Large-Restriction-Fragment Polymorphism Analysis of Mycobacterium chelonae and Mycobacterium terrae Isolates

Abstract

Mycobacterium chelonae and Mycobacterium terrae were reported to be frequently present in the environment of the Mycobacterium bovis BCG trial area in south India. Six isolates of M. chelonae and four isolates of M. terrae obtained from different sources in this area were analyzed by pulsed-field gel electrophoresis (PFGE) to examine large-restriction-fragment (LRF) polymorphism using the chromosomal DNA digested with DraI and XbaI restriction enzymes. With the exception of one isolate of M. terrae, DNA from all other isolates could be digested with DraI and XbaI and resulted in separable fragments. Visual comparison of the LRFs showed a unique pattern for each of the isolates tested. A computer-assisted dendrogram of the percent similarity demonstrated a high degree of genetic diversity in this group of isolates. This study demonstrates that species of nontuberculous mycobacteria, particularly M. chelonae and M. terrae, can be successfully typed by their LRF pattern using PFGE, which does not require species-specific DNA probes.

Assessment of genetic diversity is important in epidemiological studies of nontuberculous mycobacteria (NTM), as data from these studies could be used to monitor trends in the occurrence of new strains, identify possible sources of infection, and differentiate individual strains (17). In addition, polymorphism studies may have value in providing comparative information for the basis of human colonization, infectivity, and virulence (5). The phenotypic techniques, such as antibiotic or heavy metal susceptibility testing, serotyping, and multilocus enzyme electrophoresis, used to characterize mycobacteria are labor-intensive and have inherent limitations. The molecular techniques, such as restriction fragment length polymorphism, random amplified polymorphic DNA, and pulsed-field gel electrophoresis (PFGE), are simple to perform; of these techniques, PFGE has been used with most bacterial species (16). It has now gained acceptance as a reliable epidemiological tool for the analysis of strain relatedness of both slow- and fast-growing mycobacteria (1, 2, 6, 20).

An earlier study on the characterization of Mycobacterium avium complex (MAC) isolates, the most commonly found NTM in the Mycobacterium bovis BCG trial area in south India, using techniques such as the AccuProbe test (Gen Probe Inc., San Diego, Calif.), PCR with DT1 and DT6 probes (in-house; Institut Pasteur, Paris, France), PCR restriction analysis of the hsp65 gene, and 16S rRNA gene sequencing, showed that there was a high degree of heterogeneity in the isolates studied (3). Mycobacterium chelonae was reported to be the second most commonly found NTM in the environment of the south Indian BCG trial area (9) and the most frequent NTM associated with nosocomial disease, such as skin and soft tissue infections after outbreaks due to contaminated prosthetic valves, bronchoscopes, peritoneal dialysis equipments, injection vials, etc. (19). Mycobacterium terrae was shown to be the second most common organism, accounting for 12.5% of the NTM isolated from sputum samples from symptomatic subjects residing in the south Indian BCG trial area (13). Contamination of clinical samples with M. terrae was also reported (12). In recent years, increasing numbers of infections caused by these organisms have been diagnosed, particularly as opportunistic infections in AIDS patients (19). Characterization of these organisms is clinically important, and PFGE will be a useful technique, since it could distinguish the organism at the level of individual isolates (17). In the present study, the genetic relationships among several M. chelonae and M. terrae isolates obtained from different geographical sources was examined for large-restriction-fragment (LRF) polymorphism by PFGE.

Six isolates of M. chelonae (two each from soil, water, and sputum samples) and four isolates of M. terrae complex (one from water and three from sputum samples) obtained from the south Indian BCG trial area and maintained on Lowenstein-Jensen slopes were used for this study. All isolates were identified to the species and subspecies levels by routine biochemical methods (9). Species identification of M. terrae complex isolates was confirmed by mycolic acid analysis by high-performance liquid chromatography (18). M. chelonae strain TMC 1542 was used to optimize the extraction of genomic DNA and PFGE conditions, and this strain was tested in every batch of electrophoresis for quality assurance.

