Enterohemorrhagic Escherichia coli O157:H7 (EHEC) is a human intestinal pathogen, which can cause severe disease. EHEC O157:H7 is responsible for outbreaks of diarrhea and hemorrhagic colitis. EHEC produces a potent cytotoxin known as Vero (Shiga-like) cytotoxin, which causes diarrhea-associated hemolytic uremic syndrome (HUS), the most common cause of acute renal failure in children. Current treatment remains predominantly supportive in nature because antibiotics and non-steroidal anti-inflammatory drugs exacerbate the condition. Therefore, alternative therapeutic approaches that will prevent the EHEC colonization without the release of toxins need to be delineated. Understanding the pathobiology of disease is likely to yield novel approaches to interrupt the infectious process. My hypothesis was that pathogen-derived effectors associate with lipid rafts and, thereby, promote the recruitment of host signal transduction proteins to lipid rafts in response to EHEC O157:H7 infection. In this thesis, specific host signalling pathways hijacked by EHEC O157:H7, through lipid raft signalling platforms, to elicit pathogenic effects are studied using complementary approaches, including epithelial model cell lines and an animal model of infection (Citrobacter rodentium challenge of mice). A lack of osteopontin resulted in decreased attaching effacing lesions and reduced colonic epithelial cell hyperplasia in response to C. rodentium infection. These findings suggest that C. rodentium, mimicking EHEC O157:H7 infection, is capable of utilizing host cell components to elicit its pathogenic effects. In vitro data showed that EHEC O157:H7 effector proteins manipulate cell signalling through lipid rafts employed as platforms to recruit and activate host second messengers. PKC and PI3K activation led to attaching and effacing lesions, disruption of tight junctions, and the initiation of both innate and adaptive host immune responses. The results pointed towards a role for atypical PKC in EHEC-induced attaching and effacing lesion formation. The role of lipid rafts in EHEC O157:H7 pathogenesis was also studied using Citrobacter rodentium-infected Niemann-pick type C (NPC) mice. Infection of NPC mice, which lack lipid rafts, with C. rodentium resulted in delayed colonization and delayed onset of attaching-effacing lesion formation, compared with infected wild type mice. C. rodentium-infected NPC mice also demonstrated reduced colonic epithelial hyperplasia and decreased secretion of the pro-inflammatory cytokine, interferon-γ. Taken together, the findings presented in this thesis highlight the importance of host cell signal transduction cascades in EHEC O157:H7 disease pathogenesis, and demonstrate a role for lipid rafts and OPN in mediating host cell signaling responses to non-invasive enteric microbial pathogens.

Enterohemorrhagic Escherichia coli O157:H7 (EHEC) is a human intestinal pathogen, which can cause severe disease. EHEC O157:H7 is responsible for outbreaks of diarrhea and hemorrhagic colitis. EHEC produces a potent cytotoxin known as Vero (Shiga-like) cytotoxin, which causes diarrhea-associated hemolytic uremic syndrome (HUS), the most common cause of acute renal failure in children. Current treatment remains predominantly supportive in nature because antibiotics and non-steroidal anti-inflammatory drugs exacerbate the condition. Therefore, alternative therapeutic approaches that will prevent the EHEC colonization without the release of toxins need to be delineated. Understanding the pathobiology of disease is likely to yield novel approaches to interrupt the infectious process. My hypothesis was that pathogen-derived effectors associate with lipid rafts and, thereby, promote the recruitment of host signal transduction proteins to lipid rafts in response to EHEC O157:H7 infection. In this thesis, specific host signalling pathways hijacked by EHEC O157:H7, through lipid raft signalling platforms, to elicit pathogenic effects are studied using complementary approaches, including epithelial model cell lines and an animal model of infection (Citrobacter rodentium challenge of mice). A lack of osteopontin resulted in decreased attaching effacing lesions and reduced colonic epithelial cell hyperplasia in response to C. rodentium infection. These findings suggest that C. rodentium, mimicking EHEC O157:H7 infection, is capable of utilizing host cell components to elicit its pathogenic effects. In vitro data showed that EHEC O157:H7 effector proteins manipulate cell signalling through lipid rafts employed as platforms to recruit and activate host second messengers. PKC and PI3K activation led to attaching and effacing lesions, disruption of tight junctions, and the initiation of both innate and adaptive host immune responses. The results pointed towards a role for atypical PKC in EHEC-induced attaching and effacing lesion formation. The role of lipid rafts in EHEC O157:H7 pathogenesis was also studied using Citrobacter rodentium-infected Niemann-pick type C (NPC) mice. Infection of NPC mice, which lack lipid rafts, with C. rodentium resulted in delayed colonization and delayed onset of attaching-effacing lesion formation, compared with infected wild type mice. C. rodentium-infected NPC mice also demonstrated reduced colonic epithelial hyperplasia and decreased secretion of the pro-inflammatory cytokine, interferon-γ. Taken together, the findings presented in this thesis highlight the importance of host cell signal transduction cascades in EHEC O157:H7 disease pathogenesis, and demonstrate a role for lipid rafts and OPN in mediating host cell signaling responses to non-invasive enteric microbial pathogens.

