InteractiveFly: GeneBrief

Lysine (K)-specific demethylase 4A: Biological Overview | References

Gene name - Lysine (K)-specific demethylase 4A

Synonyms -

Cytological map position - 43F2-43F2

Function - enzyme, transcription factor

Keywords - histone demethylase - regulates H3K36me3 levels - regulates heterochromatin organization and function - Jun recruits the HP1a/KDM4A complex to its gene body region upon osmotic stress to reduce H3K36 methylation levels and disrupt H3K36 methylation-dependent histone deacetylation - along with Kdm4B, Kdm4A is essential for mediating ecdysteroid hormone signaling during larval development

Symbol - Kdm4A

FlyBase ID: FBgn0033233

Genetic map position - chr2R:7,922,775-7,924,988

NCBI classification - JmjC: JmjC domain, hydroxylase

Cellular location - nuclear

NCBI links: EntrezGene, Nucleotide, Protein

Eukaryotic genomes are broadly divided between gene-rich euchromatin and the highly repetitive heterochromatin domain, which is enriched for proteins critical for genome stability and transcriptional silencing. This study shows that Drosophila KDM4A (dKDM4A), previously characterized as a euchromatic histone H3 K36 demethylase and transcriptional regulator, predominantly localizes to heterochromatin and regulates heterochromatin position-effect variegation (PEV), organization of repetitive DNAs, and DNA repair. dKDM4A demethylase activity is dispensable for PEV. In contrast, dKDM4A enzymatic activity is required to relocate heterochromatic double-strand breaks outside the domain, as well as for organismal survival when DNA repair is compromised. Finally, DNA damage triggers dKDM4A-dependent changes in the levels of H3K56me3, suggesting that dKDM4A demethylates this heterochromatic mark to facilitate repair. It is concluded that dKDM4A, in addition to its previously characterized role in euchromatin, utilizes both enzymatic and structural mechanisms to regulate heterochromatin organization and functions (Colmenares, 2017).

Heterochromatin (HC) comprises 20% and 30% of human and Drosophila genomes, respectively, and remains condensed throughout the cell cycle. HC composition is distinct from euchromatin (EC), with its low gene count and high enrichment for repetitive sequences, di- and tri-methylated histone H3 K9 (H3K9me2 and 3), and Heterochromatin Protein 1a (HP1a). HC is concentrated at pericentromeric and telomeric regions, where it plays important roles in genome stability. Disruption of HC structure impairs chromosome segregation, replication timing, transposon silencing, gene expression, and DNA repair. However, the mechanisms by which HC components mediate these diverse processes remain poorly understood (Colmenares, 2017).

Central to HC structure is the enrichment for H3K9me2 and me3, primarily catalyzed by the histone methyltransferase (HMTase) Su(var)3-9. These methyl marks form the epigenetic basis for HC maintenance by providing binding sites for HP1a. HP1a, in turn, recruits other proteins that mediate the many chromosomal and nuclear functions of HC. Loss of Su(var)3-9, HP1a, or other HC proteins leads to defects in HC function, which can be observed as changes in the silencing of genes inserted into or near HC (position-effect variegation or PEV). Reporter genes placed in proximity to HC undergo stochastic silencing due to variable spreading of HC proteins, and become derepressed or further silenced upon disruption or augmentation of HC components, respectively. Genetic screens for PEV modifiers have identified ~150 genes that regulate HC function, but the identities of the majority of these genes and their molecular roles in HC structure and function are unknown (Colmenares, 2017).

Accumulating evidence shows that HC components are critical for genome integrity and play important roles in DNA repair. Loss of the Su(var)3-9 HMTase leads to chromosome segregation defects, repetitive DNA instability, and accumulation of DNA repair protein foci, including phosphorylated H2A variant (γH2Av). Su(var)3-9, HP1a, and the Smc5/6 complex also facilitate repair of heterochromatic double-strand breaks (DSBs) by the homologous recombination (HR) pathway, which utilizes a distinct and dynamic spatiotemporal regulatory mechanism. Early steps in HR repair, such as end resection, occur within minutes after DNA damage inside the HC domain. However, Rad51 recruitment, which is required to complete HR repair, only occurs after DSBs are relocalized outside the HC domain and associate with the nuclear periphery. This spatial partitioning of HR events enables the separation of repeats with DSBs from the rest of the HC, which likely reduces the probability of reciprocal exchange that results in genome instability (e.g., translocations that form acentric and dicentric chromosomes), and promotes less harmful HR repair from homologs or sister chromatids (Colmenares, 2017).

The Drosophila KDM4A (dKDM4A) protein belongs to the jumonji family of Fe(II)- and α-ketoglutarate-dependent lysine demethylases (Whetstine, 2006). Members of this family play vital roles in epigenetic mechanisms that govern gene expression and development, regulate DNA repair and genome stability, and are misregulated in many types of cancers. Despite its name, the closest dKDM4A mammalian homolog is KDM4D, since both contain the JmjN and JmjC domains responsible for enzymatic activity, and lack the PHD and Tudor domains found in human KDM4A. dKDM4A catalyzes the demethylation of H3K36me3 and H3K36me2 in vitro and in vivo (Lin, 2008; Crona, 2013), suggesting a transcriptional role, as these modifications are hallmarks of active gene bodies. Consequently, recent studies have focused on how dKDM4A regulates gene activity in EC (Colmenares, 2017).

dKDM4A homologs, as well as the closely related fly homolog KDM4B, have also been reported to demethylate H3K9me3. However, an in vitro study (Lin, 2008) concluded that dKDM4A lacks the H3 K9 demethylase activity associated with mammalian family members and dKDM4B; thus dKDM4A-dependent changes in H3K9me3 levels observed in vivo in flies are likely indirect or require additional cofactors (Lloret-Llinares, 2008; Tsurumi, 2013). KDM4A homologs have also been shown to demethylate other HC-associated modifications, including H1.4K26me3 and H3K56me3. However, Drosophila H1 lacks the lysine methylated in other species, and H3K56me3 has not been tested as a dKDM4A substrate in flies. Interestingly, dKDM4A contains a PxVxL motif that mediates its interaction with HP1a (Lin, 2008). Surprisingly, despite these links to HC, a potential role for dKDM4A in regulating HC structure or function has not been reported (Colmenares, 2017).

This study demonstrates that Drosophila KDM4A is highly enriched in HC and is required for normal HC structure and function, including repair of DNA damage. dKDM4A is show to affect transcription of some EC genes, but limited evidence was found for transcriptional regulation of heterochromatic genes and transposons. Instead, this study shows that dKDM4A is required for the spatial organization of repetitive elements, PEV, and the mobilization and repair of HC DSBs. It was further determined that dKDM4A contributes a structural, non-catalytic role in maintenance of HC integrity, as assayed by PEV. In contrast, dKDM4A catalytic activity was shown to be important for relocalization of DSBs outside the HC domain and for organismal survival of DNA repair mutants. Finally, DNA damage was found to trigger dKDM4A-dependent changes in the levels of heterochromatic H3K56me3. Altogether, it is concluded that dKDM4A is an HC component vital to the stability and organization of repetitive DNA and HC-mediated gene silencing, through a combination of structural and enzymatic functions (Colmenares, 2017).

