356 Biochimica et Biophysics Acta, 795 (1984) 356-362

Elsevier

BBA 51747

THE CRITICAL STEADY-STATE HYPOXIC CONDITIONS IN CARBON TETRACHLORIDE-INDUCED LIPID PEROXIDATION IN RAT LIVER MICROSOMES

THOMAS NOLL and HERBERT DE GROOT

Institut ftir Physiologische Chemie I der Universitiit Diisseldorf; Moorenstrasse 5, D-4000 Diisseldorf (F. R.G.)

(Received January 31st, 1984)

Key worak Oxygen partial pressure; Hypoxia; Carbon tetrachloride; Lipid peroxidation; (Rat liver microsome)

Defined steady-state oxygen partial pressures (po,) were maintained constant with an oxystat system to study carbon tetrachloride (Ccl,)-induced lipid peroxidation and oxygen uptake in rat liver microsomes. The initial rates of oxygen uptake and malondialdehyde formation indicated drastically increasing lipid peroxida- tion by decreasing po,, attaining a maximum between l-10 mmHg (0.1-1.3 kPa). Under these conditions, at the hypoxic end of the physiological po, in liver, Ccl, caused a 5-fold increase in the oxygen uptake rate and a 20-fold increase in the malondialdehyde formation rate while, at 80 mmHg (10.7 kPa) the haloalkane caused only an increase of 2- and 4-fold, respectively; in comparison, there was only a slight increase in NADPH-induced lipid peroxidation with increasing po,. These data clearly demonstrate the critical role of low steady-state po, in Ccl,-induced lipid peroxidation and support lipid peroxidation as a key factor in Ccl, hepatotoxicity.

Introduction

Lipid peroxidation mediated by free radicals is considered to be a primary mechanism of cell membrane destruction and cell damage [l]. One of the most thoroughly investigated examples is the lipid peroxidation stimulated by the model hepa- totoxin carbon tetrachloride (Ccl,) [2-41. In this case, lipid peroxidation is initiated by the interac- tion of reactive, free ‘Ccl, (or, following addition of dioxygen, ‘CCl,O,) radicals with unsaturated fatty acids of membrane lipids [5,6]. The ‘Ccl, radicals are reductively formed by cytochrome P-

450 (P-450), especially those isoenzymes induced by phenobarbital [7,8]. Since this metabolic activa- tion takes place at the haem moiety of P-450, the physiological binding site of dioxygen during the monooxygenase cycle, the formation of ‘Ccl, radi- cals is inhibited by dioxygen [9]. Since, on the other hand, the propagation reaction of lipid per- oxidation necessarily depends on dioxygen, a com-

plex oxygen dependence of Ccl,-induced lipid peroxidation is to be expected.

In line with these considerations, a favourable effect of hypoxia on Ccl,-induced lipid peroxida- tion has been proposed recently [lo]. However, experiments in the literature on the oxygen depen- dence of Ccl,-induced lipid peroxidation gave contradictory results. In vivo experiments pro- vided evidence that, in comparison with hyperoxia (100% 0,), normoxia (20% 0,) [ll] and, even more so, hypoxia (10% 0,) [12] increase CCl,-in- duced lipid peroxidation as indicated by an en- hanced exhalation of ethane. On the other hand, from in vitro experiments [ll] which demonstrated continuous decline of Ccl,-induced lipid per- oxidation from a maximum at a po, of 50 mmHg (6.7 kPa) down to a po, of 14 mmHg (1.9 kPa), the lowest po, studied, one can conclude that CC1 ,-induced lipid peroxidation should preferen- tially proceed under normoxia or even slight hy- peroxia. Under normoxia the po, in the liver

0005.2760/84/$03.00 0 1984 Elsevier Science Publishers B.V.

351

varies between l-56 mmHg (0.1-7.5 kPa) with the lowest values centrolobular [13,14].

Because of these discrepancies we investigate here the Ccl,-induced lipid peroxidation with spe- cial emphasis on low physiological po, under de- fined steady-state conditions. In analogy with the results from experiments with another haloalkane, the anaesthetic halothane [15], we expected a marked increase of Ccl,-induced lipid peroxida- tion under hypoxic conditions. To undertake these experiments, an incubation system was designed which allows not only the maintainance of a desired po2, but also the measurement of the rate of oxygen uptake on line.

