THE J O U R N A L OF BIOLOGICAI. CHEMISTRY Val. 258. No. 1, Issue of danuary IO, pp. 91-96, 1983 Printed in Lr S A

Purification and Properties of a Unique Superoxide Dismutase from Nocardia asteroides”

(Received for publication, July 9, 1982)

Blaine L. Beamant, Steve M. Scates, Stephen E. Moring, Richard Deem, and Hara P. Misrag From the Department of Medical Microbiology and Immunology, School of Medicine, and the §Department of Physiological Sciences, School of Veterinary Medicine, University of California, Davis, California 95616

A unique form of superoxide dismutase was isolated and characterized from Nocardia asteroides GUH-2. This enzyme contains 1 to 2 g atoms each of Fe, Mn, and Zn per mol and exhibits spectral properties sugges- tive of Fe- or Mn-containing superoxide dismutases. Its M, = 100,000, and it is composed of four subunits of equal size which are not covalently joined. The amino acid composition of the enzyme was more closely re- lated to the Mn- or Fe-containing enzymes of Mycobac- terium species and was least related to the Cu-Zn en- zyme of eukaryotes. Azide at 1 and 20 mM inhibits the activity 10 and 41%, respectively, and 5 mM HzOz in- hibits 40%, but 1 or 5 mM cyanide caused trivial effect. The immunofluorescent staining, which was specific for superoxide dismutase of N. asteroides, indicated the association of this enzyme to the outer cell wall of the organism. Further, the enzyme was shown to be selec- tively secreted into the medium.

Superoxide dismutases, which catalytically scavenge 02-, serve a protective role in all aerobic organisms against oxygen toxicity (1-3). The structural and functional relationships of superoxide dismutases have raised interesting and unresolved questions about their evolution. The copper-zinc enzymes are found in various species of animals and plants (4-13), and they are considered to be characteristic of the cytosols of eukaryotic cells (2). A similar enzyme has been found in a prokaryote, Photobacterium leiognathi (14); however, it ap- pears that this organism may have received the gene for this Cu-Zn enzyme from its ponyfkh host (15). Manganese-con- taining superoxide dismutases have been isolated from several prokaryotes (16-18) as well as from the mitochondria of chicken liver (7) and of yeast (19). Structural analyses have demonstrated a close relationship between the bacterial and the mitochondrial enzymes (20, 21), thus supporting the hy- pothesis of a symbiotic origin of mitochondria (20-22). How- ever, substantial quantities of Mn enzyme have also been found in the cytosol of baboon liver (23). The cytosol of the unicellular alga, Porphyridium cruentum, which is considered to be perhaps the most primitive eukaryote, contains a Mn enzyme (24) whereas the cyanobacteria (blue-green algae), which are considered to be among the most advanced prokar- yotes, have an iron-containing superoxide dismutase (25, 26). Iron-containing enzymes have also been found in several other

* This investigation was supported by Public Health Service Grant AI-13167 from the National Institute of Allergy and Infectious Dis- eases. The costs of publication of this article were defrayed in part hy the payment of page charges. This article must therefore he hereby marked “adr?ertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom correspondence should he addressed.

bacteria (27-29). A survey of progressively more advanced plants has failed to find copper-zinc superoxide dismutase in marine plants, whereas it has been found in land plants such as mosses and ferns (30). Thus, the facts are not easily arranged into a coherent theory of descent for the origin and distribution of the different classes of superoxide dismutase.

It has been demonstrated that Nocardia asteroides, a strictly aerobic and facultatively intracellular pathogen, can grow within macrophages, and is resistant to phagocytic at- tack (31). In addition, N . asteroides induces an oxidative metabolic burst in human polymorphonuclear neutrophils and monocytes, yet these organisms are not killed by the produc- tion of superoxide radicals (32). Since exogenously added superoxide dismutase was shown to protect some bacteria against the attack of neutrophils and macrophages (33-35), we explored whether Nocardia produces and secretes a su- peroxide dismutase that could explain this organism’s unusual resistance to the oxidative metabolic burst of polymorphonu- clear phagocytes. We here report the thorough purification and characterization of a unique superoxide dismutase from N . asteroides GUH-2 along with evidence of the association of this enzyme with the outer cell wall of the organism and its secretion into the growth medium.

