Protein Damage and Degradation by Oxygen Radicals

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society of Biological Chemists, Inc. Vol. 262, No. 20, Issue of July 15,pp. -9913, 1987 Pri...
Author: Allan Marsh
0 downloads 2 Views 2MB Size
THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society of Biological Chemists, Inc.

Vol. 262, No. 20, Issue of July 15,pp. -9913, 1987 Printed in

U.S. A.

Protein Damage and Degradationby Oxygen Radicals 111. MODIFICATION OF SECONDARY AND TERTIARYSTRUCTURE* (Received for publication, December 30, 1986)

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

Proteins which have been exposed to the hydroxyl Oxygen radicals can modify proteins and enhance their radical ('OH) or to the combination of'OH plus the degradation by intracellular proteolytic systems (1). This superoxide anion radicaland oxygen ('OH + 0; + 0,) paper is the third ina series of four (1-3) which attempts to exhibit altered primary structure and increased pro- relate oxidative modification to increased proteolytic suscepteolytic susceptibility. The present work reveals that tibility. alterations to primary structure result in gross distor-In thepreceding paper (2),the hydroxyl radical ( ' OH) and tions of secondary and tertiary structure, Denaturation/increased hydrophobicity of bovine serum albu- the combination of 'OH the superoxide anion radical (0;) were shown to induce dose-dependent alterations to the primin (BSA) by*OH, or by 'OH + 0; + 0, was maximal mary structure of bovine serum albumin (BSA).' Albumin at a radical/BSA molar ratio of 24 (all 'OH or 50% was chosen as a representative model protein on the basis of -OH + 50% 0;). BSA exposed to *OH also underwent progressive covalent cross-linking to form dimers, tri- comparative studies (I), purity, ubiquity, and a lack of (commers, and tetramers, partiallydue to the formationof plicating) prosthetic groups. In contrast with 'OH alone or 0; (+02), exposure to 0; (+02)did not affect intermolecular bityrosine. In contrast, *OH + 0; + 0, 'OH the primary structure of BSA (2). The previous results (1, 2) caused spontaneous BSA fragmentation. Fragmentation of BSA produced new carbonyl groups with no and those of other investigators (4-10) indicate that 'OH is apparent increase in free amino groups. Fragmenta- a powerful protein-modifying agent. In contrast, 0; would tion may involve reaction of (*OH-induced)a-carbon appear to be relatively unreactive (1, 2). The combination of may represent biological exposure to oxygen radicals with0, to form peroxyl radicals which decom- 'OH + 0; (+02) pose to fragment the polypeptide chain at the a-carbon, radicals (e.g. from xanthine oxidase, quinone oxidation, mirather than atpeptide bonds. BSA fragments induced tochondrial/microsomal electron transport, etc.) better than by *OH + 0; + 0, exhibited molecular weights of either 'OH alone or 0; alone. In agreement with other reports 7,000-60,000 following electrophoresis under dena(4, 7,8), we find that 0; and O2 significantly alter theeffects turing conditions, but could bevisualized as hydropho- of 'OH on protein primary structure (2). bic aggregates in nondenaturing gels (confirmed with The oxidative modifications to protein primary structure [3H]BSAfollowing treatment with ureaor acid). Com- previously reported (2) could be expected to have significant binations of various chemical radical scavengers(mannitol, urate, t-butyl alcohol, isopropyl alcohol) and effects on protein secondary and tertiary structure.Such gross structural modifications might well be recognized byintracelgasses (N20,02,N,) revealed that *OH is the primary species responsible for alteration ofBSA secondary lular proteases and peptidases, resulting in increased proteoand tertiary structure. Oxygen, and 0; serve only to lytic susceptibility. Increased degradation due to recognition of abnormal secondary and tertiary structure could explain modify the outcome of *OH reaction. Furthermore, direct studies of 0;+ 0, (in the absence of 'OH) re- the proteolysis observed when red cells (11-17), Escherichia vealed no measurable changes in BSA structure. The coli (12, 13, 15, 18-23), muscle cells (12, 13, 15), or mitochonprocess of denaturation/increased hydrophobicity was dria (12, 13, 15, 24, 25) are exposed to oxygen radicals and foundto precede either covalent cross-linking (by other active oxygen species (e.g. H202).In addition, abnormal *OH)or fragmentation (by *OH+ 0; + 0,).Denatura- secondary and tertiary structuremay underlie the breakdown tion was half-maximal at a radical/BSA molar ratio of of mitochondria which occurs during reticulocyte maturation 9.6, whereas half-maximal aggregation or fragmen- and which may be initiated by oxidative damage/modification tation occurredat a ratio of 19.4. Denaturationbydro- (26). phobicity mayhold important clues for the mechaIn thispaper, we have tested the dose-dependent effects of nism(s) by which oxygen radicals can increaseproteo- 'OH, 0; (+02), and 'OH + 0; (+02) on the secondary and lytic susceptibility. tertiary structure of BSA. Conditions have been chosen to exactly match our previous study of BSA primary structure (2) and our survey report of 16 other proteins (1). In the * This work was supported by Grant ES 03598 from the National subsequent paper (the fourth and last of this series) (3), we Institutes of Health (to K. J. A.D.). Part of this work has been published in preliminary form (Davies, K. J. A., and Delsignore, M. have carefully correlated oxidative modifications to primary, with proteolytk susceptibilE. (1984)Fed. Proc. 43,1858 (Abstr. 2579).The costs of publication secondary, and tertiary structure of this article were defrayed in part by the payment of page charges. ity in an attempt touncover a causal relationship.



