DNA methylation and chromatin structure: a view from below

DNA cytosine methyltransferases differentially regulate genome-wide hypermutation and interhomolog recombination in Trichoderma reesei meiosis

Trichoderma reesei is an economically important enzyme producer with several unique meiotic features. spo11, the initiator of meiotic double-strand breaks (DSBs) in most sexual eukaryotes, is dispensable for T. reesei meiosis. T. reesei lacks the meiosis-specific recombinase Dmc1. Rad51 and Sae2, the activator of the Mre11 endonuclease complex, promote DSB repair and chromosome synapsis in wild-type and spo11Δ meiosis. DNA methyltransferases (DNMTs) perform multiple tasks in meiosis. Three DNMT genes (rid1, dim2 and dimX) differentially regulate genome-wide cytosine methylation and C:G-to-T:A hypermutations in different chromosomal regions. We have identified two types of DSBs: type I DSBs require spo11 or rid1 for initiation, whereas type II DSBs do not rely on spo11 and rid1 for initiation. rid1 (but not dim2) is essential for Rad51-mediated DSB repair and normal meiosis. rid1 and rad51 exhibit a locus heterogeneity (LH) relationship, in which LH-associated proteins often regulate interconnectivity in protein interaction networks. This LH relationship can be suppressed by deleting dim2 in a haploid rid1Δ (but not rad51Δ) parental strain, indicating that dim2 and rid1 share a redundant function that acts earlier than rad51 during early meiosis. In conclusion, our studies provide the first evidence of the involvement of DNMTs during meiotic initiation and recombination.

Neurospora crassa, a model system for epigenetics research

The filamentous fungus Neurospora crassa has provided a rich source of knowledge on epigenetic phenomena that would have been difficult or impossible to gain from other systems. Neurospora sports features found in higher eukaryotes but absent in both budding and fission yeast, including DNA methylation and H3K27 methylation, and also has distinct RNA interference (RNAi)-based silencing mechanisms operating in mitotic and meiotic cells. This has provided an unexpected wealth of information on gene silencing systems. One silencing mechanism, named repeat-induced point mutation (RIP), has both epigenetic and genetic aspects and provided the first example of a homology-based genome defense system. A second silencing mechanism, named quelling, is an RNAi-based mechanism that results in silencing of transgenes and their native homologs. A third, named meiotic silencing, is also RNAi-based but is distinct from quelling in its time of action, targets, and apparent purpose.

Widespread plasticity in CTCF occupancy linked to DNA methylation

CTCF is a ubiquitously expressed regulator of fundamental genomic processes including transcription, intra- and interchromosomal interactions, and chromatin structure. Because of its critical role in genome function, CTCF binding patterns have long been assumed to be largely invariant across different cellular environments. Here we analyze genome-wide occupancy patterns of CTCF by ChIP-seq in 19 diverse human cell types, including normal primary cells and immortal lines. We observed highly reproducible yet surprisingly plastic genomic binding landscapes, indicative of strong cell-selective regulation of CTCF occupancy. Comparison with massively parallel bisulfite sequencing data indicates that 41% of variable CTCF binding is linked to differential DNA methylation, concentrated at two critical positions within the CTCF recognition sequence. Unexpectedly, CTCF binding patterns were markedly different in normal versus immortal cells, with the latter showing widespread disruption of CTCF binding associated with increased methylation. Strikingly, this disruption is accompanied by up-regulation of CTCF expression, with the result that both normal and immortal cells maintain the same average number of CTCF occupancy sites genome-wide. These results reveal a tight linkage between DNA methylation and the global occupancy patterns of a major sequence-specific regulatory factor.

The fungus Neurospora crassa displays telomeric silencing mediated by multiple sirtuins and by methylation of histone H3 lysine 9

Background

Silencing of genes inserted near telomeres provides a model to investigate the function of heterochromatin. We initiated a study of telomeric silencing in Neurospora crassa, a fungus that sports DNA methylation, unlike most other organisms in which telomeric silencing has been characterized.

Results

The selectable marker, hph, was inserted at the subtelomere of Linkage Group VR in an nst-1 (neurospora sir two-1) mutant and was silenced when nst-1 function was restored. We show that NST-1 is an H4-specific histone deacetylase. A second marker, bar, tested at two other subtelomeres, was similarly sensitive to nst-1 function. Mutation of three additional SIR2 homologues, nst-2, nst-3 and nst-5, partially relieved silencing. Two genes showed stronger effects: dim-5, which encodes a histone H3 K9 methyltransferase and hpo, which encodes heterochromatin protein-1. Subtelomeres showed variable, but generally low, levels of DNA methylation. Elimination of DNA methylation caused partial derepression of one telomeric marker. Characterization of histone modifications at subtelomeric regions revealed H3 trimethyl-K9, H3 trimethyl-K27, and H4 trimethyl-K20 enrichment. These modifications were slightly reduced when telomeric silencing was compromised. In contrast, acetylation of histones H3 and H4 increased.

Conclusion

We demonstrate the presence of telomeric silencing in Neurospora and show a dependence on histone deacetylases and methylation of histone H3 lysine 9. Our studies also reveal silencing functions for DIM-5 and HP1 that appear independent of their role in de novo DNA methylation.

A cytosine methyltransferase homologue is essential for sexual development in Aspergillus nidulans

BACKGROUND

The genome defense processes RIP (repeat-induced point mutation) in the filamentous fungus Neurospora crassa, and MIP (methylation induced premeiotically) in the fungus Ascobolus immersus depend on proteins with DNA methyltransferase (DMT) domains. Nevertheless, these proteins, RID and Masc1, respectively, have not been demonstrated to have DMT activity. We discovered a close homologue in Aspergillus nidulans, a fungus thought to have no methylation and no genome defense system comparable to RIP or MIP.

