LightCycler-Based Differentiation of Mycobacterium abscessus and Mycobacterium chelonae

Abstract

In this study we introduce a rapid procedure to identify Mycobacterium abscessus (types I and II) and M. chelonae using LightCycler-based analysis of the hsp65 gene. Results from 36 clinical strains were compared with hsp65 gene restriction analysis and biochemical profiles of bacilli. As all three methods yielded identical results for each isolate, this procedure offers an excellent alternative to previously established nucleic acid amplification-based techniques for the diagnosis of mycobacterial diseases.

Among frequently isolated clinical strains of rapidly growing mycobacteria are Mycobacterium abscessus and Mycobacterium chelonae (2). Patients typically present with local abscesses, lymphadenitis, and orbital or pulmonary infections (10, 11). M. abscessus is also recovered from respiratory specimens of patients with cystic fibrosis (1, 4). The differences in antimicrobial susceptibilities between M. abscessus and M. chelonae demand an easy differentiation of these two closely related species (14). Restriction analysis of an hsp65 gene fragment has been used to this end for over 10 years (7, 12, 13). However, this method requires subsequent gel analysis of amplified DNA. Likewise, interpretation of fragment sizes is sometimes difficult because of the similarity of restriction patterns. Alternative diagnostic procedures include glycan analysis of the bacterial capsule using high-performance liquid chromatography, which is highly reliable but can be performed only in specialized laboratories (13). In this study, we developed a simple and rapid method based on LightCycler technology to distinguish M. abscessus and M. chelonae. We compared the biochemical profiles, the restriction patterns of the hsp65 gene, and partial sequences of the hsp65 gene of 36 clinical strains, including M. abscessus strain ATCC 19977 and M. chelonae strain ATCC 35752 as reference strains. All strains had previously been identified as belonging to the M. chelonae-M. abscessus group by 16S rRNA gene analysis as described previously (5).

Biochemical profiling included testing for growth on 5% sodium chloride and growth on citrate as the sole carbon source as suggested by Yakrus and colleagues (13). Briefly, bacteria were precultured in liquid medium using mycobacterial growth indicator tubes (MGIT) (Becton Dickinson, Oxford, United Kingdom). Positive MGIT cultures were subcultured with 0.1 ml of a 1:10 dilution on Lowenstein-Jensen selective medium (L-J) containing 5% NaCl (Remel, Lenexa, Kans.). Utilization of sodium citrate was determined as described by Silcox et al. (8). Briefly, a basal medium was prepared by dissolving 2.4 g of (NH₄)₂SO₄ (Merck, Darmstadt, Germany), 0.5 g of KH₂PO₄ (Merck), and 0.5 g of MgSO₄ · 7H₂O (Merck) in 950 ml of distilled water; the pH was adjusted to 7.0, and after adding 20 g of Noble agar (Difco Laboratories, Detroit, Mich.), the mixture was autoclaved. Sodium citrate (5.6 g) (J. T. Baker, Deventer, The Netherlands) was dissolved in 50 ml of distilled water, filter sterilized, and added aseptically to the cooled basal medium. Plates were inoculated with 0.1 ml of a 1:10 dilution of a MGIT culture and incubated for 3 weeks.

Cultures on L-J with high (5%) NaCl concentration and on plates with citrate revealed that 27 strains, including the M. abscessus reference strain, grew on L-J with 5% NaCl but failed to grow on sodium citrate medium (Table 1). In contrast, 11 strains, including the M. chelonae reference strain, grew on sodium citrate medium but not on L-J with 5% NaCl (Table 1). We controlled for growth by inoculating bacteria simultaneously on standard L-J slants and plating them on 7H10 complete medium (Difco Laboratories). Thus, biochemical profiling allowed unambiguous identification of M. abscessus and M. chelonae; yet it was time-consuming, since at least 3 weeks of growth at 37°C was required.

TABLE 1.

Comparison of bacterial strains by biochemical profile, hsp65 gene fragment restriction, and LightCycler 16S rRNA and hsp65 gene fragment analyses

Bacterial strain (n)	Biochemical profilea		Fragment length(s) (bp) by PRAb with:		LightCycler melting point temp (°C) + SD for:
	5% NaCl tolerance	Citrate utilization	BstEII	HaeIII	16S rRNA	hsp65
M. abscessus type I (14)	+	−	245, 220	160, 60	66.1 + 0.1	61.9 + 0.1
M. abscessus ATCC 19977 (1)	+	−	245, 220	160, 60	66.1	61.9
M. abscessus type II (12)	+	−	245, 220	210, 60	65.9 + 0.1	66.3 + 0.1
M. chelonae (10)	−	+	325, 140	210	66.1 + 0.1	58.4 + 0.1
M. chelonae ATCC 35752 (1)	−	+	325, 140	210	65.9	58.2

^a+, growth; −, no growth.

