Generation of hydrogen peroxide during the oxidation of l -phenylalanine by Proteus mirabilis isolated membranes

BIOCHIMIE, 1982, 64, 891-897.

Generation of hydrogen peroxide during the oxidation of L-phenylalanine by Proteus mirabilis isolated membranes. G. SAURET-IGNAZI, A. M. LABOURE-ROSSAT, H. M. JOUVE and J. PELMONT ~.

Laboratoire de Biochimie, C.E.R.M.O., E . R . A . n ° 673, Domaine Universitaire, B.P. 53 X , 38041 Grenoble CJdex, France.

(Re~u le 14-5-1982, acceptd le 22-6-1982).

R~sum&

Summary.

L'oxydation de la L-phJnylalanine par la membrane cytoplasmique de Proteus mirabilis lib~re du peroxyde d'hydrogkne fi raison de 1-3 p. cent du flux total d'Jlectron~ (30-110 nmoles mn -1 mg-9. Le peroxyde est mis en Jvidence par fluorimJtrie en presence de la peroxydase du raifort, ou par oxydation anodique sur ~lectrode de platine. L'action inhibitrice de la superoxyde dismutase sur la dJtection du peroxyde [ormJ, dans le cas du test par fluorescence, pourrait sugg&er qu'une partie du H~02 vient de la dismutation de radicaux superoxyde. En r~alitJ cet effet peut s' expliquer par la formation secondaire de superoxyde au cours de la r~action de la peroxydase, car il n' a pas Jtj possible de montrer, en utilisant l'adrJnaline comme r~actif de dJtection, une quantit~ mesurable de superoxyde li~ ?t l'oxydation de la phJnylalanine par la membrane (moins de 1 ?t 2 nmoles mn Img-9. II semble donc que le peroxyde observ~ soit engendr~, au moins pour la plus grande partie, sans passer par le stade superoxyde. Les faits sont en faveur d'une source de peroxvde placJe en amont de la cha~ne des cytochromes. Une faible r~activitJ de la ph~nylalanine dJshvdrog~nase membranaire avec l'oxygkne de l'air pourrait ~tre ~ l' origine de ce peroxyde.

In the process of z-phenylalanine oxidation by Proteus mirabilis cytoplasmic membrane, hydrogen peroxide was produced at a rate corresponding to 1-3 per cent of the total electron flow (30110 nmoles min-lmg-1). Peroxide was estimated using a fluorimetric assay with horseradish peroxidase, or by anodic oxidation on a platinum electrode. When using the ]ormer method, superoxide dismutase decreased the apparent yield of peroxide, a fact suggesting that H~Oe was in part the dismutation product of superoxide radicals. However the superoxide dismutase effect could be an artefact due to the generation of some superoxide during the peroxidatic reaction in the assay. Adrenaline was the reagent used for the detection of superoxide. There was no significant emergence of superoxide as the result of phenylalanine oxidation by the membrane (specific activity lower than 1-2 nmoles min-lmg-1). Thus it seemed that superoxide was not an intermediate for the bulk of H~Oe formed in this system. According to the results, peroxide was probably formed at a stage of electron transport earlier than the cytochrome level. The membrane phenylalanine dehydrogentt~e could be a site where peroxide was evolved in these experiments.

Mots-cl~s : peroxyde d'hydrog~ne / ph~nylalanine d~shydrog~nase / superoxyde.

Key-words : hydrogen peroxide / phenylalanine dehydrogenase / superoxyde.

Introduction.

