A key prerequisite for cellular and organismal health is a functional proteome. A variety of human protein misfolding diseases are associated with the occurrence of amyloid protein aggregates, such as amyotrophic lateral sclerosis (ALS) or Huntington’s disease. The proteins involved in disease manifestation all contain aggregation-prone sequences of low compositional complexity. Such sequences are also known as prion-like, because of their sequence similarity to yeast prions. Yeast prion proteins are a specific subset of amyloid forming proteins with distinct physio-chemical and functional features, which give them transmissible properties. The aggregation properties of yeast prions and disease-related prion-like proteins reside in structurally independent, prion-forming domains (PrDs). These domains are highly enriched for uncharged polar amino acids, such as glutamine (Q) and asparagine (N). These compositional features can be used to predict prion-like proteins bioinformatically. To investigate the prevalence of prion-like proteins across different organisms, we analyzed a range of eukaryotic proteomes. Our analysis revealed that the slime mold D. discoideum contains the highest number of prion-like N/Q-rich proteins of all organisms. Based on this finding, we hypothesized that D. discoideum could be a valuable model system to study protein homeostasis (proteostasis) and the molecular basis of protein misfolding diseases. To explore how D. discoideum manages its highly aggregation-prone proteome, we analyzed the behavior of several well-characterized misfolding-prone marker proteins (variants of the disease-causing exon 1 of the huntingtin protein as well as wildtype and variant versions of the Q/N-rich yeast prion Sup35NM). Intriguingly, these proteins did not form cytosolic aggregates in D. discoideum, as they do in other organisms. Aggregates, however, formed as a result of heat stress, which indicates that the tested proteins have the capacity to aggregate, but are kept under tight control under normal conditions. Furthermore, when the stress level was reduced, the stress-induced aggregates dissolved, suggesting that D. discoideum has evolved mechanisms to reverse aggregation after a period of acute stress. Together, these findings reveal an unusual resilience of D. discoideum to aggregation-prone proteins, which very likely results from specific adaptations in its proteostasis network. By studying these specific adaptations, we could get important insight into the strategies that nature employs to control and maintain a highly aggregation-prone proteome. So far, our experimental investigations have revealed evidence for three specific adaptations. First, we identified the disaggregase Hsp101 as a key player in the acute stress response of D. discoideum. A functional analysis of Hsp101 in yeast and D. discoideum revealed that it supports thermotolerance. Second, we found evidence for an important role of the nucleus and nucleolus in proteostasis. We discovered that a small fraction of highly aggregation-prone proteins accumulated in the nucleus or nucleolus of D. discoideum cells. The magnitude of this nuclear accumulation could be increased by proteasome impairment, which suggests that the ubiquitin-proteasome system (UPS) is involved. This finding is consistent with previous studies in other organisms and hints at the possibility that D. discoideum disposes of aggregation-prone proteins by degrading them in the nucleus/nucleolus. Third and finally, we found that cells containing nuclear accumulations are asymmetrically distributed in the multicellular developmental stage (slug), suggesting that D. discoideum employs cell-sorting mechanisms to dispose of cells with accumulated protein damage. Although our current understanding of proteostasis in D. discoideum is preliminary, we have gained important insight into the molecular mechanisms and cellular pathways that D. discoideum uses to counteract protein aggregation. Findings from this work will inform similar comparative studies in other organisms and will impact our molecular understanding of protein misfolding diseases and aging. / Eine wesentliche Voraussetzung für die Gesundheit von Zellen und Organismen ist ein funktionales Proteom. Eine Reihe von humanen Protein- Missfaltungs-Erkrankungen, wie Chorea Huntington und Amyotrophe Lateralsklerose (ALS) werden mit dem Auftreten von amyloiden Protein- Aggregaten in Verbindung gebracht. Sämtliche Proteine, die in der Pathogenese dieser Krankheiten eine Rolle