PRINCIPAL FINDINGS

We report the cloning and characterization of the DNA methyltransferase homologue A (dmtA) gene from Aspergillus nidulans. We found that the dmtA locus encodes both a sense (dmtA) and an anti-sense transcript (tmdA). Both transcripts are expressed in vegetative, conidial and sexual tissues. We determined that dmtA, but not tmdA, is required for early sexual development and formation of viable ascospores. We also tested if DNA methylation accumulated in any of the dmtA/tmdA mutants we constructed, and found that in both asexual and sexual tissues, these mutants, just like wild-type strains, appear devoid of DNA methylation.

CONCLUSIONS/SIGNIFICANCE

Our results demonstrate that a DMT homologue closely related to proteins implicated in RIP and MIP has an essential developmental function in a fungus that appears to lack both DNA methylation and RIP or MIP. It remains formally possible that DmtA is a bona fide DMT, responsible for trace, undetected DNA methylation that is restricted to a few cells or transient but our work supports the idea that the DMT domain present in the RID/Masc1/DmtA family has a previously undescribed function.

Melatonin seems to be a mediator that transfers the environmental stimuli to oocytes for inheritance of adaptive changes through epigenetic inheritance system

Possibility of inheritance of epigenetic modifications have led us to consider that adaptive geographic variations in humans may result from interactions between environmental factors and epigenetic inheritance system. In this system melatonin seems to be a mediator that transfers the environmental stimuli to germ cells (oocytes). While environmental factors produce modifications in the body, they simultaneously induce epigenetic modifications in the oocytes with the help of melatonin, and these changes are inherited to offspring. In this way, adaptive changes could be passed on to the next generation. This kind of heritable long-term changes is generally labeled biological adaptation. But, how can melatonin cause epigenetic changes in oocytes? We suggest that melatonin induces epigenetic modifications by affecting the nuclear melatonin receptors that can in turn change the superstructure of DNA. It was previously suggested that biological adaptation is limited to neural crest derivatives such as, craniofacial tissues, melanocytes, and structures related to stature, hair form and body proportions. Thus, inheritance of adaptive changes is possible only where environmental factors affect the neural crest derivatives, including the cells that produce the next generation.

Use of cDNA microarrays to monitor transcriptional responses of the chestnut blight fungus Cryphonectria parasitica to infection by virulence-attenuating hypoviruses

Hypoviruses are a family of cytoplasmically replicating RNA viruses of the chestnut blight fungus Cryphonectria parasitica. Members of this mycovirus family persistently alter virulence (hypovirulence) and related fungal developmental processes, including asexual and sexual sporulation. In order to gain a better understanding of the molecular basis for these changes, we have developed a C. parasitica cDNA microarray to monitor global transcriptional responses to hypovirus infection. In this report, a spotted DNA microarray representing approximately 2,200 C. parasitica genes was used to monitor changes in the transcriptional profile after infection by the prototypic hypovirus CHV1-EP713. Altered transcript abundance was identified for 295 clones (13.4% of the 2,200 unique cDNAs) as a result of CHV1-EP713 infection-132 up-regulated and 163 down-regulated. In comparison, less than 20 specific C. parasitica genes were previously identified by Northern analysis and mRNA differential display as being responsive to hypovirus infection. A 93% validation rate was achieved between real-time reverse transcription-PCR results and microarray predictions. Differentially expressed genes represented a broad spectrum of biological functions, including stress responses, carbon metabolism, and transcriptional regulation. These findings are consistent with the view that infection by a 12.7-kbp hypovirus RNA results in a persistent reprogramming of a significant portion of the C. parasitica transcriptome. The potential impact of microarray studies on current and future efforts to establish links between hypovirus-mediated changes in cellular gene expression and phenotypes is discussed.

Lessons from the genome sequence of Neurospora crassa: tracing the path from genomic blueprint to multicellular organism

We present an analysis of over 1,100 of the approximately 10,000 predicted proteins encoded by the genome sequence of the filamentous fungus Neurospora crassa. Seven major areas of Neurospora genomics and biology are covered. First, the basic features of the genome, including the automated assembly, gene calls, and global gene analyses are summarized. The second section covers components of the centromere and kinetochore complexes, chromatin assembly and modification, and transcription and translation initiation factors. The third area discusses genome defense mechanisms, including repeat induced point mutation, quelling and meiotic silencing, and DNA repair and recombination. In the fourth section, topics relevant to metabolism and transport include extracellular digestion; membrane transporters; aspects of carbon, sulfur, nitrogen, and lipid metabolism; the mitochondrion and energy metabolism; the proteasome; and protein glycosylation, secretion, and endocytosis. Environmental sensing is the focus of the fifth section with a treatment of two-component systems; GTP-binding proteins; mitogen-activated protein, p21-activated, and germinal center kinases; calcium signaling; protein phosphatases; photobiology; circadian rhythms; and heat shock and stress responses. The sixth area of analysis is growth and development; it encompasses cell wall synthesis, proteins important for hyphal polarity, cytoskeletal components, the cyclin/cyclin-dependent kinase machinery, macroconidiation, meiosis, and the sexual cycle. The seventh section covers topics relevant to animal and plant pathogenesis and human disease. The results demonstrate that a large proportion of Neurospora genes do not have homologues in the yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe. The group of unshared genes includes potential new targets for antifungals as well as loci implicated in human and plant physiology and disease.

Specific and common alterations in host gene transcript accumulation following infection of the chestnut blight fungus by mild and severe hypoviruses

We report the use of a cDNA microarray to monitor global transcriptional responses of the chestnut blight fungus, Cryphonectria parasitica, to infection by mild and severe isolates of virulence-attenuating hypoviruses that share 87 to 93% and 90 to 98% identity at the nucleotide and amino acid levels, respectively. Infection by the mild hypovirus isolate CHV1-Euro7 resulted in differential expression of 166 of the ca. 2,200 genes represented on the microarray (90 upregulated and 76 downregulated). This is roughly half the number of genes scored as differentially expressed after infection by the severe isolate, CHV1-EP713 (295 genes; 132 upregulated and 163 downregulated). Comparison of the lists of genes responsive to infection by the two hypovirus isolates revealed 80 virus-common responsive genes. Infection by CHV1-EP713 also caused changes in gene transcript accumulation that were, in general, of greater magnitude than those observed with CHV1-Euro7 infections. Thus, the host transcriptional response to infection by severe hypovirus CHV1-EP713 appears to be considerably more dynamic than the response to infection by the mild isolate CHV1-Euro7. Real-time reverse transcription-PCR was performed on 39 different clones, with false-positive rates of 3 and 8% observed for the microarray-predicted list of genes responsive to CHV1-EP713 and CHV1-Euro7 infections, respectively. This analysis has allowed an initial assignment for ca. 2,200 unique C. parasitica-expressed genes as being unresponsive to hypovirus infection, selectively responsive to a specific hypovirus, or generally responsive to hypovirus infection.

