SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Mycobacterium tuberculosis and Mycobacterium bovis cause tuberculosis, which is responsible for the deaths of more people each year than any other bacterial infectious disease. Disseminated disease with Mycobacterium bovis BCG, the only currently available vaccine against tuberculosis, occurs in immunocompetent and immunodeficient individuals. Although mycobacteria are obligate aerobes, they are thought to face an anaerobic environment during infection, notably inside abscesses and granulomas. The purpose of this study was to define a metabolic pathway that could allow mycobacteria to exist under these conditions. Recently, the complete genome of M. tuberculosis has been sequenced, and genes homologous to an anaerobic nitrate reductase (narGHJI), an enzyme allowing nitrate respiration when oxygen is absent, were found. Here, we show that the narGHJI cluster of M. tuberculosis is functional as it conferred anaerobic nitrate reductase activity to Mycobacterium smegmatis. A narG mutant of M. bovis BCG was generated by targeted gene deletion. The mutant lacked the ability to reduce nitrate under anaerobic conditions. Both mutant and M. bovis BCG wild type grew equally well under aerobic conditions in vitro. Histology of immunodeficient mice (SCID) infected with M. bovis BCG wild type revealed large granulomas teeming with acid-fast bacilli; all mice showed signs of clinical disease after 50 days and succumbed after 80 days. In contrast, mice infected with the mutant had smaller granulomas containing fewer bacteria; these mice showed no signs of clinical disease after more than 200 days. Thus, it seems that nitrate respiration contributes significantly to virulence of M. bovis BCG in immunodeficient SCID mice.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

In recent years, an increased incidence of tuberculosis in both developing and developed countries, the emergence of drug-resistant strains and a deadly synergy with the human immunodeficiency virus (HIV) have been reported. One-third of the world’s population is thought to be infected with Mycobacterium tuberculosis. It is estimated that 90 million new tuberculosis cases resulting in 30 million deaths can be expected between 1990 and 1999 (Raviglione et al., 1995). In 1993, the World Health Organization (WHO) declared tuberculosis a global emergency: novel therapeutic approaches and a better understanding of the pathogenesis of tuberculosis are needed to develop new treatment concepts and a more effective and safer vaccine.

We used Mycobacterium bovis BCG and its virulence in immunodeficient mice (SCID) as a model to explore the physiology of the Mycobacterium tuberculosis complex in vivo. M. bovis BCG, the only currently available vaccine against tuberculosis, was attenuated from virulent Mycobacterium bovis by repeated passaging in vitro. However, the strain retained at least some of its original virulence. After vaccination with M. bovis BCG, disseminated disease developed in immunocompromised individuals, which proved to be fatal in several cases (Reyn et al., 1987; Weltman and Rose, 1993; Casanova et al., 1996; Jouanguy et al., 1996; Newport et al., 1996; Emile et al., 1997; Rosenfeldt et al., 1997; Talbot et al., 1997; Altare et al., 1998). Like M. tuberculosis, M. bovis BCG forms granulomas in multiple organs after intravenous infection of mice, and both grow progressively in immunodeficient (SCID) mice, eventually leading to a fatal outcome (North and Izzo, 1993).

This study was undertaken to explore nitrate metabolism and the role it might play in mycobacterial pathogenesis. In contrast to pathogens that cause short-lived infections, mycobacteria become firmly established within tissues, making the organism dependent on specific nutrients at particular sites for survival. As a response to infection, granulomas and abscesses arise in various organs. As all mycobacteria are obligate aerobes, the availability of oxygen seems crucial, yet granulomas and abscesses are believed to be deprived of oxygen (Barclay and Wheeler, 1989). Even in the lung, oxygen appears to be limited: superinfection with obligately anaerobic organisms, indicating an anaerobic environment, can occur in cavities of pulmonary tuberculosis (Garay, 1996).

The recent determination of the complete genome sequence of M. tuberculosis provided a tremendous opportunity for investigating molecular mechanisms of disease that are overlapping between M. bovis BCG and M. tuberculosis as they share 99.9% of DNA (Gordon et al., 1999). It turned out that within the M. tuberculosis genome, genes homologous to the anaerobic nitrate reductase of Bacillus subtilis (narGHJI) were present (Cole et al., 1998). Detailed studies of nitrate respiration were carried out in Bacillus subtilis, an overview of which has been published recently (Nakano and Zuber, 1998). Anaerobic nitrate reductase activity is encoded by four genes, narGHJI, clustered together in an operon. NarG, H and I are subunits of nitrate reductase, with NarG probably being the catalytic subunit, whereas NarJ is required for the assembly of the enzyme.

