Lipid rafts are nanoscale entities in the membranes of eukaryotic cells which provide a mechanism for the functional membrane segregation vital for several cellular processes. This lateral segregation of specific lipid and protein components provides the facilitative platforms for a variety of signaling and trafficking events at the plasma membrane and in the Golgi. Rafts are distinguished from the surrounding membranes by their physical properties and composition - they are relatively tightly packed and enriched in saturated lipids, sterols, and lipid-anchored proteins. Although the existence of rafts has been conclusively confirmed by several independent techniques, questions concerning various aspects of membrane heterogeneity are still to be addressed. Typical experiments investigating raft composition have been designed to evaluate the affinity of a given component for raft domains. In such experiments, the results are usually interpreted in a Boolean fashion, i.e., the component is either a raft molecule, or not. However, this binary point of view overlooks potential complexity that may underlie the nature of membrane heterogeneity. In this work, we systematically investigated the nature of functional cellular membrane heterogeneity. We started by characterizing the model membranes and fluorescent lipid analogs widely used in research into membrane domains. After extensively evaluating the potentials/limits of these approaches and the artifacts that must be avoided or alternatively could be exploited, we applied these tools to understand whether the cell membrane has multiple kinds of raft domains with distinct compositions and physical properties, rather than only one. We found that cell membranes have the potential to form various kinds of functional domains having different physicochemical properties, compositions, and functional outputs. Therefore, we propose continuously variable lipid packing as the principle of the functional membrane lateral heterogeneity. According to this principle, the membrane is not composed of a single variety of raft domain with strictly defined properties coexisting alongside a specific and uniform non-raft environment; rather it is composed of entities having continuously variable lipid packing. Finally, we show that this spectrum of membrane packing modulates the orientation of membrane lipid receptors, which ultimately influences their specific bioactivity. Our results showing continuously variable lipid packing and its ability to fine-tune the activity of membrane molecules comprise a novel model for the structure and function of eukaryotic membranes.

Intracellular transport of membrane-bound vesicles and organelles is a process fundamental for many cellular functions including cell morphogenesis and signaling. The transport is mediated by ensembles of motor proteins, such as kinesins, walking on microtubule tracks. When transporting membrane-bound cargo inside a cell, the motors are linked to diffusive lipid bilayers either directly or via adaptor molecules. The fluidity of the lipid bilayers induces loose inter-motor coupling which is likely to impact the collective motor dynamics and may induce cooperativity. Here, we investigate the influence of loose coupling of kinesin motors on its transport characteristics. In the first part of this thesis, we used truncated kinesin-1 motors with a streptavidin-binding-peptide (SBP) tag and performed gliding motility assays on streptavidin-loaded biotinylated supported lipid bilayers (SLBs), so called ‘membrane-anchored’ gliding motility assays. We show that the membrane-anchored motors act cooperatively; the microtubule gliding velocity increases with increasing motor density. This is in contrast to the transport behavior of multiple motors rigidly bound to a substrate. There, the motility is either insensitive to the motor density or shows negative interference at higher motor density, depending on the structure of the motors. The cooperativity in transport driven by membrane-anchored motors can be explained as following: while stepping on a microtubule, membrane-anchored motors slip backwards in the viscous membrane, thus propelling the microtubule in the solution at a velocity, given by the difference of the motor stepping velocity and the slipping velocity. The motor stepping on the microtubule occurs at maximal stepping velocity because the load on the membrane-anchored motors is minute. Thus, the slipping velocity of membrane-anchored motors determines the microtubule gliding velocity. At steady state, the drag force on the microtubule in the solution is equal to the collective drag force on