No description available.

La particule de reconnaissance du signal (signal recognition particle-SRP) et son récepteur (FtsY chez Escherichia coli) médiatise le processus simultané de traduction-ciblage de la protéine en dirigeant le complexe ribosome-nascent chain (RNCs) vers la membrane de destination. La reconnaissance par la SRP d'une charge RNC à transporter dépend de la présence de la partie N-terminale. L'assemblage de Ftsy au complexe RNC-PRS entraine plusieurs changements de configuration de SRP et de FtsY durant le cycle de direction. D'abord un stade « précoce » sans GTP est adopté. Celui-ci est stabilisé par le RNC. Ensuite une configuration « fermée » avec GTP est formée. Cette dernière peut s'activer pour hydrolyser GTP, elle entre alors dans sa configuration « active ». La succession de ces trois étapes conduit à la libération du complexe SRP-récepteur d'avec le ribosome et de sa protéine en cours de traduction, et leur mise à disposition au pore de la membrane. Dans ce projet, notre intérêt se limite à la traduction par le ribosome de la séquence signale EspP (RNCEspP). In vivo, EspP est une protéine dont le ciblage vers le récepteur membranaire se réalise après la traduction. Cependant il arrive que RNCEspP se lie au complexe SRP-FtsY, faisant échouer le ciblage. Nous avons étudié les bases structurales du rejet de RNCEspP par SRP et FtsY. Pour cela nous avons effectué la comparaison de la structure RNCEspP-SRP-FtsY obtenue par observation au cryo-microscope électronique avec d'autres complexes ribosome-SRP-récepteurs traduisant la charge FtsQ, qui est elle normalement ciblé par SRP. Nous avons cherché à observer la différence de structure entre les complexes SRP-FtsY dans les deux cas. Deux différences majeurs entre les complexes de ciblages contenants les séquences RNCFtsQ et RNCEspP ont été observés. Premièrement, dans le cas de la structure de RNCEspP le domaine M -Ffh est attaché à l'hélice 59 du ribosome, alors que celui-ci est détaché dans le cas de la structure de RNCFtsQ. Nous pensons que le domaine M empêche la libération de la séquence de signal, étape nécessaire à la réalisation du ciblage. Deuxièmement, dans le cas de la structure du complexe avec RNCEspP l'arrangement Ffh-FtsY avec le domaine NG était flexible. Ceci indiquerait que le complexe “précoce” formé sur RNCEspP est moins stable que celui formé sur RNCFtsQ. Une étude biochimique utilisant le transfert d'énergie via résonance fluorescente a corroboré ce résultat, montrant que FTS Y est lié avec une affinité moindre dans le cas du complexe précoce formé sur RNCEspP et que la reconfiguration au stade de complexe fermé est moins efficace. Une analyse biochimique plus poussée des variantes de la séquence de EspP montre que la partie N-Terminale de la séquence est la principale cause de rejet du cycle de ciblage via SRP.Dans un second projet, nous avons étudié la configuration “fermée” de SRP et ftsY en complexe avec une charge RNC stabilisée par un analogue non-hydrolysable de GTP (GMP-PCP). Pour franchir la barrière cinétique qui permet de passer du complexe précoce au complexe fermé, nous avons utilisé une version tronquée de FtsY, à laquelle la séquence terminale avait été amputée de tout le domaine acide (A-) ainsi que de la première hélice alpha du domaine NG. De plus, pour la formation du complexe, nous avons utilisé une construction contenant les 50 premiers acides aminés du leader peptidase (RNCLep50). En l'absence de nucléotides, notre reconstruction au cryo-EM a montré une configuration similaire à celle du stade précoce, dans laquelle Ftsy et Ffh- domaine NG, sont proche du tetraloop de la 4.5 S ARN. Une incubation avec GMP-PCP induit un détachement du domaine NG d'avec la queue du tetraloop. Il semblerait que les domaines NG soient flexibles dans l'état clos, et non attaché à la terminaison ouverte de l'ARN. / The signal recognition particle (SRP) and its receptor (FtsY in Escherichia coli) mediate co-translational protein targeting by delivering ribosome nascent chain complexes (RNCs) to the target membrane. Recognition of an RNC cargo by SRP is dependent on an N-terminal signal sequence. Binding of FtsY to the RNC-SRP complex leads to several conformational changes of SRP and FtsY during the targeting cycle: first, an “early” GTP-independent state is adopted which is stabilized by the RNC, subsequently a “closed” GTP- dependent conformation is formed which can activate itself to hydrolyze GTP (the “activated” state). Faithful completion of all three steps leads to release of the cargo from SRP-FtsY