spielen, enthalten aggregations-anfällige Sequenzen mit geringer Sequenzkomplexität. Solche Sequenzen werden als Prion-ähnlich bezeichnet, da sie in ihrer Zusammensetzung den Prionen aus der Hefe S. cerevisiae gleichen. Die Prion-Proteine der Hefe gehören zu einer Unterart von amyloid-aggregierenden Proteinen, die durch bestimmte physikochemische und funktionelle Eigenschaften einen infektiösen Charakter erhalten. Die Aggregations-Eigenschaften von Hefeprionen und aggregationsanfällige Proteinen, die mit Erkrankungen in Verbindung gebracht werden, basieren auf strukturell unabhängigen, Prion-bildenden Domänen (prion domain, PrD). Diese Domänen sind angereichert mit polaren Aminosäuren wie Glutamin und Asparagin. Diese Zusammensetzung kann dazu verwendet werden prion-ähnliche Proteine bioinformatisch vorherzusagen. Um die Verbreitung von Prion-ähnlichen Proteinen in verschiedenen Organismen zu untersuchen, analysierten wir eine Reihe von eukaryotischen Proteomen. Unsere Analyse zeigte, dass der Schleimpilz D. discoideum die höchste Anzahl von Prion-ähnlichen N/Q-reichen Proteinen aufzeigt. Aufgrund dieser Erkenntnisse erstellten wir die Hypothese, dass D. discoideum ein nützlicher Modellorganismus sein könnte, um Protein Homöostase (Proteostase) sowie die molekulare Basis von Proteins-Missfaltungs-Erkrankungen zu ergründen. Um zu analysieren, wie D. discoideum mit seinem höchst aggregations-anfälligen Proteom umgehen kann, untersuchten wir das Verhalten mehrerer bereits charakterisierter aggregations-anfälliger Marker-Proteine in D. discoideum. Hierbei verwendeten wir Varianten des krankheits-erzeugenden Exon 1 des humanen Huntingtin Protein sowie den wild-typ und Varianten des N/Q-reichen Hefe Prions Sup35. Interessanterweise bildeten diese Proteine, anders als in anderen Organismen, keine zytosolischen Aggregate in D. discoideum aus. Aggregate wurden jedoch unter Hitzestress-Bedingungen gebildet. Dies deutet darauf hin, dass die getesteten Proteine durchaus das Vermögen zu aggregieren besitzen, jedoch unter normalen Wachstumsbedingungen streng kontrolliert werden. Wenn, darüberhinaus das Stress- Level gesenkt wurde, kam es zur Auflösung der stress-induzierten Aggregate. Dies deutet darauf hin, dass D. discoideum Mechanismen entwickelt hat, um Aggregate nach Perioden von akutem Stress wieder aufzulösen. Zusammengenommen enthüllen diese Erkenntnisse eine ungewöhnliche Widerstandsfähigkeit gegenüber aggregations-anfälligen Proteinen. Diese beruht höchstwahrscheinlich auf spezifischen Modifikationen im Proteostase Netzwerk. Durch die Analyse dieser spezifischen Anpassungen könnten wichtige Einblicke in die Strategien gewährt werden, welche die Natur benutzt, um ein höchst aggregations-anfälliges Proteom zu erhalten und zu kontrollieren. Bisher erbrachten unsere Experimente Anhaltspunkte für drei spezifische Anpassungen. Erstens zeigten wir, dass die Disaggregase Hsp101 eine Schlüsselrolle in der akuten Stressantwort in D. discoideum einnimmt. Eine funktionale Analyse von Hsp101 in D. discoideum und Hefe zeigte, dass die Disaggregase Thermotoleranz fördert. Zweitens haben wir Anhaltspunkte, dass der Nukleus und der Nukleolus eine wichtige Rolle in der Proteostase einnehmen. Eine geringe Fraktion der überaus aggregations-anfälligen Proteine akkumuliert im Nukleus oder Nukleolus von D. discoideum. Das Ausmaß der nuklearen Akkumulation konnte erhöht werden, wenn das Proteasom beeinträchtigt wird. Dies deutet darauf hin, dass das Ubiquitin-Proteasom-System involviert sein könnte. Diese Beobachtung ist im Einklang mit jüngsten Berichten aus anderen Organismen und daraus folgt, dass D. discoideum möglicherweise aggregations-anfällige Proteine durch Abbau im Nukleus entsorgt. Drittens konnten wir feststellen, dass Zellen, die nukleare Akkumulationen enthalten, asymmetrisch in der multizellulären Entwicklungs-Struktur des Pseudoplasmodiums verteilt sind. Dies deutet darauf hin, dass D. discoideum möglicherweise den Zellsortierungsmechanismus während der Entwicklung nutzen kann, um Zellen mit angereicherten Protein-Schäden zu beseitigen. Auch wenn das gegenwärtige Verständnis der Proteostase in D. discoideum nur vorläufig ist, haben wir wichtige Einblicke in die molekularen Mechanismen und zellulären Prozesse erhalten, die D. discoideum verwendet, um Protein-Aggregation zu verhindern. Die Ergebnisse dieser Arbeit werden ähnliche vergleichende Studien in anderen Organismen beeinflussen und Auswirkungen auf unser molekulares Verständnis über Protein-Missfaltungs-Erkrankungen und das Altern haben.

