Extracellular and Cytosolic Iron Superoxide Dismutase from Mycobacterium bovis BCG

Abstract

Two forms of iron superoxide dismutase (SOD) were purified from cell extract (CE) and culture filtrate (CF) of Mycobacterium bovis BCG, respectively. The molecular weight of both enzymes was estimated to be approximately 84,000 by gel filtration, whereas that of their subunits was 21,500, as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, indicating that each of purified enzymes is composed of four identical subunits. The specific activities of CE SOD and CF SOD were 3,850 and 4,040, respectively. The purified enzymes were not joined by disulfide bonds and were, to some extent, resistant to sodium dodecyl sulfate. Their activities were lost by H₂O₂, but not by azide and cyanide, indicating iron SODs. Enzyme activities were detectable over a broad range of pHs, from 5.0 to 9.0, and were stable for 6 months at −20°C. Each value of pI was 4.5. In Western blots, both enzymes reacted with sera of tuberculosis patients, but not with normal sera. The N-terminal amino acid sequences of CE SOD and CF SOD were the same, suggesting that there is no N-terminal signal sequence.

Superoxide dismutases (SODs), which catalytically scavenge the superoxide radical (O₂⁻) to hydrogen peroxide and molecular oxygen, serve a protective role against oxygen toxicity in all aerobic organisms (13, 37). Because the superoxide radical is a normal product of the univalent reduction of molecular oxygen, SODs are thought to be the primary defense against its potential cytotoxicity (13). There are three common forms of SODs, which differ in the metal ion cofactor at the active site. Manganese-containing SODs (MnSODs) are found in bacteria and mitochondria, while iron-containing SODs (FeSODs) are mainly found in the cytosol of prokaryotes, in primitive eukaryotes, and in some green plants (4, 5). In contrast, copper-zinc SODs (Cu-ZnSODs) are mostly found in the cytosol of eukaryotes, surprisingly in an increasing although limited number of bacteria (10, 18).

Until now, numerous papers have reported the distribution, characterization, and biological significance of SODs in mycobacteria (17, 19, 23). However, only a few bacteria have been shown to secrete SOD during growth. It was shown previously that Mycobacterium tuberculosis, a human pathogen, secreted iron SOD, while the nonpathogenic species, Mycobacterium smegmatis and Mycobacterium phlei, did not (19). Similarly, it was shown previously that the virulent strain Nocardia asteroides GUH-2 secretes the enzyme into the growth medium (7). SODs could also be involved in the pathogenicity of Mycobacterium leprae (33), although the secretion of the enzyme has not been demonstrated and remains controversial.

Members of the mycobacteria have emerged as major opportunistic pathogens in humans, with the advent of the AIDS epidemic (16). Recently, such a case was also reported in a human immunodeficiency virus-infected individual who displayed reactivation of Mycobacterium bovis BCG 30 years after vaccination (3), although M. bovis BCG has been used in many countries for vaccination to prevent tuberculosis. BCG vaccination may cause disseminated mycobacterial infection, an illness caused by the vaccine itself in patients suffering from severe immune deficiency.

In this study, we explored the production and secretion of SOD by M. bovis BCG by investigating the purification and some properties of M. bovis BCG SOD along with evidence of the association of this enzyme with the cytosol and the growth medium of the organism.

MATERIALS AND METHODS

Bacterial strain and culture condition.

The M. bovis BCG strain used in this study was Pasteur strain 1173P2 and was obtained from the Korea Institute of Tuberculosis. The bacterium was cultured at 37°C in Sauton’s medium (30) without Triton.

Preparation and fractionation of CEs and CFs.

Cells were collected on Whatman filter paper, washed three times with phosphate-buffered saline, and suspended with sterilized distilled water. After homogenization, cells were disrupted with a French press (Aminco, Rochester, N.Y.) at a pressure of 18,000 lb/in² and centrifuged at 15,000 rpm for 30 min at 4°C. Culture filtrates (CFs) were filtered through a 0.2-μm-pore-diameter membrane (Gelman, Ann Arbor, Mich.) and concentrated by an Amicon concentrator with a YM-10 membrane (Amicon, Beverly, Mass.). Each of the cell extracts (CEs) and CFs was precipitated by 70 to 90% ammonium sulfate, suspended with D.W., lyophilized, and stored at −20°C.

Enzyme assay and protein determination.

