Ca2+ entry through voltage-gated calcium channels (CaVs) triggers a range of physiological events, including synaptic neurotransmission and muscular excito-contraction coupling. CaVs are often localised to discrete membrane microdomains and are required to be targeted to such fine structures in order to perform their cellular functions. CaVs are multi-subunit protein complexes that consist of a core, pore-forming α1 subunit and auxiliary β and α2/δ subunits. The α2/δ subunit is required for the optimal cell surface expression and function of CaVs and is itself localised to cholesterol-rich membrane microdomains called lipid rafts. What is unclear is whether the α2/δ subunit is required for whole CaV complexes to be localised to lipid rafts and what effects lipid raft association has on the cell surface distribution and function of CaVs. By a combination of cellular imaging, biochemistry and electrophysiology, this project shows that the auxiliary α2/δ-1 subunit is both necessary and sufficient to target CaV2.2 to lipid rafts in the COS-7 cell heterologous expression system (Robinson et al, 2010). In addition, α2/δ is localised at the cell surface in discrete puncta and co-localises with two endogenous lipid raft resident proteins, caveolin and flotillin-1. While the punctate cell surface distribution of α2/δ is co-incident with that of caveolin and flotillin-1, its distribution is not dependent on cellular cholesterol, but rather the integrity of the actin cytoskeleton. Additional structure-function analysis by employment of the pIN-α2/δ series of deletion and substitution mutants has shown that the association of α2/δ with lipid rafts is bestowed by an extracellular region of the delta peptide, contrary to other evidence supporting the notion that α2/δ may be a GPI-anchored protein. The exact physiological and functional significance of α2/δ and CaV association with lipid rafts remains poorly understood, but the fact that CaVs are enriched within these fine structures provides a potential mechanism for targeting and access to lipid raft associated signalling pathways.

Lipid rafts are operationally defined as cholesterol-rich membrane microdomains resistant to solubilization in nonionic detergents at low temperatures. Lipid rafts, which are quite different in lipid composition from the surrounding membranes, are of great importance to signal transduction, protein sorting and membrane transport. They have been implicated in a range of biosynthetic and endocytic processes and systems-signaling, molecular trafficking, diseases as well as being involved in the immune, vascular, digestive and reproductive systems. Dietary nutrients like fatty acids and vitamins of different types also play a critical and decisive role in the regulation of lipid rafts. / text

Myelodysplastic syndromes (MDS) include a spectrum of stem cell malignancies characterized by ineffective hematopoiesis and predisposition to acute myeloid leukemia (AML) transformation. Patients are predominantly older (greater than 60 years old), with progressive cytopenias resulting from ineffective and cytologically dysplastic hematopoiesis. MDS subtypes are classified by morphologic features and bone marrow blast percentage, as well as cytogenetic pattern, as is the case for deletion 5q MDS. Interstitial deletion of the long arm of chromosome 5, del(5q), is the most common chromosomal abnormality in patients with MDS, and the 5q- syndrome, represents a distinct subset of del(5q) MDS characterized by an isolated deletion, megakaryocyte dysplasia, hypoplastic anemia, and an indolent natural history. MDS risk stratification is most commonly based on the International Prognostic Scoring System (IPSS) with survival outcomes ranging from a few months to many years based on risk factors. There are several therapeutic options for MDS including hematopoietic growth factors, immunosuppressive therapy, azanucleosides, and allogeneic stem cell transplant, however, there is still a need for more effective treatment options, particularly targeted therapeutics. One of the most effective treatments for MDS is selective for del(5q) MDS, and is the second generation immunomodulatory agent, lenalidomide (LEN). LEN is an analog of the known teratogen, thalidomide, and has broad biological effects including selective cytotoxicity to del(5q) clones, activation of T-cells, and expansion of erythroid precursors. In patients with del(5q) MDS, LEN is effective in up to 75% of patients, however, 50% of patients will become resistant within 2-3 years of treatment response. Studies in