MATERIALS AND METHODS

N . asteroides GUH-2 was isolated from a fatal human infection at Georgetown University Hospital, Washington, D. C., and was main- tained as previously described (36). Fresh animal isolates were grown in 50 ml of brain-heart infusion broth (Difco) in a 250-mi Erlenmeyer flask at 37 “C, and with 150 rpm agitation in a Psycrotherm Environ- mental Incubator (New Brunswick). After 1 week of incubation, 0.5 ml of the starter culture was transferred to 500 ml of the broth in 12 2800-ml Fernhach flasks and grown for 12, 16, 24,48, 72, and 168 h as previously described (37, 38). Cells from each time point were har- vested by filtration on a 0.45-pm pore size membrane filter (Nalge) and washed twice with 0.85% sterile saline.

For the isolation of superoxide dismutase from the growth medium, Middlehrook 7H9 (Difco) + 0.5% glycerol (Mallinckrodt) + 0.5% glucose were substituted for brain-heart infusion broth. Growth con- ditions were as specified above, except 5.0 ml of culture were used to inoculate the Fernhach flasks. This chemically defined medium was used because brain-heart infusion broth interfered with the assays of superoxide dismutase secreted into the medium.

Cytochrome c (type III), xanthine, and xanthine oxidase were products of Sigma. Microgranular diethylaminoethyl cellulose and Sepharose 6B were obtained from Pharmacia.

Superoxide dismutase was assayed and units were defined as pre- viously described (4). Electrophoresis was performed using a Phar- macia GE-4 electrophoresis apparatus. Analytical gels were run using 10% acrylamide with the buffering system of Davis (39). Zones of protein were localized by staining with Coomassie blue R-250 (40), while zones of superoxide dismutase were negatively stained by the photochemical procedure previously described (41). Molecular weight of superoxide dimutase from both the cytoplasm and the growth medium was determined by gel filtration chromatography on a Seph- arose 6B column (1.5 X 75 cm) equilibrated with 0.01 M Tris-HC1, pH

91

92 A Unique Superoxide Dismutase from Nocardia

7.8, and containing 0.1 mM EDTA. Subunit weight was estimated by polyacrylamide gel electrophoresis in the presence of 1% sodium dodecyl sulfate with and without 1% P-mercaptoethanol (42).

Amino acid analysis was performed on a Durrum Model D-500 amino acid analyzer using a single column, three-buffer elution pro- gram. Protein samples were hydrolyzed for 24, 48, and 72 h at 100 "C in Vacuo in the presence of a small crystal of phenol (43). Tryptophan was determined by the spectrophotometric method of Edelhoch (44). Spectrophotometric assays were performed at 25 "C in a Gilford Model 260 spectrophotometer and optical spectra were recorded with a Cary Model 219 spectrophotometer. Protein was measured by the spectrophotometric procedure of Lowry (45) with recrystallized bo- vine serum albumin as the standard.

Antisera were prepared by injecting rabbits with three booster doses of purified superoxide dismutase in polyacrylamide adjuvant or Freund's incomplete adjuvant (46). The rabbit antiserum was frac- tionated with ammonium sulfate and the precipitate was collected at 40% saturation, dissolved in water, dialyzed in 20 mM Tris-HCI, pH 7.5, and absorbed onto a column of DE52 (2.6 X 15 cm) and eluted with 0.0175 M phosphate buffer, pH 6.3. The fractions containing the antibody were pooled, concentrated on a PM-10 Amicon membrane, and then stored at -80 "C.

Immunofluorescent staining of the organisms was done by utilizing rabbit anti-superoxide dismutase and fluorescein-labeled goat anti- rabbit IgG (Miles Laboratories). Single cell suspensions of the orga- nism were prepared from log phase cultures, grown in brain-heart infusion broth or in the Middlebrook 7H9-glycerol-glucose media, by filtering through a glass wool column to remove the clumps, followed by differential centrifugation as previously described (36). The cells were washed two times with phosphate-buffered saline, pH 7.2, and incubated with the purified antibody (1:lO volume) at 37 "C for 30 min. These cells were then washed five times with the same buffer and treated with goat anti-rabbit IgG. Wet mounts prepared on glass slides were immediately observed with a Zeiss research microscope equipped with a mercury vapor epifluorescent illuminator using flu- orescein isothiocyanate fiiters and photographed using Ektachrome 400 fiim (Kodak).

Metal analysis was carried out using a Perkin-Elmer Model 306 atomic absorption spectrophotometer.

