Protein Damage and Degradation by Oxygen Radicals · Vol. 262. No .20, Issue of July 15, pp. 9895-9901.1987 Printed in U.S.A. Protein Damage and Degradation by Oxygen Radicals I

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society of Biological Chemists, Inc.

Vol. 262. No .20, Issue of July 15, pp. 9895-9901.1987 Printed in U.S.A.

Protein Damage and Degradation by Oxygen Radicals I. GENERAL ASPECTS*

(Received for publication, December 30, 1986)

Kelvin J. A. Davies From the Institute for Toxicology and the Department of Biochemistry, The University of Southern California, Los Angeles, California 90033

Aggregation, fragmentation, amino acid modification, and proteolytic susceptibility have been studied following exposure of 17 proteins to oxygen radicals. The hydroxyl radical (‘OH) produced covalently bound protein aggregates, but few or no fragmentation products. Extensive changes in net electrical charge (both + and -) were observed. Tryptophan was rapidly lost with ‘OH exposure, and significant production of bityrosine biphenol occurred. When incubated with cell- free extracts of human and rabbit erythrocytes, rabbit reticulocytes, or Escherichia coli, most ‘OH-modified proteins were proteolytically degraded up to 50 times faster than untreated proteins. The exceptions were a- casein and globin, which were rapidly degraded without ‘OH modification. ATP did not stimulate the degradation of ‘OH-modified proteins, but a-casein was more rapidly degraded. Leupeptin had little effect under any condition, and degradation was maximal at pH 7.8. The data indicate that proteins which have been denatured by ‘OH can be recognized and degraded rapidly and selectively by intracellular proteolytic systems. In both red blood cells and E. coli, the degradation appears to be conducted by soluble, ATP-independent (nonlysosomal) proteolytic enzymes. In contrast with the above results, superoxide (0;) did not cause aggregation or fragmentation, tryptophan loss, or bityrosine production. The combination of ‘OH + 0; (+02), which may mimic biological exposure to oxygen radicals, induced charge changes, tryptophan loss, and bityrosine production. The pattern of such changes was similar to that seen with ‘OH alone, al- though the extent was generally less severe. In contrast with ‘OH alone, however, ‘OH + 0; (+02) caused extensive protein fragmentation and little or no aggregation. More than 98% of the protein fragments had molecular weights greater than 5000 and formed clus- ters of ionic and hydrophobic bonds which could be dispersed by denaturing agents. The results indicate a general sensitivity of proteins to oxygen radicals. Ox- idative modification can involve direct fragmentation or may provide denatured substrates for intracellular proteolysis.

Recent reports from this laboratory and others have suggested that oxidatively modified proteins are rapidly and

* This work was supported by Grant ES 03598 from the National Institutes of Health. Part of this work has been published in prelim- inary form (Lin, S. W., and Davies, K. J. A. (1985) Fed. Proc. 44, 1093 (Abstr. 3990)). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

selectively degraded by intracellular proteolytic systems (1- 17). Such observations may have important implications for the regulation of protein turnover in all aerobic organisms (4, 6-12). The mechanisms by which cellular proteinases, proteases, and peptidases recognize oxidatively modified proteins are not understood. Furthermore, we have only a cursory understanding of the types of protein modifications which various active oxygen species can induce. For detailed work on mechanisms of proteolytic recognition to be meaningful, it is also necessary to undertake in-depth studies of protein modification by oxygen radicals. In particular, those modifications which appear to be most favorable and those which appear to be most common (to various proteins) should be investigated since they may provide “universal signals” for proteolysis. Protein modification should also be compared with proteolytic susceptibility in an effort to identify potential relationships.

In this paper, I have attempted to catalogue a series of major oxidative modifications which are common to a wide variety of proteins. Several of these proteins have also been studied for proteolytic susceptibility during incubation with cell-free extracts of erythrocytes, reticulocytes, and Esche- richia coli. In the succeeding two papers (1, 2 ) , oxidative modifications have been studied in great detail with a representative model protein, bovine serum albumin (BSA).’ Fi- nally, the major oxidative modifications to BSA have been carefully correlated with proteolytic susceptibility in the fourth paper of this series (3).

A fairly wide variety of active oxygen species are known to exist, and most have been suggested as biological mediators by various investigators. The superoxide anion radical (O;), the hydrodioxyl radical (HO;), and the hydroxyl radical (‘OH) comprise the biologically relevant oxygen radicals. Of these radicals, HO; has a pK of 4.8 (18) and may therefore have limited biological significance. The other active oxygen species are hydrogen peroxide (HzO,) and singlet molecular oxygen (loz), which are not radicals. Hydrogen peroxide is the least reactive form of oxygen (other than 02), and ’Oz may only be important in photoreactions.

Previous work indicated that ‘OH (or a species with similar reactivity) might be responsible for initiating protein degradation in red blood cells (4-7). Superoxide is known to be generated by several enzymes and by the (auto)oxidation of numerous compounds (18). The combination of ’OH + 0; + 0, may best reflect biological exposure to oxygen radicals. In this paper and the accompanying three manuscripts (1-3), the oxygen radicals ‘OH and 0; have therefore been studied alone and in combination with 0,.

The abbreviations used are: BSA, bovine serum albumin; RBC, red blood cells; Hb, hemoglobin; SDS-PAGE, sodium dodecyl sulfate- polyacrylamide gel electrophoresis; HEPES, 4-(2-hydroxyethyl)-l- piperazineethanesulfonic acid.

