Proteomics Unveils Fibroblast−Cardiomyocyte Lactate Shuttle and Hexokinase Paradox in Mouse Muscles † † ‡ Dariusz Rakus,*, Agnieszka Gizak, and Jacek R

Home , BRENDA

This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Article

pubs.acs.org/jpr

Proteomics Unveils Fibroblast−Cardiomyocyte Lactate Shuttle and Hexokinase Paradox in Mouse Muscles † † ‡ Dariusz Rakus,*, Agnieszka Gizak, and Jacek R. Wisniewskí *, † Department of Animal Molecular Physiology, Wroclaw University, Wroclaw 50-205, Poland ‡ Biochemical Proteomics Group, Department of Proteomics and Signal Transduction, Max-Planck-Institute of Biochemistry, Martinsried 82152, Germany

*S Supporting Information

ABSTRACT: Quantitative mapping, given in biochemically interpretable units such as mol per mg of total protein, of tissue-specific proteomes is prerequisite for the analysis of any process in cells. We applied label- and standard-free proteomics to characterize three types of striated muscles: white, red, and cardiac muscle. The analysis presented here uncovers several unexpected and novel features of striated muscles. In addition to differences in protein expression levels, the three muscle types substantially differ in their patterns of basic metabolic pathways and isoforms of regulatory proteins. Importantly, some of the conclusions drawn on the basis of our results, such as the potential existence of a “fibroblast−cardiomyocyte lactate shuttle” and the “hexokinase paradox” point to the necessity of reinterpretation of some basic aspects of striated muscle metabolism. The data presented here constitute a powerful database and a resource for future studies of muscle physiology and for the design of pharmaceutics for the treatment of muscular disorders.

KEYWORDS: striated muscle, heart muscle, hexokinase, glycolysis, proteomic ruler, total protein approach

■ INTRODUCTION allowed for studying the molecular basis of muscle dys- 4−6 fi Striated muscles constitute approximately 40% of the whole trophy. A few recent studies have demonstrated that lter- aided sample preparation (FASP)7 and multienzyme digestion body mass, and being the largest organ in the mammalian 8 organism, they consume the majority of energetic compounds FASP (MED-FASP) methods facilitate proteomic analysis of circulating in body fluids. From this, it is not unexpected that muscles, enabling the creation of a quantitative picture of the disturbances and defects in muscle energy metabolism affect the cellular organization and providing titers of individual 9−11 energetic homeostasis and physiology of the whole organism. proteins Recently, we have described a simple analytical One of the best known muscle metabolism-related diseases is and computational approach to estimate titers of enzymes of diabetes type II, which is caused by improper insulin signal basic metabolic pathways and proteins of the contractile transduction in skeletal muscles and which results in multiple machinery in skeletal muscles of the mouse.10 Here, we have organ failure. extended our previous investigation providing an absolute Striated muscles are composed of fibers with distinct quantitative picture of the red (soleus), white (white part of physiological properties. On the basis of biochemical studies gastrocnemius), and cardiac muscle proteomes. We focused on as well as histochemical and physiological analyses, striated the characterization of proteins with regulatory functions fi muscles are broadly classi ed into the red (oxidative, type I, governing the major metabolic mechanisms. Comparison of also called the slow muscle), white (glycolytic, type II, the fast the three proteomes led us to several unexpected findings. muscle), and cardiac muscle. These include the “hexokinase paradox”, a phenomenon of the Although skeletal and cardiac muscle metabolism has been lack of correlation between the capacity of a given muscle type intensively studied over the last century, the quantitative to take up glucose and phosphorylate it to glucose-6-phosphate characterization of enzymes constituting energy metabolism and the concentrations of glycolytic enzymes. On the basis of pathways and proteins involved in regulation of these pathways the results of proteomic and histochemical analyses, we also is far from being complete. Moreover, comparison of the data, often dispersed across thousands of publications, is difficult hypothesize that, in the heart, an unexpectedly high expression of glucose transporters and hexokinase may be explained by the because of the fact that they have been acquired by a variety of fi methods. Thus, system wide insights into muscle cell protein tight coupling of cardiac broblast and cardiac myocyte compositions and physiology has been missing. metabolism in a process called the lactate shuttle. Proteomic technologies have already provided initial insights in the composition of skeletal1,2 and heart3 muscles to a depth Received: December 22, 2015 of 1,500−3,500 proteins. Label-free proteomic approaches Published: June 15, 2016

Figure 1. Proteomic analysis of mouse soleus, gastrocnemius, and heart muscles. (A) Muscle tissue was solubilized in 2% SDS, and the extracted proteins were processed by the MED-FASP procedure. Peptides were analyzed on LTQ-Orbitrap instruments. Spectra were searched and analyzed using MaxQuant software. The protein intensities were processed by the “total protein approach” and the “proteomic ruler”. (B) Number of identified and quantified proteins. (C) Overlaps of statistically significant differences in protein concentration titers observed between the three sample groups. (D) Principal component analysis of the proteomic data. (E) Determination of virtual protein content per nucleus of the muscle cells using the “proteomic ruler” method. ■ MATERIAL AND METHODS used for sequential digestion of proteins. The enzyme to protein ratios were 1/50. Peptide concentrations were assayed Isolation and Lysis of Heart Muscle, Gastrocnemius, and 12 ∼ μ Soleus Muscles using the WF assay. Aliquots containing 8 g of peptide were separated on a reverse phase column (20 cm × 75 μm Ten-week-old female C57BL/6 mice were bred on a 12:12-h inner diameter) packed with 1.8 μm C18 particles (Dr. Maisch light-dark cycle and had free access to standard chow diet. The GmbH, Ammerbuch-Entringen, Germany) using a 3 h fi soleus and the super cial, white part of gastrocnemius muscles acetonitrile gradient in 0.1% formic acid at a flow rate of 250 were dissected from sacrificed mice. The muscle samples were fi nL/min. The LC was coupled to a QExactive HF mass taken from ve animals and were analyzed separately. Muscle spectrometer (Thermo Fisher Scientific, Germany) via a samples were homogenized with an Ultra Turrax T8 ff nanoelectrospray source (Proxeon Biosystems). The QExactive homogenizer (IKA Labotechnik) in ice-cold bu er: 20 mM HF was operated in data dependent mode with survey scans of Tris-HCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 1 mM 300−1650 m/z acquired at a resolution of 60,000. Up to the PMSF, pH 7.4, at 4 °C. For the proteomic analysis, the lysates fi top 15 most abundant isotope patterns with charge m/z 2 from were supplemented with SDS and DTT to nal concentrations the survey scan were selected with an isolation window of 1.4 of 2% and 0.05 M, respectively. The protein lysates were boiled Th and fragmented by HCD with normalized collision energies at 100 °C for 5 min. Total protein was determined by fl “ ” 12 of 25. The maximum ion injection times for the survey scan and measuring tryptophan uorescence in a WF assay . the MS/MS scans were 20 and 60 ms, respectively. The ion Proteomic Analysis target values for MS1 and MS2 scan modes were set to 3 × 106 Protein lysates and extracts containing 100 μg of total protein and 1 × 105, respectively. The dynamic exclusion was 30 s. The were processed in the 30k filtration units (Cat No. mass spectrometry data have been deposited to the MRCF0R030, Millipore)13 centrifuged at 10,000g using the ProteomeXchange Consortium14 via the PRIDE partner MED-FASP protocol.8 Endoproteinase Lys-C and trypsin were repository with the data set identifier PXD002152.

