Phycoerythrin fluorescence-based assay for peroxy radicals: A screen for biologically relevant protective agents

ANALYTICAL

BIOCHEMISTRY

177.300-306

(1989)

Phycoerythrin Fluorescence-Based Radicals: A Screen for Biologically Protective Agents’ Robert

J. DeLange2

Department

Received

of Microbiology

September

and Alexander and Immunology,

Assay for Peroxy Relevant

N. Glazer3 University

of California,

Berkeley,

California

94720

281988

fenses against free radical formation and damage. The nonenzymatic defenses depend on the differences in the rates of reaction of various molecules with reactive oxygen species and on the relative concentrations of these molecules. Protective scavengers such as ascorbate or vitamin E react very much faster than DNA, lipids, and proteins and at physiological concentrations provide excellent protection against such free radicals. Other important small molecule scavengers such as urate (4) and bilirubin (5) have been discovered recently. We have recently demonstrated that the time-dependence of the change in the fluorescence of phycoerythrin could be used as a measure of peroxy radical damage, as well as of metal ion dependent site-specific damage due to hydroxyl radicals generated in the Cu2+-ascorbate system (6). This assay provides a fast method for the quantitation of the rapidly reacting peroxy radical scavengers in a sample and allows relative ranking of the reactivity of macromolecules and small molecules toward peroxy free radicals. It should be emphasized that absolute rates of reaction are not determined and that this assay serves as a screen for the detection of potentially important scavengers of peroxy radicals. The assay sysDamage to cellular constituents such as DNA, RNA, tem depends on the unique properties of phycoerythrins. membranes, enzymes, and structural proteins is believed These proteins function as light-harvesting accessory to be involved in cancer, aging, and a number of disease proteins in cyanobacteria and red algae. Phycoerythrins states such as inflammatory-immune injuries, ischemiahave a molecular weight of about 250,000, a subunit reflow damage, and degenerative arterial disease (1,2). structure (c&r, and carry 34 covalently attached openIn particular, the pathological effects of peroxy free radichain tetrapyrroles (bilins) (7-9). The fluorescence cals have received much attention in connection with quantum yield of these proteins is >0.9 and is an exthe free radical chain mechanism of lipid peroxidation tremely sensitive measure of the physical and chemical (3). Cells possessboth enzymatic and nonenzymatic deintegrity of the protein. In this report, we describe how the assay can be used i This work was supported by National Institutes of Health Grant to measure the total amount of rapidly reacting peroxy GM 28994 and by National Science Foundation Grant DMB 8518066 scavengers in a biological fluid such as plasma. We show to A.N.G. that proteins and DNA are relatively inefficient in pro’ On sabbatical leave from the Department of Biological Chemistry, tecting against peroxy radicals and discuss possible reaUCLA School of Medicine, Los Angeles, CA 90024. 3 To whom correspondence should be addressed. sons for this behavior. We also report on the relative re-

Under the conditions of this assay, antioxidants that react rapidly with peroxy free radicals (e.g., ascorbate, vitamin E analogs, urate), protect phycoerythrin completely from damage by such radicals generated by thermal decomposition of 2,2’-azobis(2-amidinopropane); other compounds provide partial concentrationdependent protection. Change in phycoerythrin fluorescence emission with time provides a measure of the rate of free radical damage. The assay exploits the unusual reactivity of phycoerythrin toward these peroxy radicals. On a molar basis, phycoerythrin reacts with these radicals over loo-fold slower than do ascorbate or vitamin E analogs, but over go-fold faster than other proteins. Applications of this assay to the estimation of the peroxy radical scavenging capacity of human plasma are described, and to the comparison of the scavenging properties of several proteins and of DNA, of vitamins and their derivatives, of catecholamine neurotransmitters, and of a variety of other low molecular weight biological compounds. o 1~8s Academic PWS, IN.

300 All

Copyright 0 1989 rights of reproduction

0003-2697/89 $3.00 by Academic Press, Inc. in any

form

reserved.