RESULTS

Purification of the Enzyme-Ten grams of washed cells of N . asteroides were suspended in 40 ml of 0.01 M Tris-HC1,O.l mM EDTA, pH 7.8, and were disrupted by agitation with 0.1- mm diameter glass beads (VMR) in a Cos-cooled Braun homogenizer (Allen-Bradley). Cell debris was removed by centrifugation at 30,000 X g for 30 min. The supernatant was brought to 65% saturation, at 23 "C, with solid ammonium sulfate. The precipitate which formed within 1 h was removed by centrifugation and the supernatant solution was brought to 85% saturation with (NH4),S04. The precipitate, collected by centrifugation, was suspended in a minimum volume of 10 mM Tris-HC1,O.l mM EDTA, pH 7.8, and was dialyzed against several changes of the same buffer at 4 "C. The dialysate was clarified by centrifugation and 100 ml of the supernatant were passed onto a column of DEAE-cellulose (25 ml of DEAE- cellulose packed under 10 p s i . of Nz in a 60-ml disposable syringe), which had been equilibrated at 4 "C with the above buffer. Batches of DEAE-cellulose were successively washed with 50 ml of the Tris-HC1 buffer and with 50 ml of 0.05 M (NH,)$O., in the Tris-HC1 buffer and were finally eluted with

100 ml of 0.2 M (NH&S04. The eluted fractions containing superoxide dismutase activity were pooled, concentrated to about 5 ml by ultrafiltration over a PM-10 membrane (Ami- con), dialyzed exhaustively against several changes of 0.01 M Tris-HC1,O.l mM EDTA, pH 7.8. These pooled fractions were purified further by gradient DEAE-cellulose chromatography. The samples were applied to a column (2 X 25 cm) and washed with 25 ml of 0.1 M Tris-HC1, 0.1 mM EDTA, pH 7.8, buffer and then with 100 ml of 0.1 M (NH4)2S04 in Tris-HC1 buffer. Next, a linear gradient of (NH4)zS04 (0.1 + 0.25 M in the Tris- HC1 buffer, 200 ml each) was then applied and 5-ml fractions were collected. Superoxide dismutase activity was eluted as a single symmetrical peak which was congruent with a peak of 280-nm absorbance. The enzyme was concentrated by ultra- filtration using a PM-10 membrane (Amicon) and was stored frozen a t -80 "C before characterization. Table I summarizes the results of this purification.

Polyacrylamide Gel Electrophoresis-Both the crude sol- uble cytoplasmic extract and the purified superoxide dismu- tase of N . asteroides were analyzed by gel electrophoresis. Protein was visualized by staining with Coomassie blue R-250, whereas superoxide dismutase activity was negatively stained by a photochemical procedure as described (41). The crude extract of N . asteroides GUH-2 exhibited several protein bands but there was only one band of cyanide-insensitive superoxide dismutase activity which had an RF value of 0.4. The purified enzyme gave only one discernible band of protein which coincided with the zone of enzymatic activity. A den- sitometric scan of a gel loaded with the purified enzyme and stained with Coomassie blue R-250 after electrophoresis is presented in Fig. 1 A .

Molecular Weight and Quaternary Structure-Molecular weight was determined by gel filtration on a Sepharose 6B column (1.5 X 75 cm). The column was equilibrated with 0.01 M Tris-HC1, 0.1 mM EDTA buffer, 0.1 M KC1, pH 7.8, and calibrated with the following molecular weight standards: aldolase (158,000), amylase (97,600), bovine serum albumin (67,000), and ovalbumin (45,000). The molecular weight of purified superoxide dismutase was found to be 100,000 by this method.

Subunit molecular weight was determined by sodium do- decyl sulfate gel electrophoresis as previously described (42). The standards used to make a plot of log molecular weight versus mobility of protein band were: aldolase (158,000), am- ylase (97,000), phosphorylase b (94,000), bovine serum albu- min (67,000), ovalbumin (45,000), carbonic anhydrase (30,000), and 2-lactalbumin (17,400). The superoxide dismutase ex- hibited mobility consistent with M , = 25,000 in the presence or absence of /3-mercaptoethanol. These results are presented in Fig. 1B and imply that the superoxide dismutase of N . asteroides GUH-2 is composed of four subunits of equal size held together by noncovalent interactions. No other bands were visible on these gels.

