Refine
Has Fulltext
- yes (30)
Is part of the Bibliography
- yes (30)
Year of publication
- 2005 (30) (remove)
Document Type
- Doctoral Thesis (29)
- Journal article (1)
Keywords
- Taufliege (4)
- BMP (3)
- Biene (3)
- Drosophila (3)
- Genanalyse (3)
- Molekulargenetik (3)
- Signaltransduktion (3)
- Bioinformatik (2)
- Domäne <Biochemie> (2)
- Gehirn (2)
Institute
- Theodor-Boveri-Institut für Biowissenschaften (30) (remove)
Chemical neurotransmission is a complex process of central importance for nervous system function. It is thought to be mediated by the orchestration of hundreds of proteins for its successful execution. Several synaptic proteins have been shown to be relevant for neurotransmission and many of them are highly conserved during evolution- suggesting a universal mechanism for neurotransmission. This process has checkpoints at various places like, neurotransmitter uptake into the vesicles, relocation of the vesicles to the vicinity of calcium channels in order to facilitate Ca2+ induced release thereby modulating the fusion probability, formation of a fusion pore to release the neurotransmitter and finally reuptake of the vesicles by endocytosis. Each of these checkpoints has now become a special area of study and maintains its own importance for the understanding of the overall process. Ca2+ induced release occurs at specialized membrane structures at the synapse known as the active zones. These are highly ordered electron dense grids and are composed of several proteins which assist the synaptic vesicles in relocating in the vicinity of Ca2+ channels thereby increasing their fusion probability and then bringing about the vesicular fusion itself. All the protein modules needed for these processes are thought to be held in tight arrays at the active zones, and the functions of a few have been characterized so far at the vertebrate active zones. Our group is primarily interested in characterizing the molecular architecture of the Drosophila synapse. Due to its powerful genetics and well-established behavioural assays Drosophila is an excellent system to investigate neuronal functioning. Monoclonal antibodies (MABs) from a hybridoma library against Drosophila brain are routinely used to detect novel proteins in the brain in a reverse genetic approach. Upon identification of the protein its encoding genetic locus is characterized and a detailed investigation of its function is initiated. This approach has been particularly useful to detect synaptic proteins, which may go undetected in a forward genetic approach due to lack of an observable phenotype. Proteins like CSP, Synapsin and Sap47 have been identified and characterized using this approach so far. MAB nc82 has been one of the shortlisted antibodies from the same library and is widely used as a general neuropil marker due to the relative transparency of immunohistochemical whole mount staining obtained with this antibody. A careful observation of double stainings at the larval neuromuscular junctions with MAB nc82 and other pre and post-synaptic markers strongly suggested an active zone localization of the nc82 antigen. Synaptic architecture is well characterized in Drosophila at the ultrastructural level. However, molecular details for many synaptic components and especially for the active zone are almost entirely unknown. A possible localization at the active zone for the nc82 antigen served as the motivation to initiate its biochemical characterization and the identification of the encoding gene. In the present thesis it is shown by 2-D gel analysis and mass spectrometry that the nc82 antigen is a novel active zone protein encoded by a complex genetic locus on chromosome 2R. By RT-PCR exons from three open reading frames previously annotated as separate genes are demonstrated to give rise to a transcript of at least 5.5 kb. Northern blots produce a prominent signal of 11 kb and a weak signal of 2 kb. The protein encoded by the 5.5 kb transcript is highly conserved amongst insects and has at its N-terminus significant homology to the previously described vertebrate active zone protein ELKS/ERC/CAST. Bioinformatic analysis predicts coiled-coil domains spread all over the sequence and strongly suggest a function involved in organizing or maintaining the structure of the active zone. The large C-terminal region is highly conserved amongst the insects but has no clear homologues in veretebrates. For a functional analysis of this protein transgenic flies expressing RNAi constructs under the control of the Gal4 regulated enhancer UAS were kindly provided by the collaborating group of S.Sigrist (Gِttingen). A strong pan-neuronal knockdown of the nc82 antigen by transgenic RNAi expression leads to embryonic lethality. A relatively weaker RNAi expression results in behavioural deficits in adult flies including unstable flight and impaired walking behavior. Due to this peculiar phenotype as observed in the first knockdown studies the gene was named “bruchpilot” (brp) encoding the protein “Bruchpilot (BRP)” (German for crash pilot). A pan-neuronal as well as retina specific downregulation of this protein results in loss of ON and OFF transients in ERG recordings indicating dysfunctional synapses. Retina specific downregulation also shows severely impaired optomotor behaviour. Finally, at an ultrastructural level BRP downregulation seems to impair the formation of the characteristic T-shaped synaptic ribbons at the active zones without significantly altering the overall synaptic architecture (in collaboration with E.Asan). Vertebrate active zone protein Bassoon is known to be involved in attaching the synaptic ribbons to the active zones as an adapter between active zone proteins RIBEYE and ERC/CAST. A mutation in Bassoon results in a floating synaptic ribbon phenotype. No protein homologous to Bassoon has been observed in Drosophila. BRP downregulation also results in absence of attached synaptic ribbons at the active zones. This invites the speculation of an adapter like function for BRP in Drosophila. However, while Bassoon mutant mice are viable, BRP deficit in addition to the structural phenotype also results in severe behavioural and physiological anomalies and even stronger downregulation causes embryonic lethality. This therefore suggests an additional and even more important role for BRP in development and normal functioning of synapses in Drosophila and also in other insects. However, how BRP regulates synaptic transmission and which other proteins are involved in this BRP dependant pathway remains to be investigated. Such studies certainly will attract prominent attention in the future.