2480 DOI: 10.1021/acs.jproteome.5b01149 J. Proteome Res. 2016, 15, 2479−2490 Journal of Proteome Research Article

Figure 2. Glycolysis and glyconeogenesis in striated muscles. (A) Scheme of glucose metabolism. (B) Abundance of glycolysis, glycogen metabolism, and pentose phosphate pathways measured as a percentage of total protein. (C) Lactate-pyruvate conversion and plasma membrane lactate transporters. (D) Regulation of glyconeogenesis. (E) Glycogen metabolism. (F) Glucose transport and phosphorylation. Statistical significance of differences is indicated by stars. For comparison of single proteins, the significance refers to FDR < 0.05 and for protein groups P < 0.01.

Proteomic Data Analysis where MSsignal and total MSsignal refer to the total MS1 signal The MS data was analyzed using the software environment intensity of protein i and the total protein MS1 signal, MaxQuant version 1.2.6.20. Proteins were identified by respectively. searching MS and MS/MS data against the UniProtKB/ The total protein content of mitochondria was obtained by Swiss-Prot database (May 2013) containing 50807 sequences. summing the total protein content of proteins matching Gene The FDR was derived by analyzing the decoy database. Ontology categories according to the relationship Carbamidomethylation of cysteine was set as a fixed modification. The initial allowed mass deviation of the {mitochondrion}∉∪ [{nucleus} {endoplasmic reticulum} precursor ion was up to 6 ppm, and for the fragment masses, ∪∪{Golgi body} {plasma membrane}] it was up to 20 ppm. The maximum false peptide discovery rate fi was speci ed as 0.01. Protein concentrations were calculated on The selection of mitochondrial proteins is shown in Table the basis of spectral protein intensity using the total protein S4. approach (TPA)15 using the equation Statistical Analysis protein concentration(i ) Results are presented as mean ± SEM unless otherwise stated. ⎡ ⎤ We used nonpaired Student’s t-test for comparisons between MSsignal (i ) mol = ⎢ ⎥ two experimental groups. The minimal number of values per total MS× MW(i )⎣ g total protein ⎦ signal group was 4. The missing values were imputed using the

2481 DOI: 10.1021/acs.jproteome.5b01149 J. Proteome Res. 2016, 15, 2479−2490 Journal of Proteome Research Article parameter values of 0.3 for width and 1.8 for down shift. P < five peptides (Table S1), and all proteins discussed in this paper 0.005 was considered statistically significant. belong to this subset. Immunohistochemical Studies The Fate of Glucose Molecules: Cardiac but not Skeletal All primary antibodies used in the experiment were purchased Muscle Glycolysis might be Hormonally Stimulated on the from Abcam. First, 3 μm wax sections of formalin-fixed mouse Level of Pfk-Catalyzed Reaction heart were dewaxed and immersed in 0.3% Sudan Black in 70% Glucose, after its uptake and esterification to glucose-6- ethanol for 30 min at room temperature to reduce phosphate (G6P), might be used for the synthesis of glycogen autofluorescence of the tissue sections. This incubation was or oxidized in glycolysis and/or the pentose phosphate pathway followed by 3 washes of 5 min each in PBS with 0.02% Tween (Figure 2A). Glycolysis is a basic and, presumably, one of the 20. Tissue sections were then blocked in 3% BSA in PBS and evolutionarily oldest metabolic pathways. In this pathway, a incubated with rabbit anti-hexokinase 1 primary antibody glucose molecule is oxidized in the absence of oxygen, (1:50) or anti-Slc16a3 (Mct4) primary antibody (1:20). The providing ATP and the reducing force NADH. In contrast to antibodies were then detected with antirabbit secondary glycolysis, PPP is an anabolic pathway that produces NADPH antibody conjugated to Alexa Fluor 405 (Thermo Fisher and pentoses, 5-carbon sugars that may serve as a substrate for fi fi Scienti c; pseudocolored in magenta). For cardiac broblast nucleic acid synthesis. counterstaining, the sections were incubated with murine The quantitative mapping of these pathways resulting from FITC-labeled antivimentin antibody (1:20). To counterstain our analysis confirmed the well-established pattern of basic fi lamentous actin in cardiomyocytes, the sections were glucose metabolism in which white muscle has the highest incubated with phalloidin-FITC. In the control reactions, capacity both to degrade glucose in glycolysis and to store it as primary antibodies against Hk1 and/or Mct4 were omitted. glycogen (Figure 2B). On the other hand, white muscle fl The tissue sections were embedded in uorescent mounting contains the lowest concentration of enzymes constituting PPP. medium and examined in a FluoView 1000 confocal micro- The final product of glycolysis−pyruvate might be oxidized scope (Olympus). in mitochondria and/or reduced by lactate dehydrogenase A18 and released outside the cell via the Slc16a3 (Mct4) lactate RESULTS AND DISCUSSION ■ transporter (Figure 2C). Conversion of pyruvate to lactate is Proteomic Analysis characteristic for tissues with high glycolytic rate and low The proteomic procedure involved tissue lysis in SDS and the accessibility of oxygen, e.g., for white muscles. Our data are in MED-FASP procedure for protein digestion (Figure 1A). The excellent agreement with this metabolic pattern: white muscle analysis of five biological replicates of soleus (red fibers), the has the highest and heart muscle (the most oxygenated striated white part of gastrocnemius, and heart muscles allowed for the muscle) has the lowest capacity to produce and release lactate. identification of 7,207 proteins (Figure 1B, Table S1). The On the other hand, the titer of enzymes involved in uptake of highest number of 5,993 proteins per sample were identified in lactate and its oxidation to pyruvate, Slc16a1 (Mct1) and Ldhb, respectively, were the highest in the heart (Figure 2C), which the heart, whereas the lowest (5,260) were found in fl gastrocnemius. There were 6,275 proteins identified in at re ects the phenomenon of lactate consumption by this organ least in four biological replicates of one tissue type, which were (for review, see ref 19). used for statistical analysis (Table S2). Missing values were It is a common belief that the regulation of glycolysis by Pfk imputed. Multivariate analysis revealed 3,500, 1,881, and 1,129 does not depend on the total amount/maximal activity of the significant protein titer changes between heart/gastrocnemius, enzyme but relies on the allosteric activation of Pfk by fructose- heart/soleus, and gastrocnemius/soleus pairs at FDR < 0.005, 2,6-bisphosphate (F2,6P2) (for review, see ref 20). This respectively (Figure 1C, Table S2). Titers of 486 proteins were metabolite is synthesized by 6-phosphofructo-2-kinase/fruc- significantly changed between all three tissues (Figure 1C). tose-2,6-bisphosphatase (Pfkfb), a hormonally regulated bifunc- Principal component analysis showed a clear separation of all tional enzyme. Under physiological conditions and in the groups (Figure 1D). A proteomic ruler allowed for calculation absence/low concentration of F2,6P2, Pfk is saturated with of the absolute protein content and the volume per nucleus ATP, an inhibitor, and one of the Pfk substrates, which results 20 (Figure 1E, Table S3). We found a similar total protein content in almost complete inactivation of the enzyme. An increase in of approximately 1.5 ng per cell nucleus for soleus and F2,6P2 titer activates Pfk and accelerates the glycolytic rate. gastrocnemius muscle (Figure 1E). These values are in a good Our data show that all three types of striated muscle express agreement with our previously published data.10 For heart significant amounts of the main Pfkfb isoform, Pfkfb1 (Figure muscle, we obtained 0.5 ng per nucleus. Assuming an average 2D). In skeletal muscles, hormone-dependent phoshorylation total protein concentration in eukaryotic cells of 20% (m/v),16 inhibits kinase and activates phosphatase activity of Pfkfb1, the estimated volume of skeletal muscle fiber is approximately leading to a drop in the F2,6P2 concentration. This suggests 10 pL per nucleus, whereas the heart muscle cells have a that, in skeletal muscle, activity of Pfk is subjected to hormone- volume of approximately 2.5 nL. regulated inhibition rather than activation. In white muscle, this The accuracy of the protein quantitation in proteomics inhibition is additionally strengthened by a significantly higher increases with the number of peptides used for either relative or concentration of Tigar protein (Figure 2D), which is known to absolute protein quantitation. In the “total protein approach”, hydrolyze F2,6P2 to fructose-6-phosphate. In other words, the accuracy of the protein concentration also depends on the skeletal muscle glycolysis appears not to be subjected to depth of proteomic analysis. In large data sets, such as the hormone-mediated activation on the level of the reaction muscle proteomes here, titers of protein identified with more catalyzed by Pfk. than five peptides can on average deviate less than 1.5-fold from In contrast to skeletal muscles, cardiac muscle contains a high titers obtained by targeted proteomics using standards.17 Our titer of the Pfkfb2 isoform (Figure 2D), whose hormone- data set contains nearly 4,000 proteins matched by more than (epinephrine, insulin) and/or hypoxia-induced phoshorylation

