Cysteine Modification by Lipid Peroxidation Products Inhibits Protein

A current “scanning and escape” model proposed by Walker and Gilbert describing the function of PDI suggests that the N-terminal active site, whic...
4 downloads 0 Views 318KB Size
Download PDF
1324

Chem. Res. Toxicol. 2005, 18, 1324-1331

Cysteine Modification by Lipid Peroxidation Products Inhibits Protein Disulfide Isomerase David L. Carbone,† Jonathan A. Doorn,‡ Zachary Kiebler,† and Dennis R. Petersen*,† Department of Pharmaceutical Sciences, School of Pharmacy, The University of Colorado Health Sciences Center, Denver, Colorado 80262, and Division of Medicinal and Natural Products Chemistry, College of Pharmacy, The University of Iowa, Iowa City, Iowa 52242 Received March 18, 2005

A proteomic approach was applied to mitochondrial protein isolated from the livers of rats fed a combination high-fat and ethanol diet to identify proteins modified by 4-hydroxynonenal (4-HNE). Using this approach, the endoplasmic reticulum chaperone, protein disulfide isomerase (PDI), which participates in the maturation of newly synthesized proteins through promoting correct disulfide formation, was consistently found to be modified by 4-HNE. Further mass spectral analysis of PDI isolated from the animals revealed modification of an active site Cys residue thought to be involved in client protein binding. To test the hypothesis that 4-HNE inhibits the chaperone, purified bovine PDI was treated with concentrations of 4-HNE ranging from 20 to 200 µM (10-100-fold molar excess aldehyde), resulting in 14-56% inhibition, respectively. Similar treatments with the lipid peroxidation products acrolein (ACR) and 4-oxononenal (4-ONE) resulted in 60 and 100% inhibition, respectively, suggesting inactivation of the chaperone via Cys modification. Thiol sensitivity was confirmed through concentrationdependent inhibition of PDI by the Cys modifier N-ethylmaleimide (NEM). While some degree of sensitivity to these lipid aldehydes is suggested by the data, when compared to inactivation of other proteins by 4-HNE, PDI has demonstrated a relative resistance. It was also observed that physiologic (e.g., 4 mM) concentrations of GSH were capable of removing the 4-HNE adducts, likely serving as a protective mechanism against inactivation by 4-HNE and other lipid peroxidation products. However, because an active site Cys was found to be modified by 4-HNE on PDI in vivo, it is possible that the protective effect of GSH on the chaperone decreases under conditions of sustained oxidative stress, such as during chronic alcohol consumption, as GSH is depleted. The data presented here thus suggest potential impairment of an important molecular chaperone during oxidative stress.

Introduction Pathologic accumulation of reactive oxygen species (ROS), or oxidative stress, is a central component of alcoholic liver disease (ALD),1 diabetes, reperfusion injury (e.g., stroke or myocardial infarction), and many other diseases (1-4). While the ROS are typically very shortlived, they can readily react with the double bonds on proximal polyunsaturated fatty acids (PUFAs), such as linoleic and arachadonic acid, resulting in PUFA peroxidation (5, 6). Spontaneous cleavage of the C-C bond on either side of the peroxide group, followed by several steps of decomposition yields the formation of multiple electrophilic aldehyde species. Many of these compounds, such as the R,β unsaturated aldehydes 4-hydroxynonenal (4-HNE) and 4-oxononenal (4-ONE), are capable of modifying proteins and other cellular nucleophiles through * To whom correspondence should be addressed. † Department of Pharmaceutical Sciences, School of Pharmacy, University of Colorado Health Sciences Center. ‡ Division of Medicinal and Natural Products Chemistry, College of Pharmacy, University of Iowa. 1 Abbreviations: ACN, acetonitrile; ACR, acrolein; ALD, alcoholic liver disease; PDI, protein disulfide isomerase; ER, endoplasmic reticulum; Hsp72, heat shock protein 72 (inducible); 4-HNE, 4-hydroxynonenal; 4-ONE, 4-oxononenal; MDA, malondialdehyde; NEM, Nethylmaleimide; LC-MS/MS, liquid chromatography and tandem mass spectrometry; PUFA, polyunsaturated fatty acid.

Michael addition, typically at Cys residues, leading to interference with protein function (7-9). Because lipid aldehyde species are comparatively longer-lived than the ROS, they often migrate beyond the site at which they are formed, thus propagating ROS toxicity (6). Elevated levels of lipid aldehydes are therefore an important component of ROS-mediated toxicity due to their relative longevity and documented ability to disrupt protein function through the formation of stable aldehydeprotein adducts. ALD is often accompanied by the presence of insoluble protein aggregates, which likely form as a result of oxidative protein damage and decreased protein degradation (10, 11). However, protein misfolding may also contribute to this observation, suggesting impairment of cellular chaperoning mechanisms. In a previous publication, it has been demonstrated that the molecular chaperone heat shock protein 72 (Hsp72) is modified by 4-HNE in the livers of rats which had received a combination high-fat and ethanol diet (12). The same report also demonstrated that modification of this heat shock protein by Cys reactive aldehydes prevented substrate binding, lending credibility to the claim that lipid aldehydes contribute to the progression of ALD through interference with chaperoning mechanisms.

