My talk will focus on recent results in my laboratory, concerning the envelope proteins of dengue virus and the discovery of a conserved epitope that is the target of broadly neutralizing antibodies, suggesting that a single immunogen can elicit antibodies that neutralize all four serotypes of dengue virus simultaneously. These results suggest in turn that it should be possible to engineer the dengue virus envelope protein such that it can elicit essentially cross-neutralizing antibodies and minimize the non-protective antibodies that can aggravate the disease via antibody dependent enhancement. In a second part of the talk I will show that the dengue virus envelope protein, which is the prototypic class II viral membrane fusion protein, has homologs in viruses of from unrelated families. Our findings suggest that the same gene has been captured during evolution by viruses that are otherwise totally unrelated. As the corresponding proteins evolved within different virus families, the homology can only be detected by their structure and not by amino acid sequence analyses, even when comparing the protein from different genera within a family. More recently, we have found that cellular proteins involved in cell-cell fusion are also homologous to class II viral membrane fusion proteins – and were likely acquired via a retro-transcription and integration event, as has been shown in the case of syncytin in mammals. I will discuss the possible implications of these finding.