2482 DOI: 10.1021/acs.jproteome.5b01149 J. Proteome Res. 2016, 15, 2479−2490 Journal of Proteome Research Article

Figure 3. Signaling pathways regulating glucose transporter Glut4 (Slc2a4) translocation in striated muscles. Pathways involved in the cell’s response to insulin (the left part of the figure) and ischemia- and/or contractile activity-regulated pathways (the right part). activates kinase activity and F2,6P2 production. This means glycogen stores, most studies demonstrated that the sharp that heart muscle glycolysis may be significantly accelerated increase in glycolytic rate during intense muscle contraction both by stress-induced hormonal stimulation (e.g., mediated by strongly correlates with the rate of glycogen degradation.21 In epinephrine) and by hypoxic conditions that block oxidative line with this, in white muscle, we observed much higher metabolism. concentrations of proteins involved in glycogen hydrolysis (Figure 2E) than those engaged in glucose uptake and Glyconeogenesis Plays a Crucial Role in the Replenishment of Glycogen Stores in White Muscle phosphorylation (Slc2a transporters and Hk, respectively) (Figure 2F). Hormonal (epinephrine) stimulation and muscle contraction We found that the importance of white muscle glycogen induce glycolysis in white muscles,increasing the concentration turnover for glycolytic flux was also reflected in the level of of glucose-6-phosphate. Although G6P may be produced from proteins regulating storage and degradation of this glucose two sources, glucose taken up from the bloodstream and from polymer (Figure 2E). The titer of kinases that may accelerate