Somites are repeated epithelial segments that are generated in a rhythmic manner from the presomitic mesoderm (PSM) in the embryonic tailbud. Later, they differentiate into skeletal muscle, cartilage and dermis. Somitogenesis is regulated by a complex interplay of different pathways. Notch/Delta signaling is one of the pathways well characterized in zebrafish through mutants affected in its different components. Previous work in mouse, chicken and zebrafish has shown that also additional components are required during somitogenesis, most importantly through an FGF and Retinoic acid (RA) gradient, as well as Wnt signaling. However, no zebrafish mutants with defects in these pathways showing specific somite malformations are described. This was explained by functional redundancies among related genes that have resulted from a whole genome duplication which occurred in a teleost fish ancestor 350 million years ago. As distinct duplicates exist in different teleost species, a large scale mutagenesis screen in the medaka (Oryzias latipes) has been performed successfully in Kyoto, Japan. I analyzed nine of the isolated medaka mutants that show variable aspects of somitic phenotypes. This includes a complete or partial loss of somite boundaries (e.g. bms and sne), somites with irregular sizes and shapes (e.g. krz and fsl) or partially fused and enlarged somites (e.g. dpk). Although some of these medaka mutants share characteristics with previously described zebrafish somite mutants, most of the mutants represent unique phenotypes, not obtained in the zebrafish screens. In-situ hybridization analyses with marker genes implicated in the segmentation clock (e.g. her7), establishment of anterior-posterior (A-P) polarity (e.g. mesp) and differentiation of somites (e.g. myf5, lfng) revealed that the medaka mutants can be separated into two classes. Class I shows defects in tailbud formation and PSM prepatterning, and lateron somite boundary formation was impaired in these mutants. A unique member of this class with a novel phenotype is the doppelkorn (dpk) mutant that has single fused or enlarged somites. This phenotype has not been reported till now in zebrafish somite mutants. In-situ analyses on dpk showed that stabilization of the cyclically expressed somitogenesis clock genes must be affected in this mutant. This is accompanied by a disrupted regulation of A-P polarity genes like mesp. This suggests that dpk is a mutant deficient in the wave front, which is necessary for the down-regulation of oscillating genes in the anterior PSM. Furthermore, as the initiation of oscillation of all three cyclic her genes was unaffected in dpk embryos, I could exclude that this mutant in affected in the Notch/Delta pathway. Another mutant that belongs to this class is the samidare (sam) mutant. Morphologically, sam mutants are similar to zebrafish after eight (aei). In both cases, the first 7-9 somites are formed properly, but after this somite formation ceases. Different to the situation in aei, sam mutant embryos presented an additional defect in the mid-hindbrain boundary (MHB) region. Similar MHB defects were described in the zebrafish fgf8 mutant acerebellar (ace). In ace zebrafish mutant, somites were only slightly defective, although FGF signaling has been shown to be important for somite formation in chicken, mouse and zebrafish. This was explained by functional redundancy between fgf8 and fgf24 ligands in the tailbud of zebrafish. Thus, it is interesting to suggest that the sam mutant, based on the parallel defects in somites and MHB, is a potential member of the FGF signaling pathway muatnts. It was shown that FGF plays a crucial role during MHB formation in medaka. In addition, I showed that fgf8 acts non-redundantly during tailbud formation and somitogenesis in medaka. Furthermore, I showed that FGF signaling regulates somite size also in medaka and that fgfr1 is the only FGF receptor expressed in the tailbud and somites. In class II medaka somite mutants, PSM prepatterning appears normal, whereas A-P polarity, boundary formation, epithelialization or the later differentiation of somites appears to be affected. Such mutants have not been isolated so far in zebrafish, mice or chicken. Therefore, medaka class II somite mutants seem to be a novel group of mutants that opens new perspectives to analyze A-P polarity regulation, determination and boundary formation in the presence of a normally functioning clock in the PSM. Identifying the encoding genes for all analyzed medaka somite mutants will contribute to the understanding of the molecular interactions of different signaling pathways involved during somitogenesis, and is expected to result in the identification of new components.