This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Present address: Colgate-Palmolive Co., 909 River Rd., Piscataway, NJ 08854.

The abbreviations used are: BSA, bovine serum albumin; SDSPAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; acid. HEPES, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic


Protein Damage and Degradation by Oxygen Radicals



Materials-The experiments reported in this paper were performed with a fatty acid- and globulin-free BSA ( M . 66,200) from Sigma (A 0281). In preliminary experiments with other BSA preparations (not shown), Sigma product A 4378 and Miles Scientific (Naperville, IL) products 81-001-2, 81-028-2, and 81-018-1 gave very similar results. Cross-linked BSA (Sigma A 9392) was used as an electrophoretic standard for BSA dimer, trimer, and trimer bands, in addition to conventional molecular weight markers. was exposed to Exposure of Proteins toOxygenRadicakr-BSA ‘OH alone, to 02 alone, or to ‘OH + 0; using @ C ‘o radiation. Exposure to ’OHalone was achieved by irradiation under 100% N20. Exposure to 0;alone was achieved by irradiation under 100% 0,in the presence of 0.01 M sodium formate. Exposure to ‘OH + 0; involved irradiation under 100% 02.Irradiation was conducted a t 25 “C with 5.0 PM solutions of BSA (0.33 mgof protein/ml) in double distilled and deionized water (no buffer). The dose rate was 634 & 5 rads/min as measured by Fricke and Hart dosimetry (27). Radiation times were varied in order to achieve oxygen radical exposures of approximately 1-120 nmol of radicals/nmol of BSA (total doses of 0.8-100 kilorads). These procedures are described in detail inthe first paper of this series (1). Protein Denaturation-The denaturation of BSA was assessed by studies of solubility in high salt buffer. In high salt, the repulsion of hydrophobic groups by the solvent cage is maximized. In the native (folded) state, hydrophobic residues are shielded from the aqueous environment, but denaturation and unfolding diminish this protection. At ita isoelectric point, a protein has no net electrical charge, and thesolubilizing properties of hydrophilic residues are minimized. Thus, a denatured, unfolded protein exhibits decreased solubility a t its isoelectric point in high salt buffer because its exposed hydrophobic groups cluster together and cause precipitation. In this study, we assessed denaturation (following exposure to oxygen radicals) by loss of soluble protein in 50 mM sodium succinate buffer (pH 4-6) containing 3.0 M KCl. Protein solutions were diluted with this buffer and left for 60 min on ice. Subsequently, the solutions were centrifuged (2500 X g) for 10 min, and the remaining soluble protein was measured by both the assays of Bradford (28) and Lowry et al. (29). Protein Aggregation and Fragmntatwn2-Aggregation and fragmentation of BSA were measured by polyacrylamide gel electrophoresis (PAGE). These experimentsutilized nondenaturing PAGE (30), as well as sodium dodecyl sulfate-PAGE (SDS-PAGE)in theabsence (31) or presence (32) of urea. All three gel types were run with and without dithiothreitol or 8-mercaptoethanol, and gels were stained either with Coomassie Brilliant Blue R-250or with silver stain (BioRad). Loss of the monomeric BSA staining band and production of aggregation or fragmentation productswere quantitated by scanning densitometry and computer-assisted analysis (1). Aggregation and fragmentation were also measured by electrophoresis gels of 3H-labeledBSA using autoradiography. BSA was labeled, prior to oxygen radical exposure, by reductive methylation with sodium borohydride and [3H]formaldehyde (33) as previously described (1). Additionally, production oflow molecular weight fragmentation products was measured by increases in acid-soluble counts following treatment with 5% trichloroacetic acid and centrifugation (2500 X g ) for 10 min. Free amino groups were assayed by reaction with fluorescamine (34) both before and after acid precipitation of intact protein. Carbonyl groups were measured by reaction with 2,4-dinitrophenylhydrazine as described by Lappin and Clark (35).

