Tandemly repeated genes promote RNAi-mediated heterochromatin formation via an antisilencing factor, Epe1, in fission yeast.

In most eukaryotes, constitutive heterochromatin, defined by histone H3 lysine 9 methylation (H3K9me), is enriched on repetitive DNA, such as pericentromeric repeats and transposons. Furthermore, repetitive transgenes also induce heterochromatin formation in diverse model organisms. However, the mechanisms that promote heterochromatin formation at repetitive DNA elements are still not clear. Here, using fission yeast, we show that tandemly repeated mRNA genes promote RNA interference (RNAi)-mediated heterochromatin formation in cooperation with an antisilencing factor, Epe1. Although the presence of tandemly repeated genes itself does not cause heterochromatin formation, once complementary small RNAs are artificially supplied in trans, the RNAi machinery assembled on the repeated genes starts producing cognate small RNAs in cis to autonomously maintain heterochromatin at these sites. This "repeat-induced RNAi" depends on the copy number of repeated genes and Epe1, which is known to remove H3K9me and derepress the transcription of genes underlying heterochromatin. Analogous to repeated genes, the DNA sequence underlying constitutive heterochromatin encodes widespread transcription start sites (TSSs), from which Epe1 activates ncRNA transcription to promote RNAi-mediated heterochromatin formation. Our results suggest that when repetitive transcription units underlie heterochromatin, Epe1 generates sufficient transcripts for the activation of RNAi without disruption of heterochromatin.

Cotranscriptional demethylation induces global loss of H3K4me2 from active genes in Arabidopsis.

Based on studies of animals and yeasts, methylation of histone H3 lysine 4 (H3K4me1/2/3, for mono-, di-, and tri-methylation, respectively) is regarded as the key epigenetic modification of transcriptionally active genes. In plants, however, H3K4me2 correlates negatively with transcription, and the regulatory mechanisms of this counterintuitive H3K4me2 distribution in plants remain largely unexplored. A previous genetic screen for factors regulating plant regeneration identified Arabidopsis LYSINE-SPECIFIC DEMETHYLASE 1-LIKE 3 (LDL3), which is a major H3K4me2 demethylase. Here, we show that LDL3-mediated H3K4me2 demethylation depends on the transcription elongation factor Paf1C and phosphorylation of the C-terminal domain (CTD) of RNA polymerase II (RNAPII). In addition, LDL3 binds to phosphorylated RNAPII. These results suggest that LDL3 is recruited to transcribed genes by binding to elongating RNAPII and demethylates H3K4me2 cotranscriptionally. Importantly, the negative correlation between H3K4me2 and transcription is significantly attenuated in the ldl3 mutant, demonstrating the genome-wide impacts of the transcription-driven LDL3 pathway to control H3K4me2 in plants. Our findings implicate H3K4me2 demethylation in plants as chromatin records of transcriptional activity, which ensures robust gene control.

Fast co-evolution of anti-silencing systems shapes the invasiveness of Mu-like DNA transposons in eudicots.

Transposable elements (TEs) constitute a major threat to genome stability and are therefore typically silenced by epigenetic mechanisms. In response, some TEs have evolved counteracting systems to suppress epigenetic silencing. In the model plant Arabidopsis thaliana, two such anti-silencing systems have been identified and found to be mediated by the VANC DNA-binding proteins encoded by VANDAL transposons. Here, we show that anti-silencing systems have rapidly diversified since their origin in eudicots by gaining and losing VANC-containing domains, such as DUF1985, DUF287, and Ulp1, as well as target sequence motifs. We further demonstrate that these motifs determine anti-silencing specificity by sequence, density, and helical periodicity. Moreover, such rapid diversification yielded at least 10 distinct VANC-induced anti-silencing systems in Arabidopsis. Strikingly, anti-silencing of non-autonomous VANDALs, which can act as reservoirs of 24-nt small RNAs, is critical to prevent the demise of cognate autonomous TEs and to ensure their propagation. Our findings illustrate how complex co-evolutionary dynamics between TEs and host suppression pathways have shaped the emergence of new epigenetic control mechanisms.