Rhodococcus equi is a Gram-positive intracellular pathogen which can cause severe bronchopneumonia in foals. In recent years, the role of this bacterium as human pathogen has been noted, as R.equi infections in humans have increase in frequency. This increase is associated with the rise in immunosupressed individuals, specially AIDS patients, where infection leads to symptoms and pathology similar to those seen in foals with a high mortality rate. Due to its capability to survive and multiply in murine and equine macrophages, R.equi has been classified as a facultative intracellular bacterium. R.equi is found frequently in macrophages in alveolar infiltrate from infected animals. The pathogenicity of R.equi depends on its ability to exist and multiply inside macrophages and has been associated with the presence of virulence plasmids. It has been observed that, inside foal alveolar macrophages, R.equi-containing vacuoles (RCVs) do not mature into phagolysosomes. However, most of the intracellular events during R.equi infection have not been investigated in detail. The aim of this study was to elucidate the intracellular compartmentation of R.equi and the mechanism by which the bacteria avoid destruction in host macrophages. The importance of the virulence-associated plasmids of R.equi for the establishment of RCVs was also evaluated. Furthermore, the intracellular fate of viable and non-viable R.equi was compared in order to study whether viability of R.equi influeciantes the establishment of RCVs. In this study, the RCV was characterized by using a variety of endocytic markers to follow the path of the bacteria trhough murine macropages. Transmission electron microscopy-base analysis showed that R.equi was found equally frequently in phagosomes with loosely or thightly apposed membranes, and RCV often contains numerous membranous vesicles. Laser scanning microscopy of infected macrophages showed that the majority of phagosomes containing R.equi acquired transiently the early endosomal markers Rab5, Ptlns3P, and EEA-1, suggesting initially undisturbed phagosome maturation. Although the RCV acquired some late endosomal markers, such as Rab7, LAMP-1, and Lamp-2, they did not acquired vATPase, did not interact with pre-labeled lysosomes, and failed to acidify. These data clearly suggest that the RCV is a compartment which has left vacuoles that resemble multivesicular body compartments (MVB), which are transport intermediates between early and late endosomes and display internal vesicles very similar to the ones observed within RCVs. Analyisis of several R.equi strains containing either VapA- or VapB-expressing plasmids or neither demonstrated that the possession of the virulence-associated plasmids does not affect phagosome trafficking over a two hour period of infection. The finding that non-viable R.equi was still able to inhibit phagosome maturation (although not to the same extent as viable R.equi did) suggests that heat-insensitive factors, such as cell periphery lipids, may play a major role in inhibition of phagosome maturation, although heat-sensitive factors may also be involved.
Diese Arbeit untersucht zelluläre Netzwerke mit dem Ziel, die so gewonnenen Einsichten medizinisch beziehungsweise biotechnologisch zu nutzen. Hierzu müssen zunächst Proteindomänen und wichtige regulatorische RNA Elemente erkannt werden. Dies geschieht für regulatorische Elemente in Nukleinsäuren am Beispiel von Iron Responsive Elements (IREs) in Staphylococcus aureus, wobei sich solche Elemente in viel versprechender Nähe zu exprimierten Sequenzen finden lassen (T. Dandekar, F. Du, H. Bertram (2001) Nonlinear Analysis 47(1): 225-34). Noch bedeutsamer als Ziele zur Medikamentenentwicklung gegen Parasiten sind Domänenunterschiede in Struktur und Sequenz bei Proteinen (T. Dandekar, F. Du, H. Bertram (2001) Nonlinear Analysis 47(1): 225-34). Ihre Identifikation wird am Beispiel eines potentiellen Transportproteins in Plasmodium falciparum exemplarisch dargestellt. Anschließend wird das Zusammenwirken von regulatorischen Elementen und Domänen in Netzwerken betrachtet (einschließlich experimenteller Daten). Dies kann einerseits zu allgemeineren Schlussfolgerungen über das Netzwerkverhalten führen, andererseits für konkrete Anwendungen genutzt werden. Als Beispiel wählten wir hier Redoxnetzwerke und die Bekämpfung von Plasmodien als Verursacher der Malaria. Da das gesamte Redoxnetzwerk einer lebenden Zelle mit Methoden der pH Wert Messung nur unzureichend zu erfassen ist, werden als alternative Messmethode für dieses Netzwerk Mikrokristalle der Glutathionreduktase als Indikatorsystem nach digitaler Verstärkung experimentell genutzt (H. Bertram, M. A. Keese, C. Boulin, R. H. Schirmer, R. Pepperkok, T. Dandekar (2002) Chemical Nanotechnology Talks III - Nano for Life Sciences). Um komplexe Redoxnetzwerke auch bioinformatisch zu modulieren, werden Verfahren der metabolischen Fluxanalyse vorgestellt und verbessert, um insbesondere ihrer Verzahnung besser gerecht zu werden und solche Netzwerke mit möglichst wenig elementaren Flussmoden zutreffend beschreiben zu können. Die Reduktion der Anzahl von Elementarmoden bei sehr großen metabolischen Netzwerken einer Zelle gelingt hier mit Hilfe unterschiedlicher Methoden und führt zu einer vereinfachten Darstellungsmöglichkeit komplexer Stoffwechselwege von Metaboliten. Dabei dient bei jeder dieser Methoden die biochemisch sinnvolle Definition von externen Metaboliten als Grundlage (T. Dandekar, F. Moldenhauer, S. Bulik, H. Bertram, S. Schuster (2003) Biosystems 70(3): 255-70). Allgemeiner werden Verfahren der Proteindomänenklassifikation sowie neue Strategien gegen mikrobielle Erreger betrachtet. In Bezug auf automatisierte Einteilung von Proteinen in Domänen wird ein neues System von Taylor (2002b) mit bekannten Systemen verglichen, die in unterschiedlichem Umfang menschlichen Eingriffs bedürfen (H. Bertram, T. Dandekar (2002) Chemtracts 15: 735-9). Außerdem wurde neben einer Arbeit über die verschiedenen Methoden aus den Daten eines Genoms Informationen über das metabolische Netzwerk der Zelle zu erlangen (H. Bertram, T. Dandekar (2004) it 46(1): 5-11) auch eine Übersicht über die Schwerpunkte der Bioinformatik in Würzburg zusammengestellt (H. Bertram, S. Balthasar, T. Dandekar (2003) Bioforum 1-2: 26-7). Schließlich wird beschrieben, wie die Pathogenomik und Virulenz von Bakterien der bioinformatischen Analyse zugänglich gemacht werden können (H. Bertram, S. Balthasar, T. Dandekar (2003) Bioforum Eur. 3: 157-9). Im letzten Teil wird die metabolische Fluxanalyse zur Identifikation neuer Strategien zur Bekämpfung von Plasmodien dargestellt: Beim Vergleich der Stoffwechselwege mit Glutathion und Thioredoxin in Plasmodium falciparum, Anopheles und Mensch geht es darum, gezielte Störungen im Stoffwechsel des Malariaerregers auszulösen und dabei den Wirt zu schonen. Es ergeben sich einige interessante Ansatzpunkte, deren medizinische Nutzung experimentell angestrebt werden kann.
