Tandem repeats in the CpG islands of imprinted genes

NanoImprint: A DNA methylation tool for clinical interpretation and diagnosis of common imprinting disorders using nanopore long-read sequencing

INTRODUCTION

Long-read whole genome sequencing like Oxford Nanopore Technology, is increasingly being introduced in clinical settings. With its ability to simultaneously call sequence variation and DNA modifications including 5-methylcytosine, nanopore is a promising technology to improve diagnostics of imprinting disorders.

METHODS

Currently, no tools to analyze DNA methylation patterns at known clinically relevant imprinted regions are available. Here we present NanoImprint, which generates an easily interpretable report, based on long-read nanopore sequencing, to use for identifying clinical relevant abnormalities in methylation levels at 14 imprinted regions and diagnosis of common imprinting disorders.

RESULTS AND CONCLUSION

NanoImprint outputs a summarizing table and visualization plots displays methylation frequency (%) and chromosomal positions for all regions, with phased data color-coded for the two alleles. We demonstrate the utility of NanoImprint using three imprinting disorder samples from patients with Beckwith-Wiedemann syndrome (BWS), Angelman syndrome (AS) and Prader-Willi syndrome (PWS). NanoImprint script is available from https://github.com/carolinehey/NanoImprint.

High-fidelity, large-scale targeted profiling of microsatellites

Microsatellites are highly mutable sequences that can serve as markers for relationships among individuals or cells within a population. The accuracy and resolution of reconstructing these relationships depends on the fidelity of microsatellite profiling and the number of microsatellites profiled. However, current methods for targeted profiling of microsatellites incur significant "stutter" artifacts that interfere with accurate genotyping, and sequencing costs preclude whole-genome microsatellite profiling of a large number of samples. We developed a novel method for accurate and cost-effective targeted profiling of a panel of more than 150,000 microsatellites per sample, along with a computational tool for designing large-scale microsatellite panels. Our method addresses the greatest challenge for microsatellite profiling-"stutter" artifacts-with a low-temperature hybridization capture that significantly reduces these artifacts. We also developed a computational tool for accurate genotyping of the resulting microsatellite sequencing data that uses an ensemble approach integrating three microsatellite genotyping tools, which we optimize by analysis of de novo microsatellite mutations in human trios. Altogether, our suite of experimental and computational tools enables high-fidelity, large-scale profiling of microsatellites, which may find utility in diverse applications such as lineage tracing, population genetics, ecology, and forensics.

Deletion of Meg8-DMR Enhances Migration and Invasion of MLTC-1 Depending on the CTCF Binding Sites

The Dlk1-Dio3 imprinted domain on mouse chromosome 12 contains three well-characterized paternally methylated differentially methylated regions (DMRs): IG-DMR, Gtl2-DMR, and Dlk1-DMR. These DMRs control the expression of many genes involved in embryonic development, inherited diseases, and human cancer in this domain. The first maternal methylation DMR discovered in this domain was the Meg8-DMR, the targets and biological function of which are still unknown. Here, using an enhancer-blocking assay, we first dissected the functional parts of the Meg8-DMR and showed that its insulator activity is dependent on the CCCTC-binding factor (CTCF) in MLTC-1. Results from RNA-seq showed that the deletion of the Meg8-DMR and its compartment CTCF binding sites, but not GGCG repeats, lead to the downregulation of numerous genes on chromosome 12, in particular the drastically reduced expression of Dlk1 and Rtl1 in the Dlk1-Dio3 domain, while differentially expressed genes are enriched in the MAPK pathway. In vitro assays revealed that the deletion of the Meg8-DMR and CTCF binding sites enhances cell migration and invasion by decreasing Dlk1 and activating the Notch1-Rhoc-MAPK/ERK pathway. These findings enhance research into gene regulation in the Dlk1-Dio3 domain by indicating that the Meg8-DMR functions as a long-range regulatory element which is dependent on CTCF binding sites and affects multiple genes in this domain.

The Dynamic Regulation of G-Quadruplex DNA Structures by Cytosine Methylation

It is well known that certain non B-DNA structures, including G-quadruplexes, are key elements that can regulate gene expression. Here, we explore the theory that DNA modifications, such as methylation of cytosine, could act as a dynamic switch by promoting or alleviating the structural formation of G-quadruplex structures in DNA or RNA. The interaction between epigenetic DNA modifications, G4 formation, and the 3D architecture of the genome is a complex and developing area of research. Although there is growing evidence for such interactions, a great deal still remains to be discovered. In vivo, the potential effect that cytosine methylation may have on the formation of DNA structures has remained largely unresearched, despite this being a potential mechanism through which epigenetic factors could regulate gene activity. Such interactions could represent novel mechanisms for important biological functions, including altering nucleosome positioning or regulation of gene expression. Furthermore, promotion of strand-specific G-quadruplex formation in differentially methylated genes could have a dynamic role in directing X-inactivation or the control of imprinting, and would be a worthwhile focus for future research.

Interplay between Histone and DNA Methylation Seen through Comparative Methylomes in Rare Mendelian Disorders

DNA methylation (DNAme) profiling is used to establish specific biomarkers to improve the diagnosis of patients with inherited neurodevelopmental disorders and to guide mutation screening. In the specific case of mendelian disorders of the epigenetic machinery, it also provides the basis to infer mechanistic aspects with regard to DNAme determinants and interplay between histone and DNAme that apply to humans. Here, we present comparative methylomes from patients with mutations in the de novo DNA methyltransferases DNMT3A and DNMT3B, in their catalytic domain or their N-terminal parts involved in reading histone methylation, or in histone H3 lysine (K) methylases NSD1 or SETD2 (H3 K36) or KMT2D/MLL2 (H3 K4). We provide disease-specific DNAme signatures and document the distinct consequences of mutations in enzymes with very similar or intertwined functions, including at repeated sequences and imprinted loci. We found that KMT2D and SETD2 germline mutations have little impact on DNAme profiles. In contrast, the overlapping DNAme alterations downstream of NSD1 or DNMT3 mutations underlines functional links, more specifically between NSD1 and DNMT3B at heterochromatin regions or DNMT3A at regulatory elements. Together, these data indicate certain discrepancy with the mechanisms described in animal models or the existence of redundant or complementary functions unforeseen in humans.

