| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 282, Issue 30, 22040-22051, July 27, 2007
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1
From the
Medical Research Council Dunn Human Nutrition Unit, Wellcome Trust/Medical Research Council Building, Hills Road, Cambridge CB2 0XY and the Department of Biochemistry, Cambridge System Biology Centre, University of Cambridge, Cambridge CB2 1GA, United Kingdom
Received for publication, May 1, 2007 , and in revised form, May 24, 2007.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
8.0) make modification of mitochondrial protein thiols by ROS more likely than for cytosolic thiols (15). Some mitochondrial proteins have shown ROS-dependent thiol oxidation:peroxiredoxin III (PrxIII) (16), NADP+-dependent isocitrate dehydrogenase (17), and complex I (18, 19). The role of protein thiol modification in mitochondrial redox regulation, however, remains unclear, in large part because of the difficulty of identifying mitochondrial thiol proteins responsive to the low ROS fluxes found in vivo.
To explore the role of mitochondrial protein thiols in redox signaling, we developed a sensitive proteomic approach to identify thiol proteins that are modified by low levels of endogenous ROS. We adapted the difference gel electrophoresis (DIGE) technique (20, 21) labeling protein thiols in control and redox-challenged samples with two different fluorescent dyes containing thiol-reactive maleimide groups, followed by detection of changes in the relative fluorescence on a single two-dimensional gel. This approach, which we call redox-DIGE, enabled us to identify proteins containing redox-sensitive thiols by peptide mass fingerprinting and tandem mass spectrometry. When this technique was applied to mitochondria exposed to low ROS fluxes emanating from the respiratory chain, such as occur during redox signaling in vivo, there was selective oxidation of a small number of redox-sensitive thiol proteins. These proteins were predominantly involved in fatty acid oxidation or in the regulation of pyruvate dehydrogenase. This finding indicates that the reversible modification of a small subset of mitochondrial protein thiols occurs during low endogenous ROS fluxes, suggesting that protein thiol modification may be an important but neglected aspect of mitochondrial redox signaling.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Protein Thiol, Glutathione, and Hydrogen Peroxide Measurements—Free mitochondrial protein thiols were measured using DTNB or biotinylated maleimide (BMAL). For the DTNB assay (24), pelleted mitochondria were resuspended in standard assay medium containing 1% (w/v) n-dodecyl 
-D-maltoside (DDM). To remove glutathione, the mitochondria lysates were then passed through a Micro Bio-Spin 6 chromatography column (pre-equilibrated with standard assay medium containing 1% (w/v) DDM), diluted 17-fold with 200 µM DTNB, 80 mM sodium phosphate, 1 mM EDTA, pH 8.0 (NaOH), and incubated for 20 min at room temperature before measuring A412 (
412 (thionitrobenzoic acid) = 13,600 M-1 cm-1). Standard curves were generated with GSH, and protein concentration was determined by the bicinchoninic acid method with BSA as a standard (25). For BMAL measurements (18), rat heart mitochondria after various incubations were pelleted by centrifugation and resuspended in standard assay medium containing 100 µM BMAL and 1% (w/v) DDM, for 5 min at 37 °C. The samples were then re-pelleted by centrifugation, resuspended in sample loading buffer (50 mM Tris-HCl, pH 6.8, 2% SDS, 10% (w/v) glycerol, 0.1% bromphenol blue, and 50 mM dithiothreitol (DTT)), separated on 12.5% acrylamide gels using a Mini Protean system (Bio-Rad), and transferred to polyvinylidene fluoride (0.45 µm; Millipore) with a Mini Protean transfer cell (Bio-Rad). The blots were probed with ExtrAvidin (Sigma) and visualized by enhanced chemiluminescence (GE Healthcare).