A serina/treonina fosfatase do tipo 1 (PP1) é uma enzima ubíqua nas células e nos tecidos das várias espécies em que foi pesquisada e regula vários processos como metabolismo intermediário, processamento de mRNA, transcrição e apoptose. Geralmente a holoenzima PP1 é encontrada como um dímero constituído por uma subunidade catalítica conservada (PP1c) e uma ou mais subunidades reguladoras variáveis. Em mamíferos, já foram identificados mais de 50 polipeptídeos que se associam direta ou indiretamente a PP1c, gerando holoenzimas com localizações celulares e especificidades distintas. Entre estas proteínas estão inibidores citosólicos de PP1c, tais como o inibidor-1 (I-1), o inibidor-2 (I-2) e o inibidor nuclear da PP1 (NIPP-1). Ortólogos do I-2 foram descritos em microorganismos como Saccharomyces cerevisiae e Neurospora crassa. Neste trabalho nós demonstramos que o genoma da ameba social Dictyostelium discoideum possui uma única cópia do gene que codifica um ortólogo do I-2 (DdI-2). Análise através de Northern blot mostrou que o mRNA de DdI-2 é expresso durante o crescimento e ao longo de todo o ciclo de desenvolvimento, com níveis variáveis. Também demonstramos que o gene DdI-2 codifica uma verdadeira proteína inibidora da PP1c uma vez que seu produto recombinante em bactéria é capaz de inibir, com eficácia equivalente, as atividades de fosforilase fosfatase das PP1c recombinantes selvagem (DdPP1c) e mutante (DdPP1cF269C) de D. discoideum e NcPP1c de N. crassa in vitro. A proteína DdPP1cF269C apresenta características distintas da DdPP1c incluindo maior estabilidade, maior atividade de fosforilase fosfatase e maior sensibilidade frente ao inibidor caliculina A. Estas diferenças devem-se a substituição da cisteína conservada da posição 269 por uma fenilalanina, que é verificada na enzima selvagem. DdPP1c e DdPP1cF269C foram também ensaiadas na presença de INc-1L e INc-1 que são ortólogos de I-2 em N. crassa. Ambas as proteínas recombinantes purificadas exibiram efeito inibidor sobre a atividade de fosforilase fosfatase das DdPP1c recombinantes selvagem e mutante, sendo que INc-1 foi um inibidor duas vezes mais eficiente que INc-1L. Este efeito pode ser devido a um segmento de 38 aminoácidos codificado por um íntron em fase que é retido na isoforma INc-1L. Nossos dados indicam ainda que a mutação F269C não afetou a sensibilidade da DdPP1c recombinante a nenhum dos ortólogos de I-2 testados in vitro. Ensaios de duplo-híbrido utilizando a PP1c selvagem e mutante de D. discoideum (DdPP1c e DdPP1cF269C) e de N. crassa (NcPP1c) como iscas e DdI-2 como presa mostraram que estas proteínas interagiram in vivo. Quando a presa era o INc-1L ou INc-1 a interação ocorreu apenas com a NcPP1c, sendo mais forte no caso de INc-1. As regiões de DdI-2 envolvidas na interação física com a DdPP1c foram mapeadas através da expressão de proteínas truncadas no ensaio de duplo híbrido. Os experimentos apontaram que o carbóxi-terminal de ~100 aminoácidos não é essencial para a interação, mas que o somatório das diversas regiões responde pela integridade da interação. / The serine/threonine phosphatase of type-1 (PP1) is a ubiquous enzyme in the cells and tissues from several species studied and regulates numerous processes such as intermediate metabolism, mRNA splicing, transcription, and apoptosis. PP1 holoenzymes consist of a well-conserved catalytic subunit (PP1c) and one or more variable regulatory subunits. In mammals, more than fifty polypeptides that bind PP1c have been identified, originating holoenzymes with distinct cell locations and specificities. These proteins include cytosolic PP1c inhibitors such as inhibitor-1 (I-1), inhibitor-2 (I-2) and nuclear inhibitor of PP1 (NIPP-1). I-2 orthologs have also been described in Saccharomyces cerevisiae and Neurospora crassa. In the present work, we demonstrate that the genome of the social amoeba Dictyostelium discoideum has a single gene encoding for an I-2 ortholog (DdI-2). Northern blot analyses have shown that DdI-2 mRNA is expressed throughout Dictyostelium developmental cycle at variable levels. We also demonstrated that DdI-2 is a true PP1c inhibitor as its recombinant product is capable of inhibiting the phosphorylase phosphatase activity of wild-type PP1c (DdPP1c) and mutant (DdPP1cF269C) of D. discoideum and NcPP1c of N. crassa in vitro. DdPP1cF269C protein presents distint traits including higher stability, phosphorylase phosphatase activity and sensibility to calyculin A than the wild-type. These differences are originated from the replacement of a well conserved cisteine residue by a phenylalanine found in the wild-type. The wild-type and mutant DdPP1c have also been assayed in the presence of INc-1L and INc-1 which are orthologues to I-2 in N. crassa. Both purified recombinant proteins have shown inhibitory effects over phosphorylase phosphatase activities, with INc-1 being twice more potent than INc-1L. This might be due to the presence of an intron retention event in the latter that results in a insertion of 38 aminoacids. Our data also indicate that F269C mutation did not affect DdPP1c sensitivity to inhibition by all the three recombinant I-2 orthologues in vitro. Yeast two-hybrid assays using wild type (DdPP1c) and mutant (DdPP1cF269C) D. discoideum and N. crassa (NcPP1c) PP1c as preys and the putative inhibitor DdI-2 as a bait showed inequivocally that these proteins interacted in vivo. When the prey was INc-1 or INc-1L the interaction occured only with NcPP1c and was stronger with INc-1. The domains of DdI-2 involved in the interaction with DdPP1c were mapped by two-hybrid interaction assays with DdI-2 deleted mutants. These experiments have pointed out that the DdI-2 carboxi-terminus of ~100 aminoacids is not essential for the interaction but that the sum of all regions is responsible for the integrity of the interaction.

