Native peptides derived from fibrinogen alpha chain, complement C3 and complement C4 were used in this study. The peptides and their isotopically labeled analogs were purchased from the Synthetic Peptide Application Lab at the University of Pittsburgh (FIBA, DSGEGDFLAEGGGVR) or Emory University Michrochemical Facility (CO3, SSKITHRIHWESASLL and CO4, NGFKSHALQLNNRQI). The DSGEGDFLAEGGGVR peptide analog carried one 13C isotope in each glycine residue, while the SSKITHRIHWESASLL and NGFKSHALQLNNRQI peptides carried six 13C and one 15N isotopes in a leucine at position 15 and 10, respectively. Guanidine hydrochloride, trifluoroacetic acid, acetonitrile, water, methanol, 3M Empore C18-SD cartridges, and ammonium bicarbonate were purchased from Sigma-Aldrich (St. Louis, MO). BCA assay, BSA standard, methyl methanethiosulfonate, and tris(2-carboxyethyl) phosphine hydrochloride (TCEP) were purchased from Pierce (Rockford, Il). Amicon Ultra-0.5 ml ultrafilters with 30 kDa MWCO membrane were from Millipore (Billerica, MA). Sequencing Grade Modified Trypsin was obtained from Promega (Madison, WI). Water Optima LC/MS, acetonitrile Optima LC/MS, and formic acid (FA) Optima LC/MS were purchased from Fisher Scientific (Fair Lawn, NJ). Guard column XTerra MS C18, 5 μm, 2.1 × 20 mm, analytical column XTerra MS 3.5 μm C18 2.1 × 100 mm, trap column nanoACQUITY UPLC 5 μm Symmetry C18, 180 μm × 20 mm, and nanoACQUITY UPLC column, 1.7 μm BEH C18, 75 μm × 150 mm were obtained from Waters (Milford, MA). Column ProSwift RP-1S, 4.6 × 50 mm column was purchased from Dionex (Sunnyvale, CA). PicoTip emitter tips were from New Objective (Woburn, MA). NuPAGE 10% Bis-Tris electrophoretic gels, running and sample buffers and XCell SureLock Mini-Cell were from Invitrogen (Carlsbad, CA).
Blood collection, processing and serum handling
Samples were collected in agreement with established IRB protocols. Patients diagnosed with hepatocellular carcinoma were enrolled at the National Cancer Institute of Cairo University, Egypt, from 2000 to 2002 as described previously [27, 35]. Individuals free of liver disease were recruited from the orthopedic department of Kasr El-Aini Faculty of Medicine at Cairo University and samples from the patients diagnosed with chronic liver disease were acquired at the Ain Shams University Specialized Hospital and the Tropical Medicine Research Institute, Cairo, Egypt as described [27, 35]. Briefly, adults (age 18 and older) previously diagnosed with HCC and those without a prior history of cancer were eligible to participate in the study. The presence of HCC and chronic liver disease was confirmed by pathology, cytology, imaging and elevated levels of serum AFP. Blood samples were collected by a trained phlebotomist each day at around 10 a.m. and processed within a few hours according to a standardized protocol. Aliquots of sera were frozen immediately after collection and stored at -80°C until the time of analysis.
Preparation of native peptides by denaturing ultrafiltration
Serum (100 μl) was diluted with water (300 μl), 8 M guanidine HCl (300 μl, final concentration 6 mol/l), or AcN (300 μl, final concentration 20%). The Amicon Ultra-0.5 ml ultrafilters with 30 kDa MWCO membrane were washed twice with 500 μl of 6 M guanidine HCl at 5.000 × g for 15 min. The remaining wash solution was removed by converting the filter upside down and spinning at 1.000 × g for 1 min. Samples were filtered at 8.000 × g and 30 min. Peptides were cleaned-up using C18 SPE cartridges; 3 M Empore C18-SD cartridges were washed with 0.5 ml of methanol and equilibrated with 0.5 ml of SPE solvent A (2% AcN, 0.1% trifluoroacetic acid (TFA)). After diluting 80 μl of the ultrafiltrates with up to 1 ml of the SPE solvent, samples were loaded onto the cartridges and trapped peptides were washed with 1 ml of SPE solvent A. Peptides were eluted with 0.4 ml of SPE solvent B (80% AcN, 0.1% TFA or 35% AcN, 0.1% TFA) directly into a new micro-tube. BCA assay and a UV-based HPLC assay were used to estimate the peptide yield in the ultrafiltrate. Each sample was prepared for this purpose in triplicate.
