DNA methylation in Dictyostelium discoideum

DNA methyltransferases of type Dnmt2 are a highly conserved protein family with enigmatic function. The aim of this work was to characterize DnmA, the Dnmt2 methyltransferase in Dictyostelium discoideum, and further to investigate its implication in DNA methylation and transcriptional gene silencing. The genome of the social amoeba Dictyostelium encodes DnmA as the sole DNA methyltransferase. The enzyme bears all ten characteristic DNA methyltransferase motifs in its catalytic domain. The DnmA mRNA was found by RT-PCR to be expressed during vegetative growth and down regulated during development. Investigations using fluorescence microscopy showed that both DnmA-myc and DnmA-GFP fusions predominantly localised to the nucleus. The function of DnmA remained initially unclear, but later experiment revealed that the enzyme is an active DNA methyltransferase responsible for all DNA (cytosine) methylation in Dictyostelium. Neither in gel retardation assays, nor by the yeast two hybrid system, clues on the functionality of DnmA could be obtained. However, immunological detection of the methylation mark with an α - 5mC antibody gave initial evidence that the DNA of Dictyostelium was methylated. Furthermore, addition of 5-aza-cytidine as demethylating agent to the Dictyostelium medium and subsequent in vitro incubation of the DNA isolated from these cells with recombinant DnmA showed that the enzyme binds slightly better to this target DNA. In order to investigate further the function of the protein, a gene knock-out for dnmA was generated. The gene was successfully disrupted by homologous recombination, the knock-out strain, however, did not show any obvious phenotype under normal laboratory conditions. To identify specific target sequences for DNA methylation, a microarray analysis was carried out. Setting a threshold of at least 1.5 fold for differences in the strength of gene expression, several such genes in the knock-out strain were chosen for further investigation. Among the up-regulated genes were the ESTs representing the gag and the RT genes respectively of the retrotransposon skipper. In addition Northern blot analysis confirmed the up-regulation of skipper in the DnmA knock-out strain. Bisufite treatment and sequencing of specific DNA stretches from skipper revealed that DnmA is responsible for methylation of mostly asymmetric cytosines. Together with skipper, DIRS-1 retrotransposon was found later also to be methylated but was not present on the microarray. Furthermore, skipper transcription was also up-regulated in strains that had genes disrupted encoding components of the RNA interference pathway. In contrast, DIRS 1 expression was not affected by a loss of DnmA but was strongly increased in the strain that had the RNA directed RNA polymerase gene rrpC disrupted. Strains generated by propagating the usual wild type Ax2 and the DnmA knock-out cells over 16 rounds in development were analyzed for transposon activity. Northern blot analysis revealed activation for skipper expression, but not for DIRS-1. A large number of siRNAs were found to be correspondent to the DIRS-1 sequence, suggesting concerted regulation of DIRS-1 expression by RNAi and DNA methylation. In contrast, no siRNAs corresponding to the standard skipper element were found. The data show that DNA methylation plays a crucial role in epigenetic gene regulation in Dictyostelium and that different, partially overlapping mechanisms control transposon silencing for skipper and DIRS-1. To elucidate the mechanism of targeting the protein to particular genes in the Dictyostelium genome, some more genes which were up-regulated in the DnmA knock-out strain were analyzed by bisulfite sequencing. The chosen genes are involved in the multidrug response in other species, but their function in Dictyostelium is uncertain. Bisulfite data showed that two of these genes were methylated at asymmetrical C-residues in the wild type, but not in DnmA knock-out cells. This suggested that DNA methylation in Dictyostelium is involved not only in transposon regulation but also in transcriptional silencing of specific genes.

Heterochromatin und epigenetische Kontrolle der Centromere in Dictyostelium discoideum