As proteínas fosfatases são enzimas responsáveis pela desfosforilação de resíduos de fosfoaminoácidos, principalmente fosfotirosina e fosfoserina/treonina, o que divide esta classe de proteínas nas famílias PTP (proteína tirosina fosfatase) e PP (proteína serina/treonina fosfatase). Diversos membros da família das PPs, em particular da subfamília PPP (Phosphoprotein Phosphatase), existem como holoenzimas compostas de uma subunidade catalítica associada a uma ou mais subunidades reguladoras, que lhes conferem diversidade funcional. Neste trabalho tivemos como objetivo identificar, utilizando ensaios no sistema de duplo híbrido em leveduras, proteínas que interagem com as subunidades catalíticas das serina/treonina fosfatases do tipo 1 (DdPP1c) e do tipo 4 (DdPP4c) da ameba social D. discoideum. Varreduras de bibliotecas de cDNA das fases de crescimento e desenvolvimento da linhagem AX4 de D. discoideum com as iscas DdPP1c e uma variante mutante da DdPP1c (DdPP1cF269C) possibilitaram a identificação de pelo menos 30 genes com evidência de interação com a PP1c. A varredura das bibliotecas com a isca DdPP4 propiciou a identificação de 10 potenciais genes candidatos com evidência de interação com a PP4c. Várias dos candidatos identificados nas varreduras correspondem a genes que codificam para proteínas hipotéticas que não apresentam similaridade significativa com proteínas de função conhecida. Entre essas, identificamos e caracterizamos DdI-3, um ortólogo do inibidor-3 da PP1c de mamíferos. A interação de DdI-3 com DdPP1c foi confirmada através de ensaios independentes no sistema do duplo híbrido em leveduras. Demonstramos que DdI-3 recombinante expresso em bactérias possui atividade inibidora da DdPP1c in vitro, sendo que esta enzima tem 50% de sua atividade de fosforilase fosfatase inibida por cerca de 0,55 nM de rDdI-3. Estes dados indicam que DdI-3 é 50 vezes mais potente do que DdI-2, um ortólogo do inibidor-2 previamente caracterizado em D. discoideum. Neste trabalho, também iniciamos a construção do catálogo (The Dictyostelium Phosphatome) que irá conter todas as subunidades catalíticas e reguladoras das proteínas fosfatases codificadas no genoma de D. discoideum. Até o momento, 101 genes foram catalogados e classificados nas diferentes famílias das proteínas fosfatases, sendo 16 na família das PTPs, 26 na família das DSPs (proteína fosfatase de dupla especificidade), 15 na família das PPMs (Phosphoprotein Phosphatase Magnesium-dependent) e 31 na família das PPPs, incluindo genes codificadores de subunidades catalíticas e reguladoras. / Protein phosphatases are responsible for dephosphorylating phosphoaminoacids residues, notably phosphotyrosine and phosphoserine/threonine, thus dividing these enzymes into PTP (protein tyrosine phosphatase) and PP (protein serine/threonine phosphatase) families. Several members of the PP family, in particular those belonging to PPP (Phosphoprotein Phosphatase) are composed of one catalytic subunit and one or more regulatory subunits that provide functional diversity to the holoenzyme. In this work our goal was to identify protein interactors to type 1 (DdPP1c) and type 4 (DdPP4c) phosphatase that might behave as potential regulatory subunits of these enzymes in the social amoeba Dictyostelium discoideum. For this intent, DdPP1c, a mutant isoform of DdPP1 (DdPP1cF269C) and DdPP4c were used as baits in yeast two-hybrid based screening of D. discoideum (AX-4 strain) cDNA libraries from growth as well as developmental stages. At least 30 genes were identified as potential DdPP1c interactors while 10 genes were selected as candidates to interact to DdPP4c. Most of them are currently annotated in D. discoideum genome as hypothetical proteins of unknown function. Among the potential PP1c interactors we selected DdI-3, an ortholog of mammalian inhibitor-3. Interaction of DdI-3 and DdPP1c was confirmed by independent yeast two-hybrid assays. We demonstrated that bacterial expressed recombinant DdI-3 is effective as an inhibitor of DdPP1c in vitro, since 50% of DdPP1c phosphorylase phosphatase activity is inhibited by circa 0,55 nM of purified rDdI-3. Our results also showed that DdI-3 is 50 times more effective than DdI-2, a previously characterized PP1c inhibitor in D. discoideum. In this work we began to organize The Dictyostelium Phosphatome, a catalog of all protein phosphatases, including their catalytic and regulatory subunits, encoded in D. discoideum genome. Until now, we have classified 101 genes into the protein phosphatase families, of which 16 were classified as classic PTP, 26 as DSP (dual-specificity phosphatases), 15 as PPM (Phosphoprotein Phosphatase Magnesium-dependent) and 31 as PPP, including genes for catalytic as well as regulatory subunits.