9895

9896 Protein Damage and Degradation by Oxygen Radicals

MATERIALS AND METHODS

Proteins-The following proteins were obtained from Sigma: BSA (A 0281), catalase (C-40), hemoglobin (H 2500; Hb), globin (G 3633), a-casein (C 7891), transferrin (T 2252), human serum albumin (A 1887), ovalbumin (A 5015), myoglobin (M 0630), cytochrome c (C 2506), hexokinase (H 5250), tyrosinase (T 7755). peroxidase (P 8375), trypsin (T 8253), and chymotrypsin (C 4129). Superoxide dismutase (83-500-1) was obtained from Miles Scientific (Naperville, IL), and lactate dehydrogenase (0278) was purchased from Pharmacia P-L Biochemicals. All proteins were extensively dialyzed prior to use.

Exposure of Proteim to Oxygen Radicals-Proteins were exposed to oxygen radicals by @'Co radiation of dilute (0.33 mg/ml) solutions in double distilled and deionized water. Protein solutions were satu- rated with either 100% N20 or 100% 0 2 and irradiated in a Gammacel 200 @Co source (Atomic Energy of Canada, Ltd.) a t a dose rate of 634 f 5 rads/min, as measured by Fricke and Hart dosimetry (19). @Co exposure (at 25 'C) was varied from 47 s to 158 min in order to achieve total doses of 0.5-100 kilorads.

@Co radiation induces the radiolysis of water (19-21).

HzO + 'OH + eiq + H ' (1)

Radiolytic yields are expressed as G values (radicals/100 eV), and G.OH - GeG = 3.0, whereas (&. - 0.55; in other words, H' represents less than 10% of the yield of 'OH + e& Under 100% N20, the solvated electron (e,) reacts quantitatively to produce more . OH (19, 21).

e, + N20 + H20 + 'OH + N2 + OH- (2)

Under 100% 0 2 , e, reacts to produce OF, and H . reacts to form HO (19, 21).

e; + 0, + 0, (3)

H ' + 0 2 + HO; (4)

The HO; produced in Reaction 4 is rapidly deprotonated at neutral pH (pK 4.8) to form more 0, (18).

HO; + 0, + H' (5)

Under 100% O2 with 0.01 M sodium formate, 'OH reacts to produce 0, (19,21).

'OH + HCOO- + Hz0 + COT (6)

co, + o2 + co2 + 0; (7)

Since e, and H ' react with 0, as per Reactions 3-5 to also give 0; (19, 21), the only oxygen radical generated by radiation with 100% O2 + 0.01 M formate is 0;. Possible effects of H ' can be discounted by experiments with isopropyl alcohol and t-butyl alcohol (22), as shown in the subsequent papers (1, 2). Under 02, the effects of 'OH and 0; (as well as O2 itself) must be considered.

Oxygen radical yields are based on known G values for aqueous samples since proteins a t 0.33 mg/ml do not significantly affect results (19). Under 100% NzO, G.0" = 6.0; and under 100% O2 with 10.0 mM sodium formate, G , = 6.0. Under 100% O2 (without formate), G.OH = 3.0 and Go,- 3.0, i.e. there is an equal yield of 'OH and 0;. Oxygen radical/protein ratios (nmol of radicals/mg of protein) of 9-1800 were studied by varying the radiation dose from 0.5 to 100 kilorads.

Preparation of CeU-free Extracts-Cell-free extracts of human and rabbit erythrocytes, rabbit reticulocytes, and E. coli were prepared by similar methods (all a t 4 "C). Human erythrocytes were obtained from healthy males and females, aged 20-25. Heparinized erythrocytes and reticulocytes from 8-12-week-old rabbits were obtained from Pel-Freez Biologicals (Rogers, AR). Reticulocytosis was induced by a series of phenylhydrazine injections. E. coli strain RM312, a prototrophic derivative of the W3110 strain of E. coli K12 (23), was the kind gift of Dr. Raymond D. Mosteller (University of Southern California).

RBC were washed by centrifugation at 800 X g in 4 volumes of 0.9% NaCl. The supernatant, buffy coat, and upper 15% of the packed cells were removed by aspiration, and three additional centrifugal washes were performed. RBC were resuspended in 1.5 volumes of water plus 1 mM dithiothreitol and stirred for 60 min in order to lyse the cells. Unbroken cells were removed by centrifugation at 10,000 X g. Membrane fractions and organelles were removed by centrifugation at 18,500 X g (60 min) for erythrocyte lysates and at 40,000 X g (90 min) for reticulocyte lysates. Lysates were dialyzed against 500 vol-

umes of 10 mM Tris, 5 mM MgCl,, 0.5 mM dithiothreitol, 8 mM KCl, and 10% (v/v) glycerol to produce cell-free extracts. Dialysis was conducted for 20 h against a membrane with a nominal molecular weight cutoff of 12,000, with three buffer changes.

E. coli strain RM312 was grown aerobically on M9 minimal medium plus 20.0 mM glucose and harvested at 16,000 X g (10 min). Cell-free extracts were obtained by a modification of the procedure of Murak- ami et al. (24). Packed cells were resuspended in 50 mM triethanola- mine buffer (pH 8.0) plus 10% (w/v) sucrose and frozen in a suspen- sion of dry ice + ethanol. The partially lysed cells were thawed at 4 "C and further lysed with 0.1 mg/ml lysozyme. Next, 10 mM 8- mercaptoethanol plus 0.5 mM KC1 were added. Unbroken cells were removed by centrifugation at 45,000 X g (60 min). The supernatant (lysate) was dialyzed (as described above) for 20 h against 500 volumes of 50 mM Tris (pH 7.8), 10 mM MgC12, 10 mM KCl, and 0.5 mM dithiothreitol (with three buffer changes).