These investigations show that Drosophila KDM4A is a structural component of HC, and regulates HC organization, PEV, and DNA repair. This study also identifies distinct dKDM4A functions in different nuclear domains (HC versus EC) and both structural and catalytic dKDM4A roles in HC, highlighting the significance of determining the enzymatic and non-enzymatic roles of dKDM4A homologs and their diversity in function and localization. dKDM4A is recruited to HC downstream of HP1a and Su(var)3-9, and is required for PEV in a non-catalytic manner. This suggests that dKDM4A contributes structurally to gene silencing and regulates the proper organization of HC complexes and sequences. dKDM4A is also required for relocalization and proper repair of heterochromatic DSBs. Intriguingly, normal HC DSB dynamics depend on dKDM4A catalytic activity and are associated with dKDM4A-dependent H3K56me3 demethylation, suggesting that this HC mark and its demethylated state(s) are important for DNA repair. Moreover, dKDM4A is required for organismal survival in the presence of mutations disrupting DNA repair and checkpoint pathways, further supporting a key role for this protein in maintaining genome stability (Colmenares, 2017).

The observation that dKDM4A is required for higher-order HC structure, specifically the organization of satellite DNAs, suggests that dKDM4A functions to maintain HC architecture. However, such defects in HC structure caused by loss of dKDM4A do not result from visible disruption of HP1a localization or dynamics, or from altered H3K36me3 levels in HC. Although changes in H3K9me3 levels after loss of dKDM4A were detected, these were not sufficient to alter transcription of the majority of HC elements. Whether dKDM4A directly or indirectly affects H3K9me3 levels in vivo is unclear. Regardless, this study shows that PEV does not require dKDM4A enzymatic activity, indicating that HC-mediated gene silencing is not directly regulated by dKDM4A demethylation of H3K9me3, or any other substrate. Although it is possible that the minor H3K9me3 perturbations are sufficient to disrupt HC, these results are more consistent with a direct structural role for dKDM4A in HC organization and gene silencing (Colmenares, 2017).

dKDM4A is directly recruited to HC by HP1a, suggesting that dKDM4A closely participates in HP1a-mediated HC functions. dKDM4A effects on variegation indicate that recruitment of factors downstream of HP1a assembly are required for full HC integrity and function. Although dKDM4A is not required for HP1a-mediated suppression of transposable element transcription, dKDM4A-dependent effects on satellite repeat organization were detected. This could suggest that suppression of PEV by dKDM4A mutants results from disruption of higher-order HC structure that increases accessibility of transcriptional machinery to HC domains. FISH studies have identified a host of regulatory proteins, including cohesins and condensins, that promote or antagonize pairing between HC domains. Whether dKDM4A participates in chromosome pairing, or influences cohesion or condensation in HC, remains to be determined. Alternatively, dKDM4A may assemble or link repetitive sequences into discrete domains within HC, which become unraveled and dispersed in the absence of dKDM4A, or dKDM4A may protect repetitive regions from being excised as extrachromosomal circles (Peng, 2007; Colmenares, 2017 and references therein).

A role in genome organization could also account for the impact of dKDM4A on HC DNA repair. Impaired repair of IR-induced DSBs in the absence of dKDM4A could result from defects in pairing of homologous chromosomes or sister chromatids, which would normally facilitate 'safe' HR among repeats, or from problems in folding or concatenation of HC domains that inhibit efficient exit of repair foci from HC. In fact, homozygous dKDM4A mutant adult flies are underrepresented compared with their heterozygote siblings; improper repair of spontaneous DNA breaks in HC could account for this subviability. This is consistent with observations in Caenorhabditis elegans, where loss of the dKDM4A homolog results in impaired DNA replication, accumulation of DNA damage, and increased apoptosis (Black, 2010; Colmenares, 2017 and references therein).

Support for this hypothesis comes from the synthetic lethality and infertility observed when a dKDM4A mutation is combined with mutations in the DNA damage checkpoint/DNA repair pathways, but not components of the mitotic checkpoint. Thus, spontaneous DNA damage requires dKDM4A to regulate HC structure for efficient repair. Su(var)3-9, but not dKDM4A, is also synthetically lethal with mitotic checkpoint mutants, which may reflect a higher level of spontaneous breaks that occur in the absence of Su(var)3-9, or defects in cohesin recruitment. This suggests that HC affects multiple pathways controlling genome stability, of which a subset is regulated by dKDM4A (Colmenares, 2017).

These experiments also identify a close relationship between dKDM4A and HP1a. HP1a directly recruits dKDM4A to HC, and dKDM4A overexpression can effectively sequester HP1a away from HC (Lin, 2008), which results in suppression of PEV. dKDM4A overexpression also potentially excludes other HP1a CSD-binding partners from HC, which could further exacerbate effects on HC functions. This study also identified EC genes that are co-regulated by dKDM4A and HP1a, indicating that even outside the HC domain these two proteins function together. This result contrasts with previously published data showing antagonistic effects of dKDM4A and HP1a on transcription (Crona, 2013), and may reflect the differences between immediate RNAi effects on tissue culture cells and steady-state effects that develop in mutant fly tissues (Colmenares, 2017).

Previously, HP1a binding was shown to enhance the H3K36me3 demethylase activity of dKDM4A (Lin, 2008). However, the current results suggest that dKDM4A does not exert PEV regulatory effects through demethylase activity and is not responsible for the low H3K36me3 levels in HC. Therefore, HP1a stimulation of dKDM4A H3K36me3 likely occurs in EC genes, although the possibility that local H3K36me3 levels increase at a few HC genes cannot be excluded. Instead it is proposed that the primary role of dKDM4A in HC is to ensure normal assembly of HP1a complexes that drives HC organization/structure and PEV (Colmenares, 2017).

A primarily structural role for dKDM4A in HC does not preclude the possibility that its catalytic activity regulates other HC functions. In fact, this study shows that dKDM4A enzymatic activity is required for DSBs to relocalize from HC, suggesting that a demethylated substrate facilitates HC repair dynamics and completion of DNA repair. Of three potential histone substrates tested, H3K56me3, which is enriched in HC and is demethylated by a dKDM4A homolog (KDM4D) in mammals, was identified as accumulating in the absence of dKDM4A, specifically after DNA damage induction by IR and in a Su(var)3-9-dependent manner. In mammals, KDM4D transiently localizes to DNA damage sites, but whether KDM4D demethylates H3K56me3 during DNA damage remains to be determined. Little is known about H3K56me3 function, but this residue resides at the junction between histone H3 and nucleosomal DNA and could regulate unfolding of DNA from the nucleosome. H3K56me3 demethylation also occurs during replication in mammalian cells, suggesting that the replication machinery requires removal of this mark for access to HC. Similarly, H3K56me3 demethylation by dKDM4A could also facilitate chromatin changes required for DSB relocalization and successful DNA repair (Colmenares, 2017).

Alternatively, the requirement for dKDM4A enzymatic activity in DSB relocalization may involve demethylation of other histone and non-histone proteins. Methylated peptides found in various non-histone chromatin proteins have been shown to be demethylated by a mammalian KDM4A homolog (Ponnaluri, 2009). Moreover, a fission yeast homolog lacking demethylase activity functions as an anti-silencing factor in HC, and has been proposed to act as a protein hydroxylase (Trewick, 2007). Therefore, it is proposed that dKDM4A recruitment to HC by HP1a regulates heterochromatic DSB repair and genome stability through demethylation of HC-specific mark(s), such as H3K56me3, but could also involve demethylation or hydroxylation of other HC components (Colmenares, 2017).

Many cancers acquire abnormal levels of HC components, which may increase genome instability and promote misregulation of oncogenes and tumor suppressors. Overexpression of several KDM4A family members has been shown to correlate with and drive tumor progression. Although it remains to be determined if human KDM4A family members regulate HC structure and promote HC DNA repair, disruption of such functions potentially contributes to tumorigenesis. The current findings therefore expand understanding of how this demethylase family exerts a myriad of effects in cancer tissue. Overexpression/ectopic expression of human KDM4A homologs have been shown to induce transient site-specific amplification of a cytogenetic region containing satellite DNA in tumors and cell lines (Black, 2013), antagonize 53BP1 binding to DSBs (Mallette, 2012), disrupt mismatch repair (Awwad, 2015), and produce chromosomal instability (Slee, 2012) and chromosome missegregation (Kupershmit, 2014). Further studies are required to determine whether high levels of KDM4A homologs promote tumor progression solely through altered transcriptional regulation of EC cancer-linked genes, or whether defects in HC structure and function also advance genomic evolution of cancers (Colmenares, 2017).