Length Polymorphism and Methylation Status of UPS29 Minisatellite of the ACAP3 Gene as Molecular Biomarker of Epilepsy. Sex Differences in Seizure Types and Symptoms

Epilepsy is a neurological disease with different clinical forms and inter-individuals heterogeneity, which may be associated with genetic and/or epigenetic polymorphisms of tandem-repeated noncoding DNA. These polymorphisms may serve as predictive biomarkers of various forms of epilepsy. ACAP3 is the protein regulating morphogenesis of neurons and neuronal migration and is an integral component of important signaling pathways. This study aimed to carry out an association analysis of the length polymorphism and DNA methylation of the UPS29 minisatellite of the ACAP3 gene in patients with epilepsy. We revealed an association of short UPS29 alleles with increased risk of development of symptomatic and cryptogenic epilepsy in women, and also with cerebrovascular pathologies, structural changes in the brain, neurological status, and the clinical pattern of seizures in both women and men. The increase of frequency of hypomethylated UPS29 alleles in men with symptomatic epilepsy, and in women with both symptomatic and cryptogenic epilepsy was observed. For patients with hypomethylated UPS29 alleles, we also observed structural changes in the brain, neurological status, and the clinical pattern of seizures. These associations had sex-specific nature similar to a genetic association. In contrast with length polymorphism epigenetic changes affected predominantly the long UPS29 allele. We suppose that genetic and epigenetic alterations UPS29 can modify ACAP3 expression and thereby affect the development and clinical course of epilepsy.

Stability and Lability of Parental Methylation Imprints in Development and Disease

DNA methylation plays essential roles in mammals. Of particular interest are parental methylation marks that originate from the oocyte or the sperm, and bring about mono-allelic gene expression at defined chromosomal regions. The remarkable somatic stability of these parental imprints in the pre-implantation embryo—where they resist global waves of DNA demethylation—is not fully understood despite the importance of this phenomenon. After implantation, some methylation imprints persist in the placenta only, a tissue in which many genes are imprinted. Again here, the underlying epigenetic mechanisms are not clear. Mouse studies have pinpointed the involvement of transcription factors, covalent histone modifications, and histone variants. These and other features linked to the stability of methylation imprints are instructive as concerns their conservation in humans, in which different congenital disorders are caused by perturbed parental imprints. Here, we discuss DNA and histone methylation imprints, and why unravelling maintenance mechanisms is important for understanding imprinting disorders in humans.

Genomic imprinting disorders: lessons on how genome, epigenome and environment interact

Genomic imprinting, the monoallelic and parent-of-origin-dependent expression of a subset of genes, is required for normal development, and its disruption leads to human disease. Imprinting defects can involve isolated or multilocus epigenetic changes that may have no evident genetic cause, or imprinting disruption can be traced back to alterations of cis-acting elements or trans-acting factors that control the establishment, maintenance and erasure of germline epigenetic imprints. Recent insights into the dynamics of the epigenome, including the effect of environmental factors, suggest that the developmental outcomes and heritability of imprinting disorders are influenced by interactions between the genome, the epigenome and the environment in germ cells and early embryos.

Conservation of Repeats at the Mammalian KCNQ1OT1-CDKN1C Region Suggests a Role in Genomic Imprinting

KCNQ1OT1 is located in the region with the highest number of genes showing genomic imprinting, but the mechanisms controlling the genes under its influence have not been fully elucidated. Therefore, we conducted a comparative analysis of the KCNQ1/KCNQ1OT1-CDKN1C region to study its conservation across the best assembled eutherian mammalian genomes sequenced to date and analyzed potential elements that may be implicated in the control of genomic imprinting in this region. The genomic features in these regions from human, mouse, cattle, and dog show a higher number of genes and CpG islands (detected using cpgplot from EMBOSS), but lower number of repetitive elements (including short interspersed nuclear elements and long interspersed nuclear elements), compared with their whole chromosomes (detected by RepeatMasker). The KCNQ1OT1-CDKN1C region contains the highest number of conserved noncoding sequences (CNS) among mammals, where we found 16 regions containing about 38 different highly conserved repetitive elements (using mVista), such as LINE1 elements: L1M4, L1MB7, HAL1, L1M4a, L1Med, and an LTR element: MLT1H. From these elements, we found 74 CNS showing high sequence identity (>70%) between human, cattle, and mouse, from which we identified 13 motifs (using Multiple Em for Motif Elicitation/Motif Alignment and Search Tool) with a significant probability of occurrence, 3 of which were the most frequent and were used to find transcription factor-binding sites. We detected several transcription factors (using JASPAR suite) from the families SOX, FOX, and GATA. A phylogenetic analysis of these CNS from human, marmoset, mouse, rat, cattle, dog, horse, and elephant shows branches with high levels of support and very similar phylogenetic relationships among these groups, confirming previous reports. Our results suggest that functional DNA elements identified by comparative genomics in a region densely populated with imprinted mammalian genes may be related to the regulation of imprinted gene expression.

Analysis of neonatal brain lacking ATRX or MeCP2 reveals changes in nucleosome density, CTCF binding and chromatin looping

ATRX and MeCP2 belong to an expanding group of chromatin-associated proteins implicated in human neurodevelopmental disorders, although their gene-regulatory activities are not fully resolved. Loss of ATRX prevents full repression of an imprinted gene network in the postnatal brain and in this study we address the mechanistic aspects of this regulation. We show that ATRX binds many imprinted domains individually but that transient co-localization between imprinted domains in the nuclei of neurons does not require ATRX. We demonstrate that MeCP2 is required for ATRX recruitment and that deficiency of either ATRX or MeCP2 causes decreased frequency of long-range chromatin interactions associated with altered nucleosome density at CTCF-binding sites and reduced CTCF occupancy. These findings indicate that MeCP2 and ATRX regulate gene expression at a subset of imprinted domains by maintaining a nucleosome configuration conducive to CTCF binding and to the maintenance of higher order chromatin structure.

MetaImprint: an information repository of mammalian imprinted genes

Genomic imprinting is a complex genetic and epigenetic phenomenon that plays important roles in mammalian development and diseases. Mammalian imprinted genes have been identified widely by experimental strategies or predicted using computational methods. Systematic information for these genes would be necessary for the identification of novel imprinted genes and the analysis of their regulatory mechanisms and functions. Here, a well-designed information repository, MetaImprint (http://bioinfo.hrbmu.edu.cn/MetaImprint), is presented, which focuses on the collection of information concerning mammalian imprinted genes. The current version of MetaImprint incorporates 539 imprinted genes, including 255 experimentally confirmed genes, and their detailed research courses from eight mammalian species. MetaImprint also hosts genome-wide genetic and epigenetic information of imprinted genes, including imprinting control regions, single nucleotide polymorphisms, non-coding RNAs, DNA methylation and histone modifications. Information related to human diseases and functional annotation was also integrated into MetaImprint. To facilitate data extraction, MetaImprint supports multiple search options, such as by gene ID and disease name. Moreover, a configurable Imprinted Gene Browser was developed to visualize the information on imprinted genes in a genomic context. In addition, an Epigenetic Changes Analysis Tool is provided for online analysis of DNA methylation and histone modification differences of imprinted genes among multiple tissues and cell types. MetaImprint provides a comprehensive information repository of imprinted genes, allowing researchers to investigate systematically the genetic and epigenetic regulatory mechanisms of imprinted genes and their functions in development and diseases.