Heterochromatin Protein 1 (HP1) is an evolutionarily conserved protein required for formation of a higher-order chromatin structures and epigenetic gene silencing. The objective of the present work was to functionally characterise HP1-like proteins in Dictyostelium discoideum, and to investigate their function in heterochromatin formation and transcriptional gene silencing. The Dictyostelium genome encodes three HP1-like proteins (hcpA, hcpB, hcpC), from which only two, hcpA and hcpB, but not hcpC were found to be expressed during vegetative growth and under developmental conditions. Therefore, hcpC, albeit no obvious pseudogene, was excluded from this study. Both HcpA and HcpB show the characteristic conserved domain structure of HP1 proteins, consisting of an N-terminal chromo domain and a C-terminal chromo shadow domain, which are separated by a hinge. Both proteins show all biochemical activities characteristic for HP1 proteins, such as homo- and heterodimerisation in vitro and in vivo, and DNA binding activtity. HcpA furthermore seems to bind to K9-methylated histone H3 in vitro. The proteins thus appear to be structurally and functionally conserved in Dictyostelium. The proteins display largely identical subnuclear distribution in several minor foci and concentration in one major cluster at the nuclear periphery. The localisation of this cluster adjacent to the nucleus-associated centrosome and its mitotic behaviour strongly suggest that it represents centromeric heterochromatin. Furthermore, it is characterised by histone H3 lysine-9 dimethylation (H3K9me2), which is another hallmark of Dictyostelium heterochromatin. Therefore, one important aspect of the work was to characterise the so-far largely unknown structural organisation of centromeric heterochromatin. The Dictyostelium homologue of inner centromere protein INCENP (DdINCENP), co-localized with both HcpA and H3K9me2 during metaphase, providing further evidence that H3K9me2 and HcpA/B localisation represent centromeric heterochromatin. Chromatin immunoprecipitation (ChIP) showed that two types of high-copy number retrotransposons (DIRS-1 and skipper), which form large irregular arrays at the chromosome ends, which are thought to contain the Dictyostelium centromeres, are characterised by H3K9me2. Neither overexpression of full-length HcpA or HcpB, nor deletion of single Hcp isoforms resulted in changes in retrotransposon transcript levels. However, overexpression of a C-terminally truncated HcpA protein, assumed to display a dominant negative effect, lead to an increase in skipper retrotransposon transcript levels. Furthermore, overexpression of this protein lead to severe growth defects in axenic suspension culture and reduced cell viability. In order to elucidate the proteins functions in centromeric heterochromatin formation, gene knock-outs for both hcpA and hcpB were generated. Both genes could be successfully targeted and disrupted by homologous recombination. Surprisingly, the degree of functional redundancy of the two isoforms was, although not unexpected, very high. Both single knock-out mutants did not show any obvious phenotypes under standard laboratory conditions and only deletion of hcpA resulted in subtle growth phenotypes when grown at low temperature. All attempts to generate a double null mutant failed. However, both endogenous genes could be disrupted in cells in which a rescue construct that ectopically expressed one of the isoforms either with N-terminal 6xHis- or GFP-tag had been introduced. The data imply that the presence of at least one Hcp isoform is essential in Dictyostelium. The lethality of the hcpA/hcpB double mutant thus greatly hampered functional analysis of the two genes. However, the experiment provided genetic evidence that the GFP-HcpA fusion protein, because of its ability to compensate the loss of the endogenous HcpA protein, was a functional protein. The proteins displayed quantitative differences in dimerisation behaviour, which are conferred by the slightly different hinge and chromo shadow domains at the C-termini. Dimerisation preferences in increasing order were HcpA-HcpA << HcpA-HcpB << HcpB-HcpB. Overexpression of GFP-HcpA or a chimeric protein containing the HcpA C-terminus (GFP-HcpBNAC), but not overexpression of GFP-HcpB or GFP-HcpANBC, lead to increased frequencies of anaphase bridges in late mitotic cells, which are thought to be caused by telomere-telomere fusions. Chromatin targeting of the two proteins is achieved by at least two distinct mechanisms. The N-terminal chromo domain and hinge of the proteins are required for targeting to centromeric heterochromatin, while the C-terminal portion encoding the CSD is required for targeting to several other chromatin regions at the nuclear periphery that are characterised by H3K9me2. Targeting to centromeric heterochromatin likely involves direct binding to DNA. The Dictyostelium genome encodes for all subunits of the origin recognition complex (ORC), which is a possible upstream component of HP1 targeting to chromatin. Overexpression of GFP-tagged OrcB, the Dictyostelium Orc2 homologue, showed a distinct nuclear localisation that partially overlapped with the HcpA distribution. Furthermore, GFP-OrcB localized to the centrosome during the entire cell cycle, indicating an involvement in centrosome function. DnmA is the sole DNA methyltransferase in Dictyostelium required for all DNA(cytosine-)methylation. To test for its in vivo activity, two different cell lines were established that ectopically expressed DnmA-myc or DnmA-GFP. It was assumed that overexpression of these proteins might cause an increase in the 5-methyl-cytosine(5-mC)-levels in the genomic DNA due to genomic hypermethylation. Although DnmA-GFP showed preferential localisation in the nucleus, no changes in the 5-mC-levels in the genomic DNA could be detected by capillary electrophoresis.