GSH and GSSG were assayed using the recycling assay (26, 27) adapted for a 96-well plate reader. After various incubations, mitochondria (2 mg of protein) were pelleted by centrifugation, and the pellet was resuspended in 100 µl of 5% (w/v) 5-sulfosalicylic acid and 0.2% (w/v) Triton X-100, and the samples were centrifuged for 10 min at 13 000 x g. To determine the total glutathione equivalents (GSH + 2GSSG), 10 µl of supernatant or 10 µl of standard (0-70 µM GSH in 5% (w/v) 5-sulfosalicylic acid) were incubated in triplicate in 285 µl of recycling assay buffer (125 mM sodium phosphate, pH 7.5, 5.5 mM EDTA, 183 µM NADPH, and 0.53 mM DTNB) with 0.7 units/ml glutathione reductase and read (A405 for 10 min; SpectraMax Plus 384; Molecular Devices). GSSG was measured by incubating 60 µl of supernatant or standard (0-20 µM GSSG in 5% (w/v) 5-sulfosalicylic acid) with 3.4 µl of 2-vinylpyridine and 2.8 µl of triethanolamine sealed under argon at 4 °C for 1 h with agitation. 10-20 µl of the samples in triplicate were then added to 285 µl of recycling assay buffer with 3.5 units/ml glutathione reductase and read (as above). To determine protein mixed disulfides with glutathione (28), the pellets were washed three times in 250 µl of 5% (w/v) 5-sulfosalicyclic acid. The washed pellets or 10 µl of standard (0-400 µM GSSG in 5% (w/v) 5-sulfosalicylic acid) were resuspended in 65 µl of 8 M urea, neutralized with 15 µl of 200 mM Tris-HCl, pH 7.4, and reduced by the addition of 20 µl of 10% (w/v) NaBH4 (in 200 mM Tris-HCl, pH 7.4) for 30 min at 40 °C. After the incubation, 25 µl of 5% (w/v) 5-sulfosalicyclic acid was added; the samples were incubated at 23 °C for 15 min and then on ice for 15 min before centrifugation for 5 min at 16,000 x g. The top layer (if any) was aspirated off, and the supernatant (10 µl) was assayed in the same way as the GSSG measurements. The average total glutathione concentration ± S.D. in control samples was 1.6 ± 0.2 nmol/mg mitochondrial protein.
Efflux of H2O2 from isolated mitochondria was measured by the horseradish peroxidase oxidation of Amplex Red to fluorescent resorufin (29). Mitochondria were incubated with stirring at 37 °C in standard assay medium containing 50 µM Amplex Red (Molecular Probes) and 4 units/ml horseradish peroxidase. Resorufin was monitored continuously in a fluorimeter (Shimadzu Rf-5301PC) (
ex = 560 nm, em = 590 nm), and the response was calibrated using known amounts of H2O2 (240 = 43.5 M-1 cm-1).
Sample Preparation for Redox-DIGE—After incubation 1 ml of 50 mM N-ethylmaleimide (NEM) was added, and the mitochondria samples were then pelleted. Pellets were resuspended in assay medium containing 50 mM NEM for 5 min at 37 °C; 1% SDS was then added, and the mitochondrial lysate was incubated for a further 5 min. The NEM was removed using a Micro Bio-Spin 6 chromatography column (Bio-Rad), and a 1% concentration of SDS was maintained throughout the assay. The samples were then reduced with 2.5 mM DTT for 10 min at room temperature. DTT was removed with two Micro Bio-Spin 6 chromatography columns (Bio-Rad), and the samples were labeled with 40 µM CyDyeTM DIGE Fluor CyTM3 saturation dye (GE Healthcare) or CyDyeTM DIGE Fluor CyTM5 saturation dye (GE Healthcare). After 30 min at 37 °C, the reaction was quenched with 2.5 mM DTT, and equal amounts of the Cy3 and Cy5 maleimide labeled samples were pooled and resolved by two-dimensional electrophoresis. Careful optimization experiments were performed for the initial blocking step with NEM, the DTT reduction and the subsequent Cy maleimide labeling steps of the redox-DIGE method using the DTNB assay to quantify free thiols. For the blocking step, 25 mM NEM was used because at this concentration 99% of the free thiols in the samples was alkylated. For the DTT reduction, a range of DTT concentrations were tested, and maximal free thiol concentrations were observed when 0.5 mM DTT or greater was used. Experiments were carried out with 2.5 mM DTT because concentrations greater than this interfered with Cy maleimide labeling. For the Cy maleimide labeling, careful control experiments were conducted to optimize the concentration, time, temperature, and pH of the reaction so that complete alkylation of thiol residues occurred. At higher concentrations of Cy maleimides, charge shifting occurred on the two-dimensional gels, presumably because of the Cy maleimides reacting with protein amines in the sample. The Cy3 maleimide appeared to react at a greater rate with amines than the Cy5 maleimide resulting in color changes that did not reflect changes in protein thiol state.