A distinct group of CpG islands shows differential DNA methylation between replicas of the same cell line in vitro

BACKGROUND

CpG dinucleotide-rich genomic DNA regions, known as CpG islands (CGIs), can be methylated at their cytosine residues as an epigenetic mark that is stably inherited during cell mitosis. Differentially methylated regions (DMRs) are genomic regions showing different degrees of DNA methylation in multiple samples. In this study, we focused our attention on CGIs showing different DNA methylation between two culture replicas of the same cell line.

RESULTS

We used methylation data of 35 cell lines from the Encyclopedia of DNA Elements (ENCODE) consortium to identify CpG islands that were differentially methylated between replicas of the same cell line and denoted them Inter Replicas Differentially Methylated CpG islands (IRDM-CGIs). We identified a group of IRDM-CGIs that was consistently shared by different cell lines, and denoted it common IRDM-CGIs. X chromosome CGIs were overrepresented among common IRDM-CGIs. Autosomal IRDM-CGIs were preferentially located in gene bodies and intergenic regions had a lower G + C content, a smaller mean length, and a reduced CpG percentage. Functional analysis of the genes associated with autosomal IRDM-CGIs showed that many of them are involved in DNA binding and development.

CONCLUSIONS

Our results show that several specific functional and structural features characterize common IRDM-CGIs. They may represent a specific subset of CGIs that are more prone to being differentially methylated for their intrinsic characteristics.

New insights into establishment and maintenance of DNA methylation imprints in mammals

Fundamental to genomic imprinting in mammals is the acquisition of epigenetic marks that differ in male and female gametes at 'imprinting control regions' (ICRs). These marks mediate the allelic expression of imprinted genes in the offspring. Much has been learnt about the nature of imprint marks, the times during gametogenesis at which they are laid down and some of the factors responsible especially for DNA methylation. Recent work has revealed that transcription and histone modifications are critically involved in DNA methylation acquisition, and these findings allow us to propose rational models for methylation establishment. A completely novel perspective on gametic DNA methylation has emerged from epigenomic profiling. Far more differentially methylated loci have been identified in gametes than known imprinted genes, which leads us to revise the notion that methylation of ICRs is a specifically targeted process. Instead, it seems to obey default processes in germ cells, giving rise to distinct patterns of DNA methylation in sperm and oocytes. This new insight, together with the identification of proteins that preserve DNA methylation after fertilization, emphasizes the key role played by mechanisms that selectively retain differential methylation at imprinted loci during early development. Addressing these mechanisms will be essential to understanding the specificity and evolution of genomic imprinting.

Computational studies of imprinted genes

Computational studies on imprinted genes can have very different purposes: one major aim of these studies is the identification of DNA elements that distinguish imprinted genes from biallelically expressed genes. Comparative studies may help to identify imprinting regulatory elements and to understand common mechanisms of imprinted gene regulation in mammalian species. To date, the continuously growing number of genomic and epigenetic data sets makes detailed, genome-wide analyses on imprinted genes feasible. However, imprinted genes are characterized by genomic features that can influence statistics and can make such studies difficult. Hence, comparative computational studies can get very complex and require a tight interaction between bioinformaticians and biologists. Furthermore, analyses of raw data that are generated by micro-array hybridization and high-throughput sequencing technologies require computational approaches that have been designed especially for the epigenetic field. This chapter gives an overview about databases and software that is suitable for analyses of imprinted genes. Furthermore, possible difficulties that are typical for computational and statistical analyses of imprinted genes are described.

Epigenetic mechanisms of genomic imprinting: common themes in the regulation of imprinted regions in mammals, plants, and insects

Genomic imprinting is a form of epigenetic inheritance whereby the regulation of a gene or chromosomal region is dependent on the sex of the transmitting parent. During gametogenesis, imprinted regions of DNA are differentially marked in accordance to the sex of the parent, resulting in parent-specific expression. While mice are the primary research model used to study genomic imprinting, imprinted regions have been described in a broad variety of organisms, including other mammals, plants, and insects. Each of these organisms employs multiple, interrelated, epigenetic mechanisms to maintain parent-specific expression. While imprinted genes and imprint control regions are often species and locus-specific, the same suites of epigenetic mechanisms are often used to achieve imprinted expression. This review examines some examples of the epigenetic mechanisms responsible for genomic imprinting in mammals, plants, and insects.

Genomic imprinting: recognition and marking of imprinted loci

Genomic imprinting is an epigenetic process resulting in the monoallelic parent-of-origin-specific expression of a subset of genes in the mammalian genome. The parental alleles are differentially marked by DNA methylation during gametogenesis when the genomes are in separate compartments. How methylation machinery recognizes and differentially modifies these imprinted regions in germ cells remains a key question in the field. While studies have focused on determining a sequence signature that alone could distinguish imprinted regions from the rest of the genome, recent reports do not support such a hypothesis. Rather, it is becoming clear that features such as transcription, histone modifications and higher order chromatin are employed either individually or in combination to set up parental imprints.

DNA methylation status in tissues of sheep clones

Genomic imprinting and DNA methylation play an important role in mammalian development. Many cloned animals showed heterogeneous DNA methylation profiles. However, there are fewer reports in cloned lambs because of a lack of genomic imprinting information. In this study, we investigated DNA methylation patterns in CpG islands and differentially methylated regions of putative imprinted gene Peg10 and imprinted genes Dlk1, Igf2R and H19 in cloned lambs. Five organs from two cloned lambs died shortly after birth and two normal controls were investigated. We observed normal DNA methylation profiles in cloned lambs. The imprinted genes Dlk1, Igf2R and H19 in livers, kidneys, hearts, muscles and lungs of the two cloned lambs exhibited relatively normal DNA methylation, except for Peg10 showing some differences between controls and cloned lambs. Our results indicate that somatic cell nuclear transfer-produced sheep exhibited relatively normal DNA methylation pattern and experienced normal DNA methylation reprogramming at imprinted loci.