Arms race between anti-silencing and RdDM in noncoding regions of transposable elements.

Transposable elements (TEs) are among the most dynamic parts of genomes. Since TEs are potentially deleterious, eukaryotes silence them through epigenetic mechanisms such as repressive histone modifications and DNA methylation. We previously reported that Arabidopsis TEs, called VANDALs, counteract epigenetic silencing through a group of sequence-specific anti-silencing proteins, VANCs. VANC proteins bind to noncoding regions of specific VANDAL copies and induce loss of silent chromatin marks. The VANC-target regions form tandem repeats, which diverge rapidly. Sequence-specific anti-silencing allows these TEs to proliferate with minimum host damage. Here, we show that RNA-directed DNA methylation (RdDM) efficiently targets noncoding regions of VANDAL TEs to silence them de novo. Thus, escape from RdDM could be a primary event leading to the rapid evolution and diversification of sequence-specific anti-silencing systems. We propose that this selfish behavior of TEs paradoxically could make them diverse and less harmful to the host.

Gene-body chromatin modification dynamics mediate epigenome differentiation in Arabidopsis.

Heterochromatin is marked by methylation of lysine 9 on histone H3 (H3K9me). A puzzling feature of H3K9me is that this modification localizes not only in promoters but also in internal regions (bodies) of silent transcription units. Despite its prevalence, the biological significance of gene-body H3K9me remains enigmatic. Here we show that H3K9me-associated removal of H3K4 monomethylation (H3K4me1) in gene bodies mediates transcriptional silencing. Mutations in an Arabidopsis H3K9 demethylase gene IBM1 induce ectopic H3K9me2 accumulation in gene bodies, with accompanying severe developmental defects. Through suppressor screening of the ibm1-induced developmental defects, we identified the LDL2 gene, which encodes a homolog of conserved H3K4 demethylases. The ldl2 mutation suppressed the developmental defects, without suppressing the ibm1-induced ectopic H3K9me2. The ectopic H3K9me2 mark directed removal of gene-body H3K4me1 and caused transcriptional repression in an LDL2-dependent manner. Furthermore, mutations of H3K9 methylases increased the level of H3K4me1 in the gene bodies of various transposable elements, and this H3K4me1 increase is a prerequisite for their transcriptional derepression. Our results uncover an unexpected role of gene-body H3K9me2/H3K4me1 dynamics as a mediator of heterochromatin silencing and epigenome differentiation.

Centromere-targeted de novo integrations of an LTR retrotransposon of Arabidopsis lyrata.

The plant genome evolves with rapid proliferation of LTR-type retrotransposons, which is associated with their clustered accumulation in gene-poor regions, such as centromeres. Despite their major role for plant genome evolution, no mobile LTR element with targeted integration into gene-poor regions has been identified in plants. Here, we report such targeted integrations de novo. We and others have previously shown that an ATCOPIA93 family retrotransposon in Arabidopsis thaliana is mobilized when the DNA methylation machinery is compromised. Although ATCOPIA93 family elements are low copy number in the wild-type A. thaliana genome, high-copy-number related elements are found in the wild-type Arabidopsis lyrata genome, and they show centromere-specific localization. To understand the mechanisms for the clustered accumulation of the A. lyrata elements directly, we introduced one of them, named Tal1 (Transposon of Arabidopsis lyrata 1), into A. thaliana by transformation. The introduced Tal1 was retrotransposed in A. thaliana, and most of the retrotransposed copies were found in centromeric repeats of A. thaliana, suggesting targeted integration. The targeted integration is especially surprising because the centromeric repeat sequences differ considerably between A. lyrata and A. thaliana. Our results revealed unexpectedly dynamic controls for evolution of the transposon-rich heterochromatic regions.