10.1021/tx050078z CCC: $30.25 © 2005 American Chemical Society Published on Web 07/22/2005

Inhibition of Protein Disulfide Isomerase by 4-HNE

The enzyme protein disulfide isomerase (PDI) (E.C. 5.3.4.1.) is an endoplasmic reticulum (ER) protein, which was also observed to be consistently modified by 4-HNE in mitochondrial fractions isolated from the livers of rats following the feeding study described above (12). PDI is classified as a molecular chaperone and plays an important role in the maturation of newly synthesized proteins by correcting improperly formed disulfide bonds (13-15). A current “scanning and escape” model proposed by Walker and Gilbert describing the function of PDI suggests that the N-terminal active site, which is one of two thioredoxin-like domains (e.g., CGHCK), first forms a disulfide bond with the client protein, while the C-terminal active site later induces release of the client by forming its own disulfide bond with the N-terminal site, thus creating an “escape” mechanism (16). In the presence of reducing conditions (e.g., GSH), PDI is regenerated and can repeat the binding and release cycle, while the client protein disulfide bonds are reformed. The cycle is thought to repeat until the client protein reaches the correct conformation. The “scanning and escape” mechanism is supported by the fact that mutants in which either the N-terminal or the C-terminal active site Cys residues have been replaced by Ser are comparatively less active than the wild-type protein (17). It is therefore conceivable that attachment of Cys reactive aldehyde species in the active site of the enzyme may lead to inactivation of the disulfide repair mechanism, resulting in the accumulation of misfolded proteins within the cell. The molecular mechanisms behind ALD are poorly understood; however, an accumulation of insoluble protein aggregates is typically associated with the progression of this disease. Because of their reactivity and relative longevity, the aldehyde species formed during peroxidation of PUFAs may play a significant role in this process. A previous report has indicated the possibility that chaperone inhibition by the lipid aldehydes contributes to the progression of this disease (12). The data described here present further evidence of molecular chaperone inhibition by the Cys reactive aldehydes 4-HNE and 4-ONE. Specifically, in a rat model of early stage ALD, the chaperone PDI was found to be modified by the lipid peroxidation product 4-HNE. Further in vitro experimentation has confirmed inhibition of the isomerase properties of this enzyme following modification by this aldehyde. Evidence for indirect inhibition under conditions of oxidative stress is also provided through the demonstration of PDI inactivation by reduced levels of GSH. Given the susceptibility of PDI to inactivation by lipid aldehydes, it is conceivable that decreased PDI function contributes to protein accumulation during ALD.

Materials and Methods Reagents. Unless stated otherwise, all reagents were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO). The lipid peroxidation products 4-HNE, 4-ONE, and malondialdehyde (MDA) were synthesized in our laboratory according to procedures described previously, and purity and concentration were confirmed by TLC and UV/vis spectrophotometry (7, 18, 19). Animals. All procedures involving animals were approved by the Institutional Animal Care and Use Committee of the University of Colorado and were performed in accordance with published National Institutes of Health guidelines. Male Harlan Sprague-Dawley rats were fed a modified Lieber-DeCarli diet consisting of 45% fat-derived calories, 35% ethanol-derived

Chem. Res. Toxicol., Vol. 18, No. 8, 2005 1325 calories, and 16% protein for a period of 60 days. Each ethanolfed animal was paired with a nutritional control animal that received a diet containing equivalent sucrose-derived calories. Upon completion of the feeding period, animals were anesthetized by intraperitoneal sodium pentobarbital injection and euthanized by exsanguination. Plasma ALT and ammonia determination, and tissue harvesting, sectioning, staining, and subcellular fractionation were performed as described in a previous report (12, 20). Protein was quantified using the BioRad Protein Assay reagent. Subcellular fractions were stored at -80 °C until analysis. Two-Dimensional Electrophoresis and In-Gel Digest. Electrophoretic separation of proteins from rat liver mitochondrial fractions and spot harvesting were performed as described in previous reports (12, 21). Briefly, 200 µg of rat liver mitochondrial protein was subject to isoelectric focusing using tube gels consisting of 9.2 M urea, 4.5% (v/v) acrylamide, 24 mM CHAPS, 1% (v/v) Biorad Biolyte 5/7 ampholyte (Hercules, CA), 4% (v/v) Biorad Biolyte 3/10 ampholyte, and 0.5% (v/v) Igepal CA 630 and polymerized with 2% (v/v) ammonium persulfate and 0.1% (v/v) TEMED, conducted over 20 h as follows: 200 V for 2 h, 500 V for 2 h, and 800 V for 16 h using 20 mM sodium hydroxide (anode buffer) and 5.9% (v/v) phosphoric acid (cathode buffer). At the completion of isoelectric focusing, each gel was extruded from its respective glass tube by water compression, followed immediately by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) separation. Slab gels (18 cm × 16 cm × 2 mm) used for the SDS-PAGE dimension consisted of a 6-15% linear gradient with a 5% stacking gel and were maintained at 4 °C for the duration of the process using a recirculating water bath. Protein modification by 4-HNE was determined following twodimensional electrophoresis through standard western blotting procedures using antibodies against 4-HNE adducts. Proteins were only considered for harvesting if they stained positive for 4-HNE from no less than three ethanol-fed animals, while the pair-fed control consistently stained negative. Proteins were digested overnight (approximately 16 h) at 25 °C in the presence of 100 mM ammonium bicarbonate and 0.63 µg of sequencing grade recombinant trypsin (Promega Corporation). Aldehyde Modification of PDI and Tryptic Digest. Purified bovine PDI (2.0 µM final concentration) was incubated in the presence of 0, 10, 50, or 100 µM aldehyde or Nethylmaleimide (NEM) in 50 mM sodium phosphate (pH 7.4) overnight (approximately 16 h) at 37 °C. For tryptic digest, 4-HNE-treated PDI was heat denatured at 100 °C for 5 min; however, reducing agents (i.e., β-mercaptoethanol) were not included because of the potential to destabilize aldehyde adducts. The denatured PDI was cooled on ice prior to the addition of acetonitrile (ACN) (5% final concentation) and 0.5 µg of trypsin, and PDI was digested overnight at 37 °C. Mass Spectral Analysis. Peptides (8 µL) from each digest or purified PDI, which had previously been treated with 4-HNE and subject to trypsin digest, were analyzed by liquid chromatography and tandem mass spectrometry (LC-MS/MS) using an Agilent (Palo Alto, CA) 1100 Series LC/ESI-MSD Trap equipped with a Phenomenex (Torrance, CA) Jupiter C18 column (1 mm × 150 mm, 300 Å), as described previously (12). Briefly, the mobile phase consisted of 0.2% formic acid in water (A) and 0.2% formic acid in ACN (B) with a flow rate of 50 µL/min and gradient conditions as follows: 5% B at 0 min, 5% B at 5 min, 70% B at 35 min, 90% B at 38 min and held for 2 min, and 5% B at 42 min and held for 3 min. Mass spectrometric detection and analysis were accomplished using the positive ion mode with a capillary voltage of 3.5 kV. The nebulizer pressure was set at 20 psi and dry gas flow at 8 L/min with the temperature of the dry gas set to 350 °C. The scanning range for all analyses was 400-2000 m/z. Peptides within the mass range of 5001500 Da were subject to MS/MS analysis. A MS/MS ion search was performed on deconvoluted spectra using MASCOT (22). Peptides from tryptic digest of PDI modified by 4-HNE in vitro were identified based on a mass shift of the parent peptide equal