^bPRA, PCR-restriction fragment length polymorphism analysis.

Next, the restriction pattern of the hsp65 gene was analyzed as described before by using a modified protocol for preparation of genomic DNA (1). Briefly, 1 ml of MGIT liquid sample was subjected to centrifugation at 10,000 × g for 15 min at room temperature. Supernatants were discarded, and the cellular pellets were resuspended in 150 μl of Tris-EDTA. Cells were disrupted by using acid-washed glass beads (106-μm diameter and finer; Sigma-Aldrich Chemie, Munich, Germany) and removed by centrifugation at 10,000 × g for 4 min. The supernatant containing the DNA was used in a 1:100 dilution. Amplification of a 439-bp fragment of the hsp65 gene using primers Tb 11 (ACC AAC GAT GGT GTG GCC AT) and Tb 12 (CTT GTC GAA CCG CAT ACC CT) was followed by restriction fragment length polymorphism analysis with BstEII and HaeIII.

The initial study by Telenti and colleagues showed that for M. abscessus restriction with BstEII yields a 245-bp fragment and a 200-bp fragment and restriction with HaeIII results in a 160-bp fragment and a 60-bp fragment. For M. chelonae, restriction with BstEII yields a 325-bp fragment and a 140-bp fragment and restriction with HaeIII results in a 210-bp fragment (12). In this study, all M. chelonae isolates but only 15 of 27 M. abscessus isolates showed these patterns (Table 1). For the remaining 12 strains, restriction with BstEII yielded the expected 245-bp and 200-bp fragments but restriction with HaeIII resulted in a 210-bp fragment and a 60-bp fragment.

Devallois and colleagues confirmed this alternative HaeIII restriction pattern for M. abscessus and suggested that M. abscessus type I and M. abscessus type II should be differentiated on that basis (3). Subsequently, Ringuet and colleagues attributed this variation to a T→C single-nucleotide polymorphism at position 542 in the hsp65 gene, affecting an HaeIII restriction site (7). However, M. abscessus type II is difficult to distinguish from M. chelonae because it features only the 60-bp fragment in addition to the common 210-bp fragment.

Finally, we sought to develop an easy and rapid procedure based on LightCycler technology to resolve these problems. First, the entire 439-bp fragment of the hsp65 gene was sequenced by using the forward amplification primer (12). Based on the alignment results of all 38 strains, FRET probes LC 98 (GAG CCT GGG CAA GCA CGG TGG-fluorescein) and LC 99 (LightCycler Red 640-GGT GGT GGT GCC GTC ACC) were chosen for subsequent LightCycler analyses (Fig. 1). Probes were designed to differentiate between M. abscessus type I, M. abscessus type II, and M. chelonae. Yet M. abscessus type I, M. abscessus type II, and M. chelonae could not be unambiguously distinguished from other mycobacteria because of limited variation within this region. Therefore, we included an M. chelonae-M. abscessus group-specific FRET probe, targeting the variable region A of the 16S rRNA gene. A 1,000-bp fragment of the 16S rRNA gene was amplified by using primers LC1 (GAG TTT GAT CCT GGC TCA GGA) and LC4 (TGC ACA CAG GCC ACA AGG GA) as described before (6). The melting profile of the 1,000-bp fragment was analyzed by using the FRET probes LC 68 (GGC CGC GGG CCC ATC CCA CAC-fluorescein) and LC 86 (LightCycler Red 705-CAA AAG CTT TGC ACC ACT CAC).

FIG. 1.

Alignment of 51-bp fragments of the hsp65 genes from M. chelonae, M. abscessus type I, and M. abscessus type II. Comparison of 38 strains revealed three different nucleotide sequences, corresponding to M. chelonae, M. abscessus type I, and M. abscessus type II. Nucleotide polymorphisms are in bold and underlined. Specific melting point temperatures (Tm) of the FRET probe from the various sequences are given on the right. The chosen FRET probes are shown. The HaeIII restriction site that is affected by a nucleotide polymorphism is also shown.

