thus, the distances between the lambda subloci are inversely proportional to their frequencies of rearrangement. The related gene lambda 5 is not within the 500 kb of the lambda locus mapped here.

Differential activation of the hprt gene on the inactive X chromosome in primary and transformed Chinese hamster cells

We have investigated the genetic activation of the hprt (hypoxanthine-guanine phosphoribosyltransferase) gene located on the inactive X chromosome in primary and transformed female diploid Chinese hamster cells after treatment with the DNA methylation inhibitor 5-azacytidine (5azaCR). Mutants deficient in HPRT were first selected by growth in 6-thioguanine from two primary fibroblast cell lines and from transformed lines derived from them. These HPRT- mutants were then treated with 5azaCR and plated in HAT (hypoxanthine-methotrexate-thymidine) medium to select for cells that had reexpressed the hprt gene on the inactive X chromosome. Contrary to previous results with primary human cells, 5azaCR was effective in activating the hprt gene in primary Chinese hamster fibroblasts at a low but reproducible frequency of 2 x 10(-6) to 7 x 10(-6). In comparison, the frequency in independently derived transformed lines varied from 1 x 10(-5) to 5 x 10(-3), consistently higher than in the nontransformed cells. This increase remained significant when the difference in growth rates between the primary and transformed lines was taken into account. Treatment with 5azaCR was also found to induce transformation in the primary cell lines but at a low frequency of 4 x 10(-7) to 8 x 10(-7), inconsistent with a two-step model of transformation followed by gene activation to explain the derepression of hprt in primary cells. Thus, these results indicate that upon transformation, the hprt gene on the inactive Chinese hamster X chromosome is rendered more susceptible to action by 5azaCR, consistent with a generalized DNA demethylation associated with the transformation event or with an increase in the instability of an underlying primary mechanism of X inactivation.

Segregation, viral phenotype, and proviral structure of 23 avian leukosis virus inserts in the germ line of chickens

We have artificially introduced 23 avian leukosis virus (ALV) proviral inserts into the chicken germ line by injection of wild-type and recombinant subgroup A ALV near the blastoderm of fertile eggs just before incubation. Eight viremic males were identified as germline mosaics because they transmitted proviral DNA to their generation 1 (G-1) progeny at a low frequency. Eleven female and 9 male G-1 progeny carried 23 distinct proviruses that had typical major clonal proviral-host DNA junction fragments detectable after digestion of their DNA with SacI, Southern blotting and hybridization with a probe representing the complete ALV genome. These proviruses, identified by their typical proviral-host DNA junction fragments, were transmitted to approximately 50% of their G-2 progeny after mating the G-1 parents to a line of chickens lacking endogenous ALV proviral inserts. One G-1 female carried 2 proviruses and another 3. The proviruses appeared to be scattered throughout the genome. One of the 14 proviruses carried by females was on the sex (Z) chromosome. Two of the 3 proviruses carried by a single G-1 female were linked with a recombination frequency of about 0.20. Twenty-one of the proviruses coded for infectious ALV. Two proviruses coded for envelope glycoprotein, and cell cultures carrying them were relatively resistant to subgroup A sarcoma virus, but failed to produce infectious ALV. One of these proviruses coded for internal gag proteins, had a deletion in pol, but produced non-infectious virus particles. The other failed to code for gag proteins and had no detectable internal deletions nor did it produce virus particles. Thus, we have shown that replication-competent ALV can artificially infect germ-line cells and that spontaneous defects in the inherited proviruses occur at a rather low rate.