The activity of the enzyme was assayed at 37°C for 30 min. The assay mixtures (0.5 ml) contained 50 μl of 0.5 M potassium phosphate (pH 7.5), 25 μl of 16% Triton X-100, 2.5 μl of 10 mM EDTA, 75 μl of 1.2 mM NTC, 2.5 μl of xanthine oxidase (1.0 U), the sample, 25 μl of 4 mM hypoxanthine, and distilled water. The A₅₄₀ was monitored (Shimadzu UV-240) after addition of 0.5 ml of a solution containing 1 M formate buffer (pH 3.5), 10% Triton X-100, and 40% formaldehyde. An enzymatic unit was defined as the amount of the enzyme required to cause a 50% inhibition in the rate of reduction of NTC under the assay conditions (27). The standard protein used in unit determination was SOD from bovine kidney (Sigma, St. Louis, Mo.).

The protein concentration was determined by the Lowry method (21) with bovine serum albumin as a standard.

Purification of CE SOD and CF SOD.

Crude CE and CF extracts were dialyzed against 50 mM Tris-HCl (pH 8.0) containing 0.5 M ammonium sulfate, briefly centrifuged, and chromatographed on phenyl Sepharose 4B (1.6 by 15 cm [Sigma]) at room temperature in the same buffer. The gradient procedure was performed with D.W. at a flow rate of 30 ml/h. SOD fractions were dialyzed against D.W. (pH 7.0) and lyophilized (Labconco, Kansas City, Mo.). The lyophilized material was dissolved with 50 mM Tris-HCl (pH 7.5) containing 0.15 M NaCl prior to chromatography at a flow rate of 20 ml/h on a Sephacryl S-200 (Pharmacia, Uppsala, Sweden) gel filtration (1.6 by 85 cm) system calibrated with standard proteins (Pharmacia) at 4°C.

SDS-PAGE and native electrophoresis.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed as described previously (20) on a minigel comprising a 12.5% acrylamide continuous resolving gel, and native electrophoresis was carried out by the method of Bollag and Edelstein (11).

Enzyme activity staining following native electrophoresis.

SOD activity was visualized on a polyacrylamide gel as described previously (9). The SOD-active area appeared as a clear zone on a blue-violet background.

Inhibition test.

To test inhibition, each SOD (1 μg) was preincubated with 3 mM cyanide, 20 mM azide, and 1 mM H₂O₂ for 30 min and electrophoresed on 12.5% polyacrylamide gel prior to enzyme activity staining.

Enzyme stability and isoelectric focusing.

After 12.5% polyacrylamide native gel electrophoresis, each lane was soaked for 12 h at pH 4.5 to 5.0 (sodium acetate buffer), 6.0 to 7.5 (sodium phosphate), 8.0 to 8.5 (Tris-HCl), and 9.0 (glycine-NaOH), and enzyme activity staining was carried out before the results were read at 550 nm in a densitometer (LKB, Bromma, Sweden). Isoelectric focusing was carried out automatically with Phastsystem (Pharmacia, Sweden).

N-terminal sequencing of the purified SOD.

Approximately 20 μg of the purified enzyme was applied to an SDS-PAGE gel. After electrophoresis, the protein was electrophoretically transferred to a polyvinylidene difluoride membrane. The area containing the band was cut out and subjected to N-terminal amino acid sequence analysis with a protein sequencer (Milligen 6600B).

Western blot analysis.

Western blotting was carried out as described previously (34). The sera used were from tuberculosis patients and were obtained from the Korea Institute of Tuberculosis.

RESULTS

Purification and properties of CE and CF SODs.

The SODs were effectively purified by 70 to 90% ammonium sulfate precipitation and underwent hydrophobic interaction chromatography with phenyl Sepharose CL-6B and Sephacryl S-200 gel filtration. The results showed the SDS-PAGE pattern of purified CE SOD (20 μg) and CF SOD (20 μg) (Fig. 1a). A single band with a molecular weight of about 21,500 was stained by Coomassie brilliant blue. The specific activity of CE-SOD was 3,850 U/mg of protein, representing a purification of 30.1-fold with 53.5% recovery. CF SOD was purified approximately 28.57-fold with a specific activity of 4,040 U/mg of protein and 29.1% recovery (Tables 1 and 2). The activities of purified enzymes were stable at −20°C for 6 months (data not shown). The molecular weights of the native enzymes, as determined by gel filtration on Sephacryl S-200 (Fig. 2b) and native electrophoresis (Fig. 1b), were about 84,000. These results indicated that each of the purified SODs is a tetramer composed of identical polypeptides with a molecular weight of about 21,500.

FIG. 1.