Materials and Methods

Microsomes. Microsomes were prepared from male Wistar rats (180-200 g), fed on Altromin stock diet (Lage/Lippe, F.R.G.), as described pre- viously [16,17]. Sodium phenobarbital was dis- solved in drinking water (1 g/l) for 4 days prior to death. Isolated microsomal fractions were stored at 0°C and used within l-4 h following prepara- tion.

Incubations. The incubation mixture consisted of 6 mM MgClJ104 mM KC1/50 mM Tris-HCl buffer (pH 7.4), microsomes (1.5-3.0 mg protein/ml), isocitrate (10 mM), isocitrate dehy- drogenase (EC 1.1.1.42; 300 U/l) and NADP (1 mM). The NADPH-regenerating system ensured a NADPH concentration of about 0.7 mM during the entire incubation period. Where indicated, Ccl, (final concentration 1 mM) was added as a solution of Ccl,-saturated (5.2 mM, [18]) 6 mM MgC1,/104 mM KC1/50 mM Tris-HCl buffer (pH 7.4). It was prepared as follows: 1 ml Ccl, was stirred with 250 ml deoxygenated buffer. To replace residual oxygen, the buffer phase was slowly flushed with argon. After 10 min the bottle was stoppered and stirring was continued for a further 4 h. Afterwards, stirring was stopped to separate the undissolved Ccl, from the buffer phase.

The NADPH-regenerating system was com- posed to the exclusion of oxygen and stored in gas-tight syringes until use. To achieve the desired po2, the components of the incubation system were mixed in gas-tight syringes with adequate amounts

of argon- and oxygen-saturated Tris buffer. Steady-state oxygen conditions were maintained

by an oxystat system. The incubations were per- formed at 37 o C in a closed, thermostated incuba- tion vessel equipped with inlet and outlet. The po, of the magnetically stirred incubation medium, which completely filled the reaction chamber, was continuously measured by a Clark-type oxygen sensor with low response time (Eschweiler, Kiel, F.R.G.). The output of the oxygen sensor operated an automatic control unit (Radiometer, Copenha- gen, Denmark) used in conjunction with a motor- driven piston burette (Radiometer, Copenhagen, Denmark) which added oxygen-saturated buffer (1.2 mM 0, at 22OC, calculated according to Forstner and Gnaiger [19]) via the inlet into the incubation mixture, maintaining po, at prede- termined levels. The displaced incubation medium was fractionated into cups at the outlet and used for further determinations. Oxygen uptake was calculated from the amounts of oxygen-saturated buffer added and was related to microsomal pro- tein. This feedback control system maintained steady-state po, between 7-80 mmHg (0.9-10.7 kPa) with a signal-to-noise ratio of about 50, and below 7 mmHg (0.9 kPa) with a signal-to-noise ratio of 5. The drift of the oxygen sensor, caused by Ccl,,, was compensated by a slaving system.

Because of the addition of oxygen-saturated buffer, microsomes were continuously diluted dur- ing incubation, maximally from 3 mg microsomal protein/ml to about 1.5 mg microsomal protein/ml. To prevent a similar effect on the components of the NADPH-regenerating system and on Ccl,, the oxygen-saturated buffer con- tained all substances, added in the same con- centrations as mentioned above after oxygen equi- librium.

Assays. Malondialdehyde was assayed as de- scribed by Reiner et al. [20] using 1,1,3,3-tetra- methoxypropane as a standard. Cytochrome P-450 was determined as ferrous CO complex [21]. NADPH concentration was measured according to Klingenberg [22]. Protein was determined with the method of Lowry et al. [23] using bovine serum albumin as a standard.

Materials. Materials were obtained from the following sources: Ccl,, isocitrate, thiobarbituric acid from Merck (Darmstadt, F.R.G.); isocitrate

358

dehydrogenase (EC 1.1.1.42) and NADP from Boehringer (Mannheim, F.R.G.); 1,1,3,3-tetra- methoxypropane from Fluka (Neu-Ulm, F.R.G.); bovine serum albumin from Behring (Marburg, F.R.G.); argon and oxygen from Linde (Dtissel- dorf, F.R.G.).

Results

Rat liver microsomes from phenobarbital-pre- treated male rats were incubated at steady-state po, using an oxystat system which maintained defined oxygen conditions by injecting oxygen- saturated buffer into the incubation medium. Oxygen uptake was measured continuously and accumulation of malondialdehyde was detected at discrete intervals.

Initial experiments with different protein con- centrations revealed that oxygen uptake and malondialdehyde formation were linearly related to microsomal protein concentration and that the system-dependent decrease of protein concentra- tion therefore had no effect on oxygen uptake and malondialdehyde formation (data not shown).