Gene structure and transcription in mouse cells with extensively demethylated DNA

In previous work, three clonal cell lines with extremely low DNA methylation levels were derived by multiple consecutive treatments of C3H 10T1/2 C18 (10T1/2) cells with 5-aza-2'-deoxycytidine (5-aza-CdR). In this study we examined the methylation status of genes in these three methyl-deficient clones to assess the specificity of the induced hypomethylation. Complete demethylation of virtually all 5'-CCGG-3' sites was observed in four genes examined, but some sites common to all three clones were persistently methylated even after further exhaustive 5-aza-CdR treatment. Thus, there is a subset of methylation sites within these cells which can never be stably demethylated. The extensive demethylation was not always associated with changes in the level of RNA expression of the genes examined but was strongly correlated with an altered chromatin structure of the unexpressed alpha 1-globin gene and the muscle determination gene MyoD1. These results provide a direct correlation between hypomethylation and the induction of a transcriptionally competent chromatin state.

Chromatin structure and levels of expression and DNA methylation in the E3 region of chromosomally integrated adenovirus type 12 DNA

Integrated adenovirus type 12 (Ad12) genes in Ad12-transformed cell lines were investigated for chromatin structure, expression levels and states of DNA methylation. The E3 region in the Ad12-transformed cell line HA12/7 is hypermethylated and not expressed. The same region in the Ad12-transformed hamster cell lines T637 and A2497-3 is transcribed and undermethylated (Kruczek, I. and Doerfler, W. (1982) EMBO J. 1, 409-414). There was no significant difference in the DNase I sensitivity of the E3 region when nuclei of the aforementioned cell lines were incubated with this nuclease. In contrast, incubation of these nuclei with the restriction endonuclease PstI and subsequent cleavage of the DNA with BamHI generated an additional 0.9 kbp fragment in T637 and A2497-3 DNA which was not observed after treating HA12/7 nuclei and DNA in the same way. This finding was interpreted as indicative of differences in the chromatin structure of the E3 region depending on its state of transcriptional activity and its level of methylation. The E1 and major late promoter regions, which were transcriptionally active and inactive, respectively, in all three cell lines investigated, did not exhibit differences in sensitivity towards DNase I or PstI treatment of nuclei. More refined technology will be required to compare the chromatin structure of active versus inactive genes.

DNA methylation and differentiation

The methylation of specific cytosine residues in DNA has been implicated in regulating gene expression and facilitating functional specialization of cellular phenotypes. Generally, the demethylation of certain CpG sites correlates with transcriptional activation of genes. 5-Azacytidine is an inhibitor of DNA methylation and has been widely used as a potent activator of suppressed genetic information. Treatment of cells with 5-azacytidine results in profound phenotypic alterations. The drug-induced hypomethylation of DNA apparently perturbs DNA-protein interactions that may consequently alter transcriptional activity and cell determination. The inhibitory effect of cytosine methylation may be exerted via altered DNA-protein interactions specifically or may be transduced by a change in the conformation of chromatin. Recent studies have demonstrated that cytosine methylation also plays a central role in parental imprinting, which in turn determines the differential expression of maternal and paternal genomes during embryogenesis. In other words, methylation is the mechanism whereby the embryo retains memory of the gametic origin of each component of genetic information. A memory of this type would probably persist during DNA replication and cell division as methylation patterns are stable and heritable.

Lack of evidence for a role of T-cell-associated retroviruses as an etiology of schizophrenia

The evidence that schizophrenia may involve infection by a virus (or viruses) has been indirect. The recent discovery, however, of the human retroviruses--human T-cell lymphoma-leukemia virus-I, and II (HTLV-I, -II) and human immunodeficiency virus (HIV)--now also known to affect the central nervous system (CNS), together with the development of new techniques in retrovirology, have made it possible to investigate more directly the role of this class of viruses as an etiology of schizophrenia. In our first effort to screen for the presence of a T-cell lymphotropic virus in schizophrenia, short-term tissue cultures of peripheral lymphocytes from 17 chronic schizophrenic patients and 10 normal controls were established. The cells were cultured in the presence of T-cell growth factor (TCGF, IL-2), and the culture supernatants were tested for the presence of the retroviral enzyme reverse transcriptase. No T-cell-associated reverse transcriptase activity was detected in cultures from patients or normal controls. Therefore, the data do not provide evidence for a role for T-cell lymphotropic retroviruses as an etiology of schizophrenia.

