Inactivation of red cell glutathione peroxidase by divicine and its relation to the hemolysis of favism

280

Biochimica et Biophysiea Aeta 847 (1985) 280 284 Elsevier

BBA 11591

I n a c t i v a t i o n o f red cell g l u t a t h i o n e p e r o x i d a s e by d i v i c i n e a n d its r e l a t i o n to t h e hemolysis of favism I r e n e M a v e l l i ~', M a r i a R o s a C i r i o l o b a n d G i u s e p p e R o t i l i o b,, " Institute of Applied Biochemistt~V and CNR ('enter of Molecular Biology, Uni~ersi(v of Rome "La Sapienza', Rome and h Department of Biology, 2nd Universi(v of Rome, Via O. Raimondo, 00173 Rome (ltalv) (Received April 24th, 1985)

Key words: Glutathione peroxidase; Divicine; Favism; Hemolysis; ( H u m a n erythrocyte)

A significant inactivation of red blood cell glutathione peroxidase (25% less than the physiological value) was observed after exposure of intact erythrocytes to 2 mM divicine (an autoxidizable aminophenol from Vicia faba seeds) and 2 mM ascorbate for 3 h at 37°C. Addition of catalase and conversion of Hb to the carbomonoxy derivative resulted in protection against enzyme inactivation. Oxidation of Hb was a concurrent phenomenon, and augmented the inactivating effect. In hemolysates, much stronger effects were observed at shorter times (2 h); divicine was effective also without ascorbate, and the presence of reductants (ascorbate or glutathione or NADPH) enhanced its inactivating power. Of the other antioxidant enzymes, superoxide dismutase was unaffected under the same experimental conditions. Catalase was found to be much less sensitive to the inactivation; it was almost unaffected in experiments with intact erythrocytes and specifically protected by N A D P H in experiments with hemolysates. This specific damage of glutathione peroxidase, apparently involving interaction of H202 and HbO2, may be related to the pathogenesis of hemolysis in favism.

Introduction Superoxide dismutase (superoxide:superoxide oxidoreductase, EC 1.15.1.1), glutathione peroxidase) glutathione:hydrogen-peroxide oxidoreductase, EC 1.11.1.9) and catalase (hydrogen-peroxide:hydrogen-peroxide oxidoreductase, EC 1.11.1.6), beside being independently active against potentially toxic oxygen radicals, may have a concerted role in the enzymatic antioxidative defense of the red blood cell. Superoxide dismutase is inactivated by hydrogen peroxide [1] and therefore is sensitive to lowered activity of H202-removing enzymes; catalase [2] and glutathione peroxidase [3] have been reported to be inhibited by 0 2 . It is likely that a critical balance of these enzymes is * To whom correspondence should be addressed.

instrumental to optimal enzymatic defense of the cell against reactive oxygen metabolites. A number of examples giving support to this hypothesis have been studied [4 7]. In particular, favism, a hemolytic disease occurring in some glucose-6-phosphate dehydrogenase (D-glucose-6-phosphate: N A D P 1-oxidoreductase, EC 1.1.1.49)-deficient individuals, shows an acute decrease of glutathione peroxidase during the hemolytic crisis [8]. It is not clear whether the drop of red blood cell glutathlone peroxidase is causative or consecutive to the crisis, or both. Pathogenesis of favic hemolysis has been related to the presence, in Viciafaba seeds, of certain potentially reactive molecules like vicine and convicine [9]. These compounds are glycosides that give rise, upon hydrolysis by fl-glucosidases, to the aglycones divicine and isouramil, which are susceptible to autoxidation processes [10] ulti-

0167-4889/85/$03.30 ~,) 1985 Elsevier Science Publishers B.V. (Biomedical Division)

281 H

H l OH glucosidase H

.2o/J

0;5

/

\A.,

INN

Fig. 1. Autoxidation of the aglycones derived from the pyrimidine glycosides of V. faba seeds Convicine and isouramil: R=OH; vicine and divicine: R = N H 2 ; AH 2 and A = reducing system.

mately forming H202 (Fig. 1). This reaction is potentially noxious to dehydrogenase-deficient red blood cells, which are less effective than normal cells in detoxifying H202 through the enzymatic system involving, in chain, glutathione reductase and the NADPH-producing-dehydrogenases of the pentose cycle. Which factor makes only certain dehydrogenase-deficient individuals sensitive to ingestion of V. faba is still unknown, Since the decrease of glutathione peroxidase is restricted to favic individuals during crisis [8] and seems therefore to be specifically linked to the hemolytic event, the effect of divicine on glutathione peroxidase of normal red cells in vitro was examined in a preliminary experiment [8] and was found to result in enzyme inactivation. A more detailed study of this process, concerning its possible mechanism, is the object of the present report. Experimental procedure Materials Vicine was obtained from Serva, Heidelberg; GSH and N A D P H were from Boehringer-Mannheim; L-ascorbic acid was from Fluka, Buchs,