SPE eluates were dried in a vacuum centrifuge and re-dissolved in 20 μl of 0.2% SDS solution. The BCA assay was used according to the instruction of manufacturer with minor modifications. A 5 μl aliquot of each sample was incubated with a 20 μl of a 50:1 ratio of the kit reagent A:B at 37°C for 1 h. Absorbance was measured against a BSA protein standard curve with each sample measured in triplicate. Samples outside the linear range were diluted accordingly. Absorbance at 562 nm was determined using a micro-volume UV-vis spectrophotometer ND-1000 (Thermo-Fisher, Wilmington DE).
UV-based HPLC assay
This assay was used for relative comparison of the peptide yield. Fifty microliters of the ultrafiltrate were diluted 20× by RP solvent A (2% AcN, 0.05% TFA). For each replicate, 900 μl was loaded onto the column on HP 1100 HPLC system (Hewlett Packard, Waldbronn, Germany); we used a guard-column (XTerra MS C18, 5 μm, 2.1 × 20 mm) heated to 40°C at a flow rate of 1 ml/min with the peptides eluted by a step increase from 0% to 70% of RP solvent B (98% AcN, 0.05% TFA) in 0.01 min. The chromatogram was monitored at 214 nm and a blank chromatogram resulting from the injection of 10 μl of RP solvent A was subtracted. Area of all detected peaks was summed up as an estimate of peptide quantity. This allowed us to assess the quantity independently of any secondary reaction of the peptides required in the BCA assay.
Recovery of peptides of interest after denaturing ultrafiltration
Light synthetic peptides FIBA (50 pmol), CO3 (5 pmol) and CO4 (2 pmol) dissolved in 45 μl of 8 M guanidine HCl were added to 15 μl of serum. The ultrafiltration was performed as described above. Ultrafiltrate (30 μl) was spiked with isotopically labeled analogs of FIBA (25 pmol), CO3 (2.5 pmol) and CO4 (1 pmol) and desalted on C18 SPE cartridges as stated above. The concentrated SPE eluate (approximately 50 μl) was mixed with 2.5 μl of a matrix solution (3.6 mg of CHCA/ml of 50% AcN, 0.1% TFA). From this mixture 0.7 μl was spotted in triplicate on an Opti-TOF MALDI sample plate and the MS spectra were recorded in relfectron mode on a 4800 MALDI-TOF/TOF mass spectrometer (Applied Biosystems/MDS Sciex, Foster City, CA). The intensity of peaks corresponding to m/z of non-labeled and labeled FIBA, CO3 and CO4 peptides were extracted and the ratio of light and heavy peptide analogs was calculated. The ratios were compared to ratios obtained from the direct mixture of light and heavy peptides prepared from the same starting solutions that did not undergo ultrafiltration. The recovery was calculated as the change in ratios determined in spiked ultrafiltrates of serum and directly in the peptide mixture.
MALDI-TOF and HPLC-UV analysis of the denaturing conditions in a BSA/trypsin model
Eight microliters of a BSA solution (2.5 μg/μl) was added either to water, 20% AcN, or 6 M guanidine HCl as above. BSA was TCEP reduced, blocked by methyl methanethiosulfonate, and digested with 0.4 μg of trypsin in 100 mM ammonium bicarbonate buffer, pH 7.5, at 37°C for 40 h. Instead of the trypsin solution, water was added as a control. The reaction was stopped with the addition of 380 μl of the above RP solvent A. The results were evaluated using MALDI-TOF mass spectrometry and an HPLC analysis with UV detection. For MALDI analysis, 30 μl of the sample was desalted on 3 M Empore C18-SD cartridges using the same procedure as described above. The SPE eluates were concentrated in a vacuum centrifuge to approximately 20 μl. MALDI-TOF analysis was conducted in triplicate as described in the previous section. For the HPLC-UV detection, 200 μl was loaded onto a ProSwift RP-1S, 4.6 × 50 mm column. The thermostat was set up at 60°C and the flow rate was 1.8 ml/min. The gradient was from 0% to 75% B in 6 min. The chromatogram was monitored at 214 nm.
