The essentials of DNA methylation

Difference in constitutive heterochromatin behaviour between human amniocytes and lymphocytes detected by a sequential in situ exonuclease III digestion-random primer extension procedure

Fixed chromosomes from human amniotic fluid cells and peripheral blood lymphocytes were digested in situ with exonuclease III and the single stranded DNA obtained was used as template for an in situ random primer extension. Under these conditions an R banding pattern, more evident in lymphocytes than in amniocytes, was obtained. Nevertheless, constitutive heterochromatin of chromosomes 1, 16, Yq, and mainly the pericentromeric region of chromosome 9 was far more intensely labelled in amniocytes than in lymphocytes. Fluorescence in situ hybridisation with a specific classical satellite DNA probe, showed that this differential labelling was dependent on a greater sensitivity of chromosome 9 constitutive heterochromatin to exonuclease III digestion in amniocytes than in lymphocytes, thus indicating qualitative differences in this region between both human cellular materials.

CpG island promoter region methylation patterns of the inactive-X-chromosome hypoxanthine phosphoribosyltransferase (Hprt) gene

Inactive-X-chromosome genes in mammalian females have methylated CpG islands. We have questioned whether there are variable levels of cytosine methylation at different CpG sites within the island that might indicate the presence of primary sites of methylation which may be critical for the maintenance of gene repression and candidate sites for the initiation of inactivation. To address these questions, we have analyzed the methylation patterns of 32 CpG sites of the X-linked hypoxanthine phosphoribosyltransferase (Hprt) gene on the active and inactive X chromosomes of mouse tissues and cell lines, using genomic sequencing of bisulfite-treated genomic DNA. Cytosine is deaminated by bisulfite, but methylcytosine is not affected. Cell lines that were heterozygous for the Hprt deletion mutation (Hprtb-m3) and a functional Hprt allele were selected with 6-thioguanine. The resulting cell populations uniformly carry the intact Hprt allele on the inactive X chromosome. The methylation of these CpG sites was determined either by the direct sequence analysis of bisulfite-treated and amplified DNA or by the sequence analysis of clones derived from the amplified DNA. No CpG methylation was detected on the active Hprt genes from either males or the active X chromosome of females. On average, 22 CpGs were methylated in the other 50% of female DNA, and the level of methylation at individual sites varied from 42 to 100%. Analysis of the inactive Hprt gene in two cell lines showed that averages of 14 and 18 CpGs were methylated and that the frequency of methylation at 32 individual sites ranged from 3 to 100%. The highest frequency of methylation in cell lines coincided with the sequences flanking transcription initiation sites. These results suggest that methylation patterns are heterogeneous within a tissue and even in clonal cell populations and that specific subsets of CpG sites sustain high methylation frequencies which may be critical for the maintenance of X-chromosome inactivation. The bisulfite method identified which CpG sites were methylated on the inactive X chromosome, and it provided a quantitative estimate of the frequency of methylation of these sites in genomic DNA.

Negative and positive effects of CpG-methylation on Xenopus ribosomal gene transcription in vitro

Methylation of cytosine-residues in the sequence CpG affects the expression of many genes and generally correlates with reduced transcription. The ribosomal genes of Xenopus laevis were among the first genes to be studied with respect to their DNA methylation, and a loss of methylation during embryonic development correlated with the onset of transcription. Nevertheless, highly methylated ribosomal genes were transcribed at normal levels when injected into oocyte nuclei, and thus transcription of these genes was generally assumed to be insensitive to CpG-methylation. Here I show that Xenopus ribosomal gene transcription can be repressed by cellular factors binding to meCpG, similarly as it has been described for transcription by RNA polymerase II. In the absence of these repressors, however, CpG-methylation has a direct positive effect on RNA polymerase I-promoter activity.

Methylation of the rat glial fibrillary acidic protein gene shows tissue-specific domains

The gene for glial fibrillary acidic protein (GFAP) was compared for CpG sites that are potential locations of methylated cytosine (mC). GFAP sequences in the 5'-upstream promoter and in exon 1 of rat, mouse, and human showed extensive similarity in the locations of CpG sites in the promoter and in exon 1, implying conservation. The methylation of mC at 9 CpG sites in the promoter and 10 sites in exon 1 was analyzed in F344 male rats by a quantitative application of ligation-mediated polymerase chain reaction (LMPCR). CpG sites with varying mC in different tissues were found in the GFAP promoter and in a CpG island in exon 1. In the brain, the promoter had about 40% less mC than in testis and liver. The degree of methylation varied strikingly between adjacent sites within and between tissues. Testis GFAP exon 1 had a gradient of mC from 5' to 3' across the exon that was absent in liver, brain, and cultured neurons and astrocytes. Among brain regions, the hippocampus had 10-40% less mC at 12 CpG sites than in hypothalamus; the other sites (7/19) showed smaller differences between these brain regions. In DNA from primary cultures, astrocytes had slightly less mC than neurons at all sites. Because neuron-rich hippocampal subregions and primary neurons cultures had less methylation than nonneural tissues, we hypothesize that neuroectodermal derivatives tend to be less methylated, whether or not GFAP is expressed. Four domains of methylated CpG sites are proposed on the basis of tissue and cell-type distribution: I) a constitutively methylated domain in the mid-upstream promoter; II) a testis-specific gradient of methylation in exon 1; III) a hypomethylated domain found in neuroectodermal derivatives; and IV) subsets of sites in the promoter and in exon 1 that have the least methylation in astrocytes, and therefore may be astrocyte-specific domains.

