Evaluation of the relatedness ofThe relatedness of Brucella spp. and Ochrobactrum anthropi was studied by protein profiling, Western blot, immunoelectrophoresis and 16s rRNA analysis. Whole-cell and soluble proteins of brucellae and 0. anthropi showed serological cross-reactivities quantitatively and qualitatively more intense than those existing with similar extracts of Agrobacterium spp. Numerical analysis of Western blot profiles of whole-cell extracts showed that 0. anthropi LMG 3301 was closer to Brucella spp. than to 0. anthropi LMG 3331T, a result not obtained by protein profiling. These differences were not observed by Western blot with soluble fractions, and immunoelectrophoretic analyses suggested that this was due to destruction of conformational epitopes in Western blot procedures with the subsequent simplification of antigenic profile. Analysis of the 165 rRNA sequences of strains previously used in the species definition confirmed that strain LMG 3301, and also LMG 3306, were closer to the brucellae, and that LMG 3331T was in a separate cluster. The LMG 3301 and the LMG 3331T clusters could also be separated by their different colistin sensitivity and by PCR with 16s rRNA Brucella primers, and both methods showed strains of both clusters among clinical isolates classified as 0. anthropi by conventional tests. These results and those of previous DNA-DNA hybridization studies [Holmes, B. , Popoff, M., Kiredjian, M. & Kersters, K. (1988). Int J Syst Bacterial 38,4064161 show that the LMG 3301 cluster and related clinical isolates should be given a new species status for which the name Ochrobactrum intermedium sp. nov. is proposed (type strain is LMG 3301T = NCTC 12171' = CNS 2-75').

A PCR assay with primers derived from the 16S rRNA sequence of Brucella abortus was developed. Nine different combinations between six primers were tested. One pair of primers, which amplified a 905-bp fragment, was selected. As little as 80 fg of Brucella DNA was detected by this method. DNAs from all of the representative strains of the species and biovars of Brucella and from 23 different Brucella isolates were analyzed and yielded exclusively the 905-bp fragment. No amplification was detected with DNAs from 10 strains phylogenetically related to Brucella spp., 5 gram-negative bacteria showing serological cross-reactions with Brucella spp., and 36 different clinical isolates of non-Brucella species. Only Ochrobactrum anthropi biotype D yielded a PCR product of 905 bp, suggesting a closer relationship between Brucella spp. and O. anthropi biotype D. The specificity and high sensitivity of the PCR assay may provide a valuable tool for the diagnosis of brucellosis.

A study was performed to evaluate the previously described PCR (C. Romero, C. Gamazo, M. Pardo, and I. López-Goñi, J. Clin. Microbiol. 33:615-617, 1995) for the diagnosis of brucellosis in dairy cattle. Milk samples from 56 Brucella milk culture-positive cattle and from 37 cattle from Brucella-free herds were examined for Brucella DNA by PCR and for specific antibodies by an indirect enzyme-linked immunosorbent assay (ELISA). The specificities of both tests were 100% when testing the milk samples from Brucella-free cattle. The milk samples from 49 infected cattle were positive by PCR (87.5% sensitivity), and 55 were positive by ELISA (98.2% sensitivity). A PCR-positive sample was negative by ELISA, and 7 ELISA-positive samples were PCR negative, yielding an observed proportion of agreement of 0.91 for the two tests. Although the results suggest that ELISA is a better screening test than PCR, the combined sensitivity of the two assays was 100%, and their simultaneous application could be more useful than one test alone for a rapid screening of brucellosis in dairy cattle.

Different methods of extraction of bacterial DNA from bovine milk to improve the direct detection of Brucella by PCR were evaluated. We found that the use of a lysis buffer with high concentrations of Tris, EDTA, and NaCl, high concentrations of sodium dodecyl sulfate and proteinase K, and high temperatures of incubation was necessary for the efficient extraction of Brucella DNA. The limit of detection by PCR was 5 to 50 Brucella CFU/ml of milk.