normal hematopoietic progenitors have shown that LEN induces expansion of the primitive erythroid precursors, which our laboratory has shown is accompanied by sensitization of progenitors to ligand induced erythropoietin receptor (EpoR) signaling. This sensitization is evidenced by increased and prolonged activation of the Signal Transducer and Activator of Transcription 5 (STAT5), compared to Epo stimulation alone. Although EpoR signaling is augmented by LEN, the exact mechanisms by which this is mediated to result in erythroid expansion are not fully characterized. In del(5q) MDS, we have shown that LEN selectively suppresses del(5q) clones via inhibition of the haploinsufficient phosphatases Cdc25c and PP2a, as well as stabilizing the human homolog of the murine double minute-2 protein (MDM2) to decrease expression of the tumor suppressor, p53, however, the mechanisms of action of LEN in non-del(5q) MDS remains elusive. Although most anemic MDS patients have normal or elevated endogenous levels of Epo, as well as comparable levels of progenitor EpoR density relative to healthy individuals, the biologic pathology underlying the impaired EpoR signaling in MDS is poorly defined. Recent reports have shown that membrane microdomains are important for T-cell, c-kit, and integrin signaling, however, there have been no reports on EpoR membrane localization. Lipid rafts are discrete membrane entities that provide platforms by which receptors aggregate and initiate downstream signaling. Furthermore, reports have indicated that there is a decrease in lipid raft density in GM-CSF primed MDS neutrophils, that consequently impaired production of reactive oxygen species (ROS) after fMLP stimulation, suggesting a role of rafts in MDS disease biology. Based on the role of rafts in signaling, and potential role in MDS pathogenesis, we sought to determine whether there was specific membrane localization of EpoR to the raft fractions, and whether disruption of rafts in MDS erythroids could impair EpoR signaling. To address this, we first examined the membrane localization of EpoR on the cell surface. We show here that EpoR translocates to lipid rafts in both erythroid progenitor cell lines as well as primary progenitor cells after stimulation by Epo. Furthermore, we found that Epo stimulation increases the assembly of lipid rafts, as well as the aggregation of rafts on the cell surface. Epo stimulation not only promoted the recruitment of EpoR into the raft fractions, but also downstream signaling intermediates such as Janus kinase 2 (Jak2), STAT5, and Lyn kinase. Moreover, a negative regulator of EpoR signaling, the CD45 tyrosine phosphatase, was redistributed outside of raft fractions after Epo stimulation, potentially enhancing receptor signal competence. Furthermore, disruption of lipid rafts by depletion of membrane cholesterol with MâCD (methyl-β-cyclodextrin) inhibited EpoR signaling in both cell lines and primary bone marrow progenitor cells. Additionally, we found that inhibition of Rho-associated, coiled-coil containing protein kinase (ROCK) and/or Ras-related C3 botulinium toxin substrate 1 (Rac1), blocked the recruitment of the receptor into the raft fractions indicating a critical role of these GTPases, and associated proteins, in the transport and localization of EpoR into raft microdomains. We next asked whether LEN could alter lipid raft assembly in erythroid precursors in the absence of Epo. LEN not only induced raft formation and aggregation but also increased F-actin polymerization. Similar to Epo stimulation, LEN alone was able to induce the recruitment of EpoR, Jak2, and STAT5 into raft fractions. Additionally, CD45 was redistributed outside of raft fractions after LEN treatment. Similarly, inhibition of ROCK blocked LEN induced raft formation and F-actin polymerization, indicating that LEN utilized effectors shared by Epo. Furthermore, LEN was able to increase raft density in raft deficient primary MDS erythroid progenitors. These data demonstrate that LEN may enhance erythroid expansion via induction of EpoR signaling competent raft platforms, to enhance survival and differentiation transcriptional response. Recently, ribosomal protein (RP), S-14, gene (RPS14) haplodeficiency was found to be a key determinant of the hypoplastic anemia in del(5q) MDS. Allelic loss of RPS14 compromises ribosome assembly, thereby causing nucleolar stress and release of free RPs that bind to and promote the degradation of MDM2, the principal negative regulator of p53. As a result, the accumulation of RPs causes lineage restricted