The 90-kDa heat shock protein (Hsp90) accounts for nearly 2% of total protein in most unstressed cells (Lai et al., 1984) and is involved in essential physiological processes, including protein trafficking and signal transduction (Bell and Poland, 2000), protein degradation (Goasduff and Cederbaum, 2000), and regulation and stabilization of a wide range of client proteins (Wegele et al., 2004). In humans, Hsp90 exists in either the β isoform or the more prominent α isoform, which share approximately 85% homology (Hickey et al., 1989). Each isoform consists of an ATPase domain (N-terminal), middle domain, and dimerization domain (C-terminal) (Wegele et al., 2004). Although Hsp90 typically exists as a homodimer, experiments in defined systems have documented a chaperoning activity by self-oligomerization of Hsp90 and client protein binding in a thermally denaturing environment, thus maintaining the substrate protein in a folding-competent form (Yonehara et al., 1996). Refolding of the substrate is accomplished with the addition of Hsp70 along with other essential chaperones and cochaperones such as Hsp60, chaperonin 10, and Hsp-organizing protein (Johnson et al., 1998).

In yeast-expressing nonfunctional Hsp90, it was shown that the absence of this chaperone was lethal (Borkovich et al., 1989). This observation has led to the emergence of Hsp90 as a chemotherapeutic target for the treatment of multiple carcinomas, including breast cancer (Beliakoff and Whitesell, 2004), prostate cancer (Solit et al., 2003), and acute myelogenous leukemia (George et al., 2004). However, although Hsp90 may be a lucrative target for the treatment of cancer, inhibition of Hsp90 in non-neoplastic disorders, such as alcoholic liver disease (ALD), would likely impact cell survival, further complicating the disease.

Oxidative stress, such as that documented to occur during ethanol metabolism, is typically accompanied by the production of electrophilic aldehydes, which are formed following the peroxidation of lipid membranes (Niemela, 1999). Of these compounds, 4-hydroxy-2-nonenal (4-HNE) (Fig. 1A) has been widely studied, and the ability of this aldehyde to disrupt cellular processes through 1,4 Michael addition with Cys, His, and Lys residues, as well as Schiff base formation with primary amines, has been reported (Esterbauer et al., 1991). In addition to 4-HNE, the aldehyde 4-oxo-2-nonenal (4-ONE) is emerging as an especially reactive electrophile (Doorn and Petersen, 2002), likely caused by the ketone group (Fig. 1B), which allows nucleophilic attack at both the second and third carbons, as opposed to 4-HNE, which can undergo nucleophilic attack only at the third carbon.

Lipid aldehydes such as 4-HNE are recognized elements of oxidative tissue injury (Esterbauer et al., 1989), with a well-established ability to disrupt protein function (Uchida and Stadtman, 1993; Luckey et al., 1999; Alderton et al., 2003). Because of their pathogenic potential, these aldehydes were recently the focus of a proteomic study (Carbone et al., 2004) in which it was found that several crucial proteins were modified by 4-HNE in a rat model of early-stage ALD. The current study represents an extension of the earlier report and documents consistent modification of Hsp90 by 4-HNE in the animals receiving an ethanol-containing diet. Because Hsp90 has an established role in maintaining cell homeostasis (Wegele et al., 2004), disruption of this molecular chaperone would impact multiple cell housekeeping pathways and would likely be detrimental to cell survival. Therefore, further experimentation was conducted to test the hypothesis that lipid aldehyde modification of Hsp90 results in decreased chaperoning efficiency. Specifically, in a system using purified recombinant Hsp90 with heat denatured firefly luciferase as a client protein, inhibition of Hsp90-mediated chaperoning by the Cys-reactive aldehydes 4-HNE and 4-ONE was established. Given the importance of Hsp90 in cellular homeostasis, the results presented here suggest that modification of this protein by lipid aldehydes produced by oxidative stress associated with ALD may contribute to progression of the disease. Collectively, this and the previous report (Carbone et al., 2004) provide further evidence that modification of intracellular proteins by lipid peroxidation products, such as 4-HNE and 4-ONE, may contribute to disease progression.

