Damage to human .alpha.-1-proteinase inhibitor by ... - ACS Publications

Dec 20, 1991 - Cigarette Tar Extracts and the Formation of Methionine. Sulfoxide. Mark D. Evanst 1 and William A. Pryor* t'J>§. Departments of Bioche...
0 downloads 0 Views 1MB Size
Chem. Res. Toxicol. 1992,5, 654-660

654

Damage to Human a-l-Proteinase Inhibitor by Aqueous Cigarette Tar Extracts and the Formation of Methionine Sulfoxide M a r k D. EvanstJ and William A. P r y o r * ~ + J ~ ~ Departments of Biochemistry and Chemistry and the Biodynamics Institute, Louisiana State University, Baton Rouge, Louisiana 70803 Received December 20. 1991

The effects of aqueous extracts of cigarette tar (ACT) on human a-l-proteinase inhibitor

(alPI)are examined by determining alterations in the elastase inhibitory capacity (EIC), amino acid residue content, and electrophoretic behavior of the protein. Hydrogen peroxide generated in ACT by autoxidative processes accounts for the major portion of the loss of EIC. This is indicated by several lines of evidence, including the fact that anaerobic incubations of ACT with alPI cause negligible loss of EIC. The hydrogen peroxide content of the ACT was estimated by measuring the ability of the extracts to oxidize methionine to methionine sulfoxide; hydrogen peroxide concentrations that model those in ACT cause a similar loss in EIC. Exposure of alPI to ACT leads to methionine sulfoxide as the only detectable amino acid residue modification and explains the loss of EIC. This is the first report to directly demonstrate methionine sulfoxide formation in alPIexposed to cigarette smoke components in vitro. Similar amounts of methionine sulfoxide are found in alPI exposed to hydrogen peroxide at the concentrations predicted to be formed in ACT. Nondenaturing PAGE reveals that alp1 exposed to ACT, but not hydrogen peroxide, shows changes in electrophoretic behavior; the changes are nonoxidative in nature and are not related to the loss of EIC. Studies on the effect of chelators on ACT-mediated damage to alPI indicate some role for metal ions; however, none of the chelators completely protect alPI. We suggest that water-soluble components of the cigarette smoke particulate phase (“tar”), in addition to gas-phase components, may damage a1PI and contribute to local antiproteinase deficiencies in smokers’ lungs. This damage probably is caused by oxidative modification of at least one methionyl residue necessary for proteinase inhibitory function.

Introduction Cigarette smoking is the major risk factor associated with the development of pulmonary emphysema (I). The proteinase-antiproteinase theory for the pathogenesis of emphysema proposes that emphysema develops when the ratio of proteinase to antiproteinase activity in lung epithelial lining fluid is altered in favor of proteinases,leading to the eventual destruction of alveolar integrity (2-4). a-l-Proteinase inhibitor (a1PI)’ is the major serum antiproteinase in humans (5)and constitutes the major antielastase activity of the human lower respiratory tract (6). Localized depression of a1PI activity,resulting in enhanced elastolytic activity, may be a significant factor involved in the development of emphysema ( 4 , 7 ) . Numerous studies have assessed the validity of the proteinase-antiproteinase theory, and recent studies support the theory (2-4, 8, 9). The elastase inhibitory capacity (EIC) of alp1 is substantially depressed by oxidation of an active-site me-

* Address correspondence to this author a t the Biodynamics Institute,

711 Choppin Hall, Louisiana State University, Baton Rouge, Louisiana 70803-1804[phone: (504b388-20631. + Department of Biochemistry. * Biodynamics Institute. Department of Chemistry. Abbreviations: a,PI, a-l-proteinaseinhibitor; ACT, aqueouscigarette tar extract; DDC, diethyldithiocarbamate; DTPA, diethylenetriaminepentaacetic acid; EDTA, ethylenediaminetetraaceticacid; EGTA, [ethylenebis(oxyethylenenitrolo)l tetraacetic acid; EIC, elastase inhibitory capacity; HIMDA, N-(2-hydroxyethyl)iminodiaceticacid; MetS=O, methionine sulfoxide; PPE, porcine pancreatic elastase; SANA, N-succinylL-alanyl-L-alanyl-L-alanine-p-nitroanilide; SD, standard deviation; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis.

thionyl residue (Met35s)to the sulfoxide (10-12). Evidence indicates that methionine sulfoxide (MetS=O) is present in a1PI recovered from smokers’ lungs (7). Neutrophils, macrophages, and cigarette smoke are all probable sources of oxidants for MetS=O formation in alp1 in vivo (1316).

Loss of the EIC of a1PI in vitro results after exposure t o gas-phase cigarette smoke either directly or as an aqueous extract (11, 17-19). In addition, aqueous and dimethyl sulfoxideextracts of cigarette smoke condensate also impair the EIC of a1PI in vitro (20-23). We here report studies of the effects of aqueous cigarette tar extract solutions (ACT) on the EIC of human a1PI. This in vitro study complements a previous study from this laboratory on the interaction of ACT with alp1 (24). Data in this report indicate that loss of EIC is probably associated with the appearance of MetS=O in a1PI.

Experimental Procedures Materials. Human q P I (assayed by SDS-PAGE as >95% pure by the supplier) and porcine pancreatic elastase (PPE; EC 3.4.21.36) were purchased from Calbiochem Corp. (La Jolla, CA). Chelex-100 chelating resin was purchased from Bio-Rad Laboratories (Richmond, CA). Research cigarettes (1R1) were purchased from the University of Kentucky, Tobacco and Health Research Institute, stored at -10 “C, and conditioned over saturated aqueous ammonium nitrate a t room temperature for at least 24 h prior to use.* All other chemicals were of reagent A. Vaught, University of Kentucky Tobacco & Health Research Institute, personal communication, 1982.