Dioxygen reduction by terminal cytochromes of respiratory chains is known as a four electrontransfer producing water. However, isolated heart mitochondria have been shown to generate some hydrogen peroxide [1, 2]. The yield of peroxide was the highest when mitochondria were in state 4, (> To whom all correspondence should be addressed.

and was estimated to account for about 2 per cent of the oxygen uptake [3, 4]. Peroxide was found to be derived mostly by dismutation of superoxide radicals formed at different sites of the carrier chain, mainly at the NADH-dehydrogenase level and at the ubiquinone-cytochrome b level [5-7]. 63

892

G. Sauret-lgnazi and coll.

Bacterial respiratory chains are also believed to generate small a m o u n t s of peroxide a n d other byproducts of oxygen reduction. F o r instance, some superoxide is p r o d u c e d by Paracoccus denitrificans [8]. The occurrence and the biological effects of these factors have been discussed in detail in several instances [9-11]. T h e studies of Webster et al. [12, 13] with Vitreoscilla have revealed that h y d r o g e n peroxide is formed during the N A D H cytochrome o oxidase reaction. T h e r e is n o d o u b t that 0-2, H202 a n d OH" can be generated in vitro by a variety of purified or semi-purified bacterial electron carriers, b u t the generality of their occurrence during respiration by a complete bacterial chain of electron transfers to oxygen r e m a i n s scanty. T h e fluorimetric assay devised by K e s t o n a n d B r a n d t [14] a n d used in reference [12] allowed us to check the a p p e a r a n c e of hydrogen peroxide in respiring m e m b r a n e s of Proteus mirabilis. This organism offers a favorable case where the respiratory chain is involved in the oxidation of various aminoacids with a high efficiency [15-17]. We show that significant a m o u n t s of peroxide were p r o d u c e d during p h e n y l a l a n i n e oxidation by Proteus purified m e m b r a n e s .

Material and Methods. 1) Bacterial growth. P. mirabilis, wild type, was secured from the Collection de l'Institut Pasteur, Paris. Bacteria were gro~vn as previously described [6], using minimal medium supplemented with casein hydrolysate (Merck ref. 2238, final concentration : 20 g per liter). The growth temperature was 37°C, and the culture was strongly aerated in Biolafitte fermentors (Gourdon, Poissy, France). The bacterial strain was periodically reisolated from one colony on nutrient agar, and carefully checked for urease, ornithine decarboxylase and phenylalanine oxidase.

3) Assay procedures. The production of phenylpyruvate in the phenytalanine oxidase reaction, and the oxidation of aminoacids using dichlorophenolindophenol (DCIP) as an electron acceptor were estimated according to reference [15]. Catalase activity was assayed by the peroxidase-o-dianisidine col0rimetric method according to Jouve et al. [19]. The xanthine-oxidase system was used for generating superox.ide radicals in the cytochrome c assay of superoxide dismutase according to Mc Cord et al. [201. Superoxide radicals were estimated using the epinephrin oxidation assay [20] and a Cbance-Aminco dual wavelength spectrophotometer as in [8]. A Beckman Acta M6 apparatus was used for most other absorbance measurements. The platinum electrode E2 of Thevenot et al. [21] was used for HOD.oestimation, at 650 mV versus an Ag/AgC1 reference electrode. The reaction mixture contained in 10 ml : potassium phosphate buffer (pH 7.6), 25 m M ; phenylalanine, 10 mM. The assay was started by adding the membrane particles (50 rtg) or known amounts of H~O_~. Estimations of hydrogen peroxide was performed by the fluorimetric procedure of Keston and Brandt 1141. LDADCF was made 0.1 mM in ethanol and diluted shortly before use by dilution with 4 vol. of 0.01 N NaOH. The assay mixture contained per ml (3 ml final) : LDADCF, 2 nmoles ; ZnSO,, 1.4 nmole ; horseradish peroxidase, 2 rtg; MgC1._,, 10 ~moles; L-phenylalanine or other aminoacid, 50 ~moles; potassium phosphate, 20 ~tmoles (final p H : 7.2). LDADCF oxidation was monitored at 30°C and 528 nm in a Perkin-Elmer MPF 2 fluorimeter (excitation : 490 nm), or in a Jobin-Yvon fluorimeter, type NC 321SC. Bacterial membranes (20 Ixg of inner membranes or 100 Ixg of total envelopes) were added in 50 ~tl after 2 minutes, and the increase of fluorescence in the mixture was monitored afterwards at 528 nm. The result was corrected for background oxidation of LDADCF in the absence of aminoacid. The amount of peroxidase used had to be precisely controlled in order to obtain reliable results. A standard curve was built using known amounts of H~O2. 4) Membrane particles. Cells were lysed and inner membranes of P. mirabilis were isolated by the method of Hasin et al. [22l, using osmotic shock, sonication and isopycnic zonal ultracentrifugation in a sucrose gradient as described [17]. Total protein was estimated according to reference [23].