This thesis describes the development of novel fluorescent and photocaged lipids, and their application as tools to probe the morphological effects of ceramide (Cer)-mediated membrane reorganization in supported lipid bilayers. Cer is a sphingolipid found in eukaryotic cells that plays a key role in regulating biological processes such as apoptosis, cell-to-cell communication, differentiation and some types of pathogenesis. Sphingolipid and cholesterol-rich lipid rafts in the plasma membrane are thought to be the point of origin for many of this lipid second messenger’s effects. Cer is formed in the exoplasmic leaflet of the plasma membrane via the enzymatic hydrolysis of sphingomyelin. The compositional complexity of biological membranes has prompted the adoption of simpler model systems to study the effects of Cer generation. When it is directly incorporated into model membranes, Cer segregates into highly ordered domains with physical properties that are distinct from those of the surrounding fluid environments. However, enzymatic generation of Cer induces complex and dynamic membrane heterogeneity that is difficult to interpret and reconcile with its direct incorporation. Here I describe the synthesis of 4-nitrobenzo-2-oxa-1,3-diazol-7-yl (NBD)-labelled cholesterol (Chol) and Cer analogs, and their use as probes in model membranes exhibiting liquid-disordered (Ld) and liquid-ordered (Lo) phase coexistence. The Chol probes reproduce the modest enrichment of Chol in Lo membrane domains as well as the Cer-induced displacement of cholesterol. One of the NBD Chol probes is used to provide direct visualization of Chol redistribution during enzymatic Cer generation, and assists in identifying new features as Cer-rich regions. The NBD-labelled Cer quantifies membrane order using orientational order parameter measurements derived from polarized total internal reflection fluorescence microscopy (pTIRFM) images. The probe reports on changes in membrane order upon enzymatic generation of Cer, and indicates a significant increase in the molecular order of Ld membrane regions that is consistent with the redistribution of Chol into these areas. The probe also identifies de novo Cer-rich domains as areas of particularly high molecular order. In the final project area, 6-Bromo-7-hydroxycoumarin-4-ylmethyl (Bhc)-caged Cers are shown to release Cer rapidly and efficiently upon irradiation with near-visible UV light. The caged lipids are then incorporated into supported membranes and photolyzed to release Cer with a high degree of spatial and temporal control. Controlled Cer generation is then used to drive protein-ganglioside clustering in lipid bilayers.

This thesis describes the development of novel fluorescent and photocaged lipids, and their application as tools to probe the morphological effects of ceramide (Cer)-mediated membrane reorganization in supported lipid bilayers. Cer is a sphingolipid found in eukaryotic cells that plays a key role in regulating biological processes such as apoptosis, cell-to-cell communication, differentiation and some types of pathogenesis. Sphingolipid and cholesterol-rich lipid rafts in the plasma membrane are thought to be the point of origin for many of this lipid second messenger’s effects. Cer is formed in the exoplasmic leaflet of the plasma membrane via the enzymatic hydrolysis of sphingomyelin. The compositional complexity of biological membranes has prompted the adoption of simpler model systems to study the effects of Cer generation. When it is directly incorporated into model membranes, Cer segregates into highly ordered domains with physical properties that are distinct from those of the surrounding fluid environments. However, enzymatic generation of Cer induces complex and dynamic membrane heterogeneity that is difficult to interpret and reconcile with its direct incorporation. Here I describe the synthesis of 4-nitrobenzo-2-oxa-1,3-diazol-7-yl (NBD)-labelled cholesterol (Chol) and Cer analogs, and their use as probes in model membranes exhibiting liquid-disordered (Ld) and liquid-ordered (Lo) phase coexistence. The Chol probes reproduce the modest enrichment of Chol in Lo membrane domains as well as the Cer-induced displacement of cholesterol. One of the NBD Chol probes is used to provide direct visualization of Chol redistribution during enzymatic Cer generation, and assists in identifying new features as Cer-rich regions. The NBD-labelled Cer quantifies membrane order using orientational order parameter measurements derived from polarized total internal reflection fluorescence microscopy (pTIRFM) images. The probe reports on changes in membrane order upon enzymatic generation of Cer, and indicates a significant increase in the molecular order of Ld membrane regions that is consistent with the redistribution of Chol into these areas. The probe also identifies de novo Cer-rich domains as areas of particularly high molecular order. In the final project area, 6-Bromo-7-hydroxycoumarin-4-ylmethyl (Bhc)-caged Cers are shown to release Cer rapidly and efficiently upon irradiation with near-visible UV light. The caged lipids are then incorporated into supported membranes and photolyzed to release Cer with a high degree of spatial and temporal control. Controlled Cer generation is then used to drive protein-ganglioside clustering in lipid bilayers.

No description available.