1992 American Chemical Society

Cigarette Tar Extracts a n d Oxidation of alp1 grade, purchased from Sigma Chemical Co. (St. Louis, MO), and were used without further purification. Caution: The following chemicals used in this study are hazardous and should be handledcarefully: Cyanogen bromide is highly toricand volatile and should always be handled in a fume hood with hand protection; acrylamide is a neurotoxin which should always be handled with great care using gloves. Waste acrylamide stock solutions can be disposed of after polymerization of the acrylamide using ammonium persulfate. Determination of EIC. The activity of P P E was measured a t pH 8.0 and 25 "C according to the method of Bieth et al. (25). The reaction mixture in a final volume of 1mL contained 0.2 M Tris-HC1,pH 8.0,0.14 U of PPE, 2.5 pg alPI, 1mM N-succinyl(L-Ala)S-p-nitroanilide(SANA), and 0.8% v/v N-methylpyrrolidinone. [One unit (1 U) of PPE is defined as the amount of enzyme that will hydrolyze 1.0 pmol of N-acetyl-(Alah-methyl ester per minute at 25 OC, pH 8.5.1 The PPE was preincubated with native or treated alPI for 5 min at 25 "C and the reaction started by the addition of SANA. The increase in absorbance at 410 nm was measured against a reference consisting of buffer and SANA only. The initial linear rate of absorbance change due to uninhibited PPE was taken as 100% PPE activity; the decrease in this rate in the presence of 2.5 pg of native a1PI was taken as 100% EIC. Under the conditionsdescribedabove,native alPI inhibited PPE by 81 f 6% (mean standard deviation; n = 44). Preparation of Aqueous Cigarette Tar Extract; Incubation Protocols. The aqueous cigarette tar solutions were prepared as noted previously (24). Briefly, a Cambridge filter (which retains 99.9% of all particles of >O.l-pm diameter) was used to collect cigarette smoke particulate matter from two 1R1 cigarettes, smoked down to a 23-mm butt length, using the puff protocol and smoking apparatus previously described (17). The Cambridge filter was extracted for 30 min at 37 "C with 5 mL of Chelex-100-treated 0.1 M phosphate buffer, pH 7.4. The extract was then filtered through a filter paper to remove Cambridge filter fibers. Quantitation of material in an ACT in terms of nonvolatile components gives avalue of 3.6 f 0.2 mg/mL (24);this procedure necessarilyneglects those ACT components that are volatile (24). The 5 mL of ACT therefore contains approximately 18 mg of material. Two 1R1 cigarettes should yield approximately 71 mg of total particulate matter; thus approximately 25% of the total particulate matter was extracted into the buffer as nonvolatile material. This material was diluted by half in order to add a1PI solution, buffer, and other components for the incubations. Therefore, each experiment contains about 1.8 mg/mL watersoluble nonvolatile particulate-phase components. This halfdiluted preparation is our standard ACT solution for alp1 exposure and corresponds to the concentration of water-soluble particulate matter from one 1R1 cigarette. Incubations contained 125 pg/mL alPI (2.4 pM) in ACT. Chelex-100-treated0.1 M phosphate, pH 7.4, was used throughout. Some incubations contained, in addition to ACT and alPI, appropriate volumes of additives as probes of the damaging mechanism. For controls, 0.1 M phosphate buffer, pH 7.4, was used instead of ACT or as a substitute for the additives. Unless otherwise indicated, incubations were conducted a t 37 "C, protected from light, for 24 h in a total volume of 300 pL, prior to the EIC assay. The effect of hydrogen peroxide alone on the EIC of alp1was determined by incubating 1.3 mM hydrogen peroxide with 125 pg/mL alPI under the conditions used for ACT exposure. Hydrogen peroxide was added as three aliquots, 5 , 5 , and 10 pL, of a 20 mM stock at 0,4, and 6 h, to give concentrations of 0.33, 0.66, and 1.3 mM hydrogen peroxide, respectively. The final molar ratio of Hz02 to alPI is 542:l. The final incubation volume was 315 pL. This repetitive addition protocol was adopted to avoid exposing alp1 to 1.3 mM hydrogen peroxide in one step. Studies on the interaction of hydrogen peroxide with ACT components involved exposures of 125 pg/mL alPI to ACT for 24 h under normal or low oxygen tensions in the presence of 0.9

*

Chem. Res. Toricol., Vol. 5, No. 5, 1992 655 mM hydrogen peroxide. Hydrogen peroxide was added as three aliquots, 5, 5, and 10 pL, of an 18.5 mM stock a t 0,4, and 6 h, to give concentrations of 0.23, 0.46, and 0.89 mM hydrogen peroxide, respectively. The final molar ratio of H202:aIPI is 371: 1. The final incubation volume was 415 pL. Also, incubations of alPI with hydrogen peroxide and ACT for 4 h were performed; in these cases, hydrogen peroxide was added at the start of the experiment to give a concentration of 0.4 mM in a 400-pL incubation volume. The final molar ratio of Hz0z:alPI is 167:l. In all cases where hydrogen peroxide solutions were added over time, control incubations received equal volumes of deionized water. Incubations under low oxygen tensions were continuously purged with argon, whereas incubations under normal oxygen tensions were continuouslypurged with air. alPI also was exposed to ACT plus bovine liver catalase for electrophoretic analysis. Catalase was added to the incubation of ACT and alPI at 0,4, and 12 h to maintain catalase activity as described previously (24). After the final addition of catalase, the incubations would contain 15 865 U/mL if all catalase remained active. A unit is defined as in the current Sigma catalog. Data derived from incubations containing additives are reported as percent protection. Percent protection is calculated as shown in eqs 1 and 2 (24). In eq 2, %EICAis the % EIC ?6 protection = 100 - % damage % damage = 100([100- %EIC,]/[100

