MetAP1 and MetAP2 drive cell selectivity for a potent anti-cancer agent in synergy, by controlling glutathione redox state

Fumagillin and its derivatives are therapeutically useful because they can decrease cancer progression. The specific molecular target of fumagillin is methionine aminopeptidase 2 (MetAP2), one of the two MetAPs present in the cytosol. MetAPs catalyze N-terminal methionine excision (NME), an essential pathway of cotranslational protein maturation. To date, it remains unclear the respective contribution of MetAP1 and MetAP2 to the NME process in vivo and why MetAP2 inhibition causes cell cycle arrest only in a subset of cells. Here, we performed a global characterization of the N-terminal methionine excision pathway and the inhibition of MetAP2 by fumagillin in a number of lines, including cancer cell lines. Large-scale N-terminus profiling in cells responsive and unresponsive to fumagillin treatment revealed that both MetAPs were required in vivo for M[VT]X-targets and, possibly, for lower-level M[G]X-targets. Interestingly, we found that the responsiveness of the cell lines to fumagillin was correlated with the ability of the cells to modulate their glutathione homeostasis. Indeed, alterations to glutathione status were observed in fumagillin-sensitive cells but not in cells unresponsive to this agent. Proteo-transcriptomic analyses revealed that both MetAP1 and MetAP2 accumulated in a cell-specific manner and that cell sensitivity to fumagillin was related to the levels of these MetAPs, particularly MetAP1. We suggest that MetAP1 levels could be routinely checked in several types of tumor and used as a prognostic marker for predicting the response to treatments inhibiting MetAP2.


N-terminal proteomic analysis
Cell lysates were treated as previously described [1], with protein N-terminal d3-acetylation, cysteine reduction/ alkylation, trypsin digestion and SCX fractionation. Fractions eluted from SCX columns with retention times of 3 to 22 min were analyzed as previously described [1] with an Easy Nano-LC II (Thermo Scientific) coupled to a LTQ-Orbitrap™ Velos (Thermo Scientific). SCX fractions were loaded onto an NS-MP-10 pre-column (Nanoseparation, Nieuwkoop, The Netherlands) at a maximum pressure of 220 bars and separated on a Nikkyo Technos capillary column (NTCC-360/100-5-153, Nikkyo Technos Co., Tokyo, Japan). The MS survey scan was acquired by Fourier-Transform scanning 400-2000 Da at 60,000 resolution, with internal calibration. The 20 ions giving the most intense signals were subjected to high CID-MS with an exclusion time of 20 seconds for the selected precursor. MS/MS spectra with parent ion signals of more than 1 count and with a signal-to-noise ratio greater than 1.5 were extracted with Proteome Discoverer (Thermo Scientific, Ver. 1.4) and transferred to Mascot software (Matrix Science, London, UK, Ver. 2.4) for protein identification and the characterization of the co/post-translational modifications by comparison with the Homo sapiens reference containing splicing isoforms proteome available from the UniProtKB web site (Release 2014_11, 41988 sequences). The trypsin/P rule, with up to eight missed cleavages was used, with parent and fragment mass tolerances of 10 ppm and 0.6 ppm, respectively. Carbamidomethylcysteine and d3-acetyl on Lys were considered as fixed modifications, whereas Met-oxidation, protein N-terminal myristoylation, acetylation and d3-acetylation, phosphorylation (S/T) and K + peptide cationization (D/E) [2] were considered as variable modifications. All data were filtered at a protein false discovery rate of 1% using internal decoy approach.

Accession codes
The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository [3], with the dataset identifier PXD002690.