Proteolysis Measurements-The proteolytic susceptibility of untreated and 'OH-treated BSA, superoxide dismutase, catalase, Hb, globin, and casein was tested with cell-free extracts of RBC and E. coli. BSA, superoxide dismutase, catalase, Hb, and casein were pre- labeled with 3H by the reductive methylation procedure of Rice and Means (25). A slightly different procedure was used for globin (26).

Following exposure to 'OH, the labeled proteins were incubated (37 "C) for 60 min with cell-free extract. For both RBC and E. coli experiments, 0.01 ml of each 3H-protein (0.33 mg/ml) was incubated with 0.06 ml of extract. Tris buffer (50 mM, pH 7.8) was added such that the final incubation volume was 0.13 ml.

After a 60-min incubation, all tubes were placed on ice for 5 min, and trichloroacetic acid was added to a final concentration of 10% in order to precipitate the remaining intact protein. Each sample was vortexed, left on ice for 10 min, and centrifuged at 2800 X g for 15 min. Aliquots of acid-soluble supernatant were mixed with liquid scintillation mixture and assayed for radioactivity. Background counts (before and after incubation) were measured with cold BSA (as carrier) instead of cell-free extract, and total counts were measured without acid precipitation. Percent protein degradation was measured as: (acid-soluble counts/total counts) X 100.

Protein Aggregation and Fragmentation2-Polyacrylamide gel electrophoresis (PAGE) was used to determine the aggregation or fragmentation of proteins by oxygen radicals. Nondenaturing PAGE (27) as well as sodium dodecyl sulfate (SDS)-PAGE, in the presence (28) or absence (29) of urea, were conducted with and without dithiothreitol and P-mercaptoethanol. Gels were stained with either Coo- massie Brilliant Blue R-250 or silver stain (Bio-Rad). Aggregation (increase in molecular size) and fragmentation (decrease in molecular size) were quantitated with a Bio-Rad Model 1650 densitometer connected to an IBM personal computer via a Data Translations A/ D board. The staining intensities of protein bands were converted to electrophoretic peaks for analysis using Plot-4-U software (courtesy of Dr. Michael B. Bolger, University of Southern California). Protein concentration was calculated from the integration of peak areas in comparison with standards. The method can reliably detect a 5% loss of a given protein band.

Amino Acid Modification-Isoelectric focusing gels were used to detect alterations in primary structure which affected overall electrical charge. Ultrathin isoelectric focusing gels (0.2 mm) were cast with GelBond backing film (30, 31). Ampholytes were chosen for an effective pH range of 3-10.

Tryptophan oxidation was monitored by loss of protein fluorescence peaks at 280 nm excitation and 340-350 nm emission (32) in comparison with (untreated and oxidized) free tryptophan. Bityrosine production was assessed at 325 nm excitation and 410-420 nm emission (33) in comparison with authentic bityrosine and bityramine.

RESULTS

Aggregation and Fragmentation of Proteins-SDS-PAGE in the presence of urea revealed that exposure to either 'OH alone or to 'OH + 0; (in the presence of 100% 02) caused a

The term "protein fragmentation" refers to the direct breakdown of proteins by oxygen radicals. Such processes have been found to involve both main chain scission and side chain scission by mechanisms which differ from peptide hydrolysis. In contrast, the term "protein degradation" refers to peptide bond hydrolysis by proteolytic enzymes.

Protein Damage and Degradation by Oxygen Radicals 9897

loss of protein staining bands (Table I). In contrast, exposure only to 0; (+02) caused no change in molecular weight for any protein. A greater loss of polypeptide band staining intensity was seen with 'OH, for most proteins, than with 'OH + 0; (+02). For the 17 proteins studied, the average loss of polypeptide bands was 57.9 f 8.5% with 'OH exposure and 39.3 f 7.5% with exposure to 'OH + 0 2 (+02).

Exposure to 'OH induced a generalized aggregation of proteins to higher molecular weight forms (Table 11). Dimers, trimers, and even tetramers were evident with many proteins. In contrast, exposure to .OH + 0; (+02) generally produced a dispersed pattern of lower molecular weight protein fragmentation products. More than 98% of these fragments (for each protein) were larger than M, 5000 as judged by SDS- PAGE and precipitation studies in 5-10% trichloroacetic acid.

Since 0; alone had no effect on molecular weight (Table I), it is likely that protein radicals induced by 'OH reacted either with 0; or (more probably) with the 100% 0, present during exposure to 'OH + 0;. Such reactions have been suggested by other investigators (34,35) and probably explain the fragmentation produced by *OH + 0; + O2 (Table 11). When %o radiation was conducted under room air, proteins were seen to fragment with increasing dose, and little evidence of aggregation was observed. This pattern is essentially identical to that reported for 100% 0, (Table 11) and indicates that even ambient O2 concentrations (-20%) are sufficient for fragmentation reactions. A detailed study of this obser- vation is provided for BSA in two of the subsequent papers of this series (2,3).

Each protein exhibited a characteristic sensitivity to oxygen radical exposure using ratios of 9 nmol to 1.8 pmol of radicals/ mg of protein. A ratio of 150 nmol of radicals/0.33 mg of protein (25 kilorads) was chosen for comparison of different proteins since both aggregation and fragmentation processes were easily measured at this point (Table 11). Comparisons of

TABLE I Sodium dodecyl sulfate-polyacrylumide gel electrophoresis following

exposure of proteins to oxygen radicals Proteins were exposed to 150 nmol of 'OH, to 75 nmol of 'OH +

75 nmol of 0; (+02), or to 150 nmol of 0: (+02) at a ratio of 150 nmol of radicals/0.33 mg of protein (25-kilorad radiation). Scanning densitometry (see "Materials and Methods") was used to quantify the loss of (silver-stained) protein bands following SDS-PAGE in the presence of urea (28). Decreased staining intensity of polypeptide bands was always accompanied by formation of new (aggregated or fragmented) species, as shown in Table 11. Values are means of three independent determinations for which standard errors were less than 10%.