Physical linkage of mouse lambda genes by pulsed-field gel electrophoresis suggests that the rearrangement process favors proximate target sequences

The first complete map of a mammalian immunoglobulin gene locus is presented. Mouse lambda genes were mapped by pulsed-field gel electrophoresis. The gene order is V2-Vx-C2-C4-V1-C3-C1. The distance between V2 or Vx and the C2-C4 cluster is 74 or 55 kilobases (kb), respectively, whereas that between V1 and C3-C1 is only 19 kb; V2 and C3-C1 are at least 190 kb apart. Thus, the distances between the lambda subloci are inversely proportional to their frequencies of rearrangement. The related gene lambda 5 is not within the 500 kb of the lambda locus mapped here.

Incidence of endogenous viral genes in two strains of white leghorn chickens selected for egg production and susceptibility or resistance to Marek's disease

Endogenous viral (ev) genes related to the avian leukosis virus were classified in two differentially selected strains of Leghorns in order to investigate whether such genes affect production traits. Strain K had been selected for resistance to Marek's disease (MD) and for high egg production and egg weight, whereas strain S had been selected only for MD susceptibility. Except that founders of strain K included a few commercial birds, both strains were derived from a common genetic base. DNA restriction fragment length analyses of 110 strain K and 94 strain S birds revealed the presence of 8 different ev-genes, 6 of which were identical to previously identified loci. This result was confirmed by assays for group specific antigen (gs-antigen), the product of the gag region of the ev-genes. The levels of gs-antigen in the birds closely followed what had been predicted from data obtained from previously described ev-genes. Both strains had a similar average number of ev-genes per bird (3.5 and 3.2 for strains S and K, respectively). However, strain K carried only five different ev-genes while strain S carried seven. Four of these loci were present in both strains. Among the ev-genes absent or occurring less frequently in strain K were those that code either for infectious endogenous virus (ev-10 and possibly ev-19) or for the internal viral gag-proteins (ev-3). Only those ev-genes which are transcriptionally silent or which code for the viral envelope gene were present in increased frequencies in strain K. The results indicate that selection for egg traits and/or Marek's disease resistance reduces the frequency of ev-genes which produce endogenous virus or the viral gag-proteins.

Methylation is co-ordinated on the putative replication origins of Physarum ribosomal DNA

In Physarum polycephalum, the ribosomal DNA is found as 60,000 base-pair palindromes. Each rDNA has four symmetrically arranged replication origins flanked by ribosomal RNA genes. A particular sequence, the putative replication origin, is repeated at the approximate position of each origin and nowhere else in the molecule. On a typical rDNA molecule, only one origin is active per replication cycle. We show that both the level and co-ordination of methylation result in asymmetrically methylated rDNA molecules that are particularly hypomethylated at one of their four putative replication origins. This pattern of methylation on a typical rDNA molecule is consistent with a model where hypomethylation is a determinant of origin activity.

Deoxyribonuclease I sensitivity of the T-DNA ipt gene is associated with gene expression

We have analyzed the chromatin structure of the T-DNA isopentenyl transferase gene, ipt, in four Nicotiana tabacum crown gall tumor lines. These four transformed lines contain identical T-DNA inserts and are derivatives of a single clone that did not exhibit any tumorous properties and contained a highly methylated, nonexpressed copy of T-DNA. One of the derivatives also does not exhibit tumorous properties, and the T-DNA of this line is not expressed. The other three lines have reverted to tumorous growth either spontaneously or after treatment with the inhibitor of DNA methylation, 5-azacytidine. Concomitant with this reversion to tumorous growth, expression of the ipt gene of these lines has reinitiated. In the lines that express the ipt gene, the chromatin structure of this gene exists in a conformation that is more accessible to DNase I than in the line in which this gene is not expressed. The level of ipt expression and DNase I sensitivity was independent of the process by which the transformed cell lines reverted to tumorous growth. The relationship of chromatin structure to gene expression and DNA methylation in these lines is discussed.

