If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
To whom correspondence should be addressed: Dept. of Biochemistry, Molecular Biology and Biophysics, University of Minnesota-Twin Cities, 321 Church St. S E., Minneapolis, MN 55455. Tel.: 612-624-2712; Fax: 612-625-2163
* This work was supported by National Institutes of Health Grant DK053189, the Minnesota Agricultural Experiment Station, and the Minnesota Obesity Center (to D A B.); by NHLBI, National Institutes of Health Grant T32 HL07741 (to P A G.); by National Institutes of Health Grant AG025371 (to T J G.); and by National Center for Research Resources Grant P20 RR17699-01 from the Center of Biomedical Research Excellence (to M J P.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The on-line version of this article (available at http://www.mcponline.org) contains supplemental material.
Obesity is a state of mild inflammation correlated with increased oxidative stress. In general, pro-oxidative conditions lead to production of reactive aldehydes such as trans-4-hydroxy-2-nonenal (4-HNE) and trans-4-oxo-2-nonenal implicated in the development of a variety of metabolic diseases. To investigate protein modification by 4-HNE as a consequence of obesity and its potential relationship to the development of insulin resistance, proteomics technologies were utilized to identify aldehyde-modified proteins in adipose tissue. Adipose proteins from lean insulin-sensitive and obese insulin-resistant C57Bl/6J mice were incubated with biotin hydrazide and detected using horseradish peroxidase-conjugated streptavidin. High carbohydrate, high fat feeding of mice resulted in a ∼2–3-fold increase in total adipose protein carbonylation. Consistent with an increase in oxidative stress in obesity, the abundance of glutathione S-transferase A4 (GSTA4), a key enzyme responsible for metabolizing 4-HNE, was decreased ∼3–4-fold in adipose tissue of obese mice. To identify specific carbonylated proteins, biotin hydrazide-modified adipose proteins from obese mice were captured using avidin-Sepharose affinity chromatography, proteolytically digested, and subjected to LC-ESI MS/MS. Interestingly enzymes involved in cellular stress response, lipotoxicity, and insulin signaling such as glutathione S-transferase M1, peroxiredoxin 1, glutathione peroxidase 1, eukaryotic elongation factor 1α-1 (eEF1α1), and filamin A were identified. The adipocyte fatty acid-binding protein, a protein implicated in the regulation of insulin resistance, was found to be carbonylated in vivo with 4-HNE. In vitro modification of adipocyte fatty acid-binding protein with 4-HNE was mapped to Cys-117, occurred equivalently using either the R or S enantiomer of 4-HNE, and reduced the affinity of the protein for fatty acids ∼10-fold. These results indicate that obesity is accompanied by an increase in the carbonylation of a number of adipose-regulatory proteins that may serve as a mechanistic link between increased oxidative stress and the development of insulin resistance.
The abbreviations used are: ROS, reactive oxygen species; A-FABP, adipocyte fatty acid-binding protein (also referred to as aP2); E-FABP, epithelial fatty acid-binding protein (also referred to as KLBP or Mal-1); 4-HNE, trans-4-hydroxy-2-nonenal (also referred to as 4-hydroxynonenal); 4-ONE, trans-4-oxo-2-nonenal; 1,8-ANS, 1-anilinonaphthalene 8-sulfonic acid; PBST, PBS containing 0.05% Tween 20; ER, endoplasmic reticulum; HRP, horseradish peroxidase; GSTA4, glutathione S-transferase A4 (also referred to as GSTA4-4 or GSTa4); βME, β-mercaptoethanol; JNK, c-Jun N-terminal kinase; eEF1α1, eukaryotic elongation factor 1α-1.