Tissue-specific DNA methylation patterns of the rat glial fibrillary acidic protein gene

The glial fibrillary acidic protein (GFAP) is an intermediate filament protein, specific of the cytoskeleton of astrocytes in the central nervous system. In the present work, as a preliminary step to the study of glial-specific gene expression, we cloned the rat GFAP gene, and we report the sequence of 1.9 kb of the 5' flanking region, exon 1, and the majority of the first intron. By digestion with methylation-sensitive restriction enzymes followed by Southern blot analysis, the methylation status of various CpG sites was examined in this genomic segment. We tested whether structural modification of the GFAP gene, such as DNA methylation, could be related to its tissue-specific transcriptional activity. Therefore, we compared a GFAP-expressing cell population (primary culture of astroglial cells), a mixed population of GFAP-expressing and -nonexpressing cells (adult rat cerebral hemispheres), and a GFAP-nonexpressing tissue (liver). In the 5' flanking region we identified a CpG site at position -1176 whose level of methylation is inversely correlated to GFAP expression. In primary cultured astrocytes, 75% of the GFAP gene alleles were demethylated at this site, while the corresponding value obtained for the cerebral hemispheres was 45%, and for liver only 9%. On the basis of the sequence data, a CpG-rich region (putative CpG island) was identified extending from -38 to +347 and overlapping 80% of the first exon. HhaI and HpaII sites located in the putative CpG island showed a relatively high level of methylation in all the cell populations examined, and did not show any clear correlation with the level of GFAP gene expression or with the methylation status of the -1176 site. Further in vivo developmental studies and in vitro differentiation studies are necessary to better understand the functional differences of the various methylatable CpG sites in the 5' end of the GFAP gene.

CpG island promoter region methylation patterns of the inactive-X-chromosome hypoxanthine phosphoribosyltransferase (Hprt) gene

Inactive-X-chromosome genes in mammalian females have methylated CpG islands. We have questioned whether there are variable levels of cytosine methylation at different CpG sites within the island that might indicate the presence of primary sites of methylation which may be critical for the maintenance of gene repression and candidate sites for the initiation of inactivation. To address these questions, we have analyzed the methylation patterns of 32 CpG sites of the X-linked hypoxanthine phosphoribosyltransferase (Hprt) gene on the active and inactive X chromosomes of mouse tissues and cell lines, using genomic sequencing of bisulfite-treated genomic DNA. Cytosine is deaminated by bisulfite, but methylcytosine is not affected. Cell lines that were heterozygous for the Hprt deletion mutation (Hprtb-m3) and a functional Hprt allele were selected with 6-thioguanine. The resulting cell populations uniformly carry the intact Hprt allele on the inactive X chromosome. The methylation of these CpG sites was determined either by the direct sequence analysis of bisulfite-treated and amplified DNA or by the sequence analysis of clones derived from the amplified DNA. No CpG methylation was detected on the active Hprt genes from either males or the active X chromosome of females. On average, 22 CpGs were methylated in the other 50% of female DNA, and the level of methylation at individual sites varied from 42 to 100%. Analysis of the inactive Hprt gene in two cell lines showed that averages of 14 and 18 CpGs were methylated and that the frequency of methylation at 32 individual sites ranged from 3 to 100%. The highest frequency of methylation in cell lines coincided with the sequences flanking transcription initiation sites. These results suggest that methylation patterns are heterogeneous within a tissue and even in clonal cell populations and that specific subsets of CpG sites sustain high methylation frequencies which may be critical for the maintenance of X-chromosome inactivation. The bisulfite method identified which CpG sites were methylated on the inactive X chromosome, and it provided a quantitative estimate of the frequency of methylation of these sites in genomic DNA.

A model for systemic lupus erythematosus based on chromatin disruption by polyamines

Here it is hypothesized that some autoimmune disorders, such as systemic lupus erythematosus, result in part from overexpression of polyamines which leads to disruption of chromatin structure. Disruption of inactive chromatin, such as the inactive X chromosome, exposes sites of unrepaired DNA damage. Repair is then hampered by the polyamines. Disruption also facilitates transcription at previously sequestered sites. Especially interesting are RNA polymerase III sites in highly repeated sequences such as the Alu sequence. Transcription and translation from these sites could create RNA and polypeptides not normally expressed. These could be antigenic either individually or in association with other cellular components. Interactions of polyamines in the nucleus and with the membrane could also lead to polyamine facilitated apoptosis.

New sites of methylcytosine-rich DNA detected on metaphase chromosomes

In situ immunofluorescence detection of antibodies against 5-methylcytosine on metaphase chromosomes prepared by a new procedure allows the display of new 5-methylcytosine-rich sites as compared to previously published methods. In short-term culture lymphocytes, the immunofluorescent signals give a recurrent pattern in which four types of binding sites can be distinguished. Type I sites are the secondary constrictions and a few juxtacentromeric regions, type II sites correspond to T-bands. Both types I and II sites emit a strong fluorescence. Type III sites form an R-band pattern and emit a weaker fluorescence. Type IV sites are the short arms of acrocentrics, they emit strong but polymorphic signals. The results obtained from control experiments suggest that the pattern observed is rather the expression of an uneven distribution of 5-methylcytosine-rich sites than a consequence of the various treatments used. In a lymphoblastoid cell line known to have a reduced 5-methylcytosine content, it was possible to demonstrate a heterogeneous hypomethylation among chromosome structures, principally involving type I sites. The method opens the possibility of studying in situ on chromosomes, regional variations of methylation in pathological conditions.

DNA methylation during differentiation of a lower eukaryote, Physarum polycephalum

Starvation-induced differentiation of the slime mould Physarum polycephalum is accompanied by continuous methylation of DNA. No stable changes in the overall level of DNA methylation are evident, but a gene known to be transcribed specifically during differentiation is subject to increased methylation. Inhibitors of DNA methylation preclude differentiation of P. polycephalum, although they are only marginally inhibitory to normal growth. Taken together these results indicate that methylation of DNA is involved in differentiation of this lower eukaryote.

Detection of methylcytosine-rich heterochromatin on banded chromosomes. Application to cells with various status of DNA methylation

A procedure including incorporation of 5-bromodeoxyuridine (BrdU) in DNA and a thermal denaturation step was developed to obtain both R-banding and efficient binding of anti-5-methylcytosine antibodies on metaphase chromosomes. BrdU incorporation improved the efficiency of antibody binding disclosed by immunofluorescence staining. This method allowed semiquantitative analysis of the antibody binding sites on straightforward characterized metaphase chromosomes and was applied to normal human lymphocytes and lymphoblastoid cell lines for which DNA methylation status had been previously analyzed. A correlation was established between level of DNA methylation and the semiquantitative estimate of antibody fixation. This procedure can be used to study DNA methylation on metaphase chromosomes in transformed and cancerous cell lines.