De novo DNA methylation: a germ cell perspective

DNA methylation is a fundamentally important epigenetic modification of the mammalian genome that has widespread influences on gene expression. During germ-cell specification and maturation, epigenetic reprogramming occurs and the DNA methylation landscape is profoundly remodelled. Defects in this process have major consequences for embryonic development and are associated with several genetic disorders. In this review we report our current understanding of the molecular mechanisms associated with de novo DNA methylation in germ cells. We discuss recent discoveries connecting histone modifications, transcription and the DNA methylation machinery, and consider how these new findings could lead to a model for methylation establishment. Elucidating how DNA methylation marks are established in the germline has been a challenge for nearly 20 years, but represents a key step towards a full understanding of several biological processes including genomic imprinting, epigenetic reprogramming and the establishment of the pluripotent state in early embryos.

Unique patterns of evolutionary conservation of imprinted genes

During mammalian evolution, complex systems of epigenetic gene regulation have been established: Epigenetic mechanisms control tissue-specific gene expression, X chromosome inactivation in females and genomic imprinting. Studying DNA sequence conservation in imprinted genes, it becomes evident that evolution of gene function and evolution of epigenetic gene regulation are tightly connected. Furthermore, comparative studies allow the identification of DNA sequence features that distinguish imprinted genes from biallelically expressed genes. Among these features are CpG islands, tandem repeats and retrotransposed elements that are known to play major roles in epigenetic gene regulation. Currently, more and more genetic and epigenetic data sets become available. In future, such data sets will provide the basis for more complex investigations on epigenetic variation in human populations. Therein, an exciting topic will be the genetic and epigenetic variability of imprinted genes and its input on human disease.

Dynamic stage-specific changes in imprinted differentially methylated regions during early mammalian development and prevalence of non-CpG methylation in oocytes

Mammalian imprinted genes are associated with differentially methylated regions (DMRs) that are CpG methylated on one of the two parental chromosomes. In mice, at least 21 DMRs acquire differential methylation in the germline and many of them act as imprint centres. We previously reported the physical extents of differential methylation at 15 DMRs in mouse embryos at 12.5 days postcoitum. To reveal the ontogeny of differential methylation, we determined and compared methylation patterns of the corresponding regions in sperm and oocytes. We found that the extent of the gametic DMRs differs significantly from that of the embryonic DMRs, especially in the case of paternal gametic DMRs. These results suggest that the gametic DMR sequences should be used to extract the features specifying methylation imprint establishment in the germline: from this analysis, we noted that the maternal gametic DMRs appear as unmethylated islands in male germ cells, which suggests a novel component in the mechanism of gamete-specific marking. Analysis of selected DMRs in blastocysts revealed dynamic changes in allelic methylation in early development, indicating that DMRs are not fully protected from the major epigenetic reprogramming events occurring during preimplantation development. Furthermore, we observed non-CpG methylation in oocytes, but not in sperm, which disappeared by the blastocyst stage. Non-CpG methylation was frequently found at maternally methylated DMRs as well as non-DMR regions, suggesting its prevalence in the oocyte genome. These results provide evidence for a unique methylation profile in oocytes and reveal the surprisingly dynamic nature of DMRs in the early embryo.

Transgenerational analysis of transcriptional silencing in zebrafish

The yeast Gal4/UAS transcriptional activation system is a powerful tool for regulating gene expression in Drosophila and has been increasing in popularity for developmental studies in zebrafish. It is also useful for studying the basis of de novo transcriptional silencing. Fluorescent reporter genes under the control of multiple tandem copies of the upstream activator sequence (UAS) often show evidence of variegated expression and DNA methylation in transgenic zebrafish embryos. To characterize this systematically, we monitored the progression of transcriptional silencing of UAS-regulated transgenes that differ in their integration sites and in the repetitive nature of the UAS. Transgenic larvae were examined in three generations for tissue-specific expression of a green fluorescent protein (GFP) reporter and DNA methylation at the UAS. Single insertions containing four distinct upstream activator sequences were far less susceptible to methylation than insertions containing fourteen copies of the same UAS. In addition, transgenes that integrated in or adjacent to transposon sequence exhibited silencing regardless of the number of UAS sites included in the transgene. Placement of promoter-driven Gal4 upstream of UAS-regulated responder genes in a single bicistronic construct also appeared to accelerate silencing and methylation. The results demonstrate the utility of the zebrafish for efficient tracking of gene silencing mechanisms across several generations, as well as provide useful guidelines for optimal Gal4-regulated gene expression in organisms subject to DNA methylation.

Imprinted genes show unique patterns of sequence conservation

BACKGROUND

Genomic imprinting is an evolutionary conserved mechanism of epigenetic gene regulation in placental mammals that results in silencing of one of the parental alleles. In order to decipher interactions between allele-specific DNA methylation of imprinted genes and evolutionary conservation, we performed a genome-wide comparative investigation of genomic sequences and highly conserved elements of imprinted genes in human and mouse.

RESULTS

Evolutionarily conserved elements in imprinted regions differ from those associated with autosomal genes in various ways. Whereas for maternally expressed genes strong divergence of protein-encoding sequences is most prominent, paternally expressed genes exhibit substantial conservation of coding and noncoding sequences. Conserved elements in imprinted regions are marked by enrichment of CpG dinucleotides and low (TpG+CpA)/(2·CpG) ratios indicate reduced CpG deamination. Interestingly, paternally and maternally expressed genes can be distinguished by differences in G+C and CpG contents that might be associated with unusual epigenetic features. Especially noncoding conserved elements of paternally expressed genes are exceptionally G+C and CpG rich. In addition, we confirmed a frequent occurrence of intronic CpG islands and observed a decelerated degeneration of ancient LINE-1 repeats. We also found a moderate enrichment of YY1 and CTCF binding sites in imprinted regions and identified several short sequence motifs in highly conserved elements that might act as additional regulatory elements.

CONCLUSIONS

We discovered several novel conserved DNA features that might be related to allele-specific DNA methylation. Our results hint at reduced CpG deamination rates in imprinted regions, which affects mostly noncoding conserved elements of paternally expressed genes. Pronounced differences between maternally and paternally expressed genes imply specific modes of evolution as a result of differences in epigenetic features and a special response to selective pressure. In addition, our data support the potential role of intronic CpG islands as epigenetic key regulatory elements and suggest that evolutionary conserved LINE-1 elements fulfill regulatory functions in imprinted regions.