SDS-PAGE and native electrophoresis analysis of purified enzymes. (a) SDS-PAGE (12.5% polyacrylamide) pattern. Lanes: M, standard marker proteins; 1, purified CE SOD (20 μg); 2, purified CFSOD (20 μg). Molecular weight (shown in parentheses) standard markers included the following proteins (Pharmacia): phosphorylase b (94,000), bovine serum albumin (67,000), ovalbumin (43,000), carbonic anhydrase (30,000), trypsin inhibitor (20,100), and α-lactalbumin (14,400). (b) Native (12.5% polyacrylamide) electrophoresis pattern. Lanes: M, standard marker proteins; 2, purified CE SOD; 4, purified CF SOD; 1 and 3, no protein. Standard marker proteins (Pharmacia, Sweden) were aldorase (158,000) and bovine serum albumin (66,000).

TABLE 1.

Purification of SOD from CEs of M. bovis BCG

Purification step	Total protein (mg)	Total activity (U)a	Sp act (U/mg)	Fold purification	% Recovery
Crude extractb	75.31	9,635.6	127.9	1.00	100.0
HICc	2.17	6,766.9	3,118.4	24.38	70.0
Gel filtrationd	1.34	5,159.0	3,850.0	30.10	53.5

^aA unit is defined as the amount of SOD needed to cause a 50% inhibition in the rate of oxidation of neotetrazolium chloride at 37°C. 

^bSeventy to 90% ammonium sulfate precipitation. 

^cSepharose CL-6B hydrophobic interaction chromatography. 

^dSephacryl S-200 superfine molecular sieve chromatography. 

TABLE 2.

Purification of SOD from CFs of M. bovis BCG

Purification step	Total protein (mg)	Total activity (U)a	Sp act (U/mg)	Fold purification	% Recovery
Crude extractb	10.08	1,527.1	141.4	1.00	100.0
HICc	0.21	933.3	3,794.0	26.83	60.1
Gel filtrationd	0.11	444.4	4,040.0	28.57	29.1

^aA unit is defined as the amount of SOD needed to cause a 50% inhibition in the rate of oxidation of neotetrazolium chloride at 37°C. 

^bSeventy to 90% ammonium sulfate precipitation. 

^cSepharose CL-6B hydrophobic interaction chromatography. 

^dSephacryl S-200 superfine molecular sieve chromatography. 

FIG. 2.

Molecular weight determination of purified SODs. (a) Molecular weight determination by SDS-PAGE (12.5% polyacrylamide). The marker proteins were the same as in Fig. 1a. (b) Molecular weight determination by Sephacryl S-200 molecular sieve chromatography. The molecular weight (shown in parentheses) marker proteins (Pharmacia) were as follows: aldorase (158,000), bovine serum albumin (67,000), ovalbumin (43,000), chymotrypsinogen A (25,000), and RNase A (13,700). Arrows indicated the purified SODs.

The stability of both enzymes was detectable over a broad range of pHs, 5.0 to 9.0, but CF SOD was a little more stable than CE SOD at pH 4.0 (Fig. 3). For unknown reasons, the specific activity and stability of CF SOD are higher than those of CE SOD.

FIG. 3.

Effect of stability of pH on the activity of purified enzymes. After native gel (12.5% polyacrylamide) electrophoresis, sliced gels were soaked in various pH buffers (pH 4.0 to 9.0) at 37°C for 12 h, incubated in staining solution containing 0.25 mM nitro blue tetrazolium for 30 min at 37°C, and read at 550 nm in a densitometer (LKB). Maximal activity was determined as 100%. The buffers used were 0.1 M sodium acetate (pH 4.0 to 5.0), 0.1 M sodium phosphate (pH 6.0 to 7.5), 0.1 M Tris-HCl (pH 8.0 to 8.5), and 0.1 M glycine-NaOH (pH 9.0). ■, CE; ○, CF.

The pI value of both enzymes was 4.5, and those areas of activity were visualized in the same bands by zymography (Fig. 4).

FIG. 4.

Isoelectric focusing analysis of purified enzymes. (a) Isoelectric focusing in 3-9 IEF premade gel (Pharmacia). (b) Activity staining. Lanes: M, standard markers; 1, purified CE SOD; 2, purified CF SOD. Isoelectric point (shown in parentheses) markers included the following proteins (Pharmacia): trypsinogen (9.30), lentil lectin basic band (8.65), lentil lectin middle band (8.45), lentil lectin acidic band (8.15), horse myoglobin basic band (7.35), horse myoglobin acidic band (6.85), human carbonic anhydrase B (6.55), bovine carbonic anhydrase B (5.85), β-lactoglobin (5.20), soybean trypsin inhibitor (4.55), and amyloglucosidase (3.50). Note that not all marker positions are labeled on the figure.