1The abbreviations used are: ROS, reactive oxygen species; A-FABP, adipocyte fatty acid-binding protein (also referred to as aP2); E-FABP, epithelial fatty acid-binding protein (also referred to as KLBP or Mal-1); 4-HNE, trans-4-hydroxy-2-nonenal (also referred to as 4-hydroxynonenal); 4-ONE, trans-4-oxo-2-nonenal; 1,8-ANS, 1-anilinonaphthalene 8-sulfonic acid; PBST, PBS containing 0.05% Tween 20; ER, endoplasmic reticulum; HRP, horseradish peroxidase; GSTA4, glutathione S-transferase A4 (also referred to as GSTA4-4 or GSTa4); βME, β-mercaptoethanol; JNK, c-Jun N-terminal kinase; eEF1α1, eukaryotic elongation factor 1α-1.
and reactive nitrogen species are generated as a result of normal metabolic processes in the cell, including the uncoupling of the electron transport chain in the mitochondria and the oxidation of excess NADPH by NADPH oxidase (
). Oxidative stress occurs due to increased pro-oxidative conditions or the decline of antioxidant systems and is highly correlated with a wide variety of inflammatory and metabolic disease states, including Alzheimer disease, macular degeneration, lung dysfunction, and obesity-linked insulin resistance (
). Although the precise role of oxidative stress in disease mechanisms is not completely understood, ROS are known to be highly reactive with proteins, DNA, carbohydrates, and lipids in the cell. The ROS-initiated peroxidation of polyunsaturated acyl chains of membrane phospholipids can result in a Hock cleavage and the subsequent formation of lipid-derived reactive aldehydes (
Of the variety of reactive aldehydes formed from lipid peroxidation, trans-4-hydroxy-2-nonenal (4-HNE) and trans-4-oxo-2-nonenal (4-ONE) have significant contributions to oxidative disease due to their high abundance and strong reactivity (
). 4-HNE specifically is metabolized (detoxified) by being reduced to an alcohol (aldehyde reductase), oxidized to a carboxylic acid (aldehyde dehydrogenase), or conjugated to glutathione (glutathione S-transferase) (
). However, some fraction of 4-HNE escapes metabolism and forms adducts with other nucleophiles including both protein and DNA. In the case of protein, 4-HNE reacts avidly with the side chains of cysteine and histidine residues as well as lysine albeit to a far lesser extent (
). Although the reactive lipid is formed in membranes, it can diffuse into the cytoplasm and nucleus due to its relatively high solubility, allowing it to react with proteins localized far away from the initial site of oxidative formation (
4-HNE modification of protein has been shown to have a variety of effects on its targets. Alkylation by 4-HNE frequently inhibits the activity of enzymes; examples include the Na+-K+-ATPase and NADP+-dependent isocitrate dehydrogenase (
). In some cases, however, 4-HNE modification has been suggested to increase the activity of key regulatory proteins such as the dimerization and ligand-independent activation of the epidermal growth factor receptor (
). As adipokines secreted by adipose tissue have dramatic effects on insulin sensitivity in liver and muscle, determining the effects of oxidative stress in adipose tissue is an important step in understanding the molecular mechanisms of type 2 diabetes and systemic energy metabolism (
). The modification of adipose proteins with lipid peroxidation products could be one contributing factor linking oxidative stress to insulin resistance.
In this study we investigated the modification of adipose proteins by 4-HNE and other aldehydes in lean and obese C57Bl/6J mice. To accomplish this, hydrazide chemistry was used for coupling with aldehyde adducts on proteins (
). Biotin hydrazide tagging allowed evaluation of the relative extent of carbonylation as well as the identification of novel targets of such modifications, including the adipocyte fatty acid-binding protein (A-FABP).
To assess the relative extent of carbonylation on adipose tissue proteins in lean and obese mice, biotin hydrazide was used as a chemical tag to couple proteins with free carbonyl groups. As an initial means of verifying the ability of biotin hydrazide to detect carbonylated proteins, purified E-FABP, a known target of 4-HNE modification, was evaluated (
). 4-HNE modification was confirmed by the presence of a mass shift of 156 Da when the protein was analyzed by MALDI-TOF MS (Fig. 1). The peak corresponding to a mass of 156 Da larger than that of E-FABP was not detected in the absence of incubation with 4-HNE (Supplemental Fig. 1). After 4-HNE-modified and unmodified E-FABP were incubated with biotin hydrazide, samples were resolved by SDS-PAGE, transferred to PVDF membrane, and incubated with FITC-streptavidin. As shown in Fig. 1, inset, biotin hydrazide modified only 4-HNE-modified E-FABP and not the native form.