the membrane-anchored motors slipping in the viscous membrane. As a consequence, at low motor density, membrane-anchored motors slip back faster to balance the drag force of the microtubule in the solution. This results in a microtubule gliding velocity significantly lower than the maximal stepping velocity of the individual motors. In contrast, at high motor density, the microtubules are propelled faster with velocities equal to the maximal stepping velocity of individual motors. Because, in this case, the collective drag force on the motors even at very low slipping velocity, is large enough to balance the microtubule drag in the solution. The theoretical model developed based on this explanation is in good agreement with the experimental data of gliding velocities at different motor densities. The model gives information about the distance that the diffusing motors can isotropically reach to bind to a microtubule, which for membrane-anchored kinesin-1 is ~0.3 µm, an order of magnitude higher as compared to rigidly bound motors, owing to the lateral mobility of motors on the membrane. In addition, the model can be used to predict the number of motors involved in transport of a microtubule based on its gliding velocity. In the second part of the thesis, we investigated the effect of loose inter-motor coupling on the transport behavior of KIF16B, a recently discovered kinesin motor with an inherent lipid-binding domain. Recent studies based on cell biological and cell extract experiments, have postulated that cargo binding of KIF16B is required to activate and dimerize the motor, making it a superprocessive motor. Here, we demonstrate that recombinant full-length KIF16B is a dimer even in the absence of cargo or additional proteins. The KIF16B dimers are active and processive, which demonstrates that the motors are not auto-inhibited in our experiments. Thus, in cells and cell extracts Kif16B may be inhibited by additional factors, which are removed upon cargo binding. Single molecule analysis of KIF16B-GFP reveals that the motors are not superprocessive but exhibit a processivity similar to kinesin-1 indicating that additional factors are most likely necessary to achieve superprocessivity. Transport on membrane-anchored KIF16B motors exhibited a similar cooperative behavior as membrane-anchored kinesin-1 where the microtubule gliding velocity increased with increasing motor density. Taken together, our results demonstrate that the loose coupling of motors via lipid bilayers provides flexibility to cytoskeletal transport systems and induces cooperativity in multi-motor transport. Moreover, our ‘membrane-anchored’ gliding motility assays can be used to study the effects of lipid diffusivity (e.g. the presence of lipid micro-domains and rafts), lipid composition, and adaptor proteins on the collective dynamics of different motors.

Hedgehog is a lipid-linked morphogen that is carried on lipoprotein particles and that regulates both patterning and proliferation in a wide variety of vertebrate and invertebrate tissues. Hyperactivity of Hedgehog signaling causes numerous forms of cancer. Hedgehog acts by binding to its receptor Patched, relieving the suppression of Smoothened and initiating Smoothened signaling. The mechanism by which Patched represses Smoothened has been unclear, but correlates with reduced Smoothened levels on the basolateral membrane. The structural homology of Patched with the Niemann-Pick-Type C1 protein and bacterial transmembrane transporters suggests that Patched might regulate lipid trafficking to repress Smoothened. However, no endogenous lipid regulators of Smoothened have yet been identified, nor has it ever been shown that Patched actually controls lipid trafficking. This work shows that, in Drosophila melanogaster, the Sterol-Sensing Domain of Patched regulates Smoothened trafficking from Patched-positive endosomes. Furthermore, it demonstrates that Patched recruits internalized lipoproteins to Patched-positive endosomes. Thereby, Patched regulates the efflux of specific lipoprotein-derived lipids from this compartment via its Sterol-Sensing Domain and utilizes these lipids to destabilize Smoothened on the basolateral membrane. We propose that Patched normally promotes Smoothened degradation and subsequently downregulates its activity by changing the lipid composition of endosomes through which Smoothened passes. For this purpose, Patched utilizes a specific lipid – possibly a modified sterol or sphingolipid – derived from lipoproteins. Further, we suggest that the presence of Hedgehog on lipoprotein particles inhibits utilization of their lipids by Patched.