I ‘ j L pa

4. 0ooo4. ,o.05 5. 0

5. 5 6. 0

4. 0 4. 5 5. 0 5. 5 6.

pH of 3. OM KC1 pH o f 3.0H KC1 FIG. 1. Loss of BSA solubility following exposure to oxygen radicals, as a function of pH.BSA was exposed to ‘OH or to ’OH 0; + O2 a t a molar ratio of 10 (A) or 100 (0)nmol of radicals/ nmol of BSA (all ’OH or 50% ’OH 50% OF). Samples were then made 3.0 M in KC1 and 50 mM in sodium succinate (pH 4-6). Following incubation and centrifugation (see “Experimental Procedures”), the remaining soluble protein was measured. The results express solubility as a percent of control (untreated) BSA using the averages of both the procedures of Lowry et al. (29) and Bradford (28). Values are means of three independent determinations (with each procedure) for which standard errors were less than 10%.







20 40 60 BO 100 120 nmol Oxygen Rodicals/nmol BSA

FIG.2. BSA denaturation/hydrophobicityas a function of oxygen radical exposure. BSA was exposed to ‘OH, to ‘OH + 0;+ 02,or to 0;+ O2 at theradical/BSA molar ratios indicated (all ‘OH, 50% ‘OH + 50% OF, or all OF). Samples were then suspended in 3.0 M KC1 + 50 mM sodium succinate (pH 4.0). Percent denaturation was estimated by the loss of soluble protein measured as described in thelegend to Fig. 1. Values are means of three independent determinations for which standard errors were less than 10%.

or denaturation/increased hydrophobicity (see “Experimental Procedures”). Mannitol completely inhibited solubility losses, indicating that ‘OH may have been the initiating species in RESULTS all cases. Protein Denaturation-Significant loss of solubility in KC1 The relationship between oxygen radical dose and BSA was observed when BSA was exposed to ‘OH alone or to denaturation/increased hydrophobicity was confirmed by de‘OH + 0; (+02). The decreased solubility was pH-dependent tailed studies at pH 4 (Fig. 2). Denaturation by ‘OH or by and was more pronounced at a radical/BSA ratio of 100 than ‘OH + 0; (+02) was essentially maximal at a radical/BSA a t a ratio of 10 (Fig. 1). Decreased salt solubility close to the ratio of 24. Denaturation by ‘OH alone was more severe than isoelectric point of BSA is consistent with protein unfolding denaturation by .OH + 0 2 (+02) at all radical/BSA ratios tested. Exposure to 0; (+OZ), however, caused no measurable The term “protein fragmentation”refers to thedirect breakdown denaturation (Fig. 2). The data of Figs. 1 and 2 have been of proteins by oxygen radicals. Such processes have been found to involve both main chainscission and side chain scission and therefore corrected for small changes in protein content registered by differ from peptide hydrolysis. In contrast, the term“protein degra- both the assays of Lowry et al. (29) and Bradford (28) following treatment with ‘OH or with ‘OH + 0; + O2(see below). dation” refers to peptide bond hydrolysis by proteolytic enzymes.

Protein Degradation Damage and


by Oxygen Radicals

Aggregationand Fragmentation-Nondenaturing PAGE (30) indicated that both 'OH alone and 'OH 0; (+Or) induced BSA aggregation(Fig. 3). The addition of dithiothreitol had h t l e effect, indicating that new disulfide bond formation was not a major cause of aggregation. In addition to aggregation, the 'OH + 0; (+Or)-treatedsamples also exhibitedsignificantamounts of materialwith mobility greaterthanthat of the BSA monomerband.This high mobility material consists of BSA fragmentation products as shown below. Denaturing SDS-PAGE in the presence of urea (32) revealed a similar pattern of aggregation for BSA treated with 'OH, but the combinationof 'OH + 0; (+O,) produced only low molecular weight fragmentation products (Fig. 4). SDSPAGE in the absence of urea (31) also revealed this same pattern of aggregation with 'OH andfragmentationwith ' O H 0; + 0, (not shown). Comparisonsof denaturing and nondenaturing gels indicat,e that 'OH + 0; (+Or) induced the fragmentation of BSA and that the fragments formed random conglomerates which were held together by hydrophobic and ionic bonds during nondenaturingPAGE. Dithiothreitol and Lj-mercaptoethanol had little or no effect on the aggregation or fragmentation profiles seen with SDS-PAGE in the presence or absence of urea. Densitometric analyses indicated that more than 90% of the BSA aggregates induced




