Dengue viruses (DENV) are mosquito-borne flaviviruses of global importance. DENV exist as four serotypes, DENV1-DENV4. Following a primary infection, individuals produce DENV-specific antibodies that bind only to the serotype of infection and other antibodies that cross-react with two or more serotypes. People exposed to a secondary DENV infection with another serotype are at greater risk of developing more severe forms of dengue disease. The increased risk of severe dengue in people experiencing repeat DENV infections appear to be due, at least in part, to the ability of pre-existing serotype cross-reactive antibodies to form virus-antibody complexes that can productively infect Fcγ receptor-bearing target cells. While the theory of antibody-dependent enhancement (ADE) is supported by several human and small animal model studies, the specific viral antigens and epitopes recognized by enhancing human antibodies after natural infections have not been fully defined. We used antibody-depletion techniques to remove DENV-specific antibody sub-populations from primary DENV-immune human sera. The effects of removing specific antibody populations on ADE were tested both in vitro using K562 cells and in vivo using the AG129 mouse model. Removal of serotype cross-reactive antibodies ablated enhancement of heterotypic virus infection in vitro and antibody-enhanced mortality in vivo. Further depletion studies using recombinant viral antigens showed that although the removal of DENV E-specific antibodies using recombinant E (rE) protein resulted in a partial reduction in DENV enhancement, there was a significant residual enhancement remaining. Competition ADE studies using prM-specific Fab fragments in human immune sera showed that both rE-specific and prM-specific antibodies in primary DENV-immune sera significantly contribute to enhancement of heterotypic DENV infection in vitro. Identification of the targets of DENV-enhancing antibodies should contribute to the development of safe, non-enhancing vaccines against dengue.