Effects of cytidine analogs on methylation of DNA and retrovirus induction

Several cytidine analogs with known mutagenic capability were tested for their effects on DNA methylation and on induction of endogenous murine retrovirus. For each of the compounds tested it was found that DNA methylation was inhibited at the same concentrations that were required to induce virus expression. With each compound it was observed that increased dose levels produced an increase in the ability to inhibit methylation and an increase in the ability to induce virus.

Absence of expression of a human endogenous retrovirus is correlated with choriocarcinoma

We examined the RNA expression of a human endogenous provirus, termed ERV3, in 170 human tissue and cell specimens. The highest expression was found in normal placental chorionic villi as mRNAs of 9, 7.3, and 3.5 kb. The 7.3-kb RNA species was found only in the placenta. ERV3 mRNA was expressed in most other normal and malignant tissues at a level which was 2-10% of that seen in placenta. However, several tissues and tumor cell lines had higher transcriptional levels, equal to 10-60% of the placental level. In contrast, an almost complete abrogation of ERV3 mRNA expression was noted in choriocarcinoma cell lines and in an invasive hydatidiform mole tissue biopsy. This abrogation was not linked to deletions or rearrangements of the ERV3 genome. It appeared to be unassociated with methylation because the ERV3 provirus was similarly methylated in the DNA of placental chorionic villi and choriocarcinoma cells, and ERV3 transcription in choriocarcinoma cells was not induced by 5-azacytidine. These results suggest that the loss of ERV3 mRNA expression is associated with susceptibility to choriocarcinoma.

Regulated expression of tumor-specific and MHC class I antigens in an ultraviolet radiation-induced murine skin cancer

Previous studies have shown that the ultraviolet (UV) radiation-induced tumor UV-1591 expresses a novel class I antigen that functions as a tumor-specific transplantation antigen (TSTA). The purpose of this study was to determine whether expression of novel class I TSTA and expression of endogenous class I antigens in UV-1591 are regulated independently or coordinately. Expression of both TSTA and H-2Kk antigens increased simultaneously after treatment of UV-1591 cells with either 5-azadeoxycytidine (5-azaC) or interferon-gamma (IFN-gamma). Furthermore, the kinetics and pattern of increased expression by 5-azaC and IFN-gamma were identical for both TSTA and H-2Kk antigens. The increased expression of TSTA and class I antigens induced by 5-azaC treatment was constitutive and did not require the continued presence of the drug. In contrast, the increased expression after IFN-gamma treatment was transient and required the continued presence of IFN-gamma. Neither 5-azaC nor IFN-gamma caused induction of class II antigens in UV-1591 tumor cells. These results indicate that expression of TSTA and class I antigens in UV-1591 tumor is coordinately regulated, whereas expression of class I and class II antigens is independently regulated.

A single point mutation in the envelope gene is responsible for replication and XC fusion deficiency of the endogenous ecotropic C3H/He murine leukemia virus and for its repair in culture