FLUORESCENCE-BASED

ASSAY

FOR

PEROXY

301

RADICALS

activity toward peroxy radicals of vitamins and their derivatives, of the catecholamine neurotransmitters, and of a variety of other low molecular weight biological compounds. In addition to illustrating the versatility of the assay, these studies have led to a number of novel observations. MATERIALS

AND

METHODS

Chemicals and stock solutions. Porphyridium cruenturn B-phycoerythrin (B-PE)4 was prepared as described by Glazer and Hixson (10). This protein is readily available from several commercial suppliers. The concentration of phycoerythrin solutions was determined spectroscopically using a molar extinction coefficient of 2.41 X lo6 M-’ cm-’ at 545 nm (10). The concentrations of sheared salmon testis DNA (Sigma, St. Louis, MO) were determined based on TV&,,,, = 260/cm. The concentrations of defatted human serum albumin, of egg white lysozyme, and of ovalbumin (all from Sigma) were based 1% on k30nm = 6.6/cm, 27.3/cm, and 7.35/cm, respectively. All other compounds tested were of the highest grade commercially available and their concentrations were based on weight taken. The free radical initiator, 2,2’azobis(2-amidinopropane)dihydrochloride (AAPH) was obtained from Polysciences (Warrington, PA) and Trolox (6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid) from Hoffman-La Roche (Nutley, NJ). Chelex 100 resin (Na+-form) was obtained from Bio-Rad (Richmond, CA). Metal ions were removed from the 0.075 M phosphate buffer, pH 7.0, used in the assay mixtures, by passage through a lo-ml column of Chelex 100 before use, and all substances were dissolved in this buffer, unless otherwise noted. Fluorescence assay. The final reaction mixtures contained 1.65 X lo-‘M B-PE (lo), 4 mM AAPH, and any other additives in 2.0 ml of 0.075 M phophate buffer, pH 7.0 (made from 0.075 M NaH2P04 and 0.075 M KzHP04), in lo-mm quartz fluorometer cells thermostatted at 37°C. The reactions were initiated by adding 0.2 ml 40 mM AAPH (freshly prepared, stored on ice) to the other components in 1.8 ml at 37°C. Compensation for the temperature drop due to this addition required approximately 2 min, after which reproducible linear decreases in relative fluorescence emission were observed in control samples (no additives) as measured at 565 nm with excitation at 540 nm in a Perkin-Elmer MPF 44B fluorescence spectrophotometer. Plasma and fractions. Human plasma prepared from fresh heparinized blood (11) of a healthy male was stored frozen under nitrogen as 0.25-ml aliquots, until used. For 4 Abbreviations used: AAPH, 2,2’-azobis(2-amidinopropane) drochloride; B-PE, B-phycoerythrin; DME, Dulbecco’s gle’s medium; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; HVA, homovanillic acid.

modified

dihyEa-

0

IO

20

30

’ 50

40

TIME (MIN)

30’

0

I 10

’ 20

.

4 30

.

’ 40

’

’ 50

TIME (MIN)

FIG. 1. Assay

of the antioxidant properties of human plasma. (A) 1, B-PE control; 2, plus plasma (1:267 final dilution); 3, plus 80% saturated (NH,),SO, supernatant from plasma; 4, plus 80% saturated (NH,),SO, pellet from plasma. All fractions were present in the assay at dilutions of 1:267 compared to pure plasma. For experimental details, see text. (B) Schematic representation showing substances contributing to the complete and partial protection phase seen in the presence of plasma.

separation of plasma proteins, after thawing and dilution to 0.5 ml with phosphate buffer (containing ascorbate or urate in some samples), 0.2-ml aliquots were withdrawn, 0.8 ml of a saturated solution of enzymegrade ammonium sulfate (OOC)in phosphate buffer was added, mixed, and placed in ice for 90 min. The samples were centrifuged for 3 min in an Eppendorf 5414 microcentrifuge, and the supernatants were carefully decanted and stored on ice. The pellets were thrice resuspended in 1.0 ml of 80% saturated ammonium sulfate and recentrifuged. The first supernatant was added to the original supernatant and the other two washes discarded. Finally, the pellets were redissolved to 1.0 ml with phosphate buffer. Aliquots of each fraction were assayed in appropriate amounts to give a total dilution of 1:267 compared with pure plasma. RESULTS

AND

DISCUSSION

Assay of plasma. As shown in Fig. lA, the constant rate of production of peroxy free radicals by the thermal