Materials and Methods

Reagents. Unless stated otherwise, all reagents were purchased from Sigma-Aldrich (St. Louis, MO). Untreated rabbit reticulocyte lysate (RRL) was purchased through Promega (Madison, WI). 4-HNE and 4-ONE were synthesized according to procedures described previously, and purity and concentration were confirmed by thin layer chromatography, UV/Vis spectrophotometry, and liquid chromatography/mass spectrometry (Mitchell and Petersen, 1991; Doorn and Petersen, 2002). Recombinant firefly luciferase and a luciferase assay system were purchased through Promega. Human recombinant Hsp90 was purchased from StressGen Biotechnologies (San Diego, CA).

Animals. All the 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 (Indianapolis, IN) Sprague-Dawley rats were fed a modified Lieber-Decarli diet documented previously to initiate liver injury consistent with the early stages of ALD (Carbone et al., 2004). On completion of the feeding protocol, plasma alanine aminotransferase (ALT) determination, tissue harvesting, sectioning, staining, and subcellular fractionation were performed as described in a previous report (Little and Petersen, 1983; Carbone et al., 2004). To assess liver histology, whole liver tissue was fixed in 10% (v/v) sodium phosphate-buffered formalin, pH 7.4. Fat, accumulation, and necrosis were estimated in each H&E-stained section by evaluation of four randomly chosen fields. Fat was assessed by estimating the percentage of cells in each field containing microvesicular or macrovesicular fat, with higher scores indicating a higher percentage of cells containing fat: 0 (absent), 1+ (1–25%), 2+ (25–50%), 3+ (50–75%), and 4+ (100%). Necrosis was scored as 0 (absent), 0.5+ (rare), 1+ (scattered necrotic cells), and 2+ (small foci of >10 necrotic cells). Total pathology scores were determined by summing the scores for steatosis and necrosis. The extent of oxidative injury was also evaluated through immunohistochemical detection of aldehyde-protein adducts in tissue sections from the harvested livers (Sampey et al., 2003) using antibodies shown previously to specifically detect 4-HNE sulfhydryl adducts (Hartley et al., 1999).

Two-Dimensional Electrophoresis and In-Gel Digest. Two-dimensional electrophoretic separation of proteins from rat liver mitochondrial fractions, immunodetection of 4-HNE adducts, and spot harvesting were performed as described previously (Shevchenko et al., 1996; Carbone et al., 2004).

In Vitro Hsp90 Modification by 4-HNE, 4-ONE, orN-Ethylmaleimide, and Tryptic Digest. Because the concentration of purified, recombinant Hsp90 varied between lots, pretreatment of the chaperone was performed as a function of molar ratio to preserve the extent to which the protein was modified. Using this approach, human recombinant Hsp90 was incubated in the presence of aldehyde ranging from 0 to approximately 500 μM (5:1, 10:1, or 100:1 aldehyde/protein molar ratio) in 50 mM sodium phosphate, pH 7.4, overnight (approximately 16 h) at 37°C. Treatment with concentrations of N-ethylmaleimide (NEM) ranging from 6 to 60 μM, corresponding to molar ratios ranging from 1:1 to 10:1, were performed similarly. Free aldehyde or NEM was removed before the addition of the client protein (luciferase) using protein desalting columns (Pierce, Rockford, IL), ensuring no carryover of the thiol modifiers to the protection and refolding reactions. This step was taken to prevent inhibition of luciferase or other components of the refolding system by the thiol modifiers. For tryptic digest, following the aldehyde pretreatment (500 μM 4-HNE), Hsp90 was heat denatured at 100°C for 5 min in the presence of 2 mM β-mercaptoethanol and cooled on ice. The chaperone was then digested overnight at 37°C in 10% (v/v) acetonitrile, 50 mM ammonium bicarbonate, and 0.3 μg of trypsin.