Biologische Funktionsanalyse und Identifizierung neuer Substrate der Methyltransferase Dnmt2

Obwohl die DNA Methyltransferase 2 (Dnmt2) hoch konserviert ist und zu der am weitesten verbreiteten eukaryotischen MTase-Familie gehört, ist ihre biologische Funktion nach wie vor unklar. Nachdem lange Zeit keine DNA Methylierungsaktivität nachgewiesen werden konnte, wurde vor einigen Jahren über geringe Mengen an 5-Methylcytosin (5mC) in Retroelementen der “Dnmt2-only”-Organismen D. melanogaster, D. discoideum und E. histolytica berichtet (Kunert et al. 2003; Fisher et al. 2004; Kuhlmann et al. 2005; Phalke et al. 2009). Als kurze Zeit später robuste Methylierung der tRNAAsp durch humane Dnmt2 gezeigt wurde (Goll et al. 2006), wurde zunächst eine Dualspezifität des Enzyms vorgeschlagen (Jeltsch et al. 2006). Neuere Daten zum 5mC-Status verschiedener „Dnmt2-only“-Organismen bilden Anlass für kontroverse Diskussionen über Ausmaß und Bedeutung der DNA Methyltransferaseaktivität von Dnmt2 (Schaefer et al. 2010a; Krauss et al. 2011). Die vorliegende Arbeit konzentriert sich auf die Identifizierung neuer RNA Substrate des Dnmt2-Homologs DnmA aus D. discoideum sowie die biologische Bedeutung der tRNA-Methylierung durch Dnmt2. Wie in anderen Organismen beschrieben, fungiert auch DnmA als tRNAAsp(GUC) MTase in vitro und in vivo. Zusätzlich konnte in vitro tRNAGlu(UUC) als neues Substrat der Dnmt2-Homologe aus D. discoideum und dem Menschen identifiziert werden. In einem Kooperationsprojekt wurde außerdem auch tRNAAsp-Methylierungsaktivität für das Dnmt2-Homolog aus S. pombe (Pmt1) nachgewiesen. Crosslink-RNA-Immunopräzipitationen (RNA-CLIP) mit anschließender Next-Generation-Sequenzierung der mit DnmA assoziierten RNAs zeigen, dass DnmA mit tRNA Fragmenten interagiert, die sich vom Anticodonloop bis in den T-loop erstrecken. Neben der tRNAAsp(GUC) und tRNAGlu(UUC/CUC) sind Fragmente der tRNAGly(GCC) verstärkt angereichert. Inwiefern diese Fragmente eine biologische Funktion haben oder spezifische Degradationsprodukte darstellen, ist noch ungeklärt. Interessanterweise sind von einigen tRNAs wenige Sequenzen von antisense-Fragmenten in den RNA-CLIP Daten zu finden, die etwas kürzer, jedoch exakt komplementär zu den genannten sense-Fragmenten sind. Besonders stark sind diese Fragmente der tRNAGlu(UUC) vertreten. In einem weiteren RNA-CLIP Experiment wurden U-snRNAs, snoRNA und intergenische Sequenzen mit DnmA angereichert. Bei nachfolgenden in vitro Methylierungsstudien konnte ausschließlich die U2-snRNA als potentielles Nicht-tRNA-Substrat der hDnmt2 und DnmA identifiziert werden. Da tRNA Modifikationen im Anticodonloop die Codonerkennung beeinflussen können, wurde ein System etabliert um die Translationseffizienz eines GFP-Reportergens in Wildtyp- und dnmAKO-Zellen zu messen. In D. discoideum wird das Aspartat-Codon GAU ca. zehnmal häufiger genutzt als das GAC Codon, allerdings ist nur eine tRNAAsp(GUC) im Genom der Amöbe kodiert. Aus diesem Grund wurde zusätzlich die Frage adressiert, inwiefern die DnmA-abhängige Methylierung dieser tRNA das „Wobbling“ beeinflusst. Dazu wurde dem Reportergen jeweils eine (GAU)5- und (GAC)5-Leadersequenz vorgeschaltet. Entgegen der Annahme wurde der (GAC)5-Leader in beiden Stämmen etwas effizienter translatiert. Insgesamt zeigte der dnmAKO-Stamm eine leicht erhöhte Translationseffizienz der Reportergene. Vergleichende Analysen zur Aufnahme von Fremd-DNA zeigten signifikant reduzierte Transformationseffizienzen mit einem integrierenden Plasmid in dnmAKO-Zellen. Ein weiterer dnmAKO-Stamm zeigte diesen Effekt jedoch nicht, wobei bei derselben Mutante eine deutlich reduzierte Aufnahme eines extrachromosomalen Plasmids zu verzeichnen war. Untersuchungen zum Einfluss von DnmA auf die Regulation des Retroelements skipper ergaben keinen Zusammenhang zwischen der Generierung kleiner RNAs und der erhöhten Transkription des Retrotransposons in dnmAKO-Zellen (Kuhlmann et al. 2005). Durch Kompensationsversuche sowie Experimente mit einer weiteren dnmAKO-Mutante konnte die Mobilisierung des Retrotransposons nicht eindeutig als DnmA-Funktion eingeordnet werden. In einem weiteren Projekt wurden die Bindung des m5C-bindenden Proteins EhMLBP aus E. histolytica an DNA mittels Rasterkraftmikroskopie abgebildet (Lavi et al. 2006). Neben vermutlich unspezifischen Endbindungsereignissen konnte eine bevorzugte Bindungsstelle des Proteins an LINE DNA (long intersperesed nuclear element) identifiziert werden. Möglicherweise fällt diese mit einem von zwei A/T-reichen Bereichen der LINE DNA zusammen, von denen vermutet wird, dass diese für die Bindung von EhMLBP an DNA von Bedeutung sind. Insgesamt bestätigen die Ergebnisse dieser Arbeit die tRNAAsp Methylierungsaktivität als konservierte Dnmt2-Funktion. Darüber hinaus erweitern sie das Substratspektrum der Dnmt2-Methyltransferasen im Bereich der tRNA. Außerdem wird erstmals ein potentielles Nicht-tRNA Substrat vorgeschlagen. Zusätzlich geben neu entdeckte Phänotypen Hinweise auf vielfältige zelluläre Dnmt2-Funktionen.