1326

Chem. Res. Toxicol., Vol. 18, No. 8, 2005

to that of 4-HNE (156 Da) in the treated protein using SALSA (23). The identity of the aldehyde-modified peptide and location of the adduct were determined via MS/MS analysis. Fragment ions were calculated using the MS-product feature of Protein Prospector version 4.0.5. Western Blotting. All SDS-polyacrylamide gels used for western blotting were transferred to PVDF membrane using a semidry transfer apparatus (BioRad). All membranes were blocked in 5% (w/v) nonfat dry milk (NFDM) in TBS-T for 30 min at room temperature. Antibody dilutions were performed in 5% (w/v) NFDM in TBS-T. Horseradish peroxidase-linked goat anti-rabbit secondary antibodies (Pierce) were used following a 1:6000 dilution for detection of primary antibodies generated in rabbit hosts. 4-HNE adducts were confirmed using custom antibodies generated in rabbit hosts against 4-HNEmodified keyhole limpet hemocyanin (24). Proteins were probed for 4-HNE adducts overnight at 4 °C using a 1:2000 dilution of the above-mentioned antibody. Antibodies against PDI (SigmaAldrich Chemical Co.) were used at a dilution of 1:2000. Where specified, antibodies were stripped from the membrane in 50 mM tris buffer (pH 6.8) containing 2% (v/v) SDS and 100 mM β-mercaptoethanol for 30 min at 50 °C, followed by extensive washing in TBS-T. Membranes were treated with Amersham ECL-Plus (Piscataway, NJ) enhanced chemiluminescence reagent, and bands or spots visualized using film and or a Molecular Dynamics STORM 860 (Sunnyvale, CA). Bands were quantified using the ImageQuant 5.2 (G.E. Healthcare) software package. PDI-Mediated Insulin Reduction. PDI-catalyzed reduction of insulin disulfides was employed to measure the activity of the enzyme following treatment with the aldehydes, using procedures adapted from a previously published report (25). Briefly, insulin (2 mg/mL final concentration) was added to 100 mM sodium phosphate buffer, pH 7.4, containing 4 mM EDTA and 4 mM GSH, and preheated to 37 °C. Disulfide reduction was initiated by the addition of PDI (0.2 mM), and the rate of insulin precipitation was followed at 640 nm over the course of 30 min using a Molecular Devices Spectra Max 190 UV/vis plate reader. Because insulin modification by free aldehyde or NEM, which could be introduced with the treated PDI, might potentially interfere with the assay described here, control experiments in which the insulin-containing buffer was spiked with 200 µM 4-HNE, 4-ONE, MDA, acrolein (ACR), or NEM (25 µM final concentration) prior to the addition of PDI were performed. However, because pretreating the insulin-containing buffer had no apparent effect,2 likely due to the high (4 mM) GSH concentration, free aldehyde was not removed from the PDI in later experiments. Adduct removal by GSH was performed in an identical manner following pretreatment of PDI with 100 µM 4-HNE; however, standard SDS-PAGE and immunoblotting techniques were employed to measure 4-HNE modification of PDI, as described above. Statistical Methods. Statistical analysis was performed using the software package GraphPad Prizm version 3.02 (GraphPad Software, San Diego, CA). PDI-catalyzed insulin reduction was compared between aldehyde-treated and untreated PDI using one-way ANOVA with Tukey’s post-test. Where appropriate, a significant difference from the control (e.g., untreated PDI) was indicated. 4-HNE immunoblotting data were quantified as described above and normalized to the PDI signal, and ratios were pooled and compared by two-tailed t-test. In all cases, mean differences were considered significant when p < 0.05, although the level of significance was often much greater.

Results An animal model developed in our laboratory for early stage ALD results in elevations of plasma ALT and ammonia levels, and the appearance of microsteatosis or 2

Unpublished observation.

Carbone et al.

Figure 1. Two-dimensional electrophoresis and immunoblot detection of 4-HNE-modified proteins from control (A) and ethanol-fed (B) rat liver mitochondrial fractions. PDI, indicated by the arrow, was identified following in-gel digest and LC-MS/ MS analysis interfaced with the MASCOT search engine.

small triglyceride droplets within the hepatocytes (12). Macrosteatosis, characterized by severe hepatocellular triglyceride accumulation, and inflammatory cell infiltration are typically absent during this stage. A previously published report has documented a 1.8- and 1.3-fold elevation in both plasma ALT and ammonia, respectively, and the presence of hepatocellular microsteatosis in tissue sections prepared from livers harvested from rats receiving the high-fat/ethanol diet (12). The isocaloric control rats displayed none of these characteristics. The appearance of these markers thus demonstrates induction of liver injury in animals receiving the ethanol diet described above. Oxidative stress is central to ALD and is often measured by the appearance of aldehyde species formed during the peroxidation of lipid membranes. Because of the potential role for 4-HNE in the progression of ALD, a proteomic approach consisting of two-dimensional gel electrophoresis and immunoblot with antibodies against the aldehyde was employed to measure the occurrence and magnitude of protein modification by 4-HNE. Data shown in Figure 1 demonstrate a clear elevation in the extent of mitochondrial protein modification by this aldehyde in the livers of ethanol-fed animals (A) over the pair-fed control rats (B). The identification of proteins adducted by 4-HNE was established through harvesting the proteins of interest from a two-dimensional gel, using landmark proteins as references, followed by in-gel trypsin digest and LC-MS/MS peptide analysis. PDI, indicated in Figure 1A (arrow), was among the proteins consistently modified by 4-HNE. PDI contains two thioredoxin-like domains. LC-MS/MS analysis and MASCOT ion search following in-gel digest of PDI isolated from the livers of ethanol-fed animals

