Proc. Nat. Acad. Sci. USA Vol. 68, No. 5, pp. 1024-1027, May 1971

An Enzyme-Based Theory of Obligate Anaerobiosis: The Physiological Function of Superoxide Dismutase

JOE M. McCORD, BERNARD B. KEELE, JR.*, AND IRWIN FRIDOVICH Department of Biochemistry, Duke University Medical Center, Durham, North Carolina,27706 Communicated by Philip Handler, February 11, 1971

ABSTRACT The distribution of catalase and super- not contain catalase, e.g., the streptococci, pneumococci, and oxide dismutase has been examined in various micro- lactic acid bacteria. More recently, Agromyces ramnosus, the organisms. Strict anaerobes exhibited no superoxide dis- mutase and, generally, no catalase activity. All aerobic predominant soil organism, was found to lack catalase, de- organisms containing cytochrome systems were found to spite its aerobic metabolism (8, 9). Further, some strict contain both superoxide dismutase and catalase. Aero- anaerobes have catalase activity, yet cannot tolerate ex- tolerant anaerobes, which survive exposure to air and posure to air (10). metabolize oxygen to a limited extent but do not contain cytochrome systems, were found to be devoid of catalase The recent discovery and characterization of superoxide activity but did exhibit superoxide dismutase activity. dismutase (11, 12) raised the question of its physiological This distribution is consistent with the proposal that the role. The substrate, a reactive and potentially detrimental prime physiological function of superoxide dismutase is free-radical form of oxygen, had long been implicated as an protection of oxygen-metabolizing organisms against the intermediate in the reduction of 02 by a family of metallo- potentially detrimental effects of the superoxide free radical, a biologically produced intermediate resulting flavoenzymes (13-17). Evidence that the superoxide radical is from the univalent reduction of molecular oxygen. released into free solution from the enzyme surface was not obtained until 1968 (11), but was then quickly confirmed for Pasteur's discovery that certain organisms are not only capable the xanthine oxidase system by the detection of its electron of growing in oxygen-free environments, but in many cases paramagnetic resonance signal (18). The reaction catalyzed are restricted to such environments, has never been satis- by superoxide dismutase, factorily explained. Obligately anaerobic organisms are strongly inhibited or killed by exposure to molecular oxygen 02- + 02- + 2H+ - 02 + H202, (1). The possible role of catalase in protecting aerobic micro- organisms from death by hydrogen peroxide poisoning was proceeds spontaneously at pH 7.7, with a rate constant of ap- rather quickly recognized. In 1893 Gottstein discovered that proximately 2 X 105 M-I sec- (19). The ubiquity and con- certain bacteria decomposed H202 with the liberation of a gas stancy of this enzymic activity in a wide variety of tissues and (2). In 1907 it was observed that certain anaerobic bacteria organisms led us to seek its physiological importance. Interest contain no detectable catalase, whereas all the aerobes ex- was heightened by the discovery that the superoxide dismutase amined exhibited significant catalatic activity (3). This led of Escherichia coli is a manganoprotein (20), which bears little to proposals that oxygen toxicity was occasioned by its re- resemblance to the copper- and zinc-containing mammalian duction product, hydrogen peroxide (4, 5). Proceeding from enzyme (12, 21), but which is nevertheless present in the or- this assumption investigators reasoned (4) that: (a) H202 ganism at similar concentration and displays a nearly identical should accumulate in aerobic cultures of anaerobes, as a result specific activity. either of bacterial metabolism or of the action of light on the The data presented in this report strongly implicate medium, (b) anaerobes should be sensitive to externally