The molecular basis has been determined for differences in infectivity and XC phenotype of endogenous ecotropic murine leukemia virus of the low-leukemia mouse strain C3H/He, its relative in the high-leukemia mouse strain AKR, and highly infectious, XC-positive C3H virus variants selected in vitro. Endogenous ecotropic type C virus induced by iododeoxyuridine from the nontransformed C3H/10T1/2 cell line is XC negative and replication deficient. In contrast, viruses produced late after iododeoxyuridine induction in chemically transformed C3H/10T1/2 cells (MCA5) are XC positive and infectious. XC-negative viruses can be converted to XC-positive viruses by being grown in certain transformed cell lines. We have cloned the endogenous ecotropic provirus of C3H/He from MCA5 cells, which is XC negative and replication deficient, as well as two XC-positive C3H proviruses derived by in vitro conversion. Fragment exchange between the XC-negative molecular clone p110 and the XC-positive AKR virus clone p623 revealed that the defect in p110 lies 3' of the SalI site located in the pol region. Nucleotide sequencing established that the C3H p110 provirus was integrated within the R region of an endogenous VL30 long terminal repeat (LTR) in reverse orientation and that the virus differed from the infectious AKR p623 provirus by a point mutation, substituting Lys for Arg at the potential precursor cleavage site for gp70 and p15E. In vitro-converted XC-positive C3H proviral clones 3211 and 4211 are identical to XC-negative C3H p110, except that they have Arg at this site and the normal cleavage site is thus regenerated in these clones. The XC-negative C3H p110 was blocked in processing of Pr85env, whereas clones 3211 and 4211 had normal cleavage of the env precursor into gp70. Both the XC-negative C3H provirus and the in vitro-converted XC-positive C3H proviruses had a single copy of a 99-base-pair enhancer element in the LTR, whereas two copies of this sequence are present in the AKR proviral LTR. Substitution of Arg for Lys at the envelope precursor processing site of C3H p110 by site-directed mutagenesis is sufficient by itself to convert the virus to the XC-positive replication-competent phenotype. Thus, we have established that a single point mutation at the processing site of the envelope precursor protein Pr85 is responsible for the difference in the infectivity and XC phenotype of endogenous ecotropic murine leukemia virus from C3H/He and AKR mice and that the basis for in vitro conversion is a mutation at this site.

Poorly expressed endogenous ecotropic provirus of DBA/2 mice encodes a mutant Pr65gag protein that is not myristylated

DBA/2 mice carry a single endogenous ecotropic murine leukemia provirus designated Emv-3. Although this provirus appears to be nondefective by genomic restriction enzyme mapping, weanling mice do not produce virus and only about one-third of adult mice ever express virus. 5-Iododeoxyuridine and 5-azacytidine, two potent inducers of ecotropic virus expression, are relatively ineffective at inducing Emv-3 expression. However, the chemical carcinogen 7,12-dimethylbenz(a)anthracene can induce ecotropic virus expression in approximately 95% of treated DBA/2 mice. Previous experiments involving DNA transfection and marker rescue analysis of molecularly cloned Emv-3 DNA suggested that Emv-3 carries a small defect(s) in the gag gene, not detectable by restriction enzyme mapping, that inhibits virus expression in vivo and in vitro. Using a combination of approaches, including DNA sequencing, peptide mapping, and metabolic labeling of cells with [3H]myristate, we have demonstrated that the defect in Emv-3 most likely results from a single nucleotide substitution within the gene for p15gag that inhibits myristylation of the Pr65gag N terminus. Myristylation of Pr65gag is thought to be required for this protein to associate with the plasma membrane and is essential for virus particle formation. These results provide a conceptual framework for understanding how Emv-3 expression is regulated during development and after chemical induction.

Tightly-bound non-histone proteins in different nucleosome-like subpopulations from pig kidney chromatin

By differential sucrose gradient centrifugation of pig kidney chromatin in the presence or absence of Na-EDTA and under varying ionic strength conditions, three nucleosome-like subpopulations with different buoyant densities can be obtained. These particles, on the basis of their histones and HMG protein pattern, of the 5-methylcytosine level of their DNA and of the RNA polymerase activity associated with them, can be considered as originating from chromatin fractions differently involved in gene expression. Two-dimensional electrophoresis of the tightly-bound non-histone proteins shows a distinct pattern for each subpopulation, such protein components being notably present in restricted numbers but in high amounts in the subpopulation which was apparently derived from condensed heterochromatin.

Expression of a single transfected cDNA converts fibroblasts to myoblasts

5-azacytidine treatment of mouse C3H10T1/2 embryonic fibroblasts converts them to myoblasts at a frequency suggesting alteration of one or only a few closely linked regulatory loci. Assuming such loci to be differentially expressed as poly(A)+ RNA in proliferating myoblasts, we prepared proliferating myoblast-specific, subtracted cDNA probes to screen a myocyte cDNA library. Based on a number of criteria, three cDNAs were selected and characterized. We show that expression of one of these cDNAs transfected into C3H10T1/2 fibroblasts, where it is not normally expressed, is sufficient to convert them to stable myoblasts. Myogenesis also occurs, but to a lesser extent, when this cDNA is expressed in a number of other cell lines. The major open reading frame encoded by this cDNA contains a short protein segment similar to a sequence present in the myc protein family.

