RNA interference (RNAi) is a highly conserved process of gene silencing in which Argonaute family proteins are guided by small RNA molecules to complementary targets. In the fission yeast Schizosaccharomyces pombe, RNAi is required for heterochromatin formation at centromeres. Although it seems counterintuitive, pericentromeric heterochromatin in fission yeast is transcribed. The transcripts are processed by RNAi machinery, which is in turn guided back to the pericentromeric repeats by sequence complementarity of the Argonaute-bound small interfering RNA (siRNA) and the nascent transcript. This generates a positive-feedback loop of siRNA amplification that recruits factors required for the assembly of heterochromatin. Previously, it was suggested that a fission yeast class of Dicer-independent small RNAs called primal small RNAs (priRNAs) initiates the positive-feedback loop of siRNA generation and heterochromatin assembly. However, the biogenesis of priRNAs as well as of Dicer-independent small RNAs from other organisms was not well understood.

The results presented here identify Triman, a novel 3’-5’ exonuclease that is involved in the final step of biogenesis of both priRNAs and siRNAs in fission yeast. It was observed that Argonaute binds longer priRNA and siRNA precursors from the total RNA fraction. This is followed by the recruitment of Triman to trim 3’ ends of Argonaute-bound small RNAs to the mature size. The final trimming of priRNAs and siRNAs is required for de novo heterochromatin formation at centromeres and the mating-type locus as well as for the maintenance of facultative heterochromatin islands. Furthermore, it was shown that in cells lacking Rrp6, a nuclease subunit of the exosome, RNAi targets various genes across the yeast genome. This demonstrated that the exosome protects the genome against aberrant RNAi. Spurious RNAi targeting in rrp6∆ cells at majority of loci occurs via accumulation of antisense transcripts that are processed into priRNAs in a Triman-dependent manner. These results suggest that Argonaute association with cellular degradation products which are processed into priRNAs might serve as a surveillance mechanism to guard the genome against invading genomic elements (Marasovic et al. 2013).

The 2-oxoglutarate and Fe (II)-dependent ten-eleven translocation (TET) enzymes can oxidize 5-methylcytosine (5mC) to 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC) and 5-carboxycytosine (5caC). Thus, it was hypothesized that this could be a pathway for active DNA demethylation. However, several studies have shown that 5hmC levels accumulate especially in neurons suggesting a role as an epigenetic mark. Nevertheless, the mechanism controlling TET activity and the role of 5hmC are poorly understood. In particular, it is not known how the neuronal TET3 isoform lacking a DNA binding domain is targeted to the DNA.

In this study, the role of 5hmC during retinal maturation by genome-wide mapping of 5hmC in immature (postnatal week 2) and mature mouse retina (postnatal week 3) was studied and correlated with expression data. Furthermore, interaction partners of neu-ronal TET3, the main isoform in retinal neurons, were identified.

5hmC accumulates during retinal maturation especially in retinal and neuronal genes. Furthermore, the accumulation of 5hmC is associated with increased transcription. Among the identified proteins the transcriptional repressor REST was identified as a highly enriched TET3-specific interactor. Interestingly, REST was able to enhance TET3 hydroxylase activity. Furthermore, increased 5hmC levels were detected in REST target genes during retinal maturation and overexpression of TET3 activated transcription of REST-target genes. Moreover, NSD3 and two other histone methyltransferases were found to interact with TET3 that are able to mediate H3K36 trimethylation. Finally, it was shown that TET3 is able to enhance NSD3-mediated H3K36 trimethylation to pro-mote transcriptional activation.

In conclusion, the data suggests that 5hmC is a stable epigenetic base in retinal neurons that is involved in transcriptional activation. Furthermore, it was shown that 5hmC is generated by TET3 that is recruited to the DNA by transcriptional regulators such as REST in a context-specific manner.

Even though platelets are the smallest cells in circulating blood, they play an integral role in blood clotting where they are activated, adhere to the vessel wall, and contribute to hemostasis. But over the years it was discovered that those anucleate cells have more extended functions. They organize their cellular vitality similar to nucleated cells and have an active protein metabolism performing protein de novo synthesis as well as protein degradation. One of the main degradation systems in cells is the proteasome. Besides protein quality control, the proteasome is involved in important cellular processes like cell survival, transcription, development, selective elimination of abnormal proteins and antigen processing.