A biomineralização óssea é um processo complexo e multifatorial sendo um grande desafio para a ciência à compreensão dos seus mecanismos regulatórios. Este processo é mediado pela liberação de vesículas da matriz (MVs), as quais surgem das superfícies de osteoblastos e são secretadas no local específico do início da biomineralização. MVs têm a capacidade de acumular altas concentrações de íons Ca2+ e fosfato (Pi), proporcionando um microambiente adequado para a formação inicial e propagação dos cristais de hidroxiapatita. Especial atenção deve ser dada a duas proteínas: Anexina V (AnxA5) e Fosfatase Alcalina (TNAP). As anexinas são as proteínas mais abundantes detectadas nas MVs e responsáveis pela formação de canais de cálcio. TNAP apresenta atividade fosfomonohidrolítica, produzindo Pi a partir, principalmente, de pirofosfato (PPi) e ATP. O enfoque deste projeto foi produzir e caracterizar proteolipossomos com diferentes composições lipídicas de dipalmitoil fosfatidilcolina (DPPC) e dipalmitoil fosfatidilserina (DPPS) contendo TNAP e AnxA5, e manter a funcionalidade das proteínas após incorporação nos sistemas miméticos. Foi possível incorporar AnxA5 em DPPC-proteolipossomos (11,64 µg/mL), mas na presença de DPPS houve um aumento significativo de AnxA5 incorporada (25,79 µg/mL) a DPPC:DPPS 10%-proteolipossomos (razão molar). A presença das proteínas nos proteolipossomos compostos por DPPC e DPPC:DPPS 5, 10 e 15% (razão molar) foi confirmada por SDS-PAGE e Immunoblotting. Melhores rendimentos de incorporação das duas proteínas foram obtidos quando ambas foram incorporadas concomitantemente. DPPC-proteolipossomos e DPPC:DPPS 10%-proteoliposomos revelaram conter 75% de AnxA5 e 25% de TNAP em concentração de proteína. A presença de DPPS não afetou significativamente as porcentagens de proteínas incorporadas. Os parâmetros cinéticos da TNAP na hidrólise de diferentes substratos fisiológicos (ATP, ADP e PPi) foram determinados na presença e ausência de AnxA5, em pH fisiológico, e para os diversos sistemas lipídicos. A melhor eficiência catalítica da enzima foi obtida para sistemas contendo 10% de DPPS (razão molar) (kcat/K0.5= 183,02; 776,06 e 657,08 M-1.s-1, respectivamente). A TNAP apresentou maior especificidade para a hidrólise de PPi quando comparado com ATP e ADP. Estudos utilizando Calorimetria Diferencial de Varredura (DSC) mostraram que o aumento da concentração de DPPS em DPPC-lipossomos proporcionou um progressivo alargamento no pico de transição de fase, diminuição na t1/2 e H. A pré-transição de fase só foi detectada até a concentração de 15% de DPPS em DPPC. Para 20% de DPPS e acima, observou-se uma segregação lateral de fase com a formação de possíveis microdomínios ricos em DPPS. A interação da AnxA5 com DPPC-lipossomos e DPPC:DPPS 10%-lipossomos resultou em uma redução nos valores de H (de 8,73 para 5,68 e 8,43 para 5,37 Kcal.mol-1, respectivamente). Quando a TNAP está presente nos proteolipossomos, este efeito é ainda maior. A AnxA5 incorporada em DPPC-proteolipossomos e DPPC:DPPS 10%-proteolipossomos (razão molar) foi capazes de mediar o influxo de 45Ca2+ para dentro das vesículas (~ 800 nmol Ca2+) quando utilizados faixas de concentração de cálcio em níveis fisiológicos (~2 mM). A presença da TNAP nos proteolipossomos não afetou o influxo de Ca2+ mediado pela AnxA5. Entretanto, a presença da AnxA5 afetou significativamente os parâmetros cinéticos da TNAP para os diferentes substratos. Estudos com vesículas unilamelares gigantes (GUVs) também confirmaram a inserção funcional da AnxA5 em vesículas constituídos de dioleoil fosfatidilcolina (DOPC) e DOPC:DPPS 10% (razão molar). O principal efeito causado pela AnxA5 na morfologia das GUVs foi a perda de contraste óptico devido a formação de poros nas membranas das vesículas. Neste caso, a presença de DPPS não proporcionou mudanças significativas para a incorporação da AnxA5. TNAP quando inserida em GUVs provocou intensa flutuação e excesso de área das vesículas com formação de filamentos. A presença do DPPS provavelmente dificulta a inserção da TNAP à membrana das GUVs. Quando há microdomínios lipídicos heterogêneos na composição de GUVs compostas por DOPC:Colesterol:Esfingomielina (8:1:1) e DOPC:Colesterol:Esfingomielina:Gangliosídeo (7:1:1:1) (razão molar), a inserção da TNAP provocou uma maior segregação lateral de fase evidenciada por imagens com fluorescência. A presença da TNAP e AnxA5 em DPPC:DPPS 10%-proteolipossomos proporcionou mudanças significativas nas propriedades mecânicas visco-elásticas dos proteolipossomos detectadas por imagens de Microscopia de Força Atômica. Assim, no presente trabalho foi possível obter uma inédita metodologia para a formação de proteolipossomos contendo TNAP e AnxA5 concomitantemente, os quais apresentaram uma reconstituição funcional das proteínas, apresentando capacidade de captar Ca2+ para dentro das vesículas e habilidade de hidrolisar fosfosubstratos