Fig. 1. Simultaneous determination of oxygen uptake and malondialdehyde (MDA) formation during an incubation of microsomes at a constant po, of 1 mmHg (0.1 kPa). The incubation system consisted of liver microsomes (1.5-3.0 mg protein/ml) from phenobarbital-pretreated male rats, and 6 mM MgC1,/104 mM KC1/50 mM Tris-HC1 buffer (pH 7.4). NADPH (about 0.7 mM, regenerating system) and Ccl, (1 mM) were added where indicated. Steady-state po, was main- tained constant throughout the experimental period by addition of oxygen-saturated buffer, using a feedback control system.

(- ) oxygen uptake, (O- 0) malondialdehyde forma- tion.

Fig. 1 shows a registration of an experiment at a steady-state po, of 1 mmHg (0.1 kPa). In the absence of NADPH, neither oxygen uptake nor malondialdehyde formation was detected. The ad- dition of NADPH resulted in an oxygen uptake, but was without significant effect on the formation of malondialdehyde. Large amounts of malondialdehyde, however, were formed following further supplementation of the NADPH-reduced microsomes with Ccl,, and under this condition a marked further increase of oxygen uptake oc- curred (Fig. 1). The addition of Ccl, to micro- somes in the absence of NADPH did not effect either oxygen uptake or malondialdehyde forma- tion (data not shown).

The time-courses of oxygen uptake and malondialdehyde formation rates induced by NADPH and NADPH/CCl, at a po, of 1 and 80 mmHg (0.1 and 10.7 kPa), as well as the time- course of the concomitant P-450 content, are shown in Figs. 2-4. At a po, of 1 mmHg (0.1

, -

i \

, 1

0 10 20 30

Incubotmn he /mm1

Fig. 2. Effect of Ccl, on the oxygen uptake rate during the incubation of NADPH-reduced liver microsomes at constant po, of 1 and 80 mmHg (0.1 and 10.7 kPa). Oxygen uptake rate in presence of 0.7 mM NADPH (open symbols) and in pres- ence of 0.7 mM NADPH/l mM CCI, (filled symbols); 0, l , incubations at a po, of 1 mmHg (0.1 kPa); 0, n , incubations at a po, of 80 mmHg (10.7 kPa). Vertical bars denote SE. of the mean for at least five separate incubations. Further experimen- tal details are as in Fig. 1.

359

\

i \

I I

0 10 20 30

Incubatmn tune lminl

Fig. 3. Effect of CCI, on the malondialdehyde (MDA) forma- tion rate during the incubation of NADPH-reduced liver mi- crosomes at constant po, of 1 and 80 mmHg (0.1 and 10.7 kPa). Malondialdehyde formation rate in presence of 0.7 mM NADPH (open symbols) and in presence of 0.7 mM NADPH/l mM Ccl, (filled symbols); 0, 0, incubations at a po, of 1 mmHg (0.1 kPa); q , W, incubations at a po2 of 80 mmHg (10.7 kPa). Vertical bars denote S.E. of the mean for at least five separate incubations. Further experimental details are as in Fig. 1.

kPa), the NADPH-induced oxygen uptake rate remained fairly constant throughout the incuba- tion period (Fig. 2), while no malondialdehyde formation (Fig. 3) and only a slight inactivation of P-450 (Fig. 4) could be detected. In marked con- trast, NADPH/CCl, greatly stimulated the initial rates of oxygen uptake and malondialdehyde for- mation (Figs. 2 and 3). However, in the further incubation period both rates asymptotically sub- sided to the values in the absence of Ccl,. Under these conditions, P-450 was almost completely inactivated after about 30 min (Fig. 4). Raising the setpoint of the steady-state po, to 80 mmHg (10.7 kPa) resulted in a significant increase in the NADPH-induced oxygen uptake rate (Fig. 2). Moreover, at this higher po, significant formation of malondialdehyde occurred (Fig. 3). Further supplementation of the NADPH-reduced micro- somes with CCI, again stimulated both oxygen uptake and malondialdehyde formation (Figs. 2

0 10 20 30

Incubohon hme lmrnl

Fig. 4. Decrease of cytochrome P-450 content during incuba- tion of microsomes in presence of 0.7 mM NADPH (open symbols) and in presence of 0.7 mM NADPH/l mM CCI, (filled symbols) at constant po, of 1 mmHg (0.1 kPa) 0,. and 80 mmHg (10.7 kPa) 0, n . Vertical bars denote S.E. of the mean for at least five separate incubations. Further experimen- tal details are as in Fig. 1.

and 3). However, compared to the related values at 1 mmHg (0.1 kPa), the initial rates as well as the absolute amounts of oxygen uptake and malondialdehyde formation (data not shown) were significantly lower and P-450 was inactivated more slowly (Fig. 4). Nevertheless, independent of the chosen po,, P-450 was almost completely in- activated after 30 min.