Two-dimensional Electrophoresis—Equal volumes of 2x sample buffer (7 M urea, 2 M thiourea, 2% amidosulfobetaine-14, 20 mg/ml DTT, and 2% Pharmalytes 3-10 NL) were added to the fluorescently labeled samples (25 µg of total protein per gel) and incubated for 15 min at room temperature. Rehydration buffer (7 M urea, 2 M thiourea, 2% amidosulfobetaine-14, 2 mg/ml DTT, and 1% Pharmalytes 3-10 NL) was added to make the volume up to 250 µl prior to isoelectric focusing. IPG strips (13 cm, pH 3-10 NL; GE Healthcare) were rehydrated with the samples for 10 h at 20 °C at 20 V using the IPGphor II apparatus following the manufacturer's instructions (GE Healthcare). IEF was performed for a total of 41,700 V-h at 20 °C at 50 µA. Prior to SDS-PAGE, the strips were equilibrated with rocking for 15 min in 100 mM Tris-HCl, pH 6.8, 30% glycerol, 8 M urea, 1% SDS, 5 mg/ml DTT. The strips were loaded onto a 12%, 1-mm acrylamide gel and overlaid with 1% agarose in SDS running buffer (5 mM Tris, pH 8.3, 192 mM glycine, and 0.1% SDS) containing 0.2 mg/ml bromphenol blue. The gels were run at 15 °C at 20 mV for 15 min and then at 40 mV until the bromphenol blue dye front had run off the bottom of the gels.
Gel Imaging and Analysis—After two-dimensional electrophoresis, gels were transferred to a TyphoonTM 9410 imager (GE Healthcare), and fluorescent spots were viewed using 532 and 633 nm lasers in conjunction with 580 and 670 nm emission filters (band pass 30 nm), respectively. The gels were then stained with Deep PurpleTM total protein stain (GE Healthcare), and the protein spots were imaged using a 457 nm laser in conjunction with a 610 nm band pass emission filter. All gels were scanned at 100 µm pixel size, and the photomultiplier tube (PMT) was set to ensure a maximum pixel intensity between 40,000 and 80,000 to avoid saturation. Most proteins were present as multiple spots aligned horizontally in charge trains. Each spot within a charge train was assumed to represent the same protein, and this was confirmed where possible by mass spectrometry. Gel analysis was performed using DeCyderTM DIA version 5.0 (GE Healthcare), a two-dimensional gel analysis software package designed specifically for DIGE. Using DeCyderTM software, a range approach for determining significance thresholds was established (30). Spots were co-detected between the two fluorescence images from a single gel and any with a volume less than 0.01% of the total fluorescence or a slope (the change in fluorescent intensity versus pixel width) greater than 1 were excluded as artifacts. The DeCyderTM (version 5) program then generated a frequency histogram of the log volume ratios and normalized the Cy3 fluorescent image (V1) and Cy5 fluorescent image (V2) by adjusting the spot volume of the Cy3 fluorescent image (V1') such that the frequency histogram of the log volume ratios centered on zero (i.e. a Cy5/Cy3 fluorescence ratio of 1). To do this, a normal distribution is fitted to the main peak of the histogram using a standard least squares gradient descent. Outliers were excluded from the model curve fitting procedure by discarding spots below 10% of the top of the main peak in the frequency histogram of the log volume ratios. The calculated normalized log volume ratios were then converted into expression ratios (E) using the equation E = (V2/V1'). The expression ratios from the analysis of 6 control versus control gels were then ordered from smallest to largest, and a range of expression ratios that encompassed 99% of the spots was established, which was in the ratio range from 0.47 to 1.73. Only spots with intensity outside this 99% confidence threshold on three gels were taken as significant. As less than 0.1% false positives were observed in the six control versus control gels, this indicates that the selected significant spots are significant with a p < 0.001. Three independent biological replicates were conducted for the 10 µM SNAP, 100 µM SNAP, 25 µM H2O2, and 250 µM H2O2 conditions, four for the 2.5 µM H2O2 and antimycin conditions, and six for the RET conditions.
Spot Excision and Mass Spectrometry—For mass spectrometry prior to gel casting, a glass plate was treated with 0.1% PlusOneTM Bind Silane (GE Healthcare), 2% acetic acid, and 80% ethanol, and two plastic fluorescent reference marker stickers were attached to the plate. Then 200 µg of mitochondrial protein was run (see above); a spot-picking list generated from DeCyder software was exported to Ettan Spot Picker (GE Healthcare), and the gel spots were excised using 1.4-mm diameter plugs. Excised protein spots were digested by "in-gel" cleavage (31) at 37 °C with 12.5 ng/ml sequencing grade trypsin (Roche Applied Science) in buffer consisting of 20 mM Tris-HCl, pH 8.0, and 5 mM CaCl2. Peptides were extracted from the gel with a 4% ARISTAR-grade formic acid, 60% acetonitrile (Romil) solution. All digests were examined in a MALDI-TOF-TOF mass spectrometer (4700 Proteomics Analyzer, Applied Biosystems) using 
-cyano-hydroxy-trans-cinnamic acid as the matrix. The instrument was calibrated with bovine trypsin autolysis products (m/z values, 2163.057 and 2273.160) and a calcium-related matrix ion (m/z value, 1060.048). Peptide masses and peptide fragmentation data were searched against the National Center for Biotechnology Information (NCBI) number 20060817 data base against all entries or Rattus with a peptide tolerance of ±25 ppm, tandem mass spectrometry tolerance of ±0.8 Da, allowed missed cleavages of 1, and with NEM as a variable modification using MASCOT. In an attempt to identify the specific sites of redox modification, all MASCOT searches were conducted again with the inclusion of the Cy5 and Cy3 maleimide modifications, but in no cases were these modifications observed.
