Thesis (Ph. D. in Microbiology and Immunology)--Vanderbilt University, May 2007. / Title from title screen. Includes bibliographical references.

Arsenic exposure in humans has been implicated in the development of a myriad of non-cancerous and cancerous diseases. A reductionist approach to understanding this unusual phenomenon would suggest that arsenic-induced perturbation of a small number of fundamental biological processes could manifest as this diverse array of disease endpoints. Endocytosis is a fundamental cellular process involved in the internalization and transport of various extracellular molecules and membranous components. BEAS-2B, a human bronchial epithelial cell line, was used to characterize the effects of chronic arsenite exposure on endocytosis. Fluorophore-labeled bovine albumin, human transferrin, and human low-density lipoprotein (LDL) were the substrates utilized to measure endocytosis in BEAS-2B cells. The uptake of albumin in unexposed BEAS-2B cells is both dose-dependent and temperature sensitive. Chronic arsenite exposure in BEAS-2B cells increased the uptake of albumin by 3.4-fold after 8 hours of uptake relative to unexposed BEAS-2B cells. Pharmacological studies utilizing endocytosis inhibitors suggested that the uptake of albumin in both unexposed and arsenite-exposed BEAS-2B cells occurs through a combination of receptor-mediated endocytosis and macropinocytosis. Chronic arsenite exposure in BEAS-2B cells also increased the endocytic uptake of transferrin by 2.9-fold at 30 minutes and LDL by 1.3-fold at 2 hours relative to unexposed BEAS-2B cells. Together, the data suggests that chronic arsenite exposure can increase the rate of endocytosis. This novel finding could add mechanistic insight to the conundrum of arsenic-associated human diseases.

Dictyostelium discoideum is a simple model organism that can be used to study endocytic pathways such as phagocytosis and macropinocytosis because of its homology to cells of the mammalian innate immune system, namely macrophages and neutrophils. Consequently, Dictyostelium can also be used to study the process of exocytosis. In our laboratory, we generated Dictyostelium cells lacking superoxide dismutase SodC. Our data suggest that cells that lack SodC are defective in macropinocytosis and exocytosis when compared to wild type cells. In this study I describe a regulatory mechanism of macropinocytosis by SodC via regulation of RasG, which in turn controls PI3K activation and thus macropinocytosis. Our results show that proper metabolism of superoxide is critical for efficient particle uptake, for the proper trafficking of internalized particles, and a timely exocytosis of fluid uptake in Dictyostelium cells.

