Mycobacterium algericum sp. nov., a novel rapidly growing species related to the Mycobacterium terrae complex and associated with goat lung lesions

Non-tuberculous mycobacteria are typically found in the environment, but can be the cause of occasional opportunistic infections in humans or animals (Tortoli, 2009).

Members of the Mycobacterium terrae complex (Mycobacterium terrae sensu stricto, Mycobacterium nonchromogenicum, Mycobacterium triviale and Mycobacterium senuense) are ubiquitous and potentially opportunistic pathogens. Indeed, some reports showed potential human pathogenicity (Krisher et al., 1988; Mayo et al., 1998; Smith et al., 2000). Disseminated M. terrae infections have also been reported among AIDS patients (Carbonara et al., 2000). M. senuense was recently described as a novel Mycobacterium species closely related to the M. terrae complex (Mun et al., 2008). It was isolated from a Korean patient with a symptomatic pulmonary infection.

In this report, we present a previously undescribed Mycobacterium species, for which we propose the name Mycobacterium algericum sp. nov. (type strain, TBE 500028/10^T). Previous publications have shown that phenotypic and biochemical analyses often do not provide an accurate identification of Mycobacterium species (Cloud et al., 2006; Lee et al., 2004; Mun et al., 2008; Springer et al., 1996). Therefore, using a combination of both biochemical and molecular techniques, we showed that isolate TBE 500028/10^T, collected from a pulmonary lesion on a goat in Algeria, was representative of a novel Mycobacterium species related to the M. terrae complex, in particular M. terrae sensu stricto and M. senuense.

The isolate was collected in the Bejaia department (Wilaya) in Algeria in 2008, at the slaughterhouse of Souk El Tenine. Lung lesions observed on goat carcasses were sampled and sent to the Pasteur Institute in Alger. There, Ziehl–Neelsen staining and microscopy were performed. After grinding and decontamination by Petroff’s method (4 % NaOH) (Ghosh et al., 1978; Petroff, 1915), samples were inoculated on Löwenstein–Jensen growth medium. Tubes were incubated at 37 °C (ambient atmosphere) and monitored weekly until colonies suggesting mycobacterial growth were observed. Tubes with potential mycobacterial colonies were sent to Switzerland for further biochemical and molecular characterization of the isolate.

Biochemical characterization of strain TBE 500028/10^T was performed by observing growth under various conditions, according to methods described by Murray et al. (2007); the determined characteristics were compared with those of the closely related Mycobacterium species M. terrae and M. senuense. To determine the optimal growth temperature, Middlebrook 7H10 agar plates were inoculated and incubated at temperatures ranging from 25 °C (with and without light) to 45 °C (ambient atmosphere). We also tested growth ability on MacConkey agar with and without 5 % NaCl at 37 °C (ambient atmosphere) and tellurite reduction, as well as nitrate reductase, urease, catalase and pyrazinamidase production. Colony morphology and pigmentation are described. Cell morphological analyses were conducted after growth in 7H9 liquid broth with 10 % OADC supplement (oleic acid, bovine albumin, glucose and catalase). Cells were harvested in the exponential-growth phase and stained with auramine–rhodamine. Microscopic imaging was performed on a Zeiss Axiovert 200M microscope using a 100-fold magnifying objective (Carl Zeiss). Cell dimensions were determined by using the measurement tool provided by the SlideBook version 4.1 software (Intelligent Imaging Innovations). Acid–alcohol fastness was assessed by Ziehl–Neelsen staining.

Susceptibility to the antibiotics rifampicin (1.0 and 10.0 mg l⁻¹), rifabutin (0.1 and 1.0 mg l⁻¹), ethambutol (5.0 and 50 mg l⁻¹), clarithromycin (4.0, 16.0, 32.0 and 64.0 mg l⁻¹), amikacin (1.0 and 10 mg l⁻¹), ofloxacin (2.0 and 10.0 mg l⁻¹) and moxifloxacin (0.5, 2.5 and 10 mg l⁻¹) was tested in mycobacterial growth indicator tubes (MGIT 960; Becton Dickinson) and monitored with EpiCenter software and the TB eXiST module (Springer et al., 2009).

HPLC analyses were performed to investigate the profile of cell-wall mycolic acids (Butler & Guthertz, 2001; CDC, 1996). The profile obtained for our isolate was compared with those from other closely related Mycobacterium species.

DNA was extracted from cultures using the Bio-Rad InstaGene Matrix. In order to assess the phylogenetic relationships of the isolate, we performed sequencing of the nearly complete 16S rRNA gene (1521 bp, corresponding to Mycobacterium tuberculosis H37Rv^T rrs nt 18–1534) using previously described primers and protocols (Rogall et al., 1990; Springer et al., 1996) and two new primers (16s_742_fw, 5′-AGCGTGGGGAGCGAACAGG-3′; 16s_1408_rv, 5′-CCCGAAGCCGGTGGCCTAA-3′). Two regions of the rpoB gene were amplified and sequenced: 710 bp (M. tuberculosis H37Rv^T rpoB nt 2436–3145) according to Adékambi et al. (2003), and 553 bp (M. tuberculosis H37Rv^T rpoB nt 1092–1645) according to Kim et al. (1999), with a modified reverse primer: MR2, 5′-CAGCTCGTCGTCGTCCTC-3′. A 441 bp part of the hsp65 gene (M. tuberculosis H37Rv^T groEL2 nt 145–585) was sequenced according to Telenti et al. (1993). All amplifications included the reference strain M. tuberculosis H37Rv^T as positive control, and buffers without any added mycobacterial DNA as negative control.