Here, we analysed nitrate reductase activity of the narGHJI gene cluster of M. tuberculosis under anaerobic conditions, looked for anaerobic nitrate reductase activity of M. bovis BCG and investigated the contribution of anaerobic nitrogen reductase to the pathogenesis of an M. bovis BCG infection in immunodeficient SCID mice.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Cloning of narGHJI from Mycobacterium tuberculosis and transfer of anaerobic nitrate reductase activity into Mycobacterium smegmatis

Three cosmids (pIW2, pIW10 and pIW12) carrying the entire narGHJI cluster were obtained from a cosmid library of M. tuberculosis H37Rv by screening ≈ 2000 clones with a 511 bp oligonucleotide probe directed against narGH. Genomic DNA from M. tuberculosis H37Rv, DNA from pIW2, pIW10 and pIW12, and genomic DNA from M. bovis BCG were compared by Southern blot analysis. All five DNAs shared an identical EcoRV fragment, an identical NheI fragment and two identical ApaI fragments (data not shown). Thus, Southern blot analysis confirmed the presence of the narGHJI gene cluster in pIW2, pIW10 and pIW12, and revealed a gene locus in the genome of M. bovis BCG that was identical to the narGHJI locus of M. tuberculosis.

To analyse the narGHJI-bearing cosmids in vitro, we looked for a mycobacterial strain that would not express nitrate reductase activity under anaerobic conditions. Using a standard assay that indicates production of nitrite from nitrate by forming a red dye, we found that the non-virulent fast-growing strain mc2155 of Mycobacterium smegmatis lacked anaerobic nitrate reductase activity when incubated for 5 or 10 days in an anaerobic jar in minimal basal (MB) medium supplemented with 10 mM nitrate. Figure 1 shows that M. smegmatis mc2155 and M. smegmatis mc2155 transformed with some uncharacterized cosmid of the M. tuberculosis library did not accumulate nitrite under anaerobic conditions (Fig. 1, tubes 1 and 2). In contrast, M. smegmatis transformed with pIW2, pIW10 or pIW12 produced nitrite from nitrate under anaerobic conditions (Fig. 1, tubes 3–5). None of the five strains reduced nitrate to nitrite when growing aerobically in MB medium supplemented with 10 mM nitrate that was shaken vigorously to ensure good aeration (data not shown). This suggests that narGHJI of M. tuberculosis encodes a functional anaerobic nitrate reductase whose activity can be transferred into an anaerobic nitrate reductase-negative strain of M. smegmatis.

image

Figure 1. Functional analysis of the narGHJI cluster of M. tuberculosis in vitro. Anaerobic nitrate reductase activity of M. smegmatis mc2155 not transformed (tube 1), transformed with an uncharacterized cosmid from the M. tuberculosis library (tube 2), or transformed with three cosmids (pIW2, pIW10, and pIW12) carrying the narGHJI cluster of M. tuberculosis (tubes 3–5) is shown. Bacteria (1 × 106 ml−1) were cultured for 5 days under anaerobic conditions in MB medium supplemented with 10 mM nitrate and supernatants were tested for production of nitrite. Nitrite is indicated by the red colour.

Download figure to PowerPoint

Anaerobic nitrate reductase activity of Mycobacterium bovis BCG and of a narG mutant of Mycobacterium bovis BCG in vitro

We decided to determine a possible role of anaerobic nitrate reductase activity in pathogenesis of mycobacteria by generating a narG mutant of M. bovis BCG and analysing the mutant in immunodeficient SCID mice. First, anaerobic nitrate reductase activity of M. bovis BCG was examined. Bacteria were cultured in MB medium supplemented with 10 mM nitrate in a roller bottle to ensure good aeration or in an anaerobic jar. Nitrite was produced only under anaerobic conditions, whereas bacteria cultured aerobically did not generate measurable amounts of nitrite (Fig. 2). We assumed that anaerobic nitrate reductase activity in M. bovis BCG could be mediated by the narGHJI cluster. This idea was tested by deleting the narG gene of M. bovis BCG and analysing the mutant for anaerobic nitrate reductase activity.

image

Figure 2. Anaerobic nitrate reductase activity of M. bovis BCG wild type. Bacteria (1 × 106 ml−1) were inoculated in MB medium supplemented with 10 mM nitrate. An aerobic environment was achieved by using roller bottles to ensure good aeration. Anaerobic conditions were generated in a jar using a standard anaerobic gas pack. After 5 or 10 days, the supernatants were tested for of nitrite. Absorbency of supernatant was measured and compared with a known standard of nitrite. Production of nitrite is shown (10−6 M).

