Pestiviruses are grouped in the family Flaviviridae, are among the most important pathogens of farm animals worldwide. The members of the genus Pestivirus have a broad host range (mainly pigs and ruminants), and induce a variety of clinical manifestations in farm or wild animals. Even though several good vaccines against the most important pestiviruses have been developed and a series of strict bio-safety measures like quarantine and stamping-out strategies have long been carried out, pestiviruses cause severe financial losses in the animal farming industry. Three envelope proteins are found on the surface of pestiviral virions. Among them, E1 is the least characterized. Due to the absence of specific antibodies directed against E1, both functional and structural information on E1 are still poor. E1 has only been analysed in context with the other two envelope proteins. In the present thesis work, I focused on the functional and structural characterization of the glycoprotein E1 of pestivirus with regard to its intracellular localization, the retention signal, the membrane topology and the oligomerization of E1 and E1-E2 heterodimer formation in order to gain an initial insight into the molecular and cellular biology of this interesting protein. The results of these analyses are also discussed the prerequisites for the steps leading to the assembly and budding of these viruses. First of all, we showed that there is no secretion or cell surface expression of E1. This led to the basic question about the intracellular compartment where E1 is mainly concentrated. By using colocalization analysis with marker proteins, we determined that E1 localizes predominantly in the ER and not Golgi compartment. Since this finding was obtained when E1 was expressed alone, it proved that E1 contains an ER-retention signal of its own. To characterize the determinants for ER retention of E1, a series of VSVg-E1 chimeric and mutated proteins were analysed. The intracellular retention signal was found to map to the putative TM domain (last 30aa at the C-terminal), furthermore, by using site direct mutagenesis analysis, the signal could be narrowed down to six fully conserved polar residues in the middle part of TM domain of E1. Then, the membrane topology of E1 before and after the signal peptide cleavage were determined. By using two independent biological methods, we concluded that E1 is a typical type I transmembrane protein with a hydrophobic membrane anchor at its C-terminus. Interestingly, the pre-cleavage situation is mimicked by blocking the cleavage site between E1 and E2, the transmembrane domain of E1 adopt a hairpin-like structure with the C-terminus located in the ER lumen. The prerequisites for the oligomerization of E1 and heterodimerization of E1-E2 were also explored in this study. Surprisingly, we found that pestiviral E1 formed homotrimers which has never been reported before. Both Cys123 and Cys171 in E1 affect the oligomerization in varying degrees. Co-expression analysis with E1/E2 mutants demonstrated that Cys123 in E1 and Cys295 in E2 are the critical sites for E1-E2 heterodimer formation. Meanwhile, Cys295 in E2 is also determinant for E2 homodimerization. To test for the importance of E1-E2 heterodimer formation for pestivirus viability and replication, we analysed the full-length infectious clone of BVDV CP7 bearing mutations that were not able to form E1-E2 heterodimers. Those BVDV mutants were shown to lose infectivity, further proving that those two sites in E1 and E2 play an essential role in the BVDV life cycle, most likely because of their role in heterodimer formation. In our study, pestiviral E1 exhibited some unexpected characteristics. According to the preliminary data from our lab, we observed an interesting phenomenon that E1 can overrule the retention signal of E2, so that absence of the E1 retention signal directs the heterodimer to the cell surface even though the E2 retention signal is still present. It was shown that the covalent linkage between E1 and E2 plays an essential role for this process. Further, we found that the E1-E2 heterodimer formation is independent of the TM domain of E1, showing a totally different mechanism to that of the closely related Hepatitis C virus. Surprisingly, the complete deletion of the TM region of E1 does not result in the secretion of the protein. We were able to demonstrate that the hydrophobic region in the middle part of E1 most likely binds to the membrane and reduces the secretion of E1 in the absence of the TM domain.