Nucleotide sequences were obtained for both forward and reverse primers. Sequence contingency alignment was performed for each gene with the software SeqMan version 7.0 (dnastar Inc.). The obtained consensus sequences were compared with mycobacterial sequences by similarity search in GenBank, using the blastn algorithm. Phylogenetic trees comparing our isolate with other Mycobacterium species were obtained for each gene by using the megalign software (dnastar Inc.) with default settings for clustal v neighbour-joining multiple alignments. Note that the rpoB phylogeny was calculated based on the 553 bp sequence fragment. All obtained trees were established by bootstrap analyses with 1000 resamplings and 111 seeds.

The goat from which the isolation was made presented nodular lung lesions of caseous aspect. Direct microscopy of Ziehl–Neelsen-stained tissues showed low-density (six acid–alcohol-resistant bacilli in 300 observed fields) infections with slightly curved bacilli. The isolate was inoculated into Löwenstein–Jensen tubes in Algeria. Growth rate was slow on Löwenstein–Jensen medium: small, round, white–yellow colonies were observed after 35 days. Inversely, subculture on Middlebrook 7H10 agar plates was rapid; microcolonies could be observed after 1 week at temperatures ranging between 25 and 40 °C, with or without light. Optimal growth was seen at 37 and 40 °C, whereas no growth was observed at 45 °C. On Middlebrook 7H10 agar plates, colonies were small, polymorphic, white, non-chromogenic, smooth and with irregular edges. Ziehl–Neelsen and auramine–rhodamine staining showed acid-fast, 1.4±0.2 µm long, rod-shaped bacilli. No spores, cords or filaments were observed.

No growth was observed on MacConkey agar with or without 5 % NaCl. Production of urease, a positive nitrate reductase reaction, a positive catalase reaction at 25 °C, and tellurite reductase activity were observed.

Drug-susceptibility testing in MGIT 960 showed susceptibility to all drugs tested (see Supplementary Table S1, available in IJSEM Online), with the exception of low-level resistance to aminoglycosides (1.0 mg amikacin l⁻¹), quinolones (0.5 mg moxifloxacin l⁻¹ and 2.0 mg ofloxacin l⁻¹) and rifampicin (1.0 mg l⁻¹). Pyrazinamidase production was detected. Strain TBE 500028/10^T is thus sensitive to pyrazinamide.

HPLC analysis of the cell wall’s mycolic acids demonstrated a unique profile. Although related closely to the profiles observed for other M. terrae complex subspecies, the chromatogram obtained for our isolate distinctively presented an additional peak in the late cluster (Fig. 1).

The overall biochemical profile of the isolate was similar to those of M. terrae and M. senuense, with the exception of an intermediate instead of a slow growth rate at 37 °C (Table 1). Furthermore, in contrast to the type strains of M. terrae and M. senuense, strain TBE 500028/10^T had a positive urease activity and a different antibiotic-susceptibility pattern (Mun et al., 2008; Murray et al., 2007; Smith et al., 2000). Indeed, unlike strain TBE 500028/10^T, M. terrae and M. senuense are susceptible to rifampicin, amikacin and quinolones (M. senuense only).

Based on the blast sequence match obtained for the 16S rRNA gene (GenBank accession no. GU564404), strain TBE 500028/10^T was shown to be most closely related to M. senuense 05-832^T (1492/1522 nt, 98.0 %), M. terrae ATCC 15755^T (1451/1470 nt, 98.7 %) and Mycobacterium arupense AR30097^T (1458/1495 nt, 97.5 %). The two combined partial sequences of the rpoB gene (GenBank accession no. GU564406) showed that strain TBE 500028/10^T was related most closely to M. senuense 05-832^T (403/421 nt, 95.7 %), M. terrae ATCC 15755^T (591/633 nt, 93.4 %), Mycobacterium confluentis CIP 105510^T (632/703 nt, 89.9 %) and Mycobacterium chitae CIP 105383^T (632/703 nt, 89.9 %). The hsp65 gene sequence (GenBank accession no. GU564405) showed that our strain was related most closely to M. senuense strain DSM 44999^T (421/431 nt, 97.7 %) and M. terrae strain CIP 104321^T (414/426 nt, 97.2 %).

The close relationship to M. terrae and M. senuense was confirmed by phylogenetic analyses (Fig. 2). Trees based on rpoB and hsp65 gene sequences each showed similar groupings for strain TBE 500028/10^T, but slightly different from that of the 16S rRNA gene tree. In the rpoB and hsp65 trees, TBE 500028/10^T forms a branch dependent of M. terrae together with M. senuense, whereas in the 16S rRNA gene tree, strain 500028/10^T forms a branch directly dependent from M. terrae, with M. senuense being situated in another cluster. These topologies were also supported by high bootstrap values.

The species which our isolate resembles most closely, M. terrae and M. senuense, are both slow-growing, non-chromogenic and potentially pathogenic to humans.