Article pubs.acs.org/jpr
Tissue Specific Phosphorylation of Mitochondrial Proteins Isolated from Rat Liver, Heart Muscle, and Skeletal Muscle Steffen Bak,†,‡ Ileana R. León,‡ Ole Nørregaard Jensen,*,‡ and Kurt Højlund*,† †
Section of Molecular Diabetes & Metabolism, Department of Endocrinology, Odense University Hospital, and Institute of Clinical Research, University of Southern Denmark, DK-5000 Odense C, Denmark ‡ Department of Biochemistry and Molecular Biology, University of Southern Denmark, DK-5230 Odense M, Denmark. S Supporting Information *
ABSTRACT: Phosphorylation of mitochondrial proteins in a variety of biological processes is increasingly being recognized and may contribute to the differences in function and energy demands observed in mitochondria from different tissues such as liver, heart, and skeletal muscle. Here, we used a combination of TiO2 phosphopeptide-enrichment, HILIC fractionation, and LC−MS/MS on isolated mitochondria to investigate the tissue-specific mitochondrial phosphoproteomes of rat liver, heart, and skeletal muscle. In total, we identified 899 phosphorylation sites in 354 different mitochondrial proteins including 479 potential novel sites. Most phosphorylation sites were detected in liver mitochondria (594), followed by heart (448) and skeletal muscle (336), and more phosphorylation sites were exclusively identified in liver mitochondria than in heart and skeletal muscle. Bioinformatics analysis pointed out enrichment for phosphoproteins involved in amino acid and fatty acid metabolism in liver mitochondria, whereas heart and skeletal muscle were enriched for phosphoproteins involved in energy metabolism, in particular, tricarboxylic acid cycle and oxidative phosphorylation. Multiple tissue-specific phosphorylation sites were identified in tissue-specific enzymes such as those encoded by HMGCS2, BDH1, PCK2, CPS1, and OTC in liver mitochondria, and CKMT2 and CPT1B in heart and skeletal muscle. Kinase prediction showed an important role for PKA and PKC in all tissues but also for proline-directed kinases in liver mitochondria. In conclusion, we provide a comprehensive map of mitochondrial phosphorylation sites, which covers approximately one-third of the mitochondrial proteome and can be targeted for the investigation of tissue-specific regulation of mitochondrial biological processes. KEYWORDS: Mitochondrial proteins, phosphoproteomes, tissue-specific regulation
■
INTRODUCTION Mitochondria are the main energy source in mammalian cells providing ATP by oxidative metabolism of substrates through the tricarboxylic acid (TCA) cycle, fatty acid oxidation and oxidative phosphorylation (OXPHOS). Mitochondria also play an important role in cellular processes such as production of reactive oxygen species, amino acid metabolism, calcium signaling, apoptosis, and more tissue-specific processes such as the urea cycle and heme synthesis.1,2 Impaired mitochondrial function is associated with several human disorders such as diabetes, cancer, cardiovascular disease, and neurodegenerative diseases and therefore has recently gained increased attention.3,4 Except for 13 proteins encoded by the mitochondrial genome, the vast majority of mitochondrial proteins are encoded by nuclear DNA and need to be actively imported from the cytosol into the mitochondria.5 A tissue-specific import of proteins into mitochondria seems to be an important mechanism to meet tissue-specific needs and makes the protein composition of mitochondria very dynamic.6,7 Liver, heart muscle, and skeletal muscle are major metabolic tissues with known differences in morphology, structure, and content of mitochondria.2,8 Previous proteomic studies of tissue-specific © XXXX American Chemical Society