Corynebacterium glutamicum is together with C. callunae and C. efficiens a member of the diverse group of mycolic-acid containing actinomycetes, the mycolata. These bacteria are potent producer of glutamate, lysine and other amino acids on industrial scale. The cell walls of most actinomycetes contain besides an arabinogalactan-peptidoglycan complex large amounts of mycolic acids. This three-layer envelope is called MAP (mycolyl-arabinogalactan-peptidoglycan) complex and it represents a second permeability barrier beside the cytoplasmic membrane similar to the outer membrane of Gram-negative bacteria. In analogy to the situation in the outer membrane of Gram-negative bacteria, channels are present in the mycolic acid layer of the mycobacterial cell wall for the passage of hydrophilic solutes. Molecular studies have provided far-reaching findings on the amino acid flux and its balance in C. glutamicum in general, but the L-glutamate export still remains unknown. The properties of the outer layers, typical of mycolata, seem to be of major importance in this process, and diffusion seems to play a key role for this part of the cell wall. The major aim of this thesis was to identify and study novel channel-forming proteins of the amino acid producers C. glutamicum, C. callunae and C. efficiens. Cell wall extracts of the organisms were investigated and a novel pore-forming protein, named PorH, that is homologue in all three organisms, was detected and characterized. PorHC.glut was isolated from C. glutamicum cells cultivated in minimal medium. The protein was identified in lipid bilayer experiments and purified to homogeneity by fast-protein liquid chromatography across a HiTrap-Q column. The purified protein forms cation-selective channels with a diameter of about 2.2 nm and an average single-channel conductance of about 2.5 nS in 1 M KCl in the lipid bilayer assay. Organic solvent extracts were used to study the permeability properties of the cell wall of C. callunae and C.efficiens. The cell extracts contained channel-forming activity, the corresponding proteins were purified to homogeneity by fast-protein liquid chromatography across a HiTrap-Q column and named PorHC.call and PorHC.eff. Channels formed by PorHC.call are cation-selective with a diameter of about 2.2 nm and an average single-channel conductance of 3 nS, whereas PorHC.eff forms slightly anion selective channels with an average single-channel conductance of 2.3 nS in 1 M KCl in the lipid bilayer assay. The PorH proteins were partially sequenced and the corresponding genes, which were designated as porH, were identified in the published genome sequence of C. glutamicum and C. efficiens. The chromosome of C. callunae is not sequenced, but PorHC.call shows a high homology to PorHC.eff and PorHC.glut. The proteins have no N-terminal extension, only the inducer methionine, which suggests that secretion of the proteins could be very similar to that of PorAC.glut of C. glutamicum. PorHC.glut is coded in the bacterial chromosome by a gene that is localized in the vincinity of the porAC.glut gene, within a putative operon formed by 13 genes that are encoded by the minus strand. Both porins are cotranscribed and coexist in the cell wall, which was demonstrated in RT-PCR and immunological detection experiments. The arrangement of porHC.glut and porAC.glut on the chromosome is similar to that of porBC.glut and porCC.glut and it was found that PorAC.glut, PorHC.glut, PorBC.glut and PorCC.glut coexist in the cell wall of C. glutamicum. The molecular mass of about 6 kDa of the PorH channel forming proteins is rather small and suggests that the cell wall channels are formed by oligomers. A possibly hexameric form was demonstrated for PorHC.glut in Western blot analysis with anti- PorHC.glut antibodies. Secondary structure predictions for PorHC.glut, PorHC.call and PorHC.eff predict that a stretch of about 42 amino acids of PorHC.glut and 28 amino acids of PorHC.call and PorHC.eff forms amphipathic -helices with a total length of 6.3 nm and 4.2 nm respectively. This should be sufficient to cross the mycolic acid layer. Another objective of this work was to establish an heterologous expression system for corynebacterial channel-forming proteins, to investigate the channel-forming properties of the up to now only hypothetical porins PorA, PorB, PorC from C. efficiens and PorC from C. glutamicum. We could demonstrate with recombinant expression experiments in E. coli that porBC.eff and porCC.eff encode for channel-forming proteins. They are, like PorBC.glut, anion-selective with a similar single-channel conductance of 1 nS in 1 M KCl.