Using the methods initially developed for detecting 4-HNE modification of E-FABP, soluble adipose protein extracts were coupled with biotin hydrazide, resolved by SDS-PAGE, transferred to membranes, and blotted with HRP-streptavidin. Reaction variables such as biotin hydrazide concentration, temperature, protein amount, and reaction time were altered to optimize the detection of endogenous carbonylation (Supplemental Fig. 2). The reaction was relatively insensitive to temperature or biotin hydrazide concentration but was dependent upon the amount of protein in the reaction. Standard conditions of 0.5 mm biotin hydrazide, 50 μg of protein, 2 h, and 22 °C were chosen for all subsequent biotin hydrazide/HRP-streptavidin blotting experiments. To confirm that biotin hydrazide is highly specific for aldehydes, as has been demonstrated previously for dinitrophenylhydrazine (
), soluble adipose protein was incubated with NaBH4 to reduce aldehydes to primary alcohols. After dialysis, the protein was coupled with biotin hydrazide and blotted with HRP-streptavidin (Supplemental Fig. 3). The signal intensity for the reduced protein was decreased significantly, confirming that the signal intensity from the unreduced protein coupled with biotin hydrazide was primarily from aldehyde groups on the proteins.
To examine the profile of carbonylated proteins in various tissues as well as to confirm that the endogenous carbonylation detected by biotin hydrazide included 4-HNE-modified proteins, soluble protein extracts were prepared from mouse eye, lung, tongue, and adipose tissues. The protein samples were coupled with biotin hydrazide using the standard conditions, resolved by SDS-PAGE, transferred to PVDF membrane, and blotted with HRP-streptavidin. In addition, parallel samples were subjected to immunochemical analysis using an anti-4-HNE antibody. As shown in Fig. 2, different patterns of protein carbonylation were detected in the various tissues, and in general, the profiles of modified proteins were similar, although the biotin hydrazide method appeared to be more sensitive and detected more proteins than did the anti-4-HNE antibody. As hydrazide chemistry has been shown previously to be specific for carbonyl groups (
). Moreover some bands detected by the anti-4-HNE antibody were not observed with the biotin hydrazide method and may be the result of nonspecific reactivity often observed with antibody detection (dark smear at ∼55 kDa seen in all samples).
To investigate the extent of carbonylation of proteins in obesity, equal amounts (50 μg) of soluble proteins from epididymal adipose tissue of lean and obese mice were coupled with biotin hydrazide, resolved by SDS-PAGE, transferred to PVDF membrane, and blotted with HRP-streptavidin (Fig. 3A). Quantification of signal intensities by densitometry (normalized to total protein) indicated increased (p < 0.05) carbonylation of proteins (∼2–3-fold) in the samples from the obese mice relative to the lean counterparts (Fig. 3B). This finding agrees well with those reported by Talior et al. (
) who measured a 2-fold increase in ROS in adipocytes from high fat-fed mice compared with control chow-fed animals. Consistent with increased ROS in adipose tissue of high fat-fed mice, extracts from obese animals exhibited a 3–4-fold decrease (p < 0.05) in GSTA4 protein compared with lean animals (Fig. 4). GSTA4 catalyzes the Michael addition of 4-HNE to glutathione (with high specificity for 4-HNE as a substrate), and its decreased abundance may contribute to the increased reactive aldehydes in adipose tissue from high fat-fed mice (