added superoxide dismutase as being vital to the existence of any hydrogen peroxide and, (c) anaerobes should grow aerobically organism that metabolizes oxygen. when catalase is present in the medium. Technical limitations at the time prevented detection of the low but toxic concentra- MATERIALS AND METHODS tions of H202 that anaerobes produce. It was, however, subse- Frozen or lyophilized cells of certain microorganisms were quently shown that nearly all anaerobes do produce H202 (6). generously supplied for this study as follows: Salmonella The sensitivity of anaerobes to H202 was readily shown, and a typhimurium and an unidentified pseudomonad from Dr. high degree of variation was apparent (4, 7). It could not be Henry Kamin; Halobacterium salinarium from Dr. Jayant shown that the presence of catalase would allow aerobic Joshi; Rhizobium japonicum from Dr. Gerald Elkan; My- growth of anaerobes (4, 5), although one anaerobe grew cobacterium sp. from Dr. Jerome Perry; Veillonella alcalescens better at low oxygen tension with catalase present (4). Thus, and Butyribacterium rettgeri from Dr. Charles Wittenberger; although catalase aids the survival of some microorganisms Clostridium pasteurianum, Clostridium sticklandii, Clostridium in aerobic media, catalase activity does not provide a suf- lentoputrescens, Clostridium barkeri, and Clostridium sp. ficient answer. Many organisms capable of aerobic growth do (strain M.E.) from Dr. T. C. Stadtman; Butyrivibrio fibri- solvens from Dr. Sam Tove; Clostridium cellobioparum and * Present address: Institute of Dental Research, University of N2C3, an unclassified rumen organism, from Drs. Robert Alabama Medical Center, Birmingham, Ala. 35233. Mah and R. E. Hungate; Zymobacterium oroticum from Dr. 1024 Downloaded by guest on September 24, 2021 Proc. Nat. Acad. Sci. USA 68 (1971) Theory of Obligate Anaerobiosis 1025

K. V. Rajagopalan; and Clostridium acetobutylicum from TABLE 1. Superoxide dismutase and catalase contents of a Mr. Robert Waterson. E. coli cells were obtained as a frozen variety of microorganisms paste from Miles Laboratories. Micrococcus radiodurans, Saccharomyces cerevisiae (ATCC 560), Streptococcus fecalis, Superoxide Streptococcus mutans, Streptococcus bovis, Streptococcus mitis, dismutase Catalase Streptococcus lactis, and Lactobacillus plantarum were grown (units/mg) (units/mg) by us on either trypticase soy broth or APT broth, available Aerobes: from BBL, Cockeysville, Md., or on Brain-Heart Infusion Escherichia coli 1.8 6.1 Broth obtained from Difco. Cultures were obtained as follows: Salmonella typhimurium 1.4 2.4 M. radiodurans from Dr. Jane Setlow; S. fecalis and L. plan- Halobacterium salinarium 2.1 3.4 tarum from Dr. John McNeill; and S. mutans 6715, S. lactis Rhizobium japonicum 2.6 0.7 (ATCC 19435), S. bovis (ATCC 9809), and S. mitis (ATCC Micrococcus radiodurans 7.0 289 9811) from Dr. James Sandham. Saccharomyces cerevisiae 3.7 13.5 Cells were suspended in 50 mM potassium phosphate Mycobacterium sp. 2.9 2.7 buffer, pH 7.8, containing 0.1 mnM EDTA, and sonicated Pseudomonas sp. 2.0 22.5 in an ice bath by means of a Branson sonifier at a power Strict Anaerobes: setting of 100 W to disrupt the cells. After centrifugation, Veillonella alcalescens 0 0 the cell-free extracts were assayed for superoxide dismutase Clostridium pasteurianum, sticklandii, activity as previously described (12). In certain cases, the lentoputrescens, cellobioparum, presence of low molecular weight substances capable of barkeri 0 0 reducing cytochrome c necessitated an overnight dialysis Clostridium acetobutylicum 0 of the cell-free extracts to remove these interfering substances. Clostridium sp. (strain M.C.) 0 0 The extracts were assayed for catalase activity by means of Butyrivibrio fibrisolvens 0 0.1 a Gilson Oxygraph equipped with a Clark electrode and a N2C3* 0 <0.1 thermostatted cell. The assay mixture