Test of denaturing potential using three native peptides in serum
Serum (15 μl) was diluted either with 60 μl of water or 60 μl of 8 M guanidine HCl (final concentration 6 mol/l). Synthetic peptide standards were added to each serum sample as follows to achieve comparable detector response: 187.5 pmol of CO3 peptide and 37.5 pmol of CO4 peptide. The samples were incubated for two hours and a second batch of samples was incubated overnight. After incubation, isotopically labeled peptides were added to each sample in the following amount: 25 pmol of FIBA, 250 pmol of CO3 and 50 pmol of CO4 which gives initially an approximately 1:1 response (CO3 and CO4) and 6:1 response (FIBA) with respect to the light peptide. Negative control samples were prepared in exactly the same way but without the two hour incubation. Low molecular mass polypeptides including the added synthetic peptides were enriched by ultrafiltration as described above including the C18 SPE cartridge cleanup. MALDI-TOF analysis was conducted in triplicate as described above. The intensity of peaks corresponding to m/z of non-labeled and labeled FIBA, CO3 and CO4 peptides were extracted and quantified.
2D LC-MS/MS identification of native serum peptides
Serum (200 μl) diluted with 600 μl of 8 M guanidine HCl was kept on ice for 5 min. The remaining steps were as described above without spiking with the standard peptides. For each analysis, we combined two ultrafiltrates. First dimensional HPLC separation was an RP HPLC at pH 10.0. 200 mM NH4FA, pH 10, was added to the ultrafiltrates to achieve a final concentration of 20 mM NH4FA. The diluted sample was loaded onto an XTerra MS 3.5 μm C18 2.1 × 100 mm column. The thermostat was set at 40°C and the flow rate at 0.2 ml/min. The peptides were eluted by a 2% to 62% B in 60 min formed by basic RP buffer A (20 mM NH4FA, pH 10) and basic RP buffer B (90% AcN, 20 mM NH4FA, pH 10). The chromatogram was monitored at 214 nm. Starting at 5 min, 14 fractions were collected each 5 min. Peptides in each fraction were dried and re-dissolved in 75 μl of LC-MS solvent A (2% AcN, 0.1% FA) for an injection onto a nanoACQUITY system (Waters, Milford, MA). Peptides were pre-concentrated on a nanoACQUITY UPLC trapping column, 5 μm Symmetry C18, 180 μm × 20 mm at a flow rate 20 μl/min. The peptides were resolved on an analytical nanoACQUITY UPLC column, 1.7 μm BEH C18, 75 μm × 150 mm by a gradient formed by LC-MS solvents A and B (98% AcN, 0.1% FA) at the flow rate 300 nl/min. The linear gradient was from 2% to 40% B in 33 min with on-line detection using a Q-TOF tandem mass spectrometer (QSTAR Elite, Applied Biosystems/MDS Sciex, Foster City, CA) equipped with a nanoelectrospray ion source with a PicoTip emitter. The instrument was operated in the Information Dependent Acquisition (IDA) mode in which the MS spectrum was recorded for 1 s and the 4 strongest 2+, 3+, 4+ or 5+ ions were selected for MS/MS with Dynamic Exclusion feature enabled. The MS/MS spectra were recorded in the m/z range 150 - 2000 using Automatic Collision Energy and Automatic MS/MS Accumulation (max. accumulation was 0.7 s). Each sample was injected twice.
The spectral data was searched using ProteinPilot 3.0 (Applied Biosystems/MDS Sciex, Foster City, CA) using the Paragon algorithm, which uses a sequence-tag search strategy, facilitating non-specific peptide identification without inflating search times. The search was specified using the following parameters in the Paragon method: sample type - identification; cysteine alkylation - none; digestion -none; instrument - QSTAR Elite ESI, species - Homo sapiens; search effort - thorough. The data were searched against the UniProt protein database. The Proteomics System Performance Evaluation Pipeline incorporated directly in ProteinPilot 3.0 was used to conduct target/decoy searches facilitate initial FDR estimates.