Download figure to PowerPoint

A 736 bp fragment within the narG gene was replaced by a hygromycin cassette and the deletion mutation was cloned into the vector pMP7 as described in the Experimental procedures. M. bovis BCG was transformed and, of 300 analysed recombinants, two clones (IW1 and IW2) turned out to carry the disrupted narG gene. Figure 3 shows the genetic organization of the two mutants (Fig. 3, lanes 2 and 3) in comparison to M. bovis BCG wild type (Fig. 3, lane 1) and M. tuberculosis (Fig. 3, lane 4). Allelic exchange deleted a NotI restriction site and added a SacI restriction site at the target site in the chromosome of the narG mutant. Therefore, compared with M. bovis BCG wild type and M. tuberculosis, the mutant strains showed a larger fragment after digestion of genomic DNA with NotI (Fig. 3, left) and a smaller fragment after digestion with SacI (Fig. 3, centre) when hybridized with the 511 bp probe binding outside the mutated region of narG. Hybridization with a probe homologous to the deleted narG fragment confirmed disruption of narG. M. bovis BCG wild type and M. tuberculosis showed the expected 5530 bp fragment (Fig. 3, right, lanes 1 and 4), whereas the narG mutants showed no signal (Fig. 3, right, lanes 2 and 3).

image

Figure 3. Southern blot analysis of genomic DNA from M. bovis BCG wild type (lane 1), M. tuberculosis (lane 4) and genomic DNA from IW1 and IW2, the two narG mutants of M. bovis BCG (lanes 2 and 3 respectively). DNA was digested with either NotI or SacI. Each lane was loaded with 1 µg of DNA and probed with either a 511 bp fragment encompassing parts of narGH (left and centre) binding outside the deleted narG fragment or a 395 bp fragment binding inside the deleted narG fragment. The wild-type fragment is the expected 5330 bp size for NotI (lanes 1 and 4, left and right panels) and 3796 bp size for SacI (lanes 1 and 4, centre panel), whereas the mutant has the expected 7988 bp size for NotI (lanes 2 and 3, left) and 1681 bp size for SacI (lanes 2 and 3, centre). With the second probe encompassing the deleted fragment, no hybridization signal was seen in the mutant (lanes 2 and 3, right).

Download figure to PowerPoint

To test the phenotype of the two mutants in vitro, the anaerobic nitrate reductase activity of IW1 and IW2 was compared with that of M. bovis BCG. All strains were cultured in MB medium containing 10 mM nitrate under anaerobic conditions, and aliquots were analysed at various time points. Both narG mutants were unable to reduce nitrate to nitrite, even after prolonged incubation for more than 40 days (Fig. 4). This suggests that narGHJI mediates the anaerobic nitrate reductase activity of M. bovis BCG. We tested for differences in growth and viability between the mutants and the wild type under anaerobic conditions. Using MB medium supplemented with 10 mM nitrate, we could not detect growth of M. bovis BCG wild type in the first place. The wild type persisted over weeks before viability slowly declined, and so did both mutants.

image

Figure 4. Anaerobic nitrate reductase activity of two nitrate reductase mutants of M. bovis BCG (IW1 and IW2). Bacteria (1 × 106 ml−1) were cultured anaerobically in MB medium supplemented with 10 mM nitrate and production of nitrite was measured at various time points. M. bovis BCG wild type was used as a positive control.

Download figure to PowerPoint

The mutants were cultured aerobically in 7H9 medium and plated for colony-forming units (cfu) to exclude a general growth impairment affecting the narG mutant in vitro. There was no detectable growth deficiency of the mutants compared with the M. bovis BCG wild type under aerobic conditions (Fig. 5). Both of the mutants and the wild type were also cultured under aerobic conditions on MB plates supplemented with 10 mM nitrate (data not shown). Again, no difference in growth was detected.

image

Figure 5. Growth of IW1 and IW2, the two nitrate reductase mutants of M. bovis BCG, compared with M. bovis BCG wild type under aerobic conditions. All three strains were cultured in 7H9 medium in roller bottles to ensure good aeration, and aliquots taken on days 0,1,2,3,9 and 10 were plated for colony-forming units on 7H10 medium.

Download figure to PowerPoint

Although we clearly demonstrated that the narG mutants lacked anaerobic nitrate reductase activity, we failed to find a significant difference in growth or survival between the wild type and the narG mutants in vitro under anaerobic conditions. Nevertheless, we decided to compare the anaerobic nitrate reductase mutant and M. bovis BCG wild type in SCID mice as we reasoned that in vivo additional factors might be crucial for growth under anaerobic conditions that we had not supplied in vitro.

Virulence of the anaerobic nitrate reductase mutant of Mycobacterium bovis BCG in immunodeficient SCID mice

