Steatosis-Induced Proteomic Changes in Liver Mitochondria Evidenced by Two-Dimensional Differential In-Gel Electrophoresis Pierre Douette,† Rachel Navet,† Pascal Gerkens,‡ Edwin de Pauw,‡ Pierre Leprince,§ Claudine Sluse-Goffart,† and Francis E. Sluse*,† Laboratory of Bioenergetics and Laboratory of Mass Spectrometry, Baˆt. B6c, Alle´e de la Chimie 3, 4000, and Centre de Recherche en Neurobiologie Cellulaire et Mole´culaire, Baˆt. L1, place Delcour 17, 4020, Lie`ge, Belgium. Received June 22, 2005
Steatosis encompasses the accumulation of droplets of fats into hepatocytes. In this work, we performed a comparative analysis of mitochondrial protein patterns found in wild-type and steatosis-affected liver using the novel technique two-dimensional differential in-gel electrophoresis (2D-DIGE). A total of 56 proteins exhibiting significant difference in their abundances were unambiguously identified. Interestingly, major proteins that regulate generation and consumption of the acetyl-CoA pool were dramatically changed during steatosis. Many proteins involved in the response to oxidative stress were also affected. Keywords: steatosis • liver mitochondria • proteomics • two-dimensional differential in-gel electrophoresis • ob/ob mouse
The liver is an essential organ in supporting many processes involved in energy metabolism. It regulates blood glucose levels within an appropriate range by controlling glycogenolysis and gluconeogenesis. The liver is also involved in fat storage, and is the major site of fatty acid synthesis from excess dietary carbohydrates, amino acids or ketone bodies (namely acetoacetate and β-hydroxybutarate), a process that is stimulated by high insulin level. During exercise or starvation, when the blood glucose level is low, lipid stores in adipose tissue are mobilized in the form of fatty acids. Then liver oxidizes fatty acids into acetyl-CoA through β-oxidation. Ketone bodies released from subsequent acetyl-CoA conversion can be used by other tissues as energy source. Therefore, the liver plays a major role in maintaining the balance of metabolic energy in response to endrocrine signals. Long-term imbalance between food intake and energy expenditure can lead to diet-induced obesity. This pathology is commonly associated with other diseases such as type 2 diabetes mellitus, hypertension, coronary heart disease, or some cancers.1,2 Moreover, a high fat diet leads to steatosis, i.e., a fatty liver resulting from the accumulation of droplets of fats. Steatosis results from insulin resistance that increases hepatic fatty acid synthesis.3 Association of steatosis with several liver lesions leads to nonalcoholic steatohepatitis (NASH). A link between obesity and genetic inheritance lies in the identification of leptin as a central hormonal mediator regulating energy homeostasis.4 Leptin, the appetite-suppressing hormone, is secreted by adipocytes and mediates satiety.5,6 This * To whom correspondence should be addressed. Fax. +32-4-366-2878. E-mail:
[email protected], † Laboratory of Bioenergetics. ‡ Laboratory of Mass Spectrometry. § Centre de Recherche en Neurobiologie Cellulaire et Mole´culaire.
2024
Journal of Proteome Research 2005, 4, 2024-2031
Published on Web 10/14/2005
16-kDa protein is the product of the ob gene7 and acts through specific receptors related to the class 1 cytokine receptor superfamily8 highly expressed in many tissues. Fifty-five years ago, a genetic defect in the ob gene was identified in inbred hyperphagic obese mice.9 The ob/ob mice produce an inactive form of leptin resulting in obesity and several metabolic and endocrine abnormalities such as hyperglycemia, hyperinsulinemia, and hypercorticosteronemia.10 The ob/ob model organism has been extensively used to investigate physiological and pathophysiological states,11,12 including steatosis, since these mice develop fatty livers.13,14 It is quite clear that steatosis is inextricably related to modifications of the mitochondrial function.3 Moreover, emergence of proteomics tools provides a strong technological background to study proteomes under various conditions. Therefore, we used two-dimensional differential in-gel electrophoresis (2D-DIGE) to gain better understanding on the fatty liver mitochondrial adaptation. We provide an integrated view of the metabolic adaptation occurring during steatosis and evidence changes in proteins involved in the regulation of the acetyl-CoA pool and in protective proteins in response to the steatosis-induced oxidative stress.