em sua superfície. / Bone biomineralization is a multifactorial and complex process, being a challenge for the science the understanding of their regulatory mechanisms. This process is mediated by the release of matrix vesicles (MVs), structures which arise by budding from osteoblast and chondroblast surface and are secreted in the specific site where biomineralization begins. MVs have the ability of accumulating high concentrations of Ca2+ and Pi ions, providing an adequate microenvironment for the initial formation and propagation of hydroxyapatite crystals. Two protein families present in MVs merit special attention: Annexins and Phosphatases. The annexins were the most abundant proteins detected in MVs and are responsible for the Ca2+-channels formation (especially AnxA5). Tissue-nonspecific alkaline phosphatase (TNAP) exhibits phosphomonohydrolytic activity, producing Pi mainly from PPi and ATP. Such proteins regulate the formation of calcium phosphate crystals, acting directly in the bone mineralization process. The goal of this project was to produce and characterize proteoliposomes with different lipid compositions of dipalmitoylphosphatidylcholine (DPPC) and dipalmitoylphosphatidylserine (DPPS) harboring TNAP and AnxA5, keeping the functions of both proteins after their incorporation into the mimetic systems. AnxA5 was able to incorporate into DPPC-proteoliposomes (11.64 µg/mL), but the presence of DPPS increased significantly the AnxA5 incorporation (25.79 µg/mL) into DPPC:DPPS 10%-proteoliposomes. The presence of both proteins into DPPC and DPPC:DPPS 5, 10 and 15% (molar ratios) proteoliposomes was confirmed by SDS-PAGE and Immunoblotting analysis. Better yield of TNAP and AnxA5 incorporation was observed when both proteins were reconstituted simultaneously. DPPC-proteoliposomes and DPPC:DPPS 10%-proteoliposomes (molar ratio) incorporated about 75% of AnxA5 and 25% of TNAP (protein concentration). DPPS presence did not affect significantly the yield of incorporation of both proteins. The kinetic parameters for the hydrolysis of different physiological substrates (ATP, ADP and PPi) by TNAP were determined in the presence and absence of AnxA5, at physiological pH, for the different systems. The best catalytic efficiencies were achieved with proteoliposomes containing DPPS 10% (molar ratio) (kcat/K0.5= 183.02; 776.06 and 657.08 M-1.s-1 for ATP, ADP and PPi, respectively), condition that also favored PPi hydrolysis by TNAP when compared to ATP and ADP hydrolysis. Studies by Differential Scanning Calorimetry (DSC) showed that the increasing DPPS concentrations in the DPPC-liposomes resulted in a progressive broadening of the phase transition peaks and decreased t1/2 and H values. The pre-transition was detected only in concentrations up to DPPS 15% in DPPC. Phase lateral segregation can be observed for DPPS 20% and above, suggesting the formation of DPPS-rich microdomains. The interaction of AnxA5 with DPPC and DPPC:DPPS 10%-liposomes resulted in a decrease of H values (from 8.73 to 5.68 and from 8.43 to 5.37 Kcal.mol-1, respectively). When TNAP was present in the proteoliposomes, this effect was even greater. AnxA5 incorporated into DPPC and DPPC:DPPS 10%-proteoliposomes (molar ratio) was able to mediate 45Ca2+-influx (~ 800 nmol Ca2+) into the vesicles at physiological Ca2+-concentrations (~ 2 mM), and this process was not affected by the presence of TNAP in the systems. However, AnxA5 affected significantly the hydrolysis of substrates by TNAP. Studies with Giant Unilamellar Vesicles (GUVs) also confirmed the functional reconstitution of AnxA5 in dioleoylphosphocholine (DOPC) and DOPC:DPPS 10% (molar ratio) vesicles. The main effect caused by AnxA5 in the GUVs morphology was the formation of pores in the vesicles membrane. In this case, DPPS presence did not affect the AnxA5 incorporation. The presence of TNAP in GUVs caused a several fluctuation, indicating that the vesicles acquired an excess of area and undergoes sequential budding transitions. It is suggested that the presence of DPPS makes the TNAP insertion into the GUVs membrane difficult. With the presence of heterogeneous lipid microdomains in GUVs composed of DOPC, Cholesterol (Chol), Sphingomyelin (SM) and Ganglioside (GM1) in the proportions DOPC:Chol:SM 8:1:1 and DOPC:Chol:SM:GM1 7:1:1:1 (molar ratios), the TNAP insertion caused a greater phase lateral segregation, evidenced by fluorescence analysis. Atomic Force Microscopy (AFM) analysis indicated that the presence of both proteins into DPPC:DPPS 10%-proteoliposomes (molar ratio) caused significant changes in the visco-elastic mechanical properties of the vesicles. In conclusion, the present work describes the synthesis of proteoliposomes harboring TNAP and AnxA5 concomitantly, with the functional reconstitution of both proteins, with the ability to transport Ca2+ into the vesicles and hydrolyze phosphosubstrates on their surface.