differences have shown that mitochondria from these tissues are qualitatively quite similar with only a relatively small number of proteins uniquely expressed in each tissue.8−10 However, quantitative comparison of the mitochondrial proteomes from rat liver, heart muscle, and skeletal muscle have revealed important differences in protein abundance between tissues.8,11,12 Thus, mitochondria isolated from heart muscle and skeletal muscle show an increased abundance of proteins in pathways involved in ATP synthesis including TCA cycle, electron transport chain, and OXPHOS,8,11−13 whereas liver mitochondria are enhanced for specific aspects of the urea cycle,11 and proteins related to lipid and amino acid metabolism.8,12 In addition to changes in the mitochondrial protein composition and abundance, it is likely that the functional heterogeneity of mitochondria in different tissues is controlled by post-translational modifications (PTM) such as phosphorylation.4 We among many others have reported that a large number of mitochondrial proteins involved in a variety of key regulatory biological processes in mitochondria are Received: March 29, 2013
A
dx.doi.org/10.1021/pr400281r | J. Proteome Res. XXXX, XXX, XXX−XXX
Journal of Proteome Research
Article
previously.15,33,34 Briefly, the tissue samples were weighed, and pieces of 100−200 mg from each tissue sample were cut with scissors in 1 mL isolation buffer (100 mM sucrose, 100 mM KCl, 50 mM Tris-HCl, 1 mM KH2PO4, 0.1 mM EGTA, 0.2% BSA, protein phosphatase inhibitor PhosphoSTOP tablet (Roche, Basel, Switzerland), protease inhibitor tablet (Roche, Basel, Switzerland), and 2 mM sodium pervanadate, pH 7.4). The small tissue pieces were rinsed thoroughly in the isolation buffer and transferred to 1 mL of isolation buffer supplemented with 0.2 mg/mL Nagarse (Sigma, St. Louis, MO). After 2 min, each tissue sample was homogenized using a motor-driven Plexiglas homogenizer. After homogenization, 3 mL of isolation buffer was added to the homogenate and centrifuged at 700 × g for 10 min. The supernatant was decanted and centrifuged at 10 000 × g for 10 min. The resulting pellet containing the mitochondria was washed, resuspended in 650 μL isolation buffer, and centrifuged at 7000 × g for 4 min. All procedures were carried out at 0−4 °C. To avoid the dephosphorylation of sites during the isolation procedure, PhosphoSTOP and sodium pervanadate were included in all buffers to completely block the action of any phosphatases in the mitochondrial preparations until lysis with SDS. The final mitochondrial pellet was lysed in a 4% SDS, 100 mM Tris, and 0.1 M DTT, pH 8.0 buffer and stored in −80 °C. Protein concentrations were measured using a BCA reducing agent compatibility kit (Thermo Scientific Pierce, Rockford, IL).
phosphorylated in different mammalian tissues including human skeletal muscle.14−23 However, the potential tissuespecific differences in the phosphoproteomes of important metabolic tissues such as the liver, heart muscle, and skeletal muscle remains to be established. Mass spectrometry (MS) is a potent tool to identify and quantify new PTMs and thereby get a better understanding of the dynamic nature of these modifications. Recent advances in MS instrumentation is favoring better sensitivity, resolution, and higher sequencing speed, all important capabilities for studying protein PTMs of low stoichiometry. Moreover, optimized phosphorylation specific enrichment protocols have significantly increased the sizes of the phosphoproteomic maps for a large panel of different mammalian tissues even without prior fractionation of organelles.24−31 However, the large dynamic range of protein abundance in whole tissue lysates from mammalian samples still remains a challenge and suggests that isolation of mitochondria is an important step prior to studies of the mitochondrial phosphoproteomes of different tissues.32 In the present study, we aimed to investigate tissue-specific phosphorylation of mitochondrial proteins isolated from rat liver, heart muscle, and skeletal muscle. The use of littermates kept the genomic variation as low as possible. Using a TiO2 phosphopeptide enrichment protocol optimized for small tissue samples and a state-of-the-art LC−MS/MS workflow, we identified a total of 1208 unique phosphopeptides, corresponding to 889 high confident unique phosphorylation sites and 78 ambiguous phosphorylation sites in 354 mitochondrial proteins. A number of these phosphorylation sites were uniquely assigned to the liver, heart muscle, and skeletal muscle, respectively, and pathway analysis provided insight into tissue-specific differences between the mitochondrial phosphoproteomes of these major metabolic tissues.