Amino Acid Composition-Table I1 presents the amino acid composition of N. asteroides superoxide dismutase. In comparison with the amino acid compositions reported for all

TABLE I Purification of superoxide dismutase from N . asteroides GUH-2

Step Superoxide

activitv Protein Protein dismutase Total activity 'pecific ity Yield Purifica- tion

w l m [ total mg units/rnl units units/mg B -fold Crude cytoplasm 5.75 1,017.7 130 23,010 22.6 100 1 6545% ammonium sulfate 0.72 25.9 250 9,OOO 347.2 39 15 DEAE-cellulose chromatography 0.328 8.2 324 8,100 987.0 35 44 DEAE-cellulose chromatography 0.187 6.4 250 8,550" 1,336.0 37 59

'' The apparent increased activity was due to the removal of some interfering proteins in the cytochrome c assay.

A Unique Superoxide Dismutase from Nocardia 93

A .

I BOTTOM

E.

I I I I I I TOP 94 67 45 30 20 145 BOTTOM

Molecular Weight ( X 1 0 3 )

FIG. 1. Densitometric scans. A , 10% polyacrylamide gel electro- phoresis loaded with purified superoxide dismutase and stained for protein with Coomassie blue R-250. The small peak near the bottom of gel scan is the tracking dye. B, 10% polyacrylamide-sodium dode- cylsulfate gel electrophoresis, loaded with purified superoxide dis- mutase. The small peak near 1400 is the tracking dye. The molecular weight markers were calibrated by running purified protein standards as described under "Materials and Methods."

TABLE I1 Amino acid composition of superoxide dismutase from N.

asteroides GUH-2 The data represent mean values obtained following hydrolysis at

24, 48, and 72 h and extrapolated for hydrolytic loss. Calculations are based on subunit M, = 25,000.

Amino acid Residues/25,000

Lysine 10 Histidine 9 Arginine 4 Aspartic acid 24 Threonine a Serine 7 Glutamic acid 20 Proline 7 Glycine 17 Alanine 25 Valine 10 Methionine 2 Isoleucine 6 Leucine 20 Phenylalanine 9 Tryptophan 9 Tyrosine 7

other superoxide dismutases, the Nocardia enzyme exhibits notable differences. Table I11 shows the sum of the squares of the difference ( S A Q ) between amino acids among different superoxide dismutases. The lower the number, the smaller the difference in amino acid content and hence a greater similarity (47). Based upon the values for S A Q , N. asteroides enzyme is closely related to the manganese-containing and iron-containing enzymes of Mycobacterium and is least related

to the copper-zinc superoxide dismutases of eukaryotes. Table I11 presents the S A Q values, which are measures of difference in composition, in the order of increasing phylogenetic dis- tance from N . asteroides.

Ultraviolet Spectrum-The purified enzyme at 2 mg/ml was colorless and the lyophilized enzyme was a white powder. The absorption in ultraviolet was typical of that seen with most proteins and is shown in Fig. 2. The E;':,,,,, a t 280 nm, was 24.3, which calculates to a molar extinction coefficient of 2.4 X io5 M" cm".

Metals and Inhibitors-Atomic absorption spectrophotom- etry indicated the presence of 1 to 2 g atoms each of Mn, Fe, and Zn per 100,000 daltons of this enzyme. Assays for copper, cobalt, and nickel indicated the presence of only trace amounts of these metals. These results are presented in Table IV. A relatively large standard deviation associated with the metal

TABLE 111 Comparison of S A Q values of different superoxide dismutases to N.

asteroides GUH-2 enzyme Superoxide dismutase SAQ"

Mycobacterium smegmatis, Mn (52) 11.8 Mycobacterium tuberculosis, Fe (55) 17.4 Photobacterium leiognathi, Fe (15) 26.2 Plectonema boryanum, Fe (25) 35.6 Porphyridium cruentum, Mn (24) 62.2 Photobacterium leiognathi, Cu-Zn (15) 81.4 Escherichia coli, Fe (27) 87.8 Pony fish, Cu-Zn (15) 97.7 Escherichia coli, Mn (16) 125.1 Bovine heart, Cu-Zn (6) 204.3 Neurospora crassa, Cu-Zn (5) 209.9 Drosophila melanogaster, Cu-Zn (13) 210.4 Bovine erythrocyte, Cu-Zn (4) 235.2

The parameters S A Q was defined as (46): S A Q = j ( X t , , - X k . , ) ' ,

where the subscripts i and k indicate the particular enzymes com- pared, and X, is the content of a given amino acid of type J.

Numbers in parentheses refer to the literature from which mole per cent contents of 18 amino acids were taken into account.

2'ol I .8

1.6 -

14 -

0)

2 1.2 -

n 0 1.0

a n

0

L - u)

0.8 -

240 260 280 300 320 Wavelength (Xnm)

FIG. 2. The ultraviolet absorption spectrum of the enzyme at 0.7 mg/ml in 0.01 M Tris-HC1,O.l mM EDTA, pH 7.8 at 23 "C.

