Résumé : We have identified previously a Trypanosoma cruzi gene encoding a protein named Tc52 sharing structural and functional properties with the thioredoxin and glutaredoxin protein family involved in thiol-disulphide redox reactions. Furthermore, we have reported that Tc52 also played a role in T. cruzi-associated immunosuppression observed during Chagas' disease. In an effort to understand further the biological role of Tc52, we used a gene-targeted deletion strategy to create T. cruzi mutants. Although T. cruzi tolerates deletion of one wild-type Tc52 allele, deletion of both genes is a lethal event, indicating that at least one active Tc52 gene is required for parasite survival. Monoallelic disruption of Tc52 (Tc52+/-) resulted in the production of T. cruzi lines that express less Tc52 mRNA and produced lower amounts of Tc52 protein compared with wild-type cells. In axenic cultures, growth rates of epimastigote forms bearing an interrupted allele were not different from those of wild-type parasites. Furthermore, monoallelic disruption of the Tc52 gene did not modify the growth rate of epimastigotes or their sensitivity to inhibition by benznidazole and nifurtimox, the two drugs used to treat Chagasic patients. Moreover, the antimonial drug SbIII, which is known, at least in Leishmania parasites, to be conjugated to a thiol and extruded by an ATP-coupled pump, had a similar effect on wild-type and mutant parasites, being equally sensitive. Hence, parasite drug sensitivity was also observed in clones overexpressing the Tc52 protein as well as in those carrying an antisense plasmid construct. Surprisingly, a significant impairment of the ability of epimastigotes carrying a Tc52 single gene replacement or antisense construct to differentiate into metacyclic trypomastigotes and to proliferate in vitro and in vivo was observed, whereas no significant enhancement of these biological properties was seen in the case of parasites that overexpress Tc52 protein. Moreover, functional complementation of Tc52+/- single mutant or selection of antisense revertant clones demonstrated that the phenotype observed is a direct consequence of Tc52 gene manipulation. Taken together, these results may suggest that Tc52 could participate among other factors in the phenotypic expression of T. cruzi virulence.