HP1a/KDM4A is involved in the autoregulatory loop of the oncogene gene c-Jun

The proto-oncogene c-Jun plays crucial roles in tumorigenesis, and its aberrant expression has been implicated in many cancers. Previous studies have shown that the c-Jun gene is positively autoregulated by its product. Notably, it has also been reported that c-Jun proteins are enriched in its gene body region. However, the role of c-Jun proteins in its gene body region has yet to be uncovered. HP1a is an evolutionarily conserved heterochromatin-associated protein, which plays an essential role in heterochromatin-mediated gene silencing. Interestingly, accumulating evidence shows that HP1a is also localized to euchromatic regions to positively regulate gene transcription. However, the underlying mechanism has not been defined. This study demonstrates that HP1a is involved in the positive autoregulatory loop of the Jra gene, the c-Jun homologue in Drosophila. Jra recruited the HP1a/KDM4A complex to its gene body region upon osmotic stress to reduce H3K36 methylation levels and disrupt H3K36 methylation-dependent histone deacetylation, resulting in high levels of histone acetylation in the Jra gene body region, thus promoting gene transcription. These results not only expand knowledge towards the mechanism of c-Jun regulation, but also reveal the mechanism by which HP1a exerts its positive regulatory function in gene expression (Liu, 2015).

In order to identify Jra interacting partners in Drosophila, the Jra complex was purified from S2 cells expressing FLAG-tagged Jra. Cells were first treated with or without 500 mM sorbitol for 30 minutes, which induces osmotic stress and activates the MAPK signaling pathway and Jra. The nuclear extracts prepared from these cells were subjected to complex purification using anti-FLAG antibody. The mass spectrometry data showed that, among other Jra interacting partners, heterochromatin protein HP1a co-purifies with Jra-FLAG under osmotic stress. The interaction between Jra-FLAG and HP1a was further confirmed by Western blot analysis. Interestingly, Jra only co-immunoprecipitates with HP1a under osmotic stress, but not under unstressed conditions (Liu, 2015).

To further determine whether HP1a interacts with endogenous Jra, nuclear extracts from S2 cells treated with or without osmotic stress were subjected to co-immunoprecipitation assay using anti-Jra antibody. The results confirmed that endogenous Jra co-immunoprecipitates with HP1a under osmotic stress. Taken together, these data demonstrate that HP1a interacts with Jra under osmotic stress (Liu, 2015).

Jra has been previously shown to bind to its gene body region. Having found that Jra interacts with HP1a, it was reasoned that Jra might recruit HP1a to its gene body region. To answer this question, a chromatin immunoprecipitation (ChIP) assay was performed to determine whether HP1a binds to the Jra gene locus. HP1a did not localized to either the promoter or the gene body region of Jra under unstressed conditions. However, under osmotic stress, HP1a is enriched in the gene body region of Jra, but not in the promoter region. Western blot assay confirmed that Jra is phosphorylated under osmotic stress. Phosphorylation is an important post-translational modification for c-Jun regulation, which has previously been shown to regulate its transcriptional activity and its protein stability. The ChIP result combined with co-immunoprecipitation data indicate that phosphorylated Jra, but not unphosphorylated Jra, interacts with HP1a (Liu, 2015).

Next, attempts were made to determine if HP1a binding to the Jra gene body region is dependent on Jra phosphorylation. To address this concern, S2 cells were treated with JNK dsRNA to knock down JNK expression. JNK is the protein kinase in MAPK signaling pathway, which is responsible for c-Jun phosphorylation. S2 cells were also treated with GFP dsRNA as a control. Western blot result confirmed that the expression levels of JNK were dramatically reduced in in JNK dsRNA treated cells, and that the phosphorylation of Jra was greatly reduced in JNK-depleted cells under osmotic stress. These cells were then analyzed by ChIP assay using anti-HP1a antibody. The data showed that upon the depletion of JNK, HP1a lost its binding to the Jra gene body under osmotic stress, suggesting that phosphorylated Jra recruits HP1a to its gene body region (Liu, 2015).

In the heterochromatin region, HP1a binding is dependent on H3K9 methylation. Therefore, whether H3K9 methylation also plays a role in HP1a recruitment to the Jra gene body was examined. Cells treated with or without osmotic stress were subjected to ChIP assay using anti-H3K9me2 antibody. The data showed that there was no significant change in H3K9me2 levels in the Jra gene body region, eliminating the regulatory role of H3K9 methylation in the recruitment of HP1a to the Jra gene body region. Taken together, these results showed that Jra recruits HP1a to its gene body region under osmotic stress, which is independent of H3K9 methylation (Liu, 2015).

Since an increased enrichment of HP1a was observed in the Jra gene body region under osmotic stress, attempts were made to examine the effect of HP1a on Jra gene expression. S2 cells were treated with HP1a dsRNA to knock down HP1a expression. S2 cells treated with GFP dsRNA were used as a control. The depletion of HP1a was confirmed by Western blot assay. The total RNA extracted from these cells was then analyzed by qRT-PCR. The results showed that HP1a knockdown significantly reduced Jra mRNA levels, indicating HP1a is positively involved in Jra transcription (Liu, 2015).

Having defined a role for HP1a in positively regulating Jra transcription, the mechanism by which this occurs was explored. A previous study reported that HP1a interacts with Drosophila KDM4A demethylase and stimulates its H3K36 demethylation activity (Lin, 2008). Since the enrichment of HP1a in the Jra gene body region was observed, and H3K36me3 modification has been reported to be enriched in the gene body region, it was asked whether HP1a collaborates with KDM4A to reduce H3K36me3 levels in the Jra gene body region. To test this idea, the H3K36me3 levels were measured of the Jra gene body region upon HP1a depletion. The results showed that HP1a depletion significantly elevated H3K36me3 levels in the Jra gene body region, indicating a potential involvement of KDM4A in Jra transcription. To verify the recruitment of KDM4A to the Jra gene body, an S2 cell line was established stably expressing KDM4A-FLAG. These cells were treated with or without osmotic stress and subjected to ChIP analysis. The results demonstrate that KDM4A is enriched in Jra gene body region upon osmotic stress, and the depletion of HP1a abolishes its binding to the Jra gene body region (Liu, 2015).

H3K36me3 has been shown to be essential in preventing cryptic transcription from the gene body region. Following RNAP II elongation, histone acetylation marks are removed by the Rpd3S complex, which is directed by H3K36me3 in the gene body region. Because elevated levels of H3K36me3 upon HP1a depletion were observed, the histone acetylation levels were examined in this region. As expected, the overall histone acetylation levels were significantly reduced upon HP1a depletion (Liu, 2015).

It should be noted that a previous study has demonstrated that HP1a positively regulates euchromatic gene expression through its involvement in RNA packaging and stability. To determine if the reduced mRNA levels of Jra was due the decreased mRNA stability in HP1a-depleted cells, cells were first treated with 500mM sorbitol and then with actinomycin D for 30 minutes, 60 minutes, and 90 minutes. Total RNA was subsequently isolated for quantitative real-time PCR analysis. The result showed that in both wild-type S2 cells and HP1a-depleted S2 cells, Jra mRNA levels were slightly reduced. However, HP1a depletion did not significantly accelerate Jra mRNA turnover after actinomycin D treatment. Therefore, it is proposed that the reduced Jra expression that was observed is primarily due to the disruption of the positive role of HP1a in Jra transcription (Liu, 2015).