Loss of protein staining band Protein

.OH .OH+ 0; 0,

% Lactic dehydrogenase 99 85 0 Transferrin 98 49 0 Bovine serum albumin 98 85 0 Human serum albumin 90 70 0 Ovalbumin 80 78 0 Superoxide dismutase 74 29 0 Myoglobin 65 11 0 Catalase 56 34 0 Globin 52 36 0 Cytochrome c 48 43 0 Hexokinase 42 20 0 Hemoglobin 41 32 0 Tyrosinase 36 21 0 Peroxidase 32 18 0 a-Casein 12 20 0 Trypsin 6 8 0 Chymotrypsin 4 6 0

TABLE I1 Aggregation and fragmentation of proteins following exposure to

oxygen radicals Oxygen radical exposures and SDS-PAGE quantitation were ex-

actly as described in the legend to Table I. Aggregation (increased molecular weight) and fragmentation (decreased molecular weight) products are expressed as percent of total protein staining intensity. Values are means of three independent determinations for which standard errors were less than 10%.

Aggregation products

'OH ' O H + O ;

%

Protein

Lactic dehydrogenase 89 11 Transferrin 90 20 Bovine serum albumin 98 0 Human serum albumin 90 0 Ovalbumin 80 0 Superoxide dismutase 74 0 Myoglobin 65 0 Catalase 56 0 Globin 52 0 Cytochrome c 48 0 Hexokinase 37 10 Hemoglobin 41 22 Tyrosinase 36 0 Peroxidase 32 0 a-Casein 12 0 Trypsin 6 0 Chymotrypsin 6 0

Fragmentation products

.OH 'OH+O:

% 10 74 8 29 0 85 0 70 0 78 0 29 0 11 0 34 0 36 0 43 5 10 0 10 0 21 0 18 0 20 0 8 0 4

nondenaturing PAGE and SDS-PAGE (all kdithiothreitol or +P-mercaptoethanol) revealed little difference in the basic patterns of aggregation or fragmentation for 'OH-treated proteins. Such analyses indicate that some 90% of the protein aggregates induced by 'OH can be attributed to new intermolecular covalent bonds (other than s-S bonds). Less than 10% of the aggregation products could be assigned to noncovalent interactions or disulfide bonds (data not shown).

Nondenaturing polyacrylamide electrophoresis gels of proteins exposed to ' OH + 0; (+02) revealed both fragmentation and aggregation products in approximately equal yield (not shown). The aggregation products were widely dispersed, and no discrete bands could be discerned. The addition of dithiothreitol or P-mercaptoethanol had no measurable effect, indicating that the aggregation products were not due to the formation of new disulfide bonds. When the same protein samples were analyzed by (denaturing) SDS-PAGE k urea (Table II), only fragmentation products were observed. It is therefore reasonable to conclude that the aggregation products seen in nondenaturing PAGE of ' OH + 0; ( +02) -treated proteins were random conglomerates of protein fragments held together by noncovalent attractions (ie. hydrophobic and ionic bonds).

Proteolytic Susceptibility-The proteins which had been exposed to 'OH alone were clearly still intact (i.e. not fragmented). Studies of proteolytic susceptibility therefore con- centrated on ' OH-treated proteins. When incubated with cell- free extracts of RBC or E. coli, most 'OH-treated proteins exhibited significant increases in degradation by cellular proteolytic systems (Table 111). It is important to note that no active oxygen species were present during proteolysis experiments. Thus, increased degradation represents true recognition of protein damage by cellular proteolytic enzymes.

The only proteins whose degradation was not increased by 'OH exposure were globin and casein (Table 111). Globin is a denatured protein, and casein has little, if any, secondary or tertiary structure. Untreated globin and casein were both excellent proteolytic substrates, but 'OH exposure may ac-

9898 Protein Damage and Degradation by Oxygen Radicals

TABLE 111 Proteolytic susceptibility following hydroxyl radical exposure

3H-Labeled proteins either were exposed to 150 nmol of ’OH/0.33 mg or were untreated. Untreated and ‘OH- treated proteins were then incubated with cell-free extracts (at 37 “C and pH 7.8) to measure proteolytic susceptibility, as described under “Materials and Methods.” RBC extracts were incubated at a final concentration of 6.0 mg of protein/ml, and E. coli extracts were used at a final (incubation) concentration of 1.5 mg of protein/ ml (Bradford assay (51)). Approximately 95% of RBC extract protein was Hb. Results are means f S.E. of three indeDendent determinations.