Development of multicellular organisms is one of the most fundamental but least understood biological processes. Due to its simple life cycle, the lower eukaryote Dictyostelium has been used as a model system to study several basic biological problems, such as cell differentiation, cell motility, cell adhesion, signal transduction, and especially gene regulation. Glycogen phosphorylase is the enzyme that initiates one of the key biochemical pathways, glycogen degradation, during Dictyostelium discoideum development. Two forms of glycogen phosphorylase, gpl and gp2, exist in D. discoideum with gp1 being active in vegetative cells and gp2 in differentiating cells. Study of glycogen phosphorylase gene regulation clearly will provide insight into the molecular mechanism of D. discoideum development and facilitate understanding of development in general. Two distinct genes that encode the two forms of glycogen phosphorylase were cloned. The nucleotide sequence analysis of the gp2 gene revealed an open reading frame of 2976 bp, that consists of three exons separated by two introns. An interesting feature in the gene is a 45 bp sequence in the second exon that contains 11 CAA trinucleotide repeats. The entire 5' and 3' non-coding regions of the gp2 gene and the whole 5' noncoding region of the gp1 gene have also been cloned. The regulation of the gp2 gene by Dictyostelium developmental signals was studied. Both cyclic AMP (cAMP) and Differentiation Inducing Factor (DIF) were discovered to induce gp2 gene expression during differentiation. DIF was also found to inhibit the cAMP responsiveness of the gene. Both cAMP and DIF induction of the gene were repressed by NH₃. Another developmental signalling molecule, adenosine, was involved in gp2 gene regulation through the inhibition of the DIF-mediated expression. The cell-type-specificity of the gp2 gene were also investigated. The gene was found to be expressed in both prestalk/stalk and prespore/spore cells. This is in agreement with the cAMP and DIF inducibility of the gene since the former molecule is a spore-cell morphogen, while the latter is a stalk-cell morphogen. A model of gp2 gene regulation during development is proposed, based on these findings. The two gp? introns and the 45 bp CAA repeat were studied by deletion of these elements. However, there were no alterations of gp2 gene expression observed after these deletions. Also investigated was genomic structural alteration in gp1- mutants that were obtained through homologous recombination and antisense RNA. Southern analysis revealed that the normal gp1 gene was disrupted in all homologous recombination transformants and in half of the antisense RNA transformants. Finally, for the first time, an extrachromosomal luciferase reporter vector has been established for the study of cis-acting regulatory elements in D. discoideum. / Ph. D.

A developmentally regulated cyclic AMP-dependent protein kinase has been recently reported in Dictyostelium discoideum. This report describes some of the physical and kinetic properties of the cAMP-dependent holoenzyme and its subunits. Gel filtration data suggests a holoenzyme Mr of 170,000-190,000, and catalytic and regulatory subunit Mrs of 40,000 and 49,000, respectively. These molecular weight determinations are compatible with an R₂C₂ subunit arrangement of the holoenzyme. Kinase activity required the presence of Mg²⁺ but cAMP binding to the enzyme was not dependent on divalent metal ions. The pH optimum for kinase activity was 7.5; the cAMP binding activity was not affected over a pH range of 5.0-10.0. The holoenzyme and isolated regulatory subunit had identical cAMP Kds of 28 nM. Cyclic AMP was able to dissociate the subunits when analyzed by density gradient centrifugation. Histone VII-S activated the subunits in the absence of cAMP but did not produce their dissociation. In contrast to the gel filtration data, sedimentation values indicated a dimeric holoenzyme structure. Reassociation of the subunits in the absence of cAMP occurred rapidly and was not dependent upon a preincubation with MgATP. High NaCl and low pH depressed both the total kinase activity and the ability of the subunits to reassociate as determined by activity ratio. MgATP did not decrease the ability of the holoenzyme to bind cAMP, neither did the holoenzyme possess a high affinity MgATP binding site. By the use of microdissection techniques holoenzyme levels were determined in individuals at each stage of development and in each cell type during development. Kinase activity was low and non-cAMP dependent in early aggregates but increased and became cAMP-dependent in later aggregates. Maximum activity and cAMP-dependency occurred during the slug and culmination stages. The only differential distribution of the kinase within a single-stage occurred during culmination when the activity in the stalks was approximately one-fourth that in the prespore mass. Preliminary evidence indicates that this difference is not due to an inhibitor. / Ph. D. / incomplete_metadata

Diese Arbeit beschäftigt sich mit zwei verschiedenen Modellsystemen, die Aufschluss über die Mechanik und die Dynamik von zellulären Systemen geben sollen. Zum Einsatz kommt zum einen der Modellorganismus Dictyostelium discoideum, dessen kollektives Migrationsverhalten analysiert wird und zum anderen wird die Mechanik von aktingefüllten Riesenvesikeln als artifizielles Modellsystem etabliert.