The identification of protein spots was primarily based on peptide mass fingerprinting that gave a statistically significant MASCOT score for assignment of the protein, based either on all data base entries or on Rattus entries only. In addition, the assignment was corroborated by tandem mass spectrometry sequencing of peptides, which often itself was sufficient to give a statistically significant identification of a protein. In the one case where the assignment based on peptide mass fingerprinting was not statistically significant (Complex III, Rieske FeS protein), peptide sequencing was significant. Therefore, most assignments are based on both peptide mass fingerprinting and peptide sequencing, with a few assignments based on peptide fingerprinting alone and one based on peptide sequencing. In the case of pyruvate dehydrogenase kinase, isoenzyme 2, identity was further supported by immunoblotting. Although it is impossible to entirely exclude the possibility that redox-DIGE picks a low abundant thiol protein that is misidentified because it co-migrates with a more abundant protein, none of the protein spots positively identified here contained significant amounts of peptides from other proteins.
Enzyme Assays—The presence of sulfinates or sulfonates on Cys47 of PrxIII was detected by Western blotting using a rabbit polyclonal antibody (LF-PA0004, LabFrontier) that recognizes cysteine sulfinic and sulfonic acids in the active sites of peroxiredoxins. PrxIII oligomerization was assayed by nonreducing Western blotting using a polyclonal antibody against PrxIII (LF-PA0030, LabFrontier). To prevent spontaneous dimerization of PrxIII, samples were incubated in 100 mM NEM for 15 min at 37 °C before the mitochondria were pelleted and lysed in sample loading buffer containing 100 mM NEM without DTT.
The oxidation state of Trx2 was monitored using a protein electrophoretic mobility shift assay (32). After various incubations, mitochondria (0.3 µg of protein) were pelleted by centrifugation, resuspended in 100 µl of urea buffer (8 M urea, 100 mM Tris-HCl, pH 8.0, and 1 mM EDTA) with 60 mM iodoacetic acid, and incubated for 2 min at 37 °C. The iodoacetic acid was then removed using a micro Bio-spin column (pre-equilibrated with urea buffer), and 10 mM DTT was added to the samples for 30 min at 37 °C followed by the addition of 30 mM iodoacetamide for 15 min at 37 °C. 3.5 mM DTT, 8% glycerol, and 0.1% bromphenol blue was then added to the samples, and they were resolved on a 1-mm urea-polyacrylamide gel (stacking gel contained 8 M urea, 2.5% acrylamide, 0.075% bisacrylamide, 0.12 M Tris-HCl, pH 6.8, and the separating gel consisted of 8 M urea, 9% acrylamide, 0.27% bisacrylamide, 0.037 M Tris-HCl, pH 8.8) using a Mini Protean system (Bio-Rad) at 120 V. The chamber buffer used for gel electrophoresis was 0.025 M Tris, 0.192 M glycine, pH 8.3. The gels were transferred to polyvinylidene fluoride (0.45 µm; Millipore) using a Mini Protean transfer cell (Bio-Rad), and the blots were probed with a Trx2 antibody (1:500; Abcam) and visualized by enhanced chemiluminescence (GE Healthcare).