2483 DOI: 10.1021/acs.jproteome.5b01149 J. Proteome Res. 2016, 15, 2479−2490 Journal of Proteome Research Article glycogen breakdown (phosphorylase b kinase subunits: Phkb relatively low contraction dependence of glucose transport in and Phkg1) was the highest just in the white muscle whereas this muscle. This is in line with studies showing that an the concentration of glycogen synthase kinase-3 beta (Gsk3β), increased rate of heart muscle contraction (e.g., during intense an enzyme that inhibits glycogen synthesis, was the lowest in exercise) correlated with elevated lactate consumption but not this muscle (Figure 2E). with glucose uptake (for review, see ref 19). Glycogen is a common polymer in most animal cells, and it is On the other hand, the highest concentration of thioredoxin usually synthesized from glucose coming from the bloodstream. 1 (Txn) (Figure 3), a protein that may activate Prkas by In some tissues, however, it may be produced from other reducing oxidative stress-formed disulfide bonds in Prka compounds such as lactate, some amino acids, and glycerol in isoforms,25 was found only in heart muscle. Such a redox- the process called glyconeogenesis. Glycogen synthesis from associated mode of glucose uptake stimulation may be critical lactate (and the release of glucosyl units into the bloodstream) for cardiac protection against ischemia and is in agreement with is typical for liver and kidney. However, several newer reports the findings of Shao et al.25 showing that heart muscle may have demonstrated that glyconeogenesis is a common process accelerate glucose uptake in metabolic stress conditions. in many other cells, e.g., in striated muscle.22,23 Although the titer of Txn1 is the highest in heart muscle, a In the present study, we found that, in white muscle, the titer significant concentration of this protein in other types of tested of regulatory glyconeogenic enzyme Fbp2 (Figure 2D) is muscles (Figure 3) points to this mechanism as an important significantly higher than titers of glucose transporters and adaptation to low oxygen pressure in all striated muscles. similar to that of Hks (Figure 2F). The concentration of Pcx, a Hexokinase Paradox: the Concentrations of Glycolytic second regulatory enzyme of glyconeogenesis, is similar to titers Enzymes are Suited for Glycogen Degradation of the transporters (Figure 2D). Taking into account the high Glucose-6-phosphate, the product of hexokinase reactions, may affinity of glyconeogenic enzymes to their substrates and that be isomerized to fructose-6-phosphate in a glycolytic reaction their kcat values are not lower than those for hexokinases catalyzed by Gpi, converted to glucose-1-phosphate by Pgm1, a (BRENDA database, http://www.brenda-enzymes.org/), it glycogen metabolism enzyme, or oxidized in PPP by G6pdx might be hypothesized that glyconeogenesis is the main (Figure 4C). Thus, it is reasonable to assume that the cellular process supporting glycogen synthesis in white muscle. This fi capacity for glucose uptake and phosphorylation should corroborates our previous ndings showing that the inhibition correlate with the amount of these three enzymes. The titer of glyconeogenic complex formation in Danio rerio and rat of G6pdx, as compared to those of Gpi and Pgm1, in all studied muscles strongly decreased the level of glycogen even in the 23 muscles was negligible (Figure 4C). Unexpectedly, our presence of glucose. proteomic data demonstrated a negative correlation between Muscle-Type Specific Regulation of Glucose Transport the concentrations of hexokinases (Figure 4B) and the first All striated muscles contained a complete set of proteins enzymes of glycolysis and glycogen synthesis (Figure 4C). The needed for translocation of glucose transporters, Slc2a4 (Glut4) same negative correlation was observed for glucose transporters (Figure 3), into a cell membrane in response to insulin (Figure 4A). signaling and contraction-related increase in calcium and AMP The highest discrepancy between capacities to take up and concentrations. Essentially, most of these proteins are present metabolize glucose was observed in white muscle, where low at similar levels among the striated muscles (Figure 3). concentrations of Slc2as and the hexokinases were accom- panied by high concentrations of the first enzymes of glycolysis Red skeletal muscle is the major site of insulin-stimulated ff glucose disposal (for review, see ref 24). Thus, it might be and glycogen synthesis. The lowest di erences between the two expected that red muscles contain the highest concentrations of groups of proteins were observed in the heart. proteins involved in insulin-dependent incorporation of glucose A negative correlation between the capacity for glucose transporters into a cell membrane. Our studies, however, did uptake and phosphorylation to G6P from one side, and the capacity for glycolytic degradation of G6P from the other, raises not show a significant elevation of titers of these proteins, a question regarding the metabolic significance of this except for Tbc1d4 (previously known as AS160) (Figure 3). phenomenon. The simplest explanation of this “hexokinase Evidently, the major role of red muscles in insulin-mediated paradox” is that an overabundance of glycolytic enzymes is glucose elimination from the bloodstream results from their needed for oxidation of glycogen-derived glucosyl units but not large mass and not from their unique protein composition. those produced by hexokinase reactions. Indeed, the capacity of The second mechanism regulating the translocation of the three types of striated muscles for glycogen phoshorolysis Slc2a4-containing vesicles to a cell membrane depends on (Figure 2E) correlates well with their glycolytic capacity contraction-induced increases in [Ca2+] and [AMP]. Our (Figure 2B). The above-mentioned hypothesis also explains proteomic analysis showed that concentrations of AMP- the enormously high concentrations of glycolytic enzymes (and dependent kinases (Prkas) did not differ significantly among their catalytic capacities) in comparison with the basal rate of the striated muscles (Figure 3). However, we found significant 26 ff muscle glycolysis. Evidently, the majority of glycolytic di erences in the concentrations of proteins activating Prkas in enzymes in striated, especially white, muscle is engaged in response to physiological stimuli. fi catalysis only during short episodes of glycogen breakdown that The rst manifestation of muscle contraction is a sharp rise occurs during short and intense exercises. in [Ca2+], whereas a change in [AMP] is a slower process as it is secondary to the activation of catabolic metabolism. We Cell-to-Cell Lactate Shuttle: Metabolic Coupling of observed the highest titer of calmodulin and Ca2+/calmodulin- Fibroblasts and Cardiomyocytes dependent kinases in white muscle (Figure 3), the muscle that One of the consequences of the “hexokinase paradox” is that contracts quickly and powerfully but fatigues very rapidly. heart muscle should have an approximately 2−3-fold higher Cardiac muscle, in comparison to the skeletal muscle, had a rate of basal glycolysis than that of the skeletal muscles because much lower concentration of calmodulin, which pointed to the the concentration of the hexokinases (mainly Hk1), the rate-