In Figs. 5 and 6 the oxygen dependence of the initial and the 30-min rates of oxygen uptake and malondialdehyde formation are shown. It can be seen that the initial rates of NADPH-induced oxygen uptake and malondialdehyde formation augmented with increasing po,. In marked con- trast, the initial rates of NADPH/CCl,-induced oxygen uptake and malondialdehyde formation were greatly increased with decreasing po, reach- ing a maximum between l-10 mmHg (0.1-1.3 kPa). At the maximum, Ccl, caused a 5-fold increase in the oxygen uptake rate and a 20-fold increase in the malondialdehyde formation rate, while at 80 mmHg (10.7 kPa) Ccl, only caused

360

60

0 LO 80

Oxygen porm pressure lmm FigI

Fig. 5. Effect of increasing po, on the initial and 30-min rate of the NADPH- (open symbols) and the NADPH/CCl,-induced oxygen uptake (filled symbols) of liver microsomes. 0, l , initial rates; 0, n , 30-min rates. Vertical bars denote S.E. of the mean for at least five separate incubations. Further experimen- tal details are as in Fig. 1.

increases of 2- and 4-fold, respectively (Table I).

The 30-min rates of NADPH-induced oxygen up- take and malondialdehyde formation showed

nearly the same oxygen dependences as the initial

\ \T

0 LO 60

Oxygen porlml pressure lmm Hgi

Fig. 6. Effect of increasing Po, on the initial and 30-min rate of the NADPH- (open symbols) and the NADPH/CCl,-induced malondialdehyde (MDA) formation (filled symbols) of liver microsomes. 0, 0, initial rates; 0, n , 30-min rates. Vertical bars denote S.E. of the mean for at least five separate incuba- tions. Further experimental details are as in Fig. 1.

rates (Figs. 5 and 6). In the presence of NADPH/CCl,, the 30-min rates of both oxygen uptake and malondialdehyde formation increased continuously with increasing po, and, in the case of the malondialdehyde formation, only a slight

stimulation could be observed compared to the

TABLE I

STIMULATING EFFECT OF Ccl, ON THE INITIAL RATES OF OXYGEN UPTAKE AND MALONDIALDEHYDE FORMATION IN NADPH-REDUCED MICROSOMES INCUBATED AT VARIOUS STEADY-STATE po,

The stimulating effect of Ccl, on oxygen uptake and malondialdehyde (MDA) formation (expressed as nmol.rnin-“mg microsomal ’ . protein- ) IS m dicated as the difference between the NADPH/CCl,- and the NADPH-induced rates (Figs. 5 and 6). The relative

increase of oxygen uptake and malondialdehyde formation represents the ratio of oxygen uptake and malondialdehyde formation in presence of NADPH/CCl, to the related values in presence of NADPH alone.

po, Ccl,-stimulated Ccl,-stimulated relative increase of Ratio of

WnHg) oxygen uptake MDA formation oxygen uptake MDA formation oxygen uptake: MDA formation

Anaerobic none none none none _

1 46k6 2.0 f 0.2 5 wa 23

5 44k6 1.9 f 0.2 5 20 23

10 32k7 1.6 f 0.3 3 17 20

20 21k3 1.1 io.2 2 9 19 40 18k4 0.9 f 0.2 2 6 20

80 15f4 0.7 f 0.3 2 4 21

’ Infinite value since, in the absence of Ccl, malondialdehyde formation is below detection limit (Figs. 2 and 6).

361

values in the absence of Ccl,. As would be ex- pected, under anaerobic conditions no develop- ment of malondialdehyde was detected (Table I).