Although many studies reported that HP1a is not only localized to heterochromatin, but also localized to euchromatic regions to regulate gene transcription, the mechanism by which HP1a facilitates gene transcription is still unclear. It has also been established that c-Jun gene is under the control of a positive autoregulatory loop. In addition, c-Jun proteins have also been shown to bind to its own gene body region. However, what c-Jun does in its own gene body region has yet to be uncovered. This study has presented evidence showing that HP1a is involved in the positive autoregulatory loop of the Jra gene. Based on the data, a model is proposed in which Jra recruits the HP1a/KDM4A complex to its gene body region upon osmotic stress to reduce H3K36 methylation levels and disrupt H3K36 methylation-dependent histone deacetylation, resulting in high levels of histone acetylation in the Jra gene body region, and thus promotes gene transcription. These results not only expand knowledge of the mechanism by which the c-Jun oncogene is regulated, but also reveal the mechanism by which HP1a exerts its positive regulatory function in gene expression (Liu, 2015).

dRYBP counteracts chromatin-dependent activation and repression of transcription

Chromatin dependent activation and repression of transcription is regulated by the histone modifying enzymatic activities of the trithorax (trxG) and Polycomb (PcG) proteins. To investigate the mechanisms underlying their mutual antagonistic activities this study analyzed the function of Drosophila Ring and YY1 Binding Protein (dRYBP), a conserved PcG- and trxG-associated protein. dRYBP is ubiquitylated and binds ubiquitylated proteins. Additionally dRYBP was shown to maintain H2A monoubiquitylation, H3K4 monomethylation and H3K36 dimethylation levels and does not affect H3K27 trimethylation levels. Further it was shown that dRYBP interacts with the repressive SCE (Ring) and dKDM2 (Lysine (K)-specific demethylase 2) proteins as well as the activating dBRE1 protein. Analysis of homeotic phenotypes and post-translationally modified histones levels show that dRYBP antagonizes dKDM2 and dBRE1 functions by respectively preventing H3K36me2 demethylation and H2B monoubiquitylation. Interestingly, the results show that inactivation of dBRE1 produces trithorax-like related homeotic transformations, suggesting that dBRE1 functions in the regulation of homeotic genes expression. These findings indicate that dRYBP regulates morphogenesis by counteracting transcriptional repression and activation. Thus, they suggest that dRYBP may participate in the epigenetic plasticity important during normal and pathological development (Fereres, 2014).

Gene regulation by the lysine demethylase KDM4A in Drosophila

Lysine methylation of histones is associated with both transcriptionally active chromatin and with silent chromatin, depending on what residue is modified. Histone methyltransferases and demethylases ensure that histone methylations are dynamic and can vary depending on cell cycle- or developmental stage. KDM4A demethylates H3K36me3, a modification enriched in the 3' end of active genes. The genomic targets and the role of KDM4 proteins in development remain largely unknown. KDM4A mutant Drosophila were generated, and 99 mis-regulated genes were identified in first instar larvae. Around half of these genes were down-regulated and the other half up-regulated in dKDM4A mutants. Although heterochromatin protein 1a (HP1a) can stimulate dKDM4A demethylase activity in vitro, it was found that they antagonize each other in control of dKDM4A-regulated genes. Appropriate expression levels for some dKDM4A-regulated genes rely on the demethylase activity of dKDM4A, whereas others do not. Surprisingly, although highly expressed, many demethylase-dependent and independent genes are devoid of H3K36me3 in wild-type as well as in dKDM4A mutant larvae, suggesting that some of the most strongly affected genes in dKDM4A mutant animals are not regulated by H3K36 methylation. By contrast, dKDM4A over-expression results in a global decrease in H3K36me3 levels and male lethality, which might be caused by impaired dosage compensation. These results show that a modest increase in global H3K36me3 levels is compatible with viability, fertility, and the expression of most genes, whereas decreased H3K36me3 levels are detrimental in males (Crona, 2013).

Methylation of H3K36 is primarily thought of as a histone modification that is linked to transcriptional activation. H3K36me3 is enriched at the 3' end of active genes, where it is deposited by the Set2 methyltransferase travelling with RNA polymerase II. However, recent results have shown that not all active genes contain H3K36me3. In Drosophila cells, YELLOW and RED are two types of transcriptionally active chromatin that differ in H3K36me3 and MRG15 binding (Filion, 2010). RED chromatin lacks H3K36me3 and is associated with tissue-specific gene expression, whereas YELLOW chromatin is enriched in H3K36me3 and associated with broadly expressed genes. This study compared genome-wide H3K36me3 ChIP data from embryos and third instar larvae generated by the modENCODE project, with the dKDM4A-regulated genes in first instar larvae. Surprisingly, most of the dKDM4A-regulated genes completely lacked H3K36me3. ChIP experiments at dKDM4A-regulated genes confirmed that H3K36me3 levels are very low in both wild-type and P/ΔdKDM4A mutant first instar larvae. The lack of H3K36me3 is not due to low expression of the genes, since many dKDM4A-regulated genes are highly expressed. Instead, this suggests that many dKDM4A-regulated genes are not controlled by H3K36me3 levels, but by other dKDM4A-dependent functions (Crona, 2013).

Some of the dKDM4A-regulated genes are located close to one another in the genome. Most strikingly, five down-regulated genes are present in a 67 kb region on the X-chromosome. Three down-regulated genes were also found in an 8 kb region on chromosome 3. In seven cases two dKDM4A target genes are close to one another. Perhaps these genes require a common chromatin environment for their regulation. However, since many of these loci lack H3K36me3, it is unlikely that dKDM4A directly influences the chromatin structure at these genes. Instead, they might be controlled by dKDM4A-dependent demethylation of a non-histone protein, or by a common regulator whose expression in turn is influenced by dKDM4A (Crona, 2013).

To investigate the requirement for dKDM4A demethylase activity in gene regulation, P/ΔdKDM4A mutants were rescued with wild-type or H195A catalytically inactive transgenes. The results show that the down-regulated genes examined and a few of the up-regulated genes (such as Ppcs) do not require dKDM4A demethylase activity for their expression. By contrast, some of the up-regulated genes depend on dKDM4A catalytic activity for proper expression. However, none of the dKDM4-regulated genes examined contain significantly more H3K36me3 in P/ΔdKDM4A mutant larvae. This indicates that both demethylase activity-dependent and independent genes are regulated by dKDM4A by mechanisms that do not involve H3K36me3 demethylation, or alternatively, that these genes are indirect dKDM4A-targets. Although so far few proteins except histones are known substrates for demethylases, it is likely that many proteins are targets of KDM enzymes, in the same way as lysine acetylases and deacetylases target various proteins. Using a peptide demethylation assay, non-histone substrates have been identified in vitro for human KDM4A (Ponnaluri, 2009; Crona, 2013 and references therein).

Another possibility is that other lysine residues in the histones are being targeted by dKDM4A. One candidate is H3K9me3, since other KDM4 family members can demethylate this modification. However, no detectable amounts of H3K9me3 were found in either wild-type nor in dKDM4A mutant larvae at dKDM4A-regulated genes. Thus, changes in H3K9me3 cannot explain the gene expression phenotypes in P/ΔdKDM4A mutants. Mammalian KDM4 family members can also demethylate H1.4K26me3. Due to lack of antibodies against this histone modification, it was not possible to investigate whether this form of the linker histone is present at dKDM4A-regulated genes. In summary, it cannot be excluded that demethylation of histone H1 is important for the gene expression changes in P/ΔdKDM4A mutants, but it seems as if histone H3 demethylation by dKDM4A is not required for the observed gene expression phenotypes (Crona, 2013).