2484 DOI: 10.1021/acs.jproteome.5b01149 J. Proteome Res. 2016, 15, 2479−2490 Journal of Proteome Research Article

Studies on brain metabolism performed during the last 20 years have demonstrated that astrocytes take up the majority of brain glucose and oxidize it to lactate (for review, see ref 28), which is then transported to neurons and oxidized in the Krebs cycle. This process is called cell-to-cell lactate shuttle (Figure 5). To test if the shuttle may function in the heart, we examined the cellular localization of two proteins, which with different patterns of expression in cardiomyocytes and cardiac fibroblasts, would be in agreement with the lactate shuttle hypothesis. The first protein was hexokinase 1, a rate-limiting enzyme for basal glycolysis, which is expressed much more strongly in the heart than in other striated muscles (Figure 4B). The second protein was lactate transporter Slc16a3 (Mct4) (Figure 2C), which preferentially releases lactate out of a cell. If the lactate shuttle was functional, both proteins should be expressed predominantly in fibroblasts (Figure 5A). Our studies showed that both Hk1 and Slc16a3 localized mainly in cells stained with vimentin, a fibroblast marker (Figure 5B), which corroborates the idea of a fibroblast-to-cardiomyocyte lactate shuttle. This finding requires further functional studies; however, it suggests that regulation of cardiac energy metabolism is a complex phenomenon encompassing at least two types of cells, cardiomyocytes and fibroblasts, and to understand heart metabolism and find pharmacological treat- ments to heart diseases, the role of fibroblasts must be taken into account. Striated Muscle Mitochondria are Highly Homogeneous Organelles in Terms of Main Energy Pathways Cells of various striated muscles are known to have different capacities to perform oxidative metabolism (for review, see ref 31). Our proteomic data are in line with this “common knowledge”. In cardiac cells, we found the highest amount of total mitochondrial proteins (Figure 6A) and the highest abundances of pyruvate dehydrogenese (Pdh) complex and proteins of beta-oxidation, tricaboxylic acid cycle (TCA), and “ ” Figure 4. The hexokinase paradox appears to be a lack of correlation oxidative phosphorylation (OXPHOS) (Figure 6B), whereas between the titers catalytic capacity of hexokinase and of glycolytic the lowest titers of enzymes constituting these pathways were enzymes in the muscles. (A) Concentrations of glucose transporters. (B) Concentrations of hexokinases. (C) Concentrations of the observed in white muscle, which is known to be mainly anaerobic (Figure 6AB). enzymes initiating glycolysis, glycogen synthesis, and pentose 32 phosphate pathways. Recently, Murgia et al. suggested that mitochondria from various types of skeletal muscle fibers show significant differences in the subset of proteins involved in energy limiting enzymes of glycolysis, is 2−3-times higher just in the metabolism. However, our proteomic analysis, which covered heart (Figure 4B). Such high glycolytic flux in the heart is not more than 600 proteins annotated as mitochondrial, revealed confirmed by any studies, and thus, an additional mechanism that the concentrations of proteins forming the main metabolic explaining the “hexokinase paradox” in cardiac muscle must be pathways (and comprising more than 60% of all mitochondrial considered. Our preliminary immunohistochemical studies proteins), Pdh, OXPHOS, and TCA, were practically the same suggest that the “hexokinase paradox” might be better explained in each type of striated muscle (Figure 6C). in the terms of a tight metabolic coupling of cardiomyocytes However, we also found some differences in mitochondrial and cardiac fibroblasts. composition among these muscles, and some of them were in In our previous work, we showed that the expression pattern agreement with our previous data.10 For example, we found of energy metabolism enzymes in mouse heart closely that Gpd2, a mitochondrial component of the glycerophos- resembles that in mouse brain.27 Both brain and heart are phate shuttle involved in oxidation of NADH, was several times composed of two major types of cells: astrocytes and neurons, more abundant in mitochondria of white muscle than of red and cardiomyocytes and fibroblasts, respectively. Both organs muscle and almost 40 times more abundant than in preferentially utilize lactate and are sensitive to hypoxia.19,28,29 mitochondria of the heart (Figure 6D). This reflects the need In the case of the heart, practically all studies have considered of white muscle for the regeneration of NAD during very cardiac fibroblasts as a negligible component in terms of global intense but short episodes of glycolysis. cardiac energy metabolism. This probably results from the Our data also confirmed the variation in mitochondrial relatively small, less than 25%, contribution of fibroblasts to the pyruvate carriers Mpc1 and Mpc2 (also known as Brp44l and total heart mass.30 Brp44, respectively), ATP/ADP transporters, and porins

2485 DOI: 10.1021/acs.jproteome.5b01149 J. Proteome Res. 2016, 15, 2479−2490 Journal of Proteome Research Article

Figure 5. Fibroblast−cardiomyocyte lactate shuttle. (A) The fibroblast−cardiomyocyte lactate shuttle hypothesis assumes that fibroblasts take up glucose from the bloodstream and phosphorylate it via hexokinase. After its oxidation in glycolysis, they release lactate through efficient lactate exporter Mct4. The lactate is then transported to cardiomyocytes and oxidized in their mitochondria to CO2 and H2O. One of the hallmarks of the existence of the cell-to-cell lactate shuttle is heterogeneous expression of a rate-limiting enzyme for basal glycolysis, Hk1, and the lactate transporter. (B) Hk1 and Mct4 are expressed predominantly in fibroblasts (identified by vimentin staining), as detected via colocalization with the fibroblasts marker vimentin. Bar = 20 μm. (B) Hk1 and Mct4 are expressed predominantly in fibroblasts (identified by vimentin staining, Vim), as detected via colocalization with the fibroblasts marker vimentin. Cells with weaker expression of these two proteins were identified as cardiomyocytes (by staining of filamentous actin with phalloidin-FITC). Bar = 20 μm.

(Figure 6D). The most striking differences in mitochondrial the anticipated differences in mitochondrial biogenesis pathways among striated muscles were related to the potential among the three striated muscle types (Figure 7). concentration of β-oxidation enzymes. This process generates Among these factors were GA-binding protein alpha chain acetyl-CoA from free fatty acids (FFA) and delivers them to (Gabpa), constitutive coactivators of PPAR-gamma-like TCA (Figure 6C). We found that titers of enzymes of β- protein−FAM120A/c, Prkas (also known as Ampks, AMP- oxidation were significantly, almost 2-fold, lower in white activated protein kinases), Akt1/2, and cAMP-responsive muscle mitochondria than in red, and approximately 2.5 times element-binding protein 1 (CREB1). Titers of all these lower than in heart muscle mitochondria. In line with this, we proteins were the highest in heart muscle, and lowest in found that the concentration of carnitine transporters, white muscle. Except Gabpa, the differences between heart and indispensable for FFA delivery to mitochondria, was 2-fold white muscle were statistically significant (Figure 7A, Table higher in cardiac than in white muscle mitochondria (Figure S2). Furthermore, nuclear DNA-encoded mitochondrial 6C). proteins involved in regulation of replication and transcription Regulation of Mitochondria Biogenesis and Dynamics of of the mitochondrial genome or in ribosome assembly, Tfam, Their Network: Where are the Master Regulators? Tfb1m, Mterfd3, and Mtg1 (Figure 7B), were more abundant The master regulator of mitochondrial biogenesis is Pgc-1α in heart and red muscle than in white muscle, and their (peroxisome-proliferator-activated receptor γ coactivator-1α). amounts correlated well with the expected growth potential of Together with Pgc-1β, it is supposed to be enriched in tissues the organelles in a given muscle type. that contain abundant mitochondria, e.g., in heart tissue.33 Mitochondria are highly dynamic organelles able to However, in the tested muscles, we were not able to detect continuously divide and fuse in a cell, and their morphology these factors or other proteins considered as important depends on cell type. In striated muscles, highly organized regulators of mitochondrial biogenesis acting up- or down- sarcomeric structures leave no room for the expanded network stream of Pgc (Pgc-1α-related coactivator (Prc), peroxisome- of fused mitochondria, and the fragmented organelles can be proliferator-activated receptor γ (Pparγ), and Pparα). The lack more readily recruited to subcellular domains with high ATP of detection of the canonical regulators of mitochondrial demands. On the other hand, transient fusion of mitochondria fi biogenesis may be a result of very low expression of most of might be bene cial because it results in mixing of their matrix their coactivators. However, this may also suggest that, in content and thus mixing of mtDNA gene products. This might ff murine striated muscles, there are other factors that can help to compensate for potentially deleterious e ects of substitute for Prc and Pgc in their regulatory roles. For production of abnormal versions of respiratory complex example, the role of Pparα in the translation of changes in proteins in a single organelle. Therefore, some extent of cellular lipids to the expression of FA oxidation genes might be mitochondrial dynamics seems to be advantageous in cells of taken over by mitochondrial Ucp3 protein (Slc25a9), which has high OXPHOS activity. been hypothesized to be engaged in the regulation of FA We found that, in the heart, the sum of titers of cytosolic and oxidation and, additionally, in mitigation of ROS emission (for mitochondrial proteins directly involved in regulation of review, see ref 34). We found that Ucp3 was relatively highly fission/fusion of the organelles (Table S5) was two and five expressed in heart muscle and was approximately 3 and 9 times times higher than in red and white muscle, respectively (Figure more abundant than in red and white muscle, respectively 7C). However, mitochondria of cardiomyocytes, cells of the (Figure 7A). highest respiratory capacity, have been shown to be On the other hand, in the tested muscles, titers of other hypodynamic organelles35 despite the presence of the fission/ nonmitochondrial (i.e., cytosolic and nuclear) regulatory fussion proteins. However, instead of regulating the organelles’ proteins, acting up- or downstream of Pgc-1α,reflected well dynamics, these proteins are believed to play a role in the