94 A Unique Superoxide Dismutase from Nocardia

analysis was probably due to the different length of storage time and dialysis of the enzyme. Samples of proteins from four different batches of purification were stored 1 week to several months at -80 "C and were dialyzed against 0.05 M potassium phosphate buffer, pH 7.8, containing 1 X lo-'' M EDTA before the metal analysis. We are convinced, however,

TABLE IV Metal composition of superoxide dismutase from N . asteroides GUH-I as determined bv atomic absorntion snectronhotometrv

Metals" Cram aroms/mol enzymeh

Mn Fe Zn

1.5 f 0.6 2.0 f 0.7 1.8 f 1.0

" Only trace amounts of other metals such as nickel, copper, and

" Mean and standard deviation of four analyses from four different cobalt were detectable in some samples.

batches of the purified enzyme.

TAHLE V Inhibition of N . asteroides GUHJ superoxide dismutase by specific

metal-associated inhibitors Treatment Inhibition

1 mM NaN:I 10.0 20 mM NaN:I 41.0

5 mM KCN 0 1 mM KCN 0

5 mM Hz02 + 1 mM KCN 40.0 5 mM HzOz + 5 mM KCN 40.0

I

FIG. 3 . Immunofluorescent staining of' an unfixed (live) cell using anti-superoxide dismutase IgC against N. usternides GUH-2 at log phase.

TABLE VI Cytoplasmic superoxide dismutase during growth of N . asteroides

G UH-2 Superoxide dismutase

Culture age c; extracta- Specific ac- hle proteins" tivity"

h Brain-heart infusion medium

12 2.31 11.9 16 5.15 15.5 24 5.50 21.8 48 5.75 22.6 72 4.50 13.8

168 3.13 20.8

24 2.18 NT' 72 2.28 NT

168 4.13 NT 336 5.39 NT " Computer integration from the areas under the peaks of densi-

tometric scans of polyacrylamide gels after electrophoresis and stain- ing for proteins.

Specific activity as defined under "Materials and Methods.'' Ac- tivity was determined by cytochrome c reduction assay and protein by the Lowry procedure.

Chemically defined medium

' Not tested.

' A

- 400

- 300

- 200

- 100

, G O O "_ "_ "-*"--r"- 1512 24 48 72 96 120 1 4 4 168

Time (h r )

- 1600 - 1400

1200 - 1000 - 800 - 600 - 400

-

\ \ \

L L ' V

t \

IO6' b 4 k b 6 Ib ;I I; I \ 14' b- """""" 2,200

Time (days)

FIG. 4. The rate of growth of N. asteroides and secretion of superoxide dismutase into the medium. Colony-forming units ( C F U ) and cell mass were determined during growth by dilution plate method and dry weight, respectively. For the determination of super- oxide dismutase in the medium, 500 ml of medium were filtered through a membrane filter (0.45 pm) at different stages of growth of the organism and concentrated to 8 ml by ultrafiltration (PM-IO membrane). A, brain-heart infusion ( B H I ) broth; B, minimal medium (Middlebrook 7H9).

that equimolar concentrations of Mn, Fe, and Zn are associ- ated with this enzyme.

Copper-zinc superoxide dismutase is sensitive to cyanide and H20n (48,49), while Fe superoxide dismutase is inhibited by H202 but not by cyanide (50). Manganese superoxide

A Unique Superoxide Dismutase from Nocardia 95

dismutase, on the other hand, is inhibited by azide but not by either cyanide or H202 (50). As shown in Table V, the Nocar- dia enzyme was not inhibited by either 1 or 5 mM KCN. However, addition of 5 mM Hz02 along with either 1 or 5 mM KCN led to a 40% inhibition after 15 min of incubation. Prolonged incubation, e.g. 30,60, or 90 min, did not yield any further inhibition. Addition of NaN3 at a concentration of 1 and 20 mM inhibited 10 and 41% of the enzyme activity, respectively. These inhibition studies were performed as pre- viously described (5 1 ) .

