INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Mycobacterial infections are the most serious infectious diseases in terms of human fatalities. The success of these pathogens depends mainly on their ability to survive ingestion by host macrophages (6, 12). In this environment, iron plays a key role in the balance between mycobacterial survival and host defense strategies, since it serves not only as an essential bacterial nutrient but also as a cofactor for production of antibacterial oxidants such as OH radicals (30, 34). Mycobacteria synthesize two kinds of high-affinity iron chelators (siderophores), the secreted exochelins and the cell wall-associated mycobactin (16, 46). However, the role of these siderophores in in vivo acquisition of iron remains questionable (26). In Mycobacterium paratuberculosis, an extremely slowly growing (primary cultivation takes at least 2 months) mycobactin-dependent mycobacterial species, no siderophores have been detected to date. In addition, extracellular superoxide dismutase (SOD) activity, a major factor in prevention of the formation of intracellular antibacterial oxidants (34, 46), has not been found in this organism.

M. paratuberculosis is the causative agent of paratuberculosis (Johne's disease), a chronic granulomatous nontreatable enteritis of ruminants occurring worldwide with increasing frequency (9, 23, 42). The disease causes significant economic losses, particularly in the dairy industry, as a result of reduced milk production, higher prevalence of mastitis, and reduction in weight gain (3, 23). The organism has also been detected in intestinal tissues of human patients with Crohn's disease, a chronic enteritis of unknown etiology with pathological and clinical similarities to Johne's disease (7, 25). Therefore, a potential zoonotic relevance of this pathogen is being discussed. To prevent further spread of Johne's disease, the Swedish government has taken drastic measures such as culling of all M. paratuberculosis-infected cattle herds (44). In other countries, eradication and certification programs are currently under discussion (22, 38). This indicates that the relevance of paratuberculosis as a problem in animal husbandry and public health has to be reconsidered.

The mechanisms involved in the pathogenesis of Johne's disease are poorly understood. It has been proposed that M. paratuberculosis invades the small intestine through the M cells of the dome epithelium and then enters resident macrophages and monocytes recruited from the blood (35). M. paratuberculosis, like other mycobacteria, is a facultatively intracellular pathogen which is able to survive in monocytes and macrophages. However, very little is known about the molecular mechanisms involved in entry and intracellular survival of this organism.

The increasing importance of M. paratuberculosis as an intracellular pathogen and the lack of siderophores in this species, a feature which distinguishes it from all other mycobacteria, prompted us to use M. paratuberculosis to investigate novel mycobacterial mechanisms to compete for iron within the host macrophage. Here we report the identification, purification, and characterization of a novel extracellular ferric reductase which we could specifically detect not only in in vitro-grown M. paratuberculosis but also in naturally infected bovine tissue.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains and growth conditions. A previously described (21) field isolate of M. paratuberculosis, designated strain 6783, was used throughout this study. The classification as M. paratuberculosis was confirmed by mycobactin-dependent growth and a positive PCR with the specific insertion element IS900 (17, 21). Mycobacteria were initially cultured on Herrold's egg yolk medium and subcultured in Watson-Reid broth, both supplemented with mycobactin (2 mg/liter; Rhône Merieux, Laupheim, Germany). After cultivation for approximately 3 months at 37°C, the bacteria were harvested by centrifugation (4,000 × g for 15 min). The resulting supernatant was sterile filtered through a 0.22-µm-cutoff filter membrane (Millipore GmbH, Eschborn, Germany) and either used directly or kept in lyophilized aliquots at 20°C.

Immunization and preparation of the immunoglobulin fraction. Antireductase antiserum was raised in rabbits by intracutaneous injections of 80 µg of purified reductase protein in saline mixed with 30% adjuvant (Emulsigen; MVP Laboratories, Ralston, Neb.) after collection of prebleeding blood. Production of antibodies was detected by Western blotting using crude M. paratuberculosis culture supernatants as the antigen. Immunoglobulins were purified by affinity chromatography using protein A-Sepharose (Sigma Chemie GmbH, Munich, Germany).