stabilization of p53 in erythroid precursors. Our laboratory and colleagues confirmed that cellular p53 expression levels were elevated in del(5q) erythroid precursors, and that LEN decreased expression in responding patients. However, at the time of LEN treatment failure, p53 expression was again elevated at levels exceeding those at baseline. These results suggest that LEN is initially able to reverse p53 accumulation levels and that this action may be a mechanism by which LEN is selectively cytotoxic to del(5q) clones. Subsequent studies showed that LEN inhibits the cereblon E3 ubiquitin ligase complex, the newly discovered target of LEN. Cereblon has been reported to be the principal protein involved in thalidomide induced teratogenicity. Furthermore, the cytotoxic activity of LEN in multiple myeloma is dependent on cereblon. Our laboratory found that LEN inhibits the auto-ubiquitination of MDM2, thereby stabilizing the protein, and promoting ubiquitination of and ultimately the degradation of p53. Additionally, we found that LEN blocked the binding of free ribosomal proteins to MDM2, which are liberated from the nucleosome by ribosomal stress from RPS14 haploinsufficiency, consequently stabilizing the E3-ubiquitin ligase and fostering p53 degradation. In non-del(5q) MDS there is no cytotoxicity of MDS clones by LEN, suggesting an alternative method of erythropoiesis rescue. Although we know that LEN promotes the formation of signaling platforms, and recruitment of EpoR, we wished to determine whether there was an effect of LEN on EpoR expression, as EpoR expression is controlled through ubiquitination and proteasomal degradation. Treatment of erythroid progenitor cell lines and primary erythroid precursors with LEN increased cellular expression of Jak2-associated EpoR in a concentration dependent manner. There was no change in mRNA expression, supporting a post transcriptional mechanism. We then investigated whether receptor up-regulation was limited to EpoR, or included other cytokine receptors. We found that LEN induced expression of another Jak2 associated Type I receptor, IL3-R, but did not alter cellular expression of c-kit, a Type II cytokine receptor. Because Type I cytokine receptor turnover is regulated by a shared E3-ubiquitin ligase, and LEN inhibited both MDM2 and cereblon, we evaluated the effects of LEN on the E3-ubiquitin ligase, Ring Finger Protein-41 (RNF41), which regulates steady state or ligand independent, Jak2 associated Type I receptor internalization. We found that LEN inhibited the ubiquitination activity of RNF41, ultimately stabilizing EpoR membrane residence and increasing expression. In summary, MDS patients display ineffective hematopoiesis likely in part to decreased lipid raft assembly. Stimulation by Epo, or treatment by LEN, not only induced raft formation, but also induced the recruitment of both growth factor receptor, and downstream signaling intermediates into raft fractions to enhance EpoR signal fidelity. We have shown here two methods by which LEN may augment EpoR signaling. First, LEN increases lipid rafts and promotes recruitment of signaling effectors. Second, LEN increases and stabilizes the expression of EpoR through inhibition of the E3 ubiquitin ligase, RNF41. Therefore, we suggest here that LEN may have broad E3 ubiquitin ligase inhibitory effects. These data also indicate that lipid raft upregulation by LEN is mediated through GTPases, suggesting that GTPase activation may also occur via inhibition of specific E3 ubiquitin ligases, a question to be addressed in future studies.

HIV gp120 binds CD4+ cells within plasma membrane lipid rafts inducing a conformational change in gp120 that exposes its V3 loop that binds to a chemokine co-receptor, also within lipid rafts, and initiates fusion. Glycosphingolipids (GSLs) may also be bound by gp120. Lipid rafts, enriched with GSLs and cholesterol, are required for HIV entry and therefore the binding of gp120 to GSL-containing vesicles has been studied. Most of the GSL-structures were within the theoretical raft fraction on a discontinuous sucrose gradient while gp120 binding occurred outside of this fraction where a minority of structures migrated. Gb3 fatty acid content modulated binding. Gp120 bound preferentially to structures depleted of cholesterol and binding was enhanced by treating gp120 with CD4. Two water-soluble mimics of Gb3 inhibited gp120 binding to the different structures. The results demonstrate that the aglycone modulation of GSLs alters their receptor function and that the soluble mimics inhibit binding.