Immunofluorescent Staining-Fig. 3 shows the immunoflu- orescent staining of a live cell of N. asteroides GUH-2, during log phase of growth, utilizing rabbit anti-superoxide dismutase IgG and fluorescein-labeled goat anti-rabbit IgG. The binding of purified rabbit anti-nocardia1 superoxide dismutase IgG to the surface of the organism indicates the association of this enzyme with the outer cell envelope of the bacterium. This immunofluorescent staining was specific for superoxide dis- mutase of N. asteroides because cells treated with normal rabbit IgG had no fluorescence. Further, Nocardia rubra 721- A and Nocardia caviae 112 treated with the same antibody against N. asteroides superoxide dismutase did not demon- strate labeled activity.

Rate of Synthesis and Secretion of Superoxide Dismutase

SOD 1' I

B I

I

FIG. 5. Densitometric scans of polyacrylamide gel electro- phoresis. Cells were grown in a chemically defined broth (7H9 with glucose as the carbon source). The cell-free media were concentrated by a PM-10 membrane filter from 500 ml to an 8-ml volume and 200 p1 of these concentrated media were applied to 10% polyacrylamide gels. Proteins were stained with Coomassie blue R-250 and superoxide dismutase (SOD) activity was stained with photochemical reduction of Nitro Blue Tetrazolium. A, log phase a t 24 h, stained for protein. B, log phase at 72 h, stained for protein. C , early stationary phase a t 168 h, stained for protein. D, stationary phase at 336 h, stained for protein. E, activity band of purified superoxide dismutase, approxi- mately 3 units (similar bands of enzyme activity were observed when the concentrated media were used). F, cytoplasm of N. asteroides GUH-2 grown in 7H9 media for 72 h, approximately 0.52 mg of protein, stained for protein.

during Growth Cycle of N . asteroides GUH-2"Nocardial superoxide dismutase is synthesized and secreted by the cells during all phases of their growth cycle. Thus, as shown in Table VI, at least 2% of the total cytoplasmic protein was superoxide dismutase during log phase and this enzyme rep- resented up to 5% of the total extractable protein as deter- mined by quantitative gel scans, when the cells reached sta- tionary phase. The organisms grew faster in brain-heart infu- sion medium than in chemically defined minimal medium. Thus, the time required to reach the stationary phase in brain- heart infusion medium and in minimal medium was 48 and 168 h, respectively. This was determined by monitoring both the colony-forming units and the cell mass by dilution plate method and dry weight determination, respectively, and the data are presented in Fig. 4. As shown in this figure, superoxide dismutase was preferentially secreted, per cell basis, in the early phase of the growth cycle.

The secretion of superoxide dismutase into the medium by the organism was investigated by polyacrylamide gel electro- phoresis and by cytochrome c reduction assay. Fig. 5 repre- sents the densitometric scan of 10% gel electrophoresis of proteins secreted into chemically defined medium. When 200- p1 samples of concentrated (500 + 8 ml) growth medium at different stages of growth was applied to the gels and stained for both protein and superoxide dismutase activity, an in- creased level of the enzyme was detected with time. Although a small amount of other proteins was detected by this method, superoxide dismutase remained the predominant protein in the medium at all phases of growth, and represented over 75$1 of the total proteins in the medium when cells were grown to stationary phase. No detectable level of catalase was found in this medium. Fig. 5 F presents the densitometric scan of the gels applied with 200 pl of cytoplasm (2.6 mg of protein/ml) of the organism at late log phase of growth.

DISCUSSION

The data show that N. asteroides GUH-2 produces a su- peroxide dismutase that differs significantly from those iso- lated from other bacteria. The intact nocardial enzyme has M , = 100,000, a subunit weight of 25,000, and appears to be a tetramer composed of identical subunits. Furthermore, equi- molar amounts of Fe, Mn, and Zn are bound to the enzyme. Rank order correlation analyses of amino acid differences, a statistical multiple regression analysis computed for pairwise comparisons of nocardial enzyme with other superoxide dis- mutases, indicate that this enzyme is most closely related to the Mn-containing enzyme of Mycobacterium smegmatis. Un- like the enzyme isolated from N. asteroides, however, the superoxide dismutase produced by M. smegmatis is a trimer with a total molecular weight of only 61,500 (52). The nocar- dial enzyme is least related to the Cu-Zn enzymes isolated from eukaryotic cells (4-13).