Propionyl-CoA carboxylase (PCC) activity was assayed radiochemically by counting the incorporation of [14C]bicarbonate (GE Healthcare) into methylmalonyl-CoA in mitochondrial lysates (33). After various incubations, rat heart mitochondria (50 µg of protein) were pelleted by centrifugation, resuspended in medium containing 50 mM Tris-HCl, pH 8.0 (23 °C), 2 mM ATP, 125 mM KCl, 10 mM MgCl2, 6.0 mM propionyl-CoA, 0.5 mg/ml BSA, 0.1% Triton X-100, and 10 mM NaH[14C]O3 (specific activity 0.1 Ci/mol) in a final volume of 50 µl, and incubated at 37 °C for 5 min. The reaction was terminated by adding 50 µl of 10% trichloroacetic acid; the mixture was centrifuged at 13,000 x g for 5 min, and 95 µl of the supernatant was dried in a scintillation vial in a water bath at 80 °C for 50 min. The dry residue was dissolved in 150 µl of H2O, 4 ml of scintillation fluid (Scintran Fluoran-Safe 2; BDH) was added, and the samples were counted (Liquid Scintillation Analyzer Tri-Carb 2800TR; PerkinElmer Life Sciences). The counts from samples in the assay mixture without propionyl-CoA were subtracted for those with propionyl-CoA, and protein concentration was determined by the bicinchoninic acid method using BSA as a standard (25). The incorporation of [14C]bicarbonate was linear over 10 min and proportional to the amount of mitochondrial lysate assayed. The average propionyl-CoA carboxylase activity ± S.D. in control samples was 57 ± 10 nmol/min/mg mitochondrial protein.
The activity of the pyruvate dehydrogenase kinase (PDHK) was determined by measuring the initial rate of incorporation of [32P]phosphate from [
-32P]ATP into exogenously added pyruvate dehydrogenase complex (34, 35). After various incubations mitochondria were pelleted by centrifugation and incubated in 25 µl of buffer (20 mM potassium phosphate, pH 7.0, 1.0 mM MgCl2, 0.1 mM EDTA, 25 µg/ml oligomycin, 10 mM NaF, 0.73 units/ml porcine pyruvate dehydrogenase complex), 0.25% (w/v) Triton X-100, and 0.3 mM [-32P]ATP (specific activity 333 Ci/mol) for 1 min at 30 °C. The reaction was terminated by adding 25 µl of 2x sample buffer (100 mM Tris-HCl, pH 6.8, 4% (w/v) SDS, 20% (w/v) glycerol, 0.2% (w/v) bromphenol blue, and 100 mM DTT), and the samples were analyzed immediately by SDS-PAGE (4-20% PreciseTM protein gels; Pierce) followed by phosphorimaging (TyphoonTM 9410 Imager; GE Healthcare). The phosphorylation of pyruvate dehydrogenase was linear for at least 2 min and proportional to the amount of mitochondrial lysate assayed.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
100 µM) or H2O2 (250 µM) gave no detectable loss in free protein thiols (Fig. 1, A and B), and higher concentrations were required to oxidize bulk protein thiols (data not shown). To investigate whether ROS production from the respiratory chain within mitochondria altered protein thiols, we measured the effects of endogenous mitochondrial ROS flux from the respiratory chain. This was achieved at complex I by reverse electron transport (RET), induced by incubation with succinate and compared with a control where RET was prevented by the uncoupler FCCP (Fig. 1C) (37, 38), and at complex III by antimycin inhibition (39) (Fig. 1C). In both cases superoxide was converted to H2O2, which was measured as it diffused from the mitochondria (Fig. 1C). This endogenous ROS production did not cause measurable changes to free mitochondrial protein thiols when they were measured by the DTNB assay (Fig. 1D). Therefore, either ROS modification of protein thiols has no part to play in mitochondrial redox signaling or such changes are not detected by our insensitive measures of bulk protein thiols. To test this latter hypothesis, we developed a method to identify low abundant thiol proteins that were particularly susceptible to reversible thiol oxidation by endogenous ROS. The Redox-DIGE Method—Within cells most exposed protein thiols are maintained in a reduced state by the glutathione and thioredoxin systems, therefore measuring protein thiol oxidation is far more sensitive than measuring thiol loss (40). Most such approaches have measured the gain of oxidized thiols in a redox-challenged sample on a two-dimensional gel, compared with a control on a separate two-dimensional gel. However, these techniques are limited by having to compare two gels and by the difficulty of distinguishing proteins with persistently occluded protein thiols, such as those in iron sulfur (FeS) centers, from those that become oxidized upon redox challenge. To eliminate these problems while maintaining the sensitivity of measuring thiol oxidation by fluorescence, we have adapted the DIGE technique. Early proteomic approaches to determine differences in protein expression relied on comparing pairs of two-dimensional gels; however, gel-to-gel variation hampered detection and quantification. By using fluorescently resolvable dyes, the DIGE technique enabled comparison of multiple samples on the same gel and was a significant advance in reproducibility and quantification (21, 41). Here we have used DIGE dyes that label cysteine residues through a maleimide group (42) to monitor sensitively small changes in protein thiol redox state. Proteins containing thiols susceptible to oxidation were identified by a change in fluorescence of the protein spot on a two-dimensional gel. The protein could then be identified by peptide mass fingerprinting and tandem mass spectrometry.