302

DELANGE

AND

decomposition of AAPH (4 X ~O-‘M spl at 37°C) (12) results in a linear decrease in phycoerythrin fluorescence (6). When an aliquot of plasma is added to the reaction mixture there is a period of complete protection of phycoerythrin (flat region of curve) followed by only partial protection (slope less than the control). The complete protection phase can be attributed to the rapidly reacting antioxidants present in plasma (ascorbate, bilirubin, urate, and vitamin E) (13) and the partial protection phase to the proteins (and the associated lipids) present in plasma (Fig. 1B). Evidence for this interpretation was obtained by precipitating the plasma proteins at 80% saturated ammonium sulfate, at which concentration more than 95% of plasma proteins precipitate. As shown in Fig. IA, the fast-reacting antioxidants are not detected in the protein (pellet) fraction, whereas more than 90% is recovered in the supernatant fraction. It is also evident (Fig. 1A) that all of the material contributing to the partial protection phase is present in the protein pellet fraction. This discrimination between the rapidly reacting small molecules and proteins is possible because on an equimolar basis phycoerythrin reacts with peroxy radicals some 60 times faster than proteins such as human serum albumin (see Proteins and DNA below) but, at the concentration chosen, reacts at least 100 times more slowly than the rapidly reacting compounds mentioned above. The results (Fig. 1) show that as long as any of the fast reacting species remain, there will be little free radical damage to the proteins. This conclusion is consistent with other recent studies of the reactivity of individual plasma antioxidant species with peroxy radicals generated from AAPH (13). From our replicate experiments with a single plasma sample, we calculated that the fast-reacting substances scavenged about 630 pmol of peroxy radical per liter of plasma. This is in good agreement with the capacity estimated by others using different analytical methods (13-15). In these other studies a significant contribution from the protein fraction was included in the total for fast-reacting substances. As discussed above, we find that at the plasma dilution used here proteins do not contribute to the fastreacting fraction. This fluorescence assay provides a convenient means of quantitating the level of fast-reacting antioxidants in plasma or other biological fluids (e.g., cerebrospinal fluid) rapidly on minute samples. Such a measurement may be useful in assessing oxidative stress or large departures from normal levels of major plasma antioxidants. When additional ascorbate was added to plasma, the increase in the length of the complete protection phase corresponded exactly to the time of protection provided by the additional ascorbate alone (see Fig. 2). Similar results were obtained when a known amount of urate was added to plasma. Proteins and DNA. The effect of human serum albu-

GLAZER

3oL

I

0

10

'

'

20

.

'

30

.

'

40

I

50

.

'

60

TIME (MIN)

FIG. 2. Effect of added ascorbate on the antioxidant properties of human plasma. 1, B-PE control; 2, plus plasma; 3, plus 1.7 pM ascorbate; 4, plus plasma and 1.7 pM ascorbate. The plasma was present at a 1:267 dilution. At the ascorbate concentrations present in these assays, ascorbate was found to trap two peroxy radicals per ascorbate molecule.

min, egg white ovalbumin, and lysozyme on the reaction of phycoerythrin with peroxy radicals is shown in Fig. 3. These three proteins differ widely with respect to molecular weight, amino acid composition, and sequence. The effect of these proteins was determined over a concentration range of several orders of magnitude. At a B-PE concentration of 1.76 X lOpa M, for all three of these proteins the inhibition of fluorescence loss was proportional to protein concentration up to about 2 X lop7 M. However, above that concentration the measured inhibition was sharply lower than that calculated from the inhibition seen at concentrations 12 X 10m7 M (Fig. 3). The results obtained at the low “inhibitor” protein concentrations showed that phycoerythrin reacted with peroxy radicals about 60 times faster in each case. For all three proteins, an inhibition of 20 + 2% was obtained at 2 X 10e7 M. Since the molecular weight of serum albumin is 66,500, that of ovalbumin 42,700, and that of lysozyme 14,400, it is evident that it is the number of protein molecules in the assay mixture rather than the weight of protein that is important. Over the protein molecular weight range examined, the important parameter that governs the reaction rate must therefore be the collision frequency between the peroxy free radicals and protein molecules. Moreover, there appears to be little difference in the reaction rates of these three very different proteins with peroxy radicals in spite of large differences in amino acid composition and sequence. These results suggest strongly that the tetrapyrrole prosthetic groups must contribute significantly to the much higher relative reaction rate observed with phycoerythrin. At higher protein concentrations, the observed protection is increasingly lower than that calculated (Fig. 3). It is striking that even at concentrations of added protein lo3 higher than that of B-PE complete protec-