Regulators of gene expression were sought among the 99 mis-regulated genes identified in the expression array, since they might mediate some of the gene expression changes observed in P/ΔdKDM4A mutants. Mlp60A is a LIM-domain protein of the CRP class. It is a nuclear protein that may function as an adaptor protein, and thereby influence gene expression. PPP4R2r is a regulatory subunit of protein phosphatase 4 (PP4). The PP4 complex that includes this regulatory subunit has been implicated in centrosome maturation, spliceosomal assembly, DNA repair, and in asymmetric localization of the Miranda protein during Drosophila neuroblast division. It is possible that mis-regulated expression of these or other gene products contribute to the altered expression of other genes in P/ΔdKDM4A mutant larvae (Crona, 2013).

Yet another possibility is that dKDM4A is a direct regulator of some genes, but that its catalytic function is not required. Protein-protein interactions may mediate the gene regulatory properties of dKDM4A on such genes. HP1a is one interaction partner of dKDM4A, although this study shows that these proteins antagonize each other at the dKDM4a-regulated genes examined. Instead, other dKDM4A-binding proteins may mediate its function in gene control. In mammalian cells, interactions of KDM4A with the Retinoblastoma protein, the androgen receptor, and the N-CoR co-repressor complex influences transcription. The hypothesis is favored that dKDM4A is directly regulating at least some of the mis-expressed genes, and that it does so by protein-protein interactions or by demethylating a non-histone substrate (Crona, 2013).

Although HP1a is found mainly in pericentric heterochromatin through its interaction with H3K9me2/3, recent results implicate HP1a also in euchromatic gene regulation. Binding of dKDM4A to HP1a stimulates the H3K36me3-demethylase activity of dKDM4A in vitro (Lin, 2008). Therefore up-regulated dKDM4A-targets that depend on dKDM4A demethylase activity for normal expression levels (such as Mlp60A and CG9059) were expected to be up-regulated also in HP1a mutants but to a lesser extent than in dKDM4A mutants. However, these genes are expressed at higher levels in HP1a mutants than in dKDM4A mutant larvae. This indicates that HP1a contributes to regulation of these genes in more ways than stimulation of dKDM4A-demethylase activity. Furthermore, targets that are down-regulated in dKDM4A mutants are up-regulated in HP1a mutants, and expression restored in the HP1a dKDM4A double mutant. Together, this indicates that the activity that influences dKDM4A-target gene expression is not stimulation of dKDM4A-mediated demethylation by HP1a, but rather an effect of dKDM4A on HP1a function or distribution. Indeed, over-expression of dKDM4A has previously been shown to alter the distribution of HP1a from the chromocenter into the chromosome arms on polytene chromosomes (Lloret-Llinares, 2008). Moreover, a recent report shows that HP1a targets dKDM4A to heterochromatic genes, but that the demethylase activity of dKDM4A in euchromatin is independent of HP1a targeting (Lin, 2012). These results are consistent with observations that HP1a does not stimulate dKDM4A's ability to regulate target gene expression in euchromatin (Crona, 2013).

Small changes were observed in global H3K36me3 levels in the P/ΔdKDM4A mutant. This indicates that other demethylases can substitute for dKDM4A in its absence. The most likely candidate is the dKDM4A ortholog dKDM4B, which can also demethylate H3K36me3 in vitro (Lin , 2008). This may explain why dKDM4A mutants are viable and affect expression of only a small number of genes. A similar case has been described for the Drosophila H3K4me3 demethylase Lid, whose demethylase function is not essential and probably compensated for by dKDM2. Thus, redundancy between histone demethylases may be common in Drosophila. Interestingly, life span is affected in Drosophila males but not in females in both dKDM4A and lid mutants. Normal life-span requires catalytically active Lid. Perhaps histone demethylation becomes critical for processes required specifically during male adulthood. Given the apparent redundancy between histone demethylases, it is not surprising that P/ΔdKDM4A mutants are viable. This study therefore examined survival of flies ubiquitously over-expressing dKDM4A, which results in diminished H3K36me3 levels. At 25°C, a marked reduction was found in male survival upon dKDM4A over-expression, but a much smaller effect on females. At 29 °C where the Gal4 protein is more active and thereby causes higher over-expression, no males survived dKDM4A expression. This may be due to impaired dosage compensation, since binding of the male-specific lethal (MSL) dosage compensation complex to the male X-chromosome in part depends on H3K36me3. Consistent with this interpretation, diminished expression of one dosage compensated gene was observed in males surviving dKDM4A over-expression. It is possible that even larger effects on dosage compensated genes would be observed in the males that are killed by increased dKDM4A expression. This indicates that although H3K36me3 levels may influence the ability of males to dosage compensate their single X-chromosome, it is not absolutely essential for development and survival in females (Crona, 2013).

These results show that a modest increase in global H3K36me3 levels is compatible with viability, fertility, and the expression of most genes, whereas decreased H3K36me3 levels are detrimental, particularly in males. Unexpectedly, many dKDM4A-regulated genes completely lack or contain low levels H3K36me3, indicating that dKDM4A may have other substrates that are more important for gene control. A more complete understanding of KDM4 proteins in development and the control of gene expression requires the identification of direct target genes (Crona, 2013).

Drosophila Kdm4 demethylases in histone H3 lysine 9 demethylation and ecdysteroid signaling

The dynamic regulation of chromatin structure by histone post-translational modification is an essential regulatory mechanism that controls global gene transcription. The Kdm4 family of H3K9me2,3 and H3K36me2,3 (dual specific histone demethylases) has been implicated in development and tumorigenesis. This study shows that Drosophila Kdm4A and Kdm4B, both members of the JHDM3 histone demethylase family are together essential for mediating ecdysteroid hormone signaling during larval development. Loss of Kdm4 genes leads to globally elevated levels of the heterochromatin marker H3K9me2,3 and impedes transcriptional activation of ecdysone response genes, resulting in developmental arrest. It was further shown that Kdm4A interacts with the Ecdysone Receptor (EcR) and colocalizes with EcR at its target gene promoter. These studies suggest that Kdm4A may function as a transcriptional co-activator by removing the repressive histone mark H3K9me2,3 from cognate promoters (Tsurumi, 2013).

This study has discovered a role for Kdm4 in the transcriptional regulation of a subset of ecdysone pathway components. Furthermore, an interaction was demonstrated between Kdm4A and EcR in vivo, providing evidence that Kdm4 demethylases may act as co-activators of EcR. A genetic approach has allowed facilitated the detection of a previously uncharacterized, but essential, role of Kdm4 in development, and has identified a direct Kdm4 target gene in euchromatin. Interestingly, Human Kdm4 members interact with the nuclear hormone receptors, Androgen Receptor (AR) and Estrogen Receptor α (ERα), and these receptors have been proposed to serve as co-activators, suggesting a molecular mechanism by which Kdm4 can act as an oncogene in prostate and breast cancers. Kdm4B was shown to be a direct target gene of ERα, yielding a feed-forward loop for an augmented hormonal response. The results indicate that a similar epigenetic mechanism exists in Drosophila, where a nuclear hormone receptor requires the Kdm4 family of demethylases to remove H3K9 methylation at the promoter of a target gene. Taken together, the Kdm4 family of demethylases may function as transcriptional co-factors required for transcriptional activation by nuclear hormone receptors (Tsurumi, 2013).