Heterochromatin?many flavours, common themes

Heterochromatin remains condensed throughout the cell cycle, is generally transcriptionally inert and is built and maintained by groups of factors with each group member sharing a similar function. In mammals, these groups include sequence-specific transcriptional repressors, functional RNA and proteins involved in DNA and histone methylation. Heterochromatin is cemented together via interactions within and between each protein group and is maintained by the cell's replication machinery. It can be constitutive (permanent) or facultative (developmentally regulated) and be any size, from a gene promotor to a whole genome. By studying the formation of facultative heterochromatin, we have gained information about how heterochromatin is assembled. We have discovered that there are many different architectural plans for the building of heterochromatin, leading to a seemingly never-ending variety of heterochromatic loci, with each built according to a general rule.

Induction and maintenance of nonsymmetrical DNA methylation in Neurospora

One can imagine a variety of mechanisms that should result in self-perpetuating biological states. It is generally assumed that cytosine methylation is propagated in eukaryotes by enzymes that specifically methylate hemimethylated symmetrical sites (e.g., (5')CpGGpC(5') or (5')CpNpGGpNpC(5')). Although there is wide support for this model, we and others have found examples of methylation that must be propagated by a different mechanism. Most methylated regions of the Neurospora genome that have been examined are products of repeat-induced point mutation, a premeiotic genome defense system that litters duplicated sequences with C.G to T.A mutations and typically leaves them methylated at remaining cytosines. In general, such relics of repeat-induced point mutation are capable of triggering methylation de novo. Nevertheless, some reflect a mechanism that can propagate heterogeneous methylation at nonsymmetrical sites. We propose that de novo and maintenance methylation are manifestations of a single mechanism in Neurospora, catalyzed by the DIM-2 DNA methyltransferase. The action of DIM-2 is controlled by the DIM-5 histone H3 Lys-9 methyltransferase, which in turn is influenced by other modifications of histone H3. DNA methylation indirectly recruits histone deacetylases, providing the framework of a self-reinforcing system that could result in propagation of DNA methylation and the associated silenced chromatin state.

DNA methylation and breast carcinogenesis

Knowledge about breast carcinogenesis has accumulated during the last decades but has barely been translated into strategies for early detection or prevention of this common disease. Changes in DNA methylation have been recognized as one of the most common molecular alterations in human neoplasia and hypermethylation of gene-promoter regions is being revealed as one of the most frequent mechanisms of loss of gene function. The heritability of methylation states and the secondary nature of the decision to attract or exclude methylation support the idea that DNA methylation is adapted for a specific cellular memory. According to Hanahan and Weinberg, there are six novel capabilities a cell has to acquire to become a cancer cell: limitless replicative potential, self-sufficiency in growth signals, insensitivity to growth-inhibitory signals, evasion of programmed cell death, sustained angiogenesis and tissue invasion and metastasis. This review highlights how DNA-methylation contributes to these features and offers suggestions about how these changes could be prevented, reverted or used as a 'tag' for early detection of breast cancer or, preferably, for detection of premalignant changes.

Homology-dependent gene silencing mechanisms in fungi

Homology-dependent gene silencing (HDGS) is a ubiquitous phenomenon among fungi, plants, and animals. Gene silencing can be triggered and can affect artificially introduced nucleic acid molecules, both DNA and RNA, and/or can act on endogenous duplicated sequences. Although the various HDGS phenomena may be related each other, probably deriving from an ancestral defense mechanism, relevant differences do exist between different HDGS mechanisms. Especially in fungi, a variety of HDGS phenomena have been uncovered during the past 10 years: Gene inactivation of duplicated sequences can be achieved either through DNA-methylation and block of transcription or through sequence-specific degradation of mRNA. Moreover, duplicated sequences can also be specifically mutagenized. Studying HDGS in fungi gives us the opportunity to study such complex mechanisms in relatively simple organisms in which both genetic and biochemical approaches can be easily used.

Protein binding protects sites on stable episomes and in the chromosome from de novo methylation

We have utilized the Escherichia coli lac repressor-operator system to test whether protein binding can interfere with de novo DNA methylation in mammalian cells. We find that a DNA binding protein can protect sites on the episome as well as in the genome from the de novo methylation activity of Dnmt3a. Transcriptional machinery moving through the binding sites does not affect the de novo methylation of these sites, and it does not affect the binding protein protection of these sites from de novo methylation. This study and previous studies provide a possible mechanism for the observation that an Sp1 site can serve as a cis-acting signal for demethylation and for preventing de novo methylation of the CpG island upstream of the mouse adenine phosphoribosyltransferase (Aprt) gene. These findings also support the hypothesis that protein binding may play a crucial role in changes of CpG methylation pattern in mammalian cells.