FLUORESCENCE-BASED

“.l SERUM

1 ALBUMIN

.l 1 OVALBUMIN

1

10 100 CONCENTRATION

10

LYSOZYME

IO 100 CONCENTRATION

ASSAY

1000 [x 10 7 M]

1000 10000 [x 1 O7 M]

100 CONCENTRATION

1000

10000 [x 10’ M]

FIG. 3. Effect of “protective” protein concentration on the rate of loss of phycoerythrin fluorescence. Experimental observations are indicated by open circles. Theoretical values (triangles and solid line) were calculated5 from the portion of the inhibition curve showing linear dependence on added protein concentration (~2 X 10m7 M “protective” protein) with the following pairwise relative probabilities of reaction with peroxy radicals at equimolar concentrations of phycoerythrin and the added protein: human serum albumin/B-PE = 0.015/ 0.985; ovalbumin/B-PE = 0.018/0.982; and lysozyme/B-PE = 0.016/ 0.984.

FOR

curves

were calculated

using

% Inhibition = P X [“protective”

OH. T(OH). T(OH)02.

% i5

T(OH).

+ O2 + T(OH)02.

+ enzyme + inactivation.

WNTROL 67.5 pglml DNA 270 pg/ml DNA

x loo 0

(1 - P)[B-PE] where the moles/liter tein.

+ T+

80

the equation

protein]

303

RADICALS

seen at that and higher concentrations. It is known that the reaction of peroxy radicals with proteins leads to the formation of protein radicals leading to modification of protein side chains, polypeptide chain scission, and protein-protein crosslinking (16-19). The inhibition data in Fig. 3 accord well with the following model. At low protein concentrations, the free radicals generated on the protein decay almost exclusively through intramolecular reaction. As protein concentration increases, the probability of protein-protein collisions during the lifetime of the protein free radicals also increases, favoring intermolecular reactions. The incomplete protection of B-PE seen at high concentrations of other proteins is attributable to reactions with such secondary radicals on the “protective” protein. This interpretation accords well with other studies. For example, Gee et al. (20) showed that lysine and glutamine, at a molar ratio to yeast alcohol dehydrogenase of about 4 X 104:1, strongly promoted the inactivation of the enzyme by peroxy radicals (see also ref. (18), p. 547). Examination of the protection of B-PE by DNA led to results similar to those seen with proteins. Sheared DNA at 67.5 pg/ml gave 32% protection and a fourfold higher DNA concentration provided only a 54% protection of B-PE (Fig. 4). In the study cited above (20), it was shown that thymine promoted inactivation of alcohol dehydrogenase due to radiation damage. In this instance, hydroxyl radicals were generated by radiation and the enhancement of enzyme inactivation in the presence of thymine (T) was attributed to the formation of secondary free radicals by the following reaction sequence:

tion is not achieved. From the results obtained at “protective” protein concentrations <2 X low7 M it can be calculated that above 10m6 M “protective” protein, the initial reactions of the peroxy radicals should be exclusively with the “protective” protein (see theoretical curves in Fig. 3) and 100% protection should have been ’ The theoretical

PEROXY

concentrations of “protective” protein and B-PE and P is the probability of reaction with “protective”

are in pro-

10

20

30

40

50

TIME (MIN)

FIG. thrin

4. Effect of DNA fluorescence.

concentration

on the rate of loss of phycoery-

304

DELANGE TABLE Inhibition

by Vitamins

Ascorbate Trolox NADPH Coenzyme A Thiamine Riboflavin-5-phosphate Pyridoxamine” NADP+ Nicotinic acid ’ Calculated

1 and

Their

Derivatives

Concentration (M)

Compound

from

% Inhibition

1.2 x 1o-6 1o-6 1om6 2.5 X 10-e 2 x 1o-4 1o-4 2 x 1o-3 1o-3 1o-3 the initial

linear

AND

rate of fluorescence

100 100 100 62 52 12 85 31 5 loss.