Ce travail porte sur l’étude des mécanismes de phosphorylation des protéines par les sérine/thréonine kinases chez Staphylococcus aureus. Nous avons, tout d’abord, mis en évidence et caractérisé une seconde Ser/Thr-kinase, nommée Stk2. Cette kinase présente peu d’homologies avec les autres Ser/Thr-kinases bactériennes décrites à ce jour, en particulier avec la première Ser/Thr-kinase mise en évidence précédemment chez S. aureus, Stk1. Nous avons ensuite caractérisé dix sites d’autophosphorylation de Stk2 et nous avons montré que trois sites sont nécessaires à son activité. Enfin, nous avons montré que le régulateur global de virulence, SarA, est phosphorylé à la fois par Stk1 et Stk2. La phosphorylation de SarA influence sa capacité de liaison à l’ADN. Cette étude contribue à mieux appréhender, au niveau moléculaire, le rôle des Ser/Thr-kinases dans le métabolisme des bactéries et, plus particulièrement, dans la régulation de leur virulence / We report that protein phosphorylation on serine and threonine is required for controlling staphylococcal virulence. We identified and characterized a second serine/threonine kinase, Stk2, in S. aureus. Biochemical analyses revealed that this enzyme displays autokinase activity on both threonine and serine residues. Stk2 is atypical in the sense that it exhibits a weak similarity with the first Ser/Thr-kinase previously detected, Stk1, and its undergoes a different mechanism of activation compared to the other bacterial Ser/Thr-kinases described so far. We also showed that SarA, a major transcription factor that regulates more than a hundred virulence genes, is phosphorylated by both Stk1 and Stk2. Phosphorylation of SarA leads to strong effects on its ability to bind DNA. The study of Stk1 and Stk2, at the molecular level, provides a better understanding of the role of these staphylococcal Ser/Thr-kinases in bacterial metabolism and, in particular, in the regulation of virulence

Les G-quadruplexes sont des structures non-canoniques fascinantes de l’ADN et/ou de l’ARN qui surviennent dans les régions riches en Guanines. La surreprésentation de ces structures dans des régions spécifiques comme les promoteurs des oncogènes et les télomères, suggère leur intervention dans les processus cellulaires clés comme la transcription, la réplication ou bien la maturation de l’ARN. De nouveaux outils in silico, in vitro et in cellulo pour la prédiction des G-quadruplexes ont été proposés, reflétant la pertinence croissante de ces structures. Des cibles potentielles de G-quadruplexes ont été décrites dans le génome humain, chez la levure, des bactéries, virus et bien d’autres. Cependant, un des problèmes dans l’étude des G4s dans le génome humain est le grand nombre de séquences susceptibles de former des structures G4s (370,000 PQS selon Quadparser et plus d’un million en utilisant un seuil de 1.5 selon G4Hunter). Il est alors presque impossible de déconvoluer les effets biologiques reliés aux G-quadruplexes dans les cellules humaines. Pour cela, nous avons choisi Dictyostelium discoideum – dont le génome est pauvre en G4s - comme modèle eucaryote pour compléter les études sur le génome humain. Avec une analyse in silico du génome de dicty en utilisant G4Hunter, un algorithme développé dans notre laboratoire, nous avons pu détecter entre 249 (seuil=2) et 1055 (seuil=1.5) séquences pouvant adopter une structure G4. D’une façon intéressante, bien que les promoteurs soient plus pauvres encore en GC que le reste du génome de dicty, la densité des G4s dans ces régions est significativement plus haute. En utilisant une combinaison de différentes méthodes biophysiques et biochimiques, nous avons démontré que parmi les séquences prédites, 14 séquences qui sont présentes dans des gènes susceptibles de jouer des rôles importants dans dicty forment des structures G4 stables. En plus, cinq gènes de dicty contenant des séquences G4s dans leurs promoteurs ont été étudiés pour l’effet d’un nouveau ligand G4 dérivé de Porphyrine sur leur expression. Nous avons démontré que ce nouveau ligand inhibe l’expression de ces gènes significativement. Globalement, nos résultats constituent le premier pas dans le but d’adopter Dictyostelium discoideum comme un nouveau modèle pour l’étude des G-quadruplexes. / G-quadruplexes (G4) are fascinating non-canonical DNA/RNA secondary structures that occur in genomic Guanine-rich regions. The over-representation of such structures in specific regions such as promoters of oncogenes and telomeres, suggests their involvement in key processes such as transcription, replication or RNA maturation. The development of in silico, in vitro and in cellulo tools for G4 prediction is emerging, reflecting the increasing relevance of these structures. Putative G4 forming sequences (PQS) have been reported in Homo sapiens, yeast, bacteria, viruses and many others. However, one of the problems in studying G4 structures in the human genome is indeed the high number of putative G4 forming sequences (370,000 PQS according to Quadparser and over 1 million when using a threshold of 1.5 with G4Hunter). It is therefore difficult to deconvolute G4-related biological effects in human cells. For this, we chose Dictyostelium discoideum - a G4 poor genome - as a eukaryotic model to complement the human studies. By an in silico analysis of dicty genome with G4Hunter a home-made algorithm, we detected 249 (threshold=2) to 1055 (threshold=1.5) G4-prone motifs. Interestingly, despite an even lower GC content in comparison to the whole dicty genome, the density of G4 motifs in dicty promoters is significantly higher than in the rest of the genome. By using a combination of different biophysical and biochemical methods, we demonstrated that 14 dicty sequences located in key genes fold into stable G4 structures. In addition, five dicty genes containing G4-prone motifs in their promoters were studied for the effect of a new Porphyrin derivative on their expression. Our results demonstrated that the new ligand decreased the expression of the several dicty genes significantly. Overall, our results constitute the first step to adopt Dictyostelium discoideum as a model for G4 studies.