1 7 5 5 ~ ~ 1 0 1 ? 5 1 5 ? 0 7 5 5 0 1 5 1 0 0o

nmol .OH / nmol BSA

~ ~ 5 5 r 5 ~ o ~ ~ s ~ s ? o ~ nmol . O H t O 2 -

by 'OH were due to new covalent bonds other thandisulfide bonds. The aggregation products induced by 'OH exhibitedmolecular sizes which initially corresponded with cross-linked BSA dimer and trimer standards (Fig. 4). SDS-PAGE withouturea (31) also revealed a low yield of products which may have been BSA tetramers.Withincreasing'OH exposure, the assumed dimer and trimer bands of Fig. 4 declined slightly in molecular size, although no low molecular weight fragments were observed. The BSA fragmentation products induced by 'OH + 0; (+02) appearedto be confined to amolecular weight range of 66,200 down to approximately7,000. Autoradiographic studies of['HH]BSA exposed to 'OH + 0; (+Or) and electrophoresed under identical conditions confirmed this finding. Following SDS-PAGE with urea ( 3 2 ) ,densitometric scans revealed a clear dose-dependent relationshipbetween 'OH or 'OH + 0; (+Or) production and loss of the BSA monomer band (Fig. 5 A ) . In contrast with these results, 0; (+Or) had no effect. The loss of BSA monomer with progressive 'OH or 'OH + 0; (+Or) treatment was inversely relatedtothe appearance of aggregation ( 'OH) and fragmentation ( ' OH + 0; + Or) product,s (Fig. 5 R ) . Comparisons of Figs. 2 and 5 ( A and B ) indicate that denaturation was more rapid than was aggregation or fragmentation. It appearspossible that protein unfolding may facilitate the aggregation/fragmentation processes. Protein measurements, such as the assaysof Bradford ( 2 8 ) or Lowry et al. (29),are stronglyinfluenced by molecular size andconformation.Theseassays indicated 15-20% protein losses following exposure to.OH or to.OH + 0; (+Or) (TableI).Inthe presence of thedenaturingagenturea, however, BSA treated with 'OH + 0; (+Or) exhibited much greater (40-50%) protein loss, whereas BSA treated with 'OH showed nofurther change. Similarresults were also obtained using SDS as the protein-denaturing agent not (data shown). These findings are consistent with the patternemerging from Figs. 3-5, namely that 'OH induces covalent aggre: , s o : whereas ~ ~ n ~ ~ 'OH + 0; (+Or) produces fragments which gation,

nmol BSA

FIG. 3. Nondenaturing polyacrylamide electrophoresis gels of BSA exposed to oxygen radicals. RSA was exposed to 'OH ( A ) or to ' O H + 0; + Or ( R ) at the radical/BSA molar ratios indicated (all 'OH or 50% 'OH + 50% OF). The nondenaturing gels were run as described under "Experimental Procedures" and in Ref. 30 and were visualized with silver stain.




nmol Oxygen Rodlcols/nmol BSA


1 2 . 5 5 7 . 5 10 12.515 2 0 2 5

5 0 7 5 1 0 0 0 1 2 5 5 7 5 10 12.5 15 2 0 2 5 5 0 7 5 100

nmol .OH / nmol BSA

nmol . O H t 0 2 - / nmol E S P

FIG.4. Denaturing SDS-polyacrylamide electrophoresis gels of BSA exposed to oxygen radicals. RSA was exposed to ' O H ( A ) or to 'OH + 0; 0, ( R ) at the radical/BSA molar ratios indicated (all 'OH or 50% 'OH + 50% 0;). SDS-PAGE was conducted in the presence of urea as described under "Experimental Procedures" and in Ref. 32, and protein bands were visualized with silver stain.


FIG. 5. Aggregation or fragmentation of BSAby oxygen radicals. BSA was exposed to 'OH, to 'OH + 0; + O,, or to 0; + O2 at the radical/RSA molar ratios indicated (all 'OH, 50% 'OH + 50% 0;. or all 0;). SDS-PAGE (32) was then conducted exactly as described in the legend to Fig. 4. Percent loss of the BSA monomer band ( A ) and production of aggregates or fragments ( R )were quantified by scanning densitometry (see "Experimental Procedures"). Oxygen radical exposures in A are as follows: A, 'OH; 0, 'OH + 0; + Or; and 0; + 0,. In E?, covalently cross-linked BSA aggregates produced hy 'OH (A) and BSA fragments produced by 'OH + 0; + O2 (0)are expressed (at each radical/BSA ratio) as a percentage of total aggregates or fragments observed at a radical/BSA molar ratio of 120. Values are means of three independent determinations for which standard errors were less than 10%.