The method of Leblond et al. (10) for in situ DNA extraction from Streptomyces lividans was adapted with minor modifications for mycobacterial DNA extraction. Briefly, the cultures were grown in Middlebrook 7H9 (Difco, Detroit, Mich.) medium supplemented with 10% albumin-dextrose complex, 0.2% glycerol, and 0.1% Tween 80. M. chelonae cultures were incubated for 1 week, and M. terrae cultures were incubated for up to 3 weeks. After the purity of the cultures was checked by Ziehl-Neelsen staining, the cells were harvested and washed twice in TE buffer (10 mM Tris and 1 mM EDTA [pH 8.0]), and the concentration of the cells was adjusted to an optical density at 600 nm of 2.0 using sucrose TE buffer (0.3 M sucrose, 25 mM Tris-HCl, 25 mM EDTA). The cells were mixed with an equal volume of 1.5% low-melting-point agarose (Sigma), and 100-μl portions of the cell suspension were poured into agarose molds. The solidified plugs were collected in a 50-ml Falcon tube and incubated at 37°C overnight in 10 ml of sucrose TE buffer containing 2 mg of lysozyme per ml. The solution was replaced with 5 ml of 1% sodium lauryl sarcosine (Sigma, St. Louis, Mo.) in 0.5 M EDTA (pH 8.0) containing 1 mg of proteinase K per ml and incubated at 60°C for 48 h with a change of solution after 24 h. The plugs were washed three times in TE buffer, rinsed twice with TE buffer containing 40 μg of phenylmethylsulfonyl fluoride per ml at 4°C for 4 h to remove proteinase K, and washed thoroughly in TE buffer before storing at 4°C in 20 ml of TE buffer. The restriction enzymes DraI and XbaI (Pharmacia Biotech) were selected on the basis of published reports (11, 15). The agarose plugs (10 mm wide) were cut into 1-mm-thick pieces with a sterile scalpel and immersed in restriction buffer containing 100 U of DraI or XbaI and incubated overnight at 37°C. The agarose plugs were mounted on the teeth of the electrophoretic comb, and 1% pulsed-field-grade agarose (Amresco) in 0.5× TBE was poured around the comb and allowed to solidify.

PFGE was performed with contour-clamped homogeneous electric field mapper system XA (Bio-Rad, Richmond, Calif.). DraI-digested DNA samples were electrophoresed for 24 h at 14°C at 6 V/cm, with a linear switch time of 0.47 s to 1 min 13.58 s. To separate XbaI-digested samples, the program was as described above except that the switch time was linearly ramped from 0.47 to 21.79 s for 20.18 h. Saccharomyces cerevisiae whole chromosomes (in-house preparation) and bacteriophage λ DNA concatemer (Pharmacia Biotech) were used as DNA standards. After electrophoresis, the gel was stained with 1 μg of ethidium bromide per ml and photographed using the gel documentation system (Ultra-Violet Products Ltd.), and the image of each gel was stored electronically for analysis.

PFGE pattern analysis was done by visual comparison of the number and similarity of bands. As visual analysis of PFGE profiles was not sufficient to compare highly banded patterns obtained by XbaI digestion, computer-assisted analysis was performed. The LRFs of isolates were scored manually for the presence (scored as 1) or absence (scored as 0) of all the bands, and the data were entered into Phylip35 software to generate a dendrogram. The identity of a visually scored fragment was ascertained by comparing the restriction pattern of a standard strain in each gel and calculating the molecular weight based on the standards which were also run in parallel. The results were interpreted according to the criteria of Tenover et al. (16).

Figure 1 shows the LRF patterns of chromosomal DNA from representative isolates of M. chelonae and M. terrae. DraI digestion of chromosomal DNA generated well-separated DNA fragments ranging from 64 to 1,000 kb in size. Bands between 97.5 and 800 kb in size were used for visual comparison. Lanes 1 through 6 contain DraI-digested M. chelonae DNA, and lanes 7 and 8 contain DraI-digested M. terrae DNA. The LRF pattern of the two isolates of M. chelonae from soil differed by six fragments, the major ones being a 450-kb fragment in the first isolate and a 750-kb fragment in the second isolate. The two isolates of M. chelonae from water differed by seven fragments, and the prominent fragments were the 365- and 680-kb fragments in the second isolate. The two isolates of M. chelonae from sputum differed by more than six fragments. Genomic DNA from the three M. terrae isolates differed by more than six fragments. One of the M. terrae isolates did not give any pattern. Overall, DraI digestion produced a unique LRF pattern for each of the M. chelonae and M. terrae isolates studied. DraI digestion represents the entire genome of the organism. Hence, the approximate genomic DNA size of each of the isolates could be calculated. The sizes are given in Table 1. The genome size of the majority of the isolates was approximately 4.0 Mb.

FIG. 1.

PFGE analysis of mycobacterial genomic DNA digested with DraI. Pulse time increased linearly from 0.47 to 1 min 13.58 s. Lane M1, yeast chromosomal DNA markers; lanes 1 and 2, M. chelonae isolates from soil; lanes 3 and 4, M. chelonae isolates from water; lane 5, M. chelonae isolate from sputum; lane 6, M. chelonae TMC 1542; lane 7, M. terrae isolate from water; lane 8, M. terrae isolate from sputum; lane M2, λ DNA PFGE markers. Molecular sizes (in kilobases) of the DNA standards are given at the sides of the gel.

TABLE 1.

Summary of PFGE results for M. chelonae and M. terrae isolates^a

Sample	Strain	Isolate source	Identificationb	No. of large restriction fragments		Genome sizec
				DraI	XbaI
1	A86/3	Soil	M. chelonae	10	22	4,050
2	D73/123	Soil	M. chelonae	11	20	4,410
3	B85/3ab	Water	M. chelonae	11	22	4,785
4	B86/12	Water	M. chelonae	11	20	4,160
5	TS09896	Sputum	M. chelonae	11	16	3,869
6	TS10108	Sputum	M. chelonae	10	21	4,090
7	TMC 1542		M. chelonae	10	15	4,166
8	B92/1	Water	M. terrae complex	8	15	3,510
9	TS10088	Sputum	M. terrae complex	11	21	4,295
10	TS11431	Sputum	M. terrae complex		NPd
11	TS16563	Sputum	M. terrae complex	10	19	4,226

^aM. chelonae strain TMC 1542 was used as a reference strain.