Inhibition of Protein Disulfide Isomerase by 4-HNE

Chem. Res. Toxicol., Vol. 18, No. 8, 2005 1327

Figure 3. Concentration-dependent inhibition of PDI-catalyzed insulin disulfide reduction following chaperone treatment with 4-HNE (A), ACR (B), 4-ONE (C), or MDA (D) (n ) 3 ( SD, * indicates significant difference from all groups). Figure 2. Modification of the first PDI active site in protein isolated from the liver of an ethanol-fed rat is suggested through the appearance of a triply charged peptide consistent with the mass of the N-terminal active site-containing peptide (amino acids 36-61, VSDTGSAGLMLVEFFAPWCHGCK) plus the mass of 4-HNE (156 Da) or m/z 871.5 (m/z 2611.47 singly charged) (A). Tandem mass spectrometry of m/z 871.5 was then used to confirm peptide sequence and 4-HNE modification of the active site (B). Table 1. PDI Tryptic Peptides Modified by 4-HNE species

AA range

modified AA

rata

36-61

Cys 57

312-318 341-347 439-446 34-44 288-302 233-249

Cys 314 Cys 345 His 440 His 43 His 297 His 233

bovineb bovineb bovineb bovineb bovineb bovineb

mass (Da)

parent peptide

2611.5 VSDTGSAGLMLVEFFAPWCGHCK 303.4 EECPAVR 905.4 ITEFCHR 928.5 VHSFPTLK 1172.6 GNFDEALAAHK 1833.9 ILFIFIDSDHTDNQR 1965 HNQLPLVIEFTEQTAPK

a In-gel digest (rat liver). b In vitro modification (purified bovine PDI).

demonstrated adduction of a Cys residue in the aminoterminal active site of the enzyme. Further analysis of PDI modification by 4-HNE was conducted using purified bovine enzyme treated with 4-HNE, followed by tryptic digest and SALSA analysis of the resulting peptides. Adducted peptides were identified based on the appearance of a shift in m/z of the parent peptide equal to the mass of 4-HNE (156 Da; Figure 2A). Tandem mass spectrometry of the 871.5 Da peptide confirmed the identity of the peptide, as well as the location of the adduct. Results in Figure 2 demonstrate Cys modification by 4-HNE on an active site-containing peptide (amino acids 39-61; VSDTGSAGLMLVEFFAPWCGHCK). All adducts and peptide sequences identified by SALSA were confirmed through MS/MS fragment analysis. The remaining peptides found to be consistently modified by 4-HNE have been summarized in Table 1. Interestingly, 4-HNE was found to adduct several His- and two Cyscontaining peptides. Unfortunately, modification of either

active site in the bovine PDI was not observed, as these peptides were not identified following tryptic digest. However, in vivo thiol adduction of the active site in the amino-terminal domain of the protein suggests that Cys reactive aldehydes may possess the ability to inactivate PDI through disruption of client protein binding. Although the 4-HNE concentration in lipid-rich environments such as the ER has been estimated to reach millimolar levels under pathological conditions, concentrations much lower than this were capable of inactivating PDI (5). Following pretreatment of the chaperone with 20, 100, and 200 µM 4-HNE (e.g., 10:1, 50:1, and 100:1 4-HNE/PDI ratio), a concentration-dependent inhibition of PDI-catalyzed reduction of insulin disulfide bonds was observed. The data presented in Figure 3A demonstrate up to 54% inhibition of PDI-catalyzed insulin reduction following treatment with 200 µM 4-HNE, with 20 and 100 µM 4-HNE pretreatments leading to 14 and 28% inhibition, respectively. The data in Figure 3A thus demonstrate an ability of 4-HNE to disrupt PDI-catalyzed insulin reduction through direct modification of the chaperone. Elevated levels of 4-HNE protein modification are demonstrated in Figure 1A; however, the lipid peroxidation initiated by the ethanol diet likely resulted in the formation of various other carbonyls as well, which may have modified PDI. Therefore, three other lipid-derived aldehydes, specifically ACR, 4-ONE, and MDA, were also tested for effects on PDI-catalyzed insulin reduction using the same concentrations as above, resulting in a range of inhibition. Inactivation of PDI by ACR and 4-ONE was determined as shown in Figure 3B,C, demonstrating a remarkable susceptibility of PDI to 4-ONE (e.g., 100% inhibition following modification of PDI with 200 µM 4-ONE) when compared with either 4-HNE (54% inhibition) or ACR (60% inhibition) treatment at the same concentration. In contrast, incubation of the isomerase with identical concentrations of MDA (e.g., 20, 100, or 200 µM) had little effect on PDI activity (Figure 3D), resulting in only 21% inhibition of the enzyme at the

1328

Chem. Res. Toxicol., Vol. 18, No. 8, 2005

Carbone et al.