Short TpA-rich segments of the zeta-eta region induce DNA methylation in Neurospora crassa

The mechanisms that establish DNA methylation in eukaryotes are poorly understood. In principle, methylation in a particular chromosomal region may reflect the presence of a "signal" that recruits methylation, the absence of a signal that prevents methylation, or both. Experiments were carried out to address these possibilities for the 1.6 kb zeta-eta (zeta-eta) region, a relict of repeat-induced point mutation (RIP) in the fungus Neurospora crassa. The zeta-eta region directs its own de novo methylation at a variety of chromosomal locations. We tested the methylation potential of a nested set of fragments with deletions from one end of the zeta-eta region, various internal fragments of this region, chimeras of eta and the homologous unmutated allele, theta (theta), and various synthetic variants, integrated precisely in single copy at the am locus on linkage group (LG) VR or the his-3 locus on LG IR. We found that: (1) the zeta-eta region contains at least two non-overlapping methylation signals; (2) different fragments of the region can induce different levels of methylation; (3) methylation induced by zeta-eta sequences can spread far into flanking sequences; (4) fragments as small as 171 bp can trigger methylation; (5) methylation signals behave similarly, but not identically, at different chromosomal sites; (6) mutation density, per se, does not determine whether sequences become methylated; and (7) neither A:T-richness nor high densities of TpA dinucleotides, typical attributes of methylated sequences in Neurospora, are essential features of methylation signals, but both promote de novo methylation. We conclude that de novo methylation of zeta-eta sequences does not simply reflect the absence of signals that prevent methylation; rather, the region contains multiple, positive signals that trigger methylation. These findings conflict with earlier models for the control of DNA methylation, including the simplest version of the collapsed chromatin model.

Chromatin remodelling during the life cycle of trypanosomatids

The mechanisms which control the expression of developmentally regulated genes in trypanosomatids remain unclear. The genes are grouped together into transcription units that are co-transcribed to yield polycistronic RNAs. Trans-splicing and polyadenylation give rise to mature, monocistronic mRNAs. It is difficult to imagine that expression of these genes is controlled at the level of transcription initiation because this would suggest that the genes are transcribed at the same rate. This is not the case, because at any given developmental stage in trypanosomes or Leishmania, genes transcribed from the same transcription unit are expressed at different levels within the cell. Consequently, these parasites must rely on post-transcriptional or post-translational mechanisms to generate the appropriate levels of gene product within the cell. There are no well-established examples of RNA polymerase II promoters in trypanosomes or Leishmania. However, the promoters for genes encoding the variant surface glycoprotein (VSG) and the procyclic acidic repetitive protein (PARP) have been identified and resemble ribosomal RNA polymerase I promoters. In higher eukaryotes where the mechanisms regulating transcription are clearer, there is increasing evidence that epigenetic factors, such as histones and modified bases, influence gene expression. Chemical modification of these factors can restructure chromatin and lead to gene activation or silencing. In trypanosomatids, an epigenetic mechanism for the control of developmentally expressed genes is a possibility. In this review, chromatin remodelling during the life and cell cycle of trypanosomes and Leishmania is explored, and the influence of epigenetic factors such as histones and modified bases on this process is discussed.

Modulation of DNA binding protein affinity directly affects target site demethylation

It has recently been shown that in Xenopus, DNA demethylation at promoter regions may involve protein-DNA interactions, based on the specificity of the demethylated sites. Utilizing a stable episomal system in human cells, we recently mapped the sites and dissected the steps of demethylation at oriP sites bound by EBNA1 protein. Although it is clear that protein binding is required for demethylation of the oriP sites, it is uncertain whether this is a unique feature of the replication origin or whether it is a general phenomenon for all DNA sequences to which sequence-specific proteins are bound. In the present study, we utilize the well-defined Escherichia coli lac repressor/operator system in human cells to determine whether protein binding to methylated DNA, in a region that is neither a replication origin nor a promoter, can also lead to demethylation of the binding sites. We found that demethylation specified by protein binding is not unique to the replication origin or to the promoter. We also found that transcriptional activity does not influence demethylation of the lac operator. Isopropyl-beta-D-thiogalactopyranoside (IPTG), an inhibitor of the lac repressor, can prevent demethylation of the lac operator DNA sites and can modulate demethylation of the lac operator by affecting the binding affinity of the lac repressor. Using this system, a titration of protein binding can be done. This titration permits one to infer that protein binding site occupancy is the determinant of demethylation at DNA sites and permits a determination of how this process progresses over time.

GFAP gene methylation in different neural cell types from rat brain

It is generally believed that specific demethylation processes take place in the promoter of tissue-specific genes during development. It has been suggested that hypomethylation of the -1500/-1100 domain of the 5' flanking regulatory region of the rat glial fibrillary acidic protein gene may be specific for neuroectodermal derivatives such as neurons and astrocytes. In the present work the methylation status of one of those seven CG sites (the -1176) of the 'neuroectoderm-specific domain' was analyzed. In agreement with the neuroectoderm hypothesis, the -1176 site is highly demethylated in astroglial, oligodendroglial and neuronal cells, but heavily methylated in microglial and fibroblast cells. The three different glial population are derived from the same tissue (cerebral hemispheres of newborn rats) but have a different embryological origin: oligodendrocytes and astrocytes originate from neuroectoderm, while microglia is of mesodermal origin. It is not clear if GFAP-negative neuronal cells maintain such demethylation in the advanced stage of maturation or if they undergo a second phase of de novo methylation. In order to clarify this point we used a subcellular fractionation method which allowed us to separate two different nuclear populations from adult rat cerebral hemispheres: one enriched in neuronal nuclei (called N1) and the other enriched in glial nuclei (N2). A higher methylation level of the -1176 site was detected in the N1 fraction, suggesting the GFAP gene undergo a de novo methylation process during neuronal maturation. This observation is in agreement with recent results showing a de novo methylation of the -1176 site during postnatal brain development. We hypothesize that a DNA demethylation process takes place in neuroectodermal precursor cells and that the -1176 site persists demethylated at the earlier stages of neuronal differentiation (immature neurons) and becomes fully methylated at more advanced stages of differentiation.

An alternative promoter in the mouse major histocompatibility complex class II I-Abeta gene: implications for the origin of CpG islands

Nonmethylated CpG islands are generally located at the 5' ends of genes, but a CpG island in the mouse major histocompatibility complex class II I-Abeta gene is remote from the promoter and covers exon 2. We have found that this CpG island includes a novel intronic promoter that is active in embryonic and germ cells. The resulting transcript potentially encodes a severely truncated protein which would lack the signal peptide and external beta1 domains. The functional significance of the internal CpG island may be to facilitate gene conversion, thereby sustaining the high level of polymorphism seen at exon 2. Deletions of the I-Abeta CpG island promoter reduce transcription and frequently lead to methylation of the CpG island in a transgenic mouse assay. These and other results support the idea that all CpG islands arise at promoters that are active in early embryonic cells.