To identify proteins that are targets of aldehyde modification, soluble adipose proteins from obese mice were coupled with biotin hydrazide and enriched via affinity chromatography using immobilized monomeric avidin. Bound proteins were eluted with biotin, digested with trypsin, and subjected to LC-ESI MS/MS. Proteins were identified by analyzing parent and fragment ions (Supplemental Fig. 4) using SEQUEST. All data were filtered using the SEQUEST parameters described under “Experimental Procedures.” The experiment was performed multiple times and repeated in the absence of biotin hydrazide coupling to control for nonspecific enrichment by the avidin column. After removing all proteins that were enriched without biotin hydrazide coupling, a list of soluble proteins identified as targets of aldehyde modification in adipose tissue was generated (Table I). Peptides matched to keratins, serum proteins, trypsin, and hypothetical proteins were also removed from the list. The identified proteins sorted into the following clusters: carbohydrate and lipid metabolism, signal transduction, antioxidant enzymes/cell stress response, nucleic acid metabolism, protein synthesis/degradation, and structural/motor proteins. Of particular note was A-FABP/aP2, a cytoplasmic fatty acid carrier protein and a member of the multigene family of lipid-binding proteins that includes previously characterized E-FABP (
Because A-FABP was found to be an in vivo target of aldehyde modification in adipose tissue by LC-ESI MS/MS, adipose extracts from A-FABP-null mice were used to confirm that A-FABP was modified with 4-HNE (Fig. 5). When soluble adipose protein extracts from wild type mice were subjected to aldehyde detection via biotin hydrazide coupling/HRP-streptavidin blotting, a prominently modified protein was detected at ∼15 kDa that was absent in protein extracts from A-FABP-null mice. When these same samples were subjected to immunoblotting with an anti-4-HNE antibody a similar result was observed. To further verify that A-FABP is a target of 4-HNE modification in vivo, the protein was purified from mouse adipose tissue. Both biotin hydrazide/streptavidin blotting and immunoblotting with anti-4-HNE antibody confirmed that A-FABP was modified with 4-HNE in vivo (Fig. 5).
To determine the stoichiometry of in vivo 4-HNE modification of A-FABP, quantitative immunoblotting using antibodies directed to 4-HNE and A-FABP was utilized (Supplemental Fig. 5). The fraction of A-FABP modified with 4-HNE in adipose tissue of obese mice was calculated by determining the concentrations of total A-FABP and 4-HNE-modified A-FABP in tissue extracts. Through immunodetection of either A-FABP or 4-HNE-modified A-FABP, it was estimated that 6–8% of total A-FABP is modified by 4-HNE in adipose tissue of obese mice. Given the total abundance of A-FABP is ∼250 μm in adipocytes (
), the concentration of 4-HNE-modified protein was estimated to be 15–20 μm.
To further characterize the modification of A-FABP with 4-HNE, purified protein was incubated with 0.5 mm (R)- or (S)-4-HNE or a 50:50 racemic mixture for various lengths of time, and the extent of modification was assessed using MALDI-TOF mass spectrometry (linear mode). As shown in Fig. 6, there was no statistical difference between the extent of modification of the protein by (R)- or (S)-4-HNE compared with the racemic mixture.
To identify the site of 4-HNE modification of A-FABP in vitro, His6-tagged A-FABP was incubated with 4-HNE and enzymatically digested with either trypsin or Glu-C, and the resulting peptides were subjected to MALDI-TOF MS/MS analysis (
). The m/z values of abundant ions in the initial MS spectra demonstrated A-FABP sequence coverage of 47.7% for Glu-C and 73.5% for trypsin (data not shown). For both trypsin and Glu-C digestions, parent ions with m/z values corresponding to a peptide with Cys-117 modified by 4-HNE (1076.5 for trypsin and 1628.8 for Glu-C) were isolated and subjected to fragmentation to generate MS/MS spectra (Fig. 7). Many of the ions in the MS/MS spectra observed corresponded to those in the predicted fragmentation pattern for the peptides with Cys-117 modified by 4-HNE (Table II). The 4-HNE modification itself appeared to undergo significant neutral loss fragmentation during the collision-induced dissociation process (labeled as [M + H]+1 − 156 for parent ion and with * for fragment ions in Fig. 7), a commonly observed event during MS/MS analysis of 4-HNE-modified peptides (