^bSpecies identification by biochemical methods and high-performance liquid chromatography.

^cApproximate genome size (in kilobases) based on DraI digestion results.

^dNP, no pattern.

The ability to detect LRF polymorphism varies with the restriction enzyme used (21). DraI is shown to be a suitable enzyme for analysis of mycobacterial genomes (15). However, it has been reported that minor changes in genomic DNA of mycobacteria could be detected only with XbaI (21). Hence, this enzyme was used to confirm the results obtained with DraI.

Figure 2 shows M. chelonae isolates digested with XbaI, which generated approximately 15 to 24 smaller fragments ranging from 25 to 450 kb. DNA fragments of 48.5 to 388 kb were used for comparison. Since there were several doublets in almost all the lanes, a second electrophoresis with an extended run time (26 h) was performed to separate the fragments. The two isolates from soil differed in the presence of a 339-kb DNA fragment in the second isolate. The two isolates from water differed in the presence of two DNA fragments of approximately 291 kb in the first isolate. The two isolates from sputum samples differed by more than six bands. Figure 3 shows the XbaI LRF profile of M. terrae isolates and the patterns of three M. terrae isolates, which differed by more than eight fragments. The strain that failed to give a separable digest with DraI also did not give any pattern with XbaI (lane 4). XbaI-digested DNA was similar to DraI-digested DNA and generated seven unique LRF patterns for the seven M. chelonae isolates and three unique patterns for the three M. terrae isolates. This study shows that the use of a single restriction enzyme is sufficient to type all isolates, as unique patterns were obtained with DraI or XbaI. It was reported earlier that isolates of Mycobacterium fortuitum could be reliably distinguished by using only one restriction enzyme (7). The discriminatory power can be greatly enhanced by performing two-dimensional PFGE (14).

FIG. 2.

PFGE analysis of genomic DNA of M. chelonae digested with XbaI. Total run time was 20.18 h. Pulse time ramped from 0.47 to 21.79 s. M. chelonae isolates from soil (lanes 1 and 2), water (lanes 3 and 4), and sputum (lanes 5 and 6) and strain TMC 1542 (lane 7) were used. Lane M1, yeast chromosomal DNA markers; lane M2, λ DNA PFGE markers. Molecular sizes (in kilobases) of the DNA standards are given at the sides of the gel.

FIG. 3.

PFGE analysis of M. chelonae digested with XbaI. Total run time was 24 h. Pulse time ramped from 0.47 to 21.79 s. M. chelonae isolates from water (lane 1) and sputum (lanes 2 to 4) were used. Lane M1, yeast chromosomal DNA markers; lane M2, λ DNA PFGE markers. Molecular sizes (in kilobases) of the DNA standards are given at the sides of the gel.

Figure 4 shows a schematic representation of different LRF patterns of M. chelonae and M. terrae generated by digestion of genomic DNA with DraI and XbaI and a dendrogram generated by comparing bands from XbaI digestion. The dendrogram revealed that one M. chelonae isolate from water and one M. chelonae isolate from soil had 85% genetic similarity, and they were closely related to the M. chelonae reference strain. Another isolate from water was similar to one of the isolates from sputum. One isolate from soil and one isolate from sputum were genetically distinct. Two M. terrae isolates from sputum were closely related and exhibited 75% similarity with the isolate from soil. Earlier reports on polymorphic studies using PFGE patterns of clinical isolates of Mycobacterium tuberculosis complex, Mycobacterium kansasii, and Mycobacterium simiae using software programs, such as Biosystematica, Taxotron, and Dendron, have shown that there was genetic diversity among the isolates of one species (4, 8, 11).

FIG. 4.

Schematic diagram showing distinct LRF patterns obtained after cleavage of the genomic DNA from M. chelonae and M. terrae complex with restriction enzymes. A dendrogram of the LRF types obtained with XbaI is shown to the left. The lane numbers are the sample numbers shown in Table 1.

In this study, M. chelonae and M. terrae isolates for which species-specific probes are not available commercially were successfully typed by PFGE. The isolates were tested on several occasions, and reproducible patterns were obtained. The unique pattern obtained for each of the isolates demonstrates that there is a high degree of genetic polymorphism between the isolates originating from the same geographical area. It seems likely that horizontal transfer of large segments did not occur. This also demonstrates that genetic variation may influence their antibiotic susceptibility profile and disease pathogenesis. A recent phylogenetic analysis of clinical and laboratory isolates of M. tuberculosis using large-sequence polymorphism and single-nucleotide polymorphism suggests that polymorphism among mycobacteria is more extensive than anticipated and that genetic variation may have an important role in disease pathogenesis and immunity (5). Animal pathogenicity studies of M. chelonae and M. terrae may help in determining their difference in colonization and pathogenicity.