Switzerland; superoxide dismutase was purified from bovine erythrocytes [11], catalase and/3-glucosidase were obtained from Sigma, St. Louis, MO. Preparation and treatment of samples Fresh anticoagulated blood with eparin was obtained from healthy adult volunteers. Packed red blood cells were washed three times with cold phosphate-buffered saline, care being taken in discarding the buffy coat at each washing. For the various experiments red cells were resuspended in phosphate-buffered saline containing 5 mM glucose (5% v/v), Hemolysates were prepared by sonication of 5% (v/v) suspensions. After each incubation excess reagent was removed either by repeated washings of red cells, or by exhaustive dialysis of hemolysates. The time of incubation was as to give a significant effect and was usually shorter for hemolysates than for intact erythrocytes. Conversion of the red blood cell Hb into CO-Hb or metHb 100% conversion of Hb into CO-Hb, as shown by lack of any spectral changes upon addition of sodium dithionite, was obtained by repeated deaeration-equilibration cycles of packed red blood cells in a Thumberg tube. The red cells were then diluted (5%, v / v ) with aerated buffer. Conversion of Hb into 80% metHb and 20% HbO 2 without byproduction of any degraded derivative [12] was obtained by incubating red cell suspensions (5%, v / v ) with 0.2 M K N O 2 in the presence of 5 mM glucose at room temperature for 20 min. Red blood cells were then washed with 300 vol. phosphate-buffered saline to remove both NO 2 and NO3-. No loss of superoxide dismutase , glutathione peroxidase or catalase was observed in the red cells after incubation with KNO 2 in the presence of glucose. Enzyme assay Glutathione peroxidase activity was assayed spectrophotometrically using either t-butylhydroperoxide [13] or H202 [14] as substrates. The extent of the changes observed was the same with either method. Catalase activity was measured by an ultraviolet method [15]. Both these enzyme

282

activities were assayed in a Perkin-Elmer model Lambda 3 spectrophotometer, and were calculated as units (/~mol substrate transformed per min per g Hb at 25°C). Superoxide dismutase was determined by a polarographic method [16], with an Amel (Milan, Italy) polarographic unit, Model 465. All enzyme activities were calculated with reference to Hb, which was determined according to Drabkin and Austin [17]. Results

T A B L E 11 EFFECT OF DIVICINE ON GLUTATH1ONE PERO X I D A S E A C T I V I T Y A N D O N metHb C O N T E N T O F H E MOLYZED HUMAN RED BLOOD CELLS

Hemolyzed red blood cells were incubated for 2 h at 3 7 o ( ` G S H and N A D P H concentrations were 2 m M . Concentrations of other reagents were the same as for "Fable I. Addition

MetHb

None Divicine (D)

Table I shows the effect of divicine on the glutathione peroxidase activity and metHb content of intact human red blood cells under various conditions. Divicine alone was ineffective. The presence of L-ascorbate in addition to divicine was necessary to obtain a substantial (at least 20%) decrease of glutathione peroxidase activity and increase of metHb content. N o change of superoxide dismutase and catalase activity was observed under the same conditions. Externally added catalase was effective in protecting against these effects, while superoxide dismutase was not. These results suggest that H 2 0 = was involved in the process and re-reduction of oxidized divicine by ascorbate [10] was essential. When the same experinaents were repeated on hemolysates (Table II), divicine alone was able to induce both metHb

D+GSH D+NADPH GSH NADPH L - A s c o r b a t e (A) D+A D+A+GSH D+A+ NADPH

(% total Hb)

Glutathione peroxidase (% activity)

Catalase (% activity)

4¥ 3 28 T 16 57T-21 21~ 5 2T 1 1T 1 2T 1 75T-28 95-T22 52 T- 24

100 58 ¥ 2 48T 2 44+ 1 95T 5 II0-T- 8 99g10 39+ 9 96-T- 14 58~13

100 55 T- 10 53¥ 9 105T15 99T14 99-Y 7 97-T-13 60+ 9 52-T- 7 114-T-26

increase and glutathione peroxidase decrease, although the effect was higher in the presence of L-ascorbate, GSH or N A D P H . The hemolysate experiments showed a significant decrease of the catalase activity as well, while again superoxide dismutase was insensitive. Full protection of glutathione peroxidase was obtained by addition of GSH to the incubation of hemolysates in the