M. bovis BCG causes a progressive disease in immunodeficient SCID mice and eventually lead to a fatal infection. For in vivo analysis of virulence, immunodeficient SCID mice were infected with 1 × 106 bacilli from the nitrate reductase mutant of M. bovis BCG (IW1) or 1 × 106 bacilli from M. bovis BCG wild type. To obtain histological samples from the liver and the lung, mice infected with either strain were killed 2 weeks after infection, and mice infected with IW1 were killed after 12 weeks when all mice infected with the wild type had succumbed to the infection. Figure 6 shows that after 2 weeks epitheloid cell granulomas containing acid-fast bacilli had formed in the liver of mice infected with either strain. After 12 weeks, granulomas in the liver of those mice infected with M. bovis BCG wild type were virtually filled with bacilli that appeared to be mainly intracellular. In contrast, granulomas of mice infected with the mutant contained notably fewer bacteria. The difference was even more striking in the lung. Figure 7 shows that after 2 weeks small granulomas with acid-fast bacilli had formed in mice infected with M. bovis BCG wild type. In the lungs of mice infected with the mutant, bacilli were found inside interstitial cells without discernible granulomas. After 12 weeks, large granulomas overwhelmed with bacteria had almost replaced the entire lung of mice infected with M. bovis BCG wild type. In contrast, mice infected with the nitrate reductase mutant showed small granulomas, again with notably fewer bacteria. In parallel to these experiments, two groups of mice, each containing five animals infected with 1 × 106 cfu from M. bovis BCG wild type or 1 × 106 cfu from the nitrate reductase mutant, were observed for external signs of disease. Mice infected with M. bovis BCG wild type changed in appearance after 50 days, showing wasting and becoming lethargic; eventually, all five mice succumbed between days 76 and 84 after infection (Fig. 8). In contrast, all SCID mice infected with the anaerobic nitrate reductase mutant showed no external sign of disease at any time. When this report was written, they were still alive after more than 200 days (P = 0.001 in a log rank test of survival).

image

Figure 6. Histology of the liver of SCID mice infected with 1 × 106 cfu of bacilli of M. bovis BCG wild type or the anaerobic nitrate reductase mutant IW1. Mice were sacrificed 2 and 12 weeks after infection, and sections of the liver were stained with Ziehl-Neelsen and counterstained with haematoxylin.

Download figure to PowerPoint

image

Figure 7. Histology of the lung of SCID mice infected with 1 × 106 cfu of bacilli of M. bovis BCG wild type or the anaerobic nitrate reductase mutant IW1. Mice were sacrificed 2 and 12 weeks after infection, and sections of the lung were stained with Ziehl-Neelsen and counterstained with haematoxylin.

Download figure to PowerPoint

image

Figure 8. Survival of mice infected with 1 × 106 cfu of bacilli of M. bovis BCG wild type or the anaerobic nitrate reductase mutant IW1. Five mice were infected in each group. Log rank test of survival yielded a P-value of 0.001 for the difference between the survival of the wild type and the survival of the mutant.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Once pathogenic mycobacteria gain entry into the host, they acquire and utilize nutrients present in the host to sustain themselves. One of the key features is their rapid and ready ability to respond to environmental changes by inducing enzymes and altering metabolic activity (Barclay and Wheeler, 1989). The present study has several implications for mycobacterial metabolic activity in vitro and in vivo. First, both M. bovis BCG and M. tuberculosis express a functional anaerobic nitrate reductase in vitro. Second, nitrate is an important nutrient in vivo, and is provided sufficiently within the granuloma. Third, M. bovis BCG adapts to an anaerobic environment in vivo by nitrate respiration.

It has been shown for other bacteria that nitrate reduction is carried out by different enzymes, among them a membrane-bound nitrate reductase encoded by narGHJI and a soluble nitrate reductase encoded by nasA (Gennis and Stewart, 1996). The former is expressed anaerobically and was identified in the complete genome of M. tuberculosis, the latter is expressed aerobically and was not found in the genome of M. tuberculosis (Cole et al., 1998). We demonstrated that the gene cluster of narGHJI was not only present in M. tuberculosis but also in M. bovis BCG. Moreover, it transferred anaerobic nitrate reductase activity to M. smegmatis. M. bovis BCG used in this study was nitrate reductase negative when grown aerobically, which was in agreement with previous findings. However, M. bovis BCG expressed a nitrate reductase activity under anaerobic conditions, which had not been found in a previous study addressing the same issue (Virtanen, 1960). Anaerobic nitrate reductase activity of M. bovis BCG was confirmed by generating a mutant of M. bovis BCG with a partial deletion of the narG gene. The mutant was unable to reduce nitrate to nitrite under anaerobic conditions.

There is a long-standing interest in the concept that adaptation to an anaerobic microenvironment represents an important stage in mycobacterial disease. It is generally accepted that the membrane-bound anaerobic nitrate reductase couples the reduction of nitrate to the generation of a proton electrochemical potential gradient across the cytoplasmic membrane, which in turn is used for synthesis of ATP. How mycobacteria metabolize nitrite further is open to speculation. Interestingly, nitrite seems to be toxic for mycobacteria (Ratledge, 1982). In principle, nitrite metabolism could follow two pathways. Denitrification is accomplished by a series of anaerobic enzymes that convert nitrite (NO2) through nitric oxide (NO) and nitrous oxide (N2O) to dinitrogen (N2). Alternatively, an anaerobic nitrite reductase may be used to detoxify some of the nitrite and produce ammonium. Different enzymes are involved and no additional energy is obtained (Mortlock, 1998). In practice, the latter pathway is more likely to be used as genes homologous to an anaerobic nitrite reductase (nirBD) have been identified in the genome of M. tuberculosis, whereas genes for denitrification were not found (Cole et al., 1998). In contrast to anaerobic nitrate reductase activity, it has been known for many years that M. tuberculosis is strongly positive for nitrate reductase activity under aerobic conditions whereas M. bovis and M. bovis BCG are not, and this is now a frequently used method for differentiating between the species (Virtanen, 1960; Hedgecock and Costello, 1962; Ratledge, 1982). The molecular basis for this phenomenon remains uncertain.