- %EIC,]J

(1) (2)

calculated for alPI incubated in ACT plus an additive and % EICe is the % EIC calculated for alPI incubated in ACT only. A percent damage of 100% is defined as the effect of ACT alone on a1PI. Thus for the exposure of alp1 to ACT plus additives, 100% damage implies the additive has no effect, less than 100% damage implies the additive protects alPI, and an additive that enhances ACT-mediated damage to alPI gives values greater than 100% (24). Amino Acid Analyses. Analyses of the individual amino acids, Val, Met, and MetS=O, that were exposed to ACT were used to help define the oxidative capability of ACT. Incubations of 16 mM Val and 12 mM Met with ACT, and of 16 mM Val and 8 mM MetS=O with ACT, were conducted in total volumes of 500 pL for 24 h, light protected at 37 "C. Controls consisted of incubations in 0.1 M phosphate, pH 7.4, instead of ACT. After incubation, the samples were lyophilized and then reconstituted and diluted with 0.4 N borate buffer, pH 10.2,for HPLC analysis. The HPLC analyses were performed on a Hewlett-Packard H P 1090 liquid chromatograph equipped with a diode array detector, using a CIS-microbore amino acid analysis column (Hewlett-Packard, Palo Alto, CA). Samples were subjected to an on-line precolumn derivatization procedure with 25 mM mercaptopropionic acid and 25 mM o-phthaldialdehyde (26). Analyses were performed at 35 "C, and the chromophore was detected a t 338 nm. Calibration was performed using amino acid standards from Pierce Chemical Co. (Rockford, IL). The mobile phase was composed of a binary solvent system, 20 mM sodium acetate, pH 7.2 (solvent A) and 80% acetonitrile/20% 0.1 M sodium acetate (solvent B). Solvent A was decreased nonlinearly from 98% to 0 % over 15 min, a t a flow rate of 0.44 mL/min (26). The amino acid analysesof native, ACT-exposed,and hydrogen peroxide-exposed alPI were performed using cyanogen bromide to detect MetS=O formation as described by Shechter et al. (27). The reaction of proteins with cyanogen bromide leads to the destruction of methionine, while oxidized methionine remains intact. Hydrolysisof the cyanogen bromide-treated protein under reducing conditions converts MetS=O back to methionine. Thus, methionine detected in a subsequent amino acid analysis corresponds to MetS=O in the unhydrolyzed protein. alPI was exposed to ACT or hydrogen peroxide as described in the preceding section. The alp1samples were filtered through disposable filter units (low-proteinbinding membrane; molecular weight cutoff of 10 000). The residue was washed twice with

656 Chem. Res. Toxicol., Vol. 5, No. 5, 1992 Table I. Analysis of Met, MetS=O, and Val Exposed to ACT *mol of % change in concentration* MetS=O incubation" Val Met MetS=Oc formed Val + Met + A C T -1.5 0.3 -18.9 f 1.4 +11.1 f 0.2 0.64 f 0.02d Val + MetS=O + +3.1 f 0.1 NDe -0.6 f 0.3

Evans and Pryor Table 11. Effect of ACT and Hydrogen Peroxide Alone and in Combination under Normal and Low Oxygen Tensions on the EIC of alPI exposure io21 time, h' % EICb

*

ACT

See Experimental Procedures. * After 24 h, relative to amino acids incubated in buffer only: (+) indicates increase; (-) indicates decrease; mean f SD, n = 3. 'As a percentage of initial Met concentration. Mean f SD, n = 3. e Not detected. (I

ultrapure water and then resuspended in 70% v/v formic acid for reaction with cyanogen bromide. QPI was incubated with cyanogen bromide in dark vials a t room temperature for approximately 24 h (molar ratio of a1PI to cyanogen bromide of 1:5000). After removal of the formic acid and unreactedcyanogen bromide, the protein was hydrolyzed a t 110 "C, for 24 h, invacuo, with 6 N HCl containing 0.6 mg/mL dithiothreitol. Amino acid analysis was performed as described above. Calibration curves were constructed using dilutions of commercially prepared standard mixtures of amino acids. Glycine, cysteine, tryptophan, and proline were not assessed in these analyses. Electrophoresis. a1PI samples for electrophoretic analysis were prepared by incubating as described earlier. Stacking gel compositions for SDS-PAGE and nondenaturing PAGE were as described by Laemmli (28). The separating gels, a t pH 8.8, for SDS-PAGE and nondenaturing PAGE were composed of 11% and 8% acrylamide, respectively. For SDS-PAGE, samples and molecular weight standards were heated a t 90 "C for 5 min in sample buffer containing 0.05% v/v mercaptoethanol and 2% w/v SDS. For nondenaturing PAGE, samples were not heated and the sample buffer contained neither SDS nor mercaptoethanol. In all cases 0.21 wg of cqPI was loaded per lane; where catalase was used, 1.87 I g was loaded per lane.

loo-\ 80

0

W

60 .

be

40 .

Results Quantitationof Oxidant Production in ACT. The HPLC technique used allows detection of all forms of methionine, i.e., the sulfide, the sulfoxide, and the sulfone. Methionine is oxidized to the sulfoxide, and not to the sulfone, when incubated with ACT under our usual conditions (Table I). The production of MetS=O can be used to estimate the oxidizing capacity of ACT toward methionine. (Production of MetS=O also can be used to estimate hydrogen peroxide production in ACT, as addressed in the Discussion.) After incubation for 24 h under our conditions, 1 mL of ACT oxidizes 1.28 f 0.04 pmol of methionine to MetS=O (Table I). Only 59 % of the decrease in methionine content after exposure to ACT is accounted for by the increase in MetS=O (Table I). Thus, about 41 '3 of the methionine is depleted in ACT by uncharacterized pathways. Two explanations for this incomplete loss of methionine can be eliminated. First, valine, an inert amino acid used as control, undergoes little or no change in concentration (Table I), suggesting that ACT has a negligible effect on a-amino or a-carboxyl groups of amino acids. Second, incubation of MetS=O with ACT for 24 h does not result in the measurable formation of methionine (Table I), suggesting that rereduction of MetS=O does not occur. Effect of ACT and Hydrogen Peroxide Alone and in Combination, under Normal and Low Oxygen Tensions, on the EIC of alPI. Under the same incubation conditions used for the ACT incubations, 0.9 or 1.3 mM hydrogen peroxide gives a 60-70 '3 loss of EIC (Table 11). The time course for the loss of EIC of alp1 exposed to 1.3 mM hydrogen peroxide (Figure 1)shows that EIC