Flagellar motility and chemotaxis are essential virulence traits required for the ability of Helicobacter pylori to colonize the gastric mucosa. The flagellar regulatory network and the complex chemotaxis system of H. pylori are fundamentally different from other bacteria, despite many similarities. In H. pylori expression of the flagella is controlled by a complex regulatory cascade involving the two-component system FlgR-HP244, the sigma factors 54 and 28 and the anti-sigma 28 factor FlgM. Thus far, the input signal for histidine kinase HP244, which activates the transcriptional regulator FlgR, which triggers sigma factor 54-dependent transcription of the flagellar class 2 genes, is not known. Based on a yeast two-hybrid screen a highly significant protein-protein interaction between the H. pylori protein HP137 and both the histidine kinase HP244 and the flagellar hook protein HP908 (FlgE´) has been reported recently (Rain et al., 2001). So far, no function could be assigned to HP137. Interestingly, the interaction between HP137 and histidine kinase HP244 was observed in the characteristic block N sequence motif of the C-terminal ATP-binding kinase domain. In this work a potential role of HP137 in a feedback regulatory mechanism controlling the activity of histidine kinase HP244 in the flagellar regulation of H. pylori was investigated. Although the substitution of the gene encoding HP137 by a kanamycin cassette resulted in non-motile bacteria, the failure to restore motility by the reintroduction of hp137 in cis into the mutant strain, and the observation that HP137 has no significant effect on the activity of histidine kinase HP244 in vitro indicated that HP137 is not directly involved in flagellar regulation. Therefore, it was demonstrated that HP137 does not participate in the regulation of flagellar gene expression, neither in H. pylori nor in the closely related bacterium C. jejuni. Chemotactic signal transduction in H. pylori differs from the enterobacterial paradigm in several respects. In addition to a CheY response regulator protein (CheY1) H. pylori contains a CheY-like receiver domain (CheY2) which is C-terminally fused to the histidine kinase CheA. Furthermore, the genome of H. pylori encodes three CheV proteins consisting of an N-terminal CheW-like domain and a C-terminal receiver domain, while there are no orthologues of the chemotaxis genes cheB, cheR, and cheZ. To obtain insight into the mechanism controlling the chemotactic response of H. pylori the phosphotransfer reactions between the purified two-component signalling modules were investigated in vitro. Using in vitro phosphorylation assays it was shown that both H. pylori histidine kinases CheAY2 and CheA´ lacking the CheY-like domain (CheY2) act as ATP-dependent autokinases. Similar to other CheA proteins CheA´ shows a kinetic of phosphorylation represented by an exponential time course, while the kinetics of phosphorylation of CheAY2 is characterized by a short exponential time course followed by the hydrolysis of CheAY2~P. Therefore, it was demonstrated that the presence of the CheY2-like receiver domain influences the stability of the phosphorylated P1 domain of the CheA part of the bifunctional protein. Furthermore, it was proven that both CheY1 and CheY2 are phosphorylated by CheAY2 and CheA´~P and that the three CheV proteins mediate the dephosphorylation of CheA´~P, although with a clearly reduced efficiency as compared to CheY1 and CheY2. Moreover, CheA´ is capable of donating its phospho group to the CheY1 protein from C. jejuni and to CheY protein from E. coli. Retrophosphorylation experiments indicated that CheY1~P is able to transfer the phosphate group back to the HK CheAY2 and the receiver domain present in the bifunctional CheAY2 protein acts as a phosphate sink fine tuning the activity of the freely diffusible CheY1 protein, which is thought to interact with the flagellar motor. Hence, in this work evidence of a complex phosphorelay in the chemotaxis system was obtained which has similarities to other systems with multiple CheY proteins. The role of the CheV proteins remain unclear at the moment, but they might be engaged in a further fine regulation of the phosphate flow in this complex chemotaxis system and the independent function of the two domains CheA´ and CheY2 is not sufficient for normal chemotactic signalling in vivo.