Discussion

Malondialdehyde formation at defined steady- state po, indicates significant increase of CCl,-in- duced lipid peroxidation by decreasing the po, (Fig. 6) with a maximum between 1 and 10 mmHg (0.1-1.3 kPa). This complex oxygen dependence can be explained considering that oxygen plays opposing roles in the initiation and propagation step of Ccl,-induced lipid peroxidation. In the initiation step, oxygen competes with Ccl, at the haem moiety of P-450. Consequently, the forma- tion of Ccl, metabolites increases with decreasing po, [24]. In the propagation step, however, the formation of fatty acid peroxides from fatty acid radicals increases with increasing po,. Hence, the complete sequence of Ccl,-induced lipid per- oxidation should proceed maximally at a po, low enough to permit formation of ‘Ccl, radicals at almost maximal rate, but high enough to promote rapid formation of lipid peroxides.

A further explanation for the observed oxygen dependence of Ccl,-induced lipid peroxidation is that increasing po, might contribute to an increas- ing shunt of electrons from the flavoprotein site of the microsomal electron transport chain to yield O;, which would result in a decrease in the availa- bility of reduction equivalents for Ccl, activation at the P-450 site. Moreover, increasing po, may change the Fe2+-to-Fe3+ ratio, which is considered to affect the breakdown of peroxides to malondialdehyde [3].

Previously, experiments performed under de- fined anaerobic conditions [16] demonstrated a Ccl,-induced inactivation of P-450 without in- volvement of lipid peroxidation, presumably due to a direct attack of Ccl, metabolites. Under those conditions, the inactivation of P-450 was always incomplete and never exceeded 70% after 30 min of incubation. In contrast, as shown in Fig. 4, under aerobic conditions, the inactivation of P-450 caused by Ccl, was almost complete. Since it is known that P-450 can also be inactivated by lipid peroxidation [25], the Ccl,-induced lipid per- oxidation is the most likely reason for the addi-

tional inactivation of P-450, a process which is slightly accelerated at low po, by the enhanced lipid peroxidation. Thus, our results give evidence that under aerobic conditions the inactivation of P-450 represents a combined effect of the direct attack of reactive Ccl, metabolites and the Ccl,- induced lipid peroxidation.

According to Kornbrust and Mavis [26] Ccl,- induced lipid peroxidation appears to be highly limited by the integrity of P-450. The same con- clusion can be drawn from our experiments de- picted in Figs. 3 and 4, since the effect of Ccl, on lipid peroxidation is restricted with progressing P-450 inactivation. This correlation is shown most clearly at low po, (Fig. 3) because under these conditions the NADPH-induced lipid peroxida- tion is negligible. The low affinity towards oxygen of the latter sequence points to a lipid peroxida- tion caused by traces of iron which are con- taminating the microsomal fractions or which are present inadvertently as an impurity in the added reagents.

Raising the setpoint of po,, the Ccl,-induced lipid peroxidation is increasingly superimposed by the NADPH-induced lipid peroxidation, since the further sequence decreases, while the latter gains significance under high po,. Nevertheless, the necessity of an intact P-450 for Ccl,-induced lipid peroxidation can also be observed under higher po, when the NADPH-induced lipid per- oxidation is subtracted from the one induced by NADPH/CCl, (Fig. 3, Table I).

On line determination of oxygen uptake under defined steady-state oxygen conditions demon- strates a close relationship between CCl,-stimu- lated lipid peroxidation and oxygen uptake, as indicated by the time-course (Figs. 2 and 3) and the oxygen dependence (Figs. 5 and 6, Table I) of both parameters, the rates of which show a maxi- mum at the same p. range. Furthermore, the molar ratios of CCl,-induced oxygen uptake to malondialdehyde formation (Table I) reveal that, in the initial phase of incubation, this ratio is about 20, independent of the chosen po,. In agree- ment with these results, an oxygen-to- malondialdehyde ratio of 20 has already been re- ported for lipid peroxidation [27].

These data demonstrate the critical role of hy- poxia in Ccl,-induced lipid peroxidation and pro-

362

vide evidence that the propagating effect of hypo-

xia on Ccl, hepatotoxicity [28,29] may result from a drastically increased lipid peroxidation.

For the anaesthetic halothane (CFsCHBrCl) which, in contrast to Ccl,, is normally oxidatively

detoxificated by P-450 [15], we recently presented evidence that, under hypoxia, this haloalkane is

also capable of inducing lipid peroxidation. For

this reason, it may be expected that the above- described apparently paradoxical phenomenon of

increasing lipid peroxidation with decreasing po, also holds for further substrates of P-450.

Acknowledgements

The authors wish to thank Professor H. Sies for his valuable suggestions during the preparation of this manuscript and Mr. P. Wissemann for provid-

ing useful advice and assistance in the problems regarding electronic circuits. This work was sup-

ported by Ministerium fur Wissenschaft und For- schung, Nordrhein-Westfalen.