20

'

0

5

10

15

20

25

TIME, H

Figure 1. Time course of the EIC loss of 125gg/mLalp1exposed to 1.3mM hydrogen peroxide at pH 7.4 and 37 "C. The hydrogen peroxide was added as three aliquota (see Experimental Procedures). Error bars represent SD (n= 3)where this exceedssymbol size.

is lost at a rate that is comparable to the rate of loss of EIC observed for a1PI exposed to ACT (24). Under low oxygen tensions, alPI activity is unaffected by ACT (Table 11). Coincubation of ACT and 0.9 mM hydrogen peroxide leads to abolition of EIC after 24 h under normal oxygen tensions, but not under low oxygen tensions (Table 11). Some data were obtained at 4-h incubation to study synergism between ACT and hydrogen peroxide. Under normal oxygen tensions, ACT enhances the damaging effects of 0.4 mM hydrogen peroxide (Table 11). However, under low oxygen tensions ACT only minimally increases the damaging ability of 0.4 mM hydrogen peroxide toward alp1 (Table 11). Effect of Chelators on the alPI-Damaging Ability of ACT. To probe for the involvement of transition metals in the alPI-damaging activity of ACT, we tested the effect of various metal ion chelators. The chelators only exert their effects at relatively high concentrations. For example, at 10 pM neither EDTA nor DTPA protect alp1 (data not shown) but 10 mM EDTA or DTPA protect by 48% and 31 '3, respectively (Table 111). The protective effect of 10 mM EDTA is not enhanced by 50 mM benzoate, a hydroxyl radical scavenger (Table 111). Two chelators that share similar structural features to EDTA

Chem. Res. Toxicol., Vol. 5, No. 5, 1992 657

Cigarette Tar Extracts and Oxidation of alPI

Table 111. Effect of Metal Ion Chelators. and Ferric Ions. on the Activity of Human a1PI Exposed to ACT* additive % protection* additive % protection* 22f4 20 mM HIMDA 42 f 7c EDTA 53f 1 DDC DTPAd 3112 11f6 bathucuproinedisulfonate EDTA + 50 mM benzoate 32i4 32f2 -7 4c 100 pM Fe3+ 5 mM Desferal 25 f 4c 200 p M Fe3+ 500 pM Desferal 614 28i4 200 p M Fe3++ 50 mM benzoate EGTA 14 f 6 0 Each additive alone had no measurable effect on alp1 activity: Unless otherwise indicated, each additive was present at 10 mM. Data are means f SD, n = 3, unless indicated otherwise. Mean f SD, n = 5. d Reference 24. e Mean f SD, n = 6.

and DTPA but are poorer chelators of transition metals, i.e., EGTA and HIMDA, only offer 14% and 22% protection, respectively. Desferal (desferrioxamine mesylate) offers no protection at either 0.5 or 5 mM (Table 111). The copper chelator diethyldithiocarbamate (DDC) protects alp1whereas bathocuproine disulfonate does not (Table 111). The addition of 100 or 200 pM ferric ions to the ACT system gives about 30% protection, and this protection is not increased by the presence of 50 mM benzoate (Table 111). Amino Acid Analysis of alp1 Exposed to ACT or Hydrogen Peroxide. The amino acid composition of native alp1agreed well with previously published reports, and no M e t S 4 could be detected (data not shown). The only detectable change in amino acid composition is the in both ACT and hydrogen formation of peroxide-exposed alpI (data not shown). The hydrogen peroxide-exposed alp1 contains 1.7 f 0.4 M e t S 4 residues (mean f SD; n = 5 ) and the ACT-exposed alp13.1 f 0.9 M e 6 4 residues (mean f SD; n = 7). These values of M e t S 4 content represent a value averaged for all the possible types of damaged alp1 species in the exposures, i.e., alp1 molecules containing varying amounts of Me-. Analysis of alp1 Exposed to Hydrogen Peroxide or ACT by Denaturing and Nondenaturing Polyacrylamide Gel Electrophoresis. On SDS-PAGE gels, alp1 exposed to either ACT or 1.3 mM hydrogen peroxide undergoes no detectable change (data not shown). Two bands are detected on the nondenaturing gels. Although two major isoforms of alp1 are present in serum, they differ in pZ by 0.2 unit and so are unlikely to be resolved on these gels. The relativelyminor less anionic proteins(s) could possibly arise from the approximately 5 % con-ination of the commercially supplied a1pI. However, the minor band could also be due to oligomericalPI. On SDSPAGE gels only one band is detected (data not shown). m e followingdiscussion ofthe nondenaturhg gelspertains to the intense, more anionic band (top of gel, cathode; bottom of gel, anode). The exposure of alp1to ACT results in electrophoretic heterogeneity characterized by band smearing in the anodal direction (Figure 2). Exposure to hydrogen peroxide has no effect on the mobility of alp1 under the nondenaturing conditions (Figure 2). Under argon or in the presence of catalase, exposure to ACT still produces alp1with heterogeneous electrophoretic behavior (Figure 3). Both Figures 2 and 3 show data for incubations of alp1with separate ACT or hydrogen peroxide exposures, indicating that the changes in electrophoretic behavior are reproducible. Also, the intense, more cationic catalase band does not contain species that comigrate with alp1 on these gels.