Pseudomonas aeruginosa é uma gamaproteobactéria ubíqua capaz de infectar indivíduos imunocomprometidos e causar infecções hospitalares. Entre duas repetições diretas situadas na ilha de patogenicidade PAPI-1 da linhagem PA14, encontram-se dois grupos de genes transcritos em direções opostas. O primeiro, composto de pvrS, pvrR, rcsC e rcsB, codifica proteínas de sistemas de dois componentes e está relacionado com virulência, enquanto o segundo compreende cinco genes (cupD1-D5) e codifica uma fímbria do tipo chaperoneusher, com alta similaridade com o grupo cupA, envolvido na formação de biofilme em outras linhagens de P. aeruginosa. Fímbrias da mesma família são importantes na patogenicidade de outras bactérias. Com o objetivo de estudar a relação entre esses dois grupos de genes, procurou-se caracterizar sua organização por ensaios de RT-PCR, que possibilitaram observar a disposição dos genes dos sistemas de dois componentes em dois operons distintos (pvrRS e rcsCB) e, pelo menos, cupD1-cupD2 como uma unidade transcricional, com indícios apontando para a organização de cupD1-D5 em um único operon. Os inícios de transcrição e as regiões promotoras de cada operon foram caracterizados por experimentos de RACE 5 e extensão de oligonucleotídeo marcado em busca de seqüências relevantes para a ativação da expressão desses genes. Visando investigar a regulação da expressão da fímbria CupD pelos sistemas de dois componentes codificados pelos genes adjacentes, foram realizados ensaios de qRT-PCR, os quais mostraram uma menor expressão de cupD na linhagem mutante para rcsB. RcsB apresenta um domínio de ligação a DNA e, apesar da falta de sucesso em ensaios de ligação dessa proteína à região promotora de cupD, os dados obtidos por qRT-PCR 1 indicam fortemente que essa proteína funciona como um ativador de transcrição dos genes da fímbria. Corroborando esses achados, somente quando se utilizou RNA extraído de P. aeruginosa PA14 superexpressando RcsB foi possível visualizar no gel a banda referente ao início de transcrição de cupD1-D2 no ensaio de extensão de oligonucleotídeo. Ao contrário do efeito positivo de RcsB observado na transcrição de cupD1, cupD2 e cupD5, a histidina quinase RcsC atua negativamente na expressão dos genes da fímbria, sugerindo que, nesse caso, sua atividade predominante sobre RcsB seja a de fosfatase. PvrS e PvrR parecem atuar de forma indireta e positiva sobre cupD. Como um segundo objetivo do trabalho, ensaios de virulência de P. aeruginosa no hospedeiro-modelo Dictyostelium discoideum foram otimizados, com o estabelecimento de uma técnica para se testar alguns genes estudados no laboratório que podem ser relevantes para a virulência dessa bactéria. Resultados desses ensaios confirmaram a atenuação de mutantes para uma suposta metiltransferase, já observada em modelos de planta, camundongo e drosófila. Em conjunto, os resultados desse trabalho tornam-se informações valiosas para serem usadas na pesquisa de controle de infecções por P. aeruginosa, que depende de fímbrias para colonizar com sucesso superfícies abióticas que servem de veículo de disseminação desse agente infeccioso e para que as bactérias possam persistir no organismo do hospedeiro. / Pseudomonas aeruginosa is a ubiquitous gammaproteobacteria able to infect immunocompromised individuals and to cause nosocomial infections. Between two direct repeats in the PAPI-1 pathogenicity island present in strain PA14, there are two gene clusters transcribed in opposite directions. The first, composed of pvrS, pvrR, rcsC and rcsB, encodes two-component systems proteins and is implicated in virulence, whereas the second comprises five genes (cupD1-D5) encoding a chaperone-usher fimbria with high similarity to cupA, a gene cluster involved in biofilm formation in other strains of P. aeruginosa. Fimbriae belonging to the same family of Cup are related to pathogenicity of other bacteria. In order to study the relationship between these two clusters, the organization of the genes in operons was characterized using RT-PCR, which results lead to the conclusion that the twocomponent systems genes are arranged into two different operons (pvrSR and rcsCB) and that cupD1-D2 share the same promoter, with some evidences that the operons extends from cupD1 to cupD5. The transcription start sites and the promoter regions of each operon were characterized with RACE 5 and primer extension assays to look for sequences that could be relevant to the activation of expression of these genes. Quantitative RT-PCR assays were carried out to investigate whether the expression of CupD fimbriae is regulated by the twocomponent systems encoded by the adjacent genes and the results showed a lower expression of cupD in the rcsB mutant strain than in the wild-type. RcsB bears a DNA-binding domain and, although our assays of DNA-protein interactions have failed, data obtained by qRT-PCR 1 strongly indicate that this protein functions as a transcription activator of fimbrial genes. These findings were corroborated by the primer extension assay, in which the band corresponding to the transcriptional start of cupD1-D2 was visible only when the reaction was performed with the RNA extracted of P. aeruginosa overexpressing RcsB. Unlike the effect observed for RcsB in cupD1, cupD2 and cupD5 transcription, the histidine quinase RcsC acts negatively in the fimbrial genes expression, suggesting that it might function predominantly as a RcsB phosphatase. PvrS and PvrR seem to regulate cupD positively and indirectly. As a second aim of this work, virulence assays of P. aeruginosa in the model host Dictyostelium discoideum were optimized, and a technique for testing genes studied in the laboratory that could be important for Pseudomonas virulence was established. These assays confirmed the attenuation-in-virulence of strains mutant in a putative methyltransferase gene, as observed before in plant, mouse and drosophila models. The results obtained in this work may contribute to P. aeruginosa infection control research, since this bacterium depends on fimbriae to successfully colonize abiotic surfaces that act as a dissemination vehicle, and to allow the bacteria to persist into the host organism.