TABLE I EFFECT OF DIVICINE BLOOD CELLS

ON GLUTATHIONE

PEROXIDASE

ACTIVITY

AND

O N metHb C O N T E N T

OF HUMAN

RED

Red cell suspensions (5%, v / v ) were incubated under gentle stirring for 3 h at 37°C. Divicine ( D ) - 2 m M v i c i n e + 11 U / r o t /3-glucosidase. L-Ascorbate concentration was 2 r a M . Catalase was 1 p.M. Superoxide dismutase was 2 p.M. Values are expressed as the m e a n s - T - S D ( n = 5). 100% glutathione peroxidase activity corresponds to 38.7_+ 8.4 units; this value is in accord with those reported by Beutler et al. [13]. 100% catalase activity corresponds to (38.9_+ 8.3). 104 units. Addition

None

Divicine(D) L-Ascorbate (A) D+A D + A + catalase

D + A + heat-inactivated catalase D + A + superoxide dismutase D + A + heat-inactivated superoxide dismutase D + A + catalase + superoxide dismutase

(% total Hb)

Met Hb

Glutathione peroxidase (% activity)

Catalase (% activity)

5 -T- 3 4+ 2 3 -Y- 1 44+20 17 -7- 1 50 + 4 53 T 10 47 -T-15 16 + I

100 93+1 97 T- 3 72+2 90 -T- 1 72 T 1 78 T- 2 75 T- 2 93 T- 1

100 100=6 97 7 5 90~8 98 + 1 95 + 4 97 T 6 97 + 5

104 T l

283 presence of divicine plus ascorbate, while the catalase activity was protected by addition of NADPH. In all experiments reported in Tables I and II the decrease in glutathione peroxidase activity was not reversed by either dialysis or treatment with thiols [18]. The role of the oxidation state of Hb on the glutathione peroxidase inactivation was also investigated. No effect whatsoever was seen when Hb was converted into CoHb before treatment of intact red cells with divicine and L-ascorbate. On the other hand, glutathione peroxidase inactivation was higher when the incubation was carried out with samples where practically all hemoglobin had been previously converted into the met-form. Discussion

Glutathione peroxidase was found to be inactivated by exposure to divicine and ascorbate of both intact erythrocytes and hemolysates. The effect was both concentration- and time-dependent and much more evident with hemolysates. A significant decrease (25%) of the enzyme activity was already found upon incubation of red blood cells with 2 mM reagents for 3 h at 37°C (Table I). Under the same experimental conditions, the inactivation was much higher in hemolysates (60% after 2 h incubation) (Table II). These conditions were selected for this study, since the average decrease of glutathione peroxidase in favism is 25% [81. Of the three antioxigenic enzymes only the glutathione peroxidase activity was affected by the conditions used in this work in all experiments performed with intact normal red blood cells. Neither superoxide dismutase nor catalase activity was substantially modified under any condition. Protection by externally added catalase in the experiments with intact red cells suggests that diffusion of H202 from outside is, at least in part, involved in the observed effect. Apparently this flux of peroxide overwhelms the natural enzymatic defenses of red blood cells only when the redox cycle of divicine is kept working by the presence of an appropriate reducing agent such as L-ascorbate. In fact, L-ascorbate is known [19] to enter the cell in its oxidized form, to be reduced intracellularly

and then to leak out, becoming available for another redox cycle. Permeability effects are likely to explain the differences between incubations with intact cells and those with hemolysates (Table If). In fact, in the latter case divicine alone is effective to some extent, probably because it is available to intracellular reductants for redox cycling. Furthermore, in the experiments with hemolysates catalase inactivation could be detected at incubation times comparable, and even shorter, with those of Table I. This inactivation was prevented by NADPH, in line with previous reports on protective or restorative effects of N A D P H on catalase exposed to H202 or H202 sources [20-22]. On the other hand, addition of GSH in the experiments with hemolysates to the inactivating incubation mixture, i.e., divicine plus ascorbate, protected glutathione peroxidase. This suggests that when GSH is limiting, either because of the experimental conditions used (Table lI) or natural deficiencies, like favism, peroxides may oxidize some activity linked amino acid residues of the enzyme, perhaps the active site selenocysteine [18]. However, when another reducing agent is engaged in the redox cycle of divicine, GSH may protect the enzyme from inactivation. The results of the present work emphasize the role of the oxidation state of Hb in producing the ultimate damaging species. While the inactivating mixture was ineffective when Hb was prevented from redox reaction by full conversion into CO-Hb, prior transformation of Hb into metHb increased glutathione peroxidase inactivation. A peroxidizing H202-metHb adduct [23] is the most likely candidate to explain these results, although glutathione peroxidase is known to be a molecule rather stable toward oxidative damage [18] and this certainly makes it a suitable enzyme for antioxidative defense. However, the results reported here show that the activity of the enzyme in the red cell is critically sensitive to a number of physiological conditions, i.e., the glutathione redox state, the hemoglobin oxidation state and the H 202 flux, which is determined by the balance of H202removing enzymes versus superoxide dismutase. Favism is a disease in which all of these seem to be involved, leading to a significant decrease in glutathione peroxidase activity, which may in turn play a major role in the pathogenesis of the hemo-