Previous studies have shown that the Trithorax-related (Trr) H3K4 methyltransferase, the Nurf nucleosome remodeling complex component, Nurf301, the Brahma (Brm)-containing chromatin remodeler, and the histone acetyltransferase CREB-binding protein (CBP) are also co-activators of EcR, indicating that activation of ecdysone pathway genes requires substantial regulation of the chromatin environment. Since H3K4 hyper-methylation at promoters is a marker of active transcription, and since H3K9 hypo-methylation also promotes upregulation of gene expression, it is feasible that synchronizing these two events would lead to more robust target gene activation. The mammalian Kdm4B (JMJD2B) forms a complex with the mixed-lineage leukemia (MLL) 2 H3K4 methyltransferase and serves as a co-activator of Estrogen Receptor. The complex couples H3K9 demethylation with H3K4 methylation in order to facilitate ERα target gene activation. Similar functional cross-talk between H3K9 demethylation and H3K4 methylation has been described in S. pombe, where the Lsd1 H3K9 demethylase and the Set1 H3K4 methyltransferase were found in a complex. Since, in Drosophila, the Nurf301 subunit, Brm and CBP were also found to interact with EcR, nucleosome remodeling may cooperate as well in the rapid and dynamic activation of ecdysone regulated genes (Tsurumi, 2013).

These studies of the Kdm4A and Kdm4B homozygous double mutants demonstrate a requirement for these genes in the ecdysone pathway. This observation is similar to results obtained with mutant alleles of Nurf301 and trr, two seemingly ubiquitous chromatin regulators, where specific downregulation of ecdysone signaling genes has been detected. Additionally, this study is consistent with the reports that adult male Kdm4A mutants display abnormal courtship behavior and concomitant downregulated fru, a gene speculated to be a direct downstream target of EcR. The specific defects in ecdysone signaling, rather than general transcription, exhibited by the double mutants indicate that either Kdm4 may not be essential for regulating all genes, or that the aberrant expression of ecdysone responsive genes is the earliest manifestation of loss of Kdm4. However, this study does not rule out the possibility that Kdm4 proteins regulate other crucial transcription factors that in turn regulate ecdysone pathway components by secondary effects. Further molecular and genomic studies are required to resolve this issue (Tsurumi, 2013).

H3K9 demethylation-dependent transcriptional activation of BR-C was demonstrated. It is possible however, that H3K36 demethylation also contributes to ecdysone pathway component regulation. Previous studies have shown that HP1a is recruited to developmental puffs in polytene chromosomes and that it stimulates H3K36 demethylation by Kdm4A. Perhaps H3K36 demethylation in the gene body and subsequent displacement of the HDAC complex is important for transcriptional elongation or for the activation of downstream nested promoters of ecdysone pathway components. Moreover, H3K36 plays a role in exon splice choice and thus ecdysone pathway genes that produce multiple splice variants may require Kdm4 regulation. However, immunostaining experiments show that HP1a and Kdm4A signals are mostly non-overlapping. Thus, it seems that HP1a's involvement in the demethylase activities of Kdm4 toward H3K9 or H3K36 would have to be transient and dynamic (Tsurumi, 2013).

In summary, this study has shown that double homozygous mutants of the two Kdm4 genes in Drosophila display developmental delays and lethality, with compromised activation of ecdysone related genes. Furthermore, it was found that BR-C may be a direct target of H3K9 demethylation, and that the interaction between Kdm4A and EcR may be important in transcriptional activation of BR-C. These results provide insight into the physiological functions and mechanistic roles of Kdm4 in vivo. The interaction between Kdm4 and EcR awaits further investigation. It is conceivable that EcR directs the recruitment of Kdm4A to the promoter of its target genes, or alternatively, that EcR allosterically regulates the demethylase activity of Kdm4A, allowing removal of H3K9m2,3 only upon hormone signaling (Tsurumi, 2013).

HP1a targets the Drosophila KDM4A demethylase to a subset of heterochromatic genes to regulate H3K36me3 levels

Recent discoveries of histone demethylases demonstrate that histone methylation is reversible. However, mechanisms governing the targeting and regulation of histone demethylation remain elusive. A Drosophila melanogaster JmjC domain-containing protein, dKDM4A (Histone demethylase 4A), is a histone H3K36 demethylase. dKDM4A specifically demethylates H3K36me2 and me3 both in vitro and in vivo. Affinity purification and mass spectrometry analysis revealed that Heterochromatin Protein 1a (HP1a) associates with dKDMA4A. The chromoshadow domain of HP1a and a HP1-interacting motif of dKDM4A are responsible for this interaction. HP1a stimulates the histone H3K36 demethylation activity of dKDM4A and this stimulation depends on the H3K9me binding motif of HP1a. Finally, in vivo evidence is provided suggesting that HP1a and dKDM4A interact with each other and loss of HP1a leads to increased level of histone H3K36me3. Collectively, these results suggest a function of HP1a in transcription facilitating H3K36 demethylation at transcribed and/or heterochromatin regions (Lin, 2012).

This study has identified one of the JmjC domain-containing KDM4 orthologs in Drosophila, dKDM4A. The in vitro demethylation assay shows that dKDM4A demethylates histone H3K36me3 and me2 using an oxidative demethylation mechanism which requires Fe (II) and α-ketoglutarate as cofactors. Overexpression of dKDM4A in Drosophila S2 cells reduces the level of histone H3K36me3, whereas knockdown of endogenous dKDM4A increases the level of histone H3K36me3 and me2. These results together demonstrate that dKDM4A is a bona fide histone H3K36 demethylase in vivo (Lin, 2012).

Through multidimensional protein identification technology (MudPIT) analysis of the affinity-purified native dKDM4A complex, it was found that HP1a associates with dKDM4A. More importantly, it was demonstrated that HP1a stimulates the demethylation activity of dKDM4A, while the HP1a CD mutant V26M, that cannot bind methyl-K9 histone H3, fails to stimulate dKDM4 activity. In addition, dKDM4A directly binds to the HP1 CSD and this binding requires an intact HP1 CSD dimer interface. A consensus HP1 binding PxVxL motif of dKDM4A is responsible for its interaction with CSD of HP1a. Interestingly, overexpression of dKDM4A causes the spread of HP1a to euchromatin regions, presumably through this specific interaction, and dKDM4A-V423A, which does not bind to HP1a, failed to localize HP1a to euchromatin. These data suggest HP1a-dKDM4A is a euchromatic H3K36 demethylase complex (Lin, 2012).

Set2 mediated histone H3K36 methylation is an important mark on chromatin during transcription elongation . In fungi, such as S. cerevisiae, S. pombe, and N. crassa, a sole histone lysine-methyltransferase Set2 is responsible for all three methylation states of H3K36. In Drosophila, histone H3K36 methylation is catalyzed by two enzymes, dSet2 and dMes-4. Although yeast Set2 is the only histone methyltransferase that catalyzes methylation of histone H3K36, two histone H3K36 demethylases, Jhd1 and Rph1, are responsible for demethylation of histone H3K36 at different modification states in budding yeast. In Drosophila, there are three histone demethylases that govern demethylation of histone H3K36. dKDM2 has been identified as a histone H3K36me2 demethylase (Lagarou, 2008). This study demonstrates that dKDM4A is a histone H3K36me3 and me2 demethylase, and dKDM4B has demethylation activity on both histone H3K9 and K36me3/me2 in vitro. Therefore, histone H3K36 methylation in Drosophila is likely regulated by highly specific enzymes in both directions. Since both modification and de-modification enzymes possess high modification state specificity, histone H3K36 may be subjected to much more sophisticated regulation in higher eukaryotes than in yeast (Lin, 2012).

Purification of the dKDM4A complex from S2 cells revealed a specific association of HP1a with dKDM4A. Three of the HP1-like chromatin proteins (HP1a, HP1b, HP1c) in Drosophila share high amino acid sequence similarity. Both HP1a and HP1b localize to euchromatin and heterochromatin, while HP1c is found only in euchromatin. It is unclear whether these HP1-like chromatin proteins have specific or redundant functions in transcription regulation. However, this study demonstrates that dKDM4A specifically interacts with HP1a, but not HP1b and HP1c. Furthermore, HP1b and HP1c cannot stimulate dKDM4A demethylation activity in vitro. A previous study showed that the yeast homolog of KDM4, Rph1 (ScKDM4), did not stably associate with any other protein. It was speculated that the C-terminal ZF domain of Rph1, which can potentially bind to DNA, allows Rph1 to function without associated factors. Unlike other proteins in the KDM4 family which commonly contain PHD, tudor or ZF domains, dKDM4A only has JmjN and JmjC domains. This study found that HP1a stably associates with dKDM4A and stimulates its demethylation activity. Since the H3K9 binding motif is required for this stimulation, it is proposed that the CD of HP1a might contribute to target dKDM4A to specific loci, particularly to H3K9me enriched regions, to regulate gene expression (Lin, 2012).