and hand over of the RNC to the translocation pore.It has been shown for E. coli that cargos can be rejected from the SRP pathway during all targeting steps. In the first project, our interest concentrates on ribosomes translating the EspP signal sequence (RNCEspP). In vivo, EspP is a post-translationally targeted protein, but RNCEspP has been shown to be bound by SRP and FtsY leading to a non-productive “early”-like RNCEspP-SRP-FtsY complex. Using single particle cryo-electron microscopy (EM), we analysed the structural basis for the rejection of RNCEspP by SRP and FtsY. Comparison of our RNCEspP-SRP-FtsY cryo-EM structure to other available cryo-EM structures of co-translational targeting complexes containing the correct cargo RNCFtsQ unravelled differences in the SRP-FtsY structure between a correct cargo and an incorrect cargo. Two major differences between the targeting complexes containing the cargos RNCFtsQ and RNCEspP were observed: first, the Ffh M-domain was attached to ribosomal RNA helix 59 of RNCEspP, while it was detached from this site in the case of RNCFtsQ. It could be that such an ordered M-domain is hampering the release of the signal sequence which is required for successful completion of targeting. Second, the Ffh-FtsY NG-domain arrangement was flexible in the complex with RNCEspP in comparison to RNCFtsQ indicating that the "early"-like complex formed on RNCEspP is less stable. Biochemical data using fluorescence resonance energy transfer corroborated these results, showing that FtsY is bound with lower affinity in the RNCEspP “early” complex and that the rearrangement to the “closed” conformation is less efficient. Further biochemical analysis of EspP signal sequence variants showed that mainly the N-terminal extension of the EspP signal sequence is responsible for its rejection from the SRP pathway.

Cerebral small vessel disease (SVD) is a vascular neurodegenerative disease which is the leading cause of vascular dementia and causes 20% of strokes. 20-30% of those over 80 show signs of the disease as white matter hyperintensities on MRI scans, doubling their risk of stroke and trebling their risk of dementia. Sporadic SVD is thought to be caused by hypertension but 30% of sufferers are normotensive and an alternative hypothesis implicates loss of integrity of the blood brain barrier (BBB). To investigate this, I studied brains from normotensive people with early stage SVD and found reduced capillary endothelial claudin-5 (a BBB tight junction protein), more oligodendrocyte precursor cells (OPCs; the precursors to myelinating oligodendrocytes), and more microglia/macrophages compared to controls. Furthermore, in a relevant rat model of spontaneous SVD, the Stroke Prone Spontaneously Hypertensive Rat (SHRSP; disease model; DM) I found that reduced endothelial claudin-5 was the earliest change, appearing at 3 weeks of age, followed by OPC proliferation, appearing at 4 weeks, and then increased number of microglia/macrophages, appearing at 5 weeks. Importantly, all these changes occurred at a young age (< 5 weeks), before any measurable hypertension. These changes were confirmed in an ex vivo slice culture model (i.e. removing blood flow), ruling out direct damage by leakage of blood components through an impaired BBB and suggesting an inherent endothelial cell dysfunction as the primary cause, with secondary BBB defects. This hypothesis of endothelial dysfunction is supported by increased endothelial cell proliferation in both human SVD tissue and the DM rats, and lower levels of endothelial nitric oxide synthase (eNOS) in brains of DM rats. To study this further I isolated primary brain microvascular endothelial cells (BMECs) from DM and control rats and found that those from DM rats formed less mature tight junctions (less membranous claudin-5) than control BMECs. I also found that conditioned media (CM) from DM BMECs causes OPCs in culture to proliferate more and mature less. This indicates that the endothelial dysfunction is inherent to the endothelial cells, rather than induced by other cell types, and through secreted factors causes OPC changes mirroring what is seen in vivo. Using an antibody array, I identified HSP90α as a candidate secreted factor and showed that it is necessary (by blocking the protein in CM) and sufficient (by adding recombinant HSP90α) to induce the maturation phenotype in OPCs, but not the proliferation phenotype. The idea that endothelial dysfunction causes SVD begs the question of what causes endothelial dysfunction, especially in our inbred DM