Vitamins and their derivatives. Of the vitamins and their derivatives tested (Table l), only Trolox (a watersoluble derivative of vitamin E), ascorbate, and NADPH gave 100% protection (initial flat region until the compound is consumed) in the assay. NADH is known to react rapidly with peroxy radicals (18) and was not tested. Although the NADP+/NADPH coenzymes are generally present at a lower concentration than the NAD+/NADH coenzymes (an exception is liver where the concentrations are approximately equal), the former are usually found predominantly in the reduced form, whereas the latter are usually found mainly in the oxidized form (21). The NADPH concentration could reach 10-4-10-5 M in many tissues and even 1O-3 M or higher in the liver (22). These coenzymes are present largely in enzyme-bound form where their reactivities may be modified; however, our results suggest that NADPH (and NADH) function as major antioxidants within cells in addition to their contribution in maintaining another major antioxidant, glutathione, in its reduced state. Coenzyme A gave good protection as expected because of its reactive sulfhydryl group. It is doubtful if the physiological concentrations of thiamine, riboflavin, and pyridoxamine are sufficiently high for these compounds to be effective as antioxidants (see Table l), but it should be noted that these vitamins are generally found bound to proteins where they might serve as local protective agents. Neurotransmitters and related substances.6 We have tested a number of neurotransmitters and related substances (Table 2) and found these, in general, to be effective peroxy radical scavengers in the micromolar concentration range. Homovanillate (HVA), an end product of neurotransmitter metabolism, is present in 6 A report of some of the results in this section was presented at the 41st Annual Scientific Meeting, The Gerontological Society of America, San Francisco, CA, Nov. 18-22, 1988.

GLAZER

most regions of the brain at higher concentrations (up to 54 PM, see Ref. (23)) than any of the neurotransmitters or related metabolites (in general <6 PM). Therefore, even though some of these other substances react more rapidly with peroxy radicals that HVA (Table 2), HVA would be expected to scavenge many more peroxy radicals on a quantitative basis. The reaction of HVA and of some of the other neurotransmitters (Table 2) with peroxy radicals was biphasic (Fig. 5) with the slower secondary reaction for HVA proceeding at a rate 85% of the initial reaction rate. Dopamine can also reach levels of up to 33 PM in certain regions of the brain (20) and would be expected to function as a significant free radical scavenger in those regions. However, the secondary reaction is not seen with dopamine. Because of the secondary reaction, on a molar basis, HVA and the other neurotransmitters scavenge more peroxy radicals than does ascorbate. This is evident from the data presented in Fig. 5. From the increase in the length of the complete protection phase seen in mixtures of ascorbate and HVA as compared to that seen with ascorbate alone (Fig. 5A), it can be calculated that HVA reacts with peroxy radicals 15-20 times slower than does ascorbate. The brain is one of the most metabolically active organs in the body and consumes 20% of the oxygen used in metabolism. Therefore, it might be expected that this organ has extensive antioxidant defenses. Brain injuries often release metal ions (mainly iron), and this can lead to iron-dependent lipid peroxidation, etc., because cerebrospinal fluid has little capacity for metal binding

TABLE Inhibition

by Neurotransmitters

Compound Epinephrine”

Norepinephrine” L-3,4-Dihydroxyphenylalanine” Dopamine

Homovanillate” Serotonin’

Histamine

2 and

Related

Concentration (M) 5 x 1om7 1o-6 1o-5 3.5 x 10-r 1om6 5 x 10-r 1o-6 3 x 1om7 6 X 10m7 10-s 1o-6 2 x 1o-6 1om6 3 x 1om6 1o-5 1o-6 1om5

Substances % Inhibition 62 80 100 63 83 60 78 41 62 82 43 55 75 83 88 19 41

a The kinetics of phycoerythrin fluorescence quenching observed in the presence of these compounds departed from zero order. Consequently the inhibition values given above were calculated from the initial linear rate of fluorescence loss.

FLUORESCENCE-BASED

ASSAY

80,

I

r.1

50'

Y 0

IO

20

TIME (MIN)

80 %

70

52

60

2z

50

ii!