Sample preparation
Cells were cultured as previously described in the main text, counted using Vi-cell XR (Beckman Coulter), centrifuged and stored at -80°C. Cell pellets were resuspended in 800 μL of extraction buffer (8 M urea, 2 M thiourea, 1% DTT, protease inhibitors), sonicated and the resulting suspension was centrifuged (5 min, 8000 x g, 4°C). The supernatant was collected and split into four 200 μl aliquots for the preparation of quadruplicate samples. The aliquots were stripped of non-protein contaminants by protein precipitation in acetone (1 to 9 volumes of acetone, -20°C and overnight). After centrifugation (15 minutes, 15000 x g, 4°C), the protein pellet was dried and resuspended in 200 μL of solubilization buffer (8 M urea, 0.1 M ammonium bicarbonate, pH 8). Protein concentration was determined in the RC-DC protein assay (Bio-Rad, Hercules, CA, USA). Equal amounts of protein were used for all three cell lines. Proteins were reduced (12 mM DTT, 37°C, 30 minutes) and alkylated (40 mM iodoacetamide, IAA, 25°C, 1 hour, in the dark). Samples were diluted in freshly prepared 0.1 M ammonium bicarbonate, to obtain a urea concentration of 1 M, and proteins were digested with trypsin (1:120 enzyme:substrate ratio, 37°C, overnight). Stable isotope-labeled synthetic peptides were used to spike samples. The samples were then acidified with formic acid (pH 3), desalted and concentrated by solid-phase extraction (Sep-Pak C18, 1 cc, 50 mg, Waters). The eluate volume was decreased in a vacuum centrifuge and adjusted with 0.1%

Relative quantification
Samples were prepared as described above. Equal volumes of a concentration-balanced mixture of crude stable isotope-labeled peptides were used to spike all samples before the solid phase extraction (SPE) step. This approach was used to correct signal fluctuations and biases introduced during the sample preparation steps. For each peptide, the ratios of the light to heavy peak areas were compared to obtain a relative quantification for the three cell lines. Four sample-preparation replicates were generated and each was analyzed in three injection replicates.

Absolute quantification
Samples were prepared as described above. A pool of the three cell lines was used to mimic the sample matrix. A calibration curve covering a range extending from 1.6 to 50 fmol internal standard (IS) peptide (IDFGTHISGR) in 10 μg of sample matrix injected into the column was generated with this sample pool. The calibration curve was established by plotting the mean chromatographic peak area (using the sum of peak areas for the three selected transitions) from triplicate injections against the amounts of IS peptide injected.

Target peptide selection
Using the UniProtKB/Swiss-Prot protein database containing splicing isoforms, we selected peptides that were unique and specific to MetAP1 and MetAP2. Priority was given to peptides that had already been identified in previous shotgun experiments on equivalent samples, preferentially without fractionation, with high-quality MS/ MS spectra. The chosen peptides were 10 to 15 amino acids long, and their sequences contained no missed-cleavage sites or methionine residues. Six peptides were chosen for the first relative quantification experiment (LQCPTCIK, LGIQGSYFCSQECFK and HAQANGFSVVR for MetAP1 and ALDQASEEIWNDFR, IDFGTHISGR and NLNGHSIGQYR for MetAP2) and one peptide (IDFGTHISGR) was used for the absolute quantification of MetAP2.

Transition selection
The best transitions for each peptide were selected by analyzing heavy isotope-labeled peptides by Nano-LC ion-trap coupling (Agilent 1100 Series Nanoflow LC system (Agilent Technologies, Palo Alto, USA) coupled to an amaZon ETD ion trap (Bruker Daltonics, Bremen, Germany) to generate a spectral library. From the MS/ MS spectra stored in the spectral library, we were able to extract and monitor at least six y-ion-type transitions for each peptide, with an unscheduled method, on a micro-LC triple quadrupole system (Dionex Ultimate 3000 system coupled to a TSQ Vantage Triple Quadrupole instrument (Thermo Fischer Scientific, San Jose, CA, USA)). This made it possible to determine the retention times of all the targeted peptides, checking the co-elution of endogenous and isotopically labeled peptides, eliminating transitions with interference and adjusting the isotopically labeled peptide concentrations. A concentration-balanced mixture of the crude peptides was prepared, to obtain comparable signal intensities for light and heavy transitions. For each peptide, at least three transitions were monitored for quantification. The complete list of transitions measured is presented in Supplementary Table 2.

