Relatively little is known about the vesicular stomatitis virus (VSV) glycoprotein G fusion mechanism. Vesicular stomatitis virus has a single type 1 integral membrane glycoprotein G embedded in the viral membrane. It is the only viral protein required for VSV induced low pH mediated fusion. Mutations in four regions (H2, A5, A4 and HI0) of the VSV G ectodomain have been shown to abolish the fusion activity of the viral glycoprotein (Li et al.,l993). One region H2 (a.a 117-139) has been suggested to be the fusion peptide (Zhang and Ghosh, 1994)(Fredericksen and Whitt, 1995). Amino acids 59-221 of the G protein, an area that encompasses the H2 region, has recently been shown to interact with liposomes through hydrophobic photolabeling experiments (Durrer et al., 1995), suggesting that the H2 region (fusion peptide)is able to interact with hydrophobic target bilayers at low pH. A soluble VSV G protein lacking the transmembrane anchor and cytoplasmic tail of VSV G is not fusogenic, suggesting that G must be anchored to the plasma membrane to promote syncytia (Florkiewicz and Rose, 1984). To better understand the steps involved in the fusion mechanism of VSV G it is important to identify domains within the protein that are involved in the fusion process. To determine the contributions of the transmembrane anchor and cytoplasmic tail to the VSV fusion mechanism chimeric G proteins were constructed. The transmembrane anchor alone or in conjunction with the cytoplasmic tail ofVSV G was replaced with equivalent domain from other viral proteins, HSV-1 glycoproteins gB and gD, adenovirus E3 11.6 K gene, that are not involved in low-pH fusion and the cellular protein CD4. All chimeras were expressed in COS-1 cells, glycosylated, oligomerized, transported to the cdl surface, showed a low-pH induced conformational change and were expressed on the cell surface at levels equivalent to wild-type G. The transmembrane hybrids show extensive syncytia formation at levels similar to wild-type G when induced at pH 5.6. The transmembrane-cytoplasmic tail hybrids showed reduced levels of syncytia as compared to wild-type Gat both pH 5.6 and 5.2. A glycosylphosphatidylinositollipid-anchored ectodomain of G (GGPI), which lacks both the transmembrane and cytoplasmic tail ofG, was expressed in COS-1 cells. The GGPI chimera was glycosylated, expressed on the cell surface,and oligomerized similar to wild-type G. However the chimera was fusion negative, could not promote lipid mixing and h~,d an altered tryptic digestion profile. A fusion negative chimera Gt12gBwas constructed by exchanging the TM of G with the equivalent domain from HSV-1 gB TM plus eight extra amino acids of the gB ectodomain. Deletion of the 11 extra gB amino acids (GgB3G) restored the fusogenic activity of this chimera. Another chimera G 10 DAF directly demonstrated that the fusion negative phenotype of GGPI, like chimera Gtii1Lll2gB, was a result of the 10 extra amino acids at the EC-TM interface. The ectodomain (EC)-transmembrane (TM) interface is highly conserved among 5 vesiculoviruses. Chimeras with a 9 amino acid insertion (GlODAF), deletion (G~9) or replacement (G~910DAF) were expressed in COS-1 cells. The expressed proteins were glycosylated, underwent a low-pH induced conformational change and were expressed on the cell surface at levels equivalent to wild type, but were fusion negative. Suggesting that both the sequence and spatial arrangement of amino acids at the EC-TM interface may affect VSV G fusion. Taken together the data suggests that the specific amino acid sequence of the transmembrane anchor of VSV G is not essential for fusion. Replacement of the TM of VSV G with equivalent domains from other viral and cellular proteins does not affect the fusion activity. The cytoplasmic tail of VSV G may form an entity alone or in conjunction with the transmembrane anchor that can regulate fusion. Another region in the ectodomain of VSV G renders the glycoprotein fusion sensitive in a cell-cell fusion assay and was characterized at the EC-TM interface. / Thesis / Master of Science (MS)