AIMS: Macrophages (MPs) and vascular smooth muscle cells (VSMCs) closely interact within the growing atherosclerotic plaque. An in vitro co-culture model was established to study how MPs modulate VSMC behaviour. METHODS AND RESULTS: MPs were exposed to fluorescence-labelled-acetylated LDL (FL-acLDL) prior to co-culture with VSMCs. Fluorescence microscopy visualized first transport of FL-acLDL within 6 h after co-culture implementation. When MPs had been fed with FL-acLDL in complex with fluorescence-labelled cholesterol (FL-Chol), these complexes were also transferred during co-culture and resulted in cholesterol positive lipid droplet formation in VSMCs. When infected with a virus coding for a fusion protein of Rab5a and fluorescent protein reporter (FP) to mark early endosomes, no co-localization between Rab5a-FP and the transported FL-acLDL within VSMCs was detected implying a mechanism independent of phagocytosis. Next, expression of lysosome-associated membrane glycoprotein 1 (LAMP1)-FP, marking all lysosomes in VSMCs, revealed that the FL-acLDL was located in non-acidic lysosomes. MPs infected with virus encoding for LAMP1-FP prior to co-culture demonstrated that intact fluorescence-marked lysosomes were transported into the VSMC, instead. Xenogenic cell composition (rat VSMC, human MP) and subsequent quantitative RT-PCR with rat-specific primers rendered induction of genes typical for MPs and down-regulation of the cholesterol sensitive HMG-CoA reductase. CONCLUSION: Our results demonstrate that acLDL/cholesterol-loaded lysosomes are transported from MPs into VSMCs in vitro. Lysosomal transfer results in a phenotypic alteration of the VSMC towards a foam cell-like cell. This way VSMCs may lose their plaque stabilizing properties and rather contribute to plaque destabilization and rupture.

All cells are delimited by membranes that protect the cell from the surrounding environment. In eukaryotic cells the same principle applies at subcellular level where membranes delimit functional cell organelles. The membrane structure, properties and function are defined in part by their lipid composition. Lipidomics is the large‐scale study of pathways and networks of cellular lipids in biological systems. It involves the identification and quantitation of cellular lipid molecular species and their interactions with other lipids, proteins, and other metabolites. Lipidomics has been greatly facilitated by recent advances in ionization technology and mass spectrometric capabilities which have simplified the sample processing prior to analysis, giving rise to shotgun lipidomics. Shotgun lipidomics is fast, highly sensitive, and can identify hundreds of lipids missed by other methods. However, Glycosphingolipids are an important lipid family that was out of the scope of shotgun lipidomics due to the lack of suitable analytical tools. The aim of my thesis was two‐fold. The first aim was the establishment of Glycosphingolipid identification and quantification by shotgun approach. This allowed us to perform lipidomic studies with unprecedented comprehensiveness (~300 lipid species from 15 different lipid classes) from low sample amounts and with minimal sample processing. The second was the application of this technology in studies of the role of lipids in several processes like vesicular carrier formation, cell polarization, protein delivery to the plasma membrane and viral budding. This work resulted in several findings. We found that there is sorting of sphingolipids and sterols into plasma membrane targeted vesicular carriers in budding yeast. When kidney cells change from a mesenchymal to an epithelial morphology there is a profound remodeling of their lipidome, with the synthesis of longer, more saturated, more hydroxylated, and more glycosylated sphingolipids. When these sphingolipids and sterols are depleted in epithelial cells, the apical transport in epithelial cells is impaired. These data strongly support the idea that lipid rafts play an important role in sorting and delivery of lipid and protein cargo to the plasma membrane. Finally, we found that the envelopes of vesicular stomatitis virus and Semliki forest virus assert little specificity in the incorporation of lipids from the plasma membrane. This weak specificity seems to be related to a combination of virus lipid bilayer asymmetry and curvature.

