Discovery, Identification, and Characterization of Candidate

Oct 2, 2008 - The catalytic activity of methionine aminopeptidase-2 (MetAP2) has ... Candidate MetAP2-specific protein substrates were discovered from...
0 downloads 0 Views 4MB Size
Discovery, Identification, and Characterization of Candidate Pharmacodynamic Markers of Methionine Aminopeptidase-2 Inhibition Scott E. Warder,*,† Lora A. Tucker,‡ Shaun M. McLoughlin,† Tamara J. Strelitzer,† Joseph L. Meuth,† Qian Zhang,‡ George S. Sheppard,‡ Paul L. Richardson,† Rick Lesniewski,‡ Steven K. Davidsen,‡ Randy L. Bell,‡ John C. Rogers,†,§ and Jieyi Wang‡ Advanced Technology and Cancer Research, Global Pharmaceutical Research and Development, Abbott Laboratories, 100 Abbott Park Road, Abbott Park, Illinois 60064-6202 Received May 28, 2008

The catalytic activity of methionine aminopeptidase-2 (MetAP2) has been pharmacologically linked to cell growth, angiogenesis, and tumor progression, making this an attractive target for cancer therapy. An assay for monitoring specific protein changes in response to MetAP2 inhibition, allowing pharmacokinetic (PK)/ pharmacodynamic (PD) models to be established, could dramatically improve clinical decision-making. Candidate MetAP2-specific protein substrates were discovered from undigested cell culture-derived proteomes by MALDI-/SELDI-MS profiling and a biochemical method using 35S-Met labeled protein lysates. Substrates were identified either as intact proteins by FT-ICR-MS or applying in-gel protease digestions followed by LC-MS/MS. The combination of these approaches led to the discovery of novel MetAP2-specific substrates including thioredoxin-1 (Trx-1), SH3 binding glutamic acid rich-like protein (SH3BGRL), and eukaryotic elongation factor-2 (eEF2). These studies also confirmed glyceraldehye 3-phosphate dehydrogenase (GAPDH) and cyclophillin A (CypA) as MetAP2 substrates. Additional data in support of these proteins as MetAP2-specific substrates were provided by in vitro MetAP1/MetAP2 enzyme assays with the corresponding N-terminal derived peptides and 1D/2D Western analyses of cellular and tissue lysates. FTICR-MS characterization of all intact species of the 18 kDa substrate, CypA, enabled a SELDI-MS cell-based assay to be developed for correlating N-terminal processing and inhibition of proliferation. The MetAP2specific protein substrates discovered in this study have diverse properties that should facilitate the development of reagents for testing in preclinical and clinical environments. Keywords: Methionine aminopeptidase-2 • MALDI-MS • FT-ICR-MS • thioredoxin • eukaryotic elongation factor-2 • pharmacodynamic

Introduction Discovery of disease or treatment-dependent protein markers usually involves direct interrogation of the fluid of interest, such as plasma, or screening cell-based models to provide candidates for in vivo testing. In the absence of a single instrument and technique that provides comprehensive identification and characterization of a proteome, complementary strategies are critical for most drug discovery efforts. Minimally fractionated proteomes from serum have been interrogated by 2DE1 and significant attention has focused on MALDI-/SELDIMS based profiling approaches for disease diagnosis.2,3 However, several issues affecting the potential predictive value of * To whom correspondence should be addressed: Scott E. Warder, Ph.D., Dept. of Cellular, Molecular and Exploratory Toxicology, Global Pharmaceutical Research and Development, Abbott Laboratories, R463 AP9A-224, 100 Abbott Park Road, Abbott Park, IL 60064-6202. E-mail: scott.warder@ abbott.com. † Advanced Technology, Abbott Laboratories. ‡ Cancer Research, Abbott Laboratories. § Current Address: Pierce, Thermo Fisher Scientific, 3747 N. Meridian Rd., Rockford, IL 61105. 10.1021/pr800388p CCC: $40.75

 2008 American Chemical Society

these markers have been raised, including study design, sample collection and storage, and the biological relevance of the protein markers.4,5 Even when applied to cellular lysates, the inherent nature of this MS-based approach generally results in profiles lacking proteomic breadth and depth. Regardless, putative cell culture and tissue markers have been reported, for example, to predict cell line sensitivity to PI3-kinase kinase inhibition6 and to characterize proteins differentially regulated in hepatocellular carcinoma.7 The focus on discovery of PD protein markers from high-abundance proteome levels using MALDI-MS based technologies has been minimal, primarily due to the inability of this approach to provide biomarkers that are anchored mechanistically to the drug target. Enzymes represent a class of proteins that may have tractable substrates for reporting compound inhibition. In addition to the linkage to target inhibition, novel biological mechanisms of action may be revealed. Methionine aminopeptidases (MetAPs) are metalloenzymes involved in the cotranslational removal of protein initiator methionine (Metinit). Two major forms of the enzyme, MetAP1 Journal of Proteome Research 2008, 7, 4807–4820 4807 Published on Web 10/02/2008