Subunit molecular weight.

The enzymes preincubated with sample buffer for 2 h at 37°C in the presence and absence of 2-mercaptoethanol in sample buffer were subjected to SDS-PAGE (4, 37). Comparison of their mobility to that of molecular weight of the standards yielded a subunit with a molecular mass of 21,500 Da in both the absence and presence (Fig. 5) of mercaptoethanol, indicating that the enzymes are not affected by mercaptoethanol. Even though a 66,000-Da protein band was also detected on the electrophoresis gel, it was suggested that the native enzyme structures are resistant to SDS because of their partial disruption by SDS treatment. From this result, it is clear that the enzymes are composed of four subunits of equal size and that these subunits are not joined by interchain disulfide bonds.

FIG. 5.

Determination of subunit molecular weight of purified enzymes by 12.5% acrylamide gel electrophoresis. Lanes: M, marker proteins; 1 and 2, CE SOD; 3 and 4, CF SOD. Lanes 1 and 3 show results in the absence of mercaptoethanol, and lanes 2 and 4 show results in the presence of it. Each of the samples was incubated at 37°C for 2 h. The marker proteins used were the same as in Fig. 1a.

Effect of inhibition.

In Fig. 6, both of the purified CE and CF SODs were inhibited completely by 1 mM H₂O₂ (CE and CF in lane 4), but not by 10 mM azide (CE and CF in lane 3) and 3 mM cyanide (CE and CF in lane 2), as shown by zymography. This indicated that the purified enzymes are iron-SOD (4, 35) and that CF SOD may be a secreted form of CE SOD. To confirm this hypothesis, when a 3-day culture filtrate was precipitated by 85% ammonium sulfate and dialyzed with D.W., a protein band with the same molecular weight of the purified CF SOD appeared by activity-staining electrophoresis (data not shown).

FIG. 6.

Effects of inhibitors on purified enzymes. One microgram of the enzyme was preincubated with 3 mM cyanide (lane 2), 20 mM azide (lane 3), and 1 mM H₂O₂ (lane 4), respectively, for 30 min and electrophoresed on a 12.5% acrylamide gel prior to enzyme activity staining. Lane 1 contained the control.

N-terminal amino acid sequence comparison.

When purified enzymes were subjected to Edman degradation and determination of N-terminal amino acid sequences, the N-terminal amino acid sequence of CE SOD was AXYTLPDLDX and that of CF SOD was AEYTLPDLDXDYGAL. Ten of the N-terminal amino acid sequences in CE SOD and 15 of those in CF SOD in this study showed a very high degree of similarity to those of other mycobacterial species and N. asteroides and low homology with those of other bacteria (Table 3). It is interesting that the N-terminal sequence of CE SOD was identical to that of CF SOD, implying that there are no N-terminal signal peptides responsible for secretion. Although an unidentified sequence existed, it was predicted that the first X of CE SOD is glutamic acid (E), as judged from Edman degradation data.

TABLE 3.

N-terminal sequence comparison of M. bovis BCG SOD with SODs from other bacteria characterized in the NCBI GenBank database

SOD type	Sequencea	Source or reference
M. bovis BCG CE	AEYTLPDLDX	This study
M. bovis BCG CF	AEYTLPDLDXDYGAL	This study
M. tuberculosis (Fe)	AEYTLPDLDWDYGAL	Zhang et al. (38)
M. fortuitum	AEYTLPDLDYDYGAL	Menendez et al. (25)
M. avium TMC 724	AEYTLPDLDWDYAAL	Escuyer et al. (12)
M. leprae (Mn)	AEYTLPDLDWDYAAL	Thangaraj et al. (33)
N. asteroides	AEYTLPDLDYDYSAL	Alcendor et al. (1)
E. coli	SYTLPSLPYAYDAL	Takeda and Avila (32)
P. aeruginosa	A---FELPPLPYEKNAL	Hassett et al. (14)
B. fragilis (GenBank accession no. M96560)	TYEMPKLPYANNAL	Unpublished observations (1992)

^aX, undetected amino acid sequence. Boldface indicates identical amino acids. 

Western blot analysis.

To examine whether the purified enzymes are working as antigenic molecules in tuberculosis patients, Western analysis was carried out with sera from the patients. Both CE SOD (Fig. 7, CE, lanes 1 to 8) and CF SOD (Fig. 7, CF, lanes 1 to 4) were reacted with sera of tuberculosis patients, but were not reacted with normal sera (Fig. 7, CE, lanes 9 to 12, and CF, lanes 5 to 8). Even though only one result is presented, the Western blotting data strongly indicated that both of the purified SODs have the same antigenic determinant, as shown by antigenicity against the sera of tuberculosis patients.