Identifying CpG islands by different computational techniques

CpG islands (CGIs) are generally regarded as important epigenetic regulatory elements due to their association with promoter regions. However, identification of functional CGIs is hampered by repetitive elements and species-specific particularities. Here, we compared the performance of different CGI detection programs on genomic sequences of human and mouse genes. Although mouse CGIs are shorter and G+C poorer than their human counterparts, the different tools tested in our study reliably identify CGIs in promoter regions in both species. Our study confirms that substantially fewer murine than human CGIs coincide with repetitive elements and indicates that such CGIs are subject to accelerated cytosine deamination. In addition, CpG depletion appears to anticorrelate with the epigenetic features of functional regulatory CGIs. Taking into account different deamination rates in unmethylated CGIs versus those in methylated CGIs might support the detection of functional CGIs in other species for which there is little epigenetic information available.

Genomic imprinting in germ cells: imprints are under control

The cis-acting regulatory sequences of imprinted gene loci, called imprinting control regions (ICRs), acquire specific imprint marks in germ cells, including DNA methylation. These epigenetic imprints ensure that imprinted genes are expressed exclusively from either the paternal or the maternal allele in offspring. The last few years have witnessed a rapid increase in studies on how and when ICRs become marked by and subsequently maintain such epigenetic modifications. These novel findings are summarised in this review, which focuses on the germline acquisition of DNA methylation imprints and particularly on the combined role of primary sequence specificity, chromatin configuration, non-histone proteins and transcriptional events.

The primate-specific microRNA gene cluster (C19MC) is imprinted in the placenta

Imprinted genes play crucial roles in mammalian development and disruption of their expression is associated with many human disorders including tumourigenesis; yet, the actual number of imprinted genes in the human genome remains a matter of debate. Here, we report on the unexpected finding that the chromosome 19 microRNA cluster (C19MC), the largest human microRNA gene cluster discovered so far, is regulated by genomic imprinting with only the paternally inherited allele being expressed in the placenta. DNA methylation profiling identified a differentially methylated region (C19MC-DMR1) that overlaps an upstream CpG-rich promoter region associated with short tandem repeats. It displays a maternal-specific methylation imprint acquired in oocytes and generates a complex population of large, compartimentalized non-coding RNA (ncRNA) species retained in close proximity to the C19MC transcription site. This occurs adjacent to, but not within, a poorly characterized nuclear Alu-rich domain. Interestingly, C19MC maps near another imprinted gene, the maternally expressed ZNF331 gene, and therefore may define a novel, previously unrecognized large imprinted primate-specific chromosomal domain. Altogether, our study adds C19MC to the growing list of imprinted repeated small RNA gene clusters and further strengthens the potential involvement of small ncRNAs in the function and/or the evolution of imprinted gene networks.

Divergence of imprinted genes during mammalian evolution

BACKGROUND

In contrast to the majority of mammalian genes, imprinted genes are monoallelically expressed with the choice of the active allele depending on its parental origin. Due to their special inheritance patterns, maternally and paternally expressed genes might be under different evolutionary pressure. Here, we aimed at assessing the evolutionary history of imprinted genes.

RESULTS

In this study, we investigated the conservation of imprinted genes in vertebrate genomes and their exposition to natural selection. In a genome-wide comparison, orthologs of imprinted genes show a stronger divergence on cDNA and protein level in mammals. This pattern is most pronounced for maternally expressed genes in rodents in comparison to their non-rodent orthologs. The divergence is not attributable to increased mutation of CpG positions. It is contrasted by strong conservation of paternally expressed genes in mouse and rat. Interestingly, we found that the early divergence of imprinted genes was accompanied by an unusually strict conservation of their paralogs.

CONCLUSIONS

The apparent degeneration of maternally expressed genes may reflect a relaxation of selective pressure due to counteracting effects on maternal and embryonic fitness. Functional redundancy provided by the presence of highly conserved (non-imprinted) paralogs may have facilitated the divergence. Moreover, intensification of imprinting in modern rodents seems to have shifted the evolutionary fate of imprinted genes towards strong purifying selection.

Differential decay of parent-of-origin-specific genomic sharing in cystic fibrosis-affected sib pairs maps a paternally imprinted locus to 7q34

Cystic fibrosis (CF) is a monogenic disease characterized by a high variability of disease severity and outcome that points to the role of environmental factors and modulating genes that shape the course of this multiorgan disease. We genotyped families of cystic fibrosis sib pairs homozygous for F508del-CFTR who represent extreme clinical phenotypes at informative microsatellite markers spanning a 38 Mb region between CFTR and 7qtel. Recombination events on both parental chromosomes were compared between siblings with concordant clinical phenotypes and siblings with discordant clinical phenotypes. Monitoring parent-of-origin-specific decay of genomic sharing delineated a 2.9-Mb segment on 7q34 in which excess of recombination on paternal chromosomes in discordant pairs was observed compared with phenotypically concordant sibs. This 2.9-Mb core candidate region was enriched in imprinting-related elements such as predicted CCCTC-binding factor consensus sites and CpG islands dense in repetitive elements. Moreover, allele frequencies at a microsatellite marker within the core candidate region differed significantly comparing mildly and severely affected cystic fibrosis sib pairs. The identification of this paternally imprinted locus on 7q34 as a modulator of cystic fibrosis disease severity shows that imprinted elements can be identified by straightforward fine mapping of break points in sib pairs with informative contrasting phenotypes.

Chromatin mechanisms in genomic imprinting

Mammalian imprinted genes are clustered in chromosomal domains. Their mono-allelic, parent-of-origin-specific expression is regulated by imprinting control regions (ICRs), which are essential sequence elements marked by DNA methylation on one of the two parental alleles. These methylation "imprints" are established during gametogenesis and, after fertilization, are somatically maintained throughout development. Nonhistone proteins and histone modifications contribute to this epigenetic process. The way ICRs mediate imprinted gene expression differs between domains. At some domains, for instance, ICRs produce long noncoding RNAs that mediate chromatin silencing. Lysine methylation on histone H3 is involved in this developmental process and is particularly important for imprinting in the placenta and brain. Together, the newly discovered chromatin mechanisms provide further clues for addressing imprinting-related pathologies in humans.