|
|
|
-cardiac actin, complex III core protein I, and PrxIII), indicating that some of their cysteine residues were constitutively oxidized in our incubations. Sarcomeric mitochondrial creatine kinase is present in the intermembrane space and has an active site cysteine and other cysteine residues that are modified by ROS (43). The -cardiac actin protein is an expected cytosolic contaminant of mitochondria that contains cysteine residues that are known to be glutathionylated (44). Complex III core protein I protrudes into the matrix and has cysteine residues on its surface (45); however, little is known about their oxidation state. PrxIII is an abundant, thioredoxin-dependent peroxidase within mitochondria whose redox state and activity can be modified by H2O2 and Trx2 redox state (46). There were also a number of yellow spots of low volume relative to the spot's protein abundance, indicating that only a small proportion of the protein in that spot contained NEM-unreactive thiols. This may be due to low abundant proteins with oxidized protein thiols co-migrating with an abundant protein or occur because only a small fraction of the protein's thiols were oxidized in control mitochondria.
|
|
Effect of Endogenous Superoxide Production on Mitochondrial Protein Thiols Determined by Redox-DIGE—As mitochondrial redox signaling pathways in vivo are likely to involve superoxide production from the respiratory chain, we used Redox-DIGE to determine whether endogenous ROS from the respiratory chain oxidized protein thiols within mitochondria. We generated ROS within mitochondria at complex I by RET or at complex III by inhibition with antimycin. RET closely mimics ROS production from the respiratory chain during redox signaling in vivo, as it occurs spontaneously without addition of inhibitors and may be an actual mode of ROS signaling in vivo. As shown in Fig. 1, both conditions generated a significant ROS flux within mitochondria without affecting bulk mitochondrial thiols. Endogenous mitochondrial ROS production from RET led to significant redox changes to six thiol proteins, and antimycin led to the oxidation of three of these proteins, and a further three were modified by antimycin but not by RET (Fig. 4). All of the proteins modified by RET were also altered by exogenous H2O2, but one of the proteins modified by antimycin was unaffected by H2O2, suggesting a superoxide-specific effect. Therefore, endogenous ROS production from the respiratory chain can reversibly modify mitochondrial protein thiols in the absence of bulk changes to protein thiols.
|
The redox state of the Trx2 pool was assessed by separating oxidized and reduced Trx2 on a urea gel followed by Western blotting (Fig. 5B) (32). Under control conditions the ratio of reduced Trx2 to partially or fully oxidized Trx2 was similar to that under RET, but the ratio became more oxidized on incubation with antimycin (Fig. 5B). The ratio of fully reduced PrxIII to the intersubunit disulfide form was measured by nonreducing electrophoresis in the presence of NEM followed by Western blotting (48) (Fig. 5C). Under control conditions and RET, PrxIII was predominantly dimerized, and during antimycin inhibition it was almost entirely dimerized (Fig. 5C). Supporting this, redox-DIGE identified PrxIII as constitutively oxidized under our control incubation conditions (Table 1). Hyperoxidation of the catalytic cysteine residues of PrxIII to sulfinic/sulfonic acids occurred on exposure to 25-250 µM exogenous H2O2 but was negligible with 2.5 µM H2O2 and during RET or antimycin inhibition (Fig. 5D). Together these data suggest that during RET the Trx2 and PrxIII pools are not markedly more oxidized than control incubations.