Sickle cell anemia and fetal hemoglobin

Fetal hemoglobin, the predominant hemoglobin of the fetus, is good for sickle cell anemia. This hemoglobin inhibits the polymerization of sickle hemoglobin. Clinical studies have shown that at any level of fetal hemoglobin, the more that is present, the better off is the patient. We are now able to increase fetal hemoglobin levels by pharmacologic means. We should know shortly if this is associated with clinical benefit.

Cell death and cell cycle perturbation in the developmental toxicity of the demethylating agent, 5-aza-2'-deoxycytidine

DNA methylation is a probable mechanism for regulating gene expression, and alterations in methylation may significantly affect embryonic development. We administered the cytidine analogue 5-aza-2'-deoxycytidine (dAZA), a specific and potent demethylator of DNA, to pregnant mice to determine its teratogenicity and effects on embryonic cell death and cell cycle. Groups of females were dosed intraperitoneally on gestation day 10 with doses of 0.05-3 mg/kg dAZA and killed at 4, 8, or 28 hr later. Two embryos per litter were immediately stained with Nile blue sulfate (NBS) to identify areas of cell death; the remaining embryos were frozen and stored for subsequent flow cytometric (FCM) analysis of the cellular DNA synthetic cycle in limb buds. A dose-related accumulation of cells in the S and G2/M phases was observed at 4 and 8 hr after maternal dosing. S-phase accumulation was the most sensitive indicator of effect; a dose-related increase in the percentage of hindlimb bud cells in S-phase was evident at all dosages 4 hr after maternal dosing. By 28 hr postdosing, a normal cell cycle phase distribution was observed at doses of < 0.3 mg/kg. However, cell cycle perturbations persisted at higher dosages. NBS staining demonstrated increased cell death in areas of rapid cell division, indicative of replication-associated cytotoxicity, at doses of > or = 0.1 mg/kg. Observation of litters from additional dams killed at term revealed that at dosages of > or = 0.3 mg/kg, cleft palate and hindlimb defects were significantly elevated. In addition, above 0.3 mg/kg, fetal weight was significantly decreased.(ABSTRACT TRUNCATED AT 250 WORDS)

Two mutations causing vitamin D resistant rickets: modelling on the basis of steroid hormone receptor DNA-binding domain crystal structures

OBJECTIVE

Hereditary vitamin D resistant rickets (HVDRR) has been shown to be due to mutations in the gene encoding the vitamin D receptor (VDR). In two patients with the characteristic phenotype we have investigated the functional defect and sequenced the VDR cDNA. We report two new mutations in the DNA binding domain of the VDR gene and we have used the crystallographic structure of the glucocorticoid and oestrogen receptors (GR and ER respectively) as models to explain the stereochemical consequences of these mutations.

DESIGN

Patient and control cell lines prepared from skin fibroblasts were used to measure binding of 1,25-dihydroxyvitamin D3 (1,25(OH)2D3) and functional responses to this hormone. These cells were also used to isolate VDR mRNA from which cDNA was prepared and sequenced. VDR cDNA from affected and control patients was also transfected into receptor defective cells to analyse further functional responses to 1,25(OH)2D3. Computer analysis of mutations in the VDR gene was carried out using the glucocorticoid and oestrogen receptors as model systems.

PATIENTS

Two patients with HVDRR from unrelated families.

MEASUREMENTS

Cytosolic binding and nuclear association of 1,25(OH)2D3 were determined in control and affected patients, and functional response to 1,25(OH)2D3 was assessed by measurement of 25-hydroxyvitamin D-24-hydroxylase activity (24-hydroxylase). VDR cDNA was sequenced and transfected into VDR-deficient CV-1 cells for further analysis of functional response to 1,25(OH)2D3 following cotransfection with a chloramphenicol acetyltransferase (CAT) reporter plasmid.

RESULTS

Cells from HVDRR patients I and II showed detectable numbers of VDR with normal hormone binding. However, unlike controls, the HVDRR cells did not show induction of 24-hydroxylase activity following treatment with 1,25(OH)2D3. Sequencing of cDNA revealed single mutations, in patient I (Phe44-->IIe) and in patient II (Lys42-->Glu). Both these residues are conserved in the steroid/thyroid hormone receptor superfamily and stereochemical analysis has been used to deduce the importance of these amino acids and the deleterious effect of these and other mutations in the DNA-binding domain of the VDR.

CONCLUSIONS

Two new mutations in the vitamin D receptor which cause hereditary vitamin D resistant rickets have been described and using molecular modelling we have been able to analyse the genesis of this inherited disease at the level of stereochemistry.

Densely methylated DNA islands in mammalian chromosomal replication origins

Densely methylated DNA sequence islands, designated DMIs, have been observed in two Chinese hamster cell chromosomal replication origins by using a PCR-based chemical method of detection. One of the origins, oriS14, is located within or adjacent to the coding sequence for ribosomal protein S14 on chromosome 2q, and the other, ori-beta, is approximately 17 kbp downstream of the dhfr (dihydrofolic acid reductase) locus on chromosome 2p. The DMI in oriS14 is 127 bp long, and the DMI in ori-beta is 516 bp long. Both DMIs are bilaterally methylated (i.e., all dCs are modified to 5-methyl dC) only in cells that are replicating their DNA. When cell growth and DNA replication are arrested, methylation of CpA, CpT, and CpC dinucleotides is lost and the sequence islands display only a subset of their originally methylated CpG dinucleotides. Several possible roles for DMI-mediated regulation of mammalian chromosomal origins are considered.

DNA methylation represses the murine alpha 1(I) collagen promoter by an indirect mechanism

Several lines of evidence indicate that DNA methylation plays a role in the transcriptional regulation of the murine alpha 1(I) collagen gene. To study the molecular mechanisms involved, a reporter gene construct containing the alpha 1(I) promoter and part of the first exon linked to the luciferase gene (Col3luc) was methylated in vitro and transfected into murine fibroblasts and embryonal carcinoma cells. Methylation resulted in repression of the alpha 1(I) promoter in both cell types, although it was less pronounced in embryonal carcinoma cells than in fibroblasts. The extent of repression depended on the density of methylation. DNase footprint and mobility shift assays indicated that the trans-acting factors binding to the alpha 1(I) promoter and first exon are ubiquitous factors and that their DNA binding is not inhibited by methylation. Transfection of Col3luc into Drosophila SL2 cells together with expression vectors for the transcription factors Sp1 and NF-1 showed that DNA methylation also inhibits the alpha 1(I) promoter in nonvertebrate cells, although to a much lesser extent than in murine cells. However, Sp1 and NF-1 transactivated the unmethylated and methylated reporter gene in SL2 cells equally well, confirming that these factors can bind and transactivate methylated DNA and indicating that DNA methylation represses the alpha 1(I) promoter by an indirect mechanism. This was further confirmed by cotransfection experiments with unspecific methylated competitor DNA which partially restored the activity of the methylated alpha 1(I) promoter. Our results suggest that DNA methylation can inhibit promoter activity by an indirect mechanism independent of methyl-C-binding proteins and that in vertebrate cells, chromatin structure and methyl-C-binding proteins cooperatively mediate the transcriptional inhibitory effect of DNA methylation.