contained 0.02 M Aerotolerant Anaerobes: hydrogen peroxide in 50 mrM potassium phosphate buffer at pH 7.8, containing 0.1 mM EDTA, at 25°C. Rates of Butyribacterium rettgeri 1.6 0 Streptococcus fecalis 0.8 0 oxygen production were compared to the rate obtained with Streptococcus mutans 0.5 0 a standardized solution of bovine liver catalase obtained Streptococcus bovis 0.3 0 from Sigma. Alternatively, catalase was assayed spectro- Streptococcus mitis 0.2 0 photometrically by the method of Beers and Sizer (22). Streptococcus lactis 1.4 0 Protein contents of cell-free extracts were estimated at 280 Zymobacterium oroticum 0.6 0 nm assuming Elm (c = 1%) = 10.0. Lactobacillus plantarum 0 0 RESULTS AND DISCUSSION * N2C3 is an unclassified cellulolytic Gram-negative rod is5- The superoxide dismutase and catalase contents of 26 species lated from the rumen of an African zebu steer and has been of microorganisms of three categories are reported in Table described by Margherita, S. S., and R. E. Hungate, J. Bacteriol., 1. The aerobes are the microorganisms that can utilize molec- 86, 855 (1963). ular oxygen as the terminal electron acceptor. The obligate anaerobes not only lack the cytochrome system necessary The strict anaerobes for aerobic respiration, but are unable to survive under conditions of aeration. The organisms termed aerotolerant The ten organisms in this classification represent various anaerobes do not utilize molecular oxygen as the terminal clostridial species and three organisms that were isolated electron acceptor for their energy metabolism, but, with originally from the bovine rumen. Aeration is lethal for all one exception, they are capable of reducing oxygen to a these species. Strict anaerobes are nearly always catalase- limited extent. Many aerotolerant anaerobes grow as well negative. Two species of the rumen bacteria displayed under vigorous aeration as under anaerobic conditions and, very low levels of catalase activity. The existence of signifi- at any rate, are not killed by such exposure to oxygen. cant catalase activity in certain strict anaerobes has been Aerotolerant anaerobe seems to be a more accurate description observed by others (10). The consistent absence of superoxide of most of the organisms in this category than the term dismutase activity among the obligate anaerobes contrasts microaerophile, which has frequently been used to describe with the presence of nearly constant levels of this activity this kind of behavior. in the organisms that possess oxygen-metabolizing capabili- ties. The aerobes The eight organisms in this category (Table 1), chosen at The aerotolerant anaerobes random, represent a broad spectrum of organisms, some Examination of the third class of organisms in Table 1 Gram-positive, some Gram-negative, one yeast. All of these casts additional light on the relative importance of catalase organisms contain both catalase and superoxide dismutase. and superoxide dismutase. While none of these organisms The catalase activity of these organisms ranged from 0.7 utilizes oxygen as a primary electron acceptor, most of them unit/mg (R. japonicum) to 289 units/mg (M. radiodurans), consume some oxygen by the action of flavoprotein and a 400-fold variation. The activity of superoxide dismutase metalloprotein oxidases not coupled to ATP synthesis (23). of these organisms ranged from 1.4 units/mg (S. typhimurium) The apparent product of these oxidative processes is usually to 7.0 units/mg (M. radiodurans), only a 5-fold variation. hydrogen peroxide. Despite their limited abilities to consume Downloaded by guest on September 24, 2021 1026 Biochemistry: McCord et al. Proc. Nat. Acad. Sci. USA 68 (1971)