FIG. 7.

Western blot analysis of purified enzymes. SDS-PAGE (12.5% polyacrylamide) and electrophoretic transfer of purified SOD onto nitrocellulose membranes were performed as described in Materials and Methods. Arrows indicate the purified SOD bands of CE (a) and CF (b). Lanes 1 to 8 in panel a (CE) and 1 to 4 in panel b (CF) show the reaction with sera from patients infected with M. tuberculosis; lanes 9 to 12 in panel a (CE) and 5 to 8 in panel b (CF) show results with normal sera from healthy individuals.

DISCUSSION

Iron-SOD has been purified from CE and CF of M. bovis BCG grown in Sauton’s medium. Purified enzymes showed similarities with respect to molecular weight, pH profile, subunit structure, substrate specificity, sensitivity to inhibitors, and antigenicity to tuberculosis patients. The molecular weight and tetrameric form of the enzymes (Fig. 1) were similar to those of SODs of M. tuberculosis (19), Mycobacterium avium (12), and N. asteroides (7). However, there is a slight difference from those of M. leprae, M. smegmatis, Mycobacterium lepraemurium, Mycobacterium intracellulare, and Mycobacterium duvalii, even with BCG (22). Since adoption of a tetrameric form is thought to confer the stability of the enzyme in cellular and extracellular fluids (15), the purified SODs showed a broad pH stability (Fig. 3). These results were similar to those with many other SODs (24, 28, 31).

CE SOD and CF SOD shared strong N-terminal amino acid sequence homologies with SODs from mycobacteria, such as M. tuberculosis (93%), Mycobacterium fortuitum (93%), M. avium (87%), M. leprae (87%), and N. asteroides (87%), but low homologies with those from other bacteria, such as Escherichia coli (60%), Pseudomonas aeruginosa (40%), and Bacteroides fragilis (33%) (Table 3). The correspondence in N-terminal amino acid sequence between CE SOD and CF SOD indicated that the SOD of M. bovis BCG is not preceded by signal peptides. Thus, it was suggested that the structural gene of the enzyme from M. tuberculosis is not preceded by a signal peptide sequence (38). Although M. bovis BCG iron SOD appeared to be cytosolic as well as a secreted protein, it was not the product of autolysis, because it was found in CF within 3 days of growth under identical culture conditions. SODs from CE and CF showed identical molecular weights, pIs, metal cofactors, N-terminal amino acid sequences, and Western blots, suggesting that CF SOD should be due to direct secretion from CE SOD. How SOD is exported in mycobacteria remains unknown, since there are no possible signal peptides (12, 38). Probably, some specific systems for protein exportation exist in mycobacteria. Further investigation is in progress to explain the secretion mechanism of the enzyme in the outer cell surface of the bacterium and at the gene level.

It has been reported that superoxide enhances formation of hydroxyl radical (OH·), a highly reactive molecule that will react with various biomolecules, including lipids, proteins, and DNA, both by reducing Fe³⁺ to Fe²⁺ and by serving as a source of H₂O₂ (26). Most bacteria contain SODs and catalase as means of eliminating superoxide and H₂O₂, respectively. Pathogenic microorganisms are exposed to exogenous superoxide and H₂O₂ generated by host neutrophils and other phagocytes (26). The secreted SOD has been documented to occur in a few virulent mycobacteria (2, 7, 12), showing that it could be involved in pathogenicity (6, 12, 19). The secretion of SOD in virulent forms of N. asteroides, as well as its association with the outer cell envelope, could provide protection against killing by superoxide radicals (6), which are produced during active phagocytosis. More direct evidence for such a protective role had been obtained previously via administration of a monoclonal antibody specific for the SOD of N. asteroides (8). The fact that exogenously added SOD has protected bacteria against phagocytic attack (36) also illustrates the importance of SOD. Therefore, the SOD secreted by mycobacteria is more important, because it could function in the bacteria as a first line of defense against oxygen-mediated killing.

It has been also reported that the SOD in the purified protein derivative of M. tuberculosis used for the skin test is a general antigen (29). In this study (Fig. 7), both M. bovis BCG CE SOD and CF SOD obviously showed antigenicity against sera from patients with tuberculosis and had the same antigenic determinant. This result showed that this SOD could be used as a marker protein for diagnosis of tuberculosis. However, the specificity and sensitivity need to be determined with a larger number of sera from patients with different stages of tuberculosis and other mycobacterial infections and from BCG-vaccinated or nonvaccinated individuals.