Densely methylated DNA islands in mammalian chromosomal replication origins

Densely methylated DNA sequence islands, designated DMIs, have been observed in two Chinese hamster cell chromosomal replication origins by using a PCR-based chemical method of detection. One of the origins, oriS14, is located within or adjacent to the coding sequence for ribosomal protein S14 on chromosome 2q, and the other, ori-beta, is approximately 17 kbp downstream of the dhfr (dihydrofolic acid reductase) locus on chromosome 2p. The DMI in oriS14 is 127 bp long, and the DMI in ori-beta is 516 bp long. Both DMIs are bilaterally methylated (i.e., all dCs are modified to 5-methyl dC) only in cells that are replicating their DNA. When cell growth and DNA replication are arrested, methylation of CpA, CpT, and CpC dinucleotides is lost and the sequence islands display only a subset of their originally methylated CpG dinucleotides. Several possible roles for DMI-mediated regulation of mammalian chromosomal origins are considered.

DNA methylation represses the murine alpha 1(I) collagen promoter by an indirect mechanism

Several lines of evidence indicate that DNA methylation plays a role in the transcriptional regulation of the murine alpha 1(I) collagen gene. To study the molecular mechanisms involved, a reporter gene construct containing the alpha 1(I) promoter and part of the first exon linked to the luciferase gene (Col3luc) was methylated in vitro and transfected into murine fibroblasts and embryonal carcinoma cells. Methylation resulted in repression of the alpha 1(I) promoter in both cell types, although it was less pronounced in embryonal carcinoma cells than in fibroblasts. The extent of repression depended on the density of methylation. DNase footprint and mobility shift assays indicated that the trans-acting factors binding to the alpha 1(I) promoter and first exon are ubiquitous factors and that their DNA binding is not inhibited by methylation. Transfection of Col3luc into Drosophila SL2 cells together with expression vectors for the transcription factors Sp1 and NF-1 showed that DNA methylation also inhibits the alpha 1(I) promoter in nonvertebrate cells, although to a much lesser extent than in murine cells. However, Sp1 and NF-1 transactivated the unmethylated and methylated reporter gene in SL2 cells equally well, confirming that these factors can bind and transactivate methylated DNA and indicating that DNA methylation represses the alpha 1(I) promoter by an indirect mechanism. This was further confirmed by cotransfection experiments with unspecific methylated competitor DNA which partially restored the activity of the methylated alpha 1(I) promoter. Our results suggest that DNA methylation can inhibit promoter activity by an indirect mechanism independent of methyl-C-binding proteins and that in vertebrate cells, chromatin structure and methyl-C-binding proteins cooperatively mediate the transcriptional inhibitory effect of DNA methylation.

Dependence of transcriptional repression on CpG methylation density

CpG methylation is known to suppress transcription. This repression is generally thought to be related to alterations of chromatin structure that are specified by the methylation. The nature of these chromatin alterations is unknown. Moreover, it has not been clear if the methylation repression occurs in an all-or-none fashion at some critical methylation density, or if intermediate densities of methylation can give intermediate levels of repression. Here I report a stable episomal system which recapitulates many dynamic features of methylation observed in the genome. I have determined the extent of transcriptional repression as a function of four densities of CpG methylation. I find that the repression is a graded but exponential function of the CpG methylation density such that low levels of methylation yield a 67 to 90% inhibition of gene expression. Higher levels of methylation extinguished gene expression completely. Transcription from methylated minichromosomes can be increased by butyrate treatment, suggesting that histone acetylation can reverse some of the repression specified by the methylated state. Sites of preferential demethylation occurred and may have resulted from transcription factor binding or DNA looping.

Homology-dependent gene silencing in transgenic plants: epistatic silencing loci contain multiple copies of methylated transgenes

Previous work has shown that two homologous, unlinked transgene loci can interact in plant nuclei, leading to non-reciprocal trans-inactivation and methylation of genes at one locus. Here, we report the structure and methylation of different transgene loci that contain the same construct but are variably able to inactivate and methylate a partially homologous, unlinked target locus. Silencing loci comprised multiple, methylated copies of the transgene construct, whereas a non-silencing locus contained a single, unmethylated copy. The correspondence between strength of silencing activity and copy number/degree of methylation was further demonstrated by producing novel alleles of a strong silencing locus: reducing the transgene copy number and methylation within this silencing locus decreased its ability to inactivate the target locus. The strong silencing locus, which was located close to a telomere, trans-inactivated various structural variants of the original target construct, regardless of their location in the genome. This suggests that the silencing locus can scan the entire genome for homologous regions, a process possibly aided by its telomeric location. Our data support the idea that epistatic trans-inactivation of unlinked, homologous transgenes in plants results from a pre-existing epigenetic difference between transgene loci, which is subsequently equalized by "epigene conversion" involving DNA-DNA pairing.

Dependence of transcriptional repression on CpG methylation density

CpG methylation is known to suppress transcription. This repression is generally thought to be related to alterations of chromatin structure that are specified by the methylation. The nature of these chromatin alterations is unknown. Moreover, it has not been clear if the methylation repression occurs in an all-or-none fashion at some critical methylation density, or if intermediate densities of methylation can give intermediate levels of repression. Here I report a stable episomal system which recapitulates many dynamic features of methylation observed in the genome. I have determined the extent of transcriptional repression as a function of four densities of CpG methylation. I find that the repression is a graded but exponential function of the CpG methylation density such that low levels of methylation yield a 67 to 90% inhibition of gene expression. Higher levels of methylation extinguished gene expression completely. Transcription from methylated minichromosomes can be increased by butyrate treatment, suggesting that histone acetylation can reverse some of the repression specified by the methylated state. Sites of preferential demethylation occurred and may have resulted from transcription factor binding or DNA looping.

