Metabolic energy is most efficiently stored as triacylglycerol (TAG). This neutral lipid accumulates mainly within adipose tissues, but it can be stored and used in all types of cells. Within cells it is packed in organelles called lipid droplets (LDs). They consist of a core of neutral lipids like TAG and cholesterol esters, which is surrounded by a phospholipid monolayer that mainly consists of phosphatidylcholine (PC). Attached to or inserted into this monolayer are various proteins, mainly LD specific structural proteins or lipid metabolic enzymes. Though excess uptake of nutrition leads to lipid accumulation in all kinds of body tissues, which is accompanied by the augmentation of LDs and results in cellular dysfunction and the development of metabolic diseases, relatively little is known about the biogenesis and growth of LDs. This thesis focuses on diacylglycerol acyltransferase 2 (DGAT2), an enzyme of the TAG biosynthetic pathway, and on lyso-phosphatidylcholine acyltransferases 1 and 2 (LPCAT1 and LPCAT2), both enzymes of one of the PC biosynthetic pathways called Lands cycle. The data presented in this thesis show that these enzymes can localize to LDs and that they actively synthesize TAG and PC at the surface of LDs. While the LPCATs reside on LDs independent from the nutrition status of the cell, DGAT2 accumulates on LDs upon excess availability of oleic acid. DGAT2, LPCAT1 and LPCAT2 differ in their structure from other iso-enzymes that catalyze the same reactions. This thesis shows that they exhibit a monotopic conformation and that they contain a hydrophobic stretch that presumably forms a hairpin. This topology enables them to localize to both a phospholipid bilayer like the membrane of the endoplasmic reticulum and to a phospholipid monolayer like the surface of LDs. The different biophysical properties of the structures of iso-enzymes might be responsible for their subcellular localization and the formation of distinct TAG or PC pools that are destined for different purposes. This would explain, why the iso-enzymes are often not able to replace each other. Knock-down and overexpression experiments performed in this thesis show that the activity of LPCAT1, LPCAT2 and DGAT2 influence the packaging of lipids within LDs. Knock-down of LPCAT1 and LPCAT2 leads to an increase in LD size without concomitant increase in the amount of TAG. Combined with the finding that the profile of the PC species of the LD surface reflects the substrate preferences of LPCAT1 and LPCAT2, the results suggest that these enzymes are responsible for the formation of the LD surface. Therefore, the increase in LD size upon LPCAT1 and LPCAT2 knock-down results from an adjustment of the surface-to-volume ratio in response to reduced availability of surface lipids. The connection between LPCATs and LD size was corroborated in the model organism Drosophila melanogaster. Three different knockout fly strains of the Drosophila homologue of LPCAT1 and LPCAT2, CG32699, exhibit enlarged LDs in the fat body of the L3 larvae. Furthermore, the data presented suggest that the morphology of LDs is important for the secretion of stored lipids. The reduction of LPCAT1 in liver cells leads to a reduction in lipoprotein particle release. This was shown by measuring the amount of released apolipoproteinB with two different methods, by measuring the release of lipids and by quantification of the amount of released hepatitis C virus, which is known to rely on LD interaction for replication and on lipoprotein particles for cellular release. DGAT2 is recruited to LDs upon excess availability of oleic acid and its overexpression leads to the formation of many, but relatively small LDs. Here, it is shown that DGAT2 interacts with acyl-CoA synthetase ligase 1 (ACSL1), an enzyme that catalyzes the activation of free fatty acids with Coenzyme A. This interaction does not influence the stability of DGAT2 nor does it seem to affect lipid synthesis. Nevertheless, it shows an influence on lipid packaging in LDs. While overexpression of DGAT2 results in the appearance of smaller LDs, overexpression of ACSL1 leads to an increase in LD size. Coexpression of ACSL1 and DGAT2 reverses the phenotypes obtained by single overexpression and normalizes the mean LD diameter to values observed at normal conditions. In conclusion, this thesis shows that LDs are able to synthesize the components of their core and their surface, which underlines their independent function in metabolism. Additionally, the results show that LDs can grow by local synthesis and that the responsible enzymes exhibit a monotopic membrane topology, which might be crucial for LD localization. Furthermore, the obtained data suggest that the localization and the ratio between different enzyme activities influence the packaging of lipids and affects lipid secretion and therefore impact the whole body lipid metabolism.