oxygen and to produce hydrogen peroxide, the data show oxide dismutase.) Of a variety of flavoprotein oxidases that none of these organisms contains a significant amount of and dehydrogenases, several were shown to produce super- catalase activity. With only one exception, though, all oxide to very limited extents (24). Clostridial ferredoxin, aerotolerant anaerobes contained superoxide dismutase a nonheme-iron protein, was shown to produce the radical activity, at levels averaging 30% of that shown by the upon air oxidation of. its reduced form (25). Pig kidney aerobes. The ability of these organisms to survive without diamine oxidase, a copper-containing enzyme, has recently catalase probably depends upon their low rate of production been shown to produce the superoxide radical (26). The of H202 and upon the relative chemical stability of this autoxidation of hemoglobin and myoglobin and all similar compound that enables it to diffuse from these cells without one-electron transfers to oxygen are processes that almost causing damage. An accumulation of H202 in the medium certainly produce 2- (27-29), and a variety of low molecular surrounding these cells would be prevented by the catalatic weight electron carriers such as reduced flavins and hydro- action of substances in the medium or, in a mixed culture, quinones likewise produce the radical upon autoxidation by the catalatic action of other species of cells containing (30). Since improved methods now exist for the detection catalase. This explanation has already been proposed (9). of the superoxide radical, we expect that additional biological The single aerotolerant organism that contained no sources of 2- will be discovered in the near future. superoxide dismutase was Lactobacillus plantarum. Impor- Superoxide dismutase activity appears to be a concomitant tantly, no consumption of oxygen by L. plantarum could be and perhaps necessary condition for the survival of oxygen- detected. The cells studied were grown under aerobic condi- metabolizing organisms. Other factors in addition to the tions and harvested during log phase. They were resus- lack of superoxide dismutase might render an organism pended in fresh growth medium and oxygen consumption unable to survive in the presence of oxygen, e.g., enzymes was measured using a Clark electrode with a Gilson Oxy- that are autoxidized and thereby inactivated by molecular graph. Under identical conditions, oxygen consumption oxygen. Conceivably, membranes containing easily autoxi- by a known consumer such as S. cerevisiae was easily mea- dizable unsaturated lipids could be rendered nonfunctional surable. If the rate of consumption by L. plantarum had been by contact with molecular oxygen. However, the data as much as 5% that of S. cerevisiae, it would have been presented herein strongly suggest that superoxide dismutase detected. is a factor of primary importance in enabling organisms to The presence of superoxide dismutase in the aerotolerant survive the challenge presented by the reactive intermediate anaerobes and its absence among the strict anaerobes suggests species resulting from the univalent reduction of molecular that the normal physiological function of this enzyme is, oxygen. Further study of this process may lead to a greater indeed, the dismutation of superoxide radicals. It appears understanding of the mechanisms of oxygen toxicity. possible that this activity may be the single most important enzymic activity for enabling organisms to survive in the This work was supported by grant GM-10287 from the Na- presence of molecular oxygen. Presumably, the great tional Institutes of Health, Bethesda, Md. J. M. M. and B. B. chemical reactivity of superoxide anion precludes the pos- K. were postdoctoral fellows of the National Institutes of Health. sibility of depending upon diffusion from the cell as a means of disposal of this radical and necessitates the presence, 1. Stanier, R. Y., M. Doudoroff, and E. A. Adelberg, The within the cell, of superoxide dismutase. Admittedly, were Microbial World, 3rd ed. (Prentice-Hall, Inc., Englewood an organism obligately anaerobic for some other reason, Cliffs, N.J., 1970), p. 75. 