|Home | About | Journals | Submit | Contact Us | Français|
Carbon utilization tests have proven to be useful for the identification of some species of rapidly growing mycobacteria and have been described as one of the few tests useful for the differentiation of Mycobacterium mucogenicum from other rapid growers. We have found the carbon utilization tests to be unreliable for the identification of patient isolates of this species. In this study, using 28 isolates of rapidly growing mycobacteria, we examined several variables which might have an effect on results of citrate, inositol, and mannitol utilization: inoculum concentration, incubation temperature, and medium manufacturer. None of these variables affected results obtained for most species of rapid growers or for ATCC strains of M. mucogenicum. Results for patient isolates of M. mucogenicum were found to be inconsistent regardless of the methodology employed and resulted in an ambiguous identification of these isolates or an incorrect identification as Mycobacterium chelonae. Molecular or cell wall analysis may be the best technique to employ for accurate identification of M. mucogenicum.
Rapidly growing mycobacteria have been implicated in a variety of infectious diseases (12), and correct species assignment may be important for epidemiologic and chemotherapeutic reasons. Accurate identification is now possible using molecular techniques (10) and various chromatographic techniques (2, 6, 7, 13); however, many clinical laboratories may not have the resources available to perform these procedures and may rely on various biochemical tests for identification of these organisms.
Among the rapid growers of clinical significance is Mycobacterium mucogenicum (formerly designated the Mycobacterium chelonae-like organism), a nonpigmented organism which is commonly recovered from tap water (1). M. mucogenicum has been shown to be genetically distinct from other rapidly growing mycobacteria by analysis of its 16S rRNA gene (9). It was initially isolated in 1982 from an outbreak of peritonitis in patients receiving peritoneal dialysis (1) and has subsequently been isolated from a variety of other clinical specimens. It has been shown most often to be clinically significant when isolated from blood and post-traumatic wounds (13) and is unusual among the rapidly growing mycobacteria in its relative susceptibility to antimicrobial agents (11, 13).
Because of the limited number of discriminatory tests available, the biochemical characterization of M. mucogenicum has often been problematic. Its failure to grow on 5% sodium chloride agar and its subtle reactions on iron uptake medium have resulted in false identifications as M. chelonae. The use of various carbohydrates as carbon sources has been proposed as a useful method for the identification of this organism (8, 13); it has been reported to utilize citrate and mannitol, but not inositol, as sole carbon sources.
In our laboratory initial biochemical studies of five M. mucogenicum patient isolates resulted in ambiguous identifications due to negative citrate and/or mannitol reactions on the carbon utilization media. In an effort to improve the biochemical methods used for the identification of this species, we investigated some variables involved in inoculation and incubation of carbon utilization media.
A total of 28 isolates of 6 different rapidly growing mycobacterial species were examined for growth on carbon utilization media to test two methods of inoculum preparation and two incubation temperatures. Patient and environmental isolates included 12 isolates from patients seen at the Clinical Center of the National Institutes of Health (NIH), one isolate from a patient seen at the George Washington University Medical Center, Washington, D.C., one isolate from a patient seen at the Walter Reed Army Medical Center, Washington, D.C., two isolates referred to the Mycobacteriology Laboratory of the Maryland State Health Department, Baltimore, Md., and one environmental isolate. Identifications of these 17 isolates were as follows: Mycobacterium abscessus, 4 isolates; M. chelonae, 4 isolates; Mycobacterium fortuitum, 2 isolates; and Mycobacterium mucogenicum, 7 isolates. The American Type Culture Collection (ATCC) type strains examined were M. abscessus ATCC 19977, M. chelonae ATCC 35752, M. fortuitum ATCC 6841, Mycobacterium peregrinum ATCC 14467, and M. mucogenicum ATCC 49650. The strains M. abscessus ATCC 23006, M. fortuitum third biovariant ATCC 49403 and 49404, M. peregrinum ATCC 35755, and M. mucogenicum ATCC 49649 and 49651 were examined, as well, as representatives of those species.
Patient isolates were characterized by traditional biochemical methods (5, 13) using the following parameters: positive Kinyoun reaction, growth rate of less than 7 days, growth on MacConkey agar without crystal violet (Remel, Lenexa, Kans.), iron uptake reaction (Remel), growth on 5% sodium chloride agar (Remel), and nitrate reduction (Remel).