Oxidativ modifizierte Phospholipide haben PAF-ähnliche Struktur und zeigen in vitro PAF-ähnliche Wirkung. Dabei handelt es sich hauptsächlich um fragmentierte Phosphatidylcholine, die eine lange Fettsäure- oder Alkylkette in sn-1 Stellung und einen kurzen Acylrest (C4-C9) mit oxidiertem C-Atom in sn-2 Stellung haben und als PAF-ähnliche Lipide bezeichnet werden. In der vorliegenden Arbeit wurden durch eine neue HPLC-Fluoreszenzmethode ein Vertreter dieser in vivo existierender PAF-ähnlichen Lipide im Plasma quantifiziert. Dieses fragmentierte Phosphatidylcholin (FPC) besitzt einen Palmitoylrest in sn-1 Position, sowie eine C3 oder C4 Kette in sn-2 Position und erfüllt damit die strukturelle Voraussetzungen für biologische Aktivität. Weiterhin konnte gezeigt werden, dass FPC in biologisch aktiven Fraktionen enthalten war, deren biologische Aktivität jedoch nicht von PAF bewirkt wurde, sondern von PAF-ähnlichen Lipiden. Dies unterstützt die Vermutung, dass FPC biologische Aktivität zeigt. Im Plasma wird FPC von Lipoproteinen transportiert. Oxidativer Stress in biologischen Systemen, ausgelöst durch Tabakrauch, Organischämie/ Reperfusion, Vitamin E-Mangel oder Inkubation von Zellen mit H2O2, führte zu einer Erhöhung der FPC-Konzentration. Damit wurde gezeigt, dass die PAF-ähnlichen Lipide in vivo durch Radikal-induzierte Lipidperoxidation gebildet werden können. Trotz eines Konzentrationsanstiegs von FPC nach oxidativem Stress, kam es bei einer systemischen Inflammation, die von Radikal-bildenen Prozessen begleitet ist, jedoch nicht zu einer Anreicherung von FPC-ähnlichen Lipidperoxidationsprodukten. Ursache für die fehlende Anreicherung von oxidativ modifizierten Lipiden bei einer systemischen Inflammation könnte ihr rascher Abbau durch Phospholipasen, oder der Abbau von Vorstufen, bzw. die Substratrminderung, sein. / Oxidative modified phospholipids have PAF like structure and show in vitro PAF like activity. They consist of fragmented phosphatidylcholine with a long fatty acid or alkyl chain in sn-1 position and a short acylrest (C4-C9) with a oxidized C-atom in sn-2 position. According to their structure they are called PAF like lipids. In this paper one representative of these in vivo existing PAF like lipids are quantified in plasma by means of a new fluorescence HPLC method. These fragmented phosphatidylcholines consits of a palmitoylrest in sn-1 position and a C3 or C4 chain in sn-2 position and therefore they fullfill the strutural conditions to have biological activity. We further showed, that the biological active fractions contain FPC. The biological activity is not caused by PAF but by PAF like lipids. This fact supports the hypothesis that FPC are biological active. FPC is carried by lipoproteins in plasma. Oxidative stress caused by cigarette smoke, ischemia/ reperfusion, Vitamin E deficiency or incubation of cells with H2O2 leads to an increase of the concentration of FPC. Thereby it could be shown that PAF like lipids can be formed in vivo by means of radical induced lipidperoxidation. Many radical producing processes are involved in systemic inflammation. Despite of an increase of FPC's concentration after oxidative stress, there was no accumulation of FPC like lipidperoxidation products during systemic inflammation. One reason for the missing accumulation of oxidative modified lipids during systemic inflammation can be the fast degradation by phospholipases or the reduction of substrat.