2) Chemicals. Aminoacids and common salts, reagent grade, were purchased from Merck; cytochrome c, catalase, and horseradish peroxidase from Boehringer ; xanthine oxidase, bovine blood superoxide dismutase, electron dye acceptors, L-epinephrlne (adrenaline) bitartrate, and ubiquinone (coenzyme Q6), from Sigma. Diacetyl 2',7'-dichlorofluorescin (LDADCF) was prepared from dichlorofluorescein (Merck) after reduction and acetylation, according to the procedure of Brandt and Keston [1'8]. LDADCF was recrystallized from chloroform-petroleum ether, and stored dry at + 4°C, Another sample of LDADCF, kindly provided by Dr. Webster, furnished identical results in this work. BtOCHIMIE, 1982, 64, n ° 10.

Results. The n o n - f l u o r e s c e n t diacetyldichloro-fluorescin ( L D A D C F ) , after t r e a t m e n t with dilute alkali, rapidly becomes oxidized to a fluorescent c o m p o u d i n the presence of peroxidase with H202 in the n a n o m o l a r range [14]. Figure 1 shows that the oxidation of L D A D C F in the presence of P. mirabilis m e m b r a n e s was m a r k e d l y e n h a n c e d after addition of L-phenylalanine. N o such effect was

893

Hydrogen peroxide and the bacterial membrane.

observed with 9-phenylalanine, and neither enantiomer promoted this oxidation in the absence of membranes. The L-phenylalanine promoted L D A D C F oxidation was proportional to the concentration of inner membrane particles in the range of 5 to 50 ~g of total protein in the assay.

o

CN

IXl

a

I

b

__

1ran

TIME Fro.

1. - - Oxidation o/ LDACDF by horse radish

peroxidase in the presence of membrane particles. (a) control, in the absence of aminoacid and membranes. (b) like a, after addition of membrane particles (5 p,g protein). (c) like a with 10 m M r-phenylalanine. (d) like c, after addition of membrane particles (5 p.g protein): complete system. For experimental conditions see Material and Metheds.

of H202 and no phenylalanine, and allowed to measure the extent of peroxide generation by the bacterial system when oxidizing phenylalanine. Seven different batches of membrane particles were prepared and used independently for this estimation, with all the results falling within the same range : about 110 nanomoles (-+ 35) of peroxide were produced per minute and per mg total protein. The average specific activity for the conversion of phenylalanine to phenylpyruvate under the same conditions was 3:5 ,umoles min-lmg -1. Therefore up t o 3 per cent of the electron flow from the substrate seemed to contribute to the formation of H2Oz instead of water. A somewhat lower value, in the range of 30 nanomoles min -1 mg -1, was obtained when using, instead of the fluorescent assay, the anodic oxidation of peroxide on a platinum electrode (see Methods). This was slightly less than 1 per cent of the electron flow compared to the higher value above, and this discrepancy will be discussed. Figure 2 indicates that the yield of H20.~ was optimal at pH 7.5. T

40" E

'7"