Material and Methods Materials. Cyanine dyes (CyDyes: Cy2, Cy3 and Cy5) and immobilized IPG strips were purchased from Amersham Biosciences. The sequencing-grade modified trypsin was from Roche. All other chemicals were of the highest purity grade and were purchased from Sigma. Animals. Five 12 week-old male ob/ob C57BL/6 mice and five 12 week-old lean C57BL/6 mice were purchased from Charles River Laboratories. All animals were housed in a temperature- and light-controlled breeding farm with standard 10.1021/pr050187z CCC: $30.25
2005 American Chemical Society
Steatosis-Induced Changes in Liver Mitochondria
pellet chow and water ad libitum. Livers were collected immediately after sacrifice and cooled in ice-cold isolation medium. All animal experiments were conducted in accordance with the Belgian Ethical Committee guidelines. Isolation of Liver Mitochondria. Briefly, 2 g of liver were cleaned and homogenized in 10 mL of isolation medium (210 mM mannitol, 70 mM saccharose, 0.5 mM EDTA, 20 mM TrisCl pH 7.4). Cell debris was removed by centrifugation at 600 × g for 10 min at 4 °C, and mitochondria were finally sedimented from the supernatant by centrifugation at 15 500 × g for 15 min at 4 °C. Isolated mitochondria were washed twice in isolation medium containing 0.5% BSA, resuspended at 10 mg/ mL, applied onto a discontinuous Percoll gradient (30-70%) and centrifuged at 150 000 × g for 1 h at 4 °C.15 Purified mitochondria were washed twice to remove Percoll. The Percoll step was performed twice. A total of 9.8 and 15.9 mg of mitochondrial proteins were recovered in the crude extract from 2 g of wild-type (WT) and ob livers, respectively, whereas 3.7 and 5.6 mg were finally obtained following Percoll purification of WT and ob crude extracts, respectively. With regard to the low recovery yield, mitochondrial samples from each of the five mice were pooled for subsequent analysis. The final mitochondrial pellet was suspended into the lysis buffer (7 M urea, 2 M thiourea, 2% (w/v) ASB-14 (zwitterionic detergent)) at 10 mg of proteins per ml and vortexed for 10 min. Insoluble material was removed by centrifugation at 20 800 × g for 15 min at 4 °C. A total of 400 µg of protein was precipitated using a 2D Cleanup kit (Amersham Bioscience) and the pH was adjusted to 8.5 with 100 mM NaOH after reconstitution in a minimal volume of lysis buffer supplemented with 30 mM TrisHCl, pH 8.5. Protein concentration was evaluated with the RC/ DC Protein Assay (Bio-Rad Laboratories). Labeling of Mitochondrial Proteins with CyDyes. Mitochondrial protein samples (25 µg) were set to a final protein concentration of 5 mg/mL with lysis buffer and labeled separately with 0.2 nmol of CyDye (Cy3, Cy5) (Amersham Bioscience), vortexed, and incubated 30 min in the dark. A mix sample composed of equal amount of mitochondrial proteins from wild-type and obese animals was also labeled with Cy2. After 30 min, the labeling was stopped with 10 mM lysine. To control the labeling efficiency, labeled proteins (0.5 µg) were subjected to SDS-PAGE analysis and the gels were scanned with the Typhoon 9400 scanner (Amersham Biosciences) at the wavelengths corresponding to each CyDye, namely 480 nm (Cy2), 532 nm (Cy3) and 633 nm (Cy5). Two-Dimensional In-Gel Electrophoresis. A 25-µg portion of each Cy3-, Cy5- and Cy2-labeled sample was combined, and the sample volume was set to 450 µL with standard rehydration buffer (7 M urea, 2 M thiourea, 2% w/v ASB 14, 25 mM DTT, and 0.6% v/v pH 3-10 NL IPG buffer). Three mixed samples were subjected to isoelectric focusing (24-cm IPG strips, pH 3-10 NL) on an IPGphor isoelectric focusing unit (Amersham Bioscience). Strips were rehydrated and focused under paraffin oil at 200 V for 200 Vh, 500 V for 500 Vh, 1 kV for 1 kVh, and 8 kV for 60 kVh at 20 °C and a maximum current setting of 50 µA per strip. Prior to SDS-PAGE, each strip was equilibrated according to Go¨rg et al.16 Strips were loaded and run on 12% acrylamide gels overnight at 20 °C in an Ettan Dalt II system (Amersham Biosciences) at 1 W per gel. For protein identification and peptide sequencing, a preparative gel including 300 µg of mixed sample (150 µg each) was performed in parallel. Gel Scanning and Image Analysis. Gels were scanned directly between low-fluorescence glass plates with the Ty-
research articles phoon 9400 scanner at the three wavelengths specific of the CyDyes. The resolution was of 100 µm. Determination of protein spot abundance was performed using the DeCyder software (Amersham Biosciences). In brief, the three CyDye-labeled forms of each spot were co-detected within each gel. Ratios between samples and internal standard abundances were calculated for each protein spot with the DIA (Differential In-gel Analysis) module. Inter-gel variability was corrected by matching and normalization of the internal standard spot maps by the BVA (Biological Variance Analysis) module of the DeCyder software. Protein spots that showed a statistically significant (p < 0.01) Student’s t-test for an increased or decreased expression level were accepted as being differentially expressed between the extracts under comparison. CyDye, DeCyder, Ettan are trademarks of Amersham Biosciences Ltd. Protein Identification and Mass Spectrometry. The preparative gel was stained using a silver-stain MS compatible protocol.17,18 Proteins spot were manually excised from silverstained gel. Following protein reduction (135 mM DTT) and alkylation (55 mM iodoacetamide), protein digestion was performed overnight at 37 °C with modified trypsin (Roche) in 50 mM NH4HCO3. The peptides were extracted twice in acetonitrile +0.1% formic acid, lyophilised and resuspended in 0.1% formic acid. Peptide separation by reversed-phase liquid chromatography was performed on an Ultimate LC system (LC Packings, Dionex) supplemented with Famos auto-sampler and Switchos II microcolumn switching device for sample cleanup and preconcentration.. Sample (30 µL) was loaded at a flow rate of 200 nL/min on a micro-precolumn cartridge (300 µm × 5 mm, packed with C18PepMap, 5 µm, 100Å). The precolumn was connected with the separating nano-column (75 µm × 15 cm, packed with C18 PepMap100, 3 µm, 100Å). Elution gradient varied from 0 to 30% (buffer A, 0.1% formic acid in 2% acetonitrile to buffer B, 0.1% formic acid in 20% acetonitrile) over 30 min. The outlet of the LC system was directly connected to the nano-electrospay source of an Esquire HCT ion trap mass spectrometer (Bruker Daltonics, Germany). Mass data acquisition was performed in the mass range of 50 to 1700 m/z using the Standard-Enhanced mode (8100 m/z per sec). For each mass scan, a data-dependent scheme picked the four most intense doubly or triply charged ions to be selectively isolated and fragmented in the trap. The resulting fragments were mass analyzed using the Ultra Scan mode (50-3000 m/z at 26 000 m/z per sec). Raw data were analyzed and formatted (Data Analysis software, Bruker) for subsequent protein identification against the NCBI nonredundant protein database through MS/ MS ions search algorithm on the Mascot search engine (www.matrixscience.com). Single hit protein identification with a mascot total score (Mowse Score) greater than 50 was accepted as positive identification. Protein subcellular location was performed using PSORT II (psort.nibb.ac/form2.html) and TargetP (www.cbs.dtu.dk/ services/TargetP). 2D-HPLC Proteomic Analysis. Purified mitochondrial proteins (100 µg) were dissolved in 100 µL of 50 mM NH4HCO3. The protein sample was then reduced by addition of 5 µL of 200 mM DTT in 100 mM NH4HCO3 followed by boiling during 10 min. The sample was alkylated for 90 min in the dark by adding 4 µL of 100 mM iodoacetamide in 100 mM NH4HCO3. The remaining iodoacetamide was then neutralized by adding 20 µL of 200 mM DTT in 100 mM NH4HCO3 for 45 min. Journal of Proteome Research • Vol. 4, No. 6, 2005 2025
research articles
Douette et al.