Zelluläre und biochemische Charakterisierung der bifunktionalen Methyltransferase Dnmt2

Seit der Entdeckung der Methyltransferase 2 als hoch konserviertes und weit verbreitetes Enzym sind zahlreiche Versuche zur vollständigen Charakterisierung erfolgt. Dabei ist die biologische Funktion des Proteins ein permanent umstrittener Punkt. In dieser Arbeit wird dnmA als sensitiver Oszillator bezüglich des Zellzyklus und weiterer Einflüsse gezeigt. Insgesamt liegt der Hauptfokus auf der Untersuchung der in vivo Charakterisierung des Gens, der endogenen subzellulären Verteilung, sowie der physiologischen Aufgaben des Proteins in vivo in D. discoideum. Um Hinweise auf Signalwege in vivo zu erhalten, in denen DnmA beteiligt ist, war es zunächst notwendig, eine detaillierte Analyse des Gens anzufertigen. Mit molekularbiologisch äußerst sensitiven Methoden, wie beispielsweise Chromatin‐IP oder qRT‐PCR, konnte ein vollständiges Expressionsprofil über den Zell‐ und Lebenszyklus von D. discoideum angelegt werden. Besonders interessant sind dabei die Ergebnisse eines ursprünglichen Wildtypstammes (NC4), dessen dnmA‐Expressionsprofil quantitativ von anderen Wildtypstämmen abweicht. Auch auf Proteinebene konnten Zellzyklus‐abhängige Effekte von DnmA bestimmt werden. Durch mikroskopische Untersuchungen von verschiedenen DnmA‐GFP‐Stämmen wurden Lokalisationsänderungen während der Mitose gezeigt. Weiterhin wurde ein DnmA‐GFP‐Konstrukt unter der Kontrolle des endogenen Promotors generiert, wodurch das Protein in der Entwicklung eindeutig als Zelltypus spezifisches Protein, nämlich als Präsporen‐ bzw. Sporenspezifisches Protein, identifiziert werden konnte. Für die in vivo Analyse der katalytischen Aktivität des Enzyms konnten nun die Erkenntnisse aus der Charakterisierung des Gens bzw. Proteins berücksichtigt werden, um in vivo Substratkandidaten zu testen. Es zeigte sich, dass von allen bisherigen Substrat Kandidaten lediglich die tRNA^Asp als in vivo Substrat bestätigt werden konnte. Als besondere Erkenntnis konnte hierbei ein quantitativer Unterschied des Methylierungslevels zwischen verschiedenen Wildtypstämmen detektiert werden. Weiterhin wurde die Methylierung sowie Bindung an einen DNA‐Substratkandidaten ermittelt. Es konnte gezeigt werden, dass DnmA äußerst sequenzspezifisch mit Abschnitten des Retrotransposons DIRS‐1 in vivo eine Bindung eingeht. Auch für den Substrakandidaten snRNA‐U2 konnte eine stabile in vitro Komplexbildung zwischen U2 und hDnmt2 gezeigt werden. Insgesamt erfolgte auf Basis der ermittelten Expressionsdaten eine erneute Charakterisierung der Aktivität des Enzyms und der Substrate in vivo und in vitro.

