ABSTRACT
Polyclonal infection by Mycobacterium avium was detected by hsp65 PCR-restriction enzyme analysis (PRA) in a bone marrow isolate from an AIDS patient. Two M. aviumstrains, differing in colony morphology, PRA HaeIII digestion pattern, insertion element (IS) 1245amplification, and restriction fragment length polymorphism fingerprints with IS1245 and IS1311 probes, were isolated.
Polyclonal Mycobacterium avium infections have been detected in AIDS and non-AIDS patients by serotyping and highly discriminatory genetic typing methods, pulsed-field gel electrophoresis and restriction fragment length polymorphism (RFLP) (1, 3, 12, 17). A Brazilian study revealed the occurrence of mixed M. avium complex infections in 34 of 90 isolates (37.8%) from 75 patients analyzed by serotyping (11). In spite of these and other reports, polyclonalM. avium infections are not commonly detected, in part because discrimination of M. avium strains by serotyping, pulsed-field gel electrophoresis, and RFLP is technically demanding and expensive for the clinical laboratory.
In this report, polyclonal M. avium infection in a bone marrow isolate from an AIDS patient was detected by PCR-restriction enzyme analysis (PRA), an identification method based on amplification of the hsp65 gene by PCR, followed by cleavage of the amplified product with endonucleases BstEII andHaeIII (14). To our knowledge, this is the first time that a polyclonal infection caused by two strains of the same mycobacterial species could be detected by PRA.
The patient, a 30-year-old white male, was admitted to the Instituto de Infectologia Emı́lio Ribas, São Paulo, Brazil, in March 1998. He had a positive human immunodeficiency virus test result and a CD4+ count of 3 cells/mm3. A bone marrow aspirate was obtained and inoculated into a flask of biphasic medium. This medium is composed of a solid phase of Lowenstein-Jensen (LJ) medium and a liquid phase of Middlebrook 7H9 modified medium (Difco, Detroit, Mich.) (8). The growth was analyzed by Ziehl-Neelsen staining and subcultured in LJ medium for identification. A loopful of growth from the LJ medium was diluted in 10 mM Tris (pH 8.0)–1 mM EDTA–1% Triton X-100 and submitted to three cycles of freezing and boiling. Ten microliters of supernatant was used for identification by PRA. A complex six-band pattern on HaeIII digestion of the hsp65 amplicon suggested that the culture on LJ medium contained more than one mycobacterial species. To evaluate this hypothesis, we studied single colonies obtained on 7H10 agar supplemented with oleic acid, albumin, dextrose, and catalase (Middlebrook OADC enrichment; Difco). Two morphotypes were observed: translucent (TL) and opaque (OP) colonies. Both were identified as species of the M. avium complex by biochemical tests and asM. avium by the specific AccuProbe test (GenProbe, San Diego, Calif.) (Table 1).
Results of molecular identification by PCR and AccuProbe test
PRA with DNA from the isolated TL and OP colonies resulted in identicalBstEII digestion patterns but two differentHaeIII patterns. Amplicons from TL colony DNA presented anM. avium prototype (variant I) HaeIII digestion pattern: 140 bp, 105 bp, and two bands smaller than 50 bp (14). Amplicons from OP colonies showed 145-bp, 140-bp, and 60-bp bands, plus two bands smaller than 50 bp, on HaeIII digestion (M. avium unreported variant IV). Amplicons from both variants were sequenced in an automated ABI Prism 377 sequencer (Perkin-Elmer, Foster City, Calif.). Sequences were aligned toM. avium ATCC 25291- and Mycobacterium intracellulare ATCC 13951-corresponding sequences (7), using the DNAsis program (Hitachi Software Engineering, San Francisco, Calif.). The third and sixth HaeIII sites on amplicons from DNA of OP colony-forming bacilli showed substitutions that were responsible for the modified PRA pattern (Fig.1).
Alignment of hsp65 sequences (total length, 380 positions) from OP colony-forming bacilli, TL colony-forming bacilli, M. avium ATCC 25291, and M. intracellulare ATCC 13950. Nucleotide substitutions in relation to the M. avium sequence are underlined. HaeIII sites are boxed, and asterisks indicate sites with point mutations.
We have recently reported the existence of three PRA variants inM. avium isolates obtained from pigs and humans (7). Here we present evidence of a fourth M. avium PRA variant, suggesting that this DNA region in M. avium is significantly more variable than previously reported (4, 13).
Molecular characterization of TL and OP strains included amplification of DT1 and DT6, single-copy sequences identified in the genome ofM. avium serotype 2 (15), insertion element (IS)1245 (a sequence consistently present in M. aviumstrains [2, 5, 9]), and IS1311 (an M. avium-specific insertion sequence showing 85% sequence similarity with IS1245 [10]). Amplification with DT6 primers but not with DT1 primers was observed in DNA from both morphotype colonies. DNA from TL colony-forming bacilli could be amplified with IS1245 primers, but PCR with DNA from OP colony-forming bacteria was repeatedly negative. DNA from both colony types was amplified with IS1311 primers (Table 1). Taken together, these results confirmed the identification of M. avium and detected a second genetic difference between the two morphotypes, related to the amplification with IS1245primers.
IS1245 and IS1311 RFLP analysis was performed with DNA from five TL and five OP colony-forming bacilli, according to a standardized protocol (16). Identical two-band IS1245 RFLP patterns were observed in all OP colonies, in spite of negative amplification with IS1245 primers. Identical multibanded IS1245 RFLP patterns were observed with all TL colonies. The same clustering was observed with the IS1311 probe (Fig. 2).
RFLP with DNA isolated from five OP and five TL colony-forming bacilli. The same membrane was hybridized with peroxidase-labeled IS1245 and IS1311 probes and developed with the ECL labeling and detection kit (Amersham, Pharmacia Biotech, Inc., Piscataway, N.J.). Normalization was performed with GelCompar II software (Applied Maths, Kortrijk, Belgium).
A relevant aspect was the finding of two different morphotypes representing two genetically distinct strains. It was previously reported that colonies with different morphotypes can be derived from a single strain (1, 19), and TL-OP or OP-TL transitions have been observed (18). Two colony morphotypes isolated from this clinical sample corresponded to two different strains. Besides that, no transitions were observed in vitro after five passages on 7H10 agar with incubations at 37°C, 30°C, and 45°C over a 6-month period (data not shown).
There is evidence that TL clones are more resistant to antimicrobial agents and more virulent in animal models (6). Susceptibility patterns of both strains, evaluated by determination of the MICs of 12 antimicrobial agents, showed minimal differences. OP and TL colonies diverged only in susceptibility to clofazimine (Table2).
MICs of antimicrobial agents for M. avium morphotypes
M. avium genetic variability in the hsp65 locus and different results of IS1245 PCR illustrated here and in a previous report (7) indicate that M. aviumstrains that cause disease in humans and in animals can differ in several genetic and phenotypic aspects. Characterization of these differences will certainly lead to better understanding of M. avium pathogenesis and epidemiology.
Nucleotide accession number.The nucleotide sequence ofM. avium PRA variant IV was deposited in GenBank under accession number AF234261 .
ACKNOWLEDGMENTS
David Hadad is acknowledged for obtaining patient's data and Robert Arbeit for helpful discussion. S.C.L. and M.R.S.B. received grants from Fundação de Amparo à Pesquisa do Estado de São Paulo. M.R.S.B. received an International Research Scholar grant from the Howard Hughes Medical Institute.
FOOTNOTES
- Received 20 July 2000.
- Returned for modification 17 August 2000.
- Accepted 20 September 2000.
- Copyright © 2000 American Society for Microbiology