DNA methylation analysis of the mammalian PEG3 imprinted domain

In this study, we performed the first systematic survey of DNA methylation status of the CpG islands of the PEG3 (Paternally expressed gene 3) imprinted domain in the mouse, cow, and human genomes. Previous studies have shown that the region surrounding the first exon of PEG3 contains a differentially methylated CpG island. In addition, we have discovered two previously unreported differentially methylated regions (DMR): one in the promoter region of mouse Zim3 and another in the promoter region of human USP29. In the cow, the Peg3-CpG island was the only area that showed DMR status. We have also examined the methylation status of several CpG islands in this region using human tumor-derived DNA. The CpG islands near PEG3 and USP29 both showed hypermethylation in DNA derived from breast and ovarian tumors. Overall, this study shows that the PEG3 imprinted domain of humans, cows, and mice contains differing numbers of DMRs, but the PEG3-CpG island is the only DMR that is conserved among these three species.

A first-stage approximation to identify new imprinted genes through sequence analysis of its coding regions

In the present study, a positive training set of 30 known human imprinted gene coding regions are compared with a set of 72 randomly sampled human nonimprinted gene coding regions (negative training set) to identify genomic features common to human imprinted genes. The most important feature of the present work is its ability to use multivariate analysis to look at variation, at coding region DNA level, among imprinted and non-imprinted genes. There is a force affecting genomic parameters that appears through the use of the appropriate multivariate methods (principle components analysis (PCA) and quadratic discriminant analysis (QDA)) to analyse quantitative genomic data. We show that variables, such as CG content, [bp]% CpG islands, [bp]% Large Tandem Repeats, and [bp]% Simple Repeats, are able to distinguish coding regions of human imprinted genes.

SNP-based prediction of the human germ cell methylation landscape

Base substitution occurs at a high rate at CpG dinucleotides due to the frequent methylation of CpG and the deamination of methylated cytosine to thymine. If these substitutions occur in germ cells, they constitute a heritable mutation that may eventually rise to polymorphic frequencies, hence resulting in a SNP that is methylation associated. In this study, we sought to identify clusters of methylation associated SNPs as a basis for prediction of methylation landscapes of germ cell genomes. Genomic regions enriched with methylation associated SNPs, namely "methylation associated SNP clusters", were identified with an agglomerative hierarchical clustering algorithm. Repetitive elements, segmental duplications, and syntenic tandem DNA repeats were enriched in methylation associated SNP clusters. The frequency of methylation associated SNPs in Alu Y/S elements exhibited a gradient pattern suggestive of linear spreading, being higher in proximity to methylation associated SNP clusters and lower closer to CpG islands. Interestingly, methylation associated SNP clusters were over-represented near the transcriptional initiation sites of immune response genes. We propose a de novo DNA methylation model during germ cell development whereby a pattern is established by long-range chromatic interactions through syntenic repeats combined with regional methylation spreading from methylation associated SNP clusters.

Transcription is required for establishment of germline methylation marks at imprinted genes

Genomic imprinting requires the differential marking by DNA methylation of genes in male and female gametes. In the female germline, acquisition of methylation imprint marks depends upon the de novo methyltransferase Dnmt3a and its cofactor Dnmt3L, but the reasons why specific sequences are targets for Dnmt3a and Dnmt3L are still poorly understood. Here, we investigate the role of transcription in establishing maternal germline methylation marks. We show that at the Gnas locus, truncating transcripts from the furthest upstream Nesp promoter disrupts oocyte-derived methylation of the differentially methylated regions (DMRs). Transcription through DMRs in oocytes is not restricted to this locus but occurs across the prospective DMRs at many other maternally marked imprinted domains, suggesting a common requirement for transcription events. The transcripts implicated here in gametic methylation are protein-coding, in contrast to the noncoding antisense transcripts involved in the monoallelic silencing of imprinted genes in somatic tissues, although they often initiate from alternative promoters in oocytes. We propose that transcription is a third essential component of the de novo methylation system, which includes optimal CpG spacing and histone modifications, and may be required to create or maintain open chromatin domains to allow the methylation complex access to its preferred targets.

[Differential epigenetic marking on imprinted genes and consequences in human diseases]

At the time of fertilisation, the parental genomes have a strikingly different organisation. Sperm DNA is packaged globally with protamines, whereas the oocyte's genome is wrapped around nucleosomes. The maternal and paternal genomes are functionally different as well, and are both required for normal uterine and postnatal development. The functional requirement of both parental genomes is a consequence of differential epigenetic marking by DNA methylation during oogenesis and spermatogenesis, on a group of genes called imprinted genes. After fertilisation, these parental marks persist throughout development and convey the allelic expression of imprinted genes. Pathological perturbation of methylation imprints, before fertilisation in the germ cells, or during development, gives rise to growth-related syndromes, and is frequently observed in cancer as well. Alteration of imprints is thought to occur early in carcinogenesis and shows similarities with the acquisition of aberrant DNA methylation at tumour suppressor genes. This suggests that similar underlying mechanisms could be involved.

Characterization of the differentially methylated region of the Impact gene that exhibits Glires-specific imprinting

Comparative genomic analysis of the Impact locus, which is imprinted in Glires but not in other mammals, reveals features required for genomic imprinting.

Background

Imprinted genes are exclusively expressed from one of the two parental alleles in a parent-of-origin-specific manner. In mammals, nearly 100 genes are documented to be imprinted. To understand the mechanism behind this gene regulation and to identify novel imprinted genes, common features of DNA sequences have been analyzed; however, the general features required for genomic imprinting have not yet been identified, possibly due to variability in underlying molecular mechanisms from locus to locus.

Results

We performed a thorough comparative genomic analysis of a single locus, Impact, which is imprinted only in Glires (rodents and lagomorphs). The fact that Glires and primates diverged from each other as recent as 70 million years ago makes comparisons between imprinted and non-imprinted orthologues relatively reliable. In species from the Glires clade, Impact bears a differentially methylated region, whereby the maternal allele is hypermethylated. Analysis of this region demonstrated that imprinting was not associated with the presence of direct tandem repeats nor with CpG dinucleotide density. In contrast, a CpG periodicity of 8 bp was observed in this region in species of the Glires clade compared to those of carnivores, artiodactyls, and primates.

Conclusions

We show that tandem repeats are dispensable, establishment of the differentially methylated region does not rely on G+C content and CpG density, and the CpG periodicity of 8 bp is meaningful to the imprinting. This interval has recently been reported to be optimal for de novo methylation by the Dnmt3a-Dnmt3L complex, suggesting its importance in the establishment of imprinting in Impact and other genes.

Signaling silence--breaking ground and spreading out

Among the persistent mysteries in epigenetics are the criteria by which specific regions of the genome are chosen for deposition of distinguishing chromatin marks. Once a particular region is modified, will the newly acquired epigenetic state spill onto neighboring regions of the genome or be confined to tidy patches? The report by Henderson and Jacobsen in the the previous issue of Genes & Development (1597-1606) addresses these questions while providing insight into the utility of DNA methylation for a eukaryotic genome.