1

2

3

4

5

6

_. .

- -

7 0 .- ------

Figure 2. Nondenaturing PAGE of alp1 exposed to ACT and alPI exposed to 1.3 mM hydrogen peroxide. Gel lanes are as follows: lanes 1 and 3, a1PI exposed to ACT; lanes 2,4,6, and 8,native alPI; lanes 5 and 7, alp' exposed to 1.3 mM peroxide. 1

2

3

4

5

6

7 ---

8

~

Figure 3. NondenaturingPAGE of alp1 exposed to ACT in the presenceand absence of catalase and alp1exposed to ACT under argon. Gel lanes me 89 follows: lanes 1 and 6, alp1 exposed to ACT plus catalase; lanes 2 and 8, alPI exposed to ACT; lanes 3 and 7, alp1 exposed to ACT under argon; lane 4, native wPI; lane 5, catalase only. Catalase was added according to the procedure outlined in the Experimental Procedures.

Discussion Hydrogen Peroxide Involvement in ACT-Mediated Damage to alPI. Hydrogen peroxide is produced from the reduction of oxygen by the oxidation of ACT components, particularly polyphenols (29-32), and is the major source of oxidizingability in cigarette smoke extracts. Our previous studies have shown that the hydrogen peroxide produced in ACT is 'an important mediator of damage to alp1(24). The involvement of hydrogen peroxide in ACTmediated damage to alp1 is further supported by the inability of ACT to damage alp1 under an argon atmosphere. If hydrogen peroxide in ACT is the primary oxidant of methionine, then MetS=O formation can be used to estimate the hydrogen peroxide content of ACT.

Evans and Pryor

658 Chem. Res. Toricol., Vol. 5, No. 5, 1992

Peroxides oxidize sulfides with a 1:l stoichiometry (33). Using this stoichiometric factor and the yield of MetS=O shown in Table I, we calculate that 1.2-1.3 mM hydrogen peroxide is produced in the ACT in 24 h. This value compares well with that measured in our earlier studies (29). Using the data from our previous quantitation of hydrogen peroxide in ACT (29) and assuming that hydrogen peroxide accumulates in ACT3 the amount of cigarette tar used in our exposures should generate at least 0.84 0.21 mM hydrogen peroxide over 24 h. Also, the rates of loss of EIC of alp1 during exposure to either ACT or 1.3mM hydrogen peroxide are similar (Figure 1and ref 24). Hydrogen peroxide probably accumulates in ACT and damages a1PI slowly. [This agrees with refs 17, 18, and 29, but we suggested the contrary in our previous publication on the interaction of alPI and ACT (241.1 A concentration of hydrogen peroxide comparable to that formed in ACT over 24 h depresses alp1 activity to a similar extent to that observed for ACT in 24 h. The data in Table I1 and Figure 1 suggest that hydrogen peroxide is directly responsible for most of the damage to alp1 caused by ACT. The remainder of the damage due to ACT could be caused by synergisticinteractions between hydrogen peroxide and other ACT components (see discussion below). Some synergism is observed between the damaging actions of hydrogen peroxide and ACT. The synergism between 0.9 mM hydrogen peroxide and ACT (Table 11) probably is due to the additive damaging effect of the exogenous (Le,,added) and endogenous (Le,, from autoxidation of ACT components) hydrogen peroxide. Under argon, the loss of EIC due to the ACT/O.S mM hydrogen peroxide mixture is attributable to the exogenous hydrogen peroxide alone. However, the damaging effect of the ACT/ 0.4 mM hydrogen peroxide mixture cannot be explained by an additive effect of exogenous and endogenous hydrogen peroxide, since 0.4mM hydrogen peroxide alone has no effect on the EIC. The incubation of alp1 with ACT/0.4 mM hydrogen peroxide under argon results in a 13% loss of EIC (Table 11). Note that the loss of EIC is 15% greater when alPI is incubated with ACT/0.4 mM hydrogen peroxide under air as compared to alp1 incubated under air with ACT alone (Table 11). Thus hydrogen peroxide acts synergistically with an unidentified component@) in ACT to cause enhanced damage to a1PI. Cohen and James have speculated that cigarette smoke might enhance the damaging ability of hydrogen peroxide produced from inflammatory cells in smokers' lungs (23). The Role of Transition Metals in the aIPI-Damaging Activity of ACT. Although none of the chelators we tested protect a1PI completely, metals in ACT could participate in damage to alp1 in three ways. First, more reactive oxidants (such as hydroxyl radicals or hypervalent transition metal species)could be generated from reactions of metal complexes with hydrogen peroxide. Second, a1PI could be damaged in a site-specific manner by the interaction of hydrogen peroxide with transition metalalp1complexes (34). Third, transition metals are catalysts for polyphenol autoxidation (35). Some of our earlier studies have shown that EDTA increases the yield of hydroxyl radical spin adduct signals

*

3 If hydrogen peroxide is continuously formed and destroyed, one would expect 22.8 m M hydrogen peroxide to be formedin our ACTpreparations over 24 h (29). This value is considerably larger than our estimate of hydrogen peroxide using the oxidation of methionine.

in ACT (30); this increase in hydroxyl radicals was proposed to occur via metal-catalyzed Haber-Weiss reactions (30, 36): 0,'

-

-

+ M(n+l)+

Mn++ H,O,

-

Mn++ 0,

(3)

+

(4)

M("+')+ HO-

+ HO'