A significant contribution of anaerobic nitrate reductase to virulence in vivo has not been shown before for any other bacterial species. With the knock-out mutant available, it was possible to test the role of anaerobic nitrate reductase in the pathogenesis of an M. bovis BCG infection in immunodeficient mice. Mycobacteria have come to be regarded as a paradigm for an obligately aerobic organism. It has been speculated that in vivo tubercle bacilli are oxygen-limited in their metabolism, explaining the predilection of M. tuberculosis for growing in lung tissues (Wheeler and Ratledge, 1994). However, during infection, tubercle bacilli seem to cope with oxygen depletion and reach high numbers in abscesses and granulomas. The ability to switch to nitrate respiration in an anaerobic environment would reconcile this apparent paradox. Histology of mice infected with M. bovis BCG showed formation of large granulomas that were loaded with bacteria, whereas mice infected with the mutant had only small granuloma with notably fewer bacteria. These results may be explained by burgeoning granulomas that become depleted of oxygen but supply enough nitrate, a nutrient that enables M. bovis BCG to grow, whereas the narG mutant, being incapable of using nitrate under anaerobic conditions, stops replicating. We are currently setting up experiments that will answer the question of whether immunodeficient mice eventually clear the infection with an anaerobic nitrate reductase mutant of M. bovis BCG or whether the mutant persists without causing disease.

In a practical sense, an anaerobic nitrate reductase mutant of M. bovis BCG could be useful as a safer live vaccine against tuberculosis. However, it has been speculated that M. tuberculosis might express unique protective antigens and epitopes not present in M. bovis BCG that could provide greater protection than M. bovis BCG (Kaufmann, 1993; Guleria et al., 1996). If an anaerobic nitrate reductase mutant of M. tuberculosis was attenuated in both immunocompetent and immunodeficient mice, it might even be preferable as a safe and highly protective live vaccine against tuberculosis. In addition, M. bovis BCG has been used for immunotherapy of superficial transitional cell carcinoma of the bladder, where it remains superior to all other treatments (Nseyo and Lamm, 1997). Besides prolonged persistence of bacilli in the urinary tract sometimes requiring palliative cystectomy, systemic complications including M. bovis BCG sepsis, pneumonitis, hepatitis and osteomyelitis after intravesical M. bovis BCG administration have been described (Katz et al., 1992; Lamm et al., 1992; Bowyer et al., 1995; Guerra et al., 1998; Aljada et al., 1999). If it were demonstrated that a nitrate reductase mutant of M. bovis BCG could induce local inflammation of the epithelium of the bladder without causing adverse reactions, such as invasive disease, it would improve acceptance of this approach for treatment of transitional cell carcinoma of the bladder.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Bacteria

Mycobacterium bovis bacillus Calmette-Guérin Pasteur (BCG) (Pasteur Vaccine strain; Statens Serum Institute) and M. smegmatis mc2155 (efficient plasmid transformation mutant; Snapper et al., 1990) were used in this study. 7H9 broth or 7H10 plates (Difco Laboratories) supplemented with 0.2% glycerol, 0.05% Tween 80 and 10% ADS (0.5% bovine albumin fraction V, 0.2% glucose, 140 mM NaCl) were used for culturing of mycobacteria unless indicated otherwise.

Nitrate reductase activity test

Nitrate is reduced to nitrite by nitrate reductase. The presence of nitrite can be demonstrated by naphthylamide and sulphanilic acid reagents, which form a red diazonium dye when reacting with nitrite (Metchock et al., 1999). Before testing for nitrate reductase activity, bacteria were grown under aerobic conditions in 7H9 broth to an OD600 of between 0.7 and 1.0. To test for nitrate reductase activity, 1 × 106 bacilli ml−1 were incubated under aerobic or anaerobic conditions in MB medium with 0.5 mM MgCl2, 0.5 mM CaCl2, 0.2% glycerol, 0.05% Tween 80, and 10% ADS and 10 mM NO3 (1 l of MB medium contained 1 g of KH2PO4, 2.5 g of Na2HPO4, 2.0 g of K2SO4 and 2 ml of trace elements; 1 l of trace elements contained 40 mg of ZnCl2, 200 mg of FeCl3·6H2O, 10 mg of CuCl2·4H2O, 10 mg of MnCl2·4H2O, 10 mg of Na2B4O7·10H2O and 10 mg of (NH4)6Mo7O24·4H2O). Anaerobic conditions were achieved with the AnaeroGen anaerobic system (Oxoid) in a standard anaerobic jar. As recommended by the manufacturer, an indicator strip was used to confirm anaerobic conditions. At various time points, 100 µl of sulphanilic acid and 100 µl of N,N-dimethyl-1-naphthylamine (both reagents were taken from the api system, bioMerieux) were added to 1 ml of culture and centrifuged at 15 000 g for 15 min at room temperature. Absorbency of supernatant was measured at 530 nm and compared with a known standard of nitrite (ranging from 1 ∞ 10−6 M to 1 ∞ 10−3 M).