Lipoprotein binding and metabolism in actively-dividing (subconfluent) and quiescent (postconfluent) bovine aortic endothelial cells (ECs) were qualitatively investigated by fluorescence microscopy using dioctadecylindocarbocyanine-labelled lipoproteins and by indirect immunofluorescence microscopy. LDL and acetylated-LDL (AcLDL) were seen bound to the surfaces of subconfluent ECs (at 4°C or at 37°C), as a random distribution of punctate foci. ECs therefore closely resembled fibroblasts in the distribution of LDL receptors on their surfaces. No binding of LDL was seen on postconfluent EC surfaces by either direct or indirect fluorescence microscopy. The patterns of AcLDL binding on postconfluent ECs resembled those on subconfluent ECs. Intracellular LDL and AcLDL occurred as perinuclear accumulations of large fluorescent disc-shaped profiles in subconfluent ECs. These accumulations were shown to arise from surface-bound material by pulse-chase experiments. Intracellular LDL was absent in the majority of postconfluent ECs, while AcLDL accumulation was massive. "Wounding" of cultures allowed simultaneous assessment of lipoprotein metabolism in quiescent and actively-dividing areas of the same culture. Quantitative assessments of the above-mentioned phenomena were made using ¹²⁵I-labelled lipoproteins. Receptor-mediated binding of LDL decreased five to ten-fold as the cultures modulated from subconfluent to postconfluent morphology. No receptor-bound LDL was detected in postconfluent ECs. Conversely, the amount of AcLDL bound increased at least fivefold during EC growth in parallel cultures. The amounts of lipoproteins endocytosed and metabolised were generally related proportionately to the amounts bound in each case. The distribution of LDL receptors on cultured cells was also investigated at the ultrastructural level using colloidal gold-conjugated LDL as a probe, and similarly labelled antibodies as probes. Whole-mounted cells with receptor probes bound to them were examined directly in the transmission electron microscope. The topographical distribution of LDL receptors has not been investigated by these techniques before. A novel method of preparing cytochemically-labelled, whole-mounted cells from styrene culture dishes was developed and used in this study. LDL Receptors expressed on the surfaces of human skin fibroblasts served to standardise these colloidal gold techniques and fortuitously led to new information on receptor distribution. Normal (FGo) and LDL receptor-negative mutant fibroblasts (GM 2000) acted as positive and negative controls respectively. Normal fibroblast LDL receptors were grouped into clusters consistent in size with coated pits (200 - 500 nm in diameter). A novel finding was the presence of a diffuse population of receptors scattered randomly amongst the clustered receptors. Another mutant fibroblast, GM 2408A, known to have an aberrant LDL receptor distribution, was also examined. Its receptors were shown to be dispersed singly, and in occasional groups of two and three, at random over the cell surfaces. No clusters were detected. The receptor-negative GM 2000 bound virtually no probes. While not as sensitive as the colloidal gold-conjugated LDL probe, an antireceptor monoclonal antibody (IgG-C7), localised by indirect immunogold labelling, gave similar results when applied to the above cells. This was taken as strong corroborative evidence that the LDL receptor distributions as determined by colloidal gold-conjugated LDL were correct. It is suggested that the dispersed population of receptors on normal fibroblasts may represent newly-emerged recycling receptors which have yet to cluster in coated pits. A further new finding reported here is the existence of the same two patterns of LD L receptors, dispersed and clustered, on the surface of subconfluent ECs. It was noted, from the study of whole-mounted and thin-sectioned cells, that the receptors were preferentially arranged in rings following the circumference of coated pit areas on the cell surface. Often these rings associated in groups or even coalesced into compound clusters. The significance of these groupings is not yet understood. In sharp contrast to the situation on subconfluent ECs, no LDL receptors (probed with the extremely sensitive colloidal-gold conjugated LDL) could be detected at the EM level on the surface of postconfluent ECs. Active cells in wounded postconfluent monolayers expressed abundant receptors detected at the EM level. It is concluded that postconfluent quiescent bovine aortic ECs in vitro metabolise virtually no LDL via the LDL-receptor pathway due to a vanishingly low number of LDL receptors. This contrasts with the ability of postconfluent cells to metabolise relatively large amounts of AcLDL via a receptor-mediated mechanism. The significance of these conclusions is discussed with respect to the interaction of plasma lipoproteins with the endothelium in vivo.

Extracellular components can be internalized by either receptor-mediated or fluid-phase endocytosis. Receptor-mediated endocytosis involves the internalization of receptor-ligand complexes into coated vesicles of about 0.1 μm in diameter. The average diameter of primary pinocytic vesicles has been calculated to be 0.24 - 0.28 μm. The discrepancy in size between coated vesicles and the average pinosome diameter can be explained if, in addition to coated vesicles, another endocytic process involving vesicles larger than 0.28 μm in diameter takes place. These two vesicle types could together produce an average diameter of 0.24 μm. This hypothesis suggests that coated vesicles cannot fully account for fluid-phase uptake. Hypertonic conditions can selectively inhibit receptor-mediated endocytosis, leaving fluid-phase uptake unaffected, again suggesting that an alternative to coated pit-mediated uptake exists. In this study we determined the volume-weighted average diameter of primary pinocytic vesicles under hypertonic conditions (0.52 osm) where receptor-mediated uptake of transferrin was selectively inhibited by 42%. Fluid-phase uptake of FITC-dextran was unaffected by 0.52 osm medium. The internalization rate of ³H-galactose-labelled plasma membrane was reduced from 2.6 %/min to 1.5 %/min. The decrease in the rate of membrane internalization, without a reduction in the rate of fluid uptake at hypertonicity, implied a reduced surface to volume ratio of the pinocytic vesicles formed under these conditions. This suggested an increase in the average diameter of primary pinocytic vesicles. Membrane internalization rates were calculated on the assumption that all labelled cell-surface constituents were internalized to the same relative extent, as has been shown previously for isotonic conditions. This assumption was also shown to hold true under isotonic conditions. The reduced rate of membrane internalization under hypertonic conditions was shown not to be due to the exclusion of any labelled protein species from internalized vesicles. The larger average vesicle size determined under conditions of selective reduction of coated vesicle formation (i.e. hypertonicity), demonstrates the existence of a population of larger pinosomes involved in a possible alternative mechanism to coated-pit-mediated endocytosis.