40

; s E

30 20

SYMBOLS

0

IO

AS IN

20

A

ABOVE

30

40

50

60

70

FIG. 5. Effect of ascorbate (ASC) and of homovanillic acid (HVA) on the rate of loss of phycoerythrin fluorescence. (A) In expanded form, the first 24 min of the kinetic data presented in (B). Note the increased length of the complete protection phase produced by ascorbate (1.2 pM) plus HVA (1.5 pM) versus that produced by ascorbate (1.2 @M) alone.

(24,25). Ascorbate (and possibly other substances) can participate in metal-catalyzed production of free radicals (1,24). Our results indicate that neurotransmitters and their metabolic breakdown products may serve as important antioxidants in the brain. Other compounds. Table 3 lists results with various other compounds screened for antioxidant activity. The 100% inhibition seen with urate indicates that no decrease in phycoerythrin fluorescence was seen until the urate was completely oxidized. It is of interest that Hepes buffer is also an effective antioxidant in the millimolar range, and this could be one reason why this compound and the related “Good buffers” (26) have been found particularly suitable for cell and tissue culture. The antioxidant properties of these buffers may be due to the presence of trialkyl nitrogen groups which may react with peroxy radicals to form N-oxides and even triethylamine gives protection at high concentrations (Table 3). We also tested a cell culture medium (Dulbecco’s modified Eagle’s mediumRef. (27)) that contained no Hepes or serum and found effective protection even at dilutions of 1:200. It would appear that the antioxidant capacity of this medium is

FOR

PEROXY

305

RADICALS

attributable to the presence of several amino acids (see methionine in Table 3), vitamins (see Table 2), glucose (Table 3), and probably pyruvate and the phenol red indicator (we have not assayed these). The polyamines spermidine and spermine are present in eukaryotic cells at concentrations that can exceed 1 mM, whereas putrescine is present in much lower amounts (28). Our results (Table 3) suggest that, in addition to their other functions, these polyamines could conceivably function as antioxidants in protecting nucleic acids and other components from free radical damage. Such a role has been proposed previously (29). Concluding remarks. The sensitivity and speed of the phycoerythrin assay and its fluorescence readout permit assessment of the protection afforded by various compounds over a concentration range spanning several orders of magnitude. The results show that at low concentrations (&2 X lop6 M) a small group of compounds (such as ascorbate and Trolox) afford complete protection of phycoerythrin against peroxy radicals and that at such low concentrations the period of complete protection is directly proportional to the concentration of the protective compound. The majority of other compounds examined, whether proteins, nucleic acids, or small molecules, showed protective effects which decreased proportionately at higher concentrations. From these results, and a number of reports in the literature, it is evident that radicals (and other reactive species) are generated by the reaction of peroxy radicals with these compounds. These secondary radicals then damage phy-

TABLE

Inhibition

by Various Compounds Concentration (M)

Compound Urate Methionine

Hepes

buffer

DME

culture

3

medium”

Triethylamine Glucose Putrescine* Spermine* Carnitine’ ’ The values given are dilutions phosphate buffer, pH 7.0. This ’ Assayed in 0.1 M NaHCO,, polyamine. ’ Additionally, . . no . inhibition acetylcarmtine, both at 5 X lo-”

% Inhibition

1o-6 2 x 1om6 6 X 10m6 2 x 1o-5 5 x 1om4 2.5 X 1O-3 0.5% 1.5% 1om2 4 x 1om3 6 X 1O-3 2 x 1o-3 6 X 1O-3 5 x 1o-3

100 39 50 67 84 92 58 75 41 29 58 53 76 0

of medium, by volume, into 0.075 M medium contains no serum (see text). pH 7, to prevent precipitation of the was observed M.

with

L-propionyl-

or L-

306

DELANGE

AND

coerythrin. Such intermolecular reactions are much less important in the case of antioxidants which react with peroxy radicals very rapidly and can in consequence protect at very low concentrations. Even an excellent scavenger such as ascorbate shows concentration-dependent antioxidant efhciency because of a peroxy radical-initiated chain autoxidation process which increases in proportion to ascorbate concentration (30).

2. Ames,

B. (1987)

FASEB

J. 1,358-364.

B. N. (1983)

Science

221,1256-1263.