Protein Damage and Degradation by Oxygen Radicals


age, in all cases, was 'OH. As with denaturation and alterations to primary structure reported in the previous paper (21, it would appear that thepresence of 0: and 0, merely modifies + the effects of 'OH. Similar protection by isopropyl alcohol and t-butyl alcohol implies that neither the solvated election (e,) nor H ' played a major role in the BSA modifications reported in this study. Tris, HEPES, and carbonate buffers were effective inhibiChange tors of BSA modification, as judged by SDS-PAGE with urea Protein assay 'OH'OH + 0; (32). Effects of these buffers were obvious over the entire 7% radical/BSA dose range of 1-120; but for brevity, only the -19.8 1 -19.9 Method 120 ratioresultsare shown (Table 11). Phosphate buffer -19.2 -41.3 +Urea produced results which were most similar to those obtained in water. It should be noted, however, that the BSA exposed -14.62 -14.8 Method -50.2 -13.7 to 'OH in phosphate buffer formed only one high molecular +Urea weight band (>lSO,OOO), rather than the two bands seen in water (Fig. 2). Additionally, exposure to 'OH + 0; + 02 in phosphate buffer induced only half the BSA fragmentation observed with exposure in water (Table 11).In agreement with our results for BSA primary structure (2), the present data indicate that Tris, HEPES, and carbonate buffers are unsuitable for studies of protein modification by oxygen radicals. Phosphate salts may be suitable alternatives, although such buffers (orcontaminants such asiron)can clearly affect results. X Il . . . . . I I BSA Fragmentation by ' O H + 0; (+O&"he results of 0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30 Figs.3-5do not reveal whether BSA fragmentation was Free Radical/BSA Molor Ratio caused by peptide bond hydrolysis or by some other form of FIG. 6. Inhibition of BSA aggregation and fragmentation chain scission. Since peptide bond hydrolysis produces free by radical scavengers. BSA was exposed to 'OH (---) or to 'OH amino groups, we tested for fluorescamine reactivity. Rather + 0;+ 0,( . . . .) at theradical/BSA molar ratios indicated (all 'OH than an increase in free amino groups, the decreased fluoresor 50% 'OH + 50% O F ) .In A , the results of exposure with 1.0 M t- camine reactivity observed indicates that some amino groups butyl alcohol + 100% N,O (m) and exposure with 1.0 M isopropyl alcohol + 100% N, (A)are shown (these conditions are detailed in were derivatized or otherwise obstructed by exposure to OH 0; + O2 (Table 111). A small, but significant the precedingpapers (1,2)). B shows the results of exposure to 'OH, or to 'OH and C shows the results of exposure to 'OH + 0; + 0,in the presence increase in acid-soluble fluorescamine-reactive material was of 1.0 mM mannitol (o), 10 mM mannitol (o), 50 p M urate (o), or 0.5 observed with 'OH 0; (+02) treatment. The scale of this mM urate (A). Data represent the percent loss of the BSA monomer increase agreed well with the measurements of acid-soluble band in SDS-PAGE. The gels were run in the presence of urea (32) as described in the legends to Figs. 4 and 5 and were quantified by radioactivity (from [3H]BSA)discussed above. Thus, it would scanning densitometry. Data are means of three independent deter- appear that peptide bond hydrolysis did not contributegreatly to the fragmentation induced by 'OH + 0; (+O,). minations for which standard errors were less than 10%. A mechanism for proteinfragmentation by 'OH in the form conglomerates of hydrophobic and ionic bonds. presence of 0, and 0, has been suggested by Garrison et al. To be sure that electrophoresis data did not mask the (4, 36). According to this proposal, chain scission can occur production of very low M , fragments (i.e. less than 5000), we at the a-carbonposition following addition of O2to an('OHmeasured the effects of oxygen radicals on the acid solubility induced) a-carbon radical. The products of such scission of ['HIBSA. Intact proteins and peptides of M , greater than reactions would be an amide and a carbonyl group. Measureapproximately 5000 are readily precipitated by treatment with 0; (+02) 5% trichloroacetic acid. The combination 'OH TABLEI1 and, to a lesser extent, 'OH alone induced small increases in Effects of various buffering agents on BSA aggregation and acid-soluble counts. Even at a radical/BSA molar ratio of 120, fragmentation by oxygen radicals however, the results indicated that species with molecular BSA was disolved in water or in 100 mM buffer and exposed to weights lower than approximately 5000 represent less than 'OH alone or to 'OH + 0;+ 0,.A molar ratio of 120 nmolof oxygen 2% of total products following exposure to 'OH and less than radicals/nmol ofBSA was used (all 'OH or 50% 'OH + 50% OF). 9% of total products following exposure to 'OH + 0; + 100% Samples were then subjected to SDS-PAGE in the presence of urea (321, and aggregation or fragmentationwas quantitated by densitom0,. Also tested were the results of exposure to 'OH 0; etry (asdescribed in the legends to Figs. 5 and 6). Results are means air. These data were essentially identical to those obtained f S.E. of three independent determinations. under 100% O2 and indicate that 20% 0, is sufficient for Fragmentation Aggregation Buffer fragmentation reactions to occur. The electrophoresis data of by 'OH + 0; by 'OH Figs. 3-5 and theacid precipitation data described above were % % further confirmed by both liquid chromatography (Sephadex None 85 f 6 98 f 8 G-100) and high pressure liquid chromatography (Bio-Si1 Tris 3 f l 9 f 2 TSK-250). HEPES 4+1 7f1 Isopropyl alcohol t-butyl alcohol, mannitol, and uric acid Carbonate 29 f 3 45 f 2 Phosphate 97 f 7" 43 f 5 all protected against the loss of the BSA monomer band and production of aggregation or fragmentation products (Fig. 6, "These aggregates had M , >180,000 and would not penetrate the A-C). These findings suggest that the initiator of BSA dam- stacking gel.