Das four-jointed (fj) Gen in Drosophila ist zum einen am proximo-distalen Längenwachstum der Extremitäten beteiligt, zum anderen spielt es auch eine Rolle in dem in neuerer Zeit verstärkt untersuchten planaren Zellpolaritätssignalweg (PCP-Signalweg). Über das in der Maus identifizierte homologe fjx1 Gen ist dagegen vergleichsweise wenig bekannt. Ziel dieser Arbeit war daher die nähere Charakterisierung von fjx1 sowie die Identifizierung möglicher Interaktionspartner. Durch RNA in situ Hybridisierung wurde zunächst das räumliche und zeitliche Expressionsmuster von fjx1 in Embryonen und adulten Organen untersucht. Dabei zeigte sich, dass fjx1 in allen Stadien vor allem im Gehirn, aber auch in epithelialen Strukturen verschiedener Organe exprimiert war. Obwohl die Expression von fjx1 ebenso wie die von fj über den Notch-Signalweg reguliert wird, konnte im Gegensatz zu Drosophila jedoch keine Regulation von fjx1 über den Wnt- und/oder den JAK/STAT-Signalweg nachgewiesen werden. Da Fj in Drosophila zumindest teilweise sezerniert wird und nicht-zellautomome Effekte zeigt, wurde ein Fjx1-Rezeptor gesucht. Mit Hilfe eines Fjx1-AP Fusionsproteins konnten Bindungsstellen überlappend bzw. angrenzend zu Regionen mit fjx1-Expression gefunden werden. Beispielsweise zeigten in der embryonalen Lunge und der Niere sowohl die in situ Hybridisierung (fjx1-Expression) als auch die Inkubation mit dem Fusionsprotein (Lokalisation des Bindungspartners) Färbung in epithelialen Strukturen, während im adulten Gehirn die Färbungen in jeweils benachbarten Schichten des Hippocampus und des Kleinhirns detektiert wurden. Durch Expressionsklonierung bzw. Coimmunpräzipitation konnte der Rezeptor jedoch nicht identifiziert werden. Aufgrund der Tatsache dass fj in Drosophila in enger Beziehung zu dachsous (ds) und fat (ft) steht, wurden die homologen Gene in der Maus gesucht und deren Expressionsmuster analysiert. In Embryonalstadien war dchs1 komplementär zu fjx1 in mesenchymalen Geweben zu finden, ähnlich der Situation in Drosophila, wo fj und ds in gegenläufigen Gradienten exprimiert sind. Das homologe Gen von ft, fat-j, war hingegen nicht ubiquitär exprimiert, sondern wie dchs1 im Mesenchym. Ergänzend dazu wurden die fat-like (ftl) Homologen, fat1-3, epithelial detektiert. Die Expression in adulten Organen wurde mit Real-Time-PCR untersucht, die zeigte, dass alle Gene (fj, ds und fat Homologe) relativ stark im adulten Gehirn zu finden sind. Mit Hilfe von RNA in situ Hybridisierungen konnten die Gene im Riechhirn, im Hippocampus und im Kortex des Großhirns sowie in der Körnerschicht des Kleinhirns lokalisiert werden. Um Hinweise auf die Funktion von Fjx1 zu erhalten, wurde in Datenbanken nach Proteinen mit ähnlicher Aminosäuresequenz gesucht, die eventuell Auskunft über mögliche Proteindomänen geben sollten. Bei den gefundenen fünf Mausproteinen handelte es sich jedoch um hypothetische bzw. noch nicht untersuchte Proteine, so dass Rückschlüsse auf die Funktion von Fjx1 nicht möglich waren. Die Expression dieser Gene war nach Datenbankangaben entweder sehr spezifisch, beschränkt auf ein bestimmtes Gewebe (z.B. Milchdrüse oder Nebenniere) oder schwach und dafür ubiquitär, was sich auch durch eine schwache, einheitliche Färbung in der RNA in situ Hybridisierung bestätigte. Die Proteinstruktur von Fjx1 und der Fjx1-ähnlichen Proteine sowie die Art der konservierten Reste geben Grund zu der Annahme, dass es sich um (sezernierte) Glykosyltransferasen handeln könnte, was durch die zumindest zeitweise Lokalisation von Fjx1 im Golgi-Apparat bestärkt wird. Auch die in Drosophila gefundenen Ergebnisse sprechen für eine derartige Funktion von Fj, obwohl auch hier noch keine konkreten biochemischen Belege vorliegen. Die Ergebnisse dieser Arbeit deuten auf eine Konservierung des in Drosophila entdeckten Fj/Ds/Ft-Siganlwegs in Vertebraten hin, wenn auch der genaue Mechanismus der Interaktion zwischen den Proteinen noch nicht geklärt ist und weiterer Untersuchungen bedarf.
Cloning and functional characterization of novel genes expressed preferentially in the human retina
(2005)
The human retina is a multi-layered neuronal tissue specialized for the reception and processing of visual information. The retina is composed of a great diversity of neuronal cell types including rod and cone photoreceptors, bipolar cells, ganglion cells, amacrine cells, horizontal cells and Müller glia. In response to light, a coordinated series of molecular events, the so-called phototransduction cascade, is triggered in photoreceptor cells and the signals from the photoreceptors are further processed by the bipolar and ganglion cells to the higher centers of the brain. The retina as highly complex system may be greatly susceptible to genetic defects which can lead to a wide range of disease phenotypes. Therefore, isolation and characterisation of the genes active in the human retina will facilitate our deeper understanding of retinal physiology and mechanisms underlying retinal degeneration and provide novel candidates for the retinal disease genes. To identify novel genes that are specifically or predominantly expressed in the human retina, a cDNA library enriched for retina specific transcripts was generated using suppression subtractive hybridization (SSH) technique. In total, 1113 clones were randomly isolated from the retina SSH cDNA library and partially