Mass Spectral Analysis. Peptides (8 μl) from each in-gel digest, or purified Hsp90 that had previously been treated with 4-HNE and subjected to trypsin digest, were analyzed by liquid chromatography/tandem mass spectrometry (LC-MS/MS) as described previously (Carbone et al., 2004). Following in-gel digest, peptides within the mass range of 500 to 1500 Da were subjected to MS/MS analysis, and MS/MS ion search was performed on deconvoluted spectra using MASCOT for protein identification (Perkins et al., 1999). Peptides from tryptic digest of Hsp90 modified by 4-HNE in vitro were identified based on a mass shift of the parent peptide equal to that of 4-HNE (156 Da). 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 ProteinProspector version 4.0.5.

Western Blot Analysis. Immunodetection of 4-HNE adducts following electrophoresis and transfer to polyvinylidene difluoride membrane was performed according to procedures published previously (Carbone et al., 2004). Immunodetection of Hsp90 was performed using antibodies purchased from StressGen Biotechnologies, and protein was visualized using GE Healthcare (Piscataway, NJ) STORM 860 imaging system along with the software package ImageQuant version 5.2 (GE Healthcare).

Luciferase Refolding. Recombinant firefly luciferase was diluted to a concentration of 100 nM in 25 mM Tricine, pH 7.8, 8 mM magnesium sulfate, 10% (v/v) glycerol, and 10 mg/ml of bovine serum albumin, and was stored in aliquots at –80°C (Raynes and Guerriero, 1998). Hsp90-mediated chaperoning was measured using an adaptation of procedures published elsewhere (Minami et al., 2001). Briefly, 20 nM luciferase was denatured in the presence of control (i.e., untreated) or 4-HNE-treated Hsp90 by heating the enzyme to 50°C for 5 min. Luciferase refolding was performed by adding 10 μl of the luciferase/Hsp90 mixture to a solution consisting of 50% RRL, 30 mM HEPES, pH 7.0, 50 mM potassium chloride, 2 mM dithiothreitol, 5 mM magnesium chloride, and 1 mM ATP in a final volume of 65 μl. Refolding was allowed to proceed for up to 40 min, at which time 5-μl aliquots were removed from the refolding buffer and added to 100 μl of luciferase assay reagent (Promega) in borosilicate glass tubes. Luminescence was immediately read using a Los Alamos Diagnostics (Los Alamos, NM) 535 luminometer.

Statistical Methods. Statistical analysis was performed using the software package GraphPad Prism version 3.02 (GraphPad Software Inc., San Diego, CA). Plasma ALT, body-to-weight ratio, and body weight were compared using a two-tailed t test. Representative figures showing luciferase refolding were subsampled no fewer than three times and compared by one-way analysis of variance with Bonferroni post test. These representative figures show a consistent trend observed in no fewer than three independent experiments. In all the cases, mean differences between the treated and control (untreated) groups are indicated (*p < 0.05; **p < 0.01; ***p < 0.001).

Results

The extent of oxidative injury induced by the ethanol-containing diet was assessed through the appearance of adduct formation between hepatic proteins and the lipid peroxidation product 4-HNE. Figure 2, A–D, is representative of the immunohistochemical detection of 4-HNE protein modification in liver tissue sections harvested from control (Fig. 2, A and C) and ethanol-fed rats (Fig. 2, B and D). The immunopositive staining in Fig. 2, B and D, clearly shows extensive protein modification by this aldehyde in the ethanol-fed animals compared with controls. Additionally, protein modification by 4-HNE within the hepatocytes appears to be pan-lobular because immunoreactive staining is observed in zones 1 to 3 between the hepatic triad and the central vein. Localization of 4-HNE adduction was observed under higher magnification (40×), in which lipid accumulation was detected in proximity to the staining. Intense immunoreactivity was especially clear in the areas surrounding several larger lipid droplets and is indicated by the arrows in Fig. 2D. The severe staining observed in liver tissue sections harvested from rats fed the ethanol-containing diet thus shows a marked increase in lipid peroxidation.