Transposons in filamentous fungi--facts and perspectives

Transposons are ubiquitous genetic elements discovered so far in all investigated prokaryotes and eukaryotes. In remarkable contrast to all other genes, transposable elements are able to move to new locations within their host genomes. Transposition of transposons into coding sequences and their initiation of chromosome rearrangements have tremendous impact on gene expression and genome evolution. While transposons have long been known in bacteria, plants, and animals, only in recent years has there been a significant increase in the number of transposable elements discovered in filamentous fungi. Like those of other eukaryotes, each fungal transposable element is either of class or of class II. While class I elements transpose by a RNA intermediate and employ reverse transcriptases, class II elements transpose directly at the DNA level. We present structural and functional features for such transposons that have been identified so far in filamentous fungi. Emphasis is given to specific advantages or unique features when fungal systems are used to study transposable elements, e.g., the evolutionary impact of transposons in coenocytic organisms and possible experimental approaches toward horizontal gene transfer. Finally, we focus on the potential of transposons for tagging and identifying fungal genes.

Epigenetic phenomena in filamentous fungi: useful paradigms or repeat-induced confusion?

Epigenetic mechanisms can serve as genome defense systems. In haploid nuclei of special sexual cells of fungi, such as Neurospora and Ascobolus, duplicated genes are silenced by hypermutation, DNA methylation, or both. In some cases, DNA introduced into the genome of Neurospora cells by transformation can also inhibit homologous genes by a silencing mechanism that does not involve DNA pairing or methylation and appears to be post-transcriptional. Transforming DNA can also trigger de novo methylation in vegetative cells, which then causes transcriptional silencing. The rules governing silencing in vegetative cells of fungi are undefined, but repeated sequences seem particularly susceptible to these processes. Thus, fungi exhibit both repeat-induced and repeat-associated silencing mechanisms. Additionally, some native genes depend on homologous pairing in the diplophase for proper regulation. Together, these processes should limit the proliferation of transposable elements and serve to preserve the overall structure of the genome.

Structure and function analysis of the human myeloid cell nuclear differentiation antigen promoter: evidence for the role of Sp1 and not of c-Myb or PU.1 in myelomonocytic lineage-specific expression

The human myeloid nuclear differentiation antigen (MNDA) is expressed specifically in maturing cells of the myelomonocytic lineage and in monocytes and granulocytes. Epitope enhancement was used to confirm the strict lineage- and stage-specific expression of MNDA in bone marrow as well as in other paraffin-embedded fixed tissues. A 1-kb region of the gene that includes 5' flanking sequence was reported earlier to contain functional promoter activity and was specifically demethylated in expressing cells in contrast to null cells. Further analysis has revealed that this 1-kb fragment promotes higher reporter gene activity in MNDA-expressing cells than non-expressing cells, indicating cell-specific differences in transactivation. This sequence contains consensus elements consistent with myeloid-specific gene expression, including a PU.1 consensus site near the major transcription start site and a cluster of c-Myb sites located several hundred bases upstream of this region. However, analysis of deletion mutants localized nearly all of the promoter activity to a short region (-73 to -16) that did not include the cluster of c-Myb sites. A 4-bp mutation of the core Sp1 consensus element (GC box) (-20) reduced overall promoter activity of the 1-kb fragment. Mutation of the PU.1 site did not significantly affect promoter activity. Only a small region (-35 to +22) including the Sp1 element and transcription start site, but not the PU.1 site was footprinted. The 4-bp mutation of the core Sp1 consensus element abolished footprinting at the site and an antibody super-shift reaction showed that Sp1 is one of the factors binding the consensus site. The Sp1 site also co-localizes with a DNase I hypersensitive site. The results indicate that DNA methylation, chromatin structure, and transactivation at an Sp1 site contribute to the highly restricted expression of this myelomonocytic lineage specific gene.

DNA methylation, heterochromatin and epigenetic carcinogens

This paper will explore emerging concepts related to alternative carcinogenic mechanisms of 'non-mutagenic,' and hence epigenetic, carcinogens that may heritably alter DNA methylation without changing the underlying DNA sequence. In this review, we will touch on the basic concepts of DNA methylation, and will elaborate in greater detail on related topics including chromatin condensation, and heterochromatin spreading that is well known to induce gene silencing by position effect variegation in Drosophila and other species. Data from our model transgenic G12 cell system will be presented to support our hypothesis that certain carcinogens, such as nickel, may be carcinogenic not primarily because of their overt mutability, but rather as the result of their ability to promote DNA hypermethylation of important cancer-related genes. We will conclude with a discussion of the broader relevance of our findings and its application to other so-called 'epigenetic' carcinogens.

Formation of methylation patterns in the mammalian genome

Cytosine methylation in mammals is an epigenetic modification required for viability of the developing embryo. It has been suggested that DNA methylation plays important roles in X-chromosome inactivation, imprinting, protection of the genome from invasive DNA sequences, and compartmentalization of the genome into active and condensed regions. Despite the significance of DNA methylation in mammalian cells, the mechanisms used to establish methylation patterns during development are not understood. This review will summarize the current state of knowledge about potential roles for cis- and trans-acting factors in the formation of methylation patterns in the mammalian genome.

restless, an active Ac-like transposon from the fungus Tolypocladium inflatum: structure, expression, and alternative RNA splicing

Elements of the hAT transposon family, such as the maize activator (Ac), have been discovered in a large number of eukaryotic species. This type of class II transposon, present in both plants and animals, has not been previously detected in any fungal organism. However, using a differential screening method to detect repetitive DNA, we have identified a hAT transposon in the hyphomycete Tolypocladium inflatum. The transposon, which we named restless, is 4,097 bp long, carries 20-bp inverted repeats and an 8-bp target site duplication, and encodes a long open reading frame which is interrupted by a single intronic sequence. The derived mRNA exhibits alternative splicing, resulting in the formation of two transcripts that may be translated into polypeptides of 157 or 803 amino acids. The predicted amino acid sequence of the larger polypeptide demonstrates significant homology with transposases from the hAT transposon family. A chromosomal analysis using pulsed-field gel electrophoresis showed that all seven chromosomal bands carry copies of the 4.1-kb transposon. This was confirmed in hybridization experiments with rare-cutting restriction endonucleases which indicate that about 15 copies are present in T. inflatum. The genomic distribution of restless and its transcriptional expression, alternative mRNA splicing, and genomic mobility all imply a potential role for this element in developing a transposon tagging system for use in filamentous fungi.