Identification of Mitochondrial Thiol Proteins Sensitive to ROS and RNS—Redox-DIGE enabled us to identify protein spots containing thiols that underwent a statistically significant change in fluorescent intensity on exposure to a mild, endogenous redox challenge. The next step was to identify the proteins that contained the redox-sensitive cysteine residues. To do this, gel spots were excised, and the proteins were identified by MALDI-TOF-TOF mass spectrometry. All proteins modified under the conditions listed in Fig. 4 that could be unambiguously identified are given in supplemental Tables I-V, and supplemental Fig. 1 shows some of the spots. A protein-stained gel is shown in supplemental Fig. 2 indicating the location of the proteins identified. Table 2, parts A and B, lists those proteins containing thiols changed by the lowest levels of exogenous oxidative stress, 10 µM SNAP and 2.5 µM H2O2, respectively. The mean fluorescence intensity ratios are included in Table 2, along with the fluorescence intensities of the same spots on exposure to higher concentrations of SNAP (100 µM) and H2O2 (25 µM and 250 µM) for comparison. The mean fluorescence intensity ratio is the average normalized fluorescence intensity ratio from at least three experiments and represents the extent of thiol modification within a particular protein spot. The fluorescence intensity ratio for H2O2 did not change much on a 10-fold increase in the H2O2 concentration suggesting that most of the protein thiol oxidation was complete at very low H2O2 concentrations. In contrast, a 10-fold increase in SNAP concentration led to a 3-12-fold increase in the fluorescence intensity ratio, consistent with decreased sensitivity of these thiols to SNAP. Furthermore, the proteins modified were on the mitochondrial outer membrane and intermembrane space (creatine kinase) suggesting that these modifications were by direct transfer of NO+ to thiols on proteins that were easily accessed by SNAP (18).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Application of Redox-DIGE to mitochondria indicated that there was a small subset of proteins that contained particularly redox-sensitive thiols that were modified by the low concentrations of ROS produced endogenously from the respiratory chain by RET or by antimycin inhibition, in the absence of a general modification of protein thiols. The extent and location of ROS production mimics mitochondrial ROS production during redox signaling and under these conditions enzyme activity altered through thiol modification. Therefore, endogenous mitochondrial ROS production can act as a reversible redox signal by modifying protein thiols within mitochondria. As inhibition by antimycin disrupts the membrane potential with secondary effects on mitochondria, as indicated by the altered GSH/GSSG and Trx2 ratios, superoxide formation from complex I during RET is a better indication of how mitochondrial thiols may respond to redox signaling in vivo. Furthermore, ROS production from complex I because of RET occurs when there is a high protonmotive force and a reduced respiratory chain, conditions that are likely to occur in vivo, and consequently RET is a plausible mode of increasing mitochondrial ROS production during physiological redox signaling. During RET the GSH/GSSG and the Trx2 ratios were unaffected, further indicating that RET from complex I can selectively modify mitochondrial thiol proteins without altering the thiol redox state of bulk mitochondrial thiol proteins. Of the 6 proteins modified by RET, 3 were also modified by antimycin, and 8 of the 9 proteins modified by RET or antimycin were also modified by exogenous H2O2, indicating that the effects of the increased superoxide production within mitochondria were most likely caused by H2O2. The sole exception to modification via H2O2 is complex III core protein I, which is only affected by antimycin treatment. This may be due to its proximity to the superoxide production by complex III at the Qo site.
These redox modifications during RET may occur via the direct reaction of H2O2 with the thiol to form a sulfenic acid that can then be stabilized, react with a vicinal protein thiol to form an intraprotein disulfide or with glutathione to form a glutathionylated protein. Alternatively, the decomposition of H2O2 to a hydroxyl radical or oxidation of the glutathione or Trx2 pools by H2O2 followed by thiol-disulfide exchange could contribute. The GSH/GSSG and Trx2 ratios did not change during RET suggesting they were not involved. To further investigate the thioredoxin system, we measured the redox state of the Trx2-dependent peroxidase, PrxIII, which is likely to be the main way mitochondrial H2O2 influences the Trx2 redox state. The catalytic cycle of PrxIII requires formation of a sulfenic acid on a catalytic thiol that then forms an intersubunit disulfide bond that is reduced by Trx2 (46). Measurement of the ratio of PrxIII monomer to disulfide-linked dimer (48) and redox-DIGE (Table 1), surprisingly, showed that in our control and RET incubations PrxIII was predominantly dimerized and that during antimycin incubation dimerization was essentially complete. This oxidation of PrxIII may be a consequence of incubation of mitochondria in vitro; however, in cells it is possible that many peroxiredoxins are predominantly in the dimerized form (48). As RET did not lead to an increase in the oxidation of PrxIII, it is unlikely that PrxIII directly interacted with ROS-sensitive thiol proteins to modify them by thiol-disulfide exchange during RET. In summary, ROS production from the respiratory chain leads to protein thiol modification of only a few proteins, and this modification occurs in the absence of bulk alterations to protein thiols. The ROS responsible for this modification is H2O2, which most likely acts directly on the protein thiols themselves.