Isolation of cosmids containing narGHJI

Genes for anaerobic nitrate reductase (narGHJI) of M. tuberculosis H37Rv were identified in a genomic cosmid library by colony blot hybridization. The library was constructed by insertion of partially digested genomic DNA of M. tuberculosis H37Rv into the BclI site of the cosmid vector pYUB412 and contains inserts with an average size of between 35 and 40 kbp (Bange et al., 1999). For isolation of narGHJI, 2000 clones were screened with an oligonucleotide probe directed against narGH. The probe had been generated by DNA amplification under standard conditions using the following oligonucleotides as PCR primers: forward primer 5′-TCCAGGTGGTCGTTTCACTC-3′ and reverse primer 5′-CAACCAATGCATTTGTCGAG-3′. DNA amplification was performed in 35 cycles using 55°C for annealing, 72°C for elongation and 95°C for denaturation. Following the protocol given by the manufacturer, the 511 bp DNA probe was gel purified and labelled with a DIG DNA-labelling and detection kit and was used for colony blot hybridization and for Southern blot hybridization (no. 1093657, Boehringer Mannheim).

From 12 positive clones, cosmid DNA was prepared and further analysed by EcoRV, NheI and ApaI restriction digests. The narGHJI gene locus encompasses 6818 bp. The flanking EcoRV restriction sites produced a 8652 bp fragment containing the entire narGHJI operon. The flanking NheI restriction sites produced a 11 710 bp fragment carrying almost the entire narGHJI cluster, except for a 222 bp fragment at the 3′ end of narI. Three ApaI sites within the narGHJI cluster produced a larger fragment of 3435 bp and a smaller one of 1855 bp.

Allelic exchange of narG

Because of considerable technical difficulty in generating a knock-out mutant in slow-growing mycobacteria, mycobacterial plasmids bearing the counterselectable marker sacB have been used for allelic exchange of deletion mutations in the chromosome (Pelicic et al., 1996; Pavelka and Jacobs, 1999). We chose a type of allelic exchange methodology, applying the mycobacterial plasmid pMP7 that had been constructed and successfully used previously (W. R. Jacobs, Jr, J. D. McKinney, and M. S. Pavelka, unpublished). Plasmid pMP7, a shuttle vector replicating in E. coli and in mycobacteria, carries a kanamycin resistance cassette as a selectable marker, as well as the sacB gene as a counterselectable marker. The plasmid had been proven to be unstable in M. bovis BCG when grown in media without kanamycin and was therefore lost after allelic exchanged had occurred (J. D. McKinney, personal communication). A 3490 bp ApaI fragment encompassing narG of M. tuberculosis was cloned into the ClaI–EcoRI sites of pBSK(–). A 736 bp ClaI fragment of the narG gene was replaced by a hygromycin resistance gene and the mutated narG gene was cloned into the unique NotI site of pMP7. M. bovis BCG wild type was transformed and clones were selected on plates with 50 µg ml−1 hygromycin.

Next, several clones were grown in 10 ml of 7H9 medium without selection. From saturated cultures, clones were selected on 7H10 plates containing 50 µg ml−1 hygromycin and 2% sucrose. Of 300 clones that were resistant to hygromycin and sucrose, 33 clones were also sensitive to kanamycin, suggesting loss of the pMP7 and integration of the mutated narG gene. Clones were subjected to Southern blot analysis. Chromosomal DNA was digested with NotI and SacI. The deleted fragment of the narG site contained a NotI site and the inserted hygromycin cassette contained a SacI site. Hybridization with the 511 bp DNA probe (see above) that binds outside the deleted DNA fragment therefore resulted in a larger fragment (7988 bp) than the wild-type gene (5530 bp) when the DNA was digested with NotI and a smaller fragment (1681 bp) than the wild type (3796 bp) when the DNA was digested with SacI. In addition, we constructed a probe using 5′-TCGGACTTTGACGCATTCGC-3′ as a forward primer and 5′-GTATCGGCGTAGGTGATGCG-3′ as a reverse primer for amplification of a 395 bp DNA product which hybridized to the DNA fragment that was deleted from the narG mutant of M. bovis BCG. Eventually, two clones, named IW1 and IW2, were proven to have the expected deletion of the narG gene.