rat strain. To establish this, I reanalysed sequencing data of the DM and control rats from a previously published study, searching for mutations which lead to truncated proteins in genes expressed in brain endothelial cells. We confirmed the candidate gene Atp11b, a phospholipid flippase, was mutated as predicted. I found that knocking down Atp11b using siRNA in a control endothelial cell line caused endothelial dysfunction and a loss of tight junction maturity, and that CM from these cells causes OPCs to proliferate more and mature less, mirroring what we see in primary DM BMECs and suggesting that Atp11b has a key function in promoting normal endothelial function. Furthermore, I showed that knocking down Atp11b causes cells to secrete increased levels of HSP90α. I propose a mechanism whereby ATP11B regulates the retention of HSP90α within endothelial cells, which in turns regulates eNOS levels and activity, as has been shown previously. In summary, this work shows that there are many pre-symptomatic changes which occur in the brain in the development of SVD in DM rats, and that these are ultimately caused by endothelial dysfunction. As these changes are similar to those found in spontaneous human SVD, I propose that endothelial dysfunction is a key mechanism of human SVD, which may in the future lead to new therapies.

Bacterial infections are again becoming difficult to treat because the microbes are growing increasingly resistant to the antibiotics in use today. The need for novel antimicrobial compounds is urgent and to achieve this new targets are crucial. In this thesis we present a strategy for identification of such targets via a bioinformatics approach. In our first study we compared proteins with unknown and hypothetical function of the spirochete Treponema pallidum to five other pathogens also causing chronic or persistent infections in humans (Yersinia pestis, Neisseria gonorrhoeae, Helicobacter pylori, Borrelia burgdorferi and Streptococcus pneumoniae). T. pallidum was used as a starting point for the comparisons since this organism has a condensed genome (1.1 Mb). As we aimed at identifying conserved proteins important for in vivo survival or virulence of the pathogens we reasoned that T. pallidum would have deleted genes not important in the human host. This comparison yielded 17 ORFs conserved in all six pathogens, these were deleted in our model organism, Yersinia pseudotuberculosis, and the virulence of these mutant strains was evaluated in a mouse model of infection. Five genes were found to be essential for virulence and thus constitute possible antimicrobial drug targets. We have studied one of these virulence associated genes (vags), vagH, in more detail. Functional and phenotypic analysis revealed that VagH is an S-adenosyl-methionine dependent methyltransferase targeting Release factor 1 and 2 (RF1 and RF2). The analysis also showed that very few genes and proteins were differentially expressed in the vagH mutant compared to wild-type Yersinia. One major finding was that expression of the Type III secretion system effectors, the Yops, were down regulated in a vagH mutant. We dissected this phenotype further and found that the down regulation was due to lowered amounts of the positive regulator LcrF. This can be suppressed either by a deletion of yopD or by over expression of the Ribosomal Recycling Factor (RRF). These results indicate that YopD in addition to its role in translational regulation of the Yops also plays a part in the regulation of LcrF translation. We suggest also that the translation of LcrF is particularly sensitive to the amount of translation competent ribosomes and that one effect of a vagH mutation in Y. pseudotuberculosis is that the number of free ribosomes is reduced; this in turn reduces the amount of LcrF produced thereby causing a down regulation of the T3SS. This down regulation is likely the cause of the attenuated virulence of the vagH mutant. Finally, we set up a high throughput screening assay to screen a library of small molecules for compounds with inhibiting the VagH methyltransferase activity. Five such compounds were identified and two were found to inhibit VagH also in bacterial culture. Furthermore, analogues to one of the compounds showed improved inhibitory properties and inhibited the T3SS-dependent cytotoxic response induced by Y. pseudotuberculosis on HeLa cells. We have successfully identified five novel targets for antimicrobial compounds and in addition we have discovered a new class of molecules with antimicrobial properties.