Digestion was performed overnight at 37 °C using 2 µg of modified trypsin (Roche) (trypsin:sample ratio 1:50). Digestion was stopped with 5 µL of 500 mM HCl during 45 min at 37 °C (pH ) 2). Finally the digested sample was centrifuged at 16 000 × g for 15 min and aliquoted. A 30-µL portion of the peptide mixture was loaded into the 2D-HPLC system (LC Packings, Dionex) combining a first SCX column (LC Packings, µ-PrecolumnTM, Bio-X-SCX (500 µm × 15 mm)) followed by a reversed-phase liquid chromatography composed of a micro-precolumn cartridge (300 µm × 5 mm, packed with C18PepMap, 5 µm, 100 Å) and a separating nanocolumn (75 µm × 15 cm, packed with C18 PepMap100, 3 µm, 100 Å). Peptides were eluted from the SCX column by different steps of salt concentration (20, 30, 40, 60, 80, 100, 200, 300, 400, and 800 mM ammonium acetate) and separated on the reversed-phase system as described for in-gel digestion and protein identification.
Results and Discussion Comparative Analysis of Mitochondrial Proteomes by 2DDIGE and Protein Identification. Two-dimensional gel electrophoresis suffers from severe shortcomings, among which gelto-gel variation in spot pattern is the most important. Twodimensional differential in-gel electrophoresis (2D-DIGE) allows us to circumvent the variability that often shades biological differences or misleads quantitative comparison of protein expression levels. This recent multiplexing-based approach takes advantage of the use of different protein samples independently labeled with three spectrally resolved fluorescent dyes (Cy2, Cy3, and Cy5) prior to isoelectrofocusing. The two protein samples of interest are labeled with Cy3 or Cy5, whereas equal amounts of both samples are mixed and labeled with Cy2. The latter pooled sample generates the internal standard used for accurate quantitative comparison sustained by statistical tests.19 Incorporating a pooled internal standard means that every protein represented within all of the labeled samples can be compared to itself in the internal standard, allowing the determination of a ratio of relative expression. Since the same internal standard is run among all gels, it can be normalized across the gels, which dramatically decreases the gel-to-gel variation and enables to detect < 10% changes in protein expression with > 95% statistical confidence.20 Differentially labeled samples are pooled and resolved on the same 2D-gels.21 Three independent images of protein patterns are acquired by successively scanning the gel with fluorophore-specific excitation wavelengths. Image analysis comprises intra-gel matching and normalization of protein patterns of interest on the basis of the internal standard, and detection of differences in protein expression levels. Mitochondria were prepared from 5 lean and 5 ob/ob C57BL/6 mice. After Percoll-purification, the qualitative purity of the samples was assessed by checking the identity of a large number of proteins contained in the purified mitochondrial fractions using tryptic in-solution digestion, peptides separation by 2D-HPLC and peptide sequencing by MS/MS analysis (Figure 1). 2D-HPLC allows the identification of a large number of proteins present in the Percoll-purified mitochondrial fractions isolated from WT and fatty livers. A total of 93 and 103 were identified by MS/MS sequencing from WT and steatosisaffected mitochondrial fractions, respectively. Eighty percent of the identified proteins were from mitochondrial origin, whereas 8-9% were peroxisomal proteins. It is noteworthy that this 2D-HPLC characterization only represents a qualitative 2026
Journal of Proteome Research • Vol. 4, No. 6, 2005
Figure 1. Subcellular identity of the proteins contained in the Percoll-purified mitochondrial fractions isolated from WT and fatty livers. Proteins were digested by trypsin in solution and the peptide mixture was separated by 2D-HPLC prior to MS sequencing. 93 and 103 proteins were identified from WT and ob mitochondrial fractions, respectively.
evaluation of the purity and is not representative of the quantitative enrichment corresponding to a weight-to-weight ratio of mitochondrial proteins content compared to total proteins content. WT and steatosis-affected mitochondria were differentially labeled using Cy3 and Cy5, mixed and co-separated on broad range pH 3-10NL (nonlinear) 2D-gels. To account for experimental bias, 2D-gels were run in triplicate to validate spot changes. Figure 2 shows the characteristic 2D-gel containing proteins of wild-type and steatosis-affected mitochondria. According to statistical tests between 2D-gels (Student’s t test values of 50) (Table 1), including 40 nonredundant proteins. According to prediction of subcellular location using PSORT II and TargetP, two proteins were assigned to the cytoplasm whereas peroxisomal proteins represented the majority of contaminants, i.e., 7 proteins including catalase even though the peroxisomal/mitochondrial localization of catalase remains unclear.22 Despite Percoll gradient for mitochondrial fractionation described by Mootha
Steatosis-Induced Changes in Liver Mitochondria
Figure 2. Mitochondrial proteins pattern by 2D-DIGE. Liver mitochondria were isolated and mitochondrial proteins were extracted by using lysis buffer (7 M urea, 2 M thiourea, 2% ASB14). Wild-type and fatty liver mitochondria were differentially labeled with Cy3 and Cy5. An internal standard composed of equal amount of each mitochondrial sample and labeled with Cy2 was added to improve comparative analysis. Labeled samples (25 µg of each Cy2, Cy3, and Cy5) were loaded on 24cm 3-10 NL IPG-strips and subjected to isoelectrofocusing. Second-dimension was performed in 12% acrylamide gels. Gels were then scanned in a wavelength-selective way and subsequent image analyses were performed withDecyder (DIA and BVA softwares, Amersham Biosciences). The proteins that were found to vary significantly (p < 0.01, Student’s t test in three independent gels) in fatty liver mitochondria compared to wildtype are marked with the number allocated by the DeCyder software.