Redox-DIGE enabled the modified proteins to be identified by peptide mass fingerprinting and tandem mass spectrometry. Although redox-DIGE gave reasonable coverage of mitochondrial proteins, its reliance on two-dimensional gels means that very hydrophobic membrane proteins may be under-represented. Furthermore, the strong labeling by both fluorescent dyes of FeS proteins could mask alterations to other thiols on these proteins, possibly explaining why no modifications to the complex I 75- and 51-kDa subunits were found, even though their thiol modification has been demonstrated previously (19, 36). Redox-DIGE enabled us to identify candidate proteins affected by endogenous mitochondrial ROS production. Surprisingly, six of the proteins affected by RET or antimycin are involved in mitochondrial fatty acid oxidation. These include very long chain acyl-CoA dehydrogenase and the -subunit of mitochondrial trifunctional protein, which act in concert converting long fatty acyl-CoAs to 3-ketoacyl-CoAs thereby catalyzing four steps of -oxidation (51) and short-chain 2-enoyl-CoA dehydrogenase, which carries out -oxidation of short chain acyl-CoAs. Three of the other enzymes modified are also associated with -oxidation. Carnitine acetyltransferase transfers short-chain acyl groups between carnitine and CoA and may be important in maintaining the CoA/acetyl-CoA balance (51). Propionyl-CoA carboxylase is the terminal step in the breakdown of odd numbered fatty acids and converts propionyl-CoA to succinyl-CoA for entry into the tricarboxylic acid cycle. Mitochondrial acyl-CoA thioesterase is thought to be involved in the constitutive metabolism of acyl-CoA (52), perhaps to ensure CoA availability during -oxidation or to prevent excessive acylcarnitine accumulation when -oxidation is slower than fatty acid transport. The other main enzyme modified was PDHK2 which phosphorylates and thus inactivates PDH, the entry point for pyruvate from glycolysis into the citric acid cycle and a critical regulation point for carbohydrate metabolism (35). This clustering of modified proteins in -oxidation and PDHK2 was unexpected and may suggest that endogenous mitochondrial ROS production modulates carbohydrate and fatty acid oxidation. Related to this, -oxidation is inhibited by endogenous H2O2 (53), and this inhibition can be reversed by thiol reagents. Furthermore, PDHK activity is inhibited by disulfides, and this inhibition can be reversed by thiols (35). As mitochondria oxidizing fatty acids generate more ROS than when oxidizing carbohydrates (38), thiol regulation could act as a feedback mechanism enabling ROS production during fatty acid oxidation to reciprocally modulate fatty acid and carbohydrate metabolism. Although this proposal is clearly speculative, our findings from redox-DIGE suggest that a link between mitochondrial ROS production and the balance between carbohydrate and fatty acid utilization in mitochondria is worth investigating.
In summary, using redox-DIGE, we found that endogenous ROS production from the mitochondrial respiratory chain led to the sensitive oxidation of a small subset of mitochondrial thiol proteins, in the absence of bulk thiol changes. As ROS production from the respiratory chain is an important component of redox signaling, this finding indicates how mitochondrial ROS production may interact with proteins to change their function. The mitochondrial proteins most sensitive to endogenous ROS production were clustered in -oxidation and in the regulation of PDH. Although considerably more work is required to explore the physiological significance of this unexpected finding, it suggests a possible link between mitochondrial ROS production and the modulation of mitochondrial fatty acid and carbohydrate metabolism. We conclude that the reversible modification of mitochondrial thiol proteins in response to ROS produced by the respiratory chain is a potentially important mode of redox signaling.
|
FOOTNOTES |
|---|
The on-line version of this article (available at http://www.jbc.org) contains supplemental Tables I-V and Figs. 1 and 2.
1 To whom correspondence should be addressed. Tel.: 44-1223-252900; Fax: 44-1223-252905; E-mail: mpm{at}mrc-dunn.cam.ac.uk.
2 The abbreviations used are: ROS, reactive oxygen species; BMAL, biotinylated maleimide; BSA, bovine serum albumin; DDM, n-dodecyl -D-maltoside; DIGE, difference gel electrophoresis; DTNB, 5,5'-dithiobis(2-nitrobenzoic) acid; DTT, dithiothreitol; FCCP, carbonylcyanide-p-(trifluoromethoxy)phenylhydrazone; FeS, iron-sulfur center; MTP, mitochondrial trifunctional protein; NEM, N-ethylmaleimide; PCC, propionyl-CoA carboxylase; PDHK, pyruvate dehydrogenase kinase; redox-DIGE, redox difference gel electrophoresis; RET, reverse electron transport; RNS, reactive nitrogen species; SNAP, S-nitroso-N-acetyl-DL-penicillamine; Trx2, thioredoxin 2; MALDI-TOF-TOF, matrix-assisted laser desorption ionization time-of-flight-time-of-flight.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, to further support this identification, this spot was also verified as pyruvate dehydrogenase kinase, isoenzyme 2 by immunoblotting (data not shown).