Protein degraded during 60-min incubation Protein studied Rabbit erythrocyte

extract E. coli extract Human erythrocyte extract

%

Rabbit reticulocyte extract

Bovine serum albumin Untreated ’OH-treated

Superoxide dismutase Untreated ‘OH-treated

Catalase Untreated ’OH-treated

Hemoglobin Untreated ‘OH-treated

Globin Untreated ’ OH-treated

cy-Casein Untreated ’ OH-treated

0.06 f 0.01 2.76 f 0.05

0.05 f 0.01 3.01 k 0.07

0.27 f 0.05 11.82 f 0.24

0.08 f 0.01 3.25 f 0.11

0.32 f 0.01 1.73 f 0.06

0.55 f 0.01 1.27 f 0.05

0.12 f 0.02 3.31 f 0.19

0.30 f 0.01 1.71 f 0.01

0.23 f 0.01 2.27 f 0.04

0.10 4 0.01 2.58 f 0.03

2.70 f 0.03 13.44 f 0.37

3.69 f 0.38 8.58 f 0.33

0.06 f 0.01 2.60 f 0.07

Not tested Not tested

0.45 f 0.03 12.93 f 0.54

2.32 f 0.17 4.93 f 0.16

13.15 f 0.53 12.44 & 0.44

3.16 f 0.05 3.35 f 0.13


11.21 -+ 0.37 10.06 f 0.45


13.56 f 0.38 12.27 f 0.36

19.67 f 0.88 17.34 f 0.73

3.50 f 0.12 2.64 f 0.04

TABLE IV Degradation of ’ OH-treated proteins in the presence and absence of ATP and leupeptin

All proteins were first exposed to 150 nmol of ‘OH/0.33 mg as described in the legends to Tables 1-111 and were then incubated with E. coli or rabbit RBC cell-free extracts (at 37 “C and pH 7.8) to measure proteolytic susceptibility. Proteolysis was measured as described under “Materials and Methods” and in the legend to Table 111. Results are means f S.E. of three independent determinations.

Cell-free extracts Protein degraded during 60-min incubation and additions Bovine serum albumin Superoxide dismutase Catalase Hemoglobin Globin Casein

% E. coli

No additions 3.5 f 0.1 1.9 f 0.1 8.6 f 0.9 5.6 f 0.2 14.7 f 1.4 18.5 f 1.1 3.0 mM ATP 3.1 f 0.1 1.7 f 0.1 7.7 f 0.3 4.7 -C 0.4 13.3 f 1.8 27.1 f 0.8

Erythrocyte No additions 2.7 f 0.1

0.1 mM leupeptin 2.6 f 0.1 3.0 mM ATP 2.5 zk 0.1

1.8 f 0.1 2.2 f 0.1 2.7 f 0.1 4.2 f 0.1 3.2 f 0.1

1.7 f 0.1 2.2 f 0.1 2.1 f 0.2 3.9 f 0.1 3.7 f 0.2 1.9 f 0.1 2.0 f 0.2 2.9 * 0.1 3.9 f 0.3 7.1 f 0.7

Reticulocyte No additions 12.3 f 0.4 3.6 f 0.2 14.4 f 0.4 12.9 f 0.5 11.1 f 0.4 13.6 f 0.7 3.0 mM ATP 12.1 f 0.4 3.4 * 0.2 15.1 f 0.4 12.1 f 0.6 10.8f 0.5 27.2 f 1.8 0.1 mM leupeptin 12.4 f 0.2 3.3 f 0.3 13.5 f 0.6 13.6 f 0.7 10.6f 0.6 11.2 f 0.6

tually have induced slight decreases in proteolytic susceptibility (Table 111).

Degradation of ‘OH-treated proteins was optimal at pH 7.8, indicating the involvement of alkaline (nonlysosomal), proteolytic systems. Degradation of ’ OH-treated proteins in the RBC-soluble extracts was also not significantly inhibited by leupeptin (Table IV), indicating that lysosomal (36) or Ca2+-activated thiol (37) proteases do not play major roles. Furthermore, since all extracts were dialyzed against a (nominal) M , cutoff 12,000 membrane, the results would suggest that ’OH-modified proteins need not be conjugated with

ubiquitin (38, 39) to stimulate proteolysis. The degradation of untreated casein was increased by addition of ATP, as previously reported (7, 40); but the breakdown of other untreated proteins was not affected (data not shown). ATP also did not affect the degradation of most ’OH-treated proteins (Table IV), thus differentiating this process from ATP-stimulated proteolytic pathways in E. coli and RBC (24, 40). The degradation of ’ OH-treated casein was somewhat stimulated by ATP (Table IV). This ATP stimulation, however, was identical to that seen with untreated casein, suggesting that the “extra” degradation could be explained by breakdown of


unmodified protein molecules. A variety of purified proteases were also able to recognize and degrade 'OH-treated proteins selectively at significantly higher rates than untreated proteins. This finding is further explored in the fourth paper of this series (3).

Amino Acid Modifications-Isoelectric focusing gels revealed alterations in net electrical charge following exposure to oxygen radicals (Fig. 1). Both negative charge changes and positive charge changes were observed with various proteins. Transferrin was the least affected of all proteins studied, but careful inspection revealed minor changes. BSA, which does not appear in Fig. 1, became significantly more basic following exposure either to 'OH or to 'OH + 0; (1). Globin, like Hb, became more acidic with oxygen radical exposure (globin was actually more sensitive than Hb). Such charge changes are indicative of modified amino acid residues. The only consistent pattern to emerge is that 'OH alone and 'OH + 0; (+OJ inducted charge modifications of the same sign (+ or -) in each protein. Since 0; alone had no measurable effect on electrical charge (not shown), the data suggest that all charge changes were initiated by 'OH. Oxygen or 0; may have reacted with protein radicals induced by 'OH (during exposure to 'OH + 0;) to modify further the primary charge changes.

Exposure to 'OH or to 'OH + 0; induced a clear loss of native tr-yptophan fluorescence in the 11 proteins which could be studied (Table V). The fluorescence method used is not suitable for many proteins (especially heme proteins) due to fluorescence quenching. Tryptophan damage was similar for both ' OH-treated proteins (35.7 f 6.2% loss) and 'OH + 0;- treated proteins (33.1 f 6.4% loss). In contrast with these results, 0; alone had no effect on the tryptophan fluorescence of the 11 proteins studied (data not shown).