~ 30" dID E 20" 0 c C

r-"~ 10, OI

Since z-phenylalanine was converted to phenylpyruvate at a high rate under the conditions used (see below), control experiments were made, indicating that the ketoacid formed did not interfere significantly with the fluorescence values. Other aminoacids were used as substitutes for phenylalanine in the above experiments, z-methionine and z-leucine were actively oxidized to the corresponding ketoacids by P. mirabilis membranes, and were found to promote L D A D C F oxidation. L-serine, L-threonine and L-valine were not substrates for oxidation by the membranes, and had no effect on the LDADCF-peroxidase system. It was thus concluded that some hydrogen peroxide was liberated during the membrane-catalyzed aminoacid oxidation, and L-phenylalanine was used as the substrate for the following work. The effect of L-phenylalanine on the rise of fluorescence showed the ordinary michaelian type of dependence, with a Km coefficient of 1.7 mM. A standard curve was built using known amounts BIOCH1MIE, 1982, 64, n ° 10.

o

04 |

6.8

|

I

|

7.2

|

7.6

8.0

PH

Fro. 2. - - Ellect of pH on the generation o] H,O~ by membrane particles in the presence o/L-phenylalanine. Peroxide formation is estimated using the platinum electrode, as described in Material and Methods.

Table I shows that the L D A D C F oxidation mediated by the membrane aminoacid oxidase was only partially inhibited by catalase. High amounts of bovine catalase were necessary to get approximatively 90 per cent inhibition. Thus it became very likely that the low amounts of contaminating catalase, usually present in the membrane suspension, interfered very little with the results This was easily understandable since most of the H:~.O2 molecules were trapped efficiently by the peroxidase in the assay mixture.

894

G. Sauret-Ignazi and coll.

Figure 3 shows that small amounts of superoxide dismutase (SOD) partially inhibit the L D A D C F fluorescence assay. Inhibition did not

oxide could rise from the autooxidation of transient intermediates generated from the dye by peroxidase.

TABLE I. E//ect o/catalase on L D A D C F oxidation by peroxidase and phenylalanine oxidizinz inner membranes.

Total

Activity Ratio H._,O_~equivalents of added bovine of added catalase as measured membrane catalase /endogenous by LDADCF protein (~g) (units/mi) (a) catalase activity oxidation (c) (b) (nmol/min/mg) 47.5 47.5 9.5 9.5 2.0 2.0

0 1945 0 1945

0 172 0 860

4.5 3 0.9 0.2

0

0

0.02

1945

4322

Percentage of inhibition by bovine catalase (per cent) 0 35 0 78 0

0.025

88

(a) Catalase units are defined as in ref. [6]. (b) Endogenous catalase activity of the membrane suspension 238 units/rag. (c) Assuming that 1 mole of LDADCF is oxidized for each mole of peroxide produced.

exceed 50-60 per cent even with high quantities of this enzyme. In view of the known specificity of SOD for 0-2 radicals [24], it thus appeared that a significant proportion of fluorescence of the dye was due to superoxide radicals that were liberated, either during phenylalanine oxidation by the membrane particles, or during the reaction with peroxidase. The latter enzyme is known to have a high affinity for superoxide, and the complex thus formed is extremely efficient for the oxidation of luminol [24]. In the presence of peroxidase, L D A D C F was readily oxidized afted addition of xanthine and xanthine oxidase, a commonly used superoxide generating system [10, 20]. The rise of fluorescence was severely inhibited (more than 80 per cent) b~¢ 2 I~g/ml SOD, and it became evident that the L D A D C F fluorescence assay of Keston and Brandt [14] with peroxidase was not specific for H202, and could be used for the detection of 0-2 as well. The oxidation of various substrates by 0-2 or H202 catalyzed by peroxidase is believed to generate organic free radicals that are readily further oxidized by dioxygen, yielding in turn superoxide. If the latter can react with more substrate or radicals derived from it, propagation of the oxidation will occur according to a SOD-sensitive process [25]. Thus any SOD effect on the peroxidase - dependent oxidation of L D A D C F could not be considered as a definite proof of a superoxide producing event during phenylalanine oxidation by membranes. Some superBIOCH1MIE, 1982, 64, n ° 10.