Molecular-neurobehavioral associations in females with the fragile X full mutation

In this study, young females with the fragile X [fra(X)] full mutation (fM) were assessed using quantitative measures of mutation amplification size (Amp) as well as the ratio of active normal X chromosome to total normal X chromosome (activation ratio-AR). Neurobehavioral assessments of females with the fM were performed and included specific and general measures of cognitive and behavioral/developmental functioning. To investigate molecular-behavioral associations, Amp and AR were used as independent variables, while cognitive and behavioral scores were used as dependent variables. Significant correlations were observed between both molecular variables (Amp and AR) and measures of cognitive functioning, with AR showing the most consistent and robust correlations. As AR increased, overall IQ and specific subtest and area scores from the cognitive tests increased. Conversely, as Amp increased, the overall IQ and specific subtest and area cognitive scores decreased. No significant associations were observed between AR or Amp and measures of behavior or development. The molecular-cognitive associations were generally consistent with the cognitive profile previously described in studies comparing females with fra(X) to age-matched controls. Amp and AR were not associated with one another, nor were they associated with the same cluster of cognitive measures. Though this report does not conclusively show that AR and Amp can be used to clinically assess the risk of a female with the fM for cognitive disability, the evidence presented does suggest that these molecular variables, especially AR, reflect important underlying genetic factors contributing to the fra(X) phenotype.

Amyloid precursor protein gene expression in neural cell lines: influence of DNA cytosine methylation

Transcription of the gene encoding amyloid precursor protein (APP) varies in a cell-specific and developmentally regulated manner. The 5' region of this gene possesses a high frequency of CpG dinucleotides as well as copies of a GC-rich sequence, a potential trans factor binding element. These findings raise the possibility that DNA cytosine methylation could participate in the regulation of APP gene expression. We examined APP mRNA/18S rRNA ratio in three neural cell lines (N18TG2, SN6, SN17) cultured in 5-azacytidine (5-AZA), an inhibitor of maintenance methylase which results in loss of cytosine methylation in proliferating cells. Culture in 5-AZA globally reduced methylation in genomic DNA as assessed by an increase in HpaII restriction sites, reduced cytosine methylation in the APP gene as assessed by Southern blotting of HpaII digests, and increased APP mRNA steady state abundance in all studied cell lines. Cell lines re-acquired APP gene methylation 48 h after removal of 5-AZA from media. These results indicate that in vitro alteration of DNA methylation can affect APP gene expression, and suggest that the APP gene in neuronal cell lines may be rapidly inactivated in vitro, perhaps to neutralize its potential toxicity.

MyoD1 promoter autoregulation is mediated by two proximal E-boxes

We show that in mouse myoblasts the MyoD1 promoter is highly stimulated by MyoD1 expression, suggesting that it is controlled by a positive feedback loop. Using deletion and mutation analyses, we identified the targets for MyoD1 promoter autoregulation as the two proximal E-boxes located close to the MyoD1 core promoter. Gel mobility shift competition assays with MyoD1 antibodies as competitor suggest that the MyoD1 protein is binding directly to these E-boxes. Autoregulation did not occur in fibroblasts cotransfected with the expression vector of MyoD1. It is assumed that autoregulation is controlled by the stoichiometry between the MyoD1 protein and negatively regulatory proteins like Id, which is known to be highly expressed in fibroblasts. When the MyoD1 promoter was methylated, autoregulation only occurred when the density of methylated sites was low. The density of DNA methylation, therefore, can determine the accessibility of the MyoD1 promoter to transcription factors and interfere with the auto- and crossregulatory loop. The MyoD1 promoter in vivo was found to be only partially methylated in all tissues tested except in skeletal muscle where it was demethylated. We propose that high level expression of the MyoD1 gene is a result of release from constraints such as negative regulatory factors and/or DNA methylation interfering with MyoD1 autoregulation.

HOVERGEN: a database of homologous vertebrate genes

Comparison of homologous genes is a major step for many studies related to genome structure, function or evolution. Similarity search programs easily find genes homologous to a given sequence. However, only very tedious manual procedures allow the retrieval of all sets of homologous genes sequenced for a given set of species. Moreover, this search often generates errors due to the complexity of data to be managed simultaneously: phylogenetic trees, alignments, taxonomy, sequences and related information. HOVERGEN helps to solve these problems by integrating all this information. HOVERGEN corresponds to GenBank sequences from all vertebrate species, with some data corrected, clarified, or completed, notably to address the problem of redundancy. Coding sequences have been classified in gene families. Protein multiple alignments and phylogenetic trees have been calculated for each family. Sequences and related information have been structured in an ACNUC database which permits complex selections. A graphical interface has been developed to visualize and edit trees. Genes are displayed in color, according to their taxonomy. Users have directly access to all information attached to sequences and to multiple alignments simply by clicking on genes. This graphical tool gives thus a rapid and simple access to all data necessary to interpret homology relationships between genes. HOVERGEN allows the user to easily select sets of homologous vertebrate genes, and thus is particularly useful for comparative sequence analysis, or molecular evolution studies.

DNA methylation prevents the amplification of TROP1, a tumor-associated cell surface antigen gene

We tested the hypothesis that different genes can have different abilities to be amplified after transfection under comparable selection conditions. DNA from human lymphoid or choriocarcinoma cell lines was transfected into L cells. Transfectants for CD5, CD8A, TROP1, and TROP2, genes expressed on lymphocytes or trophoblast and carcinomas, were selected by fluorescence-activated cell sorting. To select for amplification of the transfected gene we cloned twice by fluorescence-activated cell sorting the transfectants with the highest expression. We analyzed a total of 38 families (1768 clones) derived from the original transfectants. We then analyzed by Southern blotting the clones with the highest increase in surface expression and determined the copy number of each transfected gene. CD5, CD8A, and TROP2 were amplified with high frequency and progressively, whereas TROP1 essentially was not amplified at all. We examined the hypothesis that DNA methylation prevents the amplification of the TROP1 gene by treating JAR choriocarcinoma cells with 5-azacytidine to decrease DNA methylation. DNA extracted at different times after the treatment was used for transfection. When DNA that showed demethylation of the TROP1 gene was used, 16 Trop-1 transfectants were obtained and 6 of them were found to contain up to 40 copies of the TROP1 gene per haploid genome. Thus, we showed that transfectants obtained from a demethylated TROP1 gene were amplified efficiently and progressively. We propose that DNA methylation affects DNA amplification either by altering the recognition of methylated DNA sequences or by changing the conformation of the chromatin of methylated segments. We speculate that DNA methylation is a determinant of gene amplification in vivo, for example in tumor cells.