Metabolic energy is most efficiently stored as triacylglycerol (TAG). This neutral lipid accumulates mainly within adipose tissues, but it can be stored and used in all types of cells. Within cells it is packed in organelles called lipid droplets (LDs). They consist of a core of neutral lipids like TAG and cholesterol esters, which is surrounded by a phospholipid monolayer that mainly consists of phosphatidylcholine (PC). Attached to or inserted into this monolayer are various proteins, mainly LD specific structural proteins or lipid metabolic enzymes. Though excess uptake of nutrition leads to lipid accumulation in all kinds of body tissues, which is accompanied by the augmentation of LDs and results in cellular dysfunction and the development of metabolic diseases, relatively little is known about the biogenesis and growth of LDs. This thesis focuses on diacylglycerol acyltransferase 2 (DGAT2), an enzyme of the TAG biosynthetic pathway, and on lyso-phosphatidylcholine acyltransferases 1 and 2 (LPCAT1 and LPCAT2), both enzymes of one of the PC biosynthetic pathways called Lands cycle. The data presented in this thesis show that these enzymes can localize to LDs and that they actively synthesize TAG and PC at the surface of LDs. While the LPCATs reside on LDs independent from the nutrition status of the cell, DGAT2 accumulates on LDs upon excess availability of oleic acid. DGAT2, LPCAT1 and LPCAT2 differ in their structure from other iso-enzymes that catalyze the same reactions. This thesis shows that they exhibit a monotopic conformation and that they contain a hydrophobic stretch that presumably forms a hairpin. This topology enables them to localize to both a phospholipid bilayer like the membrane of the endoplasmic reticulum and to a phospholipid monolayer like the surface of LDs. The different biophysical properties of the structures of iso-enzymes might be responsible for their subcellular localization and the formation of distinct TAG or PC pools that are destined for different purposes. This would explain, why the iso-enzymes are often not able to replace each other. Knock-down and overexpression experiments performed in this thesis show that the activity of LPCAT1, LPCAT2 and DGAT2 influence the packaging of lipids within LDs. Knock-down of LPCAT1 and LPCAT2 leads to an increase in LD size without concomitant increase in the amount of TAG. Combined with the finding that the profile of the PC species of the LD surface reflects the substrate preferences of LPCAT1 and LPCAT2, the results suggest that these enzymes are responsible for the formation of the LD surface. Therefore, the increase in LD size upon LPCAT1 and LPCAT2 knock-down results from an adjustment of the surface-to-volume ratio in response to reduced availability of surface lipids. The connection between LPCATs and LD size was corroborated in the model organism Drosophila melanogaster. Three different knockout fly strains of the Drosophila homologue of LPCAT1 and LPCAT2, CG32699, exhibit enlarged LDs in the fat body of the L3 larvae. Furthermore, the data presented suggest that the morphology of LDs is important for the secretion of stored lipids. The reduction of LPCAT1 in liver cells leads to a reduction in lipoprotein particle release. This was shown by measuring the amount of released apolipoproteinB with two different methods, by measuring the release of lipids and by quantification of the amount of released hepatitis C virus, which is known to rely on LD interaction for replication and on lipoprotein particles for cellular release. DGAT2 is recruited to LDs upon excess availability of oleic acid and its overexpression leads to the formation of many, but relatively small LDs. Here, it is shown that DGAT2 interacts with acyl-CoA synthetase ligase 1 (ACSL1), an enzyme that catalyzes the activation of free fatty acids with Coenzyme A. This interaction does not influence the stability of DGAT2 nor does it seem to affect lipid synthesis. Nevertheless, it shows an influence on lipid packaging in LDs. While overexpression of DGAT2 results in the appearance of smaller LDs, overexpression of ACSL1 leads to an increase in LD size. Coexpression of ACSL1 and DGAT2 reverses the phenotypes obtained by single overexpression and normalizes the mean LD diameter to values observed at normal conditions. In conclusion, this thesis shows that LDs are able to synthesize the components of their core and their surface, which underlines their independent function in metabolism. Additionally, the results show that LDs can grow by local synthesis and that the responsible enzymes exhibit a monotopic membrane topology, which might be crucial for LD localization. Furthermore, the obtained data suggest that the localization and the ratio between different enzyme activities influence the packaging of lipids and affects lipid secretion and therefore impact the whole body lipid metabolism.