TABLE I Effects of oxygen radicals on measurements of total protein BSA was exposed to ' OH or to .OH 0;+ 02 at a molar ratio of 25 nmol of radicals/nmol of BSA (all 'OH or 50% 'OH + 50% OF). Protein assays (furea) were as follows: Method1, the Bradford method (28); and Method 2, the Lowry et al. method (29). Results are means of three independent determinationsfor which standard errors were less than 10%.







Protein Darnuge and Degradation by Oxygen Radicals


TABLEI11 Determination of amino groups with fluorescamine following treatment of BSA with 'OH or 'OH 01(+Od BSA was exposed to 'OH or to 'OH + 0;+ 0,at a molar ratio of 25 nmol of radicals/nmol of BSA (all 'OH or 50% 'OH + 50% 0;). Following exposure, freeamino groups were assayedby reaction with



fluorescamine (32). For total fluorescamine reactivity, BSA was added to sodiumphosphatebuffer(pH 8.0). Whilevortexing, 0.5 ml of fluorescarnine (0.4 mg/ml in acetone) was added. Acid-soluble fluorescamine reactivity was measured following precipitation of protein with trichloroaceticacid. The acid samples were vortexed, lefton ice for 20 min, and centrifuged (2800X g for 10 min). The supernatants ; LL 2 0 2 4 6 8 1 0 were buffered to pH 8.0with 0.1 M sodium borate (pH 10.0).FluoresBSA Concentrotion (mg/ml) camine was then added as described above.Note that the dilution of both the total and acid-soluble sampleswas equal. The fluorescence FIG. 8. Bityrosine production by 'OH as a function ofBSA of bothsampleswasmeasured at 390 nm excitation and 475 nm concentration. BSA, at the concentrations indicated, was exposed emission. The results are absolute fluorescence intensities (meansf to 100 nmol of 'OH, andbityrosineproduction was assessed by S.E.) from three independent determinations. fluorescence emission, as described previously (2). Values are means of three independent determinationsfor which standard errors were Fluorescarnine reactivity Protein less than 10%. treatment None 'OH

'OH + 0;



4.67 f 0.26

0.028 k 0.001 f 0.001 k 0.002

4.18 f 0.23 0.016 3.75 f 0.20 0.056

contributed to BSA aggregation with 'OH treatment (at 0.33 mg of BSA/ml). DISCUSSION