Infection of SCID mice

Balb/c SCID mice (6–10 weeks old) were obtained from the Zentrales Tierlaboratorium of the Medical School of Hannover, Germany. Before infection, M. bovis BCG and IW1 were grown in 7H9 medium. The infecting dose was established and confirmed by plating on 7H10 medium for colony counts both before and immediately after injection of mice. Intravenous infection of SCID mice was performed through needle puncture of the tail vein. Organs were harvested at the indicated times and were homogenized in PBS with 0.05% Tween 80 using a homogenizer (Ultra-Turrax T8; IKA-Werke), diluted and plated onto supplemented 7H10 media (Difco). For survival studies of SCID mice, five mice were injected with M. bovis BCG wild type and five mice were injected with the nitrate reductase mutant of M. bovis BCG. Animal work in this article has been approved by the ethics committee of the county government of Hannover, Germany, on 18 March 1999 (reference number 509c42502-99/163).

Histology

Small pieces of the lungs and the livers from the mice were fixed for 24 h in 10% buffered (pH 7.0) formalin and were embedded in paraffin. Sections (3–4 µm thick) were cut with steel knives using an ultramicrotome. The sections were stained for acid-fast bacteria with Ziehl-Neelsen stain and counterstained with haematoxylin.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We are grateful to W. R. Jacobs, Jr, who provided pMP7 and mc2155. We wish to thank M. S. Pavelka and J. D. McKinney for their advice on allelic exchange in mycobacteria, and M. S. Pavelka for constructing pMP7. We thank D. Bitter-Suermann and E. C. Böttger for their interest. F.-C.B. was supported by a postdoctoral fellowship of the ‘Infektionsforschung und AIDS – Stipendiumprogramm’ of the German government. The work was supported by the Deutsche Forschungsgemeinschaft.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Aljada, I.S., Crane, J.K., Corriere, N., Wagle, D.G., Amsterdam, D. (1999) Mycobacterium bovis BCG causing vertebral osteomyelitis (Pott’s disease) following intravesical BCG therapy J Clin Microbiol 37: 21062108.
  • Altare, F., Lammas, D., Revy, P., Jouanguy, E., Döffinger, R., Lamhamedi, S., et al. (1998) Inherited interleukin 12 deficiency in a child with bacille Calmette-Guerin and Salmonella enteritidis disseminated infection J Clin Invest 102: 20352040.
  • Bange, F.-C., Collins, F.M., Jacobs, Jr, W.R., (1999) Survival of SCID mice infected with Mycobacterium smegmatis containing large DNA fragments from Mycobacterium tuberculosis. Tubercle Lung Dis 79: 171180.
  • Barclay, R. & Wheeler, P.R. (1989) Metabolism of mycobacteria in tissues. In The Biology of the Mycobacteria, Vol. 3. Ratledge, C., Standford, J., and Grange, J.M. (eds). London: Academic Press, pp. 37196.
  • Bowyer, L., Hall, R.R., Reading, J., Marsh, M.M. (1995) The persistence of bacille Calmette-Guerin in the bladder after intravesical treatment for bladder cancer. Br J Urol 75: 188192.
  • Casanova, J.L., Blanche, S., Emile, J.F., Jouanguy, E., Lamhamedi, S., Altare, F., et al. (1996) Idiopathic disseminated bacillus Calmette-Guerin infection: a French national retrospective study Pediatrics 98: 774778.
  • Cole, S.T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C., Harris, D., et al. (1998) Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence Nature 393: 537544.
  • Emile, J.F., Patey, N., Altare, F., Lamhamedi, S., Jouanguy, E., Boman, F., et al. (1997) Correlation of granuloma structure with clinical outcome defines two types of idiopathic disseminated BCG infection J Pathol 181: 2530.
  • Garay, S.M. (1996) Pulmonary tuberculosis. In Tuberculosis, 1st edn. Rom, W.N., and Garay, S.M.(eds). Boston: Little, Brown and Company, pp. 373412 .
  • Gennis, R.B. & Stewart, V. (1996) Respiration. In Escherichia coli and Salmonella, Vol. 1, 2nd edn. Neidhardt, F.C. (ed.). Washington, DC: American Society for Microbiology Press, pp. 408433.
  • Gordon, S.V., Brosch, R., Billaut, A., Garnier, T., Eigelmeier, K., Cole, S.T. (1999) Identification of variable regions in the genomes of tubercle bacilli using bacterial artificial chromosome arrays Mol Microbiol 32: 643655.