We confirmed the identifications of patient isolates by PCR amplification of a 383-bp fragment of a portion of the 65-kDa heat shock protein gene with the primers described by Hance et al. (4). Restriction endonuclease analysis was performed with the PCR products by digestion with BstEII and HaeIII (New England Biolabs, Beverly, Mass.) (10). The patterns obtained with patient isolates were compared to those obtained with the ATCC type strains. The identifications of four of the six patient isolates and the environmental isolate of M. mucogenicum were also confirmed by high-performance liquid chromatography (HPLC) of cell wall mycolic acid esters by the Centers for Disease Control and Prevention (CDC), Atlanta, Ga., or the VA Reference Laboratory for Tuberculosis and Other Mycobacteria Diseases, West Haven, Conn. The M. mucogenicum isolates chosen for use in this study were evaluated for epidemiologic relatedness and were found to be unrelated (on the basis of date and source of isolation).
A single colony of representative isolates of each species to be tested was inoculated into 7H9 broth (Remel) to which 3-mm glass beads had been added. Suspensions were shaken vigorously and incubated at 28 ± 1°C in ambient air for 7 days. The resulting growth was inoculated onto carbon utilization media using two suspensions with different organism concentrations. For suspension A, an aliquot of the 7-day 7H9 culture was diluted 1:10 in 7H9 broth (5, 8). For suspension B, an aliquot of the 7-day culture was spectrophotometrically adjusted to approximate a 1 McFarland turbidity standard (35.8% ± 2% transmission at 540 nm) (3); this adjusted suspension was subjected to 10-fold serial dilutions in normal saline until no turbidity was visually detected (usually at the 1:100 dilution).
The following experiments were performed. (i) For all the isolates of M. abscessus, M. fortuitum, M. fortuitum third biovariant, and M. peregrinum, 0.1 ml of the final suspension from each of the two suspensions (A and B) was inoculated onto each tube of two sets of the carbon utilization media (Remel). Each set included a base control, citrate, inositol, and mannitol. One set inoculated with each suspension was incubated at 35 ± 1°C, the other set at 28 ± 1°C, both in ambient air. (ii) Four isolates of M. chelonae (the ATCC type strain and three patient-derived isolates) and five isolates of M. mucogenicum (two ATCC strains [ATCC 49649 and ATCC 49651] and three patient isolates) were inoculated as noted above, utilizing suspensions A and B, onto two sets of carbon utilization media from Remel. Additionally, two sets of media from a second manufacturer (NIH Media Unit, Bethesda, Md.) were inoculated as noted above, and one set was incubated at each temperature. (iii) The M. mucogenicum type strain (ATCC 49650) and a fourth M. mucogenicum patient isolate were tested with only the NIH media using the two inoculum concentrations and the two incubation temperatures. (iv) After the results from the three experiments described above were evaluated, all 10 M. mucogenicum isolates (6 patient isolates, 1 environmental isolate, and 3 ATCC isolates) were tested on media from a third manufacturer (Hardy Diagnostics, Santa Maria, Calif.) using inoculum A and an incubation temperature of 28 ± 1°C.
The inoculum suspension for each isolate was tested for purity on sheep blood agar (Remel), and the concentration of organisms in each inoculum suspension was determined by subculture of appropriate dilutions onto Middlebrook agar plates (Remel) in a manner which would facilitate counting of colonies. Slants were examined for growth after 1 and 2 weeks of incubation. Tests were considered to be valid if the base control for each set of carbon utilization media showed no growth. Individual carbon utilization tests (citrate, inositol, and mannitol) were considered to be positive if the slants showed any amount of growth. Correct results were those which agreed with the expected results for a particular organism (5, 13).
The range and average values of the organism concentration of the various suspensions tested were as follows: suspension A, 1.0 × 107 to 3.1 × 109 CFU/ml (average inoculum per slant, 8.8 × 106 CFU), and suspension B, 2.0 × 105 to 4.5 × 107 CFU/ml (average inoculum per slant, 7.3 × 105 CFU).
With both suspensions and at both incubation temperatures, all ATCC and patient isolates of M. abscessus, M. fortuitum, M. fortuitum third biovariant, and M. peregrinum and all three ATCC isolates of M. mucogenicum gave consistent, correct results for carbon utilization. (Expected results were as follows: for M. abscessus and M. fortuitum, citrate, inositol, and mannitol negative; for M. fortuitum third biovariant, citrate negative and inositol and mannitol positive; for M. peregrinum, citrate and inositol negative and mannitol positive; for M. mucogenicum, citrate positive, inositol negative, and mannitol positive.) Results with the media prepared by the three different manufacturers showed essentially no differences (data not shown).