Genome-wide negative feedback drives transgenerational DNA methylation dynamics in Arabidopsis.

Epigenetic variations of phenotypes, especially those associated with DNA methylation, are often inherited over multiple generations in plants. The active and inactive chromatin states are heritable and can be maintained or even be amplified by positive feedback in a transgenerational manner. However, mechanisms controlling the transgenerational DNA methylation dynamics are largely unknown. As an approach to understand the transgenerational dynamics, we examined long-term effect of impaired DNA methylation in Arabidopsis mutants of the chromatin remodeler gene DDM1 (Decrease in DNA Methylation 1) through whole genome DNA methylation sequencing. The ddm1 mutation induces a drastic decrease in DNA methylation of transposable elements (TEs) and repeats in the initial generation, while also inducing ectopic DNA methylation at hundreds of loci. Unexpectedly, this ectopic methylation can only be seen after repeated self-pollination. The ectopic cytosine methylation is found primarily in the non-CG context and starts from 3' regions within transcription units and spreads upstream. Remarkably, when chromosomes with reduced DNA methylation were introduced from a ddm1 mutant into a DDM1 wild-type background, the ddm1-derived chromosomes also induced analogous de novo accumulation of DNA methylation in trans. These results lead us to propose a model to explain the transgenerational DNA methylation redistribution by genome-wide negative feedback. The global negative feedback, together with local positive feedback, would ensure robust and balanced differentiation of chromatin states within the genome.

Mobilization of a plant transposon by expression of the transposon-encoded anti-silencing factor.

Transposable elements (TEs) have a major impact on genome evolution, but they are potentially deleterious, and most of them are silenced by epigenetic mechanisms, such as DNA methylation. Here, we report the characterization of a TE encoding an activity to counteract epigenetic silencing by the host. In Arabidopsis thaliana, we identified a mobile copy of the Mutator-like element (MULE) with degenerated terminal inverted repeats (TIRs). This TE, named Hiun (Hi), is silent in wild-type plants, but it transposes when DNA methylation is abolished. When a Hi transgene was introduced into the wild-type background, it induced excision of the endogenous Hi copy, suggesting that Hi is the autonomously mobile copy. In addition, the transgene induced loss of DNA methylation and transcriptional activation of the endogenous Hi. Most importantly, the trans-activation of Hi depends on a Hi-encoded protein different from the conserved transposase. Proteins related to this anti-silencing factor, which we named VANC, are widespread in the non-TIR MULEs and may have contributed to the recent success of these TEs in natural Arabidopsis populations.

DNA Methylation within Transcribed Regions.

DNA methylation within transcribed genes is commonly found in diverse animals and plants. Here, we provide an overview of recent advances and the remaining mystery regarding intragenic DNA methylation.

Control of transposable elements in Arabidopsis thaliana.

Arabidopsis thaliana serves as a very good model organism to investigate the control of transposable elements (TEs) by genetic and genomic approaches. As TE movements are potentially deleterious to the hosts, hosts silence TEs by epigenetic mechanisms, such as DNA methylation. DNA methylation is controlled by DNA methyltransferases and other regulators, including histone modifiers and chromatin remodelers. RNAi machinery directs DNA methylation to euchromatic TEs, which is under developmental control. In addition to the epigenetic controls, some TEs are controlled by environmental factors. TEs often affect expression of nearby genes, providing evolutionary sources for epigenetic, developmental, and environmental gene controls, which could even be beneficial for the host.

Bursts of retrotransposition reproduced in Arabidopsis.