Comparative analyses of human single- and multilocus tandem repeats

Using the compiled human genome sequence, we systematically cataloged all tandem repeats with periods between 20 and 2000 bp and defined two subsets whose consensus sequences were found at either single-locus tandem repeats (slTRs) or multilocus tandem repeats (mlTRs). Parameters compiled for these subsets provide insights into mechanisms underlying the creation and evolution of tandem repeats. Both subsets of tandem repeats are nonrandomly distributed in the genome, being found at higher frequency at many but not all chromosome ends and internal clusters of mlTRs were also observed. Despite the integral role of recombination in the biology of tandem repeats, recombination hotspots colocalized only with shorter microsatellites and not the longer repeats examined here. An increased frequency of slTRs was observed near imprinted genes, consistent with a functional role, while both slTRs and mlTRs were found more frequently near genes implicated in triplet expansion diseases, suggesting a general instability of these regions. Using our collated parameters, we identified 2230 slTRs as candidates for highly informative molecular markers.

Evolution and control of imprinted FWA genes in the genus Arabidopsis

A central question in genomic imprinting is how a specific sequence is recognized as the target for epigenetic marking. In both mammals and plants, imprinted genes are often associated with tandem repeats and transposon-related sequences, but the role of these elements in epigenetic gene silencing remains elusive. FWA is an imprinted gene in Arabidopsis thaliana expressed specifically in the female gametophyte and endosperm. Tissue-specific and imprinted expression of FWA depends on DNA methylation in the FWA promoter, which is comprised of two direct repeats containing a sequence related to a SINE retroelement. Methylation of this element causes epigenetic silencing, but it is not known whether the methylation is targeted to the SINE-related sequence itself or the direct repeat structure is also necessary. Here we show that the repeat structure in the FWA promoter is highly diverse in species within the genus Arabidopsis. Four independent tandem repeat formation events were found in three closely related species. Another related species, A. halleri, did not have a tandem repeat in the FWA promoter. Unexpectedly, even in this species, FWA expression was imprinted and the FWA promoter was methylated. In addition, our expression analysis of FWA gene in vegetative tissues revealed high frequency of intra-specific variation in the expression level. In conclusion, we show that the tandem repeat structure is dispensable for the epigenetic silencing of the FWA gene. Rather, SINE-related sequence is sufficient for imprinting, vegetative silencing, and targeting of DNA methylation. Frequent independent tandem repeat formation events in the FWA promoter led us to propose that they may be a consequence, rather than cause, of the epigenetic control. The possible significance of epigenetic variation in reproductive strategies during evolution is also discussed.

Imprinted expression of the canine IGF2R, in the absence of an anti-sense transcript or promoter methylation

Imprinted genes are epigenetically modified in a parent of origin-dependent manner, and as a consequence, are differentially expressed. Although the evolution of genomic imprinting is a subject of intense debate, imprinted genes have been studied primarily in mice and humans and in a small number of marsupial mammals. Comparative studies involving rodents and primates are of limited value, as they belong to the same superordinal group of eutherian mammals (Euarchontoglires). On the other hand, comparisons involving marsupials may not be informative, due to phylogenetic distance. Canis familiaris belongs to Laurasiatheria, a sister-group of Euarchontoglires, and should prove useful in comparative studies of imprinted genes. Using RT-PCR we demonstrate monoallelic expression of the canine IGF2R in several tissues, including uterus and umbilical cord. In the case of umbilical cord, we identify the expressed allele as maternally derived. The canine IGF2R is thus an imprinted gene. Using bisulfite sequencing, we show that the canine IGF2R resembles the imprinted mouse Igf2r in having a CpG island in intron 2 that is hemi-methylated. However, it differs from the mouse gene in that maintenance of the monoallelic expression of canine IGF2R does not require expression of an anti-sense transcript from the paternally derived allele, or methylation of the repressed IGF2R promoter. In these two important features, the imprinted canine gene resembles the imprinted opossum IGF2R. Our data suggest that these features were properties of the ancestral imprinted IGF2R and that the anti-sense transcript (Air) and promoter methylation observed in mouse are derived features of the mouse Igf2r locus.

The PcG gene Sfmbt2 is paternally expressed in extraembryonic tissues

Genomic imprinting has dramatic effects on placental development, as has been clearly observed in interspecific hybrid, somatic cell nuclear transfer, and uniparental embryos. In fact, the earliest defects in uniparental embryos are evident first in the extraembryonic trophoblast. We performed a microarray comparison of gynogenetic and androgenetic mouse blastocysts, which are predisposed to placental pathologies, to identify imprinted genes. In addition to identifying a large number of known imprinted genes, we discovered that the Polycomb group (PcG) gene Sfmbt2 is imprinted. Sfmbt2 is expressed preferentially from the paternal allele in early embryos, and in later stage extraembryonic tissues. A CpG island spanning the transcriptional start site is differentially methylated on the maternal allele in e14.5 placenta. Sfmbt2 is located on proximal chromosome 2, in a region known to be imprinted, but for which no genes had been identified until now. This possibly identifies a new imprinted domain within the murine genome. We further demonstrate that murine SFMBT2 protein interacts with the transcription factor YY1, similar to the Drosophila PHO-RC.

Comparative analysis of sequence characteristics of imprinted genes in human, mouse, and cattle

Genomic imprinting is an epigenetic mechanism that results in monoallelic expression of genes depending on parent-of-origin of the allele. Although the conservation of genomic imprinting among mammalian species has been widely reported for many genes, there is accumulating evidence that some genes escape this conservation. Most known imprinted genes have been identified in the mouse and human, with few imprinted genes reported in cattle. Comparative analysis of genomic imprinting across mammalian species would provide a powerful tool for elucidating the mechanisms regulating the unique expression of imprinted genes. In this study we analyzed the imprinting of 22 genes in human, mouse, and cattle and found that in only 11 was imprinting conserved across the three species. In addition, we analyzed the occurrence of the sequence elements CpG islands, C + G content, tandem repeats, and retrotransposable elements in imprinted and in nonimprinted (control) cattle genes. We found that imprinted genes have a higher G + C content and more CpG islands and tandem repeats. Short interspersed nuclear elements (SINEs) were notably fewer in number in imprinted cattle genes compared to control genes, which is in agreement with previous reports for human and mouse imprinted regions. Long interspersed nuclear elements (LINEs) and long terminal repeats (LTRs) were found to be significantly underrepresented in imprinted genes compared to control genes, contrary to reports on human and mouse. Of considerable significance was the finding of highly conserved tandem repeats in nine of the genes imprinted in all three species.

