The protein antigen MPB70 is a major component of culture supernatants of Mycobacterium bovis and is an active ingredient of bovine PPD used for skin-testing cattle for tuberculosis. We have shown that Mycobacterium kansasii possesses a similar gene that cross-reacts in a PCR test for M. bovis. Single strand conformational polymorphism analysis, and the DNA sequence of the PCR product, shows differences between M. kansasii strains, supporting the suggestion that M. kansasii is not a homogeneous species.
Mycobacterium kansasii is an opportunist human pathogen, causing disease mainly in individuals with predisposing respiratory disease; apart from Mycobacterium avium infections in AIDS patients, M. kansasii is the commonest non-tuberculous mycobacterium as a cause of disease in humans . Mycobacterium kansasii infections also occur occasionally in cattle, causing sensitisation to tuberculin and leading to slaughter of cattle under the UK bovine tuberculosis control programme. It is therefore important to have a reliable test to differentiate between M. bovis and M. kansasii infections.
The protein antigen known as MPB70 is a major component of the culture supernatant of M. bovis and some strains of BCG [2, 3] and is an active ingredient of bovine PPD, the reagent used for skin-testing cattle. Monoclonal antibodies to MPB70 have been produced that are specific for M. bovis with little reactivity to M. tuberculosis (due to the low level of production of this antigen by M. tuberculosis) and no cross-reactivity with other mycobacterial species [4, 5]. On this basis, MPB70 has been proposed as a specific immunodiagnostic reagent for bovine tuberculosis, and a PCR test using the mpb70 gene has also been developed . During evaluation of mpb70-based PCR for detection of M. bovis, we observed a positive result with some strains of M. kansasii. In this paper we describe the identification of an mpb70 analogue in M. kansasii (referred to as mpk70), and the occurrence of variants of this sequence in different strains of M. kansasii.
Materials and methods
The majority of the M. kansasii strains were isolated from animals between 1984 and 1992 at the Central Veterinary Laboratory. Human isolates 8376 and 8388 were from the Mycobacterial Reference Laboratory, Cardiff, and the type strain NCTC10268 was obtained from the National Collection of Type Culture, PHLS, Colindale, London. Mycobacterium bovis strain AN5, BCG Pasteur, and a clinical isolate of M. tuberculosis were from our own stocks.
Primers and PCR
Bacterial colonies were resuspended in 500 μl of water and boiled for 5 min. After cooling to room temperature, the tubes were centrifuged for 5 min at 12.000 rpm (MSE Microcentaur), and 1–2 μl of the supernatant was used for PCR, using an Omnigene thermal cycler (Hybaid). Denaturation for 2 min at 95°C was followed by 30 cycles of 64°C (1.5 min), 72°C (2 min) and 95°C (1 min), ending with one cycle of 64°C (5 min) and 72°C (7 min).
Southern blotting and hybridisation
Southern blotting and hybridisation were carried out essentially as described by Zainuddin and Dale . DNA probes were labelled with digoxigenin by random priming according to the manufacturer's instructions (Boehringer Mannheim). After overnight hybridisation at 65°C, the membranes were washed twice (15 min) in 2×SSC, 0.1% (w/v) SDS at room temperature, and twice (30 min) in 0.1×SSC, 0.1% SDS (w/v) at 65°C. Hybridised probe was detected by chemiluminescence as described by the manufacturers (Boehringer Mannheim).
Single strand conformational polymorphism (SSCP) analysis
PCR amplification for SSCP, using primers A2 and B2, was carried out in the presence of Taq antibody (Clontech), using 20 μM dCTP and 200 μM dATP, dTTP and dGTP. Each reaction contained 1 μCi of 33P-labelled dCTP (NEN). After initial denaturation at 95°C (5 min), we used 30 cycles of 94°C (20 s), 56°C (20 s) and 72°C (50 s) with a final holding temperature of 4°C, using a Perkin Elmer 9600 thermal cycler.
The labelled PCR products were denatured at 95°C for 10 min and rapidly cooled on ice. SSCP analysis was carried out by electrophoresis at 30 W for 5 h through a 6% polyacrylamide gel containing 5% (v/v) glycerol, in an S2 sequencing apparatus (Gibco BRL). The gel was then dried and exposed to X-ray film.
Single strand templates of PCR products were prepared by separation of the PCR product strands using streptavidin-coated magnetic beads (M-280, Dynal). PCR was carried out using standard conditions except that one primer was biotinylated at the 5′ end. After amplification, the products were purified using a Pharmacia micro-spin column S-300 HR and attached to 30 μl of washed streptavidin-coated beads. The non-biotinylated strand was separated from the biotinylated strand as described by the manufacturers (Dynal protocol for single strand preparation from PCR products). The biotinylated strand was resuspended in 5 μl of water and sequenced manually using Sequenase version 2 (USB).