HIV gp120 binds CD4+ cells within plasma membrane lipid rafts inducing a conformational change in gp120 that exposes its V3 loop that binds to a chemokine co-receptor, also within lipid rafts, and initiates fusion. Glycosphingolipids (GSLs) may also be bound by gp120. Lipid rafts, enriched with GSLs and cholesterol, are required for HIV entry and therefore the binding of gp120 to GSL-containing vesicles has been studied. Most of the GSL-structures were within the theoretical raft fraction on a discontinuous sucrose gradient while gp120 binding occurred outside of this fraction where a minority of structures migrated. Gb3 fatty acid content modulated binding. Gp120 bound preferentially to structures depleted of cholesterol and binding was enhanced by treating gp120 with CD4. Two water-soluble mimics of Gb3 inhibited gp120 binding to the different structures. The results demonstrate that the aglycone modulation of GSLs alters their receptor function and that the soluble mimics inhibit binding.

Many crucial cellular processes take place at the plasma membrane. The latter is a complex, two-dimensional medium exhibiting significant lateral structure. As a result, a number of non-classical processes, including anomalous diffusion, compartimentalisation and fractal kinetics take place at the membrane surface. The evaluation of various hypotheses and theories about the membrane is currently very difficult because no general modelling framework is available. In this thesis, we present a stochastic, spatially explicit Monte Carlo model for the plasma membrane that accounts for illmixedness, mobile lipid microdomains, fixed proteins, cytoskeletal fence structures and other interactions. We interrogate this model to obtain three classes of results, regarding (1) the effect of lipid microdomains on protein dynamics on the membrane (2) the effects of microdomains, cytoskeletal fences and fixed proteins on the nature of the (anomalous) diffusion on the membrane and (3) the effects of obstructed diffusion on reaction kinetics at the membrane. We find that the presence of lipid microdomains can lead to nonclassical phenomena such as increased collision rates and differences between long-range and short-range diffusion coefficients. Our results also suggest that experimental techniques measuring long-range diffusion may not be sufficiently discriminating and hence cannot be used to infer quantitative information about the presence and characteristics of microdomains. With regard to anomalous diffusion in particular, we find that to explain this phenomenon at the levels observed in vivo, a number of interactions are required, including (but not necessarily limited to) obstacle-induced diffusion and segregation, or exclusion from microdomains. The effects of these different interactions upon the nature of the diffusion appear to be approximately additive. Finally, we show that a widely used non-spatial method, the Stochastic Simulation Algorithm, can be modified to take into account anomalous diffusion and that this significantly increases its predictive accuracy. The model presented in this thesis is expected to be of future value in evaluating different models of cell surface processes.

Many crucial cellular processes take place at the plasma membrane. The latter is a complex, two-dimensional medium exhibiting significant lateral structure. As a result, a number of non-classical processes, including anomalous diffusion, compartimentalisation and fractal kinetics take place at the membrane surface. The evaluation of various hypotheses and theories about the membrane is currently very difficult because no general modelling framework is available. In this thesis, we present a stochastic, spatially explicit Monte Carlo model for the plasma membrane that accounts for illmixedness, mobile lipid microdomains, fixed proteins, cytoskeletal fence structures and other interactions. We interrogate this model to obtain three classes of results, regarding (1) the effect of lipid microdomains on protein dynamics on the membrane (2) the effects of microdomains, cytoskeletal fences and fixed proteins on the nature of the (anomalous) diffusion on the membrane and (3) the effects of obstructed diffusion on reaction kinetics at the membrane. We find that the presence of lipid microdomains can lead to nonclassical phenomena such as increased collision rates and differences between long-range and short-range diffusion coefficients. Our results also suggest that experimental techniques measuring long-range diffusion may not be sufficiently discriminating and hence cannot be used to infer quantitative information about the presence and characteristics of microdomains. With regard to anomalous diffusion in particular, we find that to explain this phenomenon at the levels observed in vivo, a number of interactions are required, including (but not necessarily limited to) obstacle-induced diffusion and segregation, or exclusion from microdomains. The effects of these different interactions upon the nature of the diffusion appear to be approximately additive. Finally, we show that a widely used non-spatial method, the Stochastic Simulation Algorithm, can be modified to take into account anomalous diffusion and that this significantly increases its predictive accuracy. The model presented in this thesis is expected to be of future value in evaluating different models of cell surface processes.