Both -OH and .OH + 0; (+Oz) have now been shown to affect the secondary and tertiary structure of BSA. Presumably, such effects are consequences of the oxidative modifications to primary structure outlined in the previous paper (2). Exposure to 'OH alone causes BSA denaturation/increased hydrophobicity, followedby formation of covalent bonds between BSA molecules (covalent aggregation). The combination of 'OH + 0; (+OJ also causes denaturation; but in contrast with 'OH alone, the protein is then fragmented. Fragmentation appears to occur predominantly at the a-carbon rather than at peptide bonds. The BSA fragments induced by * OH + 0; (+Oz) form conglomerates which are loosely held together by noncovalent forces (hydrophobic and ionic bonds). In contrast with these results, but inagreement with our studies of BSA primary structure and proteo0 20 40 60 80 100 120 lytic susceptibility (1-3), exposure to 0; 0, had no measnmol Oxygen Rodicols/nmol BSA urable effects on thesecondary or tertiary structureof BSA. The overall pattern of covalent aggregation by 'OH and FIG. 7. Carbonyl content of BSA following exposure to oxfragmentation by 'OH + 0; (+Oz) was also observed for 16 ygen radicals.BSA was exposed to 'OH (A)or to 'OH + 0;+ 0, (0)at the radical/BSA molar ratios indicated (all 'OH or 50% 'OH widely different proteins (1).Modifications to primary struc+ 50% 0;).Carbonylgroups were measured as describedunder ture, such as tryptophan loss and bityrosine production, were "Experimental Procedures." Values are means of three independent also observed in all 16 proteins (1, 2). We therefore propose determinationsfor which standard errorswere less than 10%. that BSA is a good model for the effects of oxygen radicals on protein secondary and tertiary structure. Careful comparisons of the results of Fig. 2 (denaturation/ ments of carbonyl content revealed a large and linear increase with 'OH 0; (+Oz) treatment (Fig. 7). Exposure to 'OH hydrophobicity) and Fig. 5 (aggregation/fragmentation) realone, however, induced a much smaller carbonyl increase veal that protein denaturation precedes aggregation or fragwhich peaked at an 'OH/BSA ratio of approximately 50. It mentation. The same conclusion was reached by comparisons is also possible that the carbonyl groups produced by treat- with densitometric scans of native gels (from Fig. 1)and Tris/ ment with 'OH alone resulted from limited side chain modi- glycine gels (not shown). Evidence of denaturation and unfolding was obtained at radical/BSA ratios as low as 1.0 (the fication or cleavage. Bityrosine and BSA Aggregation-As shown in the previous lowest tested). Half-maximal denaturation was observed at a paper (2), bityrosine formation increases with 'OH exposure. radical/BSA ratio of approximately 9.6, whereas half-maximal aggregation or fragmentation occurred at a ratio of approxiBityrosine can form within a protein (intramolecularbonding) mately 19.4. It appears likely that denaturation andunfolding or between two proteins (intermolecular bonding). The formay potentiate subsequent aggregation or fragmentation promation of intermolecular bityrosine bonds could contribute cesses. significantly to 'OH-induced aggregation of BSA. IntermoIn related studies,3 we have found that protein aggregation lecular bityrosine formation should be affected by BSA con- and fragmentation can also be induced by horseradish percentration. whereas intramolecular bityrosine should not. Bi- oxidase, tyrosinase, xanthine oxidase, and HZOz ATP-Fez+. tyrosine content increased with BSA concentration to a max- Such systems are widely used to model free radical reactions imum at approximately 0.33 mg/ml and declined thereafter which are thought to occur in uiuo. Our present studies of (Fig. 8). These data indicate that intermolecular bityrosine K. J. A. Davies and M. E. Delsignore, unpublished observations. formation was at least one of the covalent forces which




Protein Damage and Degradation by Oxygen Radicals “clean” radical systems (1-3) will be of great value in interpreting the more complex results obtained with enzymatic or transition metal-catalyzed reactions. The process of aggregation by ‘OH appears to involve intermolecular bityrosine formation. It is highly unlikely, however, that bityrosine is the only (non-disulfide) covalent modification. Essentiallyany amino acid radical formed within a peptide chain could cross-link with an amino acid radical in another Protein (4-10). Alternatively, Some radicals may cross-link with amino acid molecules (4-10). The selectivities and rate constants for such reactions are not well understood. For the present, bityrosine provides a sensitive and simple index of cross-linking. Protein fragmentation was explained by Garrison et al. (4, 36) as a result of 0, addition to a-carbon radicals induced by ‘OH. The resulting peroxyl radical would furtherreactto produce a peroxide whose decomposition could cause chain scission to produce a carbonyl and an amide. The peptide bond would remain intact during such reactions. Our finding of linear carbonyl production, in the absence of increased fluorescamine (amino group) reactivity, tends to support the model of Garrison et al. (4, 36). It also appears possible, however, that 0; may add to ue~ectron-poor77 amino acid radicals (formed during hydrogen abstraction by .OH) to form peroxides directly. A modified version of the Scheme of Garrison et al. was recently proposed by Schuessler and Schilling (8).In Schuessler and Schilling’s model, BSA is cleaved (by ’OH + 0; + 0,) by oxidative destruction of proline residues. Wolff et al. (25) have further explored the question of proline attackand have suggested that peptide bond OCcurs. Although we have no firm answer for theseapparently discrepant conclusions, it is possible that buffers may have influenced product profiles (e.g. see Table I1 in this paper or Table I1 in Ref. 2). In the first Paper of this Series (11, several proteins were shown to exhibit increased proteolytic susceptibility following modification by oxygen radicals. This finding is in agreement with reports from several laboratories that oxygen radicals (and/orother active oxygen species)cause protein desa&tion in red blood cells (11-17), E. coli (12, 13, 15, 18-23), muscle cells (12, 13, 15), and mitochondria (12, 13, 15, 2426). In the fourthandlastpaper of this series (3), we have measured the kinetics ofBSA denaturation by cell-free proteolflic systems and proteases. Combined with an understanding of oxidative modification(s) to primary, secondary, andtertiarystructure,thesestudies provide new insights into potential mechanisms for oxidant-induced protein degradation. REFERENCES 1. Davies, K. J. A. (1987) J. Biol. Chem. 262,9895-9901 2. Davies, K. J. A., Delsignore, M.E. and Lin, s. w.(1987) J . Bid. Chem. 262,9902-9907 3. Davies, K. J. A., Lin, S. W., and Pacifici, R.E. (1987) J. Biol. Chem. 262,9914-9920