Indicators of liver injury evaluated in this study include elevated levels of plasma ALT, increased liver-to-body weight ratio, and decreased body weight. These parameters were measured in animals from both ethanol-fed and control groups, and the data are presented in Table 1. Compared with isocaloric controls, the body weights of ethanol-treated animals were approximately 10% less at the end of the study. The data also reveal a significant elevation in liver-to-body weight ratio (1.3-fold) and plasma ALT (4.0-fold). Microscopically, liver injury in animals receiving the ethanol diet was confirmed through H&E staining of tissue sections taken from harvested livers (Fig. 2, E and F). Specifically, the results shown in these panels indicate the presence of microsteatosis and, to a lesser extent, macrosteatosis in the livers of animals receiving the ethanol diet, whereas significant necrosis and the appearance of inflammatory infiltrate were notably absent. These data are summarized in Table 2, showing significant pathology in the livers of animals that had received the ethanol-containing diet. Collectively, these findings of microsteatosis and macrosteatosis are consistent with the hepatocellular changes taking place during the early stages of alcohol-induced liver injury (Diehl, 2001; Tsukamoto and Lu, 2001). Thus, the data presented in Tables 1 and 2 show elevated tissue damage in the livers isolated from rats receiving the ethanol-containing diet accompanied by an elevation in protein modification by the lipid peroxidation product 4-HNE (Fig. 2), showing increased lipid peroxidation.

It is well established that aldehyde modification can dysregulate protein function (Esterbauer et al., 1991). Thus, a proteome-wide scan was used to locate and identify proteins modified by 4-HNE. Two-dimensional electrophoresis of proteins isolated from rat livers followed by immunoblot against 4-HNE-modified proteins has revealed an array of consistently modified targets. Thus, the aldehyde-modified proteins located were identified using in-gel tryptic digest and LC-MS/MS ion search. Figure 3, A and B, presents typical two-dimensional electrophoresis and immunodetection of 4-HNE protein adducts in control (Fig. 3A) or ethanol-fed animals (Fig. 3B), confirming the presence of more severe aldehydic protein modification in animals fed the ethanol-containing diet. Additionally, the location of the chaperone Hsp90 is indicated in Fig. 3B (arrow); however, immunoreactivity of this protein is noticeably absent in protein harvested from animals receiving the control diet. Although several comigrating proteins were also identified along with Hsp90 (Fig. 3B; Table 3), these proteins were not assigned a MASCOT score deemed significant, nor was their location consistent with the actual mass and isoelectric point of the spot (Table 3). However, both mass and isoelectric point are consistent with that which would be expected of Hsp90. Collectively, Fig. 3B and Table 3 suggest that Hsp90 is indeed the prominent protein in the spot harvested from the gel.

Finally, a previous report by Carbone et al. (2004) showed a slight induction of Hsp72 in the same animals used for the present study. Therefore, to ensure that the elevated immunoreactivity observed in Fig. 3B was not simply the result of Hsp90 induction, hepatocellular concentrations of this chaperone were compared via Western blot analysis between animals receiving the control diet (Fig. 3C, lanes 2–5) and their respective ethanol-fed pairs (Fig. 3C, lanes 6–9), showing no detectable induction of the protein. A recombinant control was loaded in lane 10, which served as a positional control. Therefore, Fig. 3 indicates that the enhanced immunoreactivity by Hsp90 likely represents aldehydic modification of the protein by 4-HNE, which is consistent with the increased immunodetection in tissue sections observed in Fig. 2, B–D.