The structure and functions of lipid rafts and the mechanisms of intracellular membrane trafficking are major topics in current cell biological research. Rafts have been proposed to act as sorting platforms during biosynthetic transport, especially along pathways that deliver proteins to the apical membrane of polarised cells. Based on this, the aim of this work was to contribute to the understanding of apical sorting in epithelial cells. The study of how lipid rafts are structured has been hampered by the scarcity of techniques for their purification. Rafts are thought to be partially resistant to solubilisation by mild detergents, which has made the isolation of detergent-resistant membranes (DRMs) the primary method to characterise them biochemically. While a growing number of detergents is being used to prepare DRMs, it is not clear what can be inferred about the native structure of cell membranes from the composition of different DRMs. This issue was addressed by an analysis of DRMs prepared with a variety of mild detergents. The protein and lipid content of different DRMs from two cell lines, Madin-Darby canine kidney (MDCK) and Jurkat cells, was compared. It was shown that the detergents differed considerably in their ability to selectively solubilise membrane proteins and lipids. These results make it unlikely that different DRMs reflect the same underlying principle of membrane organisation. Another obstacle for understanding apical sorting is that the evidence implicating certain proteins in this process has come from various disparate approaches. It would be helpful to re-examine the putative components of the apical sorting machinery in a single experimental system. To this end, a retroviral system for RNA interference (RNAi) in MDCK cells was established. Efficient suppression of thirteen genes was achieved by retroviral co-expression of short hairpin RNAs and a selectable marker. In addition, the system was extended to simultaneously target two genes, giving rise to double knockdowns.Retroviral RNAi was applied to deplete proteins implicated in apical sorting. Surprisingly, none of the knockdowns analysed caused defects in surface delivery of influenza virus hemagglutinin, a common marker protein for apical transport. Therefore, none of the proteins examined is absolutely required for transport to the apical membrane of MDCK cells. Cells may adapt to the depletion of proteins involved in membrane trafficking by activating alternative pathways. To avoid such adaptation, a visual transport assay was established. It is based on the adenoviral expression of fluorescent marker proteins whose surface transport can be followed microscopically as soon as RNAi has become effective. With this assay, it should now be possible to screen the knockdowns for defects in surface transport. Taken together, this work has provided a number of experimental tools for the study of membrane trafficking in epithelial cells. First, the biochemical analysis of DRMs highlighted that DRMs obtained with different detergents are unlikely to correspond to distinct types of membrane microdomains in cell membranes. Second, the retroviral RNAi system should be valuable for defining the function of proteins, not only in membrane transport, but also in processes like epithelial polarisation. Third, the visual assay for monitoring the surface transport of adenovirally expressed marker proteins should be suitable to detect defects in polarised sorting.

A biomineralização óssea é um processo complexo e multifatorial sendo um grande desafio para a ciência à compreensão dos seus mecanismos regulatórios. Este processo é mediado pela liberação de vesículas da matriz (MVs), as quais surgem das superfícies de osteoblastos e são secretadas no local específico do início da biomineralização. MVs têm a capacidade de acumular altas concentrações de íons Ca2+ e fosfato (Pi), proporcionando um microambiente adequado para a formação inicial e propagação dos cristais de hidroxiapatita. Especial atenção deve ser dada a duas proteínas: Anexina V (AnxA5) e Fosfatase Alcalina (TNAP). As anexinas são as proteínas mais abundantes detectadas nas MVs e responsáveis pela formação de canais de cálcio. TNAP apresenta atividade fosfomonohidrolítica, produzindo Pi a partir, principalmente, de pirofosfato (PPi) e ATP. O enfoque deste projeto foi produzir e caracterizar proteolipossomos com diferentes composições lipídicas de dipalmitoil fosfatidilcolina (DPPC) e dipalmitoil fosfatidilserina (DPPS) contendo TNAP e AnxA5, e manter a funcionalidade das proteínas após incorporação nos sistemas miméticos. Foi possível incorporar AnxA5 em DPPC-proteolipossomos (11,64 µg/mL), mas na presença de DPPS houve um aumento significativo de AnxA5 incorporada (25,79 µg/mL) a DPPC:DPPS 10%-proteolipossomos (razão molar). A presença das proteínas nos proteolipossomos compostos por DPPC e DPPC:DPPS 5, 10 e 15% (razão molar) foi confirmada por SDS-PAGE e Immunoblotting. Melhores rendimentos de incorporação das duas proteínas foram obtidos quando ambas foram incorporadas concomitantemente. DPPC-proteolipossomos e DPPC:DPPS 10%-proteoliposomos revelaram conter 75% de AnxA5 e 25% de TNAP em concentração de proteína. A presença de DPPS não afetou significativamente as porcentagens de proteínas incorporadas. Os parâmetros cinéticos da TNAP na hidrólise de diferentes substratos fisiológicos (ATP, ADP e PPi) foram determinados na presença e ausência de AnxA5, em pH fisiológico, e para os diversos sistemas lipídicos. A melhor eficiência catalítica da enzima foi obtida para sistemas contendo 10% de DPPS (razão molar) (kcat/K0.5= 183,02; 776,06 e 657,08 M-1.s-1, respectivamente). A TNAP apresentou maior especificidade para a hidrólise de PPi quando comparado com ATP e ADP. Estudos utilizando Calorimetria Diferencial de Varredura (DSC) mostraram que o aumento da concentração de DPPS em DPPC-lipossomos proporcionou um progressivo alargamento no pico de transição de fase, diminuição na t1/2 e H. A pré-transição de fase só foi detectada até a concentração de 15% de DPPS em DPPC. Para 20% de DPPS e acima, observou-se uma segregação lateral de fase com a formação de possíveis microdomínios ricos em DPPS. A interação da AnxA5 com DPPC-lipossomos e DPPC:DPPS 10%-lipossomos resultou em uma redução nos valores de H (de 8,73 para 5,68 e 8,43 para 5,37 Kcal.mol-1, respectivamente). Quando a TNAP está presente nos proteolipossomos, este efeito é ainda maior. A AnxA5 incorporada em DPPC-proteolipossomos e DPPC:DPPS 10%-proteolipossomos (razão molar) foi capazes de mediar o influxo de 45Ca2+ para dentro das vesículas (~ 800 nmol Ca2+) quando utilizados faixas de