INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Bovine tuberculosis (TB) is a chronic zoonotic disease whose etiological agent is Mycobacterium bovis. It constitutes a serious animal health problem, causing economic losses due to decreased meat and milk production and to low exportation of cattle products.

While the main host of M. bovis is cattle, other animals, including humans, may be affected. Argentina, with approximately 5% of its cattle affected, has one of the highest prevalence rates in South America (9). Tuberculous lesions were also found in 3% of pigs and recently in goats and deer (S. Underwood, S. Pinto, M. Rey Moreno, and J. C. Carfagnini, Abstr. 1st Workshop on Human and Animal Infections Provoked by Chlamydia, Mycobacteria, Brucella, and Borrelia, 1997). In addition, M. bovis can affect domestic animals such as dogs and cats, where the infection route may be food of bovine origin, such as uncooked lung tissue (F. Fernández and E. Morici, Abstr. 1st Workshop on Human and Animal Infections Provoked by Chlamydia, Mycobacteria, Brucella, and Borrelia, 1997). A study performed in the main milk production region of Argentina showed that in the period 1984 to 1989, M. bovis was responsible for 2.4 to 6.2% of the TB cases in human beings, of which 64% were rural and meat workers (11). The AIDS epidemic has also increased the risk of transmission of M. bovis to humans. Nosocomial transmission of TB produced by multiresistant M. bovis strains among human immunodeficiency virus-positive individuals was recently described in Spain (16).

Phenotypic typing methods (serotyping, biotyping, etc.) cannot efficiently discriminate among strains of the M. tuberculosis complex. On the other hand, molecular biology tools, such as restriction fragment length polymorphism (RFLP) or the more recent spoligotyping technique (8, 13), are highly efficient for typing of mycobacteria. In human TB, molecular typing is advanced by the use of the insertion sequence IS6110 (6, 7, 18), which is repeated many times in the M. tuberculosis genome, producing a genotypic heterogeneity of isolates. In M. bovis, IS6110 is less useful because the genome of most strains contains only one or very few IS6110 copies (6, 14, 20). The insertion element IS6110 is frequently found in a unique locus of the M. tuberculosis complex genome called the direct repeat (DR) region (7). The DR region was completely sequenced in M. tuberculosis H37RV (7), in an M. bovis isolate from the United State (3), and in different BCG substrains (19). In addition, the DR region sequence from the M. bovis isolate currently undergoing genome sequencing in England is available (http://www.sanger.ac.uk/projects/M_bovis). The DR region consists of 36-bp repetitive sequences separated by nonrepetitive spacers whose lengths vary from 27 to 41 bp. Deletions in this region alter the spacer composition in each strain. Spoligotyping has been developed based on these DR region properties (8).

Several studies have demonstrated that the degree of differentiation achieved by spoligotyping is higher than that of IS6110 RFLP for strains with a low IS6110 copy number, such as M. bovis (1, 2, 5, 20, 22). In contrast, for strains with high IS6110 copy numbers, like most M. tuberculosis strains, IS6110 RFLP is the more discriminative test (4, 10).

Considering that spoligotyping is a rapid and easy-to-apply technique, it is important to improve its capacity for differentiation. The present study analyzed the DR region of M. bovis isolates for the presence of alternative spacers not included in the spoligotyping membrane. The addition of these alternative spacers to the membrane would probably improve its discriminatory power.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strains. M. bovis DNA from bovine isolates from different parts of South America were used. The strains selected for the study were M. bovis 539 and 540 (pattern B), 554 (pattern E), 541 (pattern A), and 563 (pattern C). M. bovis BCG Pasteur was used as the reference strain. Mycobacterial DNA from wild seals found on the coast of Argentina was also included (15). Alternative-spacer frequency analysis was performed on 20 M. tuberculosis strains from humans and 20 M. bovis strains from cattle.

DNA extraction. Genomic DNA was extracted using proteinase K, lysozyme, and N-cetyl-N,N,N,-trimethylammonium bromide as described by Kamerbeek et al. (8) and Bunschoten et al. (4).

PCR. Primers used for amplification corresponded to the 25 bp of the spacer sequence adjacent to the gaps observed in spoligotyping patterns. Each of these was named according to the number of the spacer it hybridized with. The sequences are as follows: sp1, 5' ATAGAGGGTCGCCGGTTCTGGATCA3'; sp2, 5'CCTCATAATTGGGCGACAGCTTTTG3'; sp5, 5'TTTTCTGACCACTTGTGCGGGATTA3'; sp7, 5'GAGGAGAGCGAGTACTCGGGGCTGC3'; sp13, 5'GGGAGAGGGAATGGCAATGATGGTC3'; sp17, 5'CGGAGTCATCCGCGCGGGCCGGCGC3'; IS-left, TGACCCACCTGACATGACCCCAT (corresponds to the 5' end of IS6110). Primers corresponding to alternative spacers were as follows: sp790, 5'CATGGCACGGCAGGCGTGGCTA3'; sp863, 5'GGGCCGTGGGGCACTTACGG3'; sp1080, 5'GGAGCCGTGCACATGCCGTGGCTCAGG3'; sp1377, 5'GCATGCAGCATGCCGTCCCCGTT3'; sp1453, 5'CGCCATCATCCGGCGCCGCAGCTCCGC3'.