sequenced. On the basis of BLASTN algorithm analysis these clones were classified into four categories including those with I) significant homology to known human genes (766/1113), II) significant homology to partial transcripts and hypothetical gene predictions (162/1113), III) no homology to known mRNAs (149/1113), and IV) vector sequences and clones derived from mitochondrial genes (36/1113). After correcting for redundancy, category I represented 234 known human genes and category II a total of 92unknown transcripts. Clones from category I, were selected for expression analysis by RT-PCR in a great number of human tissues. This resulted in the identification of 16 genes which were expressed exclusively in the retina, 13 which were highly expressed in the retina compared to other tissues, 12 genes which were specifically expressed in neuronal tissues and 48 ubiquitously expressed genes. Thus, our expression analysis resulted in the identification of 29 genes exclusively or abundantly transcribed in the human retina. Of those, retina specific genes L25,L33, L35, L37, L38 and L40 were selected for further analysis. To characterize the complete mRNA sequences of these transcripts a full-length human retina cDNA library was constructed. The analysis of the L25 gene revealed three splicing variants of the ABCC5 gene, consequently named ABCC5_SV1 (SV1), ABCC5_SV2 (SV2) and ABCC5_SV3 (SV3).These isoforms comprise the first five exons of ABCC5 and additional novel exons named 5a, 5b and 5c, generated by differential exon usage. The determined lengths of the three transcripts are 2039 bp, 1962 bp, and 1887 bp in size, respectively. RT-PCR, real-time PCR and Northern blot analysis of ABCC5 as well as the isoforms SV1, SV2 and SV3demonstrated high levels of expression for all transcripts in the retina compared to other tissues. Analysis of their nucleotide sequences revealed that inclusion of exon 5a in splicing variant SV1 produced a frame shift and premature termination codon (PTC). Our data show that this splice variant is the target of nonsense mediated mRNA decay (NMD). This was shown by inhibition of protein synthesis with antibiotics puromycin and anisomycin in human cell lines A-RPE 19 and Y79. Our analysis resulted in an increase of the PTC containing transcript and a decrease of the ABCC5 transcript. Conversely, the amount of both transcripts (SV1 and ABCC5) returned to pre-treatment levels after removal of the inhibitors. Together, our results suggest that alternative splicing of the ubiquitously expressed ABCC5 gene in addition to NMD is involved in retina-specific transcriptional regulation of the mRNA level of ABCC5. In contrast, additional experiments demonstrated that the levels of expression ofSV2 and SV3 isoforms do not appear to influence ABCC5 transcription. Several of the cloned genes were selected for additional genotyping of single nucleotide polymorphisms (SNPs) in order to construct their SNP maps which are going to be used for future association studies of complex disease AMD. Thus, identification of novel retinal genes and their functional characterization will further our elucidation of retinal physiology in general and in the diseased state in particular, by providing candidate retinal disease genes.
HMG-Proteine sind Architekturelemente des Chromatins und regulieren durch ihre Bindung an das Chromatin auf verschiedene Weise DNA-abhängige Prozesse wie Replikation, Transkription und DNA-Reparatur. Um zu verstehen, wie HMG-Proteine ihre vielfältigen Funktionen erfüllen können, wurde mit Hilfe von EGFP- und DsRed2-Fusionsproteinen ihre Funktion in vivo untersucht. Im Wesentlichen wurde dabei mit Hilfe von Bleichtechniken ihr dynamisches Verhalten charakterisiert. Daneben wurde für die HMGN-Proteine ihr bislang unbekanntes Expressionsverhalten in Tumorzellen bestimmt. So konnte für die HMGN-Proteine gezeigt werden, dass bestimmte Tumorzelllinien (HT-29, FTC-133, MCF-7, RPMI 8226, 697, Ishikawa, LNCap) eine relativ erhöhte Expression von HMGN2 aufweisen, die mit der Tumordifferenzierung korreliert. Eine relativ verringerte Expression von HMGN1 steht dagegen in Mammakarzinomen und Non-Hodgkin-Lymphomen in direktem Zusammenhang mit der Aggressivität der Tumore. Somit kann die HMGN-Expression bei diesen Tumoren als diagnostischer Marker verwendet werden. FRAP-Analysen mit EGFP-Fusionsproteinen führten zu der Erkenntnis, dass HMGN1, HMGN2, HMGA1a, HMGA1b und HMGB1 sich sehr schnell durch den Zellkern bewegen und nur transient an das Chromatin gebunden sind. Es konnte gezeigt werden, dass die spezifischen DNA/Chromatin-Bindungsmotive im Wesentlichen entscheiden, wo die Bindung der HMG-Proteine in vivo erfolgt, ihre Verweildauer im Euchromatin, Heterochromatin und zellzyklusabhängig dann aber durch Modifikationen (Phosphorylierungen, Acetylierungen) reguliert wird. Dies wurde beispielhaft durch punktmutierte und deletierte Fusionsproteine, sowie durch Inkubation der Zellen mit spezifischen Drogen für die HMGA1a-Proteine gezeigt. FRAP-Analysen haben außerdem gezeigt, dass die Spleißvarianten hHMGA1a und hHMGA1b unterschiedliche kinetische Parameter besitzen. Dies zeigt, dass beiden Varianten unterschiedliche Funktionen zugesprochen werden können. Die gefundenen spezifischen, transienten Verweildauern der einzelnen HMG-Proteine führen zu einem Modell eines dynamischen Chromatin-Netzwerkes, wobei