It has been documented previously (Carbone et al., 2004) that modification of the inducible Hsp72 by thiol-reactive aldehydes such as 4-HNE and 4-ONE results in inactivation of this chaperone. Because Hsp90 plays an important role in many cell processes, a similar approach was taken to assess the effect of 4-HNE on Hsp90-mediated chaperoning of thermally denatured luciferase. Specifically, previous reports have shown the ability of Hsp90 to maintain recombinant firefly luciferase in a folding-competent form throughout coincubation under thermally denaturing conditions (Minami et al., 2001). When cooled and placed in the RRL used here, which contains Hsp/70-kDa heat shock cognate protein and other associated chaperones, the Hsp90-protected luciferase is refolded more efficiently than unprotected luciferase. Following pretreatment of Hsp90 with concentrations of 4-HNE ranging from 23 to 450 μM (5–100-fold molar excess aldehyde), a significant and concentration-dependent inhibition of Hsp90-mediated chaperoning was observed, as presented in Fig. 4. Finally, 4-HNE has a documented ability to inactivate Hsp72 (Carbone et al., 2004); therefore, free aldehyde was removed before coincubation of 4-HNE-adducted Hsp90 with luciferase, thus ensuring that any inhibition of luciferase refolding was caused by modification of Hsp90 and not the effect of 4-HNE on either luciferase or Hsp70-mediated refolding.

4-ONE is a recently discovered lipid peroxidation product, with documented thiol reactivity exceeding that of 4-HNE. Therefore, the effects of 4-HNE and 4-ONE on Hsp90 chaperoning were measured. The data presented in Fig. 4 also summarize a side-by-side comparison of inhibition following pretreatment of Hsp90 with 45 μM (10× molar excess aldehyde) 4-HNE or 4-ONE, resulting in a similar extent of protein inactivation. Interestingly, the slight difference in the extent of inhibition by 4-ONE was deemed to be significantly different (p < 0.05) from that following pretreatment of the chaperone with an identical (45 μM or 10× molar excess) concentration of 4-HNE and is likely the result of increased thiol reactivity by 4-ONE (Doorn and Petersen, 2002).

Thus, the increased extent of Hsp90 inactivation by 4-ONE suggests inhibition of this chaperone as a function of Cys modification. This notion was subsequently confirmed through pretreatment of Hsp90 with the Cys modifier NEM (Fig. 5). Specifically, the data in Fig. 5 show that treatment of Hsp90 with concentrations of NEM as low as 6 μM, representing a 1:1 molar ratio between the thiol modifier and the chaperone, almost completely abolished the activity of the Hsp. Higher concentrations were similarly effective at inhibiting Hsp90 chaperoning activity. Again, as with experiments performed with 4-HNE or 4-ONE, excess NEM was removed before coincubation with the client protein (luciferase), ensuring that the observed effect was not caused by NEM interference with luciferase or the RRL.

The nearly complete inactivation of Hsp90 chaperoning activity by a 1:1 M ratio between NEM and the Hsp suggests that consistent modification of a specific Cys residue will lead to inactivation of this chaperone. Indeed, a previous report (Nardai et al., 2000) has suggested a functional role for Cys residues based on the slight reductase activity shown by Hsp90. Therefore, tryptic digest of purified recombinant Hsp90 previously treated with 4-HNE followed by LC-MS/MS peptide analysis was performed, confirming modification of a peptide corresponding to amino acids 568 to 573 (FENLCK; 753.4 Da). A shift in m/z of the parent peptide equal to the mass of 4-HNE (156 Da) is shown by Fig. 6A, and the adduct confirmed through MS/MS of m/z 909.4 (Fig. 6B), in which neutral loss of the adduct reveals the mass of the parent peptide, or m/z 753.4. Finally, the identity of peptide 568 to 573 (FENLCK) and the site of the 4-HNE adduct were confirmed through the appearance of the following fragment ions: y₄ (477.3), b₄-NH₃ (487.2), b₄ (504.3), a₅-NH₃ (562.2), b₅-NH₃ (590.2), and b₅ (607.3). Although sequence coverage of Hsp90 following tryptic digest was incomplete and typically resulted in only 40%, modification of Cys 572 was consistently observed. Interestingly, a previous report had predicted that Cys 572 was among the least reactive of the seven Cys residues possessed by Hsp90 (Nardai et al., 2000). However, the possibility exists that other Cys residues were modified by 4-HNE as well and could not be detected because of loss of the adduct during protein digest and LC-MS/MS procedures or simply because the appropriate peptide could not be identified. Also, despite repeated attempts, because of incomplete digest of Hsp90, the modification of Cys 572 in protein isolated from the ethanol-fed animals could not be confirmed because the limited number of peptides isolated following in-gel digest of the protein did not include this residue. Thus, the spectra in Fig. 6, along with the data presented in Figs. 4, 5, 6, suggest a role for this amino acid in the chaperoning functions of client proteins.