Pseudomonas aeruginosa é uma gamaproteobactéria ubíqua capaz de infectar indivíduos imunocomprometidos e causar infecções hospitalares. Entre duas repetições diretas situadas na ilha de patogenicidade PAPI-1 da linhagem PA14, encontram-se dois grupos de genes transcritos em direções opostas. O primeiro, composto de pvrS, pvrR, rcsC e rcsB, codifica proteínas de sistemas de dois componentes e está relacionado com virulência, enquanto o segundo compreende cinco genes (cupD1-D5) e codifica uma fímbria do tipo chaperoneusher, com alta similaridade com o grupo cupA, envolvido na formação de biofilme em outras linhagens de P. aeruginosa. Fímbrias da mesma família são importantes na patogenicidade de outras bactérias. Com o objetivo de estudar a relação entre esses dois grupos de genes, procurou-se caracterizar sua organização por ensaios de RT-PCR, que possibilitaram observar a disposição dos genes dos sistemas de dois componentes em dois operons distintos (pvrRS e rcsCB) e, pelo menos, cupD1-cupD2 como uma unidade transcricional, com indícios apontando para a organização de cupD1-D5 em um único operon. Os inícios de transcrição e as regiões promotoras de cada operon foram caracterizados por experimentos de RACE 5 e extensão de oligonucleotídeo marcado em busca de seqüências relevantes para a ativação da expressão desses genes. Visando investigar a regulação da expressão da fímbria CupD pelos sistemas de dois componentes codificados pelos genes adjacentes, foram realizados ensaios de qRT-PCR, os quais mostraram uma menor expressão de cupD na linhagem mutante para rcsB. RcsB apresenta um domínio de ligação a DNA e, apesar da falta de sucesso em ensaios de ligação dessa proteína à região promotora de cupD, os dados obtidos por qRT-PCR 1 indicam fortemente que essa proteína funciona como um ativador de transcrição dos genes da fímbria. Corroborando esses achados, somente quando se utilizou RNA extraído de P. aeruginosa PA14 superexpressando RcsB foi possível visualizar no gel a banda referente ao início de transcrição de cupD1-D2 no ensaio de extensão de oligonucleotídeo. Ao contrário do efeito positivo de RcsB observado na transcrição de cupD1, cupD2 e cupD5, a histidina quinase RcsC atua negativamente na expressão dos genes da fímbria, sugerindo que, nesse caso, sua atividade predominante sobre RcsB seja a de fosfatase. PvrS e PvrR parecem atuar de forma indireta e positiva sobre cupD. Como um segundo objetivo do trabalho, ensaios de virulência de P. aeruginosa no hospedeiro-modelo Dictyostelium discoideum foram otimizados, com o estabelecimento de uma técnica para se testar alguns genes estudados no laboratório que podem ser relevantes para a virulência dessa bactéria. Resultados desses ensaios confirmaram a atenuação de mutantes para uma suposta metiltransferase, já observada em modelos de planta, camundongo e drosófila. Em conjunto, os resultados desse trabalho tornam-se informações valiosas para serem usadas na pesquisa de controle de infecções por P. aeruginosa, que depende de fímbrias para colonizar com sucesso superfícies abióticas que servem de veículo de disseminação desse agente infeccioso e para que as bactérias possam persistir no organismo do hospedeiro. / Pseudomonas aeruginosa is a ubiquitous gammaproteobacteria able to infect immunocompromised individuals and to cause nosocomial infections. Between two direct repeats in the PAPI-1 pathogenicity island present in strain PA14, there are two gene clusters transcribed in opposite directions. The first, composed of pvrS, pvrR, rcsC and rcsB, encodes two-component systems proteins and is implicated in virulence, whereas the second comprises five genes (cupD1-D5) encoding a chaperone-usher fimbria with high similarity to cupA, a gene cluster involved in biofilm formation in other strains of P. aeruginosa. Fimbriae belonging to the same family of Cup are related to pathogenicity of other bacteria. In order to study the relationship between these two clusters, the organization of the genes in operons was characterized using RT-PCR, which results lead to the conclusion that the twocomponent systems genes are arranged into two different operons (pvrSR and rcsCB) and that cupD1-D2 share the same promoter, with some evidences that the operons extends from cupD1 to cupD5. The transcription start sites and the promoter regions of each operon were characterized with RACE 5 and primer extension assays to look for sequences that could be relevant to the activation of expression of these genes. Quantitative RT-PCR assays were carried out to investigate whether the expression of CupD fimbriae is regulated by the twocomponent systems encoded by the adjacent genes and the results showed a lower expression of cupD in the rcsB mutant strain than in the wild-type. RcsB bears a DNA-binding domain and, although our assays of DNA-protein interactions have failed, data obtained by qRT-PCR 1 strongly indicate that this protein functions as a transcription activator of fimbrial genes. These findings were corroborated by the primer extension assay, in which the band corresponding to the transcriptional start of cupD1-D2 was visible only when the reaction was performed with the RNA extracted of P. aeruginosa overexpressing RcsB. Unlike the effect observed for RcsB in cupD1, cupD2 and cupD5 transcription, the histidine quinase RcsC acts negatively in the fimbrial genes expression, suggesting that it might function predominantly as a RcsB phosphatase. PvrS and PvrR seem to regulate cupD positively and indirectly. As a second aim of this work, virulence assays of P. aeruginosa in the model host Dictyostelium discoideum were optimized, and a technique for testing genes studied in the laboratory that could be important for Pseudomonas virulence was established. These assays confirmed the attenuation-in-virulence of strains mutant in a putative methyltransferase gene, as observed before in plant, mouse and drosophila models. The results obtained in this work may contribute to P. aeruginosa infection control research, since this bacterium depends on fimbriae to successfully colonize abiotic surfaces that act as a dissemination vehicle, and to allow the bacteria to persist into the host organism.