Protozoan parasites of the genus Leishmania undergo a complex life cycle involving transmission by biting sand flies and replication within mammalian macrophage phagolysosomes. A major component of the Leishmania surface coat is the glycosylphosphatidylinositol (GPI)-anchored polysaccharide called lipophosphoglycan (LPG). LPG has been proposed to play many roles in the infectious cycle, including protection against complement and oxidants, serving as the major ligand for macrophage adhesion, and as a key factor mitigating host responses by deactivation of macrophage signaling pathways. However, all structural domains of LPG are shared by other major surface or secretory products, providing a biochemical redundancy that compromises the ability of in vitro tests to establish whether LPG itself is a virulence factor. To study truly lpg(-) parasites, we generated Leishmania major lacking the gene LPG1 [encoding a putative galactofuranosyl (Gal(f)) transferase] by targeted gene disruption. The lpg1(-) parasites lacked LPG but contained normal levels of related glycoconjugates and GPI-anchored proteins. Infections of susceptible mice and macrophages in vitro showed that these lpg(-) Leishmania were highly attenuated. Significantly and in contrast to previous LPG mutants, reintroduction of LPG1 into the lpg(-) parasites restored virulence. Thus, genetic approaches allow dissection of the roles of this complex family of interrelated parasite virulence factors, and definitively establish the role of LPG itself as a parasite virulence factor. Because the lpg1(-) mutant continue to synthesize bulk GPI-anchored Gal(f)-containing glycolipids other than LPG, a second pathway distinct from the Golgi-associated LPG synthetic compartment must exist.