Micro-LC-SRM peptide characterization
Peptides were analyzed on a Dionex Ultimate 3000 system coupled to a TSQ Vantage Triple Quadrupole instrument (Thermo Fisher Scientific, San Jose, CA, USA). For each analysis, a volume of 2.5 μL of sample, corresponding to 10 μg of protein, was injected and trapped on a precolumn (Zorbax C18 stable bond, 5 μm, 1.0 × 17 mm, Agilent Technologies) and then separated on a C18 column (Zorbax 300 SB C18, 3.5 μm, 150 × 0.3 mm, Agilent Technologies). The peptides were eluted with a linear gradient of 2% acetonitrile/98% water/0.1% formic acid (solvent A) and 98% acetonitrile/2% water/0.1% formic acid (solvent B). Trapping was performed for 3 minutes, at a flow rate of 50 μL.min −1 with solvent A. Elution was performed at a flow rate of 5 μL.min −1 with the following gradient: 3 minutes of 5% B; from 5% to 35% B in 40 minutes; 5 minutes at 80% B; 16 minutes at 5% B. For optimal micro-LC-SRM, the TSQ vantage mass spectrometer was operated with the following parameters: the system was operated in positive mode, the ion spray voltage was set at 3000 V, the capillary temperature at 300°C, the nitrogen collision gas pressure was set to 1.5 mTorr, Q1 and Q3 resolution was set to 0.7 Da and the collision energy was optimized individually for each transition. Unscheduled SRM was used, with a cycle time of 3 s and a dwell time of 34 ms for each transition. The system was controlled with Chromeleon Xpress software (v. 6.8) for the liquid chromatography system and Xcalibur (v. 2.1.0) software for the mass spectrometry system.

Micro-LC-SRM data analysis
The Skyline open-source software package [4] was used to visualize the SRM data, to perform peak selection and transition peak area integration, and to verify SRM peak group identification manually by checking the co-elution and the correct relative fragmention intensities between light and heavy isotope-labeled peptides. The coefficient of variation for peptide retention time was calculated from all measurements of the light and heavy peptide over all replicates (Supplementary  Table 2). For the relative quantification experiment, the overall reproducibility of the experiment was verified by calculating light/heavy isotope area ratios for the sum of all transitions, and checking that coefficients of variation were below 20% for all sample preparation replicates. Relative protein quantification and testing for differential protein expression were performed with the R package MSstats (http://www.R-project.org). The acceptance criteria for statistically different protein abundance changes between two conditions were set at a p-value below 0.05 and a fold-change of more than 1,2. For the absolute quantification experiment, the linearity criteria required experimental dots in standard curves to display a mean CV of less than 15% between triplicate injections. Experimental dots also had to fall within the mean 80-120% accuracy range when calculating the expected injected amounts from regression equations after base 2 logarithmic transformations (Supplementary Table 2). The limit of quantitation (LOQ) is the lowest point satisfying all the above mentioned criteria. Only the points satisfying all these criteria were used to calculate the linear regression equation and correlation coefficient. Endogenous peptide levels in the three cell lines were measured in sample preparation triplicates (Supplementary  Table 2). Figure S2: Mass spectrometry pipeline for the large-scale identification of N-terminal peptides and modifications. Experimental design and procedure used to identify protein N-termini and to determine their status. The different cell lines were left untreated (CT) or were treated (Fum) before sample processing. The key steps are shown in blue and include the chemical heavy isotope acetylation of N-termini, making it possible to differentiate endogenous acetylation from free N-termini. The strong cation exchange step enriches the preparation in N-terminal peptides. MS and data processing are then used to identify the protein and its N-terminal modifications. Further details are provided in the Materials and Methods section.