Necroptosis is a pathway of inflammatory cell death that is associated with several pathologies and is induced by ligation of surface TLR or cytokine receptors in macrophages. Many signaling pathways depend on endocytosis, a process mediated by GTPases such as dynamin. We evaluated the role of dynamin-dependent endocytosis in the necroptosis of macrophages using various dynamin inhibitors. Using flow cytometry, we confirmed that during necrosome signaling, various dynamin inhibitors (e.g. Dyngo 4a and Dynasore) blocked the internalization of TLR4, which also resulted in the inhibition of cytokine production. Despite the similar impact of Dynasore and Dyngo 4a on TLR4 endocytosis and cytokine production, only Dyngo 4a prevented TLR4-induced necroptosis of macrophages. Further studies indicated that Dyngo 4a was a potent stimulator of the p38 MAPK pathway, and activation of this pathway by Dyngo 4a was responsible for the inhibition of necroptosis of macrophages following TLR4 signaling. Thus, these studies reveal the previously unknown role of the p38 MAPK pathway in regulating the activation of necrosome signaling.

The present thesis looks at the behaviour of internalised cell surface-derived membrane marker in comparison with the behaviour of endocytosed HRP (horse-radish peroxidase) as a fluid-phase contents marker. The pooling and/or segregation in the endosome was measured by determining co-localization with HRP. Colocalization of the two markers in the endosome is studied by using the ability of HRP to catalyse the crosslinking of membrane marker in endosomes with DAB (3,3'-diaminobenzidine), rendering the membrane marker detergent insoluble. To study the kinetic behaviour of membrane marker, radioactive galactose was covalently bound to cell-surface glycoconjugates on mouse macrophage-cells, P388D₁, as catalysed by galactosyltransferase. This provided a general membrane marker. After endocytosis-derived redistribution of membrane marker between the cell surface and endosomal membrane, a steady state was established with about 16% of the label on internal membranes. The bulk of the label on the cell surface was removable by subsequent treatment with β-galactosidase.

Yes / The activity-regulated cytoskeleton-associated (Arc) protein control synaptic strength by facilitating AMPA receptor (AMPAR) endocytosis. Here we demonstrate that Arc targets AMPAR to be internalized through a direct interaction with the clathrin-adaptor protein 2 (AP-2). We show that Arc overexpression overexpression in dissociated hippocampal neurons obtained from C57BL/6 mouse reduces the density of AMPAR GluA1 subunits at the cell surface and reduces the amplitude and rectification of AMPAR-mediated miniature-excitatory postsynaptic currents (mEPSC). Mutations of Arc, that prevent the AP-2 interaction reduce Arc-mediated endocytosis of GluA1 and abolish the reduction in AMPAR-mediated mEPSC amplitude and rectification. Depletion of the AP-2 subunit µ2 blocks the Arc-mediated reduction in mEPSC amplitude, effect that is restored by re-introducing µ2. The Arc/AP-2 interaction plays an important role in homeostatic synaptic scaling as the Arc-dependent decrease in mEPSC amplitude, induced by a chronic increase in neuronal activity, is inhibited by AP-2 depletion. This data provides a mechanism to explain how activity-dependent expression of Arc decisively controls the fate of AMPAR at the cell surface and modulates synaptic strength, via the direct interaction with the endocytic clathrin adaptor AP-2. / This work was supported by the BBSRC_FAPPA BB/J02127X/1 and BBSRC-BB/H018344/1 to SALC and by the FAPESP_RCUK_FAPPA 2012/50147-5 and FAPESP_Young Investigator’s grant 2009/50650-6 to LLdS. SCW was a PhD Student supported be the BBSRC/GSK PhD-CASE Studentship, LPdA is a postdoc fellow supported by FAPESP, YCJ was supported by a FAPESP scientific initiation scholarship.