) resulting in parent and fragment ions with 4-HNE modification intact in some cases and absent in others. The presence of b ions at 104.0 and 260.1 m/z in the spectrum for the Glu-C-digested peptide (Fig. 7A, labeled as b1 and b1*) verify that the 4-HNE modification is on Cys-117 as these m/z values correspond to N-terminal b1 ions consisting only of cysteine with and without the addition of 4-HNE. The 104.0 m/z peak represents the unmodified Cys-117 with the 260.1 m/z peak representing the same cysteine with the 4-HNE modification intact, giving it an additional 156 m/z units. Furthermore the additional 156 m/z units was only seen on b ions for the tryptic peptide in which Cys-117 is part of the given fragment (ions indicative of 4-HNE modification at Cys-117 are shown in bold in Table II). No fragments besides those containing Cys-117 were found to have 4-HNE adducts, and no other 4-HNE-modified peptides were detected for either digest. A-FABP contains no histidine residues and only one additional cysteine, Cys-1, making it the only other strong candidate as a potential site of in vitro 4-HNE modification. Although ions with m/z values corresponding to peptides with Cys-1 in the unmodified (alkylated with iodoacetic acid) state were detected for both trypsin and Glu-C, the corresponding ion for a 4-HNE-modified peptide was absent in each case. The reactivity of lysine with 4-HNE is much weaker than that of cysteine or histidine, and we found no evidence for a lysine modification on A-FABP at a detectable level. Furthermore the MS/MS spectra in Fig. 7 were submitted to Mascot to search against the mouse database in an unbiased way using 4-HNE as a variable modification. For both the trypsin and Glu-C digests, the Cys-117 HNE-modified peptide derived from A-FABP was the top scoring sequence match with Molecular Weight Search scores of 14 for trypsin and 39 for Glu-C (where a Molecular Weight Search score of greater than 32 indicates identity or extensive homology at p < 0.05) (
To verify that Cys-117 is the in vivo site of modification (and in agreement with the reactivity of the protein with 4-HNE in vitro), A-FABP purified from mouse adipose tissue was digested with Glu-C and subjected to MALDI-TOF MS. The spectrum obtained had 67.9% sequence coverage for A-FABP. Peaks at 1641.8 and 1530.6 m/z were detected, corresponding to unmodified peptides containing Cys-1 (acetylated) and Cys-117 alkylated with iodoacetic acid. Although no peak was observed corresponding to a peptide with Cys-1 modified with 4-HNE or 4-ONE, a cluster of peaks at 1626.6, 1627.6, and 1628.6 m/z (signal:noise between 4 and 5 for each peak) was observed (Supplemental Fig. 6) and is in agreement with the theoretical monoisotopic m/z values of the predicted Glu-C peptide with Cys-117 modified with 4-ONE (1626.8) and 4-HNE (1628.8). Although it is unclear whether the 1628.6 m/z peak is part of the isotope envelope for a monoisotopic peak at 1626.6 m/z (as the 1627.6 peak likely is) or whether it is the result of an independent modification, these m/z values are in agreement with the modified Glu-C peptide observed in the in vitro studies (Table II). In addition, this peak cluster was absent in a control Glu-C digest of E. coli-expressed recombinant A-FABP, which had 74.0% sequence coverage for A-FABP and contained the peak corresponding to the peptide with Cys-117 alkylated with iodoacetic acid (data not shown).
To investigate the effect of 4-HNE modification of A-FABP on its fatty acid binding activity, recombinant protein was modified with 4-HNE, and the binding of the surrogate fluorescent ligand 1,8-ANS was monitored (
). Fig. 8 shows the binding isotherm for the native and 4-HNE-modified protein with native A-FABP exhibiting a Kd of 2.05 ± 0.17 μm and the 4-HNE-modified A-FABP exhibiting a Kd of 23.20 ± 4.18 μm (p < 0.05).
Obesity and type 2 diabetes are correlated with an increase in ER stress (
). One possible contributing factor of obesity to insulin resistance is oxidative stress-induced lipid peroxidation and the modification of proteins by reactive aldehydes such as 4-HNE and 4-ONE. To gain insight into the extent of carbonylation of adipose proteins, biotin hydrazide was used as an affinity tag for proteins with free aldehyde groups (
), to our knowledge this is the first report of direct protein modification in adipose tissue linked to obesity. The ∼2–3-fold increase in protein carbonylation accompanying high fat feeding agrees with other measures of oxidative stress (
). In addition, 37 targets of aldehyde modification were identified in adipose tissue of the obese mice.