Electrophoresis and Western immunoblotting. Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) was performed by the method of Laemmli (24), and gels were silver stained as described by Heukeshoven and Dernick (19). Mycobacterial whole-cell lysates were prepared by mechanical treatment for 190 s with circonium beads in a Mini-bead Beater (Bio-Spec Products, Inc., Bartlesville, Okla.). Culture supernatants were prepared as described above and concentrated by trichloroacetic acid (10%, final concentration) precipitation. Protein concentration was determined by using a microassay (Micro BCA [bicinchoninic acid] protein assay; Pierce, Rockford, Ill.). Samples were prepared by boiling for 5 min in reducing sample buffer containing 62.5 mM Tris-HCl (pH 6.8), 2% (wt/vol) SDS, 5% (vol/vol) 2-mercaptoethanol, and 10% (vol/vol) glycerol, separated by SDS-PAGE, and electroblotted onto nitrocellulose membranes (43). Nonspecific binding was blocked by incubation in Tris-buffered saline-Tween (15 mM Tris, 150 mM NaCl, 0.5% Tween 80 [pH 8.0]) containing 0.5% gelatin. Serum and alkaline phosphatase conjugate were added in Tris-buffered saline-Tween and incubated for 1 h each at room temperature. Blots were developed with 5-bromo-4-chloro-3-indolylphosphate (50 µg/ml; Sigma) and nitroblue tetrazolium (100 µg/ml; Sigma) in substrate buffer (100 mM Tris-HCl, 100 mM NaCl, 5 mM MgCl₂ [pH 9.5]).

Determination of ferric reductase activity. Production of Fe²⁺ was quantified as described by Georgatsou and Alexandraki (15), using the chromogen batho-phenanthroline disulfonate (BPDS) as a ferrous iron chelator. One milliliter of culture supernatant was mixed with 10 µl of BPDS (100 mM), 25 µl of -NADH (10 mg/ml; Sigma), 10 µl of MgCl₂ (1 M), and 50 µl of Tris-HCl, pH 7.5 (1 M). To start the reaction, various amounts of ferric ammonium citrate (Sigma) were added. The reactions were incubated in the dark, and the optical density at 535 nm was determined at various time points. The amount of Fe²⁺ produced was determined based on a reference curve with standard deviations ranging from 3.55 to 11.11% of the means, depending on the concentration of the substrate. The total amount of iron in the reaction mixture was determined by addition of 10 mM (final concentration) dithiothreitol (Sigma) to the supernatant. Uninoculated medium served as a negative control.

Proteinase K sensitivity and purification of the ferric reductase. The protein character of the ferric reductase was investigated by addition of proteinase K (10 µg/ml, final concentration) to culture supernatants from M. paratuberculosis 6783. After 1 h at 37°C, the mixtures were tested for reductase activity as described above. Purification of the ferric reductase from culture supernatants was achieved by reverse-phase high-pressure liquid chromatography (HPLC; Lichrospher RP100 C18 column; Merck, Darmstadt, Germany) using a continuous linear acetonitrile gradient from 0 to 70%. Fractions were tested for reductase activity and stored at 20°C. Reductase-positive fractions were tested for purity by SDS-PAGE and subsequent silver staining. Samples loaded consisted of either 100 µg of crude culture supernatant or 1 µg of purified protein to compare similar activities on the gel. In addition, the purified reductase was applied to fast protein liquid chromatography (FPLC) using a gel filtration column (Superdex Peptide HR 10/30; Pharmacia, Freiburg, Germany) to confirm purification.

Characterization of the ferric reductase. The amino acid composition of the purified enzyme was analyzed by mass spectroscopy using an automatic analyzer (AminoAcid Analyzer Alpha Plus; Pharmacia, Uppsala, Sweden) after hydrolysis of the sample with 6 N HCl for 24 h at 110°C. The isoelectric point was determined using the Rotofor purification system (Bio-Rad, Munich, Germany). Briefly, the reductase was mixed with Biolyte 3/10 ampholyte, and focusing was done for 6 h at 12 W, using an anion-exchange membrane equilibrated with 0.1 M NaOH and a cation-exchange membrane equilibrated with 0.1 M H₃PO₄. Each fraction was tested in parallel for the pH and for reductase activity. The effects of temperature and pH on activity were determined using incubation temperatures of 20, 37, 60, and 100°C and pH values ranging from 5.0 to 10.0. For the saturation kinetics, the reductase activity was determined with 2.3 µg of purified reductase and increasing concentrations of ferric ammonium citrate at 15 min. This was in the linear range of ferric iron reduction activity of the enzyme as determined in preliminary experiments (data not shown). Substrate specificity was determined by using equine ferritin and porcine and human transferrin (all from Sigma) as substrates.