Another indication of extensive protein modification was the production of bityrosine (Table VI). Bityrosine is a covalently bound biphenol, produced by reaction of two tyrosyl radicals or a tyrosyl radical plus a tyrosine molecule (33).

FIG. 1. Charge changes induced by oxygen radicals. Proteins were exposed to 150 nmol of 'OH ( A ) or (in the presence of 1OO"O 02) to 75 nmol of 'OH + 75 nmol of Or ( R ) at a ratio of 150 nmol of oxygen radicals/0.33 mg of protein. Isoelectric focusing gels (pH 3-10) were run as described under "Materials and Methods" using 20 pg of protein/ lane. Gels were stained with silver stain. u. untreated control proteins; t , oxygen radical-treated proteins; Cat, catalase; Cyt c, ferric-ytochrome c; Hb, met-hemoglobin; Mb, metmyoglohin; Perox, peroxidase: 7:vro, tyrosinase; LDH, lactate dehydrogenase; Hex, hexokinase; HSA, human serum albumin; Oval, ovalbumin; S O D , superoxide dismutase; Trans, transferrin.

Bityrosine may be more likely to form between two protein molecules (intermolecular bonding) than within a single protein (intramolecular bonding). Rityrosine formation may thus have been an important factor in the aggregation seen with ' OH-treated proteins. The smaller bityrosine production re- sulting from treatment with 'OH + 0; (Table VI) may have been caused by competing reactions between tyrosyl radicals and 0; or OL (33). No bityrosine was observed following exposure to 0; alone (data not shown).

DISCUSSION

Recent studies have focused attention on the possible role(s) of active oxygen species in protein damage and degradation (41-44). In this paper, I have shown that 'OH can increase the proteolytic susceptibility of diverse proteins. Soluble, alkaline, proteolytic systems (of proteases and peptidases) from RBC and E. coli can recognize and rapidly degrade ' OH-treated proteins. Unlike nonsense/missense mutations or certain premature termination products (24,40, 45, 46), 'OH-treated proteins are degraded by an ATP-independent pathway. This pathway may include ATP-independent neutral and alkaline proteol-ytic enzymes which are known to exist both in the RBC cytosol (40, 47-49) and in E. coli (50). Lysosomal and Ca"-activated thiol proteases (36, 37) do not appear to be necessary for the degradation, and ubiquitin (38, 39) may not be required.

In previous work (4,6), it was shown that exposure of intact RBC to a variety of (enzymatic and chemical) oxygen radical- generating systems resulted in rapid protein damage and degradation. Furthermore, exposure of RBC proteins, casein, or BSA t.o the same oxygen radical-generating systems resulted in their recognition and rapid degradation during subsequent incubations with cell-free RBC extracts (5, 7). Alka- line, ATP-independent proteolytic systems were suggested to catalyze this degradation. On the basis of inhibitor profiles, it was proposed that 'OH might be the active oxygen species which was responsible for both protein damage and increased

I- u t u t u t u t u t u t u t ~ l l u t u t u I u t o Cat Cyt s Hb Mb Perox Tyro LDH Hex HSA Oval SOD Trans

@ Exposure to .OH+02-

8 @ U t U t U I U I U I U I u I i r 1 u t " I u I , , I Cat Cyt Hb Mb Perox Tyro LDH H e x HSA Oval SOD Trans

9900 Protein Damage and Degradation by Oxygen Radicals TABLE V

Loss of tryptophan following exposure of proteins to oxygen radicals Proteins (0.33 mg/ml) were exposed to 150 nmol of 'OH or to 75

nmol of 'OH + 75 nmol of 0; (+OJ. Results are percentage loss of tryptophan fluorescence emission intensity (which is dependent upon, but not synonymous with, the number of tryptophan residues) measured at 280 nm excitation and 340-350 nm emission. Fluorescence studies were performed at pH 7 with 20 mM HEPES buffer. An Aminco-Bowman spectrophotofluorometer (American Instrument Co.) was calibrated daily with 0.1 and 0.01 pg/ml solutions of quinine sulfate (prepared by successive dilution in 0.1 N sulfuric acid). The calibration procedure was conducted eactly as described in the 1979 Operator's Manual (catalog J4-8960A). Values are means of three independent determinations for which standard errors were less than 10%.

Tryptophan loss

' OH 'OH + 0; %

Protein

Lactic dehydrogenase Transferrin Bovine serum albumin Human serum albumin Superoxide dismutase Globin Hexokinase Peroxidase cy-Casein Trypsin Chymotrypsin

45 26 67 44 25 73 31 31 30 6

15

45 27 70 46 10 64 22 22 27 14 17

TABLE VI Production of bityrosine following exposure of proteins to oxygen

radicals Proteins (0.33 mg/ml) were exposed to 150 nmol of 'OH or to 75

nmol of 'OH + 75 nmol of 0 2 (+02). Results are bityrosine content as judged by fluorescence measurements at 325 nm excitation and 410-420 nm emission. Bityrosine fluorescence was measured in 20 mM HEPES buffer (pH 7), and results are expressed for 1 mg/ml protein solutions. An Aminco-Bowman spectrophotofluorometer was calibrated with quinine sulfate exactly as described in the legend to Table V. Values are means of three independent determinations for which standard errors were less than 10%.