Discussion

Oxidative injury is a component of many diseases, including ALD, atherosclerosis, diabetes, and ischemic injuries such as stroke and myocardial infarction. It is thought that a portion of this injury is mediated by an array of electrophilic aldehyde species, which are spontaneously formed following the peroxidation of lipid membranes (Esterbauer et al., 1989). As a logical extension of a previously published proteomic study (Carbone et al., 2004), the present report shows that the lipid peroxidation product 4-HNE consistently modified an essential molecular chaperone, Hsp90, in a rat model of alcohol-induced oxidative liver injury. Because aldehydes such as 4-HNE have a well-established ability to modify proteins and consequently disrupt protein function (Uchida and Stadtman, 1993; Luckey et al., 1999; Alderton et al., 2003), and because Hsp90 is critical to cell survival (Borkovich et al., 1989), it is conceivable that dysregulation of this Hsp could initiate or augment progression of an existing disorder such as ALD. Given the prevalence of alcoholism in the United States (Li, 2004), as well as the dismal outcome associated with later (cirrhotic) stages of the ALD (Stinson et al., 2001), characterizing the effect of the lipid peroxidation product 4-HNE on Hsp90 was the immediate focus of the present report, with the hope of gaining further insight into the molecular mechanisms behind the ALD.

In the studies described here, Hsp90-assisted firefly luciferase refolding was significantly inhibited in an RRL system following modification of Hsp90 with 45 μM 4-ONE (5:1 M ratio), as well as concentrations of 4-HNE ranging from 23 to 450 μM (5:1 to 100:1 aldehyde/protein ratio; Fig. 4). It should be noted that these concentrations were chosen strictly because of the molar ratio between the aldehyde and the chaperone and do not necessarily reflect physiological or pathological concentrations of 4-HNE or 4-ONE. In fact, intracellular concentrations of 4-HNE have been estimated to reach cytosolic concentrations of up to 10 μM (Poli and Schaur, 2000). Unfortunately, because 4-ONE has only recently been recognized as a potentially pathogenic compound, similar estimates regarding intracellular concentrations are lacking. However, although the concentrations used in the experiments presented here exceed estimates for intracellular 4-HNE concentrations, evidence is provided by this report that Hsp90 is indeed modified by 4-HNE under conditions of oxidative stress and that this modification very likely leads to inactivation of the protein.

A previous report that measured the ability of 4-ONE to modify various amino acids showed higher reactivity of this aldehyde toward Cys than that reported for 4-HNE (Doorn and Petersen, 2002). Additionally, a recent report characterizing the effects of both 4-HNE and 4-ONE on the inducible chaperone Hsp72 showed a thiol-specific mechanism of inactivation (Carbone et al., 2004). The data presented in Fig. 4 support the notion that inactivation of Hsp90 is proceeding through a Cys-based mechanism because treatment of the chaperone with 4-ONE resulted in significantly greater inhibition (p < 0.05) than observed by an identical concentration of 4-HNE. Therefore, to test the hypothesis that inactivation of Hsp90 by lipid peroxidation products such as 4-HNE and 4-ONE is a function of Cys modification, the chaperone was treated with the specific thiol modifier NEM. Confirmation of this notion is presented in Fig. 5, showing nearly complete (i.e., >99%) inactivation of Hsp90 following treatment of the protein with concentrations of NEM as low as 6 μM. Because this concentration represented a 1:1 M ratio between the Cys modifier and Hsp90 in the pretreatment, the data also suggest that modification of a single Cys residue results in inactivation of the chaperone.