Sequencing. The inserts in the pGEM-T vector were sequenced by the dideoxynucleotide chain termination method (17) using the fmol DNA Sequencing System kit (Promega).

Sequence analysis. Sequences were analyzed using the DNA Strider program for Macintosh (12) and compared with those of M. tuberculosis H37Rv and an M. bovis isolate from the United States (3). This analysis was aimed at finding alternative spacers not included in the spoligotyping membrane.

Analysis of spacer frequency. PCR analysis of spacer frequencies was performed on 20 strains of M. tuberculosis and 20 strains of M. bovis. Pairs of primers corresponding to alternative spacers 790, 863, 1080, 1377, and 1453 and to the 5' end of IS6110 (primer IS-left) were used.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Analysis of spacer frequency in M. bovis spoligotyping patterns. Analysis of the presence of individual spacers in the spoligotyping patterns of 246 M. bovis isolates, mostly from Argentina, (Fig. 1) indicated that the most frequent spacers are those downstream of IS6110. These are present in almost all of the strains, making them relatively useless for differentiation. Conversely, the spacers upstream of IS6110 are less frequent and therefore more polymorphic. Finally, spacers 39 to 43 are absent in all of the M. bovis strains studied in our laboratory. A summary of the structure of the DR region is depicted in Fig. 2.

View larger version (63K): [in this window] [in a new window]   FIG. 1.   Frequency of spacers among 246 M. bovis isolates submitted to spoligotyping. Spacer numbers are shown on the y axis, and the number of isolates containing a given spacer (frequency) is depicted on the x axis.

View larger version (3K): [in this window] [in a new window]   FIG. 2.   Gross structure of the DR region. Spacers with bound primers (===), intervening spacers (---), direct repeats [DR], and IS6110 elements are shown.

Amplification of gap fragments. To identify the presence of alternative spacers in selected strains, an approach consisting of amplification of the gap region between two known spacers was employed. Analysis of PCR was performed comparing the size of the band obtained by PCR to that of the minimum expected size (MES) estimated from the spoligotyping pattern and according to the combination of primers used. The MES of an amplicon between any pair of primers was calculated as the number of spacers that hybridized (observed spots in the spoligotype pattern) between and inclusive of the primers × 70 (the average length of the DR and spacer). Therefore, when the amplicon size exceeded the MES, the presence of an alternative spacer(s) was suspected and the amplicon was sequenced.

Spoligotyping patterns. Based on the M. tuberculosis and M. bovis spoligotype patterns observed in our laboratory, strains presenting an absence of several spacers were selected and their spoligotype patterns were arbitrarily named A with 5 missing spacers (from 8 to 12), B with 10 missing spacers (from 5 to 14), C with 20 missing spacers (from 4 to 23), and E with no major gaps. Mycobacteria isolated from wild seals found on the Argentine coast were also included. The wild-seal spoligotype pattern (D) shows a trait of 15 spacers missing (spacers 8 to 22). Strains representative of each spoligotype pattern were chosen (Fig. 3). Pattern A, called pattern 34 in a previous publication (22), is the major spoligotype pattern of Argentina.

View larger version (131K): [in this window] [in a new window]   FIG. 3.   Representative sample of spoligotype patterns observed with M. bovis isolates. The names of selected spoligotype patterns are marked at the left.

Sequence analysis. Sequences were analyzed to identify alternative spacers. For this purpose, the sequences of cloned amplified fragments were compared to the DR region sequences of M. tuberculosis H37 Rv and an M. bovis isolate from the United States. The sequence from the pattern A strain showed alternative spacer 1377 and repeated spacer 7 (Table 2) in the relative order sp7-sp7-sp1377-sp13, with spacer 1377 situated after the second spacer 7. Upon analysis of the sequence from the M. bovis strain with pattern B, alternative spacer 1080 (Table 2) was found. This spacer is located after spacer 2. Two alternative spacers, 790 and 863, were present in the M. bovis strain with pattern C (Table 2) in the order sp1-sp2-sp790-sp863. The wild-seal strain also showed alternative spacer 1453 (Table 2) located after spacer 7. The BCG fragment showed two alternative spacers, 1377 and 1453, located after spacer 7. The naming of alternative spacers corresponds to their position in the sequence of the M. bovis U.S. strain (3).