For M. chelonae, the highest percentage of correct results (expected results were citrate positive, inositol negative, and mannitol negative) was obtained with tests inoculated with suspension A (a 1:10 dilution of a 7-day broth culture; average inoculum per slant, 8.8 × 106 CFU) and incubated at 28°C (Table (Table1).1). All ATCC and patient isolates of M. chelonae inoculated and incubated under these conditions gave the expected positive citrate and negative mannitol and inositol results.
For M. mucogenicum, the highest percentage of correct results for both citrate and mannitol utilization was also obtained with test media inoculated with suspension A and incubated at 28°C (Table (Table1).1). However, for the patient and environmental isolates, using media from the various manufacturers, only 8 of 13 tests for citrate (62%) and 3 of 13 tests for mannitol (23%) gave the expected positive result (Table (Table2).2). All isolates gave the expected negative results for inositol. Of the three patient isolates tested more than once, only one patient isolate consistently gave the expected positive reaction for citrate utilization, and no patient isolates consistently gave the expected positive result for mannitol utilization. This lack of consistency suggests that small variations in test conditions may significantly affect the results of the carbon utilization test for M. mucogenicum.
It is unclear why our results differ from previously published results for this test. The majority of the initial results described by Silcox et al. (8) were obtained with isolates from a single outbreak of M. mucogenicum from peritoneal specimens. All these isolates may have been similar in their biochemical properties because of their common source. One of these isolates has become an ATCC strain of M. mucogenicum (ATCC 49649) and, as with the two other ATCC strains tested, gave consistently positive results with both citrate and mannitol in our hands. Wallace et al. comment that 20% of the nonpigmented rapidly growing mycobacteria submitted to the Mycobacterial Reference Section of the CDC in 1991 were isolates of M. mucogenicum (13). Many were evidently submitted because of unfamiliarity with this group of organisms or because of an “atypical” biochemical result. Wallace et al. also note that the nitrate test was the most variable feature of the M. mucogenicum isolates studied, with 45% of the 64 isolates tested producing a “+/−” reaction. Of those same isolates, 85% were positive for citrate utilization and 94% were positive for mannitol utilization (13). Our results suggest that even under optimum conditions, the citrate utilization test provides variable results and the mannitol utilization test may often be negative.
More recent assessment of the reliability of the carbon utilization tests with larger numbers of isolates may not have been performed, since reference laboratories currently rely on drug susceptibility testing, HPLC, or molecular methods for definitive identification of these organisms. There is some evidence that reference laboratories may also obtain variable results with these biochemical tests. One M. mucogenicum patient isolate that we referred to the Mycobacteriology Reference Laboratory of the CDC for identification was found to have a negative (+/−) citrate utilization reaction and a negative mannitol utilization reaction. The isolate was identified as an M. chelonae-like organism based on HPLC results of the cell wall mycolic acids.
There are several organism characteristics which may suggest the identification of M. mucogenicum. Typical colonies are smooth and slightly mucoid. The iron uptake test performed on ferric ammonium citrate agar has been described as having a characteristic +/− reaction, with colonies acquiring a tan or light rust color (8). However, this reaction is subtle and may not be easily recognized. Wallace et al. suggest ciprofloxacin and cephalothin disk susceptibility testing as particularly useful for the identification of M. mucogenicum and note that approximately 90% of isolates are susceptible to cephalothin (13).
Accurate identification of M. mucogenicum may be impossible using conventional biochemical testing. Laboratories should suspect the presence of M. mucogenicum if an isolate produces mucoid colonies, is relatively drug susceptible, and produces an ambiguous biochemical profile or one consistent with M. chelonae. For definitive identification, either cell wall analysis by HPLC or molecular characterization may be necessary.
We thank Nancy Hooper of the Mycobacteriology Laboratory of the Maryland State Health Department, John F. Keiser of the George Washington University Medical Center, and the Microbiology Laboratory of the Walter Reed Army Medical Center for providing us with clinical isolates for this investigation and Daniel P. Fedorko of the Microbiology Service, Department of Laboratory Medicine, Clinical Center, NIH, for critically reviewing the manuscript.