Many crucial cellular processes take place at the plasma membrane. The latter is a complex, two-dimensional medium exhibiting significant lateral structure. As a result, a number of non-classical processes, including anomalous diffusion, compartimentalisation and fractal kinetics take place at the membrane surface. The evaluation of various hypotheses and theories about the membrane is currently very difficult because no general modelling framework is available. In this thesis, we present a stochastic, spatially explicit Monte Carlo model for the plasma membrane that accounts for illmixedness, mobile lipid microdomains, fixed proteins, cytoskeletal fence structures and other interactions. We interrogate this model to obtain three classes of results, regarding (1) the effect of lipid microdomains on protein dynamics on the membrane (2) the effects of microdomains, cytoskeletal fences and fixed proteins on the nature of the (anomalous) diffusion on the membrane and (3) the effects of obstructed diffusion on reaction kinetics at the membrane. We find that the presence of lipid microdomains can lead to nonclassical phenomena such as increased collision rates and differences between long-range and short-range diffusion coefficients. Our results also suggest that experimental techniques measuring long-range diffusion may not be sufficiently discriminating and hence cannot be used to infer quantitative information about the presence and characteristics of microdomains. With regard to anomalous diffusion in particular, we find that to explain this phenomenon at the levels observed in vivo, a number of interactions are required, including (but not necessarily limited to) obstacle-induced diffusion and segregation, or exclusion from microdomains. The effects of these different interactions upon the nature of the diffusion appear to be approximately additive. Finally, we show that a widely used non-spatial method, the Stochastic Simulation Algorithm, can be modified to take into account anomalous diffusion and that this significantly increases its predictive accuracy. The model presented in this thesis is expected to be of future value in evaluating different models of cell surface processes.

Many crucial cellular processes take place at the plasma membrane. The latter is a complex, two-dimensional medium exhibiting significant lateral structure. As a result, a number of non-classical processes, including anomalous diffusion, compartimentalisation and fractal kinetics take place at the membrane surface. The evaluation of various hypotheses and theories about the membrane is currently very difficult because no general modelling framework is available. In this thesis, we present a stochastic, spatially explicit Monte Carlo model for the plasma membrane that accounts for illmixedness, mobile lipid microdomains, fixed proteins, cytoskeletal fence structures and other interactions. We interrogate this model to obtain three classes of results, regarding (1) the effect of lipid microdomains on protein dynamics on the membrane (2) the effects of microdomains, cytoskeletal fences and fixed proteins on the nature of the (anomalous) diffusion on the membrane and (3) the effects of obstructed diffusion on reaction kinetics at the membrane. We find that the presence of lipid microdomains can lead to nonclassical phenomena such as increased collision rates and differences between long-range and short-range diffusion coefficients. Our results also suggest that experimental techniques measuring long-range diffusion may not be sufficiently discriminating and hence cannot be used to infer quantitative information about the presence and characteristics of microdomains. With regard to anomalous diffusion in particular, we find that to explain this phenomenon at the levels observed in vivo, a number of interactions are required, including (but not necessarily limited to) obstacle-induced diffusion and segregation, or exclusion from microdomains. The effects of these different interactions upon the nature of the diffusion appear to be approximately additive. Finally, we show that a widely used non-spatial method, the Stochastic Simulation Algorithm, can be modified to take into account anomalous diffusion and that this significantly increases its predictive accuracy. The model presented in this thesis is expected to be of future value in evaluating different models of cell surface processes.