Native DNA repeats and methylation in Ascobolus

We identified two classes of native dispersed DNA repeats in the Ascobolus genome. The first class consisted of several kilobase long, methylated repeats. These repeats, named Mars (methylated Ascobolus repeated sequences), fell in one family of LINE-like elements and in three families of LTR-containing retrotransposable elements. The methylation features of Mars elements were those expected if they were natural targets for the MIP (methylation induced premeiotically) previously discovered in Ascobolus. The second class consisted of short repeats, approximately 100 bp long, corresponding to 5S rRNA and tRNA genes. As expected from their size, which was too small to allow MIP to occur, they were unmethylated, as were 26 kb of unique sequences tested. These observations are consistent with the hypothesis that MIP is targeted at natural DNA repeats and constitutes a defensive process against the detrimental consequences of the spreading of mobile elements throughout the genome. The 9 kb tandem repeats harbouring the 28S, 18S and 5.8S rRNA genes displayed methylation features suggesting that rDNA methylation proceeds through a process other than MIP.

Methylation of CpG island transcription factor binding sites is unnecessary for aberrant silencing of the human MGMT gene

Aberrant transcriptional inactivation of the non-X-linked human O-6-methylguanine DNA methyltransferase (MGMT) gene has been associated with loss of open chromatin structure and increases in cytosine methylation in the Sp1-binding region of the 5'-CpG island of the gene. To examine the necessity of these events for gene silencing, we have isolated and characterized a subline of human MGMT+ T98G glioma cells. The subline, T98Gs, does not express MGMT activity or MGMT mRNA, and exhibits no in vivo DNA-protein interactions at Sp1-like binding sites in the MGMT 5'-CpG island. While the MGMT CpG island is less accessible to exogenously added restriction enzymes in T98Gs nuclei than in T98G nuclei, it is similarly methylated in both T98G and T98Gs cell lines 5' and 3' to the transcription factor binding sites, and similarly unmethylated in the region encompassing the binding sites. Inappropriate transcriptional inactivation of MGMT, therefore, does not require methylation of transcription factor binding sites within the 5'-CpG island. Rather, MGMT gene silencing and transcription factor exclusion from T98Gs MGMT CpG island binding sites is most closely associated with condensed chromatin structure, which is in turn indirectly influenced by distant sites of methylation.

Methylation of DNA repeats of decreasing sizes in Ascobolus immersus

In Ascobolus immersus, DNA duplications are subject to the process of methylation induced premeiotically (MIP), which methylates the cytosine residues within the repeats and results in reversible gene silencing. The triggering of MIP requires pairing of the repeats, and its detection requires maintenance of the resulting methylation. MIP of kilobase-size duplications occurs frequently and leads to the methylation of all C residues in the repeats, including those belonging to non-CpG sequences. Using duplications of decreasing sizes, we observed that tandem repeats never escaped MIP when larger than 630 bp and showed a sudden and drastic drop in MIP frequencies when their sizes decreased from 630 to 317 bp. This contrasted with the progressive decrease of MIP frequencies observed with ectopic repeats, in which apparently the search for homology influences the MIP triggering efficiency. The minimal size actually required for a repeat to undergo detectable MIP was found to be close to 300 bp. Genomic sequencing and Southern hybridization analyses using restriction enzymes sensitive to C methylation showed a loss of methylation at non-CpG sites in short DNA segments, methylation being restricted to a limited number of CpG dinucleotides. Our data suggest the existence of two distinct mechanisms underlying methylation maintenance, one responsible for methylation at CpG sites and the other responsible for methylation at non-CpG sites.

Is genomic imprinting involved in the pathogenesis of hyperdiploid and haploid acute lymphoblastic leukemia of childhood?

Hyperdiploidy with a chromosome number between 51 and 65 and a mean peak at 55 occurs as a distinct karyotype pattern in approximately 25-30% of ALLs in childhood [1, 2]. It is considered a favorable prognostic factor. The most intriguing cytogenetic peculiarities of these leukemias are the nearly exclusive presence of nonrandom numerical abnormalities due to the gain of chromosomes 4, 6, 10, 14, 17, 18, 20, 21 and X [1, 2]. In contrast, chromosomes 1, 2, 3, 12 and 16 are rarely involved [1, 2]. Typically, the affected chromosomes are present in three copies, with chromosome 21 also often being tetrasomic.

Near-haploid cases, on the other hand, are extremely rare and have a bad prognosis [1, 2]. They contain at least one copy of each chromosome, a second copy of one of the sex chromosomes and both chromosomes 21 in most instances. In addition, two copies of chromosomes 10, 14 and 18 are commonly found.

In the majority of cases of hyperdiploid ALL, the mechanism leading to the increased number of chromosomes is unknown. However, once formed, the abnormal karyotype is uniform and stable in the malignant cell population. Molecular genetic studies performed by Onodera et al. [3] revealed that the hyperdiploid karyotype usually arises by a simultaneous event during a single abnormal cell division from a diploid karyotype. Occasionally, this can also occur by doubling of the chromosomes from a near-haploid karyotype [4]. In virtually all cases, tetrasomy of chromosome 21 was generated by a duplication of both the maternally and paternally derived homolog. This finding was one of the main arguments for the notion that hyperdiploidy cannot be caused by stepwise or sequential gains from a diploid karyotype or by consecutive losses from a tetraploid karyotype.