In S. pombe, the HP1 homolog recruits a JmjC domain-containing protein Epe1 to heterochromatin loci where they function together to counteract repressive chromatin. This study shows that HP1a directly interacts with dKDM4A through a consensus binding motif PxVxL. Most importantly, the presence of HP1a stimulates histone demethylation activity of dKDM4A in vitro, and HP1a is required for maintaining normal level of H3K36me3 in vivo as well. Since Epe1 on its own seems to have no histone demethylation activity, it would be interesting to see whether a similar scenario also occurs in S. pombe, in which Swi6 may stimulate enzymatic activity of Epe1 towards other non-histone substrates (Lin, 2012).

HP1 has been reported to associate with actively transcribed euchromatin regions. Mammalian HP1γ and histone H3K9 methylation are enriched at the coding region of active genes, implying that they may play a role during transcription elongation. In yeast, histone H3K36me3 appears to be a repressive mark at coding region of actively transcribed genes. In higher eukaryotes, histone H3K9 methylation, which is absent in the budding yeast, might replace the role of K36 methylation in the coding regions of transcribed genes. However, the mechanism by which HP1 functions in active transcription is largely unknown. The current findings suggest a possible role of HP1a in recruitment of the histone H3K36me3/me2 demethylase dKDM4A to transcribed regions to remove histone H3K36 methylation. The formation of the HP1a-dKDM4A complex may help to release HP1a from heterochromatin regions, thus targeting it to specific gene loci. It is also possible that dKDM4A, which targets histone modification marks within the 3' ORF of actively transcribed genes, recruits HP1a to euchromatic regions. A model is favored in which HP1a facilitates recruitment of dKDM4A, because the HP1a CD mutant, V26M, fails to stimulate dKDM4A activity. This result suggests that HP1a binding to histone H3 is required for the enhancement of dKDM4A demethylation activity. It is speculated that HP1a-mediated histone demethylation may serve as a regulatory mechanism to control chromatin states during active transcription elongation. Alternatively, a similar mechanism might also apply to maintaining silenced states of heterochromatin (Lin, 2012).

Heterochromatin protein 1a stimulates histone H3 lysine 36 demethylation by the Drosophila KDM4A demethylase

Recent discoveries of histone demethylases demonstrate that histone methylation is reversible. However, mechanisms governing the targeting and regulation of histone demethylation remain elusive. This study reports that a Drosophila melanogaster JmjC domain-containing protein, dKDM4A, is a histone H3K36 demethylase. dKDM4A specifically demethylates H3K36me2 and H3K36me3 both in vitro and in vivo. Affinity purification and mass spectrometry analysis revealed that heterochromatin protein 1a (HP1a) associates with dKDMA4A. The chromo shadow domain of HP1a and a HP1-interacting motif of dKDM4A are responsible for this interaction. HP1a stimulates the histone H3K36 demethylation activity of dKDM4A, and this stimulation depends on the H3K9me-binding motif of HP1a. Finally, in vivo evidence is provided suggesting that HP1a and dKDM4A interact with each other and that loss of HP1a leads to an increased level of histone H3K36me3. Collectively, these results suggest a function of HP1a in transcription facilitating H3K36 demethylation at transcribed and/or heterochromatin regions (Li, 2008).

Functions of Kdm4a orthologs in other species

KDM4A coactivates E2F1 to regulate the PDK-dependent metabolic switch between mitochondrial oxidation and glycolysis

The histone lysine demethylase KDM4A/JMJD2A has been implicated in prostate carcinogenesis through its role in transcriptional regulation. This study describes KDM4A as a E2F1 (see Drosophila E2F1) coactivator and demonstrate a functional role for the E2F1-KDM4A complex in the control of tumor metabolism. KDM4A associates with E2F1 on target gene promoters and enhances E2F1 chromatin binding and transcriptional activity, thereby modulating the transcriptional profile essential for cancer cell proliferation and survival. The pyruvate dehydrogenase kinases (PDKs; see Drosophila Pdk) PDK1 and PDK3 are direct targets of KDM4A and E2F1 and modulate the switch between glycolytic metabolism and mitochondrial oxidation. Downregulation of KDM4A leads to elevated activity of pyruvate dehydrogenase and mitochondrial oxidation, resulting in excessive accumulation of reactive oxygen species. The altered metabolic phenotypes can be partially rescued by ectopic expression of PDK1 and PDK3, indicating a KDM4A-dependent tumor metabolic regulation via PDK. These results suggest that KDM4A is a key regulator of tumor metabolism and a potential therapeutic target for prostate cancer (Wang, 2016).

Overexpression of KDM4 lysine demethylases disrupts the integrity of the DNA mismatch repair pathway

The KDM4 family of lysine demethylases consists of five members, KDM4A, -B and -C that demethylate H3K9me2/3 and H3K36me2/3 marks, while KDM4D and -E demethylate only H3K9me2/3. Recent studies implicated KDM4 proteins in regulating genomic instability and carcinogenesis. This study describes a previously unrecognized pathway by which hyperactivity of KDM4 demethylases promotes genomic instability. Overexpression of KDM4A-C, but not KDM4D, disrupts MSH6 foci formation during S phase by demethylating its binding site, H3K36me3. Consequently, it was demonstrated that cells overexpressing KDM4 members are defective in DNA mismatch repair (MMR), as evident by the instability of four microsatellite markers and the remarkable increase in the spontaneous mutations frequency at the HPRT locus. Furthermore, it was shown that the defective MMR in cells overexpressing KDM4C is mainly due to the increase in its demethylase activity and can be mended by KDM4C downregulation. Altogether, these data suggest that cells overexpressing KDM4A-C are defective in DNA MMR and this may contribute to genomic instability and tumorigenesis (Awwad, 2015).

JMJD2A promotes cellular transformation by blocking cellular senescence through transcriptional repression of the tumor suppressor CHD5

Senescence is a cellular response preventing tumorigenesis. The Ras oncogene is frequently activated or mutated in human cancers, but Ras activation is insufficient to transform primary cells. In a search for cooperating oncogenes, this study identified the lysine demethylase JMJD2A/KDM4A. JMJD2A functions as a negative regulator of Ras-induced senescence and collaborates with oncogenic Ras to promote cellular transformation by negatively regulating the p53 pathway. CHD5, a known tumor suppressor regulating p53 activity, as a target of JMJD2A. The expression of JMJD2A inhibits Ras-mediated CHD5 induction leading to a reduced activity of the p53 pathway. In addition, JMJD2A is overexpressed in mouse and human lung cancers. Depletion of JMJD2A in the human lung cancer cell line A549 bearing an activated K-Ras allele triggers senescence. It is proposed that JMJD2A is an oncogene that represents a target for Ras-expressing tumors (Mallette, 2012).