and co-workers,15 detection of peroxisomal proteins indicated that peroxisomes were co-purified with mitochondria. Steatosis-Induced Peroxisomal Changes in Fatty Acids Metabolism. In peroxisome, very long-chain fatty acids are chain-shortened. The peroxisomal membrane contains a nonspecific pore-forming protein that enables straight- and branched-chain fatty acids to be transported by passive diffusion. Interestingly, we found that the nonspecific lipid transfer protein (sterol carrier-2) was 1.5-fold increased in response to steatosis. The SCP-2 gene contains two initiation sites coding for one 58-kDa protein and one 13-kDa protein.23 Both proteins participate in uptake, oxidation and esterification of straightand branched-chain fatty acids. The first step of oxidation is catalyzed by acyl-CoA oxidase that generates hydrogen peroxide. Even though we did not find change in the latter enzyme, the 1.8-fold increase in catalase, which detoxifies hydrogen peroxide produced in the peroxisomes during this step, suggested an increase in activity of acyl-CoA oxidase. Moreover, the 2-hydroxyphytanoyl-CoA lyase that catalyzes R-oxidation of 3-methyl branched fatty acids (phytanic acid) was twice more abundant. Peroxisomal bifunctional enzyme was also 1.5-fold more abundant in fatty liver mitochondria. Peroxisomal bifunctional enzyme exhibits both enoyl CoA hydratase and 3-hydroxyacyl-CoA dehydrogenase activities and generates palmitoyl-CoA that further enters into the mitochondria for subsequent processing through mitochondrial β-oxidation. Steatosis-Induced Mitochondrial Changes in Energy Metabolism. Brady and co-workers have reported higher fatty
research articles acids oxidative capacities in ob/ob mice mitochondria.24 Higher fatty acids oxidation has been linked to increase in FFA content, carnithine palmitoyl transferase-1 (CPT-1) expression and regulation and proliferation of hepatic peroxisomes.3 However, apart from CPT-1, little information is available about other enzymes involved in fatty acids oxidation in steatosis.3 CPT-1 transports activated palmitoyl-CoA inside the mitochondria that is further consumes during β-oxidation cycles. This process involves three FAD-dependent acyl-CoA dehydrogenases and 3-hydroxyacyl-CoA dehydrogenase. Acyl-CoA dehydrogenases (medium chain (C16 to C4) and short chain) were increased ∼1.5-fold (Figure 3A) in mitochondria from obese mice. Moreover, the 3-hydroxyacyl-CoA dehydrogenase that catalyzes the NAD+-dependent formation of 3-ketoacyl-CoA from L-3hydroxyacyl-CoA was also 1.5-fold up-regulated. Concerning unsaturated fatty acids, 2,4-dienoyl-CoA isomerase was expressed in greater amount in fatty liver mitochondria. This enzyme is requires for β-oxidation of unsaturated fatty acids since they have double-bonds in the cis-configuration, whereas β-oxidation produces trans-intermediates. Therefore, in steatosis, mitochondrial fatty acids oxidation is increased not only due to CPT-1 increase, but also in response to an overall increase in fatty acids oxidative enzymes. The branched-chain amino acids (BCAA) catabolism is another process generating NADH and FADH2 reducing equivalents as well as acetyl-CoA. This pathway is initiated by the branched-chain R-ketoacid dehydrogenase that resembles pyruvate dehydrogenase. The catalytic domain responsible for BCAA catabolism, i.e., the branched chain ketoacid dehydrogenase E1, was depressed in steatosis (Figure 3B), whereas the inactivating kinase component, i.e., 3-methyl-2-oxobutanoate dehydrogenase kinase, component was up-regulated. This suggests a shutdown of BCAA catabolism in response to high fatty acids oxidative capacities. Excess of acetyl-CoA generated by high fatty acids oxidation activity leads to activation of pyruvate carboxylase and inhibition of pyruvate dehydrogenase, a situation that stimulates gluconeogenesis.25 In addition to higher activity due to high acetyl-coA, we found that pyruvate carboxylase was 2-fold greater in fatty liver mitochondria (Figure 3C). Pyruvate carboxylase catalyzes the most important anaplerotic reaction, i.e., the carboxylation of pyruvate to oxaloacetate (OAA). Mitochondrial production of OAA is the initial step of gluconeogenesis even though further conversion steps into glucose occur in the cytoplasm. For gluconeogenesis to proceed, OAA produced by pyruvate carboxylase has to be transported across the mitochondrial membrane. Since no mechanism exists for direct transfer of OAA, it is converted either in malate or in aspartate, all of which being transported. Here, the 4.7-fold increase in aspartate transminase (Figure 3D) strongly suggests that transamination of OAA to aspartate is preferentially used to export the gluconeogenic substrate. This transamination reaction requires transport of glutamate into the mitochondria in exchange for R-ketoglutarate. Regeneration of substrates for maintaining continuous exchange through the aspartate shuttle requires glutamate dehydrogenase that was also found to increase. Both up-regulations of pyruvate carboxylase and aspartate transaminase explain the aberrant gluconeogenesis occurring in steatosis. We found no evidence of changes in pyruvate dehydrogenase as well as in rate-limiting enzymes of the TCA cycle (citrate synthase and isocitrate dehydrogenase). Accordingly, Brady and co-workers reported no change in pyruvate oxidation by the Journal of Proteome Research • Vol. 4, No. 6, 2005 2027
research articles
Douette et al.