Selective activation of endogenous ecotropic retrovirus in hematopoietic tissues of B6C3F1 mice during the preleukemic phase of 1,3-butadiene exposure

1,3-Butadiene (BD), a comonomer used in the production of synthetic rubber, is a rodent carcinogen. We have observed a marked increase in the incidence of thymic lymphoma in male B6C3F1 relative to NIH Swiss mice chronically exposed to BD in the absence of demonstrable differences in bone marrow (target organ) toxicity. Increased expression of murine leukemia virus (MuLV) antigens was also observed on lymphomas from BD-exposed B6C3F1 mice. Because NIH Swiss mice do not usually express endogenous retroviruses and their ecotropic proviral sequences are not intact, these findings provide presumptive evidence of a role for endogenous retrovirus sequences in BD-induced lymphoma in the B6C3F1 mouse. The present study was conducted to examine the expression and behavior of endogenous retroviruses in these strains during the preleukemic phase of BD exposure. Chronic exposure to BD (1250 ppm) 6 hr/day, 5 days/wk for 3 to 21 weeks increased markedly the quantity of ecotropic retrovirus recoverable from bone marrow, thymus, and spleen of B6C3F1 mice. However, expression of other endogenous retroviruses (xenotropic, MCF-ERV) was not enhanced. No viruses of any type were found in similarly treated NIH Swiss mice. The mechanism of this increase in ecotropic retrovirus in B6C3F1 mice is believed to be de novo activation in greater numbers of cells because changes in the Fv-1 tropism of the replicating viruses or changes in Fv-1 host restriction were not found. Endogenous retroviruses are thus implicated in BD-induced leukemogenesis in B6C3F1 mice. Further studies will examine the role of retrovirus in BD-induced leukemogenesis and the mechanisms of activation of ecotropic proviral sequences in murine cells.

The methylation state of the proviruses in avian sarcoma virus transformed chick and rat cells

The endogenous viruses in the avian cells are not completely methylated, nor are the Schmidt-Ruppin RSV-D (SRD) proviruses in the infected cells completely unmethylated. Avian sarcoma proviruses integrated in rat transformed cloned cells are heavily methylated. In these cells, a region in the 3' end of the env gene is unmethylated in all the src-containing proviruses but not in the transformed defective (td) proviruses. A possible role for the hypomethylation of the 3' end of the env region is proposed.

Gene reactivation: a tool for the isolation of mammalian DNA methylation mutants

We report the isolation and characterization of a mammalian strain (tsm) that has a temperature-sensitive mutation in DNA methylation. The isolation procedure was based on the observation that treatment of a CHO TK- MT- cell line with demethylating agents introduces up to 46% demethylation, resulting in phenotypic reversion and transcriptional activation of the thymidine kinase (TK) and metallothionein (MT) genes at frequencies ranging from 1% to 59%. Seven thousand individual colonies from an EMS-mutagenized CHO TK- MT- population were screened for spontaneous reversion to TK+ phenotype after treatment at 39 degrees C. Successful isolates were subsequently examined for MT+ reversion. A single clone (tsm) was obtained that showed temperature-dependent reactivation of both TK and MT genes at frequencies of 7.2 X 10(-4) and 6 X 10(-4), respectively. The tsm cells were viable at 39 degrees C and showed no increased mutation frequency. Reactivation correlated with transcriptional activation of the respective genes, whereas backreversion to the TK- phenotype was associated with transcriptional inactivation. TK- backrevertants were reactivable again with demethylating agents. Although demethylation in tsm cells was not detectable by HPLC, Southern blot analysis revealed that reactivants, irrespective of their mode of generation, showed specific demethylation of both TK and MT genes. Also, after about 150 cell generations after treatment, reactivants from both temperature-induced tsm and cells exposed to demethylating agents gained 60% and 23%, respectively, in 5-methylcytosine (5mC). It is proposed that the phenotype of tsm cells is due to a mutation involved in the regulation of DNA methylation. The further characterization of this and other mammalian mutants should help to clarify the physiological role of DNA methylation, as well as its regulation.