Os mastócitos são células multifuncionais do sistema imunológico que participam em diversos processos biológicos. As funções dos mastócitos estão diretamente relacionados com a sua ativação e, subsequente, liberação de mediadores químicos. Os eventos iniciais da ativação dos mastócitos e da transdução de sinais ocorrem em microdomínios lipídicos (lipid rafts) da membrana plasmática. Os gangliosídeos derivados do GD1b são constituintes dos lipid rafts de mastócitos de roedores. O intercruzamento destes gangliosídeos pelo mAb AA4, resulta na formação de agregados (caps) na superfície celular e promove uma ativação parcial dos mastócitos, sem que ocorra a desgranulação. A ativação é semelhante a observada quando os FcRIs são intercruzados por antígenos multivalentes ligados a IgEs, mas neste caso ocorre a desgranulação. O presente estudo tem como objetivo caracterizar o papel dos gangliosídeos derivados do GD1b na liberação de mediadores de mastócitos da linhagem RBL-2H3. O intercruzamento dos gangliosídeos derivados do GD1b resulta na ativação dos fatores de transcrição NFAT e NFB e esta ativação é mediada pela proteína quinase Syk. A ativação destes fatores de transcrição resulta na liberação de mediadores neo-sintetizados, tais como: TNF-, interleucina (IL)-4. Por outro lado, o intercruzamento dos gangliosídeos derivados de GD1b não induz a liberação dos mediadores neoformados como o leucotrieno B4 (LTB4) e o leucotrieno C4 (LTC4). A agregação dos gangliosídeos derivados do GD1b resulta na desorganização dos lipid rafts e na redistribuição de seus componentes, como demostrado pela análise proteômica. Estes dados mostraram proteínas capazes de desencadear uma ativação parcial dos mastócitos e proteínas reguladoras negativas da desgranulação estão up reguladas, enquanto que proteínas críticas para a transdução do sinal estão down reguladas. Os resultados obtidos neste trabalho demonstram que os gangliosídeos derivados do GD1b desempenham papel crucial na integridade dos lipid rafts modulando a ativação e liberação de mediadores de mastócitos. / Mast cells are immunoregulatory cells that participate in diverse biological events. The action of mast cells is directly related to their activation and subsequent mediator release. Early signal transduction events occur in lipid rafts in the plasma membrane. GD1b-derived gangliosides are known constituents of lipid rafts in rodent mast cells. The cross-linking of these gangliosides by mAb AA4 results in a partial activation of mast cells similar to that observed when FcRIs are cross-linked, but does not result in the mast cell degranulation. With time, the gangliosides bound to mAb AA4 cap on the cell surface. The present study aims to characterize the role of the rodent mast cell specific gangliosides derived from GD1b in mediator release from RBL-2H3 mast cells. Cross-linking the GD1b-derived gangliosides activated the transcription factors NFAT and NFB and this activation was mediated by Syk. The activation of theses transcription factors by cross-linked GD1b-derived gangliosides results in the release of the neo-synthesized mediators TNF- and interleukin (IL)-4. However, cross-linking GD1b-derived gangliosides did not stimulate release of the newly formed mediators leukotriene B4 (LTB4) and leukotriene C4 (LTC4). Capping of GD1b-derived gangliosides disorganized lipid rafts and resulted in a redistribution of lipid raft components. Proteomic analysis showed that proteins that trigger mast cell activation and negative regulatory proteins of degranulation are up regulated, whereas proteins critical for signal transduction are down regulated in mast cells where the gangliosides are capped. The results of this work demonstrate that the mast cell-specific GD1b-derived gangliosides are crucial in maintaining the functional integrity of the lipid rafts and modulate cell activation and subsequent mediator release from mast cells.