Distribution of 5-methylcytosine-rich regions in the metaphase chromosomes of Vicia faba

The DNA methylation pattern of Vicia faba metaphase chromosomes was examined with a specific monoclonal antibody. 5-methylcytosine (5-mC) residues are present in different chromosomal sites, and are particularly abundant in telomeric and/or subtelomeric regions and in certain intercalary bands. Chromosomal localization of methylated regions enables a better knowledge of the lengthwise differentiation of this chromosome complement. Our results also indicate that there may be differences in monoclonal antibody binding between corresponding regions of homologous chromosomes in V. faba. This behaviour is detectable in specific regions with different frequencies. The data support results previously obtained for Allium cepa metaphase chromosomes using the same monoclonal antibody.

Gene silencing in transgenic tobacco hybrids: frequency of the event and visualization of somatic inactivation pattern

We have investigated the stability of the expression of different T-DNA-borne genes in hybrid tobacco lines. These lines were constructed to rescue rolC-induced male sterility in kanamycin-resistant P35s-rolC transgenic tobacco plants by expression of rolC antisense genes. Using five different tester lines, a total of 158 hybrids was obtained. We observed inactivation of transgene expression in 20% of the F1 progeny and in 35% of the backcrossed F2 progeny, as indicated by the loss of kanamycin resistance. In 3% of all crosses complete loss of antibiotic resistance was noted, while in most affected hybrid progeny only part of the population became kanamycin sensitive. Single genes could be selectively inactivated on T-DNAs harboring several genes. Gene inactivation was not restricted to one of the two T-DNAs examined. Somatic silencing, visualized by a cell-specific 35SGUSINT marker gene, occurred in a random fashion or exhibited an inherited specific pattern. The type of somatic silencing pattern observed indicated developmental control of the process. Two phenotypic classes could be distinguished with respect to frequency and timing of the inactivation process. Rapid gene inactivation, occurring within a few weeks after germination of hybrid seedlings, was characterized by complete methylation of restriction sites in the promoter of the silenced gene, resetting of gene expression during meiosis, heredity of the developmentally controlled program of gene silencing in subsequent generations, and rapid reactivation of gene expression after genetic separation of the different T-DNAs. In contrast, a slow type of gene inactivation was of a more stochastic nature and was recognized only in hybrids of the backcrossed F2 generation. In this case the degree of promoter methylation, which could extend beyond the T-DNA borders, was not correlated with the reduction in steady-state poly(A)+ mRNA levels, the silenced state was transmitted through meiosis and reactivation lasted several generations. The implications of the observations for our understanding of the gene inactivation process are discussed.

A microassay for measuring cytosine DNA methyltransferase activity during tumor progression

The cytosine DNA methyltransferase (MT) enzyme, which catalyzes DNA methylation at CpG sites, is overexpressed at the mRNA level during the progressive stages of colon cancer. This paper describes the adaption of a sensitive microassay for determining MT enzyme activity during tumor progression in human colon and murine lung. MT activity was progressively elevated in mucosa from familial adenomatosis polyposis patients, mucosa adjacent to cancers, and in colonic adenocarcinomas when compared to colonic mucosa from control patients. In addition, the activity of this enzyme was increased in alveolar type II but not Clara cells isolated from A/J mice following carcinogen exposure and continued to increase during tumor progression. The use of a microassay for measuring MT activity indicates that changes in enzyme activity were in general agreement with previous findings of increased MT mRNA levels during colon cancer progression and also implicates the involvement of this pathway in lung cancer development.

DNA methylation associated with repeat-induced point mutation in Neurospora crassa

Repeat-induced point mutation (RIP) is a process that efficiently detects DNA duplications prior to meiosis in Neurospora crassa and peppers them with G:C to A:T mutations. Cytosine methylation is typically associated with sequences affected by RIP, and methylated cytosines are not limited to CpG dinucleotides. We generated and characterized a collection of methylated and unmethylated amRIP alleles to investigate the connection(s) between DNA methylation and mutations by RIP. Alleles of am harboring 84 to 158 mutations in the 2.6-kb region that was duplicated were heavily methylated and triggered de novo methylation when reintroduced into vegetative N. crassa cells. Alleles containing 45 and 56 mutations were methylated in the strains originally isolated but did not become methylated when reintroduced into vegetative cells. This provides the first evidence for de novo methylation in the sexual cycle and for a maintenance methylation system in Neurospora cells. No methylation was detected in am alleles containing 8 and 21 mutations. All mutations in the eight primary alleles studied were either G to A or C to T, with respect to the coding strand of the am gene, suggesting that RIP results in only one type of mutation. We consider possibilities for how DNA methylation is triggered by some sequences altered by RIP.

Physiological aspects of genome variability in tissue culture. I. Growth phase-dependent differential DNA methylation of the carrot genome (Daucus carota L.) during primary culture

Investigations were performed on growth phase-dependent EcoRII site-specific DNA methylation of the carrot genome during primary culture to elucidate physiological aspects of genome DNA variability in tissue culture. While DNA methylation of the root cambium and the secondary phloem and petioles of carrot leaves were strikingly different, the methylation level of the secondary phloem seemed to be independent of cultivar origin, the age of the plants and the extent of secondary root growth. As was shown earlier a change in the differentiated state of the secondary phloem by tissue culture leads to changes in genome modification. Whereas de novo methylation was observed during the first 2 weeks of growth initiation, the results presented demonstrate genome de-methylation during the transition to stationary growth indicating differential εnome methylation during different phases of culture. The presence of kinetin in the nutrient medium of the primary culture was found to be antagonistic to changes in genome modification in general. De novo methylation and subsequent de-methylation of the carrot genome are discussed as gross changes obviously essential to molecular genome differentiation during tissue culture.

X-chromosome inactivation and cell memory

Mammalian X-chromosome inactivation is an excellent example of the faithful maintenance of a determined chromosomal state. As such, it may provide insight into the mechanisms for cell memory, defined as the faithful maintenance of a determined state in clonally derived progeny cells. We review here the aspects of X-chromosome inactivation that are relevant to cell memory and discuss the various molecular mechanisms that have been proposed to explain its occurrence, with emphasis on DNA methylation and a recently proposed mechanism that depends on the timing of replication.