Retrotransposons, which proliferate by reverse transcription of RNA intermediates, comprise a major portion of plant genomes. Plants often change the genome size and organization during evolution by rapid proliferation and deletion of long terminal repeat (LTR) retrotransposons. Precise transposon sequences throughout the Arabidopsis thaliana genome and the trans-acting mutations affecting epigenetic states make it an ideal model organism with which to study transposon dynamics. Here we report the mobilization of various families of endogenous A. thaliana LTR retrotransposons identified through genetic and genomic approaches with high-resolution genomic tiling arrays and mutants in the chromatin-remodelling gene DDM1 (DECREASE IN DNA METHYLATION 1). Using multiple lines of self-pollinated ddm1 mutant, we detected an increase in copy number, and verified this for various retrotransposons in a gypsy family (ATGP3) and copia families (ATCOPIA13, ATCOPIA21, ATCOPIA93), and also for a DNA transposon of a Mutator family, VANDAL21. A burst of retrotransposition occurred stochastically and independently for each element, suggesting an additional autocatalytic process. Furthermore, comparison of the identified LTR retrotransposons in related Arabidopsis species revealed that a lineage-specific burst of retrotransposition of these elements did indeed occur in natural Arabidopsis populations. The recent burst of retrotransposition in natural population is targeted to centromeric repeats, which is presumably less harmful than insertion into genes. The ddm1-induced retrotransposon proliferations and genome rearrangements mimic the transposon-mediated genome dynamics during evolution and provide experimental systems with which to investigate the controlling molecular factors directly.

Autocatalytic differentiation of epigenetic modifications within the Arabidopsis genome.

In diverse eukaryotes, constitutively silent sequences, such as transposons and repeats, are marked by methylation at histone H3 lysine 9 (H3K9me). Although selective H3K9me is critical for maintaining genome integrity, mechanisms to exclude H3K9me from active genes remain largely unexplored. Here, we show in Arabidopsis that the exclusion depends on a histone demethylase gene, IBM1 (increase in BONSAI methylation). Loss-of-function ibm1 mutation results in ectopic H3K9me and non-CG methylation in thousands of genes. The ibm1-induced genic H3K9me depends on both histone methylase KYP/SUVH4 and DNA methylase CMT3, suggesting interdependence of two epigenetic marks--H3K9me and non-CG methylation. Notably, IBM1 enhances loss of H3K9me in transcriptionally de-repressed sequences. Furthermore, disruption of transcription in genes induces ectopic non-CG methylation, which mimics the loss of IBM1 function. We propose that active chromatin is stabilized by an autocatalytic loop of transcription and H3K9 demethylation. This process counteracts a similarly autocatalytic accumulation of silent epigenetic marks, H3K9me and non-CG methylation.

Arabidopsis HDA6 regulates locus-directed heterochromatin silencing in cooperation with MET1.

Heterochromatin silencing is pivotal for genome stability in eukaryotes. In Arabidopsis, a plant-specific mechanism called RNA-directed DNA methylation (RdDM) is involved in heterochromatin silencing. Histone deacetylase HDA6 has been identified as a component of such machineries; however, its endogenous targets and the silencing mechanisms have not been analyzed globally. In this study, we investigated the silencing mechanism mediated by HDA6. Genome-wide transcript profiling revealed that the loci silenced by HDA6 carried sequences corresponding to the RDR2-dependent 24-nt siRNAs, however their transcript levels were mostly unaffected in the rdr2 mutant. Strikingly, we observed significant overlap of genes silenced by HDA6 to those by the CG DNA methyltransferase MET1. Furthermore, regardless of dependence on RdDM pathway, HDA6 deficiency resulted in loss of heterochromatic epigenetic marks and aberrant enrichment for euchromatic marks at HDA6 direct targets, along with ectopic expression of these loci. Acetylation levels increased significantly in the hda6 mutant at all of the lysine residues in the H3 and H4 N-tails, except H4K16. Interestingly, we observed two different CG methylation statuses in the hda6 mutant. CG methylation was sustained in the hda6 mutant at some HDA6 target loci that were surrounded by flanking DNA-methylated regions. In contrast, complete loss of CG methylation occurred in the hda6 mutant at the HDA6 target loci that were isolated from flanking DNA methylation. Regardless of CG methylation status, CHG and CHH methylation were lost and transcriptional derepression occurred in the hda6 mutant. Furthermore, we show that HDA6 binds only to its target loci, not the flanking methylated DNA, indicating the profound target specificity of HDA6. We propose that HDA6 regulates locus-directed heterochromatin silencing in cooperation with MET1, possibly recruiting MET1 to specific loci, thus forming the foundation of silent chromatin structure for subsequent non-CG methylation.

