There is no effective vaccine against Buruli ulcer. In experimental footpad infection of C57BL/6 mice with M. ulcerans, a prime-boost vaccination protocol using plasmid DNA encoding mycolyltransferase Ag85A of M. ulcerans and a homologous protein boost has shown significant, albeit transient protection, comparable to the one induced by M. bovis BCG. The mycolactone toxin is an obvious candidate for a vaccine, but by virtue of its chemical structure, this toxin is not immunogenic in itself. However, antibodies against some of the polyketide synthase domains involved in mycolactone synthesis, were found in Buruli ulcer patients and healthy controls from the same endemic region, suggesting that these domains are indeed immunogenic. Here we have analyzed the vaccine potential of nine polyketide synthase domains using a DNA prime/protein boost strategy. C57BL/6 mice were vaccinated against the following domains: acyl carrier protein 1, 2, and 3, acyltransferase (acetate) 1 and 2, acyltransferase (propionate), enoylreductase, ketoreductase A, and ketosynthase load module. As positive controls, mice were vaccinated with DNA encoding Ag85A or with M. bovis BCG. Strongest antigen specific antibodies could be detected in response to acyltransferase (propionate) and enoylreductase. Antigen-specific Th1 type cytokine responses (IL-2 or IFN-γ) were induced by vaccination against all antigens, and were strongest against acyltransferase (propionate). Finally, vaccination against acyltransferase (propionate) and enoylreductase conferred some protection against challenge with virulent M. ulcerans 1615. However, protection was weaker than the one conferred by vaccination with Ag85A or M. bovis BCG. Combinations of these polyketide synthase domains with the vaccine targeting Ag85A, of which the latter is involved in the integrity of the cell wall of the pathogen, and/or with live attenuated M. bovis BCG or mycolactone negative M. ulcerans may eventually lead to the development of an efficacious BU vaccine.