Immunoelectron microscopy. Postembedding immunoelectron microscopic localization of the ferric reductase of M. paratuberculosis was done essentially as described previously (41). Tissue samples were taken from the ileal Peyer's patch region of an M. paratuberculosis-infected cow with clinical symptoms. Histopathology had revealed large numbers of intracellular acid-fast bacteria in the mucosa, including those regions from which the sample had been taken. As a negative control, similar samples were taken from a healthy cow. After fixation in glutaraldehyde and osmium tetroxide, samples were dehydrated and low-temperature embedded in Lowicryl K4M (Plano, Wetzlar, Germany). Cells of M. paratuberculosis 6783 were treated similarly except that embedding was done in Lowicryl HM20 (Plano). Samples were incubated either with specific antireductase immunoglobulin G (IgG) antibodies or preimmune IgG antibodies (negative control) for 2 h at 30°C, washed, and then incubated with protein A-gold particles (gold particle size, 10 nm).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Identification and purification of the ferric reductase. The culture supernatants of different M. paratuberculosis isolates grown for approximately 12 weeks were screened for iron reductase activities by using ferric ammonium citrate as a substrate and BPDS as a ferrous iron chelator. One isolate, designated M. paratuberculosis 6783, was then selected for purification and further characterization of the reducing compound. Crude culture supernatant of this strain contained 5.5 mg of protein/ml, which was lowered to 4.6 mg/ml by ultrafiltration. In HPLC, one fraction eluted at 55 to 60% acetonitrile was positive in the reductase assay (Fig. 1). This fraction contained 43 µg of protein/ml and retained 93.6% of the original reductase activity, resulting in an approximately 120-fold purification. This preparation appeared >95% pure and gave a single band with a relative molecular weight of ca. 17,000 when separated on an SDS-polyacrylamide gel (Fig. 1, inset). Further purification of the HPLC preparation by FPLC did not enhance purity. FPLC and HPLC preparations gave bands of similar sizes in SDS-PAGE and reacted similarly in Western immunoblot assays using the polyclonal antireductase antibody raised in rabbits (data not shown). Analysis of the purified protein by mass spectroscopy revealed glycine, serine, asparagine or aspartic acid, and glutamine or glutamic acid as the major components (Fig. 2).

View larger version (32K): [in this window] [in a new window]   FIG. 1.   Purification of the ferric reductase of M. paratuberculosis 6783 by reverse-phase HPLC. A single fraction, which eluted at 55 to 60% acetonitrile, was positive in the reductase assay using ferric ammonium citrate as a substrate and BPDS as a ferrous iron chelator. The amount of ferrous iron produced was determined spectroscopically and is expressed as enzymatic activity (dotted line). This fraction was separated by SDS-PAGE and gave a single band after silver staining of the gel (inset [lane 1, crude supernatant; lane 2, purified reductase; MW, molecular weight in thousands]). For comparison with crude supernatant, equal amounts of enzyme activities were loaded (see Materials and Methods for details).

View larger version (29K): [in this window] [in a new window]   FIG. 2.   Amino acid composition of the purified ferric reductase of M. paratuberculosis 6783 as revealed by mass spectroscopy using an automated amino acid analyzer.

Functional characterization of the reductase. The iron-reducing activity of the purified enzyme was highest at 37°C (tested temperatures were 20, 37, 60, and 100°C) and at pH values between 5.0 and 10.0. The isoelectric point of the reductase was 9.0 as determined by isoelectric focusing. Combining a variety of potential coenzymes and divalent ions in the reductase assays revealed that enzyme activity was dependent on NADH as an electron donor and was enhanced by the divalent cation Mg²⁺ (Fig. 3). Kinetic analysis done with constant reductase and various substrate (ferric ammonium citrate) concentrations revealed a saturable process, thus indicating the formation of an enzyme-substrate complex. Plotting of the velocities of iron reduction against substrate concentrations resulted in a sigmoidal curve (Hanes diagram [Fig. 4]). From the resulting Lineweaver-Burk plot (Fig. 4, inset), a K_m of 0.213 mM was calculated. The calculated V_max was 0.345 mM min¹ mg¹; hence, 1 µM reductase could reduce 5.75 mM Fe³⁺ per min. Relating the enzymatic activity of a standard preparation from the culture supernatant to the wet weight of the respective bacterial cell pellet, we calculated that M. paratuberculosis 6783 could generate approximately 11.42 µM Fe²⁺/min/g (wet weight) from ferric ammonium citrate.