The major sorting station of biosynthetic material destined for the cell surface or secretion is the trans Golgi Network, TGN. This organelle sorts proteins and lipids into vesicular transport carriers that are targeted via different pathways to distinct membrane compartments of the cell. The molecular principles that operate in cargo sorting at the TGN are still not very well understood. Especially, we know very little about the sorting of lipids. It was postulated that a sorting mechanism based on clustering of lipid rafts, dynamic membrane domains enriched in sphingolipids and sterols, could be an important part of the picture. My thesis study dealt with the elucidation of the molecular sorting principles at the TGN and their exploitation for cell surface polarity in the yeast Saccharomyces cerevisiae. To this end, we conducted a genome wide screen that identified yeast mutants defective in cell surface delivery of the model cargo protein FusMid-GFP. The most striking result of this screen was that mutant strains with defects in ergosterol (the major yeast sterol) and sphingolipid biosynthesis lost sorting competence. To elucidate a direct role for sphingolipids and ergosterol in cargo sorting and secretion we sought to characterize the lipid composition of secretory vesicles. Hence, we established a vesicle purification protocol based on an immunoisolation strategy. Additionally, in collaboration with the group of A. Shevchenko, we developed a mass spectrometry methodology that allows the comprehensive and quantitative lipid analysis of subcellular organelles. Preliminary results corroborate our genetic evidence. The data show that the vesicles are enriched in sphingolipids and decreased in phosphatidylcholine indicating a role for raft clustering in cargo sorting at the TGN. The studies of cell polarity during yeast mating also unraveled a role for raft clustering. We could identify that the lipid bilayer at the tip of the mating projection was more ordered than at the plasma membrane enclosing the cell body and that this was dependent on sphingolipid synthesis. The results of my thesis suggest that in the yeast Saccharomyces cerevisiae fundamental cell biological processes such as cargo sorting and vesicle formation at the TGN as well as cell surface polarity during mating employ raft clustering mechanisms.

The physical properties of glycerophospholipids (GPLs) are not only determined by the head group (HG), but also by their fatty acid (FA) chains, which affect their distribution and function within membranes in the cell. Understanding the microheterogenity of lipid membranes on a molecular level requires qualitative and quantitative characterization of individual lipids and identification of their FA moieties. The aim of my study was to introduce the new technology of multiple precursor ion scanning (MPIS) on a QSTAR Pulsar time-of-flight mass spectrometer (QqTOF) to analyze lipids. Detailed information on fatty acid composition of individual GPL molecules could be obtained in parallel with conventional profiling of lipid classes, and this could be done by direct analysis of total lipid extracts. This method was termed Fatty Acid Scanning (FAS) and Head Group Scanning HGS, respectively. In this way the molecular GPL composition of total lipid extracts could be charted in a single analysis accurately and rapidly at a low picomole concentration level. Furthermore, combining FAS and HGS together with ion trap MS3 analysis allowed complete charting of the molecular composition of PCs, including quantification of their positional isomers, thus providing a detailed and comprehensive characterization of molecular composition of the pool of PCs. Development of the Lipid Profiler software allowed full automation and rapid processing of complex data, including identification and quantification of molecular GPLs. This approach was evaluated by preliminary applications. First, the molecular composition of PCs of total lipid extracts of MDCK cells and of human red blood cells (RBC) could accurately be charted. Significant presence of positional isomers was observed increasing the total number of individual PC species close to one hundred. Secondly, the molecular PC and SM species distribution in detergent resistant membranes (DRMs) prepared by Triton X-100 DRMs were analyzed and were found to be enriched in distinct GPLs. The distribution in PCs and SMs of Triton X-100 DRMs of RBC were compared with those of the DRMs of MDCK cells. Finally, combining the use of a 96 well plate and a robotic system demonstrated that these analyses can be automated and analyzed with high throughput. This system we termed Shotgun Lipidomics. Taken together, this mass spectrometric methodology provides rapid and detailed insight into the distribution of the molecular GPLs of membranes and membrane sub-fractions.