Methylation status and chromatin structure of an early response gene (ornithine decarboxylase) in resting and stimulated NIH-3T3 fibroblasts

The early response gene ornithine decarboxylase (odc) is indispensable for normal and malignant cell growth. Although DNA methylation is generally associated with chromatin condensation and gene inactivation, the odc gene is heavily methylated at CCGG-sequences in animal cell lines. In this work we analyzed the chromatin structure and the DNA methylation status at the CpG-rich promoter sequences at the odc locus in mouse 3T3 fibroblasts. We show that the proximal promoter region of the odc locus is not hypermethylated, while the distal promoter sequences appear to have a few methylated CCGG-sites and display methylation polymorphism. Furthermore, it was found that the 5' promoter region of odc is constitutively more sensitive to micrococcal nuclease than the coding and 3' regions of the odc gene. Stimulation of the cells with serum resulted in an appearance of a DNase I sensitive site at the promoter region. The chromatin structure of the mid-coding and 3' regions of the odc gene also underwent structural changes that were accompanied by the rapid accumulation of odc mRNA. Such changes were not detected in the chromatin structure of glyceraldehyde-3-phosphate dehydrogenase (gadph) gene, whose expression remains invariant upon serum stimulation. These data suggest that the chromatin structure may play an important role in the rapid transcriptional activation of odc and other immediate early genes during serum stimulation.

Evidence that random and imprinted Xist expression is controlled by preemptive methylation

The mouse Xist gene is expressed exclusively from the inactive X chromosome and may control the initiation of X inactivation. We show that in somatic tissues the 5' end of the silent Xist allele on the active X chromosome is fully methylated, while the expressed allele on the inactive X is completely unmethylated. In tissues that undergo imprinted paternal Xist expression and imprinted X inactivation, the paternal Xist allele is unmethylated, and the silent maternal allele is fully methylated. In the male germline, a developmentally regulated demethylation of Xist occurs at the onset of meiosis and is retained in mature spermatozoa. This may be the cause of imprinted expression of the paternal Xist allele. A role for methylation in the control of Xist expression is further supported by the finding that in differentiating embryonic stem cells during the initiation of X inactivation, differential methylation of Xist alleles precedes the onset of Xist expression.

Neuropeptide role of both peptide YY and neuropeptide Y in vertebrates suggested by abundant expression of their mRNAs in a cyclostome brain

The evolution of the neuropeptide Y (NPY) family of peptides has been unclear despite sequence information from many vertebrates. We describe here two NPY-related peptides deduced from cDNA clones of the river lamprey (Lampetra fluviatilis), a cyclostome providing one of the best models of a primitive vertebrate brain. One peptide corresponds to NPY as it has 83% identity to human NPY and its mRNA is expressed in the lateral brainstem, dorsal spinal cord and retina. The second lamprey peptide corresponds anatomically to peptide YY (PYY) as its mRNA is found in gut cells and in medial brainstem neurons. Its sequence is 60-70% identical to both PYY and NPY of mammals. These data suggest that the gene duplication leading to NPY and PYY had already occurred in the ancestral vertebrate 450 million years ago. The expression of the presumed PYY homolog in both gut and central nervous system indicates that PYY has served the dual role as a hormone and a neuropeptide from an early stage in vertebrate evolution. The similarities in the location of NPY- and PYY-expressing cells between lamprey and mammals suggest that the functions of these peptides may have been conserved.

Parental imprinting of autosomal mammalian genes

The molecular mechanisms introducing epigenetic modifications that lead to differential silencing of some autosomal alleles depending on their parental legacy are still largely unknown, but recent results from studies of endogenously imprinted genes and particular transgenes make DNA methylation a strong candidate. At the same time, these results have raised new questions about the details of the imprinting process.

Alu transcripts: cytoplasmic localisation and regulation by DNA methylation

Full length Alu transcripts in HeLa cells are detected by primer extension using reverse transcriptase and are also analyzed as cloned cDNA sequences. The 5' end of these transcripts corresponds to the transcriptional start site for RNA polymerase III indicating that these RNAs are transcribed from their internal polymerase III promoters. The Alu transcripts found in cytoplasmic poly A+ RNAs appear to be organized into RNPs as assayed by sucrose gradient sedimentation. Present at about one hundred to one thousand copies per cell, the Alu transcripts are rare as compared to 7SL RNA. In agreement with previous reports that methylation inhibits Pol III-directed transcription of Alu in vitro, treatment of HeLa cells with 5-azacytidine results in Alu DNA hypomethylation and an increase in the abundance of the Alu transcript. Sequence analysis shows that many different Alu repeats including members of all subfamilies are transcribed by Pol III in vivo. cDNA sequences of the Pol III-directed transcripts exactly match the A box of the Pol III promoter element whereas in other Alu transcripts this element is not faithfully conserved.

Regulation of folate-binding protein gene expression by DNA methylation in methotrexate-resistant KB cells

Folate-binding protein (FBP) is responsible for the cellular transport of folate and methotrexate (MTX) in human KB (nasopharyngeal epidermoid carcinoma) cells. The levels of membrane-associated FBP and FBP mRNA are decreased 70-80% in an MTX-resistant KB subline (KB1BT) (Hsueh C-T and Dolnick BJ, Oncol Res 4: 497-505, 1992). Southern blot analysis did not reveal any differences in FBP gene organization or copy number between KB1BT and KB cells. However, there was a 70% decrease in the FBP gene transcription rate and no change in FBP mRNA stability in KB1BT cells. Assessing genomic DNA methylation by MspI and HpaII restriction analysis suggested that the FBP gene in KB1BT cells was more methylated than in KB cells. These alterations in the expression, transcription rate and DNA methylation state of the FBP gene did not change when KB1BT cells were grown in the absence of MTX for 8 months (MTX-free KB1BT). When MTX-free KB1BT cells were exposed to 2.5 microM 5-aza-2'-deoxycytidine for 72 hr, the FBP gene became hypomethylated and the levels of membrane-associated FBP and FBP mRNA increased by 2- to 3-fold. These data indicate that decreased FBP gene expression in KB1BT cells results from increased DNA methylation.

