Misdiagnosis of Mycobacterium brumae infection.

We read with interest the paper “Catheter-Related Bloodstream Infection Caused by Mycobacterium brumae” by Lee and colleagues (4) because we had described this species in 1993 and no other isolates had been reported from this first description (5). So we contacted the corresponding author, Dr. Han, who gently sent us the M. brumae strain causing the bloodstream infection (MDA0695 strain). In order to confirm the identification of the MDA0695 strain, we performed a total of 24 biochemical tests and growing characteristics by following standard procedures (8). In vitro antibiotic susceptibility was performed in Middlebrook 7H10 agar by the proportion method (8). The M. brumae type strain and the MDA0695 strain were in disagreement in the following eight phenotypic characteristics: colony pigmentation (Fig. 1), growth at 42°C, growth in the presence of 5% NaCl, nitrate reductase (2 h), arylsulfatase (3 days), iron uptake, use of inositol as a sole carbon source, and sensitivity to streptomycin (4 g/ml) (Table 1). For mycolic acid analyses, the two strains were grown on Middlebrook 7H10 agar at 30°C for 2 weeks. Mycolic acids were liberated by saponification, methylated with diazomethane, and analyzed by thin-layer chromatography as previously described (2). All the M. brumae strains contain only -mycolates (5); nevertheless, the MDA0695 strain showed small amounts of -mycolates and large amounts of dicarboxymycolates with their corresponding long-chain alcohols. The pattern of mycolic acids is widely used in taxonomy and identification of mycobacterial species. With few exceptions (1, 6), all the strains that belong to the same species share the same pattern of mycolic acids. So a different pattern of mycolic acids must arouse suspicions on whether the strains compared really belong to the same species. The sequencing of the 16S rRNA gene was performed with PCR amplicons (7) and analyzed using EditSeq and MegAlign software (DNAStar). By comparing the species-specific hypervariable regions A (39 bp) and B (45 bp) of the 16S rRNA gene (3), we found mismatches in 5 and 3 nucleotides, respectively. The MDA0695 isolate showed only 85% and 93% similarity for both regions with the M. brumae type strain (GenBank/EMBL/DDBJ accession number AF547907). Thus, the nucleotide sequence of the MDA0695 strain shows substantial differences from the corresponding sequence of the type strain. The sequences of hypervariable regions are very much conserved within species. The homology should be 100%, or very close to it, in strains of the same species. However, we found significant differences in both of the hypervariable regions. The results of the genotype study show that the MDA0695 strain is not an M. brumae strain. Hypervariable A or B 16S rRNA region’s sequence alignments of the MDA0695 strain revealed it to be closer to noncultivable mycobacteria species and other known rapid-grower mycobacteria than to M. brumae (see Fig. S1 in the supplemental material). An unrooted neighbor-joining tree based on the concatenated hypervariable A and B sequences of MDA0695 and concatenated sequences of other related Mycobacterium species indicated a close relationship between MDA0695 and Mycobacterium flavescens, Mycobacterium monacense, Mycobacterium novocastrense, Mycobacterium smegmatis, and Mycobacterium goodii (see Fig. S2 in the supplemental material). Therefore, the genotypic differences found between the M. brumae type strain and the MDA0695 strain, together with differences found in mycolic acid patterns and the other phenotypic characteristics, allow us to conclude that the MDA0695 strain cannot be considered a strain of the M. brumae species. As a result of the misidentification of the MDA0695 strain, the species M. brumae is now considered a bacterial pathogen and has been classified as biosafety level 2 by the ATCC. Except for the MDA0695 strain, no other M. brumae strains