Two arguments were in favor of this hypothesis. The first came from the fluorescence assay itself, by measuring the molar stoechiometric ratio of the oxidized L D A D C F to the amount of peroxide

100

500 .,,e,

I

0.1 S.O.D

I

I

0.2 0.3 (pg.ml'l)

FIG. 3. - - E[lect o/ SOD on LDADCF oxidation by horseradish peroxidase in the presence of membrane particles and L-phenylalanine (protein 30 ~g). For experimental conditions see Material and Methods.

895

Hydrogen peroxide and the bacterial membrane.

added. Known amounts of peroxide were supplied to the mixture in the absence of membrane particles, and the fluorescence was allowed to increase until it levelled off, when all H~.O~ was used up. In all cases this ratio was found to be much higher than one, in the range of 5 to 7. The second observation came from the use of the platinum electrode : with phenylalanine and membrane particles present, an excess of SOD had no significant effect on the electrode response. Thus it was clear that peroxide was actually generated by the membranes, and was not entirely the product of superoxide dismutation. However the simultaneous occurrence of some 0-2 during phenylalanine oxidation could not be ruled out. In order to check this point we had to use an other system of detection. Epinephrine (adrenaline) oxidation to adrenochrome by O-2 offers a sensitive assay and requires no peroxidase [20]. The conditions and equipment used were those of Henry and Vignais [8]. Careful calibration of this system with xanthine and xanthine oxidase showed that we should have been able to detect less than 2 nanomoles of superoxide per minute in the presence of membranes at a concentration of 0.5 g of total protein per ml. The result was entirely negative. We concluded that superoxide was either absent or scavenged very actively by the membrane particles themselves. Hydrogen peroxide is formed during the oxidation of N A D H by cytochrome o purified from Vitreoscilla [12, 13]. A cytochrome o is also present as a terminal oxidase in the case of Proteus [15]. Oxidative deamination of phenylalanine by P. mirabilis is catalyzed by a membrane dehydrogenase, and electrons are channelled from it to oxygen via ubiquinone and the cytochromes ; the dehydrogenase can use D C I P as an electron acceptor [15, 16]. The whole process corresponded to the so-called <>. It was severely inhibited by Triton X100, although a high proportion of the dehydrogenase activity was preserved, as indicated by D C I P reduction (fig. 4). As previously shown, phenylalanine dehydrogenase was released under soluble form by the detergent with subsequent blockage of cytochrome reduction [16]. Triton X100-treated membranes were still able to promote L D A D C F oxidation (fig. 4). The addition of ubiquinone Q6 to these membranes restored partially the phenylalanine oxidase activity [16], with an increase in the L D A D C F oxidation response. The latter effect could be explained by the auto-oxidation of a reduced form of the quinone. However since t h e respiratory flow of electrons to oxygen was p a r -

BIOCHIMIE, 1982, 64, n ° 10.

tially restored by ubiquinone, oxygen by-products could also be evolved at a later stage during electron transport.

,oo-[ >

\

I...-

~.,,,

>

",~. A

o"¢501 ~

='"'"= B

lo]

c i

I

i

!

2

4

6

8

TRITON XlO0/PROTEIN (W/W) FIG. 4. - - E]]ects of Triton X 100 on enzymatic activities.

A (0 i ) L-phenylalanine-dependent LDADCF oxidation. Inner membranes : 40 txg protein. B (ll . . . . II) L-phenylalanine DCIP reductase. The reaction mixture contained per ml = L-phenylalanine : 50 Ixmoles; Tris-maleate : 20 I~moles (pH 7.5), MgCb_: 10 ~moles; DCIP = 65 nmoles; PMS: 0,82 nmole ; inner membranes: 15 txg protein. Incubation at 30°C. Although optimal reduction of DCIP in this system requires cyanide (15), this experiment was carried out without this compound, since cyanide would block peroxidase in part A, and the phenylalanine oxidase reaction in part C. C (D D) L-phenylalanine oxidase. The reaction mixture contain per ml: L-phenylalanine, 50 ~moles ; Tris-maleate = 20 ~moles (pH 7.5), MgCb : 10 ~moles ; inner membranes: 8 ~g protein. Incubation at 30°C. t00 per cent inner membranes enzymatic activity represent. - - L-phenylalanine oxidase reaction : 1,2 ~mole phenyl-