concentração de cálcio em níveis fisiológicos (~2 mM). A presença da TNAP nos proteolipossomos não afetou o influxo de Ca2+ mediado pela AnxA5. Entretanto, a presença da AnxA5 afetou significativamente os parâmetros cinéticos da TNAP para os diferentes substratos. Estudos com vesículas unilamelares gigantes (GUVs) também confirmaram a inserção funcional da AnxA5 em vesículas constituídos de dioleoil fosfatidilcolina (DOPC) e DOPC:DPPS 10% (razão molar). O principal efeito causado pela AnxA5 na morfologia das GUVs foi a perda de contraste óptico devido a formação de poros nas membranas das vesículas. Neste caso, a presença de DPPS não proporcionou mudanças significativas para a incorporação da AnxA5. TNAP quando inserida em GUVs provocou intensa flutuação e excesso de área das vesículas com formação de filamentos. A presença do DPPS provavelmente dificulta a inserção da TNAP à membrana das GUVs. Quando há microdomínios lipídicos heterogêneos na composição de GUVs compostas por DOPC:Colesterol:Esfingomielina (8:1:1) e DOPC:Colesterol:Esfingomielina:Gangliosídeo (7:1:1:1) (razão molar), a inserção da TNAP provocou uma maior segregação lateral de fase evidenciada por imagens com fluorescência. A presença da TNAP e AnxA5 em DPPC:DPPS 10%-proteolipossomos proporcionou mudanças significativas nas propriedades mecânicas visco-elásticas dos proteolipossomos detectadas por imagens de Microscopia de Força Atômica. Assim, no presente trabalho foi possível obter uma inédita metodologia para a formação de proteolipossomos contendo TNAP e AnxA5 concomitantemente, os quais apresentaram uma reconstituição funcional das proteínas, apresentando capacidade de captar Ca2+ para dentro das vesículas e habilidade de hidrolisar fosfosubstratos em sua superfície. / Bone biomineralization is a multifactorial and complex process, being a challenge for the science the understanding of their regulatory mechanisms. This process is mediated by the release of matrix vesicles (MVs), structures which arise by budding from osteoblast and chondroblast surface and are secreted in the specific site where biomineralization begins. MVs have the ability of accumulating high concentrations of Ca2+ and Pi ions, providing an adequate microenvironment for the initial formation and propagation of hydroxyapatite crystals. Two protein families present in MVs merit special attention: Annexins and Phosphatases. The annexins were the most abundant proteins detected in MVs and are responsible for the Ca2+-channels formation (especially AnxA5). Tissue-nonspecific alkaline phosphatase (TNAP) exhibits phosphomonohydrolytic activity, producing Pi mainly from PPi and ATP. Such proteins regulate the formation of calcium phosphate crystals, acting directly in the bone mineralization process. The goal of this project was to produce and characterize proteoliposomes with different lipid compositions of dipalmitoylphosphatidylcholine (DPPC) and dipalmitoylphosphatidylserine (DPPS) harboring TNAP and AnxA5, keeping the functions of both proteins after their incorporation into the mimetic systems. AnxA5 was able to incorporate into DPPC-proteoliposomes (11.64 µg/mL), but the presence of DPPS increased significantly the AnxA5 incorporation (25.79 µg/mL) into DPPC:DPPS 10%-proteoliposomes. The presence of both proteins into DPPC and DPPC:DPPS 5, 10 and 15% (molar ratios) proteoliposomes was confirmed by SDS-PAGE and Immunoblotting analysis. Better yield of TNAP and AnxA5 incorporation was observed when both proteins were reconstituted simultaneously. DPPC-proteoliposomes and DPPC:DPPS 10%-proteoliposomes (molar ratio) incorporated about 75% of AnxA5 and 25% of TNAP (protein concentration). DPPS presence did not affect significantly the yield of incorporation of both proteins. The kinetic parameters for the hydrolysis of different physiological substrates (ATP, ADP and PPi) by TNAP were determined in the presence and absence of AnxA5, at physiological pH, for the different systems. The best catalytic efficiencies were achieved with proteoliposomes containing DPPS 10% (molar ratio) (kcat/K0.5= 183.02; 776.06 and 657.08 M-1.s-1 for ATP, ADP and PPi, respectively), condition that also favored PPi hydrolysis by TNAP when compared to ATP and ADP hydrolysis. Studies by Differential Scanning Calorimetry (DSC) showed that the increasing DPPS concentrations in the DPPC-liposomes resulted in a progressive broadening of the phase transition peaks and decreased t1/2 and H values. The pre-transition was detected only in concentrations up to DPPS 15% in DPPC. Phase lateral segregation can be observed for DPPS 20% and above, suggesting the formation of DPPS-rich microdomains. The interaction of AnxA5 with DPPC and DPPC:DPPS 10%-liposomes resulted in a decrease of H values (from 8.73 to 5.68 and from 8.43 to 5.37 Kcal.mol-1, respectively). When TNAP was present in the proteoliposomes, this effect was even greater. AnxA5 incorporated into DPPC and DPPC:DPPS 10%-proteoliposomes (molar ratio) was able to mediate 45Ca2+-influx (~ 800 nmol Ca2+) into the vesicles at physiological Ca2+-concentrations (~ 2 mM), and this process was not affected by the presence of TNAP in the systems. However, AnxA5 affected significantly the hydrolysis of substrates by TNAP. Studies with Giant Unilamellar Vesicles (GUVs) also confirmed the functional reconstitution of AnxA5 in dioleoylphosphocholine (DOPC) and DOPC:DPPS 10% (molar ratio) vesicles. The main effect caused by AnxA5 in the GUVs morphology was the formation of pores in the vesicles membrane. In this case, DPPS presence did not affect the AnxA5 incorporation. The presence of TNAP in GUVs caused a several fluctuation, indicating that the vesicles acquired an excess of area and undergoes sequential budding transitions. It is suggested that the presence of DPPS makes the TNAP insertion into the GUVs membrane difficult. With the presence of heterogeneous lipid microdomains in GUVs composed of DOPC, Cholesterol (Chol), Sphingomyelin (SM) and Ganglioside (GM1) in the proportions DOPC:Chol:SM 8:1:1 and DOPC:Chol:SM:GM1 7:1:1:1 (molar ratios), the TNAP insertion caused a greater phase lateral segregation, evidenced by fluorescence analysis. Atomic Force Microscopy (AFM) analysis indicated that the presence of both proteins into DPPC:DPPS 10%-proteoliposomes (molar ratio) caused significant changes in the visco-elastic mechanical properties of the vesicles. In conclusion, the present work describes the synthesis of proteoliposomes harboring TNAP and AnxA5 concomitantly, with the functional reconstitution of both proteins, with the ability to transport Ca2+ into the vesicles and hydrolyze phosphosubstrates on their surface.