Enzyme inhibition analyses demonstrated that the nocar- dial superoxide dismutase was not inhibited by cyanide which inhibits Cu-Zn superoxide dismutases (48, 49). However, a maximum of 40% of the enzyme's activity was destroyed by H,Oz which inactivates Fe superoxide dismutases, but has little effect on the Mn enzymes (48, 49). Thus, these obser- vations suggest that the remaining 60970 activity of the nocar- dial enzyme reflects active Mn sites. Since Mn- and Fe-con- taining enzymes are not sensitive to CN-, and Mn-containing enzymes are more sensitive to azide than are either Cu-Zn enzymes or Fe enzymes (50), and since 20 mM azide inhibited only 40% activity of the nocardial enzyme, it appears that the superoxide dismutase produced by N. asterozdes GUH-2 re- tains some characteristics of both Mn and Fe superoxide dismutases. Although bacteria have been shown to produce

96 A Unique Superoxide Di

superoxide dismutase which contain either Fe-Zn (53), Cu-Zn (14, 54), Mn (16-la), or Fe (27-29), none has been reported previously to contain three different metals in a single, elec- trophoretically distinct enzyme.

Few bacteria have been demonstrated to secrete superoxide dismutase during growth; however, it was shown that Myco- bacterium tuberculosis, a human pathogen, secreted a Fe superoxide dismutase while the nonpathogenic strains, M. smegmatis and Mycobacterium phlei, did not (55). Similarily, it was shown that the virulent strain N, asteroides GUH-2 secretes the enzyme into the growth medium (Fig. 4) while in contrast, extracellular enzyme could not be detected in the growth medium of nonpathogenic strains of Nocardia (data not shown). Recent studies have indicated that log phase cells of N. asteroides GUH-2 are a t least 1000 times more virulent for mice than stationary phase cells of the same organism (38). We have shown that the rate of secretion of superoxide dismutase into the medium by this organism was significantly higher, on a per cell basis, during the early stages of growth than during stationary phase (Fig. 4). Because the addition of extracellular superoxide dismutase has been shown to protect some bacteria against polymorphonuclear phagocytes (33-35), it seems reasonable to suggest that both the secretion of this enzyme into the medium as well as its association with the outer cell envelope (Fig. 3) could offer significant protection against superoxide radical toxicity during active phagocytosis by polymorphonuclear neutrophils. Thus, these observations support an active role for this unique superoxide dismutase as a possible determinant of nocardia1 pathogenesis.

Acknowledgment-We would like to thank Neil Sharkey for his expert assistance in determining the metal analysis by atomic ahsorp- tion spectrophotometry.

REFERENCES 1. McCord, J . M., Keele, B. B., Jr., and Fridovich, I. (1971) Proc.

2. Fridovich, I. (1975) Annu. Reu. Biochem. 44, 147-159 3. Hassan, H. M., and Fridovich, I. (1977) J. Bacteriol. 129, 1574-

4. McCord, J. M., and Fridovich, I. (1969) J. Biol. Chem. 244,6049-

5. Misra, H. P., and Fridovich, I. (1972) J. Biol. Chem. 247, 3410-

6. Keele, B. B., Jr., McCord, J. M., and Fridovich, I. (1971) J. Biol.

7. Weisiger, R. A., and Fridovich, I. (1973) J . Biol. Chem. 248,3582-

8. Banister, W. H., Daigleish, D. G., Banister, J. V., and Wood, E. J .

9. Beauchamp, C. O., and Fridovich, I. (1973) Biochim. Biophys.

10. Asada, K., Urano, M., and Takahashi, M. (1973) Eur. J. Biochem.

11. Goscin, S. A,, and Fridovich, I. (1972) Biochim. Biophys. Acta

12. Sawada, Y., Ohyama, T., and Yamazaki, I. (1972) Biochim. Bio-

13. Lee, Y. M., Ayala, F. J., and Misra, H. P. (1981) J. Biol. Chem.

14. Puget, K., and Michelson, A. M. (1974) Biochem. Biophys. Res.

15. Martin, J . P., Jr., and Fridovich, I. (1981) J. Biol. Chem. 256,

Natl. Acad. Sci. U. S. A. 68, 1024-1027

1583

6055

3414

Chem. 246,2875-2880

3592

(1972) Znt. J. Biochem. 3, 560-568

Acta 317, 50-64

36,257-266

289, 276-283

phys. Acta 268, 305-312

256,8506-8509

Commun. 58,830-838

'smutase from Nocardia

6080-6089 16. Keele, B. B., Jr., McCord, J. M., and Fridovich, I. (1970) J. Biol.