pyruvate.mg-X.min-1. - - L-phenylalanine DCIP reductase : 0,3 ~mole reduced DCIP.mg-l.min-1. - - LDADCF oxidation = 14 nmoles H~O,.mg-l.min -1.

Discussion. W e h a v e s h o w n that s o m e p e r o x i d e is f o r m e d during the bacterial oxidation of L-phenylalanine.

896

G. Sauret-lgnazi and coll.

The results obtained with the fluorescence assay suggested that about 3 per cent of the electrons from the substrate were derived to the generation of H2Oe instead of water. Using a platinum electrode, a lower value, 1 per cent at the most~ was obtained. This discrepancy may be due to the lack of precision of the calibrations with known amounts of peroxide, especially when measuring the fluorescence of LDADCF, as discussed below. The assay of Keston and Brandt [14] is not specific for peroxide but could be used for the detection of superoxide as well. Superoxide is trapped efficiently by peroxidase, leading to compound III of this enzyme. The latter may oxidize organic substrates of peroxidase or break down to O-2 and H202 [24, 26, 27]. If AH2 is such a substrate, it will be oxidized by peroxidase, according to current knowledge, to AH- radicals. The fully oxidized compound will be generated by : AH" + AH" m A + AH2 (1) AH" + 02 ~ A + 0-2 + H + (2) When occurring at a sufficiently high rate, reaction (2) will produce superoxide, leading to the generation of new AH" radicals by peroxidase and propagating the oxidation of more substrate. Such an amplification of the peroxidatic reaction in the presence of free oxygen has been reported in several cases (see for instance reference [24]). The extent of t h e amplification phenomenon should depend on the steady-state balance between compound I of peroxidase (obtained with H202) and compound II. One can suspect that adding known amounts of peroxide at one time for calibrating the system may favor compound I and lead to less amplification. For this reason we consider that the level of peroxide revealed by the platinum electrode (25-30 nmoles min-lmg-1), corresponding to about 1 per cent of the electron flow to oxygen, may be closer to reality. Efforts to demonstrate the generation of superoxide by P. mirabilis membranes when oxidizing phenylalanine have been inconclusive. Dismutation of these radicals could not offer a simple explanation for this negative result, because the membrane particles used were essentially devoid of SOD activity. Apparently superoxide could not be produced at a rate higher than a few nanomoles min-lmg-~, that is to say a much lower value than found for peroxide. We propose as a hypothesis that oxygen by-products appearing during phenylalanine oxidation by Proteus are mainly under the form of hydrogen peroxide. The membrane phenylalanine dehydrogenase may be the source of peroxide, since the reduced form of v~ious flavoenzymes are well known to react BIOCH1MIE, 1982, 64, n ° 10.

with molecular oxygen according to a two-electron process, yielding H2Oe. Alternatively semiquinone radicals appear to react reversibly with oxygen, giving rise to superoxide : semiquinone + 02 ~ quinone + 0-2 (3) There is evidence that this equilibrium strongly favors the semiquinones of natural origin [28]. An extremely low rate of superoxide production within the bacterial membrane may be achieved in two ways. The high reaction rate of semiquinone with another species of the respiratory chain may keep its steady-state level at a very low value. The proper shielding of these radicals may be attained in the bacterial membrane, minimizing the risk of superoxide production.