Conserved antagonism between JMJD2A/KDM4A and HP1gamma during cell cycle progression

The KDM4/JMJD2 family of histone demethylases is amplified in human cancers. However, little is known about their physiologic or tumorigenic roles. This study has identified a conserved and unappreciated role for the JMJD2A/KDM4A H3K9/36 tridemethylase in cell cycle progression. It was demonstrated that JMJD2A protein levels are regulated in a cell cycle-dependent manner and that JMJD2A overexpression increased chromatin accessibility, S phase progression, and altered replication timing of specific genomic loci. These phenotypes depended on JMJD2A enzymatic activity. Strikingly, depletion of the only C. elegans homolog, JMJD-2, slowed DNA replication and increased ATR/p53-dependent apoptosis. Importantly, overexpression of HP1gamma antagonized JMJD2A-dependent progression through S phase, and depletion of HPL-2 rescued the DNA replication-related phenotypes in jmjd-2-/- animals. These findings describe a highly conserved model whereby JMJD2A regulates DNA replication by antagonizing HP1gamma and controlling chromatin accessibility (Black, 2010).

The JmjC domain protein Epe1 prevents unregulated assembly and disassembly of heterochromatin

Heterochromatin normally has prescribed chromosomal positions and must not encroach on adjacent regions. This study demonstrates that the fission yeast protein Epe1 stabilises silent chromatin, preventing the oscillation of heterochromatin domains. Epe1 loss leads to two contrasting phenotypes: alleviation of silencing within heterochromatin and expansion of silent chromatin into neighbouring euchromatin. Thus, it is proposed that Epe1 regulates heterochromatin assembly and disassembly, thereby affecting heterochromatin integrity, centromere function and chromosome segregation fidelity. Epe1 regulates the extent of heterochromatin domains at the level of chromatin, not via the RNAi pathway. Analysis of an ectopically silenced site suggests that heterochromatin oscillation occurs in the absence of heterochromatin boundaries. Epe1 requires predicted iron- and 2-oxyglutarate (2-OG)-binding residues for in vivo function, indicating that it is probably a 2-OG/Fe(II)-dependent dioxygenase. It is suggested that, rather than being a histone demethylase, Epe1 may be a protein hydroxylase that affects the stability of a heterochromatin protein, or protein-protein interaction, to regulate the extent of heterochromatin domains. Thus, Epe1 ensures that heterochromatin is restricted to the domains to which it is targeted by RNAi (Trewick, 2007).


Search PubMed for articles about Drosophila KDM4a

Awwad, S. W. and Ayoub, N. (2015). Overexpression of KDM4 lysine demethylases disrupts the integrity of the DNA mismatch repair pathway. Biol Open 4(4): 498-504. PubMed ID: 25770186

Black, J. C., Allen, A., Van Rechem, C., Forbes, E., Longworth, M., Tschop, K., Rinehart, C., Quiton, J., Walsh, R., Smallwood, A., Dyson, N. J. and Whetstine, J. R. (2010). Conserved antagonism between JMJD2A/KDM4A and HP1gamma during cell cycle progression. Mol Cell 40(5): 736-748. PubMed ID: 21145482

Colmenares, S. U., Swenson, J. M., Langley, S. A., Kennedy, C., Costes, S. V. and Karpen, G. H. (2017). Drosophila histone demethylase KDM4A has enzymatic and non-enzymatic roles in controlling heterochromatin integrity. Dev Cell 42(2): 156-169. PubMed ID: 28743002

Crona, F., Dahlberg, O., Lundberg, L. E., Larsson, J. and Mannervik, M. (2013). Gene regulation by the lysine demethylase KDM4A in Drosophila. Dev Biol 373(2): 453-463. PubMed ID: 23195220

Fereres, S., Simon, R., Mohd-Sarip, A., Verrijzer, C. P. and Busturia, A. (2014). dRYBP counteracts chromatin-dependent activation and repression of transcription. PLoS One 9: e113255. PubMed ID: 25415640

Filion, G. J., van Bemmel, J. G., Braunschweig, U., Talhout, W., Kind, J., Ward, L. D., Brugman, W., de Castro, I. J., Kerkhoven, R. M., Bussemaker, H. J. and van Steensel, B. (2010). Systematic protein location mapping reveals five principal chromatin types in Drosophila cells. Cell 143(2): 212-224. PubMed ID: 20888037

Kupershmit, I., Khoury-Haddad, H., Awwad, S. W., Guttmann-Raviv, N. and Ayoub, N. (2014). KDM4C (GASC1) lysine demethylase is associated with mitotic chromatin and regulates chromosome segregation during mitosis. Nucleic Acids Res 42(10): 6168-6182. PubMed ID: 24728997

Lin, C. H., Li, B., Swanson, S., Zhang, Y., Florens, L., Washburn, M. P., Abmayr, S. M. and Workman, J. L. (2008). Heterochromatin protein 1a stimulates histone H3 lysine 36 demethylation by the Drosophila KDM4A demethylase. Mol Cell 32(5): 696-706. PubMed ID: 19061644

Lin, C. H., Paulson, A., Abmayr, S. M. and Workman, J. L. (2012). HP1a targets the Drosophila KDM4A demethylase to a subset of heterochromatic genes to regulate H3K36me3 levels. PLoS One 7: e39758. PubMed Citation: 22761891

Lloret-Llinares, M., Carre, C., Vaquero, A., de Olano, N. and Azorin, F. (2008). Characterization of Drosophila melanogaster JmjC+N histone demethylases. Nucleic Acids Res 36(9): 2852-2863. PubMed ID: 18375980

Mallette, F. A. and Richard, S. (2012). JMJD2A promotes cellular transformation by blocking cellular senescence through transcriptional repression of the tumor suppressor CHD5. Cell Rep 2(5): 1233-1243. PubMed ID: 23168260

Peng, J. C. and Karpen, G. H. (2007). H3K9 methylation and RNA interference regulate nucleolar organization and repeated DNA stability. Nat Cell Biol 9(1): 25-35. PubMed ID: 17159999

Ponnaluri, V. K., Vavilala, D. T., Putty, S., Gutheil, W. G. and Mukherji, M. (2009). Identification of non-histone substrates for JMJD2A-C histone demethylases. Biochem Biophys Res Commun 390(2): 280-284. PubMed ID: 19799855

Slee, R. B., Steiner, C. M., Herbert, B. S., Vance, G. H., Hickey, R. J., Schwarz, T., Christan, S., Radovich, M., Schneider, B. P., Schindelhauer, D. and Grimes, B. R. (2012). Cancer-associated alteration of pericentromeric heterochromatin may contribute to chromosome instability. Oncogene 31(27): 3244-3253. PubMed ID: 22081068

Trewick, S. C., Minc, E., Antonelli, R., Urano, T. and Allshire, R. C. (2007). The JmjC domain protein Epe1 prevents unregulated assembly and disassembly of heterochromatin. EMBO J 26(22): 4670-4682. PubMed ID: 17948055

Tsurumi, A., Dutta, P., Shang, R., Yan, S. J. and Li, W. X. (2013). Drosophila Kdm4 demethylases in histone H3 lysine 9 demethylation and ecdysteroid signaling. Sci Rep 3: 2894. PubMed ID: 24100631

Wang, L. Y., Hung, C. L., Chen, Y. R., Yang, J. C., Wang, J., Campbell, M., Izumiya, Y., Chen, H. W., Wang, W. C., Ann, D. K. and Kung, H. J. (2016). KDM4A Coactivates E2F1 to Regulate the PDK-Dependent Metabolic Switch between Mitochondrial Oxidation and Glycolysis. Cell Rep 16: 3016-3027. PubMed ID: 27626669

Whetstine, J. R., Nottke, A., Lan, F., Huarte, M., Smolikov, S., Chen, Z., Spooner, E., Li, E., Zhang, G., Colaiacovo, M. and Shi, Y. (2006). Reversal of histone lysine trimethylation by the JMJD2 family of histone demethylases. Cell 125(3): 467-481. PubMed ID: 16603238

Biological Overview

date revised: 22 February, 2018

Home page: The Interactive Fly © 2011 Thomas Brody, Ph.D.