Conserved features of imprinted differentially methylated domains

Genomic imprinting is a conserved epigenetic phenomenon in eutherian mammals, with regards both to the genes that are imprinted and the mechanism underlying the expression of just one of the parental alleles. Epigenetic modifications of alleles of imprinted genes are established during oogenesis and spermatogenesis, and these modifications are then inherited. Differentially methylated domains (DMDs) of imprinted genes are the genomic sites of these inherited epigenetic imprints. We previously showed that CpG-rich imperfect tandem direct repeats within three different mouse DMDs (Snurf/Snrpn, Kcnq1 and Igf2r), each with a unique sequence, play a central role in maintaining the differential methylation. This finding implicates repeat-related DNA structure, not sequence, in the imprinting mechanism. To better define the important features of this signal, we compared sequences of these three DMD tandem repeats among mammalian species. All DMD repeats contain short indirect repeats, many of which are organized into larger unit repeats. Even though the larger repeat units undergo deletion and addition during evolution (most likely through unequal crossovers during meiosis), the size of DMD tandem repeated regions has remained remarkably stable during mammalian evolution. Moreover, all three DMD tandem repeats have a high-CpG content, an ordered arrangement of CpG dinucleotides, and similar predicted secondary structures. These observations suggest that a structural feature or features of these DMD tandem repeats is the conserved DMD imprinting signal.

Convergent evolution of genomic imprinting in plants and mammals

Parental genomic imprinting is characterized by the expression of a selected panel of genes from one of the two parental alleles. Recent evidence shows that DNA methylation and histone modifications are responsible for this parent-of-origin-dependent expression of imprinted genes. Because similar epigenetic marks have been recruited independently in plants and mammals, the only organisms in which imprinted gene loci have been identified so far, this phenomenon represents a case for convergent evolution. Here we discuss the emerging parallels in imprinting in both taxa. We also describe the significance of imprinting for reproduction and discuss potential models for its evolution.

Genomic imprinting in mammals: emerging themes and established theories

The epigenetic events that occur during the development of the mammalian embryo are essential for correct gene expression and cell-lineage determination. Imprinted genes are expressed from only one parental allele due to differential epigenetic marks that are established during gametogenesis. Several theories have been proposed to explain the role that genomic imprinting has played over the course of mammalian evolution, but at present it is not clear if a single hypothesis can fully account for the diversity of roles that imprinted genes play. In this review, we discuss efforts to define the extent of imprinting in the mouse genome, and suggest that different imprinted loci may have been wrought by distinct evolutionary forces. We focus on a group of small imprinted domains, which consist of paternally expressed genes embedded within introns of multiexonic transcripts, to discuss the evolution of imprinting at these loci.

A screen for retrotransposed imprinted genes reveals an association between X chromosome homology and maternal germ-line methylation

Imprinted genes undergo epigenetic modifications during gametogenesis, which lead to transcriptional silencing of either the maternally or the paternally derived allele in the subsequent generation. Previous work has suggested an association between imprinting and the products of retrotransposition, but the nature of this link is not well defined. In the mouse, three imprinted genes have been described that originated by retrotransposition and overlap CpG islands which undergo methylation during oogenesis. Nap1l5, U2af1-rs1, and Inpp5f_v2 are likely to encode proteins and share two additional genetic properties: they are located within introns of host transcripts and are derived from parental genes on the X chromosome. Using these sequence features alone, we identified Mcts2, a novel candidate imprinted retrogene on mouse Chromosome 2. Mcts2 has been validated as imprinted by demonstrating that it is paternally expressed and undergoes promoter methylation during oogenesis. The orthologous human retrogenes NAP1L5, INPP5F_V2, and MCTS2 are also shown to be paternally expressed, thus delineating novel imprinted loci on human Chromosomes 4, 10, and 20. The striking correlation between imprinting and X chromosome provenance suggests that retrotransposed elements with homology to the X chromosome can be selectively targeted for methylation during mammalian oogenesis.

The conventional view is that DNA carries all of our heritable information and our genes control development into adulthood. The discovery of epigenetics, a term coined to describe effects that are not coded for by DNA sequence, but can nonetheless affect our development and well-being, has added another layer of complexity to our understanding of genetics. One class of genes under epigenetic control are imprinted genes. Mammals inherit two copies of every gene, one from mother and one from father, and in most cases, both are active. However, for a small number of imprinted genes in mammals, only one is active, either the maternal or the paternal copy. Epigenetics amounts to a control system for switching genes on and off appropriately. We focus on a group of little-studied imprinted genes that share features that give clues to their evolutionary origins. These so-called “retrogenes” are protein-coding sequences of DNA that have undergone duplication and jumped into novel locations in the genome. Because of this, it is possible to determine where, and roughly when, many of the imprinted retrogenes originated. This provides an opportunity to study the molecular events that have generated imprinted genes during mammalian evolution.

Two-step recruitment of RNA-directed DNA methylation to tandem repeats

Tandem repeat sequences are frequently associated with gene silencing phenomena. The Arabidopsis thaliana FWA gene contains two tandem repeats and is an efficient target for RNA-directed de novo DNA methylation when it is transformed into plants. We showed that the FWA tandem repeats are necessary and sufficient for de novo DNA methylation and that repeated character rather than intrinsic sequence is likely important. Endogenous FWA can adopt either of two stable epigenetic states: methylated and silenced or unmethylated and active. Surprisingly, we found small interfering RNAs (siRNAs) associated with FWA in both states. Despite this, only the methylated form of endogenous FWA could recruit further RNA-directed DNA methylation or cause efficient de novo methylation of transgenic FWA. This suggests that RNA-directed DNA methylation occurs in two steps: first, the initial recruitment of the siRNA-producing machinery, and second, siRNA-directed DNA methylation either in cis or in trans. The efficiency of this second step varies depending on the nature of the siRNA-producing locus, and at some loci, it may require pre-existing chromatin modifications such as DNA methylation itself. Enhancement of RNA-directed DNA methylation by pre-existing DNA methylation could create a self-reinforcing system to enhance the stability of silencing. Tandem repeats throughout the Arabidopsis genome produce siRNAs, suggesting that repeat acquisition may be a general mechanism for the evolution of gene silencing.

InArabidopsis, methylation and silencing mediated by theFWA transgene are found to require the tandem repeat sequences upstream ofFWA and to be independent of siRNA association.