Acknowledgements. We thank Dr. Henry and Dr. Vignais (Grenoble) ]or their help in using the epinephrin assay, Dr. Gautheron (Lyon) for the use o] the platinum electrode. We thank also Dr. Ricard (Marseille) [or fruit]ul discussions about the properties of peroxidase. This work has received a financial support ]rom CNRS (ERA n ° 673).

REFERENCES. 1. Loschen, G., Flohe, L. & Chance, B. (1971) FEBS Letters, 18, 261-264. 2. Boveris, A., Oshino, N. & Chance, B. (t972) Biochem. J., 128, 617-630. 3. B~veris, A. & Chance, B. (1973) Biochem. J., 134, 702-716. 4. Chance, B., Sies, H. ~: Boveris, A. (1979) Physiol. Rev., 59, 527-605. 5. Dionisi, O., Galectti, T., Terranova, T. & Azzi, A. (1975) Bioehim. Biophys. Acta, 403, 292-301. 6. Turrens, J. F. & Boveris, A. (1980) Biochem. J,, 191, 421-427. 7. Nohl, H. & Jordan, W. (1980) Europ. J. Biochem., 111, 203-210. 8. Henry, M. F. & Vignais, P. M. (1980) Arch. Bioehem. Biophys., 203, 365-371. 9. Hassan, H. M. & Fridovich, I. (1979) Arch. Biochem. Biophys., 196, 385-395. 10. Fridovich, I. (1978) Photochem. Photobiology, 28, 733-741. 11. Halliwell, B. (1981)Bull. Europ. Physiopath. Res., 17 (Suppl.) 21-28. 12. Webster, D. A. (1975) J. Biol. Chem., 250, 4955-4958. 13. Tyree, B. & Webster, D. A. (1979) I. Biol. Chem., 254, 176-179: 14. Keston, A. S. & Brandt, R. (1965) Anal. Biochem., 11, 1-5. 15. Pelmont, J., Arlaud, G. & Rossat, A. M. (1972) Biochimie, 54, 1359-1374.

H y d r o g e n peroxide and the bacterial m e m b r a n e . 16. Pelmont, J., Jabbour, J. & Laboure, A. M. (1978) Biochimie, 60, 817-821. 17. Laboure, A. M., Manson, C., Jouve, H. & Pelrnont, J. (1979) J. Bact., 137, 161-168. 18. Brandt, R. 8~ Keston, A. S. (1965) Anal. Biochem., 11, 6-10. 19. Jouve, H., Sauret, G., Laboure, A. M. & Pelmont, J. (1979) Canad. J. Microbiol., 25, 302-311. 20. Mc Cord, J. M., Beauchamp, C. O., Goscin., S., Misra, H. P. & Fridovich, I. (1973) in <> ed. by King. T. E., Mason, H. S. and Morrison, University Park Press, 51-76. 21. Thevenot, D. R., Coulet, P. R. & Gautheron, D. C. (1978) in <> vol. 4 ed. by Brown, G. B., Manecke, G. & Wingard, L. B., 219-221.

BIOCH1MIE, 1982, 64, n ° 10.

897

22. Hasin, M., Rottem, S. ~ Razin, S. (1975) Biochim. Biophys. Acta, 375, 381-394. 23. Bradford, M. M. (1976) Anal. Biochem., 72, 248-254. 24. Michelsan, A. M. (1977) in ~ Superoxide and Superoxide Dismutase ~ ed. by Michelson, A. M., Mc Cord, J. M. ~ Fridovich, I., Acad. Press., 87106. 25. Halliwell, B. ~ De Rycker, J. (1978) Photochem. PhotobioL, 28, 757-763. 26. Sawada, Y. & Yamazaki, I. (1973) Biochim. Biophys. Acta, 327, 257-265. 27. Rotilio, G., Falcioni, G., Fioretti, E. & Brunori, M. (1975) Biochem. J., 145, 405-407. 28. Winterbourn, C. C. (1981) Arch. Biochem. Biophys., 209, 159-167.