2. Gottstein, A., Virchows Arch., 133, 295 (1893). there would be no reason for it to continue to carry the 3. Rywosch, D., and M. Rywosch, Zentralbl. Bakteriol. Para- genetic burden of the genes for catalase and superoxide sitenk. Infectionskr. Hyg. Abt. Orig., 44, 295 (1907). dismutase and this correlation might have some other, as yet 4. McLeod, J. W., and J. Gordon, J. Pathol. Bacteriol., 26, 332 unsuspected, explanation. (1923). 5. Callow, A. B., J. Pathol. Bacteriol., 26, 320 (1923). An organism without superoxide dismutase could survive 6. Gordon, J., R. A. Holman, and J. W. McLeod, J. Pathol. in an oxygen environment if it does not produce quantities Bacteriol., 66, 527 (1953). of the superoxide radical sufficient to jeopardize its survival. 7. McLeod, J. W., and J. Gordon, J. Pathol. Bacteriol., 26, 326 L. plantarum appears to be such as organism. Even an (1923). 8. Gledhill, W. E., and L. E. Casida, Jr., Appl. Microbiol., 18, oxygen-metabolizing organism need not possess the enzyme, 340 (1969). providing the direct products of its limited oxygen-metabo- 9. Jones, D., J. Watkins, and D. J. Meyer, Nature, 226, 1249 lizing pathways be hydrogen peroxide and water rather (1970). than the superoxide radical. 10. Prevot, A. R., and H. Thouvenot, Ann. Inst. Pasteur Paris, The ubiquity of superoxide dismutase among all aerobic 83, 443 (1952). 11. McCord, J. M., and I. Fridovich, J. Biol. Chem., 243, 5753 organisms and its uniform distribution among the various (1968). tissues of mammals implies that the superoxide free radical 12. McCord, J. M., and I. Fridovich, J. Biol. Chem., 244, 6049 is a commonly occurring but quite undesirable physiological (1969). species. Relatively few biological reactions have been shown 13. Fridovich, I., and P. Handler, J. Biol. Chem., 233, 1581 (1958). to produce the radical in vivo. The first enzyme shown to 14. Fridovich, I., and P. Handler, J. Biol. Chem., 236, 1836 release the radical into free solution was milk xanthine (1961). oxidase (11, 18), a metalloflavoprotein containing nonheme 15. Greenlee, L., I. Fridovich, and P. Handler, Biochemistry, 1, iron. Two other nonheme-iron flavoenzymes, rabbit liver 779 (1962). 16. Fridovich, I., and P. Handler, J. Biol. Chem., 237, 916 aldehyde oxidase and dihydroorotate dehydrogenase from (1962). Zymobacterium oroticum (17), also produce the radical. 17. Handler, P., K. V. Rajagopalan, and V. Aleman, Fed. Proc., (Z. oroticum is an aerotolerant anaerobe and contains super- 23, 30 (1964). Downloaded by guest on September 24, 2021 Proc. Nat. Acad. Sci. USA 68 (1971) Theory of Obligate Anaerobiosis 1027

18. Knowles, R. F., J. F. Gibson, F. M. Pick, and R. C. Bray, 24. Massey, V., S. Strickland, S. G. Mayhew, L. G. Howell, Biochem. J., 111, 53 (1969). P. C. Engel, R. G. Matthews, M. Schuman, and P. A. 19. D. Behar, G. H. Czapski, J. Rabani, L. M. Dorfman, and Sullivan, Biochem. Biophys. Res. Commun., 36, 891 (1969). H. A. Schwarz, J. Phys. Chem., 74, 3209 (1970). 25. Orme-Johnson, W. H., and H. Beinert, Biochem. Biophys. 20. Keele, B. B., Jr., J. M. McCord, and I. Fridovich, J. Biol. Res. Commun., 36, 905 (1969). Chem., 245, 6176 (1970). 26. Rotilio, G., L. Calabrese, A. Finazzi-Agro, and B. Mondovi, 21. Keele, B. B., Jr., J. M. McCord, and I. Fridovich, J. Biol. Biochim. Biophys. Ada, 198, 618 (1970). Chem., in press. 27. George, P., and C. J. Stratmann, Biochem. J., 51, 163 (1952). 22. Beers, R. F., and I. W. Sizer, J. Biol. Chem., 195, 133 (1952). 28. George, P., and C. J. Stratmann, Biochem. J., 51, 418 (1952). 23. Stanier, R. Y., M. Doudoroff, and E. A. Adelberg, The 29. George, P., and C. J. Stratmann, Biochem. J., 57, 568 (1954). Microbial World, 3rd ed. (Prentice-Hall, Inc., Englewood 30. McCord, J. M., and I. Fridovich, J. Biol. Chem., 245, 1374 Cliffs, N.J., 1970), p. 663. (1970). Downloaded by guest on September 24, 2021