Primary normal human dermal fibroblasts (NHDF) from newborn males (Camprex # CC-2509 annotated NHDF-1 and ATCC #CRL-2429 annotated NHDF-2), were cultivated at 37°C and 5% (v/v) CO2 in RPMI 1640 medium (BioWhittaker) containing 10% (v/v) fetal calf serum (BioWhittaker). Cells for experiments were used between passage 9 and 13. After preculturing, the cultures were transferred to 150 cm2 flasks and harvested at sub-confluence after approximately 72 h in RPMI 1640 medium (with 2 g/l glucose), or glucose-free RPMI 1640 medium supplemented with 2 g/l galactose, both supplemented with 10% fetal calf serum.
Cells from four 150-cm2 flasks were resuspended in 10 ml MOPS buffer (10 mM, pH 7.2) containing sucrose (200 mM), EDTA (0.1 mM) and protease inhibitor (Complete from Roche). The cells were disrupted on ice by 30 strokes in a Dounce homogenizer. Cell debris was removed by two centrifugation steps at 600 × g for 7 minutes, where the pellets were discarded and the resulting supernatant was centrifuged at 10,000 × g for 15 min. The pellet containing mitochondria was washed in the MOPS buffer (pH 7.2) without protease inhibitor, centrifuged at 10,000 × g for 15 min and stored at -80°C. After adding a sample buffer from the iTRAQ kit (Applied Biosystems, Foster City, California, USA) consisting of 0.5 M triethylammonium bicarbonate buffer (pH 8.5) with 0.1% SDS, the samples were treated with ultrasonication (Branson Sonifier 250, Branson Ultrasonics corp., Danbury, USA) at output control 3 and 30% duty cycle for three rounds of 10 seconds with one minute on ice between each round.
Western blot analysis
The mitochondrial protein samples, 6 μg per well, were separated on a 12% SDS-bis-Tris polyacrylamide gel (BioRad). Each sample was loaded in triplicate, and a standard dilution series with five concentrations in duplicate was loaded on each gel for quantitative purposes. Blotting to a PVDF membrane was performed on a Semi-Dry Transfer Cell (BioRad). The detection procedure was according to instructions from the manufacturer of ECL Plus Western Blotting Detection Reagents (GE Healthcare). The membrane was incubated over night with primary antibodies against NDUFA9 (MitoSciences, Eugene, Oregon, USA) and VDAC1/porin (Abcam, Cambridge, Massachusetts, USA). The blots were scanned on ChemiDoc (UVP, Upland, California, USA) and densitometry was performed in ImageQuant 5.0 software (Molecular Dynamics, Sunnyvale, California, USA). The ratio between the protein amount of NDUFA9 and the loading control VDAC1 was calculated for each lane and the three resulting values from glucose and galactose samples were compared.
iTRAQ labeling, IEF separation and purification of peptides
Protein concentrations in the samples enriched for mitochondria were measured by the Bradford assay (Bio-Rad Laboratories) and 100 μg of each protein sample was processed according to iTRAQ manufacturer's instruction (Applied Biosystems). Each protein sample was digested with 2 μg trypsin (Trypsin Gold from Promega, Madison, Wisconsin, USA) overnight at 30°C in iTRAQ sample buffer. Different combinations of the 4-plex iTRAQ labels, two labels per LC-MS/MS run, were used in the four different experiments, to minimize risks of systematic errors. After iTRAQ-labeling the peptide samples were combined and subsequently purified using a strong cation exchange (SCX)-cartridge; Strata from Phenomenex (Torrence, California, USA). Before loading, the samples were adjusted to pH 3.0 by dilution at least a factor ten in 10 mM phosphoric acid with 25% acetonitrile (AcN) and pH 3.0, which also served as washing buffer. The peptides were eluted with a mixture of 5% of ammonia and 30% methanol and subsequently vacuum-dried. The peptides were separated by isoelectric focusing (IEF) on a Multiphor II unit (Pharmacia Biotech AB, Uppsala, Sweden) using an Immobiline Drystrip Gradient (IPG) pH 3.5–4.5 gel (GE Healthcare, Uppsala, Sweden), a pH range previously shown to give high proteome coverage . The sample was dissolved in rehydration solution, containing 8 M urea, 0.5% IPG buffer 3.5–5 (GE Healthcare) and 0.002% bromophenol blue, and the 18 cm Drystrip was rehydrated overnight. IEF was run for 59 kVh with the following program: 1 min gradient from 0–500 V, 1.5 h gradient from 500–3500 V followed by 16 h at 3500 V. The gel strip was wiped with filter paper to remove excess cover oil from IEF and cut in ten pieces of equal size. Peptides were extracted from the gel in two steps, of one hour each, with 100 μl 5% AcN, 0.5% trifluoracetic acid (TFA), and purified on PepClean C-18 Spin Columns (Pierce, Rockford, Illinois, USA) according to manufacturer's protocol.