2486 DOI: 10.1021/acs.jproteome.5b01149 J. Proteome Res. 2016, 15, 2479−2490 Journal of Proteome Research Article

Figure 6. Mitochondrial metabolism and metabolite transporters. (A) The cellular content of mitochondria measured as the percentage of total protein. The total protein content of mitochondria was obtained by summing the total protein content, as indicated in Materials and Methods. (B) The compiled contents of proteins involved in oxidative phosphorylation, Krebs cycle, free fatty acid oxidation, and pyruvate decarboxylation expressed as a fraction of the total cellular protein or the total mitochondrial protein. (C) Porins and inner membrane transporters. regulation of mitophagy,35 the process of removing senescent One of the most interesting findings of this study is the lack or depolarized mitochondria. of correlation among the titers of hexokinase and glycolytic In our studies, we were not able to detect proteins enzymes, which we called the “hexokinase paradox”. The considered as the central controllers of the mitophagic “hexokinase paradox” indicates that the enormously high level elimination of defective mitochondria, Pink1 and Parkin,36 in of glycolytic enzymes in skeletal muscles is not an adaptation to any of the muscle types. According to literature data, mice high glucose uptake but rather is an adjustment to rapid lacking cardiomyocytic Pink exhibit abnormalities of mitochon- oxidation of glucosyl units mobilized from glycogen stores. dria and develop cardiac pathologies.37 Results of our studies, Our studies confirmed the findings of Shao et al.25 that however, question the central role of these proteins in the striated muscles, especially the heart muscle, may accelerate mitophagic process, at least in adult murine heart and skeletal glucose uptake under ischemic conditions. The mechanism of 38 muscles, and support previous observations by Kubli et al., this process is based on the thioredoxin 1-mediated activation who found that Parkin is dispensable in the adult heart under of Prka-dependent translocation of glucose transporters into a normal conditions. cell membrane. In this paper, we demonstrated that cardiac fibroblasts, but ■ CONCLUSIONS not cardiomyocytes, are characterized by high expression of Analysis of the quantitative proteomes presented here uncovers Hk1 and lactate transporter Slc16a3 (Mct4). Such location of several unexpected features of striated muscles. Importantly, the proteins resembles the distribution reported for the brain our analysis reveals that, in addition to different protein and suggests the existence of a fibroblast-to-cardiomyocyte expression levels, the three types of striated muscles lactate shuttle. The transfer of lactate from astrocytes to substantially differ in their patterns of metabolic pathways neurons was originally described by Magistretti and Pellerin and isoforms of regulatory proteins. (reviewed in ref 39) . It is now well-documented that this

2487 DOI: 10.1021/acs.jproteome.5b01149 J. Proteome Res. 2016, 15, 2479−2490 Journal of Proteome Research Article

Figure 7. Regulation of mitochondrial biogenesis and dynamics of their network. (A) Nonmitochondrial proteins involved in mitochondrial biogenesis. (B) Mitochondrial proteins involved in the regulation of replication and transcription of the mitochondrial genome or in ribosome assembly. (C) Compiled total protein content of proteins regulating mitochondrial fission and fusion. process is crucial for proper brain function39,28) and that aging- white muscles replenish glycogen stores, at least under the related neuronal plasticity defects correlate with disruption of conditions of low blood glucose (e.g., during recovery from a the shuttle.40 Although further biochemical/metabolic studies high-intensity exercise). are needed to elucidate the significance of the fibroblast-to- Quantitative proteomics approaches allow not only testing of cardiomyocyte lactate shuttle, this idea is likely to be a novel hypotheses but also provide tools that can be used to promising concept, clarifying, for example, why cardiomyocytes verify some of the most fundamental dogmas of biology (for are highly sensitive to oxygen deprivation despite the highest example, see ref 41). Quantitative mapping of tissue-specific concentration of glucose transporters and hexokinases among proteomes at molar concentrations is the precondition to striated muscles. successful determine targets for pharmacological treatment. In Results of our experiments showed that mitochondria from contrast to studies providing relative protein quantitation, all three types of striated muscles possess very similar metabolic absolute data allow discrimination between major and minor profiles and that the specific aerobic capacity of a given muscle isoforms of enzymes and identification of physiologically type is reflected in the concentrations of proteins involved in relevant changes.42 Thus, our data constitute a powerful mitochondria biogenesis and dynamics. Cardiac muscle, the database and resource for future studies of muscle physiology muscle with the highest dependence on oxidative metabolism, and for the design of pharmaceutics for treatment of muscular also has the highest mitochondrial biogenesis capacity and, as a disorders. Importantly, some conclusions drawn on the basis of consequence, the greatest potential for adaptive changes in the our results, such as the potential existence of the “fibroblast− number of the organelles in response to metabolic/environ- cardiomyocyte lactate shuttle” and the “hexokinase paradox”, mental stimuli. However, some proteins considered to be point to the necessity of reinterpretation of some basic aspects crucial for mitochondrial biogenesis were absent from the of striated muscle metabolism. muscle proteomes, which emphasizes their little importance for this process. In particular, the lack of putative mitophagy- ■ ASSOCIATED CONTENT regulating proteins in mouse heart questions their central role * in the elimination of defective mitochondria, at least under S Supporting Information normal conditions. The Supporting Information is available free of charge on the The data presented in this paper are also the first to showing ACS Publications website at DOI: 10.1021/acs.jproteo- that glyconeogenesis might be an essential process in which me.5b01149.