H3K9 methylation regulates heterochromatin silencing through incoherent feedforward loops.

Histone H3 lysine-9 methylation (H3K9me) is a hallmark of the condensed and transcriptionally silent heterochromatin. It remains unclear how H3K9me controls transcription silencing and how cells delimit H3K9me domains to avoid silencing essential genes. Here, using Arabidopsis genetic systems that induce H3K9me2 in genes and transposons de novo, we show that H3K9me2 accumulation paradoxically also causes the deposition of the euchromatic mark H3K36me3 by a SET domain methyltransferase, ASHH3. ASHH3-induced H3K36me3 confers anti-silencing by preventing the demethylation of H3K4me1 by LDL2, which mediates transcriptional silencing downstream of H3K9me2. These results demonstrate that H3K9me2 not only facilitates but orchestrates silencing by actuating antagonistic silencing and anti-silencing pathways, providing insights into the molecular basis underlying proper partitioning of chromatin domains and the creation of metastable epigenetic variation.

Molecular and structural basis of the chromatin remodeling activity by Arabidopsis DDM1.

The histone H2A variant H2A.W occupies transposons and thus prevents access to them in Arabidopsis thaliana. H2A.W is deposited by the chromatin remodeler DDM1, which also promotes the accessibility of chromatin writers to heterochromatin by an unknown mechanism. To shed light on this question, we solve the cryo-EM structures of nucleosomes containing H2A and H2A.W, and the DDM1-H2A.W nucleosome complex. These structures show that the DNA end flexibility of the H2A nucleosome is higher than that of the H2A.W nucleosome. In the DDM1-H2A.W nucleosome complex, DDM1 binds to the N-terminal tail of H4 and the nucleosomal DNA and increases the DNA end flexibility of H2A.W nucleosomes. Based on these biochemical and structural results, we propose that DDM1 counters the low accessibility caused by nucleosomes containing H2A.W to enable the maintenance of repressive epigenetic marks on transposons and prevent their activity.

RNAi-independent de novo DNA methylation revealed in Arabidopsis mutants of chromatin remodeling gene DDM1.

Methylation of histone H3 lysine 9 (H3K9me) and small RNAs are associated with constitutively silent chromatin in diverse eukaryotes including plants. In plants, silent transposons are also marked by cytosine methylation, especially at non-CpG sites. Transposon-specific non-CpG methylation in plants is controlled by small RNAs and H3K9me. Although it is often assumed that small RNA directs H3K9me, interaction between small RNA and H3K9me has not been directly demonstrated in plants. We have previously shown that a mutation in the chromatin remodeling gene DDM1 (DECREASE IN DNA METHYLATION 1) induces a global decrease but a local increase of cytosine methylation and accumulation of small RNA at a locus called BONSAI. Here we show that de novo BONSAI methylation does not depend on RNAi but does depend on H3K9me. In mutants of H3K9 methyltransferase gene KRYPTONITE or the H3K9me-dependent DNA methyltransferase gene CHROMOMETHYALSE3, the ddm1-induced de novo cytosine methylation was abolished for all three contexts (CpG, CpHpG and CpHpH). Furthermore, RNAi mutants showed strong developmental defects when combined with the ddm1 mutation. Our results revealed unexpected interactions of epigenetic modifications that may be conserved among diverse eukaryotes.

An Arabidopsis jmjC domain protein protects transcribed genes from DNA methylation at CHG sites.