Funktion und Biogenese kleiner RNAs in Dictyostelium discoideum

RNA mediated gene silencing pathways are highly conserved among eukaryotes and they have been well investigated in animals and in plants. Longer dsRNA molecules trigger the silencing pathways: RNase III proteins and their dsRNA binding protein (dsRBP) partners recognize those molecules as a substrate and process 21 nucleotide long microRNAs (miRNAs) or small interfering RNAs (siRNAs). Some organisms encode RNA dependent RNA polymerases (RdRPs), which are able to expand the pool of existing siRNAs. Argonaute proteins are able to bind small regulatory RNAs and are subsequently recruited to target mRNAs by base complementary. This leads in turn to transcriptional or posttranscriptional silencing of respective genes. The Dictyostelium discoideum genome encodes two Dicer homologues (DrnA and DrnB), five Argonaute proteins (AgnA to AgnE) and three RdRPs (RrpA to RrpC). In addition, the amoeba is known to express miRNAs and siRNAs, while the latter derive mainly from the DIRS-1 retrotransposon. One part of this work focused on the miRNA biogenesis pathway of D. discoideum. It was shown that the dsRNA binding protein RbdB is a necessary component for miRNA processing in the amoeba. There were no mature miRNAs detectable by Northern blot analysis in rbdB- strains, which is also true for drnB mutants. Moreover, primary miRNA-transcripts (pri-miRNAs) accumulated in rbdB- and drnB- strains. Fluorescence microscopy studies showed a nuclear localization of RbdB. RbdB accumulated in distinct perinucleolar foci. These were reminiscent of plant dicing bodies that contain essential protein components for miRNA processing. It is well known that RNase III enzymes and dsRBPs work together during miRNA processing in higher eukaryotes. This work demonstrated that the same is true for members of the amoebozoa supergroup. In Arabidopsis the nuclear zinc finger protein Serrate (SE) is also necessary for miRNA processing. The D. discoideum homologue SrtA, however, is not relevant which has been shown by the analysis of the respective knockdown strain. MiRNAs are known to be differentially expressed in several RNAi knockout strains. The accumulation of miRNAs in agnA- strains and a strong decrease in rbdB- strains were criteria that could thus be successfully used (among others) to identify and validate new miRNAs candidates by Illumina®-RNA sequencing. In another part of this study, the silencing and amplification of the DIRS-1 retrotransposons was analyzed in more detail. It was already known that DIRS-1 transcripts and extrachromosomal DIRS-1 DNA molecules accumulated in agnA- strains. This phenotype was correlated with the loss of endogenous DIRS-1 siRNAs in the knockout strain. By deep sequencing analysis of small RNAs from the AX2 wild type and the agnA- strain, the strong decrease of endogenous DIRS-1 siRNAs in the mutant strain (accounting for 70 %) could be confirmed. Further analysis of the data revealed an unequal distribution of DIRS-1 derived siRNAs along the retroelement in the wild type strain, since only very few of them matched the inverted terminal repeats (ITRs) and the 5’- half of the first open reading frame (ORF). Besides, sense and antisense siRNAs were asymmetrically distributed, as well. By using different reporter constructs it was shown indirectly that AgnA is necessary for the RrpC mediated production of secondary DIRS-1 siRNAs. These analyses also demonstrated an amplification of siRNAs in 5’- and in 3’-direction. Further analysis of the agnA- strain revealed that not only DIRS-1 sense transcripts but also ORF2 and ORF3 encoded proteins were enriched. In contrast, the ORF1 encoded protein GAG was equally expressed in the mutant and the wild type. This might reflect the unequal distribution of endogenous DIRS-1 siRNAs along the retrotransposon. Southern Blot and PCR-analyses showed that extrachromosomal DIRS-1 DNA molecules are present in the cytoplasm of angA- strains and that they are complementary to sense transcripts of intact DIRS-1 elements. Thus, the extrachromosomal DIRS-1 intermediates are likely incomplete cDNA molecules generated by the DIRS-1 encoded reverse transcriptase. One could hypothesize that virus like particles (VLPs) are the places of DIRS-1 cDNA synthesis. At least, DIRS-1 GAG proteins interact and fluorescence microscopy studies showed that they localize in distinct cytoplasmic foci which accumulate in close proximity to the nuclei.