The cellular membrane is a complex biological entity, far from being an inert assembly of protein and lipids which separates cells from the surrounding environment. A multitude of biological processes, ranging from active transport of ions into and out of the cell, to the immune response, are regulated at the level of the plasma membrane. The understanding of their molecular basis is among the central goals of modern biological research. In order to dissect the complexity of actual cell membranes, which involves a very complex network of intermolecular interactions, a “divide and conquer” strategy proves very useful. To this end, researchers try to isolate molecules from complex biological contexts to understand their function in simple model systems under controlled conditions. A variety of model membranes have thus been developed in order to gain insight into membrane processes. This approach has resulted in a deeper knowledge on how lipids and proteins interact and how these interactions govern the function of cellular membranes. In the recent past in fact, a connection has been established between the lateral structure of the plasma membrane and its biological function. Furthermore, a large range of biophysical techniques have been used to characterize protein-lipid microdomains. For example, atomic force microscopy (AFM) is a powerful technique which allows a highly detailed topographical characterization of lipid domains in physiological conditions. While AFM imaging offers an extremely high spatial resolution, up to the nanometer scale, the limited image acquisition speed (minutes) can pose a severe drawback in adequately studying fast dynamic processes. On the other hand, fluorescence based imaging techniques are much faster (10-3-100 s), but certainly lack the high spatial resolution that AFM offers. FCS in particular can also provide information about dynamic processes, like diffusion of fluorescent membrane components. For these reasons, implementing a combination of the above mentioned techniques on the same sample (e.g. cell membrane models) would prove extremely beneficial in the complete dynamic and structural characterization of molecular interactions. . The work described in this thesis can be summarized in two main points: i) the development of a novel combined approach of atomic force microscopy (AFM), laser scanning imaging (LSM), and fluorescence correlation spectroscopy (FCS) and ii) the study of the effects of ceramide in the lateral organization of model plasma membranes. We described one of the first simultaneous applications of AFM and FCS on biologically relevant systems. More specifically, model membranes showing complex phase separation were investigated with a combined approach of AFM, confocal fluorescence imaging, force measurements and FCS, based on commercially available instruments. AFM conveys information about the structural and mechanical properties of the different lipid phases. Different membrane domains can be distinguished based on height difference, elastic properties and line tension as measured by the AFM tip. Simultaneous optical measurements offer the correlation of these data in real time with the partition behavior and diffusion of fluorescent lipids and proteins. We established a clear link between the local membrane viscosity, probed by FCS, and the lipid-lipid interactions involved in line tension, probed by AFM force measurements. An example of a significant drawback circumvented by the AFM-FCS approach involves the use of AFM micromanipulation to eliminate unwanted interactions between lipid particles — similar to intra-cellular vesicles found in vivo experiments — and the membrane, which usually result in distorted FCS autocorrelation curves. Finally, the combined application of AFM and FCS on membrane-anchored proteins reconstituted in lipid bilayers has been instrumental in clarifying inconsistencies that arose in work that focused solely on either AFM or fluorescence techniques. We have shown that, in the case of proteins diffusing in the plane of the membrane, AFM can unambiguously detect only a small immobile fraction. Furthermore, since AFM detection of proteins might be facilitated by high local membrane viscosity (e.g. in ordered lipid phases), the measurement of protein partition between disordered and ordered membrane domains might be biased toward the latter. In this case, the use of FCS as a complementary technique allows a more thorough investigation and deeper understanding of the system of interest. The second part of this thesis dealt with the study of complex lipid mixtures which are used to model the putative lipid/proteins domains in cells, called “rafts”. Firstly, we proved how the combined fluorescence imaging/AFM approach is useful in general for studying supported lipid membranes and the role of lipid domains in biological contexts. We investigated the effect of environmental stress on biological membranes and the protective effects of several substances. Our experimental approach was shown to be a new valuable method to visualize the dehydration damage and its effects on the lateral organization of lipid domains. Our results demonstrated that disaccharides like trehalose or sucrose are effective in protecting lipid membranes, not only on a macroscopic scale — preserving the overall integrity of the bilayer — but also on a microscopic scale, preventing the clustering of microdomains. These phenomena are interesting in the context of biological damage to living cells which need to be stored for long time, like organs to be transplanted or blood platelets. Finally, a large section of this thesis focused on the effects of a specific lipid called “ceramide” on the lateral organization of proteins and lipids in the plasma membrane. Ceramide is produced by cells in several situations, like bacterial infections or apoptosis. As