MS/MS spectral library
ProteinPilot results, including decoy hits, were output in XML format using the vendor supplied tool, group2xml. The resulting XML file was parsed using an in-house script to permit filtering of the peptide identifications for score and identified post-translational modifications; to remove decoy hit peptides; and to permit re-estimation of peptide, ion, and spectrum level FDR for the remaining identifications. This script was used to eliminate peptide identifications with identification confidence (ProteinPilot Probability) less than 0.95; precursor mass more than 0.2 Da from the theoretical monoisotopic mass or the mass corresponding to the isotope cluster peak associated with one, or two, 13C isotopes; and modifications other than oxidation and Pyro-glu on Glu and Gln. After filtering, the high-quality peptide identifications and their associated tandem mass-spectra, as provided in the XML formatted output of ProteinPilot, were reformatted as mzXML and pepXML, with corresponding faux spectral identifiers. Skyline's spectral library (University of Washington, Seattle, WA) build process reads the mzXML and pepXML file pair and constructs a spectral library suitable for Skyline's MRM design tools. For each peptide ion (distinct peptide sequence, charge and modification state), Skyline clusters repeat identifications and selects a representative spectrum that is, on average, most like the others in the cluster. Only the representative spectra are retained. After ProteinPilot analysis, 28,077 unfiltered peptide identifications were formatted as XML, and 1469 were retained by the above filters. After clustering, the spectral library held MS/MS spectra for 416 peptide ions, representing 349 native peptide sequences, for browsing and MRM assay design. The library, configured for Skyline, facilitates the design of LC-MRM MS assays for native peptides.
LC-MRM MS assays
Chromatographic separations were performed on a nanoACQUITY system (Waters, Milford, MA). After loading of the samples from a 4°C cooled autosampler, peptides were pre-concentrated on a nanoACQUITY UPLC trapping column, 5 μm Symmetry C18, 180 μm × 20 mm at a flow rate 15 μl/min for one min. Fast gradient from 20% to 45% of LC-MS solvent B in 5 min at a flow rate 400 nl/min was used to resolve peptides on an analytical nanoACQUITY UPLC column, 1.7 μm BEH C18, 75 μm × 150 mm. Peptides were quantified on a 4000 QTRAP hybrid mass spectrometer (AB Sciex) equipped with a nanoelectrospray ion source with a PicoTip emitter. A spray voltage of 2400 V was used with a source temperature 150°C. The mass spectrometer was operated in MRM mode with first quadrupole (Q1) filtering with 0.7 and third quadrupole (Q3) filtering with 1.0 unit mass resolution. For all MRM analyses, 30 ms dwell time was used for each transition. MIDAS based enhanced product ion spectra were recorded with Q1 filter set to 1.0 unit resolution. Enhanced product ion spectra were recorded at a scan speed 4000 amu/s between 150 to 1200 m/z, with enabled dynamic fill time and Q0 trapping.
The five most intense transitions for each suitable precursor were selected based on data deposited in the MS/MS library using Skyline. LC-MRM MS chromatograms were obtained on samples of native peptide from serum. Detected peaks were subsequently validated by a MIDAS experiment. After we unambiguously confirmed the detectability of selected peptides in serum samples, synthetic peptides were used for confirmatory analyses and optimization of collision energy and declustering potential. Three most intense optimized transitions per precursor (FIBA and CO4) were used in the final method for peptide quantification. The confirmatory analyses of synthetic CO3 peptide revealed that the best intensity transition is produced from its quadruply charged ion. The final method for CO3 quantification therefore included two most intense optimized transitions from triply and quadruply charged precursors which represent four most intense transition overall. Carryover between injections as well as potential cross-talk between light and heavy equivalents of the peptides was subsequently examined. The final MRM method was examined for the linearity of response and the limit of quantification using relative dilution series experiments (see Additional file 2). Potential background interferences were assessed by comparing normalized peak areas of each precursor to be monitored. The values obtained from injections of heavy labeled peptides dissolved in LC-MS solvent A were compared with those originating from the spiked heavy labeled peptides as well as from the native peptides in the serum samples. The MRM data were processed and evaluated in Skyline.
Samples to be analyzed were prepared according to the protocol described above. Sera (20 μl) were mixed with 8 M guanidine HCl containing heavy labeled analogues for absolute quantification. To each microliter of serum, 500 fmol of the FIBA and CO3 peptides and 50 fmol of the CO4 peptide were added together with 625 fmol of peptide SSKITHRIHWESASLLR per μl of serum, added to decrease non-specific adsorption of the CO3 peptide. The amount of native peptides enriched from an equivalent of 2 μl of original sera was injected for the final analysis and each sample was measured in duplicate. The MRM spectra data files were imported into Skyline for evaluation. The quantity of the peptides was determined by comparison to the spiked internal standard. Data was exported to GraphPad Prism 5 (GraphPad Software, La Jolla, CA) for statistical evaluation.