Table 1. List of Proteins Exhibiting Different Abundance in Fatty versus WT Liver Mitochondriaa AC
name
Q9DBM2 P32020 Q8BLD7 P24270 P24270 P34914 Q9QXE0 Q8R128 Q921G7 Q5U5Y5 P45952 P45952 Q07417 Q61425 O35459 Q99KI0 P97807 Q9Z2I9 Q91VD9 Q91VD9 Q91WD5 Q05920 Q05920 Q05920 Q05920 P26443 P05202 Q8CBX6 P13707 Q9DBT9 Q91VA0 P38060 P38060 P54869 Q80VU5 Q80VU5 O55028 Q99L69 P12007 O88844 P20108 Q9CZS1 Q9JLJ2 Q8BWF0 Q8BWF0 Q8BU08 Q8BU08 Q8C196 Q8C196 Q8C196 Q8C196 Q8C196 Q8C196 P29758 P10126
Peroxisomal bifunctional enzyme sterol carrier protein x similar to 3-ketoacyl-CoA Thiolase B Catalase
L
Peroxisomal Proteins P P P P/M
t-test
V
0.000022 1.48 0.0059 1.46 0.00095 1.32 0.00006 1.7 0.00002 1.88 Soluble epoxide hydrolase P/C 0.00026 1.58 2-hydroxyphytanoyl-CoA lyase P 0.00039 1.91 Alanine-glyoxylate aminotransferase P/M 0.0008 -1.42 Fatty acids oxidation Electron-transfer flavoprotein-ubiquinone oxidoreductase M 0.00014 1.5 Trifunctional enzyme alpha subunit 0.0077 1.33 Acyl-coenzyme A dehydrogenase, medium chain M 0.00024 1.51 0.00092 1.63 Acyl-Coenzyme A dehydrogenase, short chain M 0.0000076 1.43 3-hydroxyacyl-CoA dehydrogenase M 0.0012 1.52 Delta3,5-delta2,4-dienoyl-CoA isomerase M 0.0018 1.65 TCA cycle Aconitate hydratase, mitochondrial M 0.00009 1.54 Fumarate hydratase M 0.0033 1.35 Succinyl-CoA ligase [ADP-forming] M 0.0001 1.54 Respiratory Chain NADH-ubiquinone oxidoreductase 75 kDa subunit M 0.0013 1.44 M 0.000049 1.45 NADH-ubiquinone oxidoreductase 49 kDa subunit M 0.0000049 1.35 Gluconeogenesis Pyruvate carboxylase M 0.00013 2.26 0.0004 2.03 0.0002 2.18 0.0024 1.4 Glutamate dehydrogenase M 0.0022 1.38 Aspartate aminotransferase M 0.00025 4.69 glycerol phosphate dehydrogenase M 0.0011 1.79 Glycerol-3-phosphate dehydrogenase [NAD+] C 0.0003 1.54 Lipids, cholesterols, choline, sterols, steroids biosynthesis Dimethylglycine dehydrogenase, mitochondrial M 0.00029 1.31 Medium-chain acyl-CoA synthetase M 0.0045 -1.41 Ketogenesis Hydroxymethylglutaryl-CoA (HMG-CoA) lyase M 0.000069 1.84 0.005 1.42 Hydroxymethylglutaryl-CoA (HMG-CoA) synthase M 0.00023 1.45 Branched-Chain Amino Acids Metabolism Propionyl-Coenzyme A carboxylase, alpha polypeptide M 0.000034 1.77 0.005 1.64 3-methyl-2-oxobutanoate dehydrogenase [lipoamide] kinase M 0.000013 1.35 Branched chain ketoacid dehydrogenase E1 M 0.00056 -1.64 Isovaleryl-CoA dehydrogenase M 0.0014 -1.36 Oxidative Stress and Detoxification isocitrate dehydrogenase 1 C 0.00046 1.7 Thioredoxin-dependent peroxide reductase M 0.0055 1.44 aldehyde dehydrogenase 1 family, member B1 M 0.0033 1.33 4-trimethylaminobutyraldehyde dehydrogenase 9A C 0.00096 2.06 Succinate semialdehyde dehydrogenase M 0.00068 1.82 M 0.0001 2.04 nudix (nucleoside diphosphate linked moiety X)-type motif 7 M 0.0013 -2.23 0.0016 -2.21 Urea Cycle weakly similar to carbamoyl-phosphate synthase M 0.000084 -1.97 0.0015 -1.9 0.0019 -2.02 0.0018 -2.13 0.00053 -1.93 0.001 -1.75 ornithine-oxo-acid transaminase precursor M 0.00036 1.3 Protein Fate Elongation factor TU M 0.00077 1.37
MW
pI
NP
C
MS
78633 59804 44481 59882 59882 63045 64588 46282
9.26 7.15 8.82 7.72 7.72 5.85 5.89 8.37
14 21 12 43 66 26 18 13
4% 143 887 9% 296 1351 11% 196 1531 33% 304 1113 35% 1099 1119 13% 582 1065 6% 215 1028 9% 250 1627
68877 83302 46921 46921 45146 27371 36437
7.33 9.24 8.56 8.56 8.68 9.1 7.6
20 36 13 55 26 6 26
12% 17% 15% 15% 21% 5% 10%
246 472 261 301 465 164 152
N
986 987 1491 1508 1618 2182 1941
86151 8.08 54564 9.12 40429 5.15
14 18% 13 6% 26 21%
588 694 213 1439 500 1534
80752 5.51 80752 5.51 52991 6.52
13 9% 5 3% 30 16%
223 735 56 740 422 1403
6.25 107 16% 6.25 13 7% 6.25 54 12% 6.25 4 3% 8.31 27 18% 9.05 15 16% 6.17 13 7% 6.83 5 5%
978 391 331 398 662 405 74 884 594 1261 391 1593 328 877 132 1847
130344 130344 130344 130344 61582 48123 81345 38045
97422 7.69 65517 6.75
68 16% 15 9%
778 262
34673 8.7 34673 8.7 54265 8.02
17 20% 16 20% 79 19%
364 1872 423 1875 628 1390
80498 80498 43651 50612 46709
6.83 6.83 6.27 8.15 8.67
6 22 8 18 60
8% 18% 13% 10% 16%
150 828 564 829 216 1690 324 1421 397 1533
47914 28337 58087 54449 56503 56503 24816 24816
6.49 7.15 6.59 6.63 8.53 8.53 5.97 5.97
19 13 17 20 23 2 7 26
12% 12% 10% 15% 17% 4% 10% 14%
356 187 332 441 393 120 114 252
83966 83966 83966 83966 83966 83966 48723
6.27 6.27 6.27 6.27 6.27 6.27 6.19
7 49 68 15 36 73 32
3% 245 237 10% 841 274 5% 364 275 5% 1193 280 10% 705 287 12% 938 289 27% 626 1417
49876 7.23
22 20%
639 988
1453 2274 1229 1304 1346 1324 2086 2100
401 1423
a Proteins that significantly varied in fatty versus wild-type liver (p < 0.01, Student’s t test) are organized according to their general function. Legend abbreviations: Ac, access name according to Swiss-Prot; L, subcellular localization; t-test, value of the Student’s t test; V, amplitude of variation where a positive value means that the amount of protein is increased in fatty liver mitochondria; MW, molecular weight; pI, isoelectric point; NP, number of sequenced peptides by LC-MS/MS; C, sequence coverage; MS, Mowse Score; N, spot number allocated by the DeCyder software.