Analysis of alternative-spacer frequency of occurrence. To study the frequency of occurrence of alternative spacers in M. tuberculosis and M. bovis strains, amplifications with primers derived from the five alternative spacers, in combination with primer IS-left, were performed (Fig. 4). Twenty different strains each of M. tuberculosis and M. bovis were used. Amplification with sp790 and IS-left was positive in 6 (30%) of the 20 M. bovis strains, while M. tuberculosis gave no amplification; with primers sp863 and IS-left, 18 (90%) of the 20 M. bovis strains and 8 (40%) of the 20 M. tuberculosis strains were amplified. The pair sp1080 and IS-left yielded positive amplification of 16 (80%) of 20 M. bovis strains and only 1 M. tuberculosis strain. Using primers sp1377 and IS-left 9 (45%) of 20 M. bovis strains were positive while only 1 M. tuberculosis strain gave an amplification product. Finally, the combination sp1453-IS-left resulted in 60% of M. bovis and 70% of M. tuberculosis strains amplified.

View larger version (51K): [in this window] [in a new window]   FIG. 4.   Agarose gel showing PCR amplification of DNA from randomly selected M. bovis and M. tuberculosis isolates with primer IS-left in combination with primers derived from alternative spacers. Amplifications of 10 strains of M. bovis and M. tuberculosis are shown. Annealing temperatures for amplifications (in parentheses): A, sp790 (63°C); B, sp863 (63°C); C, sp1080 (65°C); D, sp1377 (63°C); E, sp1453 (55°C); F, IS6110 (65°C) (positive control).

Discriminatory power of alternative spacers. To assess the discriminatory power of alternative spacers, PCRs with combinations of alternative spacers and IS-left primers were performed on 10 pattern A strains (Fig. 5). Using the pair sp790-IS-left, 8 of 10 strains were positive; with sp863-IS-left, all 10 strains were positive; with sp1080-IS-left, 9 of 10 strains were positive; with sp1377-IS-left, 9 of 10 strains were positive; and finally, using the sp1453-IS-left primer pair, 4 of 10 strains were positive (Fig. 5). These results indicated that spacer 1453 may be useful for M. bovis typing, while the other alternative spacers allow moderate discrimination.

View larger version (27K): [in this window] [in a new window]   FIG. 5.   Agarose gel showing PCR amplification of DNA of M. bovis with pattern A using primer IS-left in combination with primers derived from alternative spacers.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

M. bovis strains commonly lack several spacers. These missing spacers appear as gaps in their spoligotyping patterns. To determine whether these gaps are mere deletions of DR-spacer units or whether they contain alternative spacers, these regions were amplified using primers hybridizing to spacers adjacent to the gaps; the resulting fragments were then cloned and sequenced. Five alternative spacers were found in this way. However, all of the amplified fragments contain only one or two alternative spacers, indicating that the gaps can be explained mostly by deletions. Mutated spacers were not observed, and in only one case was a repeated spacer observed. The five alternative spacers were previously described by Beggs et al. (3) in an M. bovis isolate. These spacers were present in an average of 55% of the M. bovis isolates and in only 8% of the M. tuberculosis strains, with the exception of spacer 1453, which was present in 70% of the M. tuberculosis strains tested. It is interesting that this alternative spacer is the only one found in the wild-seal strains, which are genetically intermediate between M. tuberculosis and M. bovis. The fact that certain spacers are found more frequently, or that they are exclusive to a given M. tuberculosis complex species, is emphasized by the fact that spacers 3, 9, 16, and 39 to 43 are not observed in M. bovis (Fig. 1). In this work, the absence of an alternative spacer was detected by negative amplification using an alternative-spacer-IS-left primer pair. Another explanation for a no-amplification event could be that the IS6110 element may be inverted in a given strain. However, we believe that this is not likely because in most cases the same strain that gave negative amplification rendered positive amplification with another primer pair, meaning that the IS6110 orientation is not inverted.

As noted previously by other authors (19), the current DR region seems to have evolved from an ancestor possessing all of the present spacers and probably others. The deletions seem to involve the entire DR in addition to the spacer units. The fact that the sequence of the 36-bp DR element is highly conserved between strains and species suggests that deletions have probably arisen by recombination between DR sequences. Another probable mechanism is the duplication and insertion of primordial DR-plus-spacer units, followed later by sequence divergence. However, this mechanism seems unlikely since, with the exception of a few spacers, their sequences are significantly different from each other and in consequence they do not seem to have evolved from a common ancestor. The spacers of the DR region have a conserved order and sequence. In addition, the repeated spacers found in this study are in tandem with the original spacer. The alternative spacers also conserve the order they present in the previously studied M. bovis isolated in the United States (3).

In accordance with the aim of the present study, how much discrimination would be gained by including new spacers in the spoligotyping membrane? Isolates with pattern A, the major spoligotype pattern in Argentina, were selected to test the discriminative power of the five alternative spacers in this cluster of strains. Spacer 1453 showed the highest discriminative power, since only 40% of the strains were positive. These results suggest than some alternative spacers may be more discriminative for M. bovis typing than those already included in the membrane.