Sequences affecting the V(D)J recombinational activity of the IgH intronic enhancer in a transgenic substrate

The immunoglobulin heavy chain intronic transcriptional enhancer (E mu) is part of a complex cis-regulatory DNA region which has notably been shown to modulate V(D)J rearrangements of associated variable gene segments. We have used recombination substrates comprised of the E mu enhancer together with various lengths of additional downstream mu sequences to assess the individual contribution of those sequences to the V(D)J recombinational regulatory activity. Surprisingly, in the absence of large amounts of mu sequences, substrate rearrangements were not detected in Southern blot analyses of the lymphoid tissues from independent transgenic mice, but were readily detectable following transfection into cultured pre-B cells. A short mu segment which includes matrix association regions (MARs) was not sufficient to restore high levels of rearrangements within the reporter transgenes. However, additional experiments demonstrated that the mu sequences are dispensable for V(D)J recombination in transgenic thymuses, implying a suppressive effect exerted by the vector sequences left in the transgenic insert, when they are attached near the E mu regulatory region. This suppression of V(D)J recombination, which correlates with an hypermethylation of the transgenes, is discussed in view of previously reported transgenic and gene targeting experiments.

In vitro methylation of the human O6-methylguanine-DNA methyltransferase promoter reduces transcription

Approx. 20% of human tumor cell lines (termed Mer-) are deficient in the DNA repair enzyme O6-methylguanine-DNA methyltransferase (MGMT; E.C.2.1.1.63). Such cells possess the MGMT gene and promoter sequences but have virtually no mRNA or protein. Cytosine methylation of gene sequences has been proposed as a mechanism by which MGMT could be suppressed in Mer- cells; however, the experimental evidence does not uniformly support this idea. We therefore investigated the effect of in vitro methylation of the MGMT promoter in a reporter gene construct transfected into cultured human cells. DNA methylation by HpaII or HhaI methylases suppressed the activity of the promoter, although the effect was not absolute. The occurrence of partial intracellular demethylation of promoter sequences may account for the incomplete inhibition of transcription. A model that attempts to reconcile the opposing views on the role of cytosine methylation in MGMT gene expression is presented.

Protein-DNA interactions in the promoter region of the amyloid precursor protein (APP) gene in human neocortex

We have investigated protein-DNA interactions in the proximal promoter of the human amyloid precursor protein (APP) gene in temporal lobe neocortical nuclei isolated from control and Alzheimer disease (AD) affected brains. We report that the human APP 5' promoter sequence from -203 to +55 bp, which has been previously reported to contain essential regulatory elements for APP gene transcription, lies in a deoxyribonuclease I, micrococcal nuclease- and restriction endonuclease-sensitive, G+C-rich nucleosome-free gap flanked both 5' and 3' by typical nucleosome structures. As analyzed by electrophoretic mobility shift assay, this extended internucleosomal linker DNA is heavily occupied by nuclear protein factors, and interacts differentially with nuclear protein extracts obtained from HeLa and human brain neocortical nuclei. This suggests that the chromatin conformation of the APP gene promoter may vary in different cell types, and may correlate with differences in APP gene expression. Human recombinant transcription factors AP1, SP1 and TFIID (but not AP2 or brain histones H1, H2B and H4) interact with the -203 to +55 bp of the human APP promoter sequence. Only minor differences were observed in the chromatin structure of the immediate APP promoter between non-AD and AD affected neocortical nuclei, suggesting either that post-transcriptional processes, or that regulatory elements lying elsewhere in the APP gene may be important in the aberrant accumulation of the APP gene product.

Phenotypic stability and variation in cells of the porcine aorta: collagen and elastin production

The extracellular matrix of the developing vasculature varies in composition as a function of time and position. Cellular models of vascular biology and pathology depend on the assumption that stable phenotypic characteristics of vascular cells can be propagated through several generations of in vitro cultivation. We show that the positional and developmental heterogeneity of matrix phenotypes in the porcine aorta are expressed by explanted vascular smooth muscle cell (SMC) and adventitial cell populations for a limited number of passages. Elastin was expressed most highly by thoracic SMC while interstitial collagen production was usually maximal in abdominal segments. Parallel gradients of collagen types I, III and V, detected by specific ELISA assays, were expressed in early-passage SMC. Adventitial cell populations from the abdominal aorta of the neonatal pig accumulated significant levels of collagen, while these fibroblasts produced less than 10% of the elastin made by SMC. All cell populations expressed alpha-smooth muscle actin in vitro. Gradients of collagen and elastin expression were evident for no more than three passages, and direct outgrowth of cells without limited digestion of the matrix further reduced phenotypic stability. Variation and decline of the elastin phenotype could be due to hypermethylation of regulatory sequences in the elastin gene or trans-acting factors, but elastin production was dose-dependently stimulated to a similar extent (100%; 10 microM 5-azacytidine) in all segmental SMC populations at early (p1) and late (p3) passage. These data indicated that faithful reflection of in vivo SMC behavior was limited to a few population doublings, at least under standard culture conditions. Modification of the cellular environment by reducing serum factors, changing matrix, or adding mechanical stimulation may increase phenotypic stability.