Proteolytic processing of the amyloid precursor protein (APP) may lead to the formation of the Abeta peptide, the major constituent of amyloid plaques in Alzheimer`s disease. The full-length APP is a substrate for at least 2 different (alpha and beta) proteases ("secretases"). The beta-secretase, BACE, cleaves APP in the first step of processing leading to the formation of the neurotoxic Abeta. BACE competes for APP with alpha-secretase, which cleaves APP within its Abeta sequence, thus precluding Abeta formation. It is thus important to understand how is the access of the alpha- and beta-secretase to APP regulated and how are the individual activities of these secretases modulated. Both these regulatory mechanisms, access to substrate and direct activity modulation, can be determined by the lipid composition of the membrane. Integral membrane proteins (like APP and BACE), can be viewed as solutes in a two-dimensional liquid membrane, and as such their state, and biological activity, critically depend on the physico-chemical character (fluidity, curvature, surface charge distribution, lateral domain heterogeneity etc.) of the lipid bilayer. These collective membrane properties will influence the activity of embedded membrane proteins. In addition, activity regulation may involve a direct interaction with a specific lipid (cofactor or co-structure function). Interactions of membrane proteins are furthermore affected by lateral domain organization of the membrane. Previous results had suggested that the regulation of the activity of the alpha- and beta-secretases and of their access to APP is lipid dependent, and involves lipid rafts. Using the baculovirus expression system, we have purified recombinant human full-length APP and BACE to homogeneity, and reconstituted them in large (~100nm, LUVs) and giant (10-150microm, GUVs) unilamellar vesicles. Using a soluble peptide substrate mimicking the beta-cleavage site of APP, we have examined the involvement of individual lipid species in modulating BACE activity in LUVs of various lipid compositions. We have identified 3 groups of lipids that stimulate proteolytic activity of BACE: 1.cerebrosides, 2.anionic glycerophospholipids, 3. cholesterol. Furthermore, we have co-reconstituted APP and BACE together in LUVs and demonstrated that BACE cleaves APP at the correct site, generating the beta-cleaved ectodomain identical to that from cells. We have developed an assay to quantitatively follow the beta-cleavage in proteoliposomes, and we have shown that the rate of cleavage in total brain lipid proteoliposomes is higher than in phosphatidylcholine vesicles. We have also studied partitioning of APP and BACE in GUVs between liquid ordered (lo) and liquid disordered (ld) phases. In this system, significant part of the BACE pool (about 20%) partitions into the lo phase, and its partitioning into lo phase can be further enhanced by cross-linking of membrane components. Only negligible fraction of APP can be found in the lo phase. We continue to study the behavior of co-reconstituted APP and BACE in GUVs The work presented in this thesis has yielded some interesting results and raised further questions. One of the important assignments of this project will in the next stage be the characterization of the impact of membrane domain organization on the beta-cleavage. Different domain arrangements that can be hypothesized in cell membranes can be modeled by varying the degree of phase fragmentation in proteoliposomes comprising reconstituted APP and BACE.

Endocytosis is one of the most fundamental mechanisms by which the cell communicates with its surrounding. Specific signals are transduced through the cell membrane by a complex interplay between proteins and lipids. Clathrin depended endocytosis is one of important signalling pathways which leads to budding of the plasmalemma and a formation of endosomes. The FCHo2 is an essential protein at the initial stage of the this process. In is a membrane binding protein containing BAR (BIN, Amphiphysin, Rvs) domain which is responsible for a membrane binding. Although numerous valuable work on BAR proteins was published recently, the mechanistic description of a BAR domain functionality is missing. In present work we applied in vitro systems in order to gain knowledge about molecular basis of the activity of the FCHo2 BAR domain. In our studies we used supported lipid bilayers (SLBs) and lipid monolayers as s model membrane system. The experiments were carried out with a minimal number of components including the purified FCHo2 BAR domain. Using SLBs we showed that the BAR domain can bind to entirely flat bilayers. We also demonstrated that these interactions depend on the negatively charged lipid species incorporated in the membrane. We designed an assay which allows to quantify the membrane tubulation. We found out that the interaction of the FCHo2 BAR domain with the lipid membrane is concentration dependent. We showed that an area of the bilayer deformed by the protein depends on the amount of the used BAR domain. In order to study the relation between the mobility of lipids and the activity of FCHo2 BAR domain we designed a small-volume monolayer trough. The design of this micro-chamber allows for the implementation of the light microscopy. We demonstrated that the measured lipid diffusion in the monolayer by our new approach is in agreement with literature data. We carried out fluorescence correlation spectroscopy (FCS) experiments at different density of lipids at the water-air interface.We showed that the FCHo2 BAR domain binding affinity is proportional to the mean molecular area (MMA). We additionally demonstrated that the increased protein binding is correlated with the higher lipid mobility in the monolayer. Additionally, by curing out high-speed atomic force microscopy (hsAFM) we acquired the structural information about FCHo2 BAR domains orientation at the membrane with a high spatio-temporal resolution. Obtained data indicate the BAR domains interact witheach other by many different contact sites what results in a variety of protein orientations in a protein assemble.