2488 DOI: 10.1021/acs.jproteome.5b01149 J. Proteome Res. 2016, 15, 2479−2490 Journal of Proteome Research Article

Definitions of supporting tables (PDF) (11) Wisniewski, J. R.; Prus, G. Homogenous Phase Enrichment of Protein concentrations of identified proteins (XLSX) Cysteine-Containing Peptides for Improved Proteome Coverage. Anal. Chem. 2015, 87, 6861. Statistical analyses (XLSX) (12) Wisniewski, J. R.; Gaugaz, F. Z. Fast and sensitive total protein Proteomic ruler-based determination of total protein and Peptide assays for proteomic analysis. Anal. Chem. 2015, 87 (8), content per nucleus (XLSX) 4110−6. Selection of mitochondrial proteins (XLSX) (13) Wisniewski, J. R.; Zielinska, D. F.; Mann, M. Comparison of Mitochondrial fission and fusion proteins (XLSX) ultrafiltration units for proteomic and N-glycoproteomic analysis by the filter-aided sample preparation method. Anal. Biochem. 2011, 410 (2), 307−9. ■ AUTHOR INFORMATION (14) Vizcaino, J. A.; Cote, R. G.; Csordas, A.; Dianes, J. A.; Fabregat, Corresponding Authors A.; Foster, J. M.; Griss, J.; Alpi, E.; Birim, M.; Contell, J.; O’Kelly, G.; Schoenegger, A.; Ovelleiro, D.; Perez-Riverol, Y.; Reisinger, F.; Rios, *Tel.: +48 71 375 9539; e-mail: [email protected]. * D.; Wang, R.; Hermjakob, H. The PRoteomics IDEntifications Tel. +49 89 8578-2205; e-mail: [email protected]. (PRIDE) database and associated tools: status in 2013. Nucleic Acids Notes Res. 2013, 41, D1063−9. The authors declare no competing financial interest. (15) Wisniewski, J. R.; Rakus, D. Multi-enzyme digestion FASP and the ’Total Protein Approach’-based absolute quantification of the − ■ ACKNOWLEDGMENTS Escherichia coli proteome. J. Proteomics 2014, 109, 322 31. (16) Brown, G. C. Total cell protein concentration as an evolutionary We thank Professor Matthias Mann for continuous and constraint on the metabolic control distribution in cells. J. Theor. Biol. generous support, Katharina Zettl for technical assistance, and 1991, 153 (2), 195−203. Igor Paron and Korbinian Mayr for support in mass (17) Wisniewski, J. R.; Hein, M. Y.; Cox, J.; Mann, M. A ″proteomic spectrometric analysis. This work was supported by the Max- ruler″ for protein copy number and concentration estimation without Planck Society for the Advancement of Science, Polish National spike-in standards. Mol. Cell. Proteomics 2014, 13 (12), 3497−506. Science Centre Grant No. 2013/09/B/NZ1/ and German (18) Luftner, D.; Mazurek, S.; Henschke, P.; Mesterharm, J.; Research Foundation (DFG/Gottfried Wilhelm Leibniz Prize). Schildhauer, S.; Geppert, R.; Wernecke, K. D.; Possinger, K. Plasma levels of HER-2/neu, tumor type M2 pyruvate kinase and its tyrosine- ■ REFERENCES phosphorylated metabolite in advanced breast cancer. Anticancer Res. 2003, 23 (2A), 991−7. (1) Burniston, J. G.; Connolly, J.; Kainulainen, H.; Britton, S. L.; (19) Brooks, G. A. Cell-cell and intracellular lactate shuttles. J. Koch, L. G. Label-free profiling of skeletal muscle using high-definition Physiol. 2009, 587 (23), 5591−600. mass spectrometry. Proteomics 2014, 14 (20), 2339−44. (20) Sola-Penna, M.; Da Silva, D.; Coelho, W. S.; Marinho-Carvalho, (2) Malik, Z. A.; Cobley, J. N.; Morton, J. P.; Close, G. L.; Edwards, M. M.; Zancan, P. Regulation of mammalian muscle type 6- B. J.; Koch, L. G.; Britton, S. L.; Burniston, J. G. Label-Free LC-MS phosphofructo-1-kinase and its implication for the control of the Profiling of Skeletal Muscle Reveals Heart-Type Fatty Acid Binding metabolism. IUBMB Life 2010, 62 (11), 791−6. Protein as a Candidate Biomarker of Aerobic Capacity. Proteomes (21) Shulman, R. G.; Rothman, D. L. The ″glycogen shunt″ in − 2013, 1 (3), 290 308. exercising muscle: A role for glycogen in muscle energetics and fatigue. (3) Aye, T. T.; Scholten, A.; Taouatas, N.; Varro, A.; Van Veen, T. A.; Proc. Natl. Acad. Sci. U. S. A. 2001, 98 (2), 457−61. Vos, M. A.; Heck, A. J. Proteome-wide protein concentrations in the − (22) Fournier, P. A.; Brau, L.; Ferreira, L. D.; Fairchild, T.; Raja, G.; human heart. Mol. BioSyst. 2010, 6 (10), 1917 27. James, A.; Palmer, T. N. Glycogen resynthesis in the absence of food (4) Holland, A.; Henry, M.; Meleady, P.; Winkler, C. K.; Krautwald, ingestion during recovery from moderate or high intensity physical M.; Brinkmeier, H.; Ohlendieck, K. Comparative Label-Free Mass activity: novel insights from rat and human studies. Comp. Biochem. Spectrometric Analysis of Mildly versus Severely Affected mdx Mouse Physiol., Part A: Mol. Integr. Physiol. 2002, 133 (3), 755−63. Skeletal Muscles Identifies Annexin, Lamin, and Vimentin as Universal (23) Gizak, A.; Mazurek, J.; Wozniak, M.; Maciaszczyk-Dziubinska, Dystrophic Markers. Molecules 2015, 20 (6), 11317−44. E.; Rakus, D. Destabilization of fructose 1,6-bisphosphatase-Z-line (5) Holland, A.; Dowling, P.; Meleady, P.; Henry, M.; Zweyer, M.; interactions is a mechanism of glyconeogenesis down-regulation in Mundegar, R. R.; Swandulla, D.; Ohlendieck, K. Label-free mass vivo. Biochim. Biophys. Acta, Mol. Cell Res. 2013, 1833 (3), 622−8. spectrometric analysis of the mdx-4cv diaphragm identifies the (24) Zhang, J.; Liu, F. Tissue-specific insulin signaling in the matricellular protein periostin as a potential factor involved in − dystrophinopathy-related fibrosis. Proteomics 2015, 15 (13), 2318−31. regulation of metabolism and aging. IUBMB Life 2014, 66 (7), 485 (6) Murphy, S.; Zweyer, M.; Mundegar, R. R.; Henry, M.; Meleady, 95. P.; Swandulla, D.; Ohlendieck, K. Concurrent Label-Free Mass (25) Shao, D.; Oka, S.; Liu, T.; Zhai, P.; Ago, T.; Sciarretta, S.; Li, H.; Spectrometric Analysis of Dystrophin Isoform Dp427 and the Sadoshima, J. A redox-dependent mechanism for regulation of AMPK Myofibrosis Marker Collagen in Crude Extracts from mdx-4cv Skeletal activation by Thioredoxin1 during energy starvation. Cell Metab. 2014, − Muscles. Proteomes 2015, 3 (3), 298−327. 19 (2), 232 45. (7) Wisniewski, J. R.; Zougman, A.; Nagaraj, N.; Mann, M. Universal (26) Suarez, R. K.; Staples, J. F.; Lighton, J. R.; West, T. G. sample preparation method for proteome analysis. Nat. Methods 2009, Relationships between enzymatic flux capacities and metabolic flux 6 (5), 359−62. rates: nonequilibrium reactions in muscle glycolysis. Proc. Natl. Acad. (8) Wisniewski, J. R.; Mann, M. Consecutive proteolytic digestion in Sci. U. S. A. 1997, 94 (13), 7065−9. an enzyme reactor increases depth of proteomic and phosphoproteo- (27) Wisniewski, J. R.; Gizak, A.; Rakus, D. Integrating Proteomics mic analysis. Anal. Chem. 2012, 84 (6), 2631−7. and Enzyme Kinetics Reveals Tissue-Specific Types of the Glycolytic (9) Deshmukh, A. S.; Murgia, M.; Nagaraj, N.; Treebak, J. T.; Cox, J.; and Gluconeogenic Pathways. J. Proteome Res. 2015, 14, 3263. Mann, M. Deep proteomics of mouse skeletal muscle enables (28) Belanger, M.; Allaman, I.; Magistretti, P. J. Brain energy quantitation of protein isoforms, metabolic pathways, and transcription metabolism: focus on astrocyte-neuron metabolic cooperation. Cell factors. Mol. Cell. Proteomics 2015, 14 (4), 841−53. Metab. 2011, 14 (6), 724−38. (10) Rakus, D.; Gizak, A.; Deshmukh, A.; Wisniewski, J. R. Absolute (29) Jugdutt, B. I. Ventricular remodeling after infarction and the quantitative profiling of the key metabolic pathways in slow and fast extracellular collagen matrix: when is enough enough? Circulation skeletal muscle. J. Proteome Res. 2015, 14 (3), 1400−11. 2003, 108 (11), 1395−403.