  • Guerra, C.E., Betts, R.F., O’Keefe, J.O., Shilling, J.W. (1998) Mycobacterium bovis osteomyelitis involving a hip arthroplasty after intravesicular Bacille Calmette-Guerin for bladder cancer. Clin Infect Dis 27: 639640.
  • Guleria, I., Teitelbaum, R., McAdam, R.A., Kalpana, G., Jacobs, Jr, W.R., Bloom, B.R. (1996) Auxotrophic vaccines for tuberculosis Nature Med 2: 334337.
  • Hedgecock, L.W. & Costello, R.L. (1962) Utilization of nitrate by pathogenic and saprophytic mycobacteria J Bacteriol 84: 195205.
  • Jouanguy, E., Altare, F., Lamhamedi, S., Revy, P., Emile, J.-F., Newport, M., et al. (1996) Interferon-γ-receptor deficiency in an infant with fatal bacille Calmette-Guerin infection N Engl J Med 335: 19561961.
  • Katz, D.S., Wogalter, H., D’Esposito, R.F., Cunha, B.A. (1992) Mycobacterium bovis vertebral osteomyelitis and psoas abscess after intravesical BCG therapy for bladder carcinoma. Urology 40: 6366.
  • Kaufmann, S.H.E. (1993) Immunity to intracellular bacteria Annu Rev Immunol 11: 129163.
  • Lamm, D.L., Van Der Meijden, P.M., Morales, A., Brosman, S.A., Catalona, W.J., Herr, H.W., et al. (1992) Incidence and treatment of complications of bacillus Calmette-Guerin intravesical therapy in superficial bladder cancer J Urol 147: 596600.
  • Metchock, B.G., Nolte, F.S., Wallace, Jr, R.J. (1999) Mycobacterium. In Manual of Clinical Microbiology, 7th edn. Murray, P.R. (ed.). Washington, DC: American Society for Microbiology Press, pp. 16651685.
  • Mortlock, R.P. (1998) Bacterial growth and metabolism. In Systematic Bacteriology, Vol. 2, 9th edn. Balows, A., and Duerden, B.I. (eds). London: Arnold, pp. 85124.
  • Nakano, M.M. & Zuber, P. (1998) Anaerobic growth of a ‘strict aerobe’ (Bacillus subtilis). Annu Rev Microbiol 52: 165190.
  • Newport, M.J., Huxley, C.M., Huston, S., Hawrylowicz, C.M., Oostra, B.A., Williamson, R., Levin, M. (1996) A mutation in the interferon-γ-receptor gene and susceptibility to mycobacterial infection N Engl J Med 335: 19411949.
  • North, R.J. & Izzo, A.A. (1993) Mycobacterial virulence. Virulent strains of Mycobacteria tuberculosis have faster in vivo doubling times and are better equipped to resist growth-inhibiting functions of macrophages in the presence and absence of specific immunity J Exp Med 177: 17231733.
  • Nseyo, U.O. & Lamm, D.L. (1997) Immunotherapy of bladder cancer Semin Surg Oncol 13: 342349.
  • Pavelka, M.S. & Jacobs, Jr, W.R. (1999) Comparison of the construction of unmarked deletion mutations in Mycobacterium smegmatis, Mycobacterium bovis Bacillus Calmette-Guerin and Mycobacterium tuberculosis H37Rv by allelic exchange J Bacteriol 181: 47804789.
  • Pelicic, V., Reyrat, J.-M., Gicquel, B. (1996) Generation of unmarked directed mutations in mycobacteria, using sucrose counter-selectable suicide vectors Mol Microbiol 20: 919925.
  • Ratledge, C. (1982) Nutrition, growth and metabolism. In The Biology of the Mycobacteria, Vol. 1. Ratledge, C., and Standford, J. (eds). London: Academic Press, pp. 185271.
  • Raviglione, M.C., Snider, D.E., Kochi, A. (1995) Global epidemiology of tuberculosis: morbidity and mortality of a worldwide epidemic. J Am Med Assoc 273: 220226.
  • Von Reyn, C.F., Clements, C.J., Mann, J.M. (1987) Human immunodeficiency virus infection and routine childhood immunisation Lancet 2: 669672.
  • Rosenfeldt, V., Paerregaard, A., Valerius, N.H. (1997) Disseminated infection with Bacillus Callmette-Guerin in a child with advanced HIV disease Scand J Infect Dis 29: 526527.
  • Snapper, S.B., Melton, R.E., Mustafa, S., Kieser, T., Jacobs, Jr, W.R. (1990) Isolation and characterization of efficient plasmid transformation mutants of Mycobacterium smegmatis Mol Microbiol 4: 19111919.
  • Talbot, E.A., Perkins, M.D., Silva, S.F.M., Frotingham, R. (1997) Disseminated Bacille Calmette-Guerin disease after vaccination: case report and review. Clin Infect Dis 24: 11391146.
  • Virtanen, S. (1960) A study of nitrate reduction by mycobacteria. Acta Tuberc Scand 47 (Suppl.): 1116.
  • Weltman, A.C. & Rose, D.N. (1993) The safety of Bacille Calmette-Guerin vaccination in HIV infection and AIDS. AIDS 7: 149157.
  • Wheeler, P.R. & Ratledge, C. (1994) Metabolism of Mycobacterium tuberculosis. In Tuberculosis: Pathogenesis, Protection, and Control. Bloom, B.R. (ed.). Washington, DC: American Society for Microbiology Press.