Genomic DNA was prepared as described above, and 5 μl of a 1:100 dilution of a DNA preparation was used for each LightCycler reaction. The hot-start reaction mixture (LightCycler FastStart DNA Master Hybridization Probes; Roche Molecular Biochemicals, Mannheim, Germany), containing FastStart Taq polymerase, reaction buffer, deoxynucleoside triphosphates, and 1 mM MgCl₂, was supplemented with 2 mM MgCl₂. By use of primers LC 1, LC 4, TB 11, and TB12 (multiplex PCR) at a 1.1 μM final concentration and DNA probes at 100 nM final concentrations, the amplification program began with a denaturation step of 10 min at 95°C, followed by 35 cycles of PCR, with 1 cycle consisting of denaturation (3 s at 95°C), annealing (2 s at 56°C), and extension (40 s at 72°C). The temperature transition rate for all cycling steps was 20°C/s. The amplification program was followed by a melting program of 95°C for 30 s (denaturation), 38°C for 30 s (annealing), and then 38 to 80°C at a transition rate of 0.2°C/s with continuous monitoring of fluorescence.

The 439-bp fragment of the hsp65 gene and the 1,000-bp fragment of the 16S rRNA gene were amplified simultaneously (multiplex PCR). Figure 2 shows subsequent melting point analyses in the F2 (640 nm) and F3 (705 nm) channels. Targeting region A of the 16S rRNA gene generates identical melting points at 66°C for M. chelonae and M. abscessus (Fig. 2B). Specificity of the M. chelonae-M. abscessus probe was confirmed by melting point analysis of 33 mycobacteria, including M. chelonae and M. abscessus (Table 2). Except for M. chelonae and M. abscessus, 23 of 31 species showed melting points at temperatures between 43.4 and 55.4°C (mean, 51.0°C; standard deviation, 3.0°C), which are well below the specific melting point of 66°C for the M. chelonae-M. abscessus group. For eight species, melting points were not detectable above 38°C.

FIG. 2.

(A) Melting point analysis of the hsp65 gene fragment in the F2 channel (640 nm). Data for M. abscessus type I, M. abscessus type II, M. chelonae, and the control (H₂O) are shown. Amplification and detection of DNA were performed with primer pairs Tb11 and Tb12 and FRET probes LC98 and LC99. The melting point for M. chelonae was 58°C, for M. abscessus type I it was 62°C, and for M. abscessus type II it was 67°C. (B) Melting point analysis of the 16S rRNA gene fragment in the F3 channel (705 nm). Data for M. abscessus type I, M. abscessus type II, M. chelonae, and the control (H₂O) are shown. Amplification and detection of DNA were performed with primer pairs LC1 and LC4 and FRET probes LC68 and LC86. Melting points for M. chelonae, M. abscessus type I, and M. abscessus type II were all at 66°C.

TABLE 2.

Melting point analysis of the 16S rRNA gene fragment in the F3 channel (705 nm)

Mycobacterial species	Melting point temp (°C)
M. abscessus	66.1
M. chelonae	66.3
M. tuberculosis	—a
M. asiaticum	47.3
M. avium	48.1
M. celatum	49.2
M. conspicuum	50
M. flavescens	—
M. gadium	52.9
M. gastri	50.4
M. genavense	52.3
M. gordonae	—
M. haemophilum	52.2
M. intracellulare	47.7
M. kansasii	50.6
M. lentiflavum	48.3
M. malmoense	—
M. marinum	49.6
M. mucogenicum	54.6
M. neoaurum	53.6
M. nonchromogenicum	55.1
M. peregrinum	55.1
M. phlei	43.4
M. scrofulaceum	50.3
M. shimoidei	55.4
M. simiae	50.2
M. smegmatis	—
M. szulgai	—
M. terrae	52.7
M. thermoresistibile	—
M. triviale	55.3
M. vaccae	—
M. xenopi	49.4

^a—, not detectable.

In contrast, targeting the hsp65 gene results in three different melting points, i.e., at 58, 62, and 67°C, allowing to clearly identify M. abscessus type I, M. abscessus type II, and M. chelonae, respectively (Fig. 2B). The range for positive identification was ±0.3°C for identification of all species shown in Table 1.

In this study we developed a LightCycler multiplex PCR based on analysis of the 16S rRNA gene and the hsp65 gene for rapid differentiation between M. abscessus type I, M. abscessus type II, and M. chelonae. The advantage of this LightCycler protocol is the simultaneous detection of two different genomic regions, since neither LightCycler analysis of the hsp65 gene nor that of the 16S rRNA gene alone leads to unambiguous identification of these organisms. In addition, the ability to discriminate between M. abscessus types I and II may be useful for a preliminary epidemiological analysis in clinical infection. This LightCycler protocol provides an alternative method for differentiation between the closely related species M. abscessus and M. chelonae and complements previously published LightCycler-based procedures for the identification of mycobacteria (6, 9).