2028
Journal of Proteome Research • Vol. 4, No. 6, 2005
Steatosis-Induced Changes in Liver Mitochondria
research articles
Figure 3. Comparative analysis of eight protein spot intensities using the BVA module of the DeCyder software. The selected spots are displayed as partial view of the 2D-gel (top panel) and as three-dimensional images (bottom panel). Spot boundary of selected proteins is displayed in pink.
TCA cycle.23 Ketogenesis from acetyl-CoA occurs in liver mitochondria and is increased in steatosis in response to higher fatty acids oxidation.26 Production of ketone bodies (3-hydroxybutyrate, acetoacetate, and acetone) happens more likely to consume the excessive pool of acetyl-CoA that does not enter the TCA cycle. Accordingly, hydroxymethylglutaryl-CoA (HMGCoA) synthase and lyase were found in greater amount in fatty liver mitochondria although we did not detect higher amount of D-3-hydroxybutyrate dehydrogenase. Since D-3-hydroxybutyrate dehydrogenase is located in the inner mitochondrial membrane,27 the lack of observed variation could be due to the fact that membrane proteins are poorly resolved on 2Dgels. Indeed 2D-gel analysis of membrane proteins is frequently
hampered by the fact that hydrophobic proteins display a tendency to aggregate at basic pH during isoelectric focusing or are not properly eluted from isoelectric focusing strip during transfer from the first to the second dimension.28,29 Anyway, an increase in these enzymes as well as unchanged pyruvate dehydrogenase, citrate synthase, and isocitrate dehydrogenase explain the shunting of acetyl-CoA conversion into ketone bodies instead of citrate. NADH and FADH2 generated during fatty acids oxidation are reoxidized by the mitochondrial respiratory chain activity whereas acetyl-CoA enters the TCA cycle. As a response to the high reducing power (NADH, FADH2) produced by oxidizing fatty acids, the electron-transfer flavoprotein-ubiquinone oxiJournal of Proteome Research • Vol. 4, No. 6, 2005 2029
research articles doreductase and the respiratory chain complex I were ∼1.5fold up-regulated in steatosis-affected mitochondria (Figure 3E, F). Interestingly, the amplitude of the increase fits well with the overall change in enzymes involved in fatty acids oxidation. No variation in any subunit of other respiratory complexes or of ATP synthase was additionally found even after careful analysis of multiple hits protein identifications. Interestingly, mitochondrial uncoupling protein 2 (UCP2) mRNA was found to be increased 5- to 6-fold in these ob/ob mice.30 UCPs catalyze a protonophoretic cycle activated by free fatty acids and uncouple the mitochondrial respiration from oxidative phosphorylation. Since they compete with ATP synthase for the proton electrochemical gradient, a decrease in ATP synthesis yield is a direct consequence of their activity.31 In the steatotic metabolic context, i.e. higher fatty acids oxidation capacities leading to higher NADH/FADH2 production with no change in ATP synthase, modulation of UCP2 content could provide a way to control decrease reducing state of the respiratory chain and accelerates electrons uptake and NAD+/FAD regeneration required for both β-oxidation and TCA cycle. Mitochondrial Oxidative Stress in Steatosis. Fatty liver mitochondria overproduce oxygen free radicals.32 In the present study, we observed up-regulation of many enzymes involved in oxidative stress protection. Although we found no evidence of increase in mitochondrial superoxide dismutase, thioredoxin-dependent peroxide reductase that uses the mitochondrial thioredoxin as a source of electrons to scavenge hydrogen peroxide was 50% more abundant in mitochondria from obese mice (Figure 3G). Moreover, a possible mitochondrial localization for catalase that increased 1.8-fold in fatty liver has been suggested.22 In the presence of oxidizable fats, reactive oxygen radicals can trigger lipid peroxidation leading to the production of reactive aldehydes including 4-hydroxynonenal (HNE) that can further induce mitochondrial dysfunction.25 Whereas increase in reduced glutathione (GSH) has been reported in response to steatosis,32 NADP+-isocitrate dehydrogenase that is essential in maintaining and regenerating the reduced state of GSH33 was increased 1.7-fold. It is noteworthy that the presence of this cytosolic protein in the liver mitochondrial fraction has been previously reported.34 We also observed upregulation of three aldehyde dehydrogenases including succinate semialdehyde dehydrogenase (Figure 3H). Higher levels of succinate semialdehyde dehydrogenase could participate in detoxification of HNE,35 whereas an increase in other hepatic aldehyde dehydrogenases is able to also oxidize HNE and other toxic oxidative stress-derived aldehyde.36 Mitochondrial aconitase, which is irreversibly affected by free radicals through the Fenton chemistry,37 is more likely increased in response to the accumulation of inactive enzyme.