Protein Bityrosine fluorescence

Untreated 'OH 'OH + 0, Lactic dehydrogenase 0.29 1.95 1.19 Transferrin 0.17 2.55 1.06 Bovine serum albumin 0.25 2.68 1.78 Human serum albumin 0.55 2.85 1.66 Ovalbumin 0.21 1.02 0.64 Superoxide dismutase 0.13 0.55 0.38 Globin 4.85 12.45 5.03 Hexokinase 0.25 2.30 1.91 Peroxidase 0.17 1.40 1.19 cy-Casein 0.21 0.85 0.33 Trypsin 1.06 4.85 2.42 Chymotrypsin 0.46 3.19 2.46

proteolytic susceptibility, but no definitive evidence was obtained. The present work demonstrates that 'OH is certainly capable of inducing both damage and increased proteolytic susceptibility.

All proteins studied were susceptible to modification by 'OH or by 'OH + 0; (+02). The modifications included altered molecular weight (aggregation or fragmentation), altered net electrical charge (+ or -), loss of tryptophan, and production of bityrosine. Since 0, alone had no measurable effect on any of these parameters, I propose that 'OH was the initiating species in all cases. Oxygen and/or 0; appear to modify significantly the damage induced by 'OH.

Protein fragmentation has previously been reported following exposure to ' OH + 0; + 0, (34, 35). The process of

fragmentation was suggested to involve hydrogen abstraction by 'OH from amino acid a-carbon atoms, followed by reaction with 0, to produce peroxyl species (34). Decomposition of a- carbon peroxides was proposed as the mechanism for protein fragmentation (34). In the third paper of this series (2), it is shown that the fragmentation of BSA by exposure to 'OH and 0; (+O,) is broadly consistent with this scheme.

Several forms of protein damage resulted from 'OH exposure. Thus, it is possible to begin to relate protein modification to proteolytic susceptibility. Protein aggregation might appear to be an attractive signal for proteolysis, but no such claim could be made on the basis of this investigation, Indeed, it should be noted that ' OH-treated BSA, superoxide dismutase, catalase, and Hb were all good substrates for proteolytic digestion, yet aggregation varied from 98 to 41%. Both globin and casein exhibited aggregation following 'OH exposure, yet no increases in proteolytic susceptibility were observed; in fact, the degradation of globin and casein generally decreased after ' OH exposure.

Previously, it was reported that exposure of casein to H,O, + Fez+ or to T o radiation increases proteolytic susceptibility (7). The previous experiments were conducted under an air atmosphere which contains sufficient oxygen for casein fragmentation to occur. Recent work (43, 44) indicates that the previous results (7) can be explained by degradation of casein fragments. Casein fragmentation products are clearly formed during 'OH exposure in the presence of 0, (Table 11). Such protein fragments are excellent substrates for RBC proteases (e.g. see Ref. 3) but are not formed during 'OH treatment in the absence of oxygen (Table 11).

Changes in net electrical charge were observed for BSA, superoxide dismutase, catalase, Hb, globin, and casein (as well as most other proteins), yet the sign of the charge change (+ or -) did not correlate with proteolytic susceptibility. Tryp- tophan destruction and bityrosine production were also common features of ' OH exposure, yet the degree of modification again bore no apparent relationship to proteolytic susceptibility.

From the present work, it is only possible to state that protein denaturation by 'OH results in increased proteolytic susceptibility. Denatured proteins are generally found to be excellent substrates for intracellular proteolytic systems and for purified proteases (3,5, 7-9, 12,24,37,40-47,501. Denat- uration involves protein unfolding and increased accessibility of peptide bonds (to proteases). Thus, the simple process of denaturation could provide an attractive universal signal for the degradation of oxidatively modified proteins. Since globin and casein may be considered to be denatured proteins, it is possible that their rates of degradation were already maximal before 'OH exposure. More detailed work was clearly required, and the relationship between denaturation and proteolytic susceptibility has been further explored with the representative protein BSA in the accompanying papers (1- 3).

Acknowledgments-I am grateful to Dr. Raymond D. Mosteller for his advice and helpful discussions on the E. coli studies. I also thank Sharon W. Lin and Robert E. Pacifici for their invaluable work.

REFERENCES 1. Davies, K. J. A., Delsignore, M. E., and Lin, S. W. (1987) J. BWl.

2. Davies, K. J. A., and Delsignore, M. E. (1987) J. BWl. Chem.

3. Davies, K. J. A., Lin, S. W., and Pacifici, R. E. (1987) J. BWl.

4. Davies, K. J. A. (1985) in Cellular and Molecular Aspects of Aging: The Red Cell (1s a Model (Eaton, J. W., Konzen, D. K., and

Chem. 262,9902-9907

262,9900-9913

Chem. 262,9914-9920


White, J. G., eds) pp. 15-2’7, Alan R. Liss, Inc., New York 28. Anderson, B. L., Berry, R. W., and Telser, A. (1983) Anal. 5. Davies, K. J. A. (1986) in Superoxide and Superoxide Dismutase Biachem. 132,365-375

in Chemistry, Biology and Medicine (Rotilio, G., ed) pp. 443- 29. Laemmli, U. K. (1970) Nature 227, 880-885 450, Elsevier/North-Holland Biomedical Press, Amsterdam 30. Gorg, A., Posterl, W., and Westermeier, R. (1980) in Electropha-

6. Davies, K. J. A., and Goldberg, A. L. (1987) J. Bial. Chem. 262, resis ‘79 (Radola, B. J., ed) pp. 67-68, Walter de Gruyter & Co., 8220-8226 New York

7. Davies, K. J. A., and Goldberg, A. L. (1987) J. Biol. Chem. 262, 31. Radola, B. J. (1980) in Electropharesis ‘79 (Radola, B. J., ed) pp. 8227-8234 80-94, Walter de Gruyter & Co., New York

8. Levine, R. L., Oliver, C. N., Fulks, R. M., and Stadtman, E. R. 32. Teale, F. W. J. (1960) Biochem. J. 76,381-388 (1981) Proc. Natl. Acad. Sci. U. S. A. 78, 2120-2124 33. Priitz, W. A., Butler, J., and Land, E. J. (1983) Znt. J. Radiat.