The protection by EDTA (Table 111) could arise from the following: (i) inhibition of polyphenol autoxidation and hydrogen peroxide production; (ii) removal of proteinbound metals from sensitive sites on a1PI; (iii) decomposition of hydrogen peroxide to more reactive, and therefore less selective, species by EDTA-metal ion complexes. With regard to (iii),we have previously shown that hydroxyl radical scavengers do not protect a1PI from damage by ACT (24). Unlike EDTA, DTPA abolishes hydroxyl radical spin adduct signals in ACT (301, presumably by blocking eq 3. Thus DTPA must protect alp1 by mechanisms like (i) and (ii) noted above for EDTA. The relatively poor protective ability of HIMDA and EGTA (despite structural similarity to EDTA and DTPA) implies that EDTA and DTPA do not protect alp1 by directly scavenging a1PIdamaging species. Desferal inhibits the production of reactive oxygen species in cigarette smoke solutions (32) but, unlike DTPA, does not protect a1PI against damage by ACT. It is possible that Desferal reacts with ACTderived superoxide to produce a protein-damaging nitroxyl radical (37). Ferric ions increase the production of reactive oxygen species in cigarette smoke solutions (30, 32). The protective effect of ferric ions in ACT probably arises by decomposition of hydrogen peroxide in ACT via ferric ionACT component complexes. Benzoate does not improve the protective effect of either EDTA or ferric ions. Thus, increases in hydroxyl radical production in ACT do not compromise the protective effects of ferric ions or EDTA. The fact that bathocuproinedisulfonate does not protect a1PI implies that copper ions play a negligible role in ACTmediated damage to alPI. The known susceptibility of DDC to oxidation suggests that DDC protects a1PI from damage by ACT due to sacrificial oxidation rather than copper complexation. The involvement of metals in cigarette smoke-mediated damage to alp1 in vivo is uncertain, although cigarette smoke does possess appreciable siderophoric activity and the iron burden to the lower respiratory tract of smokers is increased compared to nonsmokers (38-41). Molecular Changes in alp1 Exposed to ACT or Hydrogen Peroxide. A number of studies strongly support the proposal that Met358in alp1 is oxidized after exposure to aqueous solutions of cigarette smoke in vitro (11, 18,21, 42). The data in this report directly demonstrate, for the first time we believe, the presence of MetS=O in alPI exposed to water-soluble cigarette smoke components in vitro; furthermore, MetS=O is the only detectable amino acid residue modification. Exposure of alPI to low molecular weight oxidants generally results in the oxidation of the two methionyl residues most accessible to the solvent, i.e., Met351and Met358(10,14). The results reported here indicate an average of 2 or 3 MetS=O/alPI are formed when alp1 is exposed to hydrogen peroxide or ACT. Therefore, hydrogen peroxide at concentrations estimated to be formed in ACT is responsible for most of the MetS=O detected in ACT-exposed a1PI. Studies on

Cigarette Tar Extracts and Oxidation of alPI

alPI recovered from smokers’ lungs show 4 MetS=O/alPI which would likely occur due to the action of both cellularly generated and smoke-borne oxidants (7). The information presented in our earlier study (24) and here suggests that oxidation of methionyl residues in alp1 by hydrogen peroxide produced in ACT occurs by direct reaction: MetS

+ H,O,

-

MetS+-OH

-

+ HO-

MetS=O + H,O ( 5 ) Other (reversible)amino acid residue modifications may be present in ACT-exposed alp1 but be undetectable by our amino acid analysis. However, evidence from our previous study (24)suggests that any modifications of lysyl, phenylalanyl, tryptophanyl, arginyl, histidyl, or prolyl residues in ACT-exposed alPI are of no consequence with respect to the loss of EIC. The results of SDS-PAGE indicate that no major aggregation or fragmentation of a1PI is detectable after exposure to either ACT or hydrogen peroxide. a1PI exposed to ACT shows electrophoretic heterogeneity.Some earlier reports on a1PI exposed to cigarette smoke in vitro noted that the protein displayed increased anionic mobility on nondenaturing gels (22,43). The formation of MetS=O alone is not responsible for the modified electrophoretic behavior of ACT-exposed q P I , since hydrogen peroxideexposed alPI does not show any change on nondenaturing gels. Furthermore, exposure of a1PI to ACT in the presence of catalase or under argon still results in altered electrophoretic behavior, despite the preservation of EIC (Table I1 and ref 24). Therefore, the electrophoretic heterogeneity caused by modification(s) in a1PI is unrelated to the loss of EIC. Conclusions. The results presented here confirm ideas from severalgroups including our own that oxidative mechanisms play a primary role in the direct damage to a1PI by cigarette smoke components. In addition, we have confirmed, by direct measurement in vitro, the previously held assumption that methionine sulfoxide is present in alPI that was exposed to cigarette smoke components. Others have suggested that cigarette smoke components are ineffective at directly damaging a1PI (44). However, we have shown that this was the result of methodological differences and was reflective of differences in the mechanisms by which gas-phase and whole smoke produce a1PI damage (45). The pattern of emphysematous lesions in the lungs of cigarette smokers correlates well with the sites of deposition of cigarette smoke particles (46). Water-soluble cigarette smoke particulate-phase components (i.e., ACT components) produce an oxidative burden on the lung. In addition, cigarette smoke components provide an impetus for the recruitment and activation of increased numbers of phagocytic cells in smokers’ lungs (47-50). Thus, the combined effects of phagocyte-derived and cigarette smoke-derived oxidants could contribute to damage to alp1 in microdomains in smokers’ lungs (51). Acknowledgment. We thank Dr. Juan J. Moreno for helpful discussions. This work was supported by NIH Grant HL-25820.

References (1) US. Public Health Service (1984) The Health Consequences of Smoking: Chronic Obstructive Lung Disease, A Report of the Surgeon General, U S . Government Printing Office, Washington, DC.