The inheritance of epigenetic defects

Evidence from many sources shows that the control of gene expression in higher organisms is related to the methylation of cytosine in DNA, and that the pattern of methylation is inherited. Loss of methylation, which can result from DNA damage, will lead to heritable abnormalities in gene expression, and these may be important in oncogenesis and aging. Transformed permanent lines often lose gene activity through de novo methylation. It is proposed that epigenetic defects in germline cells due to loss of methylation can be repaired by recombination at meiosis but that some are transmitted to offspring.

Characterization of two solitary long terminal repeats of murine leukemia virus type that are conserved in the chromosome of laboratory inbred mouse strains

Twenty molecular clones containing sequences homologous to the long terminal repeats (LTRs) of the endogenous ecotropic murine leukemia virus (MuLV) of the RFM/Un mouse were isolated from a library of RFM/Un mouse spleen DNA in phage lambda. Three of these LTRs were not associated with any viral structural genes. Nucleotide sequence analysis demonstrated that they were solitary LTRs which were flanked by 4-bp directly repeated cellular sequences and which lacked primer binding sites. Two of the three subclones were found to be identical except for their orientations in the vector pBR322. Unique-sequence regions on either side of the two nonidentical elements were used to characterize their integration sites in genomic DNA. The solitary LTRs and their flanking regions were found to be conserved in a number of inbred mouse strains, including three strains known not to harbor endogenous ecotropic MuLV-type proviruses. Comparison of cleavage by the methylation-sensitive restriction enzyme SmaI and methylation-insensitive KpnI at the characteristic LTR SmaI/KpnI site suggested that at least one of these solitary LTRs is methylated to a lesser extent than are most endogenous proviral LTRs. These particular solitary LTRs, like endogenous proviral sequences, appear to be stably transmitted genetic elements.

Methylation of the enhancer region of avian sarcoma virus long terminal repeat suppresses transcription

The effect of methylation of an enhancer on transcription was studied. A 245 bp enhancer-containing a fragment of the LTR of the avian sarcoma virus was methylated in vitro and ligated back into a vector which lacked the upstream enhancer sequence. The transient expression in QT6 cells indicated that methylation of the enhancer-containing sequence severely reduced the extent of transcription.

Theoretical mechanisms for synthesis of carcinogen-induced embryonic proteins: XVIII. Biomethylation and differentiation

Many reports have appeared describing a direct relationship between hypomethylated states of genes and gene activity. Even after the introduction of viral genomes, these new genes appear to be controlled by specific DNA methylations. A variety of other studies have shown chromatin structural changes being implicated in the activities of certain gene loci. Modifications of chromatin domains may also be initiated or under the control of methylation reactions. Embryonic genes may be controlled by particular methylations by virtue of a differential (hyper-) sensitivity to concentrations of active methyl groups, on a variety of chromatin domains thereby explaining the variation in S-adenosyl-L-methionine synthesis required in developing liver tissue. Also our finding of the ability to manipulate experimentally the activity of the alpha-fetoprotein gene by methyl group availability indicates some methyl-sensitive mechanism is operating with respect to embryonic genes.