Sequence dependent effects of CpG cytosine methylation. A joint 1H-NMR and 31P-NMR study

The impact of cytosine methylation in the central CpG step of two closely related octanucleotide duplexes d(CATCGATG)2 and d(CTTCGAAG)2 was examined by 1H-NMR and 31P-NMR experiments, and a quantitative structural analysis was performed using the NOE-derived distances, the sugar puckers and the epsilon torsion angles. The two starting oligonucleotides displayed a B-DNA conformation with, however, significant local structure differences. Although the methylated oligonucleotides retained their B-DNA conformation, different structural and thermal stability effects were observed. The magnitude of the methylation effects was to depend on the initial conformation of the CpG site, which is governed by the nature of the dinucleotide AT or TT located on the CpG flanks. As an example of sequence dependence, the methylation of CpG entailed larger conformational variation in d(CATCGATG)2 than in d(CTTCGAAG)2. In this study, the 1H and 31P chemical-shift parameters averred as extremely sensitive probes for detecting subtle conformational changes. Finally, our comparative results may aid our understanding of the structural and related biological effects produced by cytosine methylation in DNA.

DNA methylation is not involved in the structural alterations of ornithine decarboxylase or total chromatin of c-Ha-rasVal 12 oncogene-transformed NIH-3T3 fibroblasts

The ornithine decarboxylase (odc) gene is an early response gene, whose increased expression and relaxed chromatin structure is closely coupled to neoplastic growth. In various tumour cells, the odc gene displays hypomethylation at the sequences CCGG. Hypomethylation of genes is believed to correlate with chromatin decondensation and gene expression. Since a given pattern of DNA methylation may not be preserved in neoplastic cells, we studied the methylation status of odc gene at the CCGG sequences in c-Ha-rasVal 12 oncogene-transformed NIH-3T3 fibroblasts during the growth cycle and relative to their normal counterparts. We found that the methylation state of the odc gene and its promoter and mid-coding and 3' regions remain unaltered during the cell cycle. We also found that in ras oncogene-transformed cells, which display a more decondensed nucleosomal organization of chromatin than the normal cells, the CCGG sequences in bulk DNA and at the odc gene were methylated to the same extent as in the nontransformed cells. These data suggest that DNA hypomethylation at the CCGG sequences is not a prerequisite for chromatin decondensation and cell transformation by the c-Ha-rasVal 12 oncogene.

Methylation and expression of a metallothionein promoter ovine growth hormone fusion gene (MToGH1) in transgenic mice

We have examined transgene methylation in the DNA from the livers of a pedigree of mice carrying three copies of an integrated MToGH1 transgene. Utilizing the methylation-sensitive isoschizomers Msp I and Hpa II, Southern blot analysis revealed that all second generation animals derived from a transgenic female had hypermethylated DNA, whereas first generation animals sired by a transgenic male displayed a range of methylation phenotypes ranging from no methylation to hypermethylation of the transgene sequences. Of the mice that exhibited hypermethylation of the transgene in CpG dinucleotides (CmCGG), a minority of these animals also exhibited apparent CpC methylation (i.e. inhibition of Msp I cutting, presumably blocked by methylation of the outer C of CCGG). Methylation was also examined in the inner C of CC(A/T)GG sequences in the MToGH1 transgene using the isoschizomer pair BstN I and EcoR II. A minority of MToGH1 animals in the F1 generation showed clear evidence of methylation in these sites as well as in the inner and outer Cs of CCGG sites. An examination of MToGH1 expression in terms of oGH levels in serum revealed that there was a high degree of variation in the levels of circulating oGH between animals of this pedigree. There was a weak inverse relationship between the serum level of oGH and the extent of methylation of the transgene. In particular, mice exhibiting CpC together with CpG methylation were found to have very low levels of circulating oGH. Our results highlight the nature and complexity of epigenetic factors associated with transgene sequences which may ultimately influence expression of introduced genes in the mammalian genome.

Epigenetic programming of differential gene expression in development and evolution

This review covers data on changing patterns of DNA methylation and the regulation of gene expression in mouse embryonic development. Global demethylation occurs from the eight-cell stage to the blastocyst stage in preimplantation embryos, and global de novo methylation begins at implantation. We have used X-chromosome inactivation in female embryos as a model system to study specific CpG sites in the X-linked Pgk-1 and G6pd housekeeping genes and in the imprinted regulatory Xist gene to elucidate the role of methylation in the initiation and maintenance of differential gene activity. Methylation of the X-linked housekeeping genes occurs very close in time to their inactivation, thus raising the question as to whether methylation could be causal to inactivation, as well as being involved in its maintenance. A methylation difference between sperm and eggs in the promoter region of the Xist gene, located at the X-chromosome inactivation centre, is correlated with imprinted preferential inactivation of the paternal X chromosome in extra-embryonic tissues. Based on our data, a picture of the inheritance of methylation imprints and speculation on the significance of the Xist imprint in development is presented. On a more general level, an hypothesis of evolution by "adaptive epigenetic/genetic inheritance" is considered. This proposes modification of germ line DNA in response to a change in environment and mutation at the site of modification (e.g., of methylated cytosine to thymine). Epigenetic inheritance could function to shift patterns of gene expression to buffer the evolving system against changes in environment. If the altered patterns of gene activity and inactivity persist, the modifications may become "fixed" as mutations; alternatively, previously silenced gene networks might be recruited into function, thus appearing as if they are "acquired characteristics." An extension of this hypothesis is "foreign gene acquisition and sorting" (selection or silencing of gene function according to use). "Kidnapping" and sorting of foreign genes in this way could explain the observation that increased complexity in evolution is associated with more "junk" DNA. Adaptive epigenetic/genetic inheritance challenges the "central dogma" that information is unidirectional from the DNA to protein and the idea that Darwinian random mutation and selection are the sole mechanisms of evolution.

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.