Chem. Res. Toxicol., Vol. 5, No. 5, 1992 669 (2) Stockley, R. A. (1987) Alphal-antitrypsin and the pathogenesis of emphysema. Lung 165,61-77. (3) Niewoehner, D. E. (1988) Cigarette smoking, lung inflammation, and the development of emphysema. J. Lab. Clin.Med. 111,15-27. (4) Janoff, A. (1985) Elastases and emphysema. Current assessment of the protease-antiprotease hypothesis. Am. Rev. Respir. Dis. 132, 417-433. (5) Travis, J., and Salvesen, G. S. (1983) Human plasma proteinase inhibitors. Annu. Rev.Biochem. 52,655-709. (6) Gadek, J. E., Fells, G. A., Zimmerman, R. L., Rennard, S. I., and Crystal, R. G. (1981)Antielastases of the human alveolar structures. Implications for the protease-antiprotease theory of emphysema. J. Clin. Invest. 68, 889-898. (7) Carp, H., Miller, F., Hoidal, J. R., and Janoff, A. (1982) Potential mechanism of emphysema: al-proteinase inhibitor recovered from lungs of cigarette smokers contains oxidized methionine and has decreased elastase inhibitory capacity. Proc. Natl. Acad. Sci. U.S.A. 79, 2041-2045. (8) Fujita, J., Nelson, N. L., Daughton, D. M., Dobry, C. A., Spurzem, J. R., Irino, S., and Rennard, S. I. (1990) Evaluation of elastase and antielastase balance in patients with chronic bronchitis and pulmonary emphysema. Am. Rev. Respir. Dis. 142, 57-62. (9) Ogushi, F., Hubbard, R. C., Vogelmeier, C., Fells, G. A., and Crystal, R. G. (1991) Risk factors for emphysema: cigarette smoking is associated with a reduction in the association rate constant of lung al-antitrypsin for neutrophil elastase. J. Clin. Invest. 87, 10601065. (10) Johnson, D., and Travis, J. (1979) The oxidative inactivation of human a-1-proteinase inhibitor: further evidence for methionine at the reactive center. J. Biol. Chem. 254,4022-4026. (11) Janoff, A,, George-Nascimento, C., and Rosenberg, S. (1986) A genetically engineered, mutant human a-1-proteinase inhibitor is more resistant than the normal inhibitor t~ oxidative inactivation by chemicals, enzymes, cells, and cigarette smoke. Am. Reu. Respir. D ~ s 133, . 353-356. (12) Beatty, K., Bieth, J., and Travis, J. (1980) Kinetics of association of serine proteinases with native andoxidizeda-1-proteinaseinhibitor and a-1-antichymotrypsin. J. Biol. Chem. 255, 3931-3934. (13) Carp, H., and Janoff, A. (1979) In vitro suppression of serum elastaseinhibitory capacity by reactive oxygen species generated by phagocytosing polymorphonuclear leukocytes. J. Clin. Inuest. 63, 793797. (14) Matheson, N. R., Wong, P. S., Schuyler, M., and Travis, J. (1981) Interaction of human a-1-proteinase inhibitor with neutrophil myeloperoxidase. Biochemistry 20, 331-336. (15) Hubbard, R. C., Fumitaka, O., Fells, G. A., Cantin, A. M., Jallat, S., Courtney, M., and Crystal, R. G. (1987) Oxidants spontaneously released by alveolar macrophages of cigarette smokers can inactivate the active site of a1-antitrypsin, rendering it ineffectiveasan inhibitor of neutrophil elastase. J. Clin.Invest. 80, 1289-1295. (16) Shock, A., and Baum, H. (1988) Inactivation of alpha-1-proteinase inhibitor in serum by stimulated human polymorphonuclear leucocyte~. Evidence for a myeloperoxidase-dependent mechanism. Cell Biochem. Funct. 6,13-23. (17) Pryor, W. A., Dooley, M. M., and Church, D. F. (1984) Inactivation of human alpha-1-proteinase inhibitor by gas-phase cigarette smoke. Biochem. Biophys. Res. Commun. 122,676-681. (18) Pryor, W. A,, Dooley, M. M., and Church, D. F. (1986) The inactivation of alpha-1-proteinase inhibitory by gas-phase cigarette smoke: protection by antioxidanta and reducing species. Chem.Biol. Interact. 57, 271-283. (19) Janoff, A,, and Dearing, R. (1982) al-proteinase inhibitor is more sensitive to inactivation by cigarette smoke than is leukocyte elastase. Am. Rev. Respir. Dis. 126, 691-694. (20) Janoff, A,, and Carp, H. (1977) Possible mechanisms of emphysema in smokers; cigarette smoke condensate suppresses protease inhibition in uitro. Am. Rev. Respir. Dis. 116, 65-71. (21) Janoff, A., and Carp, H. (1978) Possible mechanisms of emphysema in smokers; in uitro suppression of serum elastaseinhibitory capacity by fresh cigarette smoke and ita prevention by antioxidanta. Am. Rev.Respir. Dis. 118, 617-621. (22) Ohlsson, K., Fryksmark, U., and Tegner, H. (1980) The effect of cigarette smoke condensate on al-antitrypsin, antileukoprotease and granulocyte elastase. Eur. J. Clin. Inuest. 10, 373-379. (23) Cohen, A. B., and James, H. L. (1982) Reduction of the elastase inhibitory capacity of alpha-antitrypsin by peroxides in cigarette smoke. Am. Reu. Respir. Dis. 126, 25-30. (24) Evans, M. D., Church, D. F., and Pryor, W. A. (1991) Aqueous cigarette tar extracts damage human alpha-1 proteinase inhibitor. Chem.-Biol. Interact. 79, 151-164. (25) Bieth, J., Speiss, B., and Wermuth, C. G. (1974) The synthesis and analytical use of a highly sensitive and convenient substrate of elastase. Biochem. Med. 11. 350-357.