Table 2. Comparative Aldehyde Reactivity with GSH (e.g., Cys) at pH 7.4a aldehyde

k (M-1 s-1)

ref

MDA 4-HNE ACR 4-ONE

NAb 1.09, 1.33 121 145.1

6 6, 7 6 7

a k ) rate constant for GSH-aldehyde complex formation. b No observed rate.

highest concentration. Data in Figure 3A-D thus demonstrate inactivation of PDI by multiple thiol reactive lipid aldehydes. The variation in PDI susceptibility to 4HNE, ACR, 4-ONE, and MDA suggests that inactivation of the enzyme is due to Cys modification, and previous work has documented a range of Cys reactivity between the aldehydes tested presently (6, 7). Table 2 reports published kinetic parameters in which the formation of aldeyde-GSH conjugates at physiological pH (7.4) was measured, providing a means for comparison of Cys reactivity. Specifically, MDA studies conducted by Esterbauer et al. (6) demonstrated minimal reactivity with GSH (6). In contrast, 4-HNE (k ) 1.09-1.33 M-1 s-1), ACR (k ) 121.0 M-1 s-1), and 4-ONE (k ) 145.1 M-1 s-1) react quite readily with GSH, likely through Michael addition, demonstrating a variation in Cys reactivity (e.g., 4-ONE > ACR . 4-HNE) (6, 7). The data presented in Figure 3 demonstrated a difference in the ability of each aldehyde to inactivate PDI. Therefore, a direct comparison was made between the aldehydes by treating PDI with 100 µM aldehyde and measuring the relative inactivation by MDA, 4-HNE, ACR, and 4-ONE. As expected, treatment of PDI with MDA resulted in no significant inactivation of the enzyme (7%), while treatment with 4-HNE, ACR, and 4-ONE resulted in progressively more extensive PDI inactivation (20, 39, and 69%, respectively), demonstrated by Figure 4A. Because the extent of inactivation was reflective of the relative thiol reactivity of each aldehyde, the Cys modifier NEM was used to confirm susceptibility of PDI to thiol modification. Data presented in Figure 4B thus demonstrate concentration-dependent inactivation of PDI following treatment of the enzyme with 20, 100, or 200 µM NEM (e.g., 15, 43, and 62%, respectively). Taken together, the data in Figure 4 demonstrate thiol sensitivity of PDI, suggesting that inactivation of the enzyme is the result of Cys modification. Reduced PDI would be expected to react readily with 4-HNE; however, high intracellular GSH concentrations such as those found under normal conditions (i.e., 4 mM) likely exert a protective effect on the protein through direct interaction with the aldehyde or direct removal of an adduct from PDI. To test the latter, PDI was modified with 100 µM 4-HNE as described above and treated with 4 mM GSH. Removal of the adduct was measured through immunoblotting (Figure 5A) and confirmed by stripping the membrane of anti-4-HNE antibodies and reprobing with antibodies against PDI (Figure 5B). Following normalization of 4-HNE signal to PDI, it was determined that significant removal of the adduct by GSH is possible, as demonstrated in Figure 5C. It is thus conceivable that GSH can protect PDI from 4-HNEmediated inactivation not only by directly depleting free aldehyde from the system but also through the removal of the 4-HNE/Cys Michael addition product.

Figure 4. Inactivation of PDI through modification of Cys residues. A side-by-side comparison using MDA, 4-HNE, and ACR inactivates PDI in a fashion consistent with each aldehyde’s reported Cys reactivity (e.g., MDA < 4-HNE < ACR < 4-ONE) (A). Confirmation of Cys reactivity as a mechanism behind PDI inactivation using the specific Cys modifier, NEM (B) (n ) 3 ( SD, * indicates significant difference from all groups).

Several reports have documented reduced GSH concentrations under conditions of oxidative stress (26-28). While it was shown in Figure 5 that GSH provides a protective mechanism against aldehyde-mediated PDI inactivation, this reducing agent is also required to recycle oxidized PDI. Therefore, under pathologic conditions in which GSH levels are decreased, such as through depletion by free aldehyde species, PDI may become less efficient. The data presented in Figure 6 demonstrate a significant reduction in PDI efficiency as a result of decreased GSH concentration, suggesting that oxidative conditions may be detrimental to PDI function not only through aldehydic inactivation of the chaperone but also indirectly through depletion of GSH.

Discussion In the United States, up to 18 million individuals are thought to suffer from alcoholism, and for the fraction of heavy users who progress to the cirrhotic stage of ALD, the outlook remains quite dismal (29, 30). Although the mechanisms by which ALD progresses remain largely unknown, oxidative stress is generally accepted to be an underlying factor. Given the documented pathologic nature of many lipid peroxidation products associated with oxidative stress, such as 4-HNE, 4-ONE, ACR, and MDA, it is conceivable that these aldehydes exacerbate

Inhibition of Protein Disulfide Isomerase by 4-HNE

Figure 5. GSH-mediated removal of 4-HNE adducts from PDI, supporting a protective role for GSH. Immunblot with antibodies against 4-HNE was used to demonstrate the presence or absence of adducts in PDI treated with (lanes 5-7) or without (lanes 2-4) GSH, with a 4-HNE positive control in lane 1 and a PDI control in lane 8 (A). Antibodies were stripped from the membrane, which was then reprobed with antibodies against PDI (B). A ratio of 4-HNE to PDI signal was used to compare the protective effect of GSH, demonstrating significant removal of the adduct from PDI (C).

Figure 6. GSH depletion (e.g., oxidative stress) may indirectly inactivate PDI by preventing its reduction, thus leaving it trapped in an oxidized state. The dependence of PDI on GSH is demonstrated by a concentration-dependent inactivation of the enzyme corresponding to decreasing GSH levels (n ) 3 ( SD, * indicates significant difference from all groups).