■
Digestion of Mitochondrial Proteins
Protein was digested on spin filters.35,36 From each sample, 200 μg of protein lysate was applied to a 10 KDa spin filter (VWR collection) containing 200 μL of UA (8 M urea and 100 mM Tris/HCl, pH 8.5). The protein solution was centrifuged at 14 000 × g in 45 min, followed by two washes each with 200 μL of UA and centrifugation at 14 000 × g. The protein was alkylated with iodoacetamide (15 mM in UA) at room temperature for 20 min in dark. The filter was centrifuged at 14 000 × g for 20 min, followed by two separate washes with 100 μL of UA and 20 min centrifugation at 14 000 × g. To change buffers, two further washes were conceived with 100 μL of 50 mM ammonium bicarbonate (ABC) and 20 min centrifugation at 14 000 × g. The protein samples were digested overnight with trypsin 1/100 ratio enzyme to protein, by weight at 37 °C. Peptides were collected in new collection tubes by centrifugation at 14 000 × g for 20 min and two sequential washes/elution with 50 μL ABC followed by 14 000 × g centrifugation. Peptide samples were concentrated by lyophilization in a vacuum centrifuge.
EXPERIMENTAL PROCEDURES
Animals
Four male SPRD (Sprague−Dawley) rats from the same litter (average weight of 476 g) were euthanized 17 weeks old using an overdose of an anesthetic (isoflurane) followed by cervical dislocation. The use of rats in the present study were performed in accordance to the rules of the University of Southern Denmark. Liver, heart muscle, and skeletal muscle from the soleus muscle were dissected out and put in ice cold isolation buffer (100 mM sucrose, 100 mM KCl, 50 mM Tris-HCl, 1 mM KH2PO4, 0.1 mM EGTA, 0.2% BSA, protein phosphatase inhibitor PhosSTOP tablet (Roche, Basel, Switzerland), protease inhibitor tablet (Roche, Basel, Switzerland), and 2 mM sodium pervanadate, pH 7.4). Both heart muscle and slowtwitch muscle have a high mitochondrial content and as such are more comparable than using fast-twitch or mixed skeletal muscle. In our experiments, soleus muscle was chosen to ensure that a high amount of mitochondria could be isolated from this type of muscle and that any differences between the mitochondrial phosphoproteomes of the heart muscle and skeletal muscle were not simply due to less efficient isolation of mitochondrial proteins.
Phosphopeptide Enrichment TiO2
For phosphopeptide enrichment, a titanium dioxide (TiO2) material was weighed and put into an Eppendorf tube (0.4 mg/ 100 μg protein). Samples were dissolved in 100 μL of loading buffer (80% ACN, 5% TFA, 1 M glycolic acid) and mixed with TiO2 material in the tube for 10 min. After spinning down, the supernatant was removed. The resin was washed with 100 μL of loading buffer, 100 μL of washing buffer 1 (80% ACN, 1% TFA), and 100 μL of washing buffer 2 (20% ACN, 0.1% TFA). Then the resin was mixed with 100 μL elution buffer (NH3· H2O, pH 10.5). The supernatant containing the phosphopeptides was collected. The elution was acidified with 100% formic acid and 10% TFA before desalting.