Conclusions In this work, we provide an in-depth analysis of metabolic adaptation occurring during steatosis in liver mitochondria. A total of 119 proteins were significantly altered in steatosis whereas we identified 56 of them. We observed large changes in all key-enzymes of the fatty acids oxidation pathway including proteins that consume the reducing equivalents generated by FA oxidation. Complex I was increased in response to high reducing power suggesting a higher electron input into the respiratory chain. Since no change in any subunit of ATP synthase was found, we hypothesize that UCP2 might accelerate electron flux in order to maintain high FA oxidation rate 2030
Journal of Proteome Research • Vol. 4, No. 6, 2005
Douette et al.
Figure 4. Biochemical pathways involved in acetyl-CoA generation and consumption showing different abundance in fatty liver mitochondria.
through NAD+/FAD regeneration. According to our proteomic data, we can propose a clear scheme of pathways involved in the generation and consumption of acetyl-CoA pool at the protein level (Figure 4). We also show that aspartate transaminase in conjunction with pyruvate carboxylase is the preferential way used leading to gluconeogenesis. Finally, we observe large change in proteins involved in the response to the oxidative stress. Among them, aldehyde dehydrogenases more likely act to decrease lipid peroxidation whereas the thioredoxin pathway is overexpressed in order to limit hydrogen peroxide generation. As observed with liver mitochondrial adaptation to chronic alcohol exposure reported by Venkatraman and coworkers,38 steatosis induces significant changes in liver mitochondrial physiology, thus reflecting the mitoproteome plasticity in response to various physiological and pathophysiological states. Abbreviations: 2D-DIGE, two-dimensional differential ingel electrophoresis; BCAA, branched-chain amino acids; CPT1, carnithine palmitoyl transferase-1; FA, free fatty; GSH, reduced glutathione; HNE, 4-hydroxynonenal; OAA, oxaloacetate; TCA, tricarboxylic acid; UCP, uncoupling protein.
Acknowledgment. This work was supported by grants from the Fonds National de la Recherche Scientifique (FRFC 2.4532.03, FRSM 9.4573.04) and from the Fonds Spe´ciaux de Recherche dans les universite´s. It was also co-financed by the Centre d’Analyse des Re´sidus en Traces (CART), the Re´gion Wallone (i-Maldi 114713), Fonds Social Europe´en (FSE) and the Centre of Biomedical Integrative Genoproteomics (CBIG). P.D. is recipient of a Fonds pour la Recherche Industrielle et Agronomique fellowship. P.L. is a Research Associate of the FNRS. References (1) Ravussin, A.; Bouchard, C. Human genomics and obesity: finding appropriate drug targets. Eur. J. Pharmacol. 2000, 410, 131-145. (2) Kopelman, P. G. Obesity as a medical problem. Nature 2000, 404, 635-643. (3) Pessayre, D.; Fromenty, B. NASH: a mitochondrial disease. J. Hepathol. 2005, 42, 928-940. (4) Friedman, J. M.; Halaas, J. L. Leptin and the regulation of body weight in mammals. Nature 1998, 395, 763-770. (5) Janeckova, R. The role of leptin in human physiology and pathophysiology. Physiol. Rev. 2001, 50, 443-459.
research articles
Steatosis-Induced Changes in Liver Mitochondria (6) Munzberg, H.; Bjornholm, M.; Bates, S. H.; Myers, M. G., Jr. Leptin receptor action and mechanisms of leptin resistance. Cell Mol. Life Sci. 2005, 62, 642-652. (7) Zhang, Y.; Proenca, R.; Maffei, M.; Baronne, M.; Leopold, L.; Friedman, J. M. Positional cloning of the mouse obese gene and its human homologue. Nature 1994, 372, 425-432. (8) Tartaglia, L. A.; Dembski, M.; Weng, X.; Deng, N.; Culpepper, J.; Devos, R.; Richards, G. J.; Campfield, L. A.; Clark, F. T.; Deeds, J. Identification and expression cloning of a leptin receptor OBR. Cell 1995, 83, 1263-1271. (9) Ingalls, A. M.; Dickie, M. M.; Snell, G. D. Obese, new mutation in house mouse. J. Hered. 1950, 41, 317-318. (10) Lee, G. H.; Proenca, R.; Montez, J. M.; Carroll, K. M.; Darvishzadeh, J. G.; Lee, J. I.; Friedman, J. M. Abnormal splicing of the leptin receptor in diabetic mice. Nature 1996, 379, 632-635. (11) Wolf, G. Insulin resistance associated with leptin deficiency in mice: a possible model for noninsulin-dependent diabetes mellitus. Nutr. Rev. 2001, 59, 177-179. (12) Edvardsson, U.; Alexandersson, M.; Brockenhuus von Lowenhielm, H.; Nystrom, A.-C.; Ljung, B.; Nilsson, F.; Dahllof, B. A proteome analysis of livers from obese (ob/ob) mice treated with the peroxisome proliferators WY14, 643. Electrophoresis 1999, 20, 935-942. (13) Hookman, P.; Barkin, J. S. Current biochemical studies of nonalcoholic fatty liver disease and nonalcoholic steatohepatitis suggest a new therapeutic approach. Am. J. Gastroenterol. 2003, 98, 2093-2097. (14) Koteish, A.; Diehl, A. M. Animal models of steatohepatitis. Best Pract. Res. Clin. Gastroenterol. 2001, 16, 679-690. (15) Mootha, V. K.; Lepage, P.; Miller, K.; Bunkenborg, J.; Reich, M.; Hjerild, M.; Delmonte, T.; Villeneuve, A.; Sladek, R.; Xu, F.; Mitchell, G. A.; Morin, C.; Mann, M.; Hudson, T. J.; Robinson, B.; Rioux, J. D.; Lander, E. S. Identification of a gene causing human cytochrome c oxidase deficiency by integrative genomics. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 605-610. (16) Go¨rg, A.; Boguth, G.; Obermaier, C.; Posch, A.; Weiss, W. Twodimensional polyacrylamide gel electrophoresis with immobilized pH gradients in the first dimension (IPG-Dalt): the state of the art and the controversy of vertical versus horizontal systems. Electrophoresis 1995, 16, 1079-1086. (17) Shevchenko, A.; Wilm, M.; Vorm, O.; Mann, M. Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal. Chem. 1996, 68, 850-858. (18) Gharahdaghi, F.; Weinberg, C. R.; Meagher, D. A.; Imai, B. S.; Mische, S. M. Mass spectrometric identification of proteins from silver-stained polyacrylamide gel: a method for the removal of silver ions to enhance sensitivity. Electrophoresis 1999, 20, 601605. (19) Alban, A.; David, S. O.; Bjorkesten, L.; Andersson, C.; Sloge, E.; Lewis, S.; Currie, I. A novel experimental design for comparative two-dimensional gel analysis: Two-dimensional difference gel electrophoresis incorporating a pooled internal standard. Proteomics 2003, 3, 36-44. (20) Knowles, M. R.; Cervino, S.; Skynner, H. A.; Hunt, S. P.; de Felipe, C.; Salim, K.; Meneses-Lorente, G.; McAllister, G.; Guest, P. C. Multiplex proteomic analysis by two-dimensional differential ingel electrophoresis. Proteomics 2003, 3, 1162-1171. (21) Unlu, M.; Morgan, M. E.; Minden, J. S. Difference gel electrophoresis: a single gel method for detecting changes in protein extracts. Electrophoresis 1997, 18, 2071-2077.