9. Fucci, L., Oliver, C. N., Coon, M. J., and Stadtman, E. R. (1983) Bial. Relat. Stud. Phys. Chem. Med. 44, 183-196 Proc. Natl. Acad. Sci. U. S. A. SO, 1521-1525 34. Garrison, W. M., Jayko, M. E., and Bennett, W. (1962) Radiat.

10. Levine, R. L. (1983) J. Bial. Chem. 258,11823-11827 Res. 16,483-502 11. Levine, R. L. (1983) J. Btil. Chem. 258,11828-11833 35. Schuessler, H., and Schilling, K. (1984) Znt. J. Radiat. Bial. Relat. 12. Rivett, A. J. (1985) J. Bial. Chem. 260, 300-305 Stud. Phys. Chem. Med. 45,267-281 13. Dubiel, W., Muller, M., and Rapoport, S. (1981) Biachem. Znt. 3, 36. Ivy, G. O., Schottler, F., Wenzel, J., Baudry, M., and Lynch, G.

165-171 (1984) Science 226, 985-987

14. Dean, R. T., and Pollak, J. K. (1985) Btihem. Biaphys. Res. 37. Melloni, E., Spartore, B., Salamino, F., Michetti, M., and Pon-

Commun. 126.1082-1089 tremoli, S. (1982) Biochem. Biophys. Res. Commun. 106, 731- - .^ 15. Chin, D. T., Kuehl, L., and Rechsteiner, M. (1982) Proc. Natl. ‘140

Acad. Sci. U. S. A. 79,5857-5861 38. Hershko, A., Ciechanover, A., Heller, A., Haas, A. L., and Rose,

16. Hendil, K. B. (1980) J. Cell. Physial. 105,449-460 I. (1980) Proc. Natl. Acad. Sci. U. S. A. 77, 17831786

17. Goldberg, A. L., and Boches, F. S. (1982) Science 215, 1107- 39. Ciechanover, A., Elias, S., Heller, H., Ferber, S., and Hershko, A.

1109 (1980) J. Bial. Chem. 255, 7525-7528

18. Fridovich, I. (1983) Annu. Rev. Pharmacol. Toricol. 23, 239-257 40. Tanaka, K., Waxman, L., and Goldberg, A. L. (1984) J. Biol.

19. Fricke, H., and Hart, E. J. (1966) in Radiation Dosimetry (Attix, Chem. 259,2803-2809

F. H., and Roesch, W. C., eds) p. 167, Academic Press, New 41. Stadtman, E. R. (1986) Trends Biochem. Sci. ll, ll-12 York 42. Wolff, S. P., Garner, A., and Dean, R. T. (1986) Trends Biachem.

20. Bielski, B. H. J., and Gebicki, J. M. (1977) in Free Radicals in Sci. 11,27-31

Biology (Pryor, W. A., ed) Vol. 3, pp. 2-51, Academic Press, 43. Davies, K. J. A. (1986) J. Free Radicals Biol. Med. 2, 155-173

New York 44. Davies, K. J. A. (1987) in Cellular Antioxidant Defense Mecha-

21. Alexander, P., and Lett, J. T. (1967) in Comprehensive Biachem- nisms (Chow, C. K., ed) CRC Press, Inc., Boca R&on, FL, in

istry (Florkin, M., and Stotz, E. H., eds) Vol. 27, pp. 267-356, press

Elsevier/North-Holland Biomedical Press, Amsterdam 45. Etlinger, J., and Goldberg, A. L. (1977) Proc. Natl. Acad. Sci. U.

22. Singh, H., and Vadasz, J. A. (1983) Znt. J. Radiat. Bial. Relat. S. A. 74,54-58

Stud. Phys. Chem. Mod. 44,601-606 46. Goldberg, A. L. (1972) Proc. Natl. Acad. Sci. U. S. A. 69, 422-

426 23. Simonian, M. H., Goldstein, R. V., and Mosteller, R. D. (1978) 47. p on t

J Bacterial. 133,650-660 remoli, S., Melloni, E., Salamino, F., Sparatore, B., Michetti,

24. Murakami, K., Voellmy, R., and Goldberg, A. L. (1979) J. Bial. M., Benatti, U., Morelli, A., and De Flora, A. (1980) Eur. J.

Chem. 254,8194-8200 Biachem. 110,421-430

25. Rice, R. H., and Means, G. E. (1971) J. Biol. Chem. 246, 831- 48. Daniels, R. S., McKay, M. J., Worthington, V. C., and Hipkiss,

832 A. R. (1982) Biochim. Biophys. Acta 717, 220-227

26. Jentoft, N., and Dearborn, D. G. (1979) J. Bial. Chem. 254, 49. McKay, M. J., and Hipkiss, A. R. (1982) Eur. J. Biochem. 125,

567-573 4359-4365

27. Reisfield, R. A.. Lewis. U. J.. and Williams. D. E. (1962) Nature 50. Swamy, K. H. S., and Goldberg, A. L. (1981) Nature 292, 652-

654 195,i81-283 51. Bradford, M. M. (1976) Anal. Biochem. 72,248254

Documents

Protein Damage and Degradation by Oxygen Radicals · Vol. 262. No .20, Issue of July 15, pp. 9895-9901.1987 Printed in U.S.A. Protein Damage and Degradation by Oxygen Radicals I