alle HMG-Proteine in Wechselwirkungen innerhalb eines dynamischen Chromatinprotein-Cocktails DNA-abhängige Prozesse regulieren können. Die jeweiligen, wie hier gezeigt, durch Modifikationen regulierten Verweildauern der HMG-Proteine bestimmen darüber, welche anderen Chromatinproteine wie lange am Chromatin verbleiben und bestimmte Funktionen, wie beispielsweise die Modifikation der Core-Histone, übernehmen können. Die dynamischen Parameter einzelner HMG-Proteine erklären so, wie diese Proteine ihre vielfältigen Funktionen als Architekturelemente und bei der Regulation DNA-abhängiger Prozesse erfüllen können. Einige Vertreter, wie die HMGB1-Proteine, bewegen sich so schnell durch den Zellkern, dass ihre kinetischen Parameter durch das beschränkte zeitliche Auflösungsvermögen konfokaler Mikroskope der älteren Generation nicht erfassbar sind. Die Bestimmung von Dosis-Wirkungs-Beziehungen von Drogen, welche die kinetischen Parameter von HMGB1-Proteinen beeinflussen können, ist inzwischen mit Mikroskopen der neuen Generation möglich. Im Verlaufe der Arbeit zeigte sich, dass andere verwendete Fluorophore wie DsRed2 die kinetischen Eigenschaften von HMG-Fusionsproteinen beeinflussen können. Durch eine erhöhte Verweildauer können auch sehr transiente Interaktionen sichtbar gemacht werden. Wie gezeigt wurde, kann eine erhöhte Verweildauer aber auch zur Verdrängung anderer Proteine führen, die die gleichen Bindungsstellen benutzen und so eine Modulation des Chromatins bewirken. Die Nutzung von DsRed-Fluorophoren ermöglicht interessante neue Erkenntnisse. Diese müssen aber stets vor dem Hintergrund eines veränderten dynamischen Verhaltens der Fusionsproteine interpretiert werden. Zusammengenommen liefern die hier vorgestellten Ergebnisse zur Dynamik der HMG-Proteine grundlegende Informationen, die zur Klärung ihrer Funktion bei Chromatinmodulationen, etwa bei Differenzierungsprozessen oder der Entstehung von Tumorzellen entscheidend beitragen. Die Erkenntnis, dass diese Proteine lediglich transiente Interaktionen mit ihren Bindungspartnern eingehen können, sind im Hinblick auf die Behandlung von Tumoren, bei denen HMG-Proteine im Vergleich zu Normalgewebe häufig überexprimiert sind, von großer Bedeutung.
Listeria monocytogenes ist ein fakultativ humanpathogenes Bakterium und aufgrund seiner Fähigkeit, in Zellen des Wirtes einzudringen und sich im Zytoplasma der befallenen Wirtszelle zu vermehren, ein attraktiver Träger, um heterolog exprimierte Antigene in den MHC-I- und -II-Präsentationsweg antigenpräsentierender Zellen (APC) einzuschleusen und so eine effektive zelluläre Immunreaktion zu erzeugen. Dabei hat die Art und Weise der Antigenexpression einen wesentlichen Einfluss auf die Erzeugung der antigenspezifischen Immunität. So konnte unter Verwendung extrazellulärer Trägerbakterien gezeigt werden, dass insbesondere die Verankerung von Antigenen auf der Zelloberfläche der Bakterien zu einer effektiven Induktion einer humoralen Immunantwort führt. Mit dem Ziel, mit einem derartigen Ansatz auch eine zelluläre Immunität zu erzeugen, wurde ein Plasmidsystem für die Expression von heterologen Proteinen in der Zellwand von L. monocytogenes entwickelt. Dabei gelang die Verankerung zahlreicher Proteine eukaryontischer wie auch prokaryontischer Herkunft über das LPXTG-Ankermotiv von Internalin A. Die so erzielte starke Expression in der Zellwand setzte aber sowohl die Fitness der Bakterien als auch deren Invasivität in vitro deutlich herab und verhinderte damit eine effektive MHC-I-Präsentation des verwendeten Modellantigens. Alternativ wurde L. monocytogenes bereits erfolgreich zur Übertragung von DNA-Vakzinen eingesetzt, um so auch die Synthese gegebenenfalls posttranslationell modifizierter Antigene in ihrer korrekten Konformation durch die infizierte Wirtszelle zu erzielen. Allerdings erfolgte die Expression des als Reporterprotein verwendeten EGFP insbesondere in APC sehr langsam und war von geringer Effizienz. Dabei konnte in der vorliegenden Arbeit erstmals gezeigt werden, dass nach bakterieller Übertragung von DNA-Vakzinen der Import der Plasmidmoleküle in den Kern insbesondere sich nicht teilender Zellen einen der wichtigsten Engpässe für eine möglichst frühzeitige und effektive Reportergenexpression darstellt. Einen Ausweg bietet die bakterielle Übertragung codierender mRNA, die unmittelbar nach der Freisetzung aus der Bakterienzelle im Zytoplasma der infizierten Wirtszelle zur Translation zu Verfügung steht. Dazu wurde das 5’-Ende der EGFP-codierenden Sequenz mit dem IRES-Element des Encephalomyocarditisvirus genetisch fusioniert. Um eine möglichst hohe Syntheserate zu erzielen und damit dem Abbau der mRNA in der Bakterienzelle entgegenzuwirken, erfolgte die Synthese der mRNA in L. monocytogenes mit Hilfe eines T7-RNA-Polymerase-basierten Transkriptionssystems. Im Gegensatz zur bakteriellen Übertragung von Plasmid-DNA konnte so bereits 4 h nach Infektion sowohl in epithelialen als auch in APC wie Makrophagen und humanen dendritischen Zellen eine deutliche EGFP-Expression nachgewiesen werden sowie bei Verwendung von Ovalbumin als Reporterprotein eine effektive MHC-I-Präsentation in vitro. Damit stellt die bakterielle Übertragung von mRNA einen vielversprechenden neuartigen Ansatz dar zur Erzeugung einer zellulären und gegebenenfalls auch humoralen Immunantwort gegen posttranslational modifizierte Antigene.