Several proteins we identified in this study have been reported previously to be carbonylated or 4-HNE-modified specifically in other systems. Glyceraldehyde-3-phosphate dehydrogenase and glutathione S-transferase M1 have been shown to be targets of 4-HNE (
). In addition, dihydropyrimidinase-related protein 2 has been shown to be carbonylated in brain of Alzheimer disease patients, and triose-phosphate isomerase 1 has been found to be carbonylated in yeast upon treatment with hydrogen peroxide (
). As we found similarities between the detection of proteins with biotin hydrazide and anti-4-HNE antibody in adipose and other tissues, we presume that many of the proteins identified by enrichment via biotin hydrazide coupling are 4-HNE-modified (Fig. 2). However, as amino acid side chain oxidation and other lipid peroxidation products can also result in aldehyde groups on proteins (
). A-FABP and E-FABP are members of the intracellular fatty acid-binding protein multigene family, share ∼60% amino acid identity, and have essentially superimposable secondary and tertiary structures (
). The defining feature of the FABP multigene family is a large interior water-filled cavity that serves as the hydrophobic ligand-binding domain. Lipid binding to A-FABP occurs within a defined fatty acid binding site with the carboxyl group coordinated by Arg-106, Arg-126, and Tyr-128 (
). The decreased binding affinity of 4-HNE-modified A-FABP for the hydrophobic probe 1,8-ANS (Fig. 8) agrees with the finding of Cys-117 as the site of modification (Fig. 7) and is consistent with a variety of crystallographic and NMR studies on the A-FABP apo- and holoprotein forms. X-ray crystallographic analysis of A-FABP in a variety of fatty acid-bound forms (palmitate, stearate, and arachidonate) have shown that the C3 to C6 region of a bound lipid lies in close proximity to Cys-117 (
). This report implies that the ligand binding specificity of FABPs is broader than previously considered and extends to medium chain aldehydes. It will be interesting in the future to determine whether testis FABP, which exhibits the same structurally conserved cysteine residue, is also 4-HNE-modified.
The presence of a cluster of peaks containing ions with m/z 1626.6 and 1628.6 in the MALDI-TOF MS spectrum of in vivo purified A-FABP digested with Glu-C (Supplemental Fig. 6) is consistent with the finding of Cys-117 as the in vitro site of modification (Fig. 7). As further support for in vivo modification at Cys-117, digestion of recombinant E. coli-derived A-FABP (E. coli lacks large amounts of polyunsaturated fatty acid substrates for 4-ONE and 4-HNE synthesis) revealed no peak cluster at these m/z values. It is not surprising that a significant portion of the modified species of A-FABP in vivo may be alkylated with 4-ONE (consistent with 1626.6 peak) at Cys-117 in addition to 4-HNE (consistent with 1628.6 peak). Because 4-ONE contains a more electronegative keto function relative to the hydroxyl group of 4-HNE it has been measured to be significantly more reactive than 4-HNE (
), and A-FABP-null mice are relatively insulin-sensitive compared with wild type C57Bl/6J littermates. In light of this, covalent binding of 4-HNE may promote the holoprotein structure, mimicking the fatty acid-bound state. As the interaction of A-FABP and hormone-sensitive lipase has been shown previously to be fatty acid-dependent (
), the level at which A-FABP is modified with 4-HNE could have significant effects on lipolysis in the adipocyte. In this study it was estimated that 6–8% of A-FABP is modified with 4-HNE in adipose tissue of obese mice. Because of the high level of expression of A-FABP in adipose tissue (250 μm), this seemingly modest fraction that is modified would accumulate to relatively high concentration (
μm HNE-modified A-FABP) in obese mice. The relative extents of modification of A-FABP in lean and obese mice was consistent with the ∼2–3-fold increase in total protein carbonylation (Fig. 3), and increased modification of A-FABP may in part contribute to insulin resistance in obesity through altering hormone-sensitive lipase activity.