PCR detection of mpk70
The PCR primers A2/B2 amplify a fragment of 259 bp with M. tuberculosis (or M. bovis) DNA as a template. When these primers were tested with a panel of isolates of M. kansasii, a band of similar size was obtained in the majority of cases; in one case, strain 129, a slightly smaller amplified product was detected (Fig. 1). To verify the specificity of the product, the gel was blotted and probed with the amplified products from M. bovis, and from M. kansasii strain 129. Fig. 1a shows that the M. kansasii products hybridize to the M. bovis-derived probe, although much less strongly than the M. bovis product. Conversely, with the M. kansasii 129 probe (Fig. 1B) strong hybridization was seen with M. kansasii strains 129 and 286, with the remaining M. kansasii strains, and M. bovis, hybridizing less strongly.
In a similar experiment (Fig. 1c,d), the amplified product from M. kansasii strain 684 also hybridized strongly to M. kansasii M21, but less strongly to M. kansasii strains 129 and 286, and to M. tuberculosis and M. bovis.
These results indicate the presence of an mpb70 analogue in strains of M. kansasii, provisionally designated mpk70, and further suggest that the sequence of this gene varies between M. kansasii strains.
In order to substantiate the existence of variation in the mpk70 gene, SSCP analysis was carried out, using the PCR product from primers A2/B2, with a selection of M. kansasii strains of animal and human origin, as well as M. tuberculosis, M. bovis and BCG. Fig. 2 shows that the M. kansasii strains fell into three groups: (a) strains 129, 1063, where the PCR product was significantly smaller; (b) strains 447, 516; and (c) the remaining strains, including strain 684 and the NCTC strain 10268, as well as the two human strains. The M. tuberculosis complex strains formed a distinct, homogeneous group.
Sequence analysis of the PCR products confirmed the existence of three versions of the mpk70 gene in these M. kansasii strains; in each case, strains showing similar SSCP patterns had identical sequences. An alignment of the DNA sequences of one representative of each type is shown in Fig. 3a, confirming that the product from strain 129 is shorter than those of the other M. kansasii strains. Closer inspection (Table 1) shows that the sequences from the M. kansasii strains NCTC 10268 and 516 are very similar but distinct, whereas that represented by strain 129 is not significantly closer to the other M. kansasii strains than to the M. bovis mpb70 sequence. This is also true at the amino acid sequence level (Fig. 3b).
Mycobacterium bovis and BCG have been shown to possess a second gene, mpb83, with a high degree of similarity to mpb70[9–11]. The alignments in Fig. 3 show that the amplified regions of the M. kansasii sequences are similar to both mpb70 and mpb83. Southern blot data on digested chromosomal DNA (not shown) using the M. bovis-derived probe indicates that some M. kansasii strains may also possess a second gene related to mpb70, which is not amplified by the PCR primers used here. Therefore, it is possible to regard the M. kansasii gene as an analogue of mpb83 rather than mpb70. Recombinant clones were isolated from a gene library of M. kansasii NCTC10268 by probing with the M. bovis PCR product  and the relevant portion was sequenced (accession number X99760). Alignment of this sequence with the sequences of mpb70 and mpb83 from M. bovis confirmed that the M. kansasii sequence was similar to both M. bovis sequences without a significant difference in the degree of similarity. Since the function of these genes is unclear, it is at present not possible to finally resolve whether the M. kansasii gene reported here is functionally related to mpb70 or to mpb83.
Previous reports have suggested that the mpb70 gene is specific to the M. tuberculosis complex, although only expressed at a high level in M. bovis and some strains of BCG. In particular, Cousins et al.  showed that a PCR test based on mpb70 did not amplify DNA from other mycobacterial species, including M. kansasii. In this paper, the use of a different pair of primers has shown the presence of a related gene in the majority of strains of M. kansasii. The variation in the amplification achieved, and the different hybridization properties of the products, suggests that the sequence of this gene varies between M. kansasii strains; this was borne out by SSCP and by sequence analysis of the products. The presence of this gene, closely related at both DNA and protein sequence levels, in M. kansasii strains indicates that the specificity of diagnostic tests for M. bovis based on either PCR of the mpb70 gene or antibody detection of MPB70 must be evaluated carefully, including testing representatives of different M. kansasii types.
Although some reports have claimed that the genome of M. kansasii is highly conserved , subsequent reports identified a distinct subspecies by Southern blot hybridization patterns, differential hybridization with a species-specific probe, the sequence of the 16S rRNA gene, and the presence of the insertion sequence IS1652[14–16]. The variation in the sequence of the mpk70 gene in M. kansasii is in marked contrast to the mpb70 gene in the M. tuberculosis complex which is identical in M. tuberculosis, M. bovis, and two strains of BCG, despite differences in expression in these strains [5, 7, 17, 18]. These results provide further evidence that the M. kansasii group is sufficiently heterogeneous that it should not be regarded as a single species.
We are grateful to Dr A. Jenkins for supplying human isolates of M. kansasii.