2489 DOI: 10.1021/acs.jproteome.5b01149 J. Proteome Res. 2016, 15, 2479−2490 Journal of Proteome Research Article

(30) Camelliti, P.; Borg, T. K.; Kohl, P. Structural and functional characterisation of cardiac fibroblasts. Cardiovasc. Res. 2005, 65 (1), 40−51. (31) Schiaffino, S.; Reggiani, C. Fiber types in mammalian skeletal muscles. Physiol. Rev. 2011, 91 (4), 1447−531. (32) Murgia, M.; Nagaraj, N.; Deshmukh, A. S.; Zeiler, M.; Cancellara, P.; Moretti, I.; Reggiani, C.; Schiaffino, S.; Mann, M. Single muscle fiber proteomics reveals unexpected mitochondrial specialization. EMBO Rep. 2015, 16 (3), 387−95. (33) Puigserver, P.; Wu, Z.; Park, C. W.; Graves, R.; Wright, M.; Spiegelman, B. M. A cold-inducible coactivator of nuclear receptors linked to adaptive thermogenesis. Cell 1998, 92 (6), 829−39. (34) Busiello, R. A.; Savarese, S.; Lombardi, A. Mitochondrial uncoupling proteins and energy metabolism. Front. Physiol. 2015, 6, 36. (35) Song, M.; Dorn, G. W., 2nd Mitoconfusion: noncanonical functioning of dynamism factors in static mitochondria of the heart. Cell Metab. 2015, 21 (2), 195−205. (36) Narendra, D. P.; Youle, R. J. Targeting mitochondrial dysfunction: role for PINK1 and Parkin in mitochondrial quality control. Antioxid. Redox Signaling 2011, 14 (10), 1929−38. (37) Billia, F.; Hauck, L.; Konecny, F.; Rao, V.; Shen, J.; Mak, T. W. PTEN-inducible kinase 1 (PINK1)/Park6 is indispensable for normal heart function. Proc. Natl. Acad. Sci. U. S. A. 2011, 108 (23), 9572−7. (38) Kubli, D. A.; Quinsay, M. N.; Gustafsson, A. B. Parkin deficiency results in accumulation of abnormal mitochondria in aging myocytes. Commun. Integr. Biol. 2013, 6 (4), e24511. (39) Pellerin, L.; Magistretti, P. J. Sweet sixteen for ANLS. J. Cereb. Blood Flow Metab. 2012, 32 (7), 1152−66. (40) Drulis-Fajdasz, D.; Wojtowicz, T.; Wawrzyniak, M.; Wlodarczyk, J.; Mozrzymas, J. W.; Rakus, D. Involvement of cellular metabolism in age-related LTP modifications in rat hippocampal slices. Oncotarget 2015, 6 (16), 14065−81. (41) Gizak, A.; Rakus, D. Will Quantitative Proteomics Redefine Some of the Key Concepts in Skeletal Muscle Physiology? Proteomes 2016, 4 (1), 2. (42) Wisniewski, J. R.; Koepsell, H.; Gizak, A.; Rakus, D. Absolute protein quantification allows differentiation of cell-specific metabolic routes and functions. Proteomics 2015, 15 (7), 1316−25.

2490 DOI: 10.1021/acs.jproteome.5b01149 J. Proteome Res. 2016, 15, 2479−2490