Malaria is a leading cause of morbidity and mortality, estimated to cause >1 million childhood deaths annually. Plasmodium falciparum causes the most severe form of the disease. There is as yet no licensed vaccine for this disease, despite over a half century of research. In this study, we investigated a transmission-blocking vaccine candidate, the ookinete surface protein Pfs25. Antibodies against Pfs25, drawn in during a bite, can block parasite development in the mosquito midgut, preventing transmission to other individuals. Pfs25 is a low-molecular-weight protein, by itself not immunogenic. To increase its immunogenicity, we investigated several methods of conjugating Pfs25 to itself and to other proteins: recombinant Pseudomonas aeruginosa exotoxin A, and ovalbumin, using amide, hydrazone, or thioether linkages. All conjugates were immunogenic and induced booster responses in mice. The scheme to form amide bonds between proteins by using adipic acid dihydrizide as a linker produced the most immunogenic conjugates. Adsorption of the conjugates onto aluminum hydroxide further increased the antibody response. Remarkably, the antibody levels 3 or 7 months after the last injection were significantly higher than those 1 wk after that injection. The observed transmission-blocking activity of immune sera correlated with antibody levels measured by ELISA. malaria ͉ vaccine O ne approach for a vaccine against malaria is to block transmission of the parasite from mosquitoes to humans. When ingested by a mosquito with the blood meal, antibodies against the sexual and the mosquito stage-specific surface antigens can block parasite development inside the vector (1). Four proteins have been identified as potential inducers of transmission-blocking antibodies (2-5). Two of these are expressed on the surface of gametes and in intracellular gameotocytes. The other two are the Plasmodium falciparum surface proteins with apparent molecular masses of 25 kDa (Pfs25) and 28 kDa (Pfs28), expressed exclusively on the zygote and ookinete surfaces during the mosquito stage of the infection. No antibody response to these two proteins has been shown in people infected with malaria and living in endemic countries (6). Pfs25 from different parts of the world has shown minimal variation in its amino acid sequence (7). This relative homogeneity, likely a result from not being subjected to immune pressure in the human host, makes Pfs25 an attractive candidate for a malaria transmission-blocking vaccine (8). Pfs25 is poorly immunogenic in mice and in humans, even if administered with adjuvant (9, 10). In this article, we show that Pfs25 bound onto itself or onto another protein induced high levels of transmission-blocking antibodies in mice.