A dysregulation of this multicatalytic protein complex leads to various disease developments. Proteasome inhibitors, for instance, have been studied for treating cancer. Platelets like nucleated cells contain a proteasome. However, the impact of the proteasome on platelet functions remains poorly investigated until today. A better knowledge of signaling pathways in platelets aids in understanding how alterations in proteasome functions affect platelet-mediated processes and diseases.

This study confirms the existence of a functional proteasome in human platelets and illustrates an important role in platelet biology, as well as sepsis.

With this study the role of the proteasome in anucleate platelets is demonstrated in more detail and a signaling pathway regulating its activity was observed. Here, the proteasome in platelets is linked to platelet aggregation. First, proteasome inhibitors epoxomicin and bortezomib reduce ADP- and collagen-induced aggregation. Furthermore, the 26S chymotrypsin-like activity of the proteasome is enhanced when platelets are incubated with the platelet agonist collagen. Additionally, cytoskeletal proteins Filamin A and Talin-1, which are crucial for platelet activation, were identified as proteasome substrates and increased cleavage of these proteins occurs with proteasome activation.

To investigate possible mechanisms of regulating the proteasome, the signaling pathway related to NFκB was analyzed under platelet agonist treatment. The NFκB pathway, that mediates aggregation, is initiated when platelets are treated with collagen and the inhibitory protein of NFκB, IκBα, is degraded in collagen-stimulated platelets. More interestingly, NFκB inhibitors prevent collagen-stimulated enhancement of the proteasome activity. In return the connection of the proteasome and the NFκB pathway is further demonstrated as NFκB inhibitors restrict cleavage of the proteasome substrate Talin-1. These results propose a novel pathway that involves the proteasome and that is in return connected with non-genomic functions of NFκB in regulating platelet aggregation.

In a second part this work shows for the first time that mitochondrial membrane depolarization in platelets correlates with the disease course and disease severity in patients with sepsis. Additionally, during these studies increased proteasome activity was observed in sepsis patients compared to control patients and pathogenic bacteria intensified the 26S trypsin-like activity of human platelets. Therefore, molecular markers of platelet vitality may be valuable parameters to help evaluating the clinical outcome of sepsis patients.

In summary, the study confirms the existence of a functional proteasome in human platelets, contributes to our understanding how the proteasome affects platelet functions such as aggregation and how this may be regulated on a molecular basis. Furthermore, it allows for new insights in the disease course of sepsis and identifies new molecular markers for assessing the disease severity and clinical outcome of sepsis patients.

Epithelial barriers are central to the development of metazoans by compartmentalizing the body in distinct chemical milieus essential for the function of many organs. One such barrier is the blood-brain barrier, which isolates the nervous system from the body fluid to maintain its ionic homeostasis and ensure nerve pulse transmission. In Drosophila, the blood-brain barrier is formed late in embryogenesis by a thin epithelium of subperineurial glia that ensheath the nervous system. Similar to other epithelia, subperineurial glia seal the paracellular space by forming large multiprotein complexes at the lateral membrane, the septate junctions (SJs), which impede free diffusion and mediate barrier function.

To identify novel genes required for blood-brain barrier formation, we followed a genome-wide in vivo RNAi approach. We initially screened almost the whole genome for genes required in glia for adult viability and impressively identified 3679 potential candidates. Subsequently, we tested these candidates for requirement in subperineurial glia for adult survival and identified 383 genes. At a last step, we directly asked if blood-brain barrier formation is compromised in the knock-down of the genes by performing the embryonic dye penetration assay in a selection of candidates and identified five genes that play a role during barrier development. Three of these genes, macroglobulin complement-related (mcr) and the previously uncharacterized pasiflora1 and pasiflora2 are further characterized in the context of this thesis.