4. Garrison, W. M., Jayko, M. E., and Bennett, W. (1962) Radiat. Res. 16,483-502 5. Yamamoto, o. (1977) in Protein Crosslinking.. and Molecular Aspects (Friedman, M., ed) pp. 509-556, Plenum Press, New York 6. Adams, G. E., Willson, R. L., Bisby, R. H., and Cundall, R.B. (1971) Znt. J. Radiat. Biol. Relat. Stud.Phys. Chem. Med. 2 0 , 405-415 7. Prutz, W.A., Butler, J., and Land, E. J. (1983) Znt. J. Radiat. Biol. Relat. Stud. Phys. Chem. Med. 44,183-196 8. Schuessler, H., and Schilling, K. (1984) Znt. J. Rad& Biol. Re&. Stud.Phys. Chem. Med. 45,267-281 9. %.-Nag, I., and Floyd, R. A. (19W Biochim. Bkphys. Acta 7 9 0 , 23&250 10. Gajewski, E., Dizdaroglu, M., Krutzsch, H. C., and Simic, M. G . (1984) Znt. J. Radiat. Biol. Relat. Stud. Phys. Chem. Med. 4 6 , 47-55 11. Davies, K. J. A. (1985) in Cellular and Molecular Aspects of Aging: The RedCell os a Model (Eaton, J. W., Konzen, D. K., and White, J. G., eds) pp. 15-27, Academic Press, New York 12. Davies, K. J. A. (1986) in Superoxide and Superoxide Dismutase in Chemistry, Biology, and Medicine (Rotilio, G., ed) pp. 443450, Elsevier/North-Holland Biomedical Press, Amsterdam 13. Davies, K. J. A. (1986) J. Free Radicah Bid. Med. 2,155-173 14. Fagan, J. M., Waxman, L., and Goldberg, A.L. (1986) J. Biol. Chem. 261,5705-5713 15. Davies, K. J. A. (1987) in Cellular Antioxidant Deferwe Mechunisms (Chow, C. K., ed) CRC Press, Inc., Boca Raton, FL, in press 16. Davies, K. J. A., and Goldberg, A.L. (1987) J. Biol. Chem. 2 6 2 , 8220-8226 17. Davies, K. J. A., and Goldberg, A. L. (1987) J. Biol. Chem. 2 6 2 , 8227-8234 18. Levine, R. L., Oliver, C. N., Fulks, R. M., and Stadtman, E. R. (1981) Proc. Natl. Acad. Sci. U. S. A. 7 8 , 2120-2124 19. Fucci, L., Oliver, C. N., Coon, M. J., and Stadtman, E. R. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 1521-1525 20. Levine, R.L. (1983) J. Biol. Chem. 2 5 8 , 11823-11827 21. Levinel R.L. (1983) J. Chem. 2589 11828-11833 22. Rivett, A. J. (1985) J. Biol. Chem. 260,300-305 23. Stadtman, E. R. (1986) Trends Biochem. Sci. 11,11-12 24. Dean, R. T., and p o b k , J. K. (1985) B&hern. Bkphys. Res, Commun. 126,1082-1089 25. Wolff, S. P., Gamer, A., and Dean, R. T. (1986) Trends Biochem. Sci. 11, 27-31 26. RapPoport, s. M.,Schewe,T., Wiesner, R ~ Halang, , K. w.9 Ludwig, P., Janicke-Hohne, M., Tannert, C., and Klatt, D. (1979) Eur. J. Biochem. 9 6 , 545-561 27. Fricke, H.,and Hart, E. J. (1966) in Radiation Dosimetry (Attix, F. H.,and Roesch, W. C., eds) p. 167, Academic Press, New York 28. Bradford, M. M. (1976) Anal. Biochem. 72,248-254 29. Lowry, 0.H.,Rosebrough, N. J., Fan, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193,265-275 30. hisfield, R. A., Lewis, U. J., and Williams, D. E. (1962) Nature 195,281-283 31. Laemmli, U. K. (1970) Nature 227,680-685 32. Anderson, B. L., Berry, R.W., and Telser, A. (1983) Anal. Biochem. 132,365-375 33. Rice, R. H., and Means, G . E. (1971) J. Biol. Chem. 2 4 6 , 831832 34. Bohlen, P., Stein, S., Dairman, W., and Udenfriend, s. (1973) Arch. Biochem. Biophys. 155,213-220 35. Lappin, J., and Clark, L. (1951) Anal. C b m . 23, 541-542 36. Garrison, M. W. (1968) Curr. Radiat. Top. Res. 4 , 43-94

Suggest Documents