3. Mead, J. F. (1976) in Free Radicals Vol. 1, pp. 51-68, Academic Press,

in Biology New York.

(Pryor,

W. A., Ed.),

R., Schwiers, E., and Hochstein, 4. Ames, B. N., Cathcart, Proc. Natl. Acad. Sci. USA 78,6858-6862. 5. Stocker, R., Glazer, A. N., Acad. Sci. USA 84.5918-5922.

and

6. Glazer,

J. 2,2487-2491.

A. N. (1988)

FASEB

Ames,

P. (1981)

B. N. (1987)

Proc.

of Plants (Hatch, 7. Glazer, A. N. (1981) in The Biochemistry and Boardman, N. K., Eds.), Vol. 8, pp. 51-96, Academic New York. 8. Glazer,

A. N. (1984)

Biochim.

9. Klotz, 4863.

A. V., and Glazer,

10. Glazer,

A. N., andHixson,

11. Yamamoto, (1987) Anal. 12. Terao, 201.

Biophys.

K., and Niki,

13. Frei, B., Stocker, USA 85,9748-9752.

Actu

N&l.

M. D., Press,

768,29-51.

A. N. (1985)

J. Biol.

Chem.

260,

C. S. (1977)

J. Biol.

Chem.

252,32-42.

Y., Brodsky, M. H., Baker, Biochem. 160,7-13. E. (1986)

R., and Ames,

14. Wayner, D. D. M., Burton, G. W., Ingold, K. U., and Locke, S. (1985) FEBS L&t. 187,33-37. 15. Ingold, K. U., Webb, A. C., Witter, D., Burton, G. W., Metcalfe, T. A., and Muller, D. P. R. (1987) Arch. Biochem. Biophys. 259, 224-225. 16. Wolff, S P., Garner, A., and Dean, R. T. (1986) Trends Biochem. Sci. 11,27-31. 17. Hunt, J. V., Simpson, J. A., and Dean, R. T. (1988) B&hem. J. 250,87-93. 18. Willson, R. L., Dunster, C. A., Forni, L. G., Gee, C. A., and Kittridge, K. J. (1985) Phil. Trans. R. Sot. London B 3 11,545-563. 19. Sparrow, C. P., Parthasarathy, S., and Steinberg, D. (1988) J. Lipids Res. 29,745-753.

REFERENCES 1. Halliwell,

GLAZER

4856-

J. C., and Ames,

J. Free Radicals B. N. (1988)

Biol. Proc.

Med. Natl.

B. N. 2,193-

Acad.

Sci.

20. Gee, C. A., Kittridge, K. J., and Willson, R. L. (1985) Brit. J. Radiol. 58,251-256. 21. Veech, R. L., Eggleston, L. V., and Krebs, H. A. (1969) Biochem. J. 115,609-619. 22. White, A., Handler, P., and Smith, E. L. (1968) chemistry, Fourth Edition, p. 335, McGraw-Hill, 23. Lloyd, K. G., Hornykiewicz, O., Davidson, ley, I., Goldstein, M., Shibuya, M., Kelley, (1981) N. Engl. J. Med. 305,1106-1111.

Principles of BioNew York.

L., Shannak, K., FarW. N., and Fox, I. H.

24. Halliwell, B., and Gutteridge, ology and Medicine, Clarendon

J. M. C. (1985) Press, Oxford.

Free Radicals

25, Halliwell, B., and Gutteridge, phys. 246,501-514.

J. M. C. (1986)

Arch.

Biochem.

in BiBio-

26. Good, N. E., and Izawa, S. (1972) in Methods in Enzymology (San Pietro, A., Ed.), Vol. 24, pp. 53-68, Academic Press, San Diego, CA. 27. Rutzky, L. P., and Pumper, R. W. (1974) In Vitro 9,468-469. 28. Russell, D. H., and Durie, B. G. M. (1978) Progress search and Therapy, Vol. 8, pp. l-78, Raven Press, 29. Kitada, M., Igarashi, K., Hirose, S., and Kitagawa, them. Biophys. Res. Commun. 87,388-394. 30. Wayner, D. D. M., Burton, G. W., and Ingold, K. chim. Biophys. Acta 884,119-123.

in Cancer ReNew York. H. (1979) BioU. (1986)

Bio-