The novel protein Carom is a homolog of the Drosophila protein, Nervous Wreck (Nwk). Carom was cloned as a homocysteine response protein in primary endothelial cells (EC). In order to discern the functionality of the protein in higher order mammals and potential role in homocysteine-inhibited angiogenesis, we performed basic characterization studies on Carom. Theoretical modeling of Carom matched the solved structure of Nwk. Using a lab-generated antibody against Carom’s F-BAR domain, it is evident that Carom is localized to the mitochondria and speckling in the nucleus of primary ECs. In order to perform biochemical and structural studies in primary ECs, Carom was cloned from an adenoviral shuttle vector to an adeno-associated virus (AAV) transfer vector. We created a multi-cistronic open reading frame with an N-terminal Flag and a cleavable C-terminal green fluorescent protein (mClover3). Primary ECs are difficult to transfect, so we optimized an AAV packaging system in order to get high titers and high purity AAV that can transduce our primary ECs. Carom is composed of several functional domains with the capability of binding to cell membranes and act as a scaffolding for attaching adaptor proteins. To isolate which domains are important to the partner binding and cellular localization, we serially truncated the domains from Carom starting from both the C-terminus and N-terminus. We demonstrated that the C-terminal region features some post-translational modifications creating the second band in western blots with lower mobility. Also, the F-BAR domain is responsible for translocalization of Carom from the cytoplasm to the cell membrane and nucleus. A novel mechanism is proposed for why Carom is upregulated in response to homocysteine (Hcy), an independent risk factor for cardiovascular disease. It is previously known that Hcy inhibits angiogenesis. Our data mining studies identified a potentially important receptor for angiogenesis in ECs, VEGFR2, being endocytosed and ultimately degraded. Through biotinylation assays, we determined that Carom does help enhance the endocytosis of VEGFR2 potentially leading to degradation via the lysosome. In summary, Carom is endogenously localized to the mitochondria in primary ECs, the C-terminus is post-translationally modified, the bipartite nuclear localization signal containing F-BAR domain localizes to the cell membrane and nucleus, and Carom enhances the endocytosis of VEGFR2. / Biomedical Sciences

Tube formation, or tubulogenesis, is an elaborate form of epithelial morphogenesis that includes processes such as cell migration and cell shape changes. The embryonic Drosophila heart, or dorsal vessel, is an excellent model of tubulogenesis and more specifically the signaling mechanisms required for cell migration and lumen formation. Similar to vertebrate heart formation, Drosophila heart tubulogenesis begins with the collective migration of cardioblasts that meet at the midline and adhere at specialised junctions, enclosing a lumen between them. Roundabout, and its ligand Slit, are required to restrict cell-to-cell adhesions to the junctional domains of contralateral cardioblasts, as well as maintain the integrity of the lumen. The localisation patterns of Robo, and other luminal cell surface receptors important for lumen formation are significantly modified throughout heart formation. Initial receptor expression is broadly distributed over the cardioblast surface. Receptors are then relocalised to specific cell surface domains by late embryonic development. The mechanisms by which Robo and other cell surface receptors are localised have yet to be determined. Endocytosis is a promising mechanism by which cell surface receptors are targeted and trafficked to cell surface domains. Specifically, vesicular trafficking proteins, such as Rab GTPases, are molecular switches that regulate endocytic events. Here, we investigated the roles of Rab5, Rab11, and Sec6 during heart formation. Of these, only Rab5, a regulator of the early endosome, was required for lumen formation. Particularly, gain of function, loss of function, and overexpression of rab5 resulted in reduced lumen phenotype, characterised by lumen pockets rather than a continuous lumen along the anterior-posterior axis. Perturbed Rab5 function also resulted in the mislocalisation of Robo at the basal domain. Live imaging showed that expression of rab5 dominant negative, constitutively active, and overexpression constructs did not perturb apical membrane motility of migrating cardioblasts in the developing heart. / Thesis / Master of Science (MSc)