Differential cytosine methylation of genes and transposons is important for maintaining integrity of plant genomes. In Arabidopsis, transposons are heavily methylated at both CG and non-CG sites, whereas the non-CG methylation is rarely found in active genes. Our previous genetic analysis suggested that a jmjC domain-containing protein IBM1 (increase in BONSAI methylation 1) prevents ectopic deposition of non-CG methylation, and this process is necessary for normal Arabidopsis development. Here, we directly determined the genomic targets of IBM1 through high-resolution genome-wide analysis of DNA methylation. The ibm1 mutation induced extensive hyper-methylation in thousands of genes. Transposons were unaffected. Notably, long transcribed genes were most severely affected. Methylation of genes is limited to CG sites in wild type, but CHG sites were also methylated in the ibm1 mutant. The ibm1-induced hyper-methylation did not depend on previously characterized components of the RNAi-based DNA methylation machinery. Our results suggest novel transcription-coupled mechanisms to direct genic methylation not only at CG but also at CHG sites. IBM1 prevents the CHG methylation in genes, but not in transposons.

Transgenerational epigenetic control of constitutive heterochromatin, transposons, and centromeres.

Epigenetic mechanisms are important not only for development but also for genome stability and chromosome dynamics. The latter types of epigenetic controls can often be transgenerational. Here, we review recent progress in two examples of transgenerational epigenetic control: i) the control of constitutive heterochromatin and transposable elements and ii) epigenetic mechanisms that regulate centromere specification and functions. We also discuss the biological significance of enigmatic associations among centromeres, transposons, and constitutive heterochromatin.

Identification of ecdysone receptor target genes in the worker honey bee brains during foraging behavior.

Ecdysone signaling plays central roles in morphogenesis and female ovarian development in holometabolous insects. In the European honey bee (Apis mellifera L.), however, ecdysone receptor (EcR) is expressed in the brains of adult workers, which have already undergone metamorphosis and are sterile with shrunken ovaries, during foraging behavior. Aiming at unveiling the significance of EcR signaling in the worker brain, we performed chromatin-immunoprecipitation sequencing of EcR to search for its target genes using the brains of nurse bees and foragers. The majority of the EcR targets were common between the nurse bee and forager brains and some of them were known ecdysone signaling-related genes. RNA-sequencing analysis revealed that some EcR target genes were upregulated in forager brains during foraging behavior and some were implicated in the repression of metabolic processes. Single-cell RNA-sequencing analysis revealed that EcR and its target genes were expressed mostly in neurons and partly in glial cells in the optic lobes of the forager brain. These findings suggest that in addition to its role during development, EcR transcriptionally represses metabolic processes during foraging behavior in the adult worker honey bee brain.

Retrotransposon addiction promotes centromere function via epigenetically activated small RNAs.

Retrotransposons have invaded eukaryotic centromeres in cycles of repeat expansion and purging, but the function of centromeric retrotransposons, if any, has remained unclear. In Arabidopsis, centromeric ATHILA retrotransposons give rise to epigenetically activated short interfering RNAs (easiRNAs) in mutants in DECREASE IN DNA METHYLATION1 (DDM1), which promote histone H3 lysine-9 di-methylation (H3K9me2). Here, we show that mutants which lose both DDM1 and RNA dependent RNA polymerase (RdRP) have pleiotropic developmental defects and mis-segregation of chromosome 5 during mitosis. Fertility defects are epigenetically inherited with the centromeric region of chromosome 5, and can be rescued by directing artificial small RNAs to a single family of ATHILA5 retrotransposons specifically embedded within this centromeric region. easiRNAs and H3K9me2 promote pericentromeric condensation, chromosome cohesion and proper chromosome segregation in mitosis. Insertion of ATHILA silences transcription, while simultaneously making centromere function dependent on retrotransposon small RNAs, promoting the selfish survival and spread of centromeric retrotransposons. Parallels are made with the fission yeast S. pombe, where chromosome segregation depends on RNAi, and with humans, where chromosome segregation depends on both RNAi and HELLSDDM1.