660 Chem. Res. Toxicol., Vol. 5, No. 5, 1992 (26)Schuster, R., and Apfel, A. (1986)A new technique for the analysis of primary and secondary amino acids. Hewlett-Packard, HPLC Applic. Note 12,5954-6257 (Abstract). (27)Shechter, Y., Burstein, Y., and Patchornik, A. (1975)Selective oxidation of methionine residues in proteins. Biochemistry 14,44974503. (28)Laemmli, U. K. (1970) Cleavage of structural proteins during assembly of the head of the bacteriophage T4. Nature 227,680-685. (29) Nakayama, T., Church, D. F., and Pryor, W. A. (1989)Quantitative analysis of the hydrogen peroxide formed in aqueous cigarette tar extracts. Free Radical Biol. Med. 7,9-15. (30) Cosgrove, J. P., Borish, E. T., Church, D. F., and Pryor, W. A. (1985) The metal-mediated formation of hydroxyl radical by aqueous extracts of cigarette tar. Biochem. Biophys. Res. Commun. 132, 390-396. (31)Nakayama, T., Kodama, M., and Nagata, C. (1984)Generation of hydrogen peroxide and superoxide anion radical from cigarette smoke. Gann Monogr. Cancer Res. 75,95-98. (32) Leanderson, P., and Tagesson, C. (1990)Cigarette smoke-induced DNA-damage: role of hydroquinone and catechol in the formation of the oxidative DNA-adduct.. 8-hvdroxvdeoxveuanosine. Chem .Biol. Interact. 75,71-81. (33)Barnard, D.. Bateman, L., and Cuneen, J. I. (1961)Oxidation of organic sulfides. In Organic Sulfur Compounds (Kharasch, N., Ed.) Vol. 1, pp 229-247,Pergamon Press, New York, NY. (34)Chevion, M. (1988)A site-specific mechanism for free radical induced biological damage: the essential role of redox-activetransition metals. Free Radical Biol. Med. 5, 27-37. (35) Miller, D. M., Buettner, G. R., and Aust, S. D. (1990)Transition metals as catalysts of ‘autoxidation” reactions. Free Radical Biol. Med. 8,95-108. (36)Borish, E. T., Pryor, W. A., Venugopal, S., andDeutsch, W. A. (1987) DNA synthesis is blocked by cigarette tar-induced DNA singlestrand breaks. Carcinogenesis 8, 1517-1520. (37)Davies, M. J., Donkor, R., Dunster, C. A., Gee, C. A., Jonas, S., and Willson, R. L. (1987)Desferrioxamine (Desferal) and superoxide free radicals; formation of an enzyme damaging nitroxide. Biochem. J. 246,725-729. (38) Qian, M., and Eaton, J. W. (1989)Tobacco-borne siderophoric activity. Arch. Biochem. Biophys. 275,280-288. (39)Thompson, A. B., Bohling, T., Heires, A., Linder, J., and Rennard, S. I. (1991)Lower respiratory tract iron burden is increased in association with cigarette smoking. J . Lab. Clin. Med. 117,493499. (40) Church, D. F.,Burkey, T. J., and Pryor, W. A. (1990)Preparation of human lung tissue from cigarette smokers for analysis by electron ~

Evans and Pryor spin resonance spectroscopy. In Methods in Enzymology, Volume 186,Oxygen Radicals inBiological Systems,Part B, Oxygen Radicals and Antioxidants (Packer, L., and Glazer, A. N., Eds.) pp 665-669, Academic Press, New York, NY. (41) Moreno, J. J., Foroozesh, M., Church, D. F., and Pryor, W. A. (1992) Release of iron from ferritin by aqueous extracts of cigarette smoke. Chem. Res. Toxicol. 5, 116-123. (42) Carp, H., Janoff, A., Abrams, W., Weinbaum, G., Drew, R. T., Weissbach, H., and Brot, N. (1983)Human methionine sulfoxidepeptide reductase, an enzyme capable of reactivating oxidized a-lproteinase inhibitor in vitro. Am. Reu. Respir. Dis. 127,301-305. (43) Stockley, R.A., and Afford, S. C. (1983)The interaction of cigarette smoke solution with al-antitrypsin: effect on inhibitory capacity, electrophoretic mobility and immunological measurement, Clin. S C 64, ~ . 223-230. (44) Wyss, A., Virca, G. D., and Schnebli, H. P. (1984)Cigarette smoke components are not very effective in directly inactivating alproteinase inhibitor. Hoppe-Seyler’s 2.Physiol. Chem. 365, 511516. (45)Pryor, W. A,,and Dooley, M. M. (1985)Inactivationof humanalpha1-proteinase inhibitor by cigarette smoke: effect of smoke phase and buffer. Am. Reu. Respir. Dis. 131, 941-943. (46) Dunnill, M. S. (1979)Aetiology of emphysema. Bull. Eur. Physiopathol. Respir. 15, 1015-1029. (47) MacNee, W., Wiggs, B., Belzberg, A. S., and Hogg, J. C. (1989)The effect of cigarette smoking on neutrophil kinetics in human lungs. N . Engl. J.Med. 321,924-928. (48)Pacht, E. R., Kaseki, H., Mohammed, J. R., Cornwell, D. G., and Davis, W. B. (1986)Deficiencv of vitamin E in the alveolar fluid of cigarette smokers; influence on alveolar macrophage cytotoxicity. J.Clin. Inuest. 77,789-796. (49) Greening, A. P., and Lowrie, D. B. (1983)Extracellular release of hydrogen peroxide by human alveolar macrophages: the relationship to cigarette smoking and lower respiratory tract infections. Clin. Sci. 65,661-664. (50)Davis, W. B., Pacht, E. R., Spatafora, M., and Martin, W. J., I1 (1988)Enhanced cytotoxic potential of alveolar macrophages from cigarette smokers. J. Lab. Clin. Med. 111, 293-298. (51) Evans, M. D. (1991)Dissertation, Louisiana State University, Baton Rouge, LA 70803.

Registry No. alPI, 9041-92-3; Met, 63-68-3;MetS=O, 45441-1; H202,7722-84-1; OH’, 3352-57-6; Cu, 7440-50-8; Fe, 743989-6.