Preparation of Mitochondria
Isolation of rat liver, heart, and skeletal muscle mitochondria was performed using a method, in which both subsarcolemmal and intermyofibrillar mitochondria are isolated and ensure functionally intact mitochondria of a high quality as described B
dx.doi.org/10.1021/pr400281r | J. Proteome Res. XXXX, XXX, XXX−XXX
Journal of Proteome Research
Article
Sample desalting
taxonomy on an in-house Mascot server (version 2.3.02, Matrix science Ltd., London, UK). Trypsin was chosen as enzyme with a maximum of 2 missed cleavages allowed. S-Carbamidomethyl cysteine was defined as a fixed modification and oxidation of methionine and phosphorylation of serine, threonine, and tyrosine as variable modifications. The MS and MS/MS results were searched with a peptide ion mass tolerance of ±8 ppm and a fragment ion mass tolerance of ±0.3 Da. Percolator40 was used for calculating FDR, and for phosphorylation, site validation the PhosphRS41 scoring was used. Only peptides that were identified as rank 1 peptides and with a confidence value of 1% (q < 0.01) were considered for further analysis. If the mascot phosphosite was not in accordance with the phosphoRS site location, the site was corrected. The phosphopeptides were grouped in high confident, phosphoRS scores above 78% and ambiguous site with below 78%. The 977 identified phosphorylation sites were site located using a probability scoring algorithm with a cut off at ≥78%, resulting in 899 high confident sites having a median probability score >0.99997%. Acknowledging that it is impossible to obtain 100% purity of mitochondria, the aim of our study was not to claim the identification of novel mitochondrial proteins but rather to identify the potential phosphorylation of proteins, which have previously been annotated to mitochondria in several public available databases. Annotation and classification of the identified proteins and phosphoproteins into mitochondrial and nonmitochondrial proteins was facilitated in Proteome Discoverer by the built in link to ProteinCenter (Thermo Scientific, Odense, DK), and all phosphopeptides from nonmitochondrial proteins were excluded from further analysis, except for the analysis of potential kinases in the mitochondrial preparations.
The samples were desalted using self-made microcolumns packed with Poros R3 reverse phase material. In a P10 tip was put a little disk of C8 material from a 3 M Empore (Sigma, St. Louis, MO) C8 extraction disk to plug the hole. The microcolumn was packed onto this plug with a slurry consisting of R3 material in 100% ACN. Before using the columns they were washed in buffer B (70% ACN, 0.1% TFA) and reequilibrated in buffer A (0.1% TFA). Every acidified sample was desalted on such columns by loading twice, washing twice in 20 μL buffer A, and eluting very slowly with 20 μL buffer B. Eluates were lyophilized in a vacuum centrifuge and stored at −20 °C until the next step. HILIC Peptide Fractionation
The enriched phosphopeptide samples were prefractionated by hydrophilic interaction liquid chromatography (HILIC). The separation was performed on a 1200 Series HPLC system (Agilent, Santa Clara, CA) using an in-house packed, 15 cm long, 2 mm diameter TSKGel Amide-80 HILIC column (3 μm particle size, Tosoh Bioscience, South San Francisco, CA).37,38 The column was equilibrated in buffer A (90% ACN, 0.1% TFA). The mobile phase gradient to separate the peptides was as follows: 0−10% buffer B (0.1% TFA) in 0.5 min, and 10− 40% buffer B in 26 min followed by 40−100% buffer B in 4 min. Loading was carried out with a 12 μL/min flow for 8.5 min and gradient separation at 6 μL/min. A total of 36 fractions were collected during loading and separation. The peptide fractions were dried by lyophilization and redissolved and pooled according to UV intensity chromatograms in 0.1% formic acid just before MS analysis. Data Acquisition
Nano LC−MS/MS analyses were performed using an EasyLC system (Thermo Scientific, Odense, DK) interfaced to a LTQOrbitrap Velos hybrid mass spectrometer (Thermo Scientific, Bremen, Germany). Samples were dissolved in Solvent A (0.1% FA) and loaded onto a custom-made 2 cm trap column (100 μm i.d., 375 μm o.d., packed with Reprosil C18, 5 μm reversedphase particles (Dr. Maisch GmbH, Ammerbuch-Entringen, Germany)) connected to a 15 cm analytical column (75 μm i.d., 375 μm o.d., packed with Reprosil C18, 3 μm reversedphase particles (Dr. Maisch GmbH, Ammerbuch-Entringen, Germany)) with a pulled emitter. Separation was performed at a flow rate of 250 nL/min using a 50 or 100 min gradient of 0− 34% Solvent B (90% ACN, 0.1% FA) into the nanoelectrospray ion source (Thermo Scientific, Odense, DK). The LTQOrbitrap Velos instrument was operated in a data-dependent MS/MS mode using Multistage Activation (MSA).39 The peptide masses were measured by the Orbitrap (MS scans were obtained with a resolution of 30 000 at m/z 400), and up to 15 of the most intense peptide m/z were selected and subjected to fragmentation using MSA in the linear ion trap (LTQ). Dynamic exclusion was enabled with a list size of 500 masses, duration of 40 s, and an exclusion mass width of ±10 ppm relative to masses on the list.