Here we show that all three proteins are novel components of the Drosophila SJ. Pasiflora1 and Pasiflora2 belong to a previously uncharacterized family of tetra-spanning membrane proteins, while Mcr was reported to be a secreted protein in S2 cells required for phagocytosis and clearance of specific pathogens. Through detailed phenotypic analysis we demonstrate that the mutants show leaky blood-brain and tracheal barriers, overelongated tracheal tubes and mislocalization of SJ proteins, phenotypes that are characteristic of SJ mutants. Consistent with the observed phenotypes, the genes are co-expressed in SJ-forming embryonic epithelia and glia and are required cell-autonomously to exert their function. In columnar epithelia, the proteins localize at the apicolateral membrane compartment, where they colocalize with other SJ proteins, and similar to known SJ components, their restricted localization depends on other complex members. Using fluorescence recovery after photobleaching experiments, we demonstrate for Pasiflora proteins that they are core SJ components, as they are required for complex formation and themselves show restricted mobility within the membrane of wild-type epithelial cells, but fast diffusion in cells with disrupted SJs. Taken together, our results show that Pasiflora1 and Pasiflora2 are novel integral SJ components and implicate a new family of tetraspan proteins in the development of cell junctions. In addition, we find a new unexpected role for Mcr as a transmembrane SJ protein, which raises questions about a potential intriguing link between epithelial barrier function, phagocytosis and innate immunity and has potential implications for the function of occluding junctions.

Peroxisomes are highly dynamic organelles of eukaryotic cells, carrying out essential oxidative metabolic processes. These organelles scavenge reactive oxygen species such as hydrogen peroxide (H2O2) and catabolise fatty acids, which are particular hallmarks and highly conserved features of peroxisomes among different species. Peroxisomal proteins and enzymes are encoded by nuclear DNA and therefore, targeted post-translationally into the peroxisomal matrix. A special class of proteins, collectively called peroxins, perform certain cellular tasks, such as peroxisomal matrix protein import or membrane development in order to maintain peroxisome biogenesis as well as a constant flux of matrix proteins into peroxisomes. The type II AAA+ peroxins Pex1/Pex6 are a core component of the peroxisomal matrix protein import system. ATPases of the AAA+ family of proteins generally assemble into large, macromolecular machines, structurally remodelling their substrate protein, which is driven by the hydrolysis of ATP. The main function of Pex1/6 complexes is to release the receptor Pex5 from peroxisomal membranes after matrix protein import. This relocation of Pex5 into the cytosol ensures a constant pool of available receptor molecules for subsequent cycles of protein import into peroxisomes. Accordingly, certain mutations in mammalian Pex1/Pex6 proteins compromise peroxisome biogenesis and thus, lipid metabolism, causing severe genetic Zellweger diseases in humans.

In collaboration with Professor Ralf Erdmann and colleagues at the Ruhr-Universität Bochum, we characterize the structure and function of the AAA+ Pex1/6 complex from yeast Saccharomyces cerevisiae. Single particle electron microscopy (EM) in combination with biochemical assays allows us to analyze how ATP turnover is related to the biological function of the Pex1/6 complex. This study presents EM structures of Pex1/6 complexes assembled in the presence of ADP, ATP, ADP-AlFx and ATPγS, providing a comprehensive structural characterization of the heterohexameric type II AAA+ complex in different nucleotide states. Our EM reconstructions reveal an unexpected triangular overall shape, different than observed for the closely related and well-characterized homohexameric AAA+ protein p97. We show that the heterohexameric Pex1/6 complex is composed of a trimer of heterodimers with alternating subunit arrangement of Pex1 and Pex6 moieties. Furthermore, our results suggest that conserved aromatic residues, lining the central pore of the Pex1/6 D2 ring mediate substrate interactions. These residues correspond to substrate interaction regions in related AAA+ proteins. Comparing Pex1/6 EM reconstructions in different nucleotide states implicates that the mechanical function of Pex1/6 involves an N- to C-terminal protein translocation mechanism along the central pore. The Pex1/6 EM structures resolve symmetric and asymmetric large-scale domain motions, which likely create a power stroke during cycles of ATP binding and hydrolysis. We conclude that Pex5 is probably partially or completely unfolded while it is threaded through the central pore of Pex1/6 complexes.

In addition, ATP hydrolysis assays of Pex1/Pex6 complexes containing single amino acid exchanges in individual Walker B motifs reveal that not all active sites are functionally equivalent. In isolated complexes, ATP turnover mainly occurs in Pex6 D2 domains, while Pex1 subunits sustain the structural integrity of the complex. We further resolve the structures of Pex1/6 Walker B variants and observe mutually exclusive protomer-protomer communication. In the Pex1/6 complex, a Walker B mutation induces ATP hydrolysis in the adjacent D2 domain, presenting a structural framework of protomer-protomer communication in the AAA+ heterohexamer.