References

1 McBrien, D.C.H. and Slater, T.F. (1982) Free Radicals,

Lipid Peroxidation and Cancer, Academic Press, London

2 Recknagel, R.O., Glende, E.A., Jr. and Hruszkewycz, A.M.

(1977) in Free Radicals in Biology (Pryor, W.A., ed.), Vol.

3, pp. 97-132, Academic Press, New York

3 Slater, T.F. (1978) Biochemical Mechanisms of Liver In-

jury, Academic Press, London

4 Reynolds, E.S. and Moslen, M.T. (1980) in Free Radicals in

Biology (Pryor, W.A.. ed.), Vol. 4, pp. 49-94, Academic

Press, London

5 Poyer, J.L., Floyd, R.A., McCay, P.B., Janzen, E.G. and

Davis, E.R. (1978) Biochim. Biophys. Acta 539. 402-409

6 Tomasi, A., Albano, E.. Lott, K.A.K. and Slater. T.F.

(1980) FEBS Lett. 122, 303-306

7 Noguchi, T.. Fong. K.-L., Lai, E.K., Olson, L. and McCay,

P.B. (1982) B&hem. Pharmacol. 31, 609-614

8

9

10

11

12

13

14

15

16

17

18

1

19

20

21

22

Forstner. H. and Gnaiger, E. (1983) in Polarographic

Oxygen Sensors (Gnaiger, E. and Forstner, H., eds.), pp.

321-333, Springer-Verlag, Berlin

Reiner, O., Athanassopoulos, S., Hellmer, K.H., Murray,

R.E. and Uehleke, H. (1972) Arch. Toxicol. 29, 219-233

Omura, T. and Sato, R. (1964) J. Biol. Chem. 239.2370-2378

Khngenberg, M. (1974) in Methoden der Enzymatischen

Analyse (Bergmeyer, H.U., ed.), pp. 2094-2108. Verlag

Chemie, Weinheim

23

24

25

Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall,

R.J. (1951) J. Biol. Chem. 193, 265-275

Wolf, C.R., Harrelson, W.G., Jr., Nastainczyk, W.M.,

Philpot, R.M., Kalyanaraman, B. and Mason, R.P. (1980)

Mol. Pharmacol. 18, 553-558

Levin, W., Lu, A.Y.H., Jacobson, M., Kuntzman, R., Poyer,

J.L. and McCay, P.B. (1973) Arch. B&hem. Biophys. 158,

842-852

26 Kornbrust, D.L. and Mavis, R.D. (1980) Mol. Pharmacol.

17, 408-414

27

28

Wills, E.D. (1969) B&hem. J. 113, 315-324

Strubeit, 0. and Breining, H. (1980) Toxicol. Lett. 6,

109-113

De Groot, H. and Nell, T. (1983) Hepatology 3, 601-606

De Groot, H. and Haas, W. (1980) FEBS Lett. 115.253-256

De Groot, H. and Haas, W. (1981) Biochem. Pharmacol. 30,

2343-2347

Ugazio, G. (1978) in Biochemical Mechanisms of Liver

fnjury (Slater, T.F., ed.), pp. 709-743, Academic Press,

London

29 Shen, E.S., Garry, V.F. and Anders, M.W. (1982) B&hem.

Pharmacol. 31, 3787-3793

Noguchi, T., Fong, K.-L., Lai, E.K.. Alexander, S.S., King,

M.M., Olson, L., Poyer, J.L. and McCay, P.B. (1982) Bio-

them. Pharmacol. 31, 615-624

Ahr, H.J., King, L.J., Nastainczyk, W. and Ulhich, V.

(1980) Biochem. Pharmacol. 29, 2855-2861

Kappus, H. and Sies, H. (1981) Experientia 37, 1233-1241

Kieczka, H. and Kappus, H. (1980) Toxicol. Lett. 5.191-196

Frank, H. and Diirk, H. (1983) Arch. Toxicol. 53, 213-223

Kessler, M., Lang, H., Sinagowitz, E., Rink, R. and Hoper,

J. (1973) in Oxygen Transport to Tissue, Part A (Bruley,

D.F. and Bicher, H.I., eds.), pp. 351-360, Plenum Press,

New York

Sies, H. (1977) Hoppe-Seylers 2. Physiol. Chem. 358, 1021-1032