Mitochondrial Purity
To evaluate the purity of the mitochondrial preparations, we set up a label-free MS quantification experiment for unbiased evaluation of the purity of each mitochondrial preparations and degree of contaminating proteins. Twenty micrograms of each mitochondrial preparation was digested following the same procedure as before. Peptide digests were R2 cleaned up and an amount equal to 1 μg was analyzed by nano-LC−MS/MS on an LTQ Orbitrap Velos (ThermoFisher, Bremen, Germany) operated in a DDA mode and selecting up to the top 10 most intense precursor ions for fragmentation by CID from a MS spectrum with a resolution of 30 000 at m/z 400. The raw files were processed in Proteome Discover; identified proteins were annotated, and their relative amounts in terms of extracted ion chromatograms (XICs) were obtained by the Precursor Ions Area Detector function. Mitochondrial purities were calculated by taking the total area of the XIC of all annotated mitochondrial proteins divided by the total area XIC of all identified proteins (Supplemental Table 1). Data Presentation
Details about all the phosphopeptides and phosphorylation sites assigned to mitochondrial proteins are given in Supplemental Table 2. Functional annotation was established using the tools in DAVID Bioinformatics Resources 6.7.42,43 DAVID was used to assign all proteins to GO biological processes and KEGG pathways44,45 to identify enriched biological processes and pathways as compared to the rat international protein index (IPI). P-values for enrichment were calculated by the EASE method used in DAVID. A P-value
Data Processing and Database Search
The LC−MS/MS data was processed (smoothing, background subtraction, and centroiding) using Proteome Discoverer (Version 1.3, Thermo Scientific, Rockford, IL). The processed LC−MS/MS data was submitted to database searching against the SwissProt database (version 2011_10, containing 25719 sequences after taxonomy) with Rodentia (Rodents) as C
dx.doi.org/10.1021/pr400281r | J. Proteome Res. XXXX, XXX, XXX−XXX
Journal of Proteome Research
Article
Figure 1. Overview of results. Experimental workflow used in the study (a). The number of nonredundant phosphopeptides, phosphoproteins, and phosphorylations sites identified in isolated mitochondria from rat liver and heart and skeletal muscle in each experiment (b). Pie charts of the distribution of serine, threonine, and tyrosine phosphorylation sites identified in mitochondria from each tissue (c). Number of the identified phosphorylation sites reported or not reported (new sites) in three large phosphorylation databases (Uniprot, PhosphoSitePlus, and the CPR PTM Resource produced by Lundby et al.31) (d). Venn diagrams showing the overlap between biological replicates and the phosphorylation sites identified in total and in n ≥ 2 rats in each tissue (e).
mitochondria from the different tissues showed to be essentially the same, with 95.3% for liver, 93.7% for skeletal muscle, and 91.0% for heart muscle. Using the workflow shown in Figure 1A, this study provides an extensive map of phosphorylation sites in mitochondrial proteins covering most biological processes known to take place in mitochondria. In total, we identified 1208 phophopeptides corresponding to 899 nonredundant phosphorylation sites (78 ambiguous) in 354 mitochondrial proteins with a high confidence in each tissue in 4 rats (Figure 1b). Representative spectra of two peptides from HMGCS2 with phosphorylations are presented in Figure 2. The highest number of mitochondrial phosphorylation sites was found in the liver samples with 594 nonredundant sites (50 ambiguous), followed by heart muscle with 448 sites (33 ambiguous), and skeletal muscle with 336 sites (12 ambiguous) (Supplemental Table 2). No differences were seen in the distribution of Ser, Thr, and Tyr phosphorylation sites in mitochondrial proteins between the three tissues (Ser 72−75%), (Thr 19−20%), and (Tyr 6−8%) (Figure 1c). Among the 899 mitochondrial phosphorylation sites identified, less than 420 sites were previously reported in two large phosphorylation databases, www.uniprot.org and www.phosphosite.org, or a recent large-scale phosphoproteo-