ALD through stable modification of intracellular proteins, likely leading to disruption of cellular processes. Previously, it was shown that the molecular chaperone Hsp72 was consistently modified in cytosol isolated from the livers of rats, which had received a chronic high-fat and ethanol diet and displayed characteristics of early stage ALD (e.g., hepatosteatosis) (12). This same report also demonstated the susceptibility of Hsp72 to inactivation following Cys modification (12). The work presented here is a logical extension of the previous study and has employed the same proteomic approach, consisting of two-dimensional gel electrophoresis, immunoblot, in-gel digest, and LC-MS/MS peptide analysis to demonstrate an increase in the number and severity of protein mod-

Chem. Res. Toxicol., Vol. 18, No. 8, 2005 1329

ification in mitochondrial fractions isolated from the same animals and to identify consistently modified proteins. PDI, which promotes the repair of incorrectly formed disulfide bonds and is thus an important component of protein maturation, was one of several proteins found to be consistently modified by 4-HNE in the mitochondrial fractions of livers harvested from the ethanol-fed animals. This observation is consistent with an earlier report by Suh et al. (31), in which PDI was found to be oxidatively modified in CYP2E1-expressing human hepatoma cells following ethanol adminstration (31). Further peptide analysis of PDI harvested from two-dimensional gels and subject to tryptic digest revealed that 4-HNE modification was occurring in the N-terminal thioredoxin-like domain, which according to the “scanning and release” mechanism proposed by Walker and Gilbert. (16), is also the site for initial interactions with PDI client proteins (16). Two Cys residues are located in this region, and predictably, PDI mutants lacking one or both of the N-terminal thiols demonstrate reduced function, likely as a result of decreased substrate binding. It was thus hypothesized that formation of a stable complex between either of these residues and 4-HNE would likewise result in decreased enzyme efficiency. This idea is supported by an earlier report published by Liu et al. (32), in which NEM and ACR were shown to inactivate reduced PDI, presumably through modification of active site Cys residues (32). When treated with concentrations of 4-HNE ranging from 20 to 200 µM (10-100-fold molar excess aldehyde), a corresponding decrease in the ability of PDI to catalyze disulfide reduction of insulin was observed. Interestingly, despite the predicted susceptibility of PDI to inactivation by 4-HNE, the enzyme was comparatively resistant with respect to the inhibition of other chaperones by this aldehyde. Specifically, while treatment of Hsp72 and Hsp90 with concentrations of aldehyde approaching 100fold molar excess was nearly 100% inhibitory toward chaperoning activity, similar treatment of PDI resulted in only 64% inhibition when compared with an unmodified control, demonstrated in Figure 3A (12). While the immunoblots used to detect aldehyde-modified proteins were specific for 4-HNE, lipid peroxidation yields a variety of electrophilic aldehyde species, many of which have varying reactivities toward thiols. Therefore, PDI was also treated with three other lipid aldehydes, specifically MDA, ACR, and 4-ONE, resulting in varying degrees of inactivation. While PDI inhibition by ACR was similar to that observed following treatment with similar concentrations of 4-HNE (Figure 3B), the chaperone was almost completely resistant to inactivation by MDA, as demonstrated by Figure 3D in which 200 µM MDA (100-fold molar excess) resulted in only 21% inhibition of the protein when compared to an untreated control. However, treatment of PDI with the carbonyl 4-ONE, which has recently emerged as an especially potent electrophile, resulted in significantly more severe inactivation of the enzyme. Specifically, while PDI was relatively resistant to inactivation by 20 or 100 µM 4-HNE, ACR, or MDA, these same concentrations of 4-ONE resulted in 33 and 74% enzyme inactivation, respectively, while concentrations of 4-ONE reaching 200 µM were completely inhibitory toward the protein (Figure 3C). The data presented in Figure 3 thus demonstrate varying degrees of resistance by PDI to 4-HNE, MDA, ACR, and 4-ONE. Supportive of the hypothesis that PDI

1330

Chem. Res. Toxicol., Vol. 18, No. 8, 2005

may be inactivated following modification of active site thiols by aldehydic lipid peroxidation products is the fact that a side-by-side comparison of inactivation by each aldehyde (Figure 4A) correlates with published rates regarding the reactivity of these aldehydes with GSH, presented in Table 2 (6, 7). Confirmation that the mechanism of inhibition is a function of Cys modification was achieved through successful inactivation of PDI with the Cys modifier NEM, demonstrated in Figure 4B. Together, data in Figure 4 strongly suggest that modification of critical Cys residues by thiol reactive lipid aldehydes can result in inactivation of the chaperone. While the aldehyde-mediated inactivation of PDI likely occurs following modification of the Cys residues in the thioredoxin-like domains, direct evidence through 4-HNE pretreatment and tryptic digest of purified PDI, followed by LC-MS/MS peptide analysis, was not obtained. A likely explanation behind this observation is the poor coverage of the active site regions using trypsin. Another interesting possibility is that the lack of modification of the N-terminal active site is the result of client-mediated aldehyde transfer. Because the absence of thiol modification is most likely a function of adduct stability, given the reactive nature of the N-terminal Cys residue, the concept of client-mediated adduct transfer is certainly conceivable. Interestingly, if this is indeed the case, then multiple mechanisms of PDI inactivation are suggested, as a result of protein modification at sites other than the active sites. While protein modification by 4-HNE has proven to be quite specific (e.g., 1-2 consistently modified amino acids) in earlier reports, consistent modification of six amino acids (Table 1), including four His residues, was identified using SALSA following tryptic digest of PDI treated with 4-HNE, with modification of several other unique sites per repetition (12, 33). In each case, 4-HNE adducts were confirmed through MS/MS fragmentation. While the variety of peptides modified by 4-HNE was unexpected, the susceptibility of multiple sites on the protein to modification by lipid aldehydes may serve a protective function, in that these sites may be modified instead of the active site thiol, contributing to the observed aldehyde resistance by PDI. On the other hand, if aldehydic modification of the active site is due to a client-mediated transfer, then modification of the Cys or His residues described above may also lead to PDI inactivation, possibly as a result of intraprotein disulfide disruption or other conformational effects, which may occur following modification of these residues. Additionally, steric inhibition of the active site is also a possibility if the modification is near the reactive Cys. Unfortunately, because the crystal structure for full length bovine PDI has not been solved, the proximity of the modified amino acids to the active site could not be determined. The fact that multiple residues are modified by 4-HNE may account for some measure of the resistance demonstrated by PDI; however, it is likely that high GSH concentrations also exert a protective effect. Specifically, the reduction of insulin, which was used in this report as a measure of PDI activity, was performed in the presence of high concentration (e.g., 4 mM) GSH. While GSH would be expected to prevent any interaction between PDI and free aldehyde, the data presented in Figure 5 demonstrate additionally the ability of GSH to remove bound 4-HNE from PDI, likely through formation of a GSH-HNE conjugate. Although adduct removal was slight, likely due to the stability of the initial adduct with

Carbone et al.