(22) Jiang, X.-S.; Dai, J.; Sheng, Q.-H.; Zhang, L.; Xia, Q.-C.; Wu, J.-R.; Zeng R. A Comparative Proteomic Strategy for Subcellular Proteome Research. Mol. Cell. Proteomics 2005, 4, 12-34. (23) Gallegos, A. M.; Atshaves, B. P.; Storey, S. M.; Starodub, O.; Petrescu, A. D.; Huang, H.; McIntosh, A. L.; Martin, G. G.; Chao, H.; Kier, A. B.; Schroeder, F. Gene structure, intracellular localization and functional roles of sterol carrier protein-2. Prog. Lipid Res. 2001, 40, 498-563. (24) Brady, L. J.; Brady, P. S.; Romsos, R. D.; Hoppel, C. L. Elevated hepatic mitochondrial and peroxisomal oxidative capacities in fed and starved adult obese (ob/ob) mice. Biochem. J. 1985, 231, 439-444. (25) Sugden, M. C.; Bulmer, K.; Holness, M. J. Fuel-sensing mechanisms integrating lipid and carbohydrate utilization. Biochem. Soc. Trans. 2001, 29, 272-278. (26) Fromenty, D.; Robin, M. A.; Igougjil, A.; Mansouri, A.; Pessayre, D. The ins and outs of mitochondrial dysfunction in NASH. Diabetes Metab. 2004, 30, 121-138. (27) McIntyre, J. O.; Bock, H. G.; Fleischer, S. The orientation of D-betahydroxybutyrate dehydrogenase in the mitochondrial inner membrane. Biochim. Biophys. Acta 1978, 513, 255-267. (28) Gorg, A.; Weiss, W.; Dunn, M. J. Current two-dimensional electrophoresis technology for proteomics. Proteomics 2004, 4, 3665-3685. (29) Bailey, S. M.; Landar, A.; Darley-Usmar, V. Mitochondrial proteomics in free radical research. Free Radic. Biol. Med. 2005, 38, 175-188. (30) Chavin, K. D.; Yang, SQ.; Lin, HZ.; Chatham, J.; Chacko, V. P.; Hoek, J. B.; Walajtys-Rode, E.; Rashid, A.; Chen, C.-H.; Huang, C.-C.; Wu, T.-C.; Lane, M. D.; Diehl, A. M. Obesity induces expression of uncoupling protein-2 in hepatocytes and promotes liver ATP depletion. J. Biol. Chem. 1999, 274, 5692-5700. (31) Vidal-Puig, A. J. Uncoupling expectations. Nat. Genet. 2000, 25, 387-388. (32) Yang, S. Q.; Zhu, H.; Li, Y.; Lin, H. Z.; Gabrielson, K.; Trush, M. A.; Diehl, A. M. Mitochondrial adaptation to obesity-related oxidant stress. Arch. Biochem. Biophys. 2000, 378, 259-268. (33) Lee, S. M.; Koh, H. J.; Park, D. C.; Song, B. J.; Huh, T. L.; Park, J. W. Cytosolic NADP(+)-dependent isocitrate dehydrogenase status modulates oxidative damage to cells. Free Radic. Biol. Med. 2002, 32, 1185-1196. (34) Fountoulakis, M.; Berndt, P.; Langen, H.; Suter, L. The rat liver mitochondrial proteome. Electrophoresis 2002, 23, 311-328. (35) Murphy, T. C.; Amarnath, V.; Gibson, K. M.; Picklo, M. J., Sr. Oxidation of 4-hydroxy-2-nonenal by succinic semialdehyde dehydrogenase (ALDH5A). J. Neurochem. 2003, 86, 298-305. (36) Vasiliou, V.; Pappa, A. Polymorphisms of human aldehyde dehydrogenases. Pharmacology 2000, 61, 192-198. (37) Verniquet, F.; Gaillard, J.; Neuburger, M.; Douce, R. Rapid inactivation of plant aconitase by hydrogen peroxide. Biochem. J. 1991, 276, 643-648. (38) Venkatraman, A.; Landar, A.; Davis, A. J.; Chamlee, L.; Sanderson, T.; Kim, H.; Page, G.; Pompilius, M.; Ballinger, S.; Darley-Usmar, V.; Bailey, S. M. Modification of the mitochondrial proteome in response to stress of ethanol-dependent hepatotoxicity. J. Biol. Chem. 2004, 279, 22092-22101.
PR050187Z
Journal of Proteome Research • Vol. 4, No. 6, 2005 2031