Nano-liquid chromatography and mass spectrometry (MS) analysis
The peptide mixtures were separated by liquid chromatography (Easy nLC from Proxeon, Odense, Denmark) coupled to mass spectrometry (LTQ-Orbitrap, Thermo Fisher Scientific, Waltham, USA) through a nano-electrospray source with stainless steel emitter (Proxeon). The peptides were separated on a reverse phase column, 75 μm in diameter and 100 mm long, packed with 3.5 μm Kromasil C18 particles (Eka Chemicals, Bohus, Sweden) at a flow of 300 nL/minute using a 100 minutes gradient of AcN in 0.4% acetic acid; starting with 5% and ending with 35% AcN. The mass spectrometry detection constituted of full scan (m/z 400–2000) with Orbitrap detection at resolution R = 60,000 (at m/z 400) followed by up to four data dependent MS/MS scans, with linear ion trap (LTQ) detection of the most intense ions. Dynamic exclusion of 25 s was employed as well as rejection of charge state +1 and real time recalibration  by lock mass on m/z 445.120025. Pulsed Q dissociation (PQD) fragmentation was performed with activation time of 0.1 s and activation Q of 0.7. For efficient fragmentation and detection of iTRAQ reporter ions, normalized collision energy of 33 was used since optimization experiments showed that it gave the highest number of identified peptides with iTRAQ signal. Selected ion monitoring (SIM) was designed as data dependent scanning targeting m/z values of proteotypic peptides (typically two peptides per protein), which had been identified in the previous experimental runs. SIM analyses were performed using full scan in LTQ, followed by SIM in Orbitrap (with a mass width window of ± 3 m/z units) and MS/MS in LTQ. Thus the fragmentation scans and acquisition of iTRAQ signal in the SIM analyses were performed in the same way as in the previous standard experimental runs. Approximately 35 peptides were on the inclusion list per run, with retention time limit of ± 5 minutes.
Database searches and statistics
The raw data files were processed using extract_msn.exe (Thermo Fischer Scientific) to generate peak lists of the tandem spectra. The processed data was searched with Mascot http://www.matrixscience.com version 2.2.04 (Matrix Science, London, UK), which was used for protein identification and iTRAQ reporter quantification. Full scan tolerance was 5 ppm, MS/MS tolerance was 0.9 Da, and up to two missed cleavages were accepted. Fixed modifications were those originating from iTRAQ protocol: iTRAQ-4plex of lysine and N-terminal and methylthio modification of cysteines, whereas oxidation of methionine and iTRAQ-4plex of tyrosine were set as variable modifications. The threshold of significance was set to 0.001, which resulted in a false discovery frequency of less than 0.003 when searched in Mascot against the decoy database of random sequences. In each study, all generated peak lists, from standard analyses as well as from SIM analyses, of the ten different fractions of peptides were merged together. The merged files were searched against the IPI human database version 3.45 (71,983 sequences, released 6/10/2008) using the MudPIT scoring algorithm of Mascot. Protein identification data can be found in additional files [see Additional files 1, 2, 3 and 4]. Throughout the manuscript the HGNC symbol http://www.genenames.org/ obtained from the IPI-database was used to refer to protein hits. iTRAQ values were reported for proteins with three or more measured iTRAQ values, where each peptide should have an expectation value of 0.02 or below. iTRAQ quantitation was performed in Mascot, were normalization to summed intensities was applied to compensate for possible variation in starting material. For details see http://www.matrixscience.com/help/quant_config_help.html. Three iTRAQ-studies were performed comparing galactose and glucose cultivation of the fibroblast NHDF-1. The three studies were performed at different times and on independent cultivations. The iTRAQ-ratio of galactose to glucose values were calculated for each protein from the three independent studies giving independent triplicate values. Average galactose to glucose ratios for each protein was reported as significantly different from 1.0 if they passed two tests 1) a threshold test of two times the global standard error (2 × 0.055 = 0.11) and 2) a two-tailed student's T-test for equal variance data.