PDI, this mechanism would likely protect PDI under conditions of normal GSH homeostasis. Oxidative stress, such as that resulting from ethanol metabolism, typically results in a decrease in the GSH/GSSG ratio, which could conceivably limit the protective effect of GSH on PDI (26-28). This raises the possibility then that PDI may be more susceptible to modification by 4-HNE and other lipid peroxidation products under conditions of oxidative stress not only because of higher aldehyde concentrations but also because of a decrease in the GSH/GSSG ratio. According to the “scanning and escape” mechanism proposed for PDI function, reduction of oxidized PDI, presumably by high GSH/GSSG ratios, is required for recycling of the enzyme (12). Therefore, the depletion of GSH, which occurs during ALD, may indirectly inhibit PDI by preventing PDI reduction, thus trapping the chaperone in an oxidized state. A final series of experiments were therefore performed assessing the effects of reduced GSH concentrations of the ability of PDI to reduce insulin. The data presented in Figure 6 demonstrate a concentration-dependent inhibition of PDI corresponding to decreasing GSH concentrations. Previous studies with rats chronically ingesting ethanol report reductions in hepatocellular GSH concentrations upward to 65% of control values (28). Because the data presented in Figure 6 demonstrate a significant decrease in PDIcatalyzed insulin reduction following a 50% decrease in GSH concentration (e.g., 4 to 2 mM), this figure suggests the possibility of decreased PDI activity as a function of decreased GSH concentrations as well. The work presented here is a logical extension of a previous report identifying cytosolic Hsp72 as a target for modification and inactivation by 4-HNE in the livers of rats fed an ethanol diet resulting in liver injury consistent with early stage ALD (12, 34). The present report has demonstrated modification of the chaperone PDI in the liver fractions from the same animals, an observation consistent with previous work in which the PDI was oxidatively modified in immortalized human cells challenged with oxidative stress (32). On the basis of the hypothesis that inhibition of molecular chaperones may play a role in the progression of ALD, leading to the accumulation of insoluble or improperly folded proteins, the effects of 4-HNE, ACR, 4-ONE, and MDA on PDI activity were measured. While some degree of protein inhibition was observed, likely through modification of the active site thiols, PDI demonstrated remarkable resistance to inactivation by these aldehydes. One possible mechanism of resistance is the presence of multiple aldehyde binding locations, which may compete with the active site residues for modification by the aldehydes. Additionally, it has also been demonstrated that high GSH concentrations are capable of removing 4-HNE adducts from the chaperone, thus providing another mechanism of resistance. However, identification of 4-HNE adducts at an active site Cys on PDI isolated from the livers of the ethanol-fed animals suggests that this protein may be more susceptible to modification and inactivation under conditions of oxidative stress, likely due to a decrease in GSH concentration and a corresponding decrease in any protective effect that GSH may offer. The data presented here thus demonstrate modification of another chaperone by 4-HNE in a rat model of early stage ALD and suggest inhibition of PDI under conditions of reduced GSH concentrations. Because PDI is an important factor in the maturation of

Inhibition of Protein Disulfide Isomerase by 4-HNE

newly synthesized proteins, inhibition of this chaperone may contribute to ALD by preventing the synthesis of correctly folded, functional proteins.

Acknowledgment. This work was supported in part by Grants NIH/NIAAA RO1AA09300 (D.R.P.), NIH/ NIEHS F32 ES11937 (J.A.D.), and NIH/NIAAA F31 AA014308 (D.L.C.).

References (1) Carini, R., and Albano, E. (2003) Recent insights on the mechanisms of liver preconditioning. Gastroenterology 125, 1480-1491. (2) Lieber, C. S. (2001) Alcoholic liver injury: Pathogenesis and therapy in 2001. Pathol. Biol. 49, 738-752. (3) Liu, P. K., Grossman, R. G., Hsu C. Y., and Robertson, C. S. (2001) Ischemic injury and faulty gene transcripts in the brain. Trends. Neurosci. 24, 581-588. (4) Lefer, D. J., and Granger, N. (2000) Oxidative stress and cardiac disease. Am. J. Med. 109, 316-323. (5) Esterbauer, H., Zollner, H., and Schaur, R. J. (1989) Aldehydes formed by lipid peroxidation: Mechanisms of formation, occurrence, and determination. In Membrane Lipid Oxidation (VigoPelfreys, C., Ed.) Vol. 1, pp 239-268, CRC Press, Boca Raton. (6) Esterbauer, H., Schaur, R. J., and Zollner, H. (1991) Chemistry and biochemistry of 4-hydroxynonenal, malondialdehyde, and related aldehydes. Free Radical Biol. Med. 11, 81-128. (7) Doorn, J. A., and Petersen, D. R. (2002) Covalent modification of amino acid nucleophiles by the lipid peroxidation products 4-hydroxy-2-nonenal and 4-oxo-2-nonenal. Chem. Res. Toxicol. 15, 1445-1450. (8) Uchida, K., and Stadtman, E. R. (1993) Covalent attachment of 4-hydroxynonenal to glyceraldehyde-3-phosphate dehydrogenase: A possible involvement of intra- and intermolecular crosslinking reaction. J. Biol. Chem. 268, 6388-6393. (9) Luckey, S. W., Tjalkens, R. B., and Petersen, D. R. (1999) Mechanism of inhibition of rat liver Class 2 ALDH by 4-hydroxynonenal. Adv. Exp. Med. Biol. 463, 71-77. (10) Fataccioli, V., Andraud, E., Gentil, M., French, S. M., and Rouach, H. (1999) Effects of chronic ethanol administration on rat liver proteasome activities: Relationships with oxidative stress. Hepatology 29, 14-20. (11) Grune, T., Reinheckel, T., and Davies, K. J. A. (1997) Degradation of oxidized proteins in mammalian cells. FASEB J. 11, 526-534. (12) Carbone, D. L., Doorn, J. A., Kiebler, Z., Sampey, B. S., and Petersen, D. R. (2004) Inhibition of Hsp72-mediated protein refolding by 4-hydroxy-2-nonenal. Chem. Res. Toxicol. 17, 14591467. (13) Song, J. L., and Wang, C. C. (1995) Chaperone-like activity of protein disulfide-isomerase in the refolding of rhodanese. Eur. J. Biochem. 15, 231 (2), 312-316. (14) Weissman, J. S., and Kim, P. S. (1993) Efficient catalysis of disulphide bond rearrangements by protein disulfide isomerase. Nature 365, 185-188. (15) Creighton, T. E., Hillson, D. A., and Freedman, R. B. (1980) Catalysis by protein-disulphide isomerase of the unfolding and refolding of proteins with disulphide bonds. J. Mol. Biol. 142, 4362. (16) Walker, K. W., and Gilbert, H. F. (1997) Scanning and escape during protein-disulfide isomerase-assisted protein folding. J. Biol. Chem. 272, 8845-8848.