Ausgangspunkt dieser vorliegenden Arbeit waren bisherige Untersuchungen unseres Arbeitskreises zum Memb-ranverankerungsverhalten (Phospholipidmembranen, LUV) von Nukleosiden und Oligonukleotiden, welche einen lipophilen Anker an der 5-Position der Pyrimidin- oder an der 8-Position der Purinbase tragen. Diese Nukleolipide ankern gut in der Membran, stehen aber nicht mehr für eine Watson-Crick-Basenpaarung an der Phasengrenzfläche zu Verfügung. Demnach wurde durch die Verwendung unterschiedlicher Reaktionen (Veresterung, Thioetherbildung, Carbamoylverknüpfung oder „Clickreaktion“ zu Triazolen) und verschiedener funktioneller Gruppen (Hydroxy, Thiohydroxy, Azid, Amin) an die 2´-Position der Nukleoside eine Reihe von lipophilen Resten (Alkylketten, Cholesterol, Pyren) eingeführt. Diese Konjugate verankerten ebenfalls gut in den Membranen und es zeigten sich erste Hinweise, dass durch die Einführung eines Spacers zwischen dem Nukleosid und dem lipophilen Anker, eine Basenpaarung an der Phasengrenzfläche möglich ist. Weiterhin zeigte es sich, dass Nukleolipide mit nur einem lipophilen Rest nicht stabil in Membranen verankern, vor allem, wenn dieser nicht verzweigt ist. Bei der Anwendung von Oligonukleotiden zum Ankern in Membranen ist es unbedeu-tend, an welcher Stelle der lipophile Rest am Nukleotid vorkommt, denn zum einen geht das entsprechende Nukleolipid selbst keine Basenpaarung ein und zum anderen erfolgt keine Basenpaarung über dieses hinweg. Für biotechnologische Anwendungen konnte mit Hilfe dieser synthetisierten lipophilen Oligonukleotide gezeigt werden, dass zwei vesikelmembranverankerte Oligonukleotide, welche komplementäre Enden tragen, eine Doppelhelix miteinander bilden und so diese beiden Vesikel auf einen definierten Abstand halten können. Da Nukleolipide einen amphiphilen Charakter aufweisen, sollte unter dem AFM untersucht werden, ob diese supramolekulare Strukturen zeigen. Dies wurde in der Tat auch beobachtet. Ebenso konnten mittels der LB-Technik LB-Schichten aus Nukleolipiden dargestellt werden. / The starting point of this work was found in our previous studies about anchoring behaviour of lipidated nucleo-sides and oligonucleotides in biocompatible phospholipid membranes (LUV). That nucleosides and oligonucleotides bear a lipophilic anchor at the 5-position of pyrimidine or at the 8-position of purinbases. This nucleolipi-des anchor well in such membranes, but were not longer available for a Watson-Crick base pairing at the interface to water. Therefore lipophilic groups (alkyl chain, cholesterol, Pyren etc.) were now connected to the 2''-position of nucleosides by several reactions (esterification, thioether binding, carbamoyl binding or "click reaction") and various functional groups (hydroxy, thiohydroxy, azide, amine) to the 2´-position of nucleosides. These nucleolipides also well anchored in the model membranes, and gave first evidence that by introducing a spacer between the nucleoside and the lipophilic anchor a base pairing at the interface to water is possible. However, only one anchor is not sufficient for a stable anchoring in the phospholipid membranes, especially if they are not branched. It was found out that it is insignifacant for the application of oligonucleotides in membrane anchoring, at which position of nucleotide the lipid is attached, because on the one hand, the corresponding nucleolipid can not form a pair with a corresponding nucleobase and secondly, there is no base pairing in the nucleotides situated between two lipidated positions. For biotechnology applications it might be interesting that two different vesicles each of it furnushed with a complementary lipidated oligonucleotide could be kept together in a defined distance by forming double strand DNA. Since nucleolipide possess amphiphilic character, there abillity to form supramolecular structures was investigated by atomic force microscope (AFM). In addition formation of LB-layers could be achieved by LB-technology.