284

lytic event. Such a decrease may be produced by an unusual accessibility of large amount of favic aglycones for the dehydrogenase-deficient red blood cells of favic individuals, which creates in vivo a situation such as that pertaining in the experiments described here.

Acknowledgements This work was supported by the CNR Special Project 'Genetic Engineering and Molecular Basis of Inherited Disease'. References 1 Bray, R.C.+ Cockle, S.A., Fielden, E.M., Roberts, P.B., Rotilio, G. and Calabrese, L. (1974) Biochem. J. 139, 43-48 2 Kono, Y. and Fridovich, I. (1982) J. Biol. Chem. 257, 5751 5754 3 Blum, J. and Fridovich, I. (1984) Fed. Proc. 43, 2011 4 Bozzi, A., Mavelli, 1., Mondovi, B., Strom, R. and Rotilio, G. (1981) Biochem. J. 194, 369-372 5 Mavelli, I., Rigo, A., Federico, R., Ciriolo, M.R. and Rotilio, G. (1982) Biochem. J. 204, 535-540 6 Ciriolo, M.R., Mavelli, I., Rotilio, G., Borzatta, V., Cristofori, M., and Stanzani, L. (1982) FEBS Lett. 144, 264 268 7 Mavelli, I., Ciriolo, M.R., Rotilio, G., De Sole, P.+ Castorino, M. and Stabile, A. (1982) Biochem. Biophys. Res. C o m m u n . 106, 286-290 8 Mavelli, I., Ciriolo, M.R., Rossi, L., Meloni, T. Forteleoni, G., De Flora, A., Benatti, U., Morelli, A, and Rotilio+ G. (1984) Eur. J. Biochem. 139, 13 18

9 Mager, J., Glaser+ G., Razin, A., Isak, G., Bien, S. and Noam, J. (1965) Biochem. Biophys. Res. Commun. 20, 235-240 10 Chevion, M., Navok, T., Glaser, G. and Mager, J. (1982) eur. J. Biochem. 127, 405-409 11 McCord, J.M. and Fridovich, 1. (1969) J. Biol. Chem. 244, 4049-6055 12 Kosaka, H., lmaizumi, K., lmai, K. and Tyuma, I. (1979) Biochim. Biophys. Acta 581, 184-188 13 Beutler, E.+ Blum, K.G., Kaplan, J.C+, L6hr, G.W. and Valentine, W.N. (1977) Br. J. Haematol. 35+ 331 340 14 Paglia, D.E. and Valentine, W.N. (1967) J. Lab. Clin. Med. 70, 158 169 15 Lack, H. (1963) in Methods of Enzymatic Analysis (Bergmeyer, H.U., ed.), pp. 885-888, Verlag Chemie, Academic Press, Weinheim, New York 16 Rigo, A., Viglino, P. and Rotilio, G. (1975) Anal. Biochem. 68, 1-8 17 Drabkin, D.L. and Austin, J.M. (1953) J. Biol, Chem. 112, 52--65 18 Condell, R.A. and Tappel, A+L. (1983) Arch. Biochem. Biophys. 223, 407-416 19 Orringer, E.P. and Roer, M.E.S. (1979) J. Clin. Invest. 63, 53 58 20 Kirkman, H.N. and Gaetani, G.F. (1984) Proc. Natl. Acad. Sci. USA 81, 4343-4347 21 Eaton, J.W., Boreas, M. and Etkin, N.L. (1972) Adv. Exp. Med. Biol. 28, 121-131 22 Mc Mahon, S. and Stern, A. (1979) in Molecular Diseases, (Schewe T. and Rapoport, S., eds.), pp. 41-46. Pergamon, New York 23 Weiss+ S.J. (1980) J. Biol. Chem. 255, 9912 9917