17. Vance, P. G., Keele, B. B., Jr., and Rajagopalan, K. V. (1972) J.

18. Avron, G. P., Henry, L., Palmer, J. M., and Hall, D. 0. (1976)

19. Ravindranath, S. D., and Fridovich, I. (1975) J. Biol. Chem. 250,

20. Steinman, H. M., and Hill, R. L. (1973) Proc. Natl. Acad. Sci. U.

21. Bridgen, J., Harris, J . A,, and Northrop, F. (1975) FEBS Lett. 49,

22. Fridovich, I. (1974) Life Sci. 14, 819-826 23. McCord, J . M. (1976) Adu. Exp. Med. Biol. 74, 540-550 24. Misra, H. P., and Fridovich, I. (1977) J. Biol. Chem. 252, 6421-

25. Misra, H. P., and Keele, B. B., Jr . (1975) Biochim. Biophys. Acta

26. Asada, K., Yoshikawa, K., Takahashi, M., Maeda, Y., and En-

27. Yost, F. J., Jr., and Fridovich, I. (1973) J. Biol. Chem. 248,4905-

28. Yamakura, F. (1976) Biochim. Biophys. Acta 422, 280-294 29. Puget, K., and Michelson, A. M. (1974) Biochimie 56, 1255-1267 30. Asada, K., Kanematsu, S., Takahashi, M., and Kona, Y. (1976)

31. Beaman, B. L. (1979) Infect. Zmmun. 26, 355-361 32. Filice, G. A,, Beaman, B. L., Krick, J. A,, and Remington, J. S.

(1980) J. Infect. Dis. 142,432-438 33. Johnston, K. B., Jr., Keele, B. B., Jr., Misra, H. P., Lehmeyer, J .

E., Webb, L. S., Bachner, R. L., and Rajagopalan, K. V. (1975) J. Clin. Znuest. 55, 1357-1372

34. Babior, B. M., Kipnes, R. S., and Curnutte, J. T. (1973) J. Clin. Znuest. 52, 741-744

35. Yost, F. J., and Fridovich, I. (1974) Arch. Biochem. Biophys. 161,

36. Beaman, B. L., and Maslan, S. (1977) Infect. Zmmun. 16,995-1004 37. Beaman, B. L. (1975) J. Bacteriol. 123, 1235-1253 38. Beaman, B. L., and Maslan, S. (1978) Infect. Zmmun. 20,290-295 39. Davis, B. J . (1964) Ann. N. Y. Acad. Sci. 121,404-427 40. Chrambach, A., Reisfeld, R. A,, Wyckoff, M., and Zaccari, J.

41. Beauchamp, C., and Fridovich, I. (1971) Anal. Biochem. 44,276-

42. Weber, K., and Oshorn, M. (1969) J. Biol. Chem. 244,4406-4412 43. Moore, S., and Stein, W. H. (1963) Methods Enzymol. 6,819-831 44. Edelhoch, H. (1967) Biochemistry 6, 1948-1954 45. Bailey, J. L. (1967) Techniques in Protein Chemistry, pp. 340-

46. Benacerraf, B., and Unanue, E. R. (1979) Textbook of Zmmunol-

47. Marchalonis, J . J., and Weltman, J. K. (1971) Comp. Biochem.

48. Rotilio, G., Bray, R. C., and Fielden, E. M. (1972) Biochim.

49. Fridovich, I. (1975) Annu. Reu. Biochem. 44, 147-159 50. Misra, H. P., and Fridovich, I. (1978) Arch. Biochem. Biophys.

51. Privalle, C. T., and Gregory, E. M. (1979) J. Bacteriol. 138, 139-

52. Kusunose, M., Noda, Y., Ichihara, K., and Kusunose, E. (1976)

53. Searcy, K. B., and Searcy, D. G. (1981) Biochim. Biophys. Acta

5 4 . Vignais, P. M., Terech, A,, Meyer, C. M., and Henry, M.-F. (1982)

55. Kusunose, E., Ichihara, K., Noda, Y., and Kusunose, M. (1976) J.

Chem. 245,6176-6181

Biol. Chem. 247,4782-4786

Biochem. SOC. Trans. 4,618-620

6107-6112

S. A. 70,3725-3729

392-395

6423

379, 418-425

manji, K. (1975) J. Biol. Chem. 250,2801-2807

4908

Adu. Exp. Med. Biol. 74, 551-564

395-401

(1967) Anal. Biochem. 20, 150-154

287

341, Elsevier, New York

ogy, Williams & Wilkins, Baltimore

Physiol. 38B, 609-625

Biophys. Acta 268,605-609

189, 317-322

145

Arch. Microbiol. 108, 65-73

670, 39-46

Biochim. Biophys. Acta 701,305-317

Biochem. (Tokyo) 80, 1343-1352