A potential role for A-FABP is as an antioxidant protein, scavenging 4-HNE in the cell (
). The fact that A-FABP forms adducts with both enantiomers of 4-HNE suggests that such a 4-HNE-scavenging role would not be completely overlapping with that of glutathione. As both stereoisomers of 4-HNE are formed in the cell, the ability of FABP to form adducts with (R)-4-HNE could be an important way of regulating concentrations of free 4-HNE.
Several other proteins identified in this study as targets of carbonylation may contribute to insulin resistance and inflammation. The well characterized roles of peroxiredoxin 1 and glutathione peroxidase 1 are in reducing hydroperoxides in the cell, including lipid hydroperoxides (
). However, peroxiredoxin and glutathione peroxidase enzymes have also been shown to be important for the regulation of signal transduction pathways through controlling the redox state of proteins and lipids. Specifically the nuclear localization of NFκB, a stress-activated transcription factor known to contribute to insulin resistance in adipose tissue through the control of adipokine expression, has been shown to be negatively regulated in part by peroxiredoxin and glutathione peroxidase proteins (
). Furthermore the presence of active site cysteine (peroxiredoxin 1) and selenocysteine (glutathione peroxidase) residues makes these enzymes strong candidates for being inactivated by alkylation with 4-HNE, which could play a role in activating the NFκB stress response in obesity. In addition, glutathione S-transferase and aldehyde dehydrogenase family members are important antioxidants against 4-HNE and other electrophilic lipids (
). The modification of filamin A with a lipid such as 4-HNE possibly could aid in its localization to the plasma membrane, increasing interactions with the insulin receptor and inhibiting insulin action. Calumenin, a protein involved in the regulation of calcium flux out of the ER, is also an intriguing target of carbonylation as it has been shown that ER stress is activated in obesity (
). During conditions of ER stress, misfolded proteins accumulate in the ER lumen and initiate the unfolded protein response and the activation of JNK, which is known to increase serine phosphorylation of insulin receptor substrate-1 (
). As ER stress can be initiated by calcium release from the ER, the potential inactivation of calumenin by 4-HNE could contribute to the unfolded protein response and JNK activation in an indirect manor (
). In addition, the modification of eukaryotic elongation factor 1α-1 (eEF1α1) could contribute to insulin resistance through the activation of a similar stress response. This protein has been shown to be involved in activating apoptosis under lipotoxic conditions in non-adipose cells (
). Although adipocytes do not undergo apoptosis in response to high amounts of lipid, it is possible that eEF1α1 is involved in activating other stress responses in adipocytes in obesity. Although these proteins are intriguing candidates for being targets of modification by 4-HNE or other lipid peroxidation products, additional studies are necessary to determine the specific modifications and potential effects in each case.
The decreased amount of GSTA4 protein (∼3–4-fold) in adipose tissue of obese mice seen in this study (Fig. 4) is consistent with published microarray data demonstrating decreased GSTA4 mRNA expression in adipose tissue of obese C57Bl/6J mice (
). As such, a 3–4-fold decrease in GSTA4 in adipose cells would presumably contribute to increased accumulation of 4-HNE and other lipid peroxidation products, resulting in elevated protein carbonylation as observed in this study. Further investigation is needed to determine whether protein carbonylation and GSTA4 abundance are components linking obesity to insulin resistance.
We thank all members of the Bernlohr and Griffin research groups who provided helpful suggestions regarding this work and specifically Lisa Smith and Wendy Wright for maintenance and care of the animals and Mikel Roe for assistance with instrument methods. We also thank Dr. LeeAnn Higgins and the staff of the University of Minnesota Center for Mass Spectrometry and Proteomics for instruction on instrument operation and the Minnesota Supercomputing Institute for maintenance of the SEQUEST cluster.
Evaluating preparative isoelectric focusing of complex peptide mixtures for tandem mass spectrometry-based proteomics: a case study in profiling chromatin-enriched subcellular fractions in Saccharomyces cerevisiae.