Albumin and alpha-fetoprotein gene transcription in rat hepatoma cell lines is correlated with specific DNA hypomethylation and altered chromatin structure in the 5' region

We examined DNA methylation and DNase I hypersensitivity of the alpha-fetoprotein (AFP) and albumin gene region in hepatoma cell lines which showed drastic differences in the level of expression of these genes. We assayed for methylation of the CCGG sequences by using the restriction enzyme isoschizomers HpaII and MspI. We found two methylation sites located in the 5' region of the AFP gene and one in exon 1 of the albumin gene for which hypomethylation is correlated with gene expression. Another such site, located about 4,000 base pairs upstream from the AFP gene, seems to be correlated with the tissue specificity of the cells. DNase I-hypersensitive sites were mapped by using the indirect end-labeling technique with cloned genomic DNA probes. Three tissue-specific DNase I-hypersensitive sites were mapped in the 5' flanking region of the AFP gene when this gene was transcribed. Similarly, three tissue-specific DNase I-hypersensitive sites were detected upstream from the albumin gene in producing cell lines. In both cases, the most distal sites were maintained after cessation of gene activity and appear to be correlated with the potential expression of the gene. Interestingly, specific methylation sites are localized in the same DNA region as DNase I hypersensitive sites. This suggests that specific alterations of chromatin structure and changes in methylation pattern occur in specific critical regulatory regions upstream from the albumin and AFP genes in rat hepatoma cell lines.

Albumin and alpha-fetoprotein gene transcription in rat hepatoma cell lines is correlated with specific DNA hypomethylation and altered chromatin structure in the 5' region

We examined DNA methylation and DNase I hypersensitivity of the alpha-fetoprotein (AFP) and albumin gene region in hepatoma cell lines which showed drastic differences in the level of expression of these genes. We assayed for methylation of the CCGG sequences by using the restriction enzyme isoschizomers HpaII and MspI. We found two methylation sites located in the 5' region of the AFP gene and one in exon 1 of the albumin gene for which hypomethylation is correlated with gene expression. Another such site, located about 4,000 base pairs upstream from the AFP gene, seems to be correlated with the tissue specificity of the cells. DNase I-hypersensitive sites were mapped by using the indirect end-labeling technique with cloned genomic DNA probes. Three tissue-specific DNase I-hypersensitive sites were mapped in the 5' flanking region of the AFP gene when this gene was transcribed. Similarly, three tissue-specific DNase I-hypersensitive sites were detected upstream from the albumin gene in producing cell lines. In both cases, the most distal sites were maintained after cessation of gene activity and appear to be correlated with the potential expression of the gene. Interestingly, specific methylation sites are localized in the same DNA region as DNase I hypersensitive sites. This suggests that specific alterations of chromatin structure and changes in methylation pattern occur in specific critical regulatory regions upstream from the albumin and AFP genes in rat hepatoma cell lines.

Significance of DNase I-hypersensitive sites in the long terminal repeats of a Moloney murine leukemia virus vector

A Moloney murine leukemia virus-derived retroviral vector (N4) carrying the bacterial neomycin resistance gene (neo) was used to study the chromatin configuration of integrated proviral DNA in NIH 3T3-derived cell lines containing one copy of the vector DNA per cell. Three independently obtained cell lines were examined. In two of these cell lines, the vector was introduced by viral infection, while in the third the construct was introduced by DNA transfection. Such transfected cell lines (including the one examined) usually express 10- to 50-fold less virus-specific RNA than do cell lines obtained by viral infection. All three cell lines exhibited similar patterns of DNase I-hypersensitive (HS) sites. Two strong DNase I HS sites were detected in the 5' long terminal repeat, which contains signals required for proper and efficient initiation of viral transcription. One of these sites was found to overlap the viral enhancer sequences, while the other site mapped very close to the start site for viral transcription. A third HS site was detected in nearby internal viral sequences. Only one HS site was found in the 3' long terminal repeat, which contains the signal(s) required for proper addition of a poly(A) tail to viral transcripts. This HS site was located in the region of the viral enhancer. Several weak DNase I HS sites were also found in the cellular sequences adjacent to the integration sites, at different locations in each cell line analyzed. No common pattern of cellular DNase I HS sites was found. These observations suggest that the 5' and 3' long terminal repeats of integrated retroviral proviruses exhibit different chromatin conformations, possibly reflecting the different functions encoded by the otherwise identical sequences, and the DNase I HS sites detected in these studies reflect only a potential for transcription and are not a reflection of the high transcriptional activity characteristic of retroviruses.