consequence of ceramide production in vivo, the local concentration and the dynamic behavior of lipids and membrane receptors are supposed to exhibit strong variations. In order to understand the molecular mechanisms responsible for these effects, we applied a combination of AFM, FCS and fluorescence imaging on simple model membranes containing ceramide. We could show for the first time that, in presence of raft-like Lo/Ld phase separation, physiological quantities of ceramide induced the formation of a highly ordered gel phase, constituted of ceramide and sphingomyelin. The enzymatic production of ceramide was monitored both in supported and in free-standing bilayers. In the second case, ceramide production was connected to selective vesicle budding from the raft-like phase. Since short-chain analogues are often used in both medical applications and biochemical research to mimic the effect of long-chain ceramides, we investigated the effect of chain-length on ceramide-induced membrane reorganization. We could show that only long-chain ceramides (C18 and C16) form highly ordered domains. Interestingly, FCS measurements indicated that the physical properties of the Lo raft-like domains are hardly affected by the presence of ceramide domains. Furthermore, the increased thickness of the Ld phase — as measured by AFM — and its higher viscosity — as measured by FCS — strongly support the hypothesis of ceramide-induced cholesterol displacement from rafts. On the other hand, short-chain ceramides showed completely different biophysical properties that lead to a destabilization of the raft domains, possibly acting as surfactants between the different lipid phases. Our findings contribute to the explanation of in vivo experiments where short-chain ceramides inhibit cell signaling by disrupting the lipid order in the plasma membrane. We have so far demonstrated that ceramide plays a fundamental role in lipid-lipid interactions. In a physiological context, it is also known to produce dramatic effects in living cells. Since a majority of the processes in vivo are thought to be governed by the activity of proteins, it is highly likely that ceramide not only affects lipid organization but also modifies protein-protein and protein-lipid interactions to produce its effects. To test this hypothesis, we reconstituted several membrane proteins in lipid bilayers containing Ld, Lo, and ceramide-rich domains. We were able to show that some membrane proteins are sorted into ceramide-rich domains. More specifically, the raft-associated proteins we tested were enriched in the highly ordered ceramide-rich domains, while the Ld-associated components were excluded from them. Furthermore, the inclusion of any membrane component in ceramide-rich domains is directly connected to a dramatic reduction of its in-plane diffusion. In an in vivo context, such a reorganization of membrane receptors might be used by the cell to alter the signaling process, for example, by i) separating raft receptors from inhibitors with lower raft affinity, ii) bringing both raft-associated receptors and raft-associated signaling molecules into contact, or iii) stabilizing the interactions between a receptor and its ligand by decreasing their diffusion coefficients. In conclusion, this thesis describes a novel combination of AFM, LSM, and FCS for the investigation of the lateral organization of biological membranes. Our results show that this approach applied on model membranes of increasing complexity is an effective tool for understanding the molecular mechanisms behind the organization of biological membranes. This report opens up new possibilities for further investigation in living cell membranes using the same methodology we have described.

Aims: T1α/(podoplanin) is abundantly expressed in the alveolar epithelial type I cells (ATI) of rodent and human lungs. Caveolin-1 is a classical primary structural protein of plasmalemal invaginations, so-called caveolae, which represent specialized lipid rafts, and which are particularly abundant in ATI cells. The biological functions of T1α in the alveolar epithelium are unknown. Here we report on the characteristics of raft domains in the microplicae/microvillar protrusions of ATI cells, which contain T1α. Methods: Detergent resistant membranes (DRMs) from cell lysates of the mouse epithelial ATI-like cell line E10 were prepared using different detergents followed by flotation in a sucrose gradient and tested by Western and dot blots with raft markers (caveolin-1, GM1) and nonraft markers (transferrin receptor, PDI and β-Cop). Immunocytochemistry was employed for the localization of T1α in E10 cells and in situ in rat lungs. Results: Our biochemical results showed that the solubility or insolubility of T1α and caveolin-1 differs in Triton X-100 and Lubrol WX, two distinct non-ionic detergents. Caveolin-1 was unsoluble in both detergents, whereas T1α was Triton X-100 soluble but Lubrol WX insoluble. Immunofluorescence double stainings revealed that both proteins were colocalized with GM1, while caveolin-1 and T1α were not colocalized in the plasma membrane. Cholesterol depletion modified the segregation of T1α in Lubrol WX DRMs. Cellular processes in ultrathin sections of cultured mouse E10 cells were immunogold positive. Immunoelectron microscopy (postembedding) of rat lung tissue revealed the preferential localization of T1α on apical microvillar protrusions of ATI cells. Conclusion: We conclude that T1α and caveolin-1 are located in distinct plasma membrane microdomains, which differ in their protein-lipid interactions. The raft-associated distribution of T1α may have an impact on a specific, not yet clarified function of this protein in the alveolar epithelium. / Dieser Beitrag ist mit Zustimmung des Rechteinhabers aufgrund einer (DFG-geförderten) Allianz- bzw. Nationallizenz frei zugänglich.