Reasons behind the extreme potency of NEM compared with either 4-HNE- or 4-ONE-mediated Hsp90 inactivation remain unknown. However, a plausible explanation is that these two lipid peroxidation products are reactive with amino acids other than Cys. Specifically, both aldehydes, which have the highest reactivity for Cys, are capable of modifying His and Lys, and in the case of 4-ONE, Arg as well (Doorn and Petersen, 2002; Oe et al., 2003). Therefore, it is possible that these other amino acids may also serve as targets for aldehyde modification. Attempts to test this possibility were unfortunately hindered by the relative resistance of Hsp90 to tryptic digest, which is required for LC-MS/MS peptide analysis and adduct characterization. Despite sequence coverage typically in the range of 40%, a stable Cys modification by 4-HNE was observed, lending credibility to the claim of thiol modification as the mechanism behind Hsp90 inactivation.

If indeed modification of Cys residues is responsible for inactivation of Hsp90 by the lipid peroxidation products 4-HNE and 4-ONE, as is strongly suggested through differential inhibition by 4-HNE and 4-ONE, inhibition by NEM, and mass spectral identification of Cys modification by 4-HNE, then a mechanism distinct from that observed with the classic Hsp90 inhibitor geldanamycin is likely. This rationale stems from the fact that geldanamycin interacts with Hsp90 in the ATPase (N-terminal) domain (Whitesell et al., 1992). However, because Hsp90 lacks Cys residues in this domain (Yamakazi et al., 1989), modification of the ATPase region by thiol-reactive aldehydes such as 4-HNE is unlikely. Therefore, the formation of 4-HNE adducts is probably occurring in a region other than the ATPase domain of the Hsp, leading to a mechanism of inhibition that is probably distinct from that of geldanamycin.

Although the exact mechanism of Hsp90 inhibition by 4-HNE and 4-ONE remains unknown, the data presented here show susceptibility of this chaperone to inactivation by lipid peroxidation products through a pathway distinct from classic Hsp90 inhibitors. Furthermore, the consistent appearance of 4-HNE-immunoreactive Hsp90 in the livers of rats chronically ingesting alcohol underscores the importance of characterizing the effect of lipid aldehyde modification on this protein in diseases associated with persistent oxidative stress. Potentially, inactivation of Hsp90 by 4-HNE and other Cys-reactive aldehydes may be a result of impaired substrate or chaperone binding, or ATP hydrolysis. As such, elucidation of these possible pathways is the focus of ongoing research.

Lipid peroxidation occurring under conditions of oxidative stress yields a series of reactive aldehyde species, some of which have received considerable attention regarding the disruption of protein function through modification at critical residues (Esterbauer et al., 1989). However, despite the prevalence and apparent pathogenic potential of these aldehydes, a definite link between lipid peroxidation products and disease progression has remained elusive. Both the study presented here and the previous report characterizing the effects of lipid peroxidation products on Hsp72 function represent an extension of a proteome-wide scan for aldehyde-modified proteins in a rat model of early-stage alcoholic liver injury (Carbone et al., 2004). This proteomic study was conducted under the premise of providing further evidence for a link between lipid aldehydes and disease progression. Specifically, it is the general goal of this work to provide evidence that crucial cellular proteins are targets for modification and inactivation by these aldehydes during disease conditions, thus providing a foundation for more complex experiments examining an in vivo effect.

As stated earlier, the present study represents a logical extension of a previously reported proteome-wide scan (Carbone et al., 2004) to identify substrates for 4-HNE modification in the livers of rats receiving a combination high-fat/ethanol diet. This diet has been shown here and elsewhere to induce early-stage ALD (Diehl, 2001; Tsukamoto and Lu, 2001). Among the proteins modified by this aldehyde is the essential heat shock protein Hsp90, which is involved in crucial cell processes, including protein chaperoning, protein degradation, and protein trafficking. The progression of ALD beyond the steatotic (fatty liver) stage includes the appearance of insoluble cytokeratin aggregates, or Mallory bodies, thus suggesting impairment of protein chaperoning or degradation in this disease. Because Hsp90 is involved in these processes and has been shown to be modified and inactivated by the lipid peroxidation product 4-HNE, the contribution to disease progression through impairment of Hsp90 by lipid peroxidation products is a distinct possibility.