The objective of this study was to validate an indirect enzyme-linked immunoassay (iELISA) using the recombinant proteins, malate dehydrogenase (MDH) and superoxide dismutase (SOD) [CuZn], as antigens and to evaluate its ability to discriminate antibodies produced by vaccination from those induced by infection in the diagnosis of bovine brucellosis. Sera from six groups were evaluated: G1 - culture-positive animals (52 serum samples) (naturally infected); G2 - non-vaccinated animals (28 serum samples) positive in RBT (Rose Bengal test) and 2ME (2-mercaptoethanol test) selected from brucellosis-positive herds; G3 - animals from a brucellosis-free area (32 serum samples); G4 - S19 vaccinated heifers (114 serum samples); G5 - RB51 vaccinated heifers (60 serum samples); G6 - animals inoculated with inactivated Yersinia enterocolitica O:9 (42 serum samples). Diagnostic sensitivity (DSe) and diagnostic specificity (DSp) were estimated using the frequentist approach and the confidence interval (CI) (95%) calculated by the Clopper-Pearson (exact) method. The DSe for iELISA_MDH in the G1 group was 71.7% (CI 95%: 57.6-83.2%) and for the G2 100.0% (CI 95%: 87.7-100.0%), whereas the DSp was 84.4% in the G3 (CI 95%: 67.2-94.7%). For the iELISA_SOD the DSe was estimated 67.3% for the G1 (CI 95%: 52.9-79.7%) and 71.4% for G2 (CI 95%: 51.3-86.8%), while the DSp for G3 was 87.5% (CI 95%: 71.0-96.5%). iELISA_MDH and iELISA_SOD showed potential to be used in the diagnosis of infected animals, increasing the range of serological tests available for the diagnosis of bovine brucellosis, with the advantage of being S-LPS-free. However, none of the tests could differentiate between infection and vaccination.