View larger version (11K): [in this window] [in a new window]   FIG. 3.   Effects of NADH (A) and divalent Mg2+ ions (B) on the reducing activity of the reductase, tested by using indicated amounts of either NADH (plus 10 µM Mg2+) or Mg2+ (plus 100 µM NADH). Results are expressed as enzymatic activity and represent means from three independent experiments.

View larger version (17K): [in this window] [in a new window]   FIG. 4.   Kinetics of the reductase activity, using ferric ammonium citrate as a substrate. The amounts of ferrous iron produced per minute by 2.3 µg of purified reductase from the indicated concentrations of ferric iron were determined after 15 min, which was in the linear range of ferric iron reduction activity of the enzyme as determined in preliminary time saturation experiments. Results are expressed as enzymatic activity and represent means from three independent experiments. The inset shows the resulting Lineweaver-Burk diagram.

In vitro and in vivo detection of the reductase by immunoelectron microscopy. We were able to localize the reductase both in in vitro-grown mycobacteria and in gut tissue of naturally infected cattle by immunogold electron microscopy using a specific rabbit IgG antibody which was raised against the purified enzyme. This antibody did not significantly inhibit enzyme activity. However, it specifically bound to a 17-kDa band upon Western immunoblotting using either crude mycobacterial culture supernatant, HPLC-purified reductase, or FPLC-purified reductase (data not shown), indicating its specificity. In vitro, most (ca. 90%) of the reductase was found extracellularly, as indicated by the relative poor immunogold staining of bacterial cells (Fig. 5C) and comparison of the reductase activity in the supernatant with that in the bacterial pellet (data not shown). In vivo, the enzyme was detected on the surface and in close proximity to the bacteria, as well as in the surrounding cytoplasm and compartments of the macrophage, indicating secretion of the enzyme (Fig. 5A and B). The number of gold particles, indicating presence of the reductase, associated with bacteria in infected tissue was higher than in in vitro-cultured M. paratuberculosis cells (Fig. 5). Almost no labeling could be found in the negative controls, which were (i) tissue samples incubated with preimmune antibodies (Fig. 5D), (ii) tissue samples taken from a healthy cow incubated with antireductase antibodies, and (iii) in vitro-cultured bacteria incubated with preimmune antibodies. This indicated that labeling was specific for the reductase.

View larger version (101K): [in this window] [in a new window]   FIG. 5.   Postembedding immunoelectron microscopic localization of the ferric reductase of M. paratuberculosis in low-temperature-embedded naturally infected bovine intestinal tissue and in vitro-grown M. paratuberculosis. Binding of specific antireductase IgG antibodies was visualized by using 10-nm protein A-gold particles. The bound gold particles (G, arrows) are indicative of the presence of ferric reductase on the surface and in close proximity of mycobacteria in naturally infected tissue (A) and on the cellular surface of M. paratuberculosis (C). In infected tissue, reductase was also detected to a certain extent in the surrounding cytoplasm and compartments of the macrophage (B). Infected bovine tissue which was incubated with preimmune antibodies showed almost no label (D). N, nucleus; bars, 0.25 µm.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Pathogenic mycobacteria have in common that they must survive within monocytes and macrophages in order to establish disease and thus have evolved mechanisms to either adapt to or modify the hostile intracellular environment. Well- documented strategies employed by these bacteria are (i) modification of the phagosome, resulting in inhibition of acidification, phagosome-lysosome fusion, and lysosomal enzyme activities, and (ii) resistance to or neutralization of the damaging effects caused by reactive oxygen intermediates (6, 8, 40, 45).

Secreted mycobacterial proteins appear to play an important role in the interaction between the host macrophage and the bacterial parasite (2). Therefore, these antigens have been thoroughly investigated in several mycobacterial species (1, 37). Two secreted enzymes, the SOD (1, 48) and the catalase-peroxidase (20, 36), have been described, and their function in protecting mycobacteria against antibacterial oxygen radicals within the macrophage is generally accepted. Another obstacle that mycobacteria must overcome during intracellular growth is the lack of accessible iron, most of which is stored intracellularly as ferritin. In a variety of bacterial pathogens including mycobacteria, iron availability and the sequestration of iron within the host have been shown to be associated with virulence (30). The common mycobacterial siderophore mycobactin has long been proposed as a major mechanism to obtain iron under limiting conditions. However, recent evidence suggests that mycobactin is not synthesized in vivo (26). Based on this evidence, another class of secreted iron-chelating proteins, the exochelins (31), have gained renewed interest (16). In addition, several iron-regulated envelope proteins have been found in M. avium and M. leprae with six proteins, among them one of 21 kDa, which are expressed in vivo only (39).