Chem. Res. Toxicol., Vol. 18, No. 8, 2005 1331 (17) Walker, K. W., Lyles, M. M., and Gilbert, H. F. (1996) Catalysis of oxidative protein folding by mutants of protein disulfide isomerase with a single active-site cysteine. Biochemistry 35, 1972-1980. (18) Mitchell, D. Y., and Petersen, D. R. (1991) Inhibition of rat hepatic mitochondrial aldehyde dehydrogenase-mediated acetaldehyde oxidation by trans-4-hydroxy-2-nonenal. Hepatology 13, 728-734. (19) Requena, J. R., Fu, M. X., Ahmed, M. U., Jenkins, A. J., Lyons, T. J., Baynes, J. W., and Thorpe, S. R. (1997) Quantification of malondialdehyde and 4-hydroxynonenal adducts to lysine residues in native and oxidized human low-density lipoprotein. Biochem. J. 322, 317-325. (20) Little, R. G., and Petersen, D. R. (1983) Subcellular distribution and kinetic parameters of HS mouse liver aldehyde dehydrogenase. Comp. Biochem. Physiol. 74c, 271-279. (21) Shevchenko, A., Wilm, M., Vorm, O., and Mann, M. (1996) Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal. Chem. 68, 850-858. (22) Perkins, D. N., Pappin, D. J., Creasy, D. M., and Cottrell, J. S. (1999) Probability-based protein identification by searching sequence databases using mass spectrometry data. Electrophoresis 20, 3551-3567. (23) Hansen, B. T., Jones, J. A., Mason, D. E., and Liebler, D. C. (2001) SALSA: A pattern recognition algorithm to detect electrophileadducted peptides by automated evaluation of CID spectra in LCMS-MS analysis. Anal. Chem. 73, 1676-1683. (24) Hartley, D. P., Kroll, D. J., and Petersen, D. R. (1997) Prooxidantinitiated lipid peroxidation in isolated rat hepatocytes: Detection of 4-hydroxynonenal- and malondialdehyde-protein adducts. Chem. Res. Toxicol. 10, 895-905. (25) Bonfils, C. (1998) Purification of a 58-kDa protein (ER58) from monkey liver microsomes and comparison with protein-disulfide isomerase. Eur. J. Biochem. 254, 420-427. (26) Lluis, J. M., Colell, A., Garcia-Ruiz, C., Kaplowitz, N., and Fernandez-Checa, J. C. (2003) Acetaldehyde impairs mitochondrial glutathione transport in HepG2 cells through endoplasmic reticulum stress. Gastroenterology 124, 708-724. (27) Aleynik, S. I., and Lieber, C. S. (2003) Polyenylphosphatidylcholine corrects the alcohol-induced hepatic oxidative stress by restoring S-adenosylmethionine. Alcohol Alcohol. 38, 208-212. (28) Fernandez-Checa, J. C., Ookhtens, M., and Kaplowitz, N. (1987) Effect of chronic ethanol feeding on rat hepatocytic glutathione. J. Clin. Invest. 80, 57-62. (29) Li, T. (2004) Alcohol use disorders and mood disorders: A National Institute on Alcohol Abuse and Alcoholism perspective. Biol. Psychiatry 56, 718-720. (30) Stinson, F., Grant, B. F., and Dufour, M. C. (2001) The critical dimension of ethnicity in liver cirrhosis mortality statistics. Alcohol.: Clin. Exp. Res. 25, 1181-1187. (31) Suh, S., Hood, B. L., Kim, B., Conrads, T. P., Veenstra, T. D., and Song, B. J. (2004) Identification of oxidized mitochondrial proteins in alcohol-exposed human hepatoma cells and mouse liver. Proteomics 4, 3401-3412. (32) Liu, X. W., and Sok, D. E. (2004) Inactivation of protein disulfide isomerase by alkylators including alpha, beta-unsaturated aldehydes at low physiological pHs. Biol. Chem. 385, 633-637. (33) Carbone, D. L., Doorn, J. A., and Petersen, D. R. (2004) 4-Hydroxynonenal regulates 26S proteasomal degradation of alcohol dehydrogenase. Free Radic. Biol. Med. 37, 1430-1439. (34) Tsukamoto, H., and Lu, S. (2001) Current concepts in the pathogenesis of alcoholic liver injury. FASEB J. 15, 1335-1349.

TX050078Z