Dengue viruses (DVs) cause millions of infections globally each year and yet there is not a vaccine that provides complete protection for all individuals. DVs are antigenically distinct but can evoke cross-reactive immunity to other DVs. Among this cross-reactive response, weakly neutralizing antibodies can exacerbate secondary infections through antibody-dependent endocytosis of the new viruses, which proves to be a challenge for vaccine development. Current live-attenuated virus vaccines are tetravalent formulations of chimeric DV1-4 and fail to elicit a balanced response to all four serotypes. Part of this imbalance response is attributed to uneven replication of each virus component. To bypass this challenge, subunit vaccines are attractive candidates because of their safety and customizability. For DVs, the envelope (E) glycoprotein is a natural subunit vaccine candidate as it is the main target of the antibody-mediated protection. On the virion, the E protein exists as head-to-tail homodimers. Important neutralizing antibodies that target complex epitopes within these homodimers have been identified. However, DV1-4 wildtype (WT) E proteins are unstable as recombinant soluble proteins and do not present neutralizing quaternary epitopes. As a result, they are not ideal as subunit vaccines. In this dissertation, I studied different engineering strategies to improve the E protein as subunit antigens for DV1-4. They include (1) stabilizing the homodimer conformation of the DV1-4 soluble E (sE) protein by computational design, (2) multivalently displaying sE on liposomes to amplify the immune response and (3) selectively presenting neutralizing antibody epitopes on the antigen by masking the surface. Stabilized dimers were characterized and validated by biophysical methods to maintain correct conformations down to low nanomolar concentrations and can bind dimer-epitope antibodies at physiological conditions. Immunogenicity data from mouse vaccinations showed that homodimers were immunogenic and are potential vaccine candidate when coupled to liposomes. However, dimer-epitope antibodies were not successfully elicited in mice after masking the rest of the proteins, suggesting that these antibodies require a different mechanism to form. Nevertheless, the antigen design strategies were successful in improving sE as subunit vaccines.