In M. paratuberculosis, neither siderophores nor extracellular SOD activities have been detected; the latter was also confirmed with the strain used in this study (data not shown). Therefore, we hypothesized that a possible way for these organisms to scavenge iron during intracellular growth and simultaneously prevent the formation of toxic OH radicals might be the reduction of Fe³⁺, a detoxifying mechanism common in free-living microorganisms (31). Based on this hypothesis, we looked for the presence of ferric reductase activity in the supernatants of M. paratuberculosis field isolates. We have identified, purified, and characterized an extracellular iron-reducing compound from M. paratuberculosis 6783. To our knowledge, this is the first purified extracellular reductase identified in bacteria. Other ferric reductases which have been characterized in both intracellular and extracellular pathogens such as M. smegmatis (33), Listeria monocytogenes (11), Neisseria gonorrhoeae (14, 27), and Escherichia coli (10) were either located in the cytoplasm or associated with the inner membrane. Another intracellular enzyme of M. paratuberculosis, a putative serine protease, was recently described by Cameron et al. (4).

Functional characterization of the M. paratuberculosis reductase revealed most promising features in terms of its potential role in vivo. First, enzyme activity was highest at 37°C and at pH values between 5.0 and 10.0, which is important considering that mycobacteria have to adapt to various pH values during infection. Second, ferritin, the major intracellular iron storage protein, was found to serve as a substrate for the reductase. Thirdly, though the K_m of 0.213 mM for ferric ammonium citrate indicates only a moderate affinity compared to those of other bacterial reductases (as cited above), the V_max of 0.345 mM min¹ mg¹ suggests that the reductase is an effective enzyme in terms of iron mobilization compared with other siderophores such as enterobactin (5). We calculated that M. paratuberculosis 6783 could generate ca. 11.42 µM Fe²⁺/min/g (wet weight) from ferric ammonium citrate. In comparison, a membrane-bound reductase of Saccharomyces cerevisiae, which was suggested to represent an alternative mechanism of iron acquisition in this species, could reduce only 20.4 nM Fe³⁺/min/g (wet weight) from the same substrate (29). Taken together, the functional data suggest that the M. paratuberculosis reductase could represent an effective mycobacterial mechanism to mobilize iron from environmental sources.

For many facultative intracellular parasites, such as Salmonella typhimurium (13), Yersinia enterocolitica (18), and M. tuberculosis (16), siderophores have been proposed as virulence factors since they facilitate bacterial growth under iron-limited conditions in the host. It is not clear, however, whether siderophores serve solely to ensure provision of iron for the bacterial parasite or whether they additionally interfere with iron-dependent bactericidal reactions of the macrophage, such as the Fe³⁺-dependent processing of H₂O₂ into highly toxic OH radicals. A first hint in this direction was recently reported by Leonard et al. (28), who showed that survival of Brucella abortus in macrophages could be facilitated by the addition of siderophores. Another recent report suggested that another redox system, the thioredoxin reductase of M. leprae, can inhibit oxygen-dependent killing of intracellular bacteria (47). These findings led us to speculate that the reductase of M. paratuberculosis might play a role in intracellular survival. To support this hypothesis, we confirmed the in vivo production of ferric reductase in infected macrophages. We localized the reductase on the surface of in vitro-cultured M. paratuberculosis and in gut tissue of naturally infected cattle by immunogold electron microscopy using a specific rabbit IgG antibody. The enzyme was not only detected on the surface and in close proximity of the bacteria but also in the surrounding cytoplasm and compartments of the macrophage, indicating that the enzyme was produced in vivo and, most probably, secreted into the surrounding area. Comparison of levels of reductase production in vivo and in vitro suggested that production was higher in vivo than in vitro, though it has to be considered that the experimental conditions used did not allow exact quantitation.

In conclusion, we report a novel extracellular ferric reductase of mycobacteria which might play a role in bacterial survival within macrophages by providing iron as an essential bacterial nutrient and simultaneously interfering with the production of antibacterial oxidants. Future studies will show whether the ferric reductase is an M. paratuberculosis-specific protein and, most importantly, whether this enzyme is also functional in vivo and thus represents a novel bacterial survival strategy in intracellular parasitism.