High-resolution methylation analysis of the human hypoxanthine phosphoribosyltransferase gene 5' region on the active and inactive X chromosomes: correlation with binding sites for transcription factors

DNA methylation within GC-rich promoters of constitutively expressed X-linked genes is correlated with transcriptional silencing on the inactive X chromosome in female mammals. For most X-linked genes, X chromosome inactivation results in transcriptionally active and inactive alleles occupying each female nucleus. To examine mechanisms responsible for maintaining this unique system of differential gene expression, we have analyzed the methylation of individual cytosine residues in the 5' CpG island of the human hypoxanthine phosphoribosyltransferase (HPRT) gene on the active and inactive X chromosomes. Methylation analysis of 142 CpG dinucleotides by genomic sequencing was carried out on purified DNA using the cytosine-specific Maxam and Gilbert DNA sequencing reaction in conjunction with ligation-mediated PCR. These studies demonstrate the 5' CpG islands of active and 5-azacytidine-reactivated alleles are essentially unmethylated while the inactive allele is hypermethylated. The inactive allele is completely methylated at nearly all CpG dinucleotides except in a 68-bp region containing four adjacent GC boxes where most CpG dinucleotides are either unmethylated or partially methylated. Curiously, these GC boxes exhibit in vivo footprints only on the active X chromosome, not on the inactive X. The methylation pattern of the inactive HPRT gene is strikingly different from that reported for the inactive X-linked human phosphoglycerate kinase gene which exhibits methylation at all CpG sites in the 5' CpG island. These results suggest that the position of methylated CpG dinucleotides, the density of methylated CpGs, the length of methylated regions, and/or chromatin structure associated with methylated DNA may have a role in repressing the activity of housekeeping promoters on the inactive X chromosome. The pattern of DNA methylation on the inactive human HPRT gene may also provide insight into the process of inactivating the gene early in female embryogenesis.

Liver-specific DNase I-hypersensitive sites and DNA methylation pattern in the promoter region of a 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase gene

The mRNA for the liver isozyme of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase is transcribed from the L promoter of gene A. We show here that L-promoter activity is tissue specific. To identify on the gene in situ potential cis-acting sequences, we have examined 15 kb of its 5' region for DNase I-hypersensitive sites detectable on chromatin. We have also evaluated the DNA methylation status of the 3.7-kb encompassing the L promoter. Five DNase I-hypersensitive sites were detected on liver chromatin, three upstream (M1 at position -4500, L2 at position -1000, L1 at position -200) and two downstream (I1 at position +3000, I2 at position +3500) from the L-type mRNA transcription initiation site. Their presence correlated with transcriptional activity as they were not observed on chromatin from kidney, a tissue where gene A is not expressed. Sites M1 and L1 corresponded to the M and L promoters, respectively, providing in vivo evidence for a promoter localization obtained earlier with cloned DNA only. Site I2 coincided with a glucocorticoid-responsive unit described by others, but its presence did not depend on glucocorticoids. Thus, sites L2 and I1 could correspond to novel control elements. While DNA was methylated around position -2000 both in liver and kidney, downstream from that position it was fully demethylated in liver but not in kidney. This pattern changed during development of fetal liver. The data suggest mechanisms for the lack of activity of the L promoter in kidney and for its activation in developing and adult liver.

High-resolution methylation analysis of the human hypoxanthine phosphoribosyltransferase gene 5' region on the active and inactive X chromosomes: correlation with binding sites for transcription factors

DNA methylation within GC-rich promoters of constitutively expressed X-linked genes is correlated with transcriptional silencing on the inactive X chromosome in female mammals. For most X-linked genes, X chromosome inactivation results in transcriptionally active and inactive alleles occupying each female nucleus. To examine mechanisms responsible for maintaining this unique system of differential gene expression, we have analyzed the methylation of individual cytosine residues in the 5' CpG island of the human hypoxanthine phosphoribosyltransferase (HPRT) gene on the active and inactive X chromosomes. Methylation analysis of 142 CpG dinucleotides by genomic sequencing was carried out on purified DNA using the cytosine-specific Maxam and Gilbert DNA sequencing reaction in conjunction with ligation-mediated PCR. These studies demonstrate the 5' CpG islands of active and 5-azacytidine-reactivated alleles are essentially unmethylated while the inactive allele is hypermethylated. The inactive allele is completely methylated at nearly all CpG dinucleotides except in a 68-bp region containing four adjacent GC boxes where most CpG dinucleotides are either unmethylated or partially methylated. Curiously, these GC boxes exhibit in vivo footprints only on the active X chromosome, not on the inactive X. The methylation pattern of the inactive HPRT gene is strikingly different from that reported for the inactive X-linked human phosphoglycerate kinase gene which exhibits methylation at all CpG sites in the 5' CpG island. These results suggest that the position of methylated CpG dinucleotides, the density of methylated CpGs, the length of methylated regions, and/or chromatin structure associated with methylated DNA may have a role in repressing the activity of housekeeping promoters on the inactive X chromosome. The pattern of DNA methylation on the inactive human HPRT gene may also provide insight into the process of inactivating the gene early in female embryogenesis.

Inactive chromatin spreads from a focus of methylation

The detailed mechanisms of inhibition of transcription by DNA methylation are still unknown, but it has become obvious that the formation of chromatin plays an important role in this process. Using an approach enabling us to methylate, in vitro, chosen regions in a plasmid, we now show that specific methylation of nonpromoter sequences results in transcriptional inhibition of a reporter gene construct and that this inhibition is independent of the position of the methylated region within the plasmid. In plasmid minichromosomes containing a short region of methylated DNA, both methylated and unmethylated sequences are protected from limited MspI digestion. Our results show that inactive chromatin is present at unmethylated regions in partially methylated minichromosomes and can thereby inhibit gene expression. Spreading of the inactive chromatin is not inhibited by the presence of active promoters, nor is it a consequence of transcriptional inactivity.

Genetic imprinting in the mouse: implications for gene regulation

Genetic imprinting specifies a germline marking that subsequently results in the repression of one or other parental allele at some point in development. Genetic manipulations to generate maternal and paternal duplications of specific chromosome regions have been used to screen almost the entire mouse genome for evidence of imprinting. As a result, 15 imprinting effects involving 10 regions on 6 different chromosomes have been detected that range from early embryonic lethalities to various growth and developmental defects seen only after birth. Genes with important roles in development therefore appear to be involved. Diverse studies have identified four imprinted genes, all of which show monoallelic expression in some, but not necessarily all, tissues. A correlation with methylation is indicated but the pattern of methylation is not consistent for each of the genes; methylation is therefore unlikely to be the imprinting signal. Methods being used to identify further imprinted genes are summarized and some of the difficulties posed are indicated.