research articles and MetAP2, are found in eukaryotes. At the protein level, cotranslational removal of Metinit and processing of the N-termini is an essential event for cellular viability and the normal functioning of a variety of proteins.8 For example, Metinit hydrolysis is required for the proper allosteric functioning of β-hemoglobin chains.9 Other proteins require Metinit hydrolysis as a requisite step prior to co-translational modifications that ensure proper cellular localization, including myristoylation of the TRIF-related adaptor molecule (TRAM)10 and acetylation of the small GTPase, ARFRP1.11 In general, the amino acid exposed as the result of Metinit hydrolysis significantly influences the intracellular half-life of a protein via the N-end rule and potentially impacts biological functions.12 Factors affecting MetAP substrate specificity have been extensively studied, particularly in prokaryotic proteomes such as Escherichia coli.13 Although the amino acid in position 2 (P1′) from the N-terminus has been reported as a major determinant for Metinit cleavage, the rules governing this event are not clearly delineated. The substrate specificity between the two eukaryotic MetAPs, MetAP1 and MetAP2, are even less appreciated. However, this distinction is clearly of interest since MetAP2 has been defined as the target of the fumagillan class of natural product inhibitors that potently abrogate the angiogenesis process.14,15 MetAP2 is pharmacologically linked to cytostatic growth arrest at the G1/S phase of the cell cycle in vascular endothelial cells and some tumor lines, with minimal effect on most primary cells of nonendothelial type.16 TNP470, a synthetic fumagillan analog with higher potency and lower toxicity has entered clinical trials for various cancers.17 However, dose limiting cerebellar toxicity was reported for this covalent inhibitor.18,19 MetAP2-specific reversible inhibitors, including A-357300 and an anthranilic acid sulfonamide (A800141), have demonstrated activity in animal models of angiogenesis and tumor growth.16,20 In addition, a nonselective MetAP inhibitor (LAF-389) has also shown efficacy in preclinical animal models.21 More recently, MetAP1-specific inhibitors were shown to induce apoptosis in leukemia cell lines, presumably as a consequence of their interference with the G2/M phase transition of the cell cycle.22 In contrast to the top-down approach in this current study, several complementary analytical approaches have been recently reported that enable selective monitoring of proteolytically derived N-terminal peptides. These include N-terminal chemical modification with an affinity tag for enrichment 23,24 and diagonal chromatography (COFRADIC)25 strategies. However, these methods have the strict requirement that the protease-derived N-terminal peptides will ionize well in the MS and yield informative MS/MS spectra. We have recently developed an assay that monitors the N-terminal processing of GAPDH and tested the performance using circulating mononuclear cells and tumors from in vivo MetAP2 inhibitor studies.26 This report revealed that the degree of MetAP2 inhibition reported by retention of N-terminal Met of GAPDH was highly correlated with tumor inhibition. However, the IEF-immunoblot format is not ideal for a clinical setting. In this current study, we have applied multiple approaches to identify and characterize intact MetAP2-specific substrates with diverse properties that may serve as candidates for clinical assay development.

Experimental Section Cell Culture. Cells were grown in a humidified incubator at 37 °C and 5% CO2. Human leukemic cells (K562) and mouse 4808

Journal of Proteome Research • Vol. 7, No. 11, 2008

Warder et al. erythroleukemia (MEL) cells, including the corresponding media, were obtained directly from ATCC (Manassas, VA) and cultured according to the maufacturer’s suggested protocols. HT-1080 cells were grown in DMEM (Invitrogen, Carlsbad, CA) supplemented with 10% heat inactivated fetal bovine serum. Cells were treated with DMSO or with the MetAP2 inhibitors TNP-470, A-357300, and A-800141 over a concentration range of 1.0 × 10-4 to 10 µM. Cells were harvested and analyzed at 24, 48, and 72 h post inhibitor treatment. Proliferation Assays. Cells were plated in a 96-well plate at a density of (2 × 103 to 20 × 103 cells/well in 180 µL of media. Adherent cells were incubated for 3 h to allow for attachment. A compound dilution series, typically a 1 × 10-9 to 1 × 10-4 M final concentration range, were added (20 µL/well) in duplicate. After 72 h, 100 µL of media from the adherent cell lines was removed and 100 µL of CellTiter96 Aqueous One Reagent (Promega, Madison, WI) was added to each well. For nonadherent cell lines, CellTiter96 Aqueous One Reagent (6× stock; 20 µL) was added directly to the cells suspended in media. The plate was incubated at 37 °C for 1-3 h, depending on the cell line. The plates were read at 490 nM and the data analyzed with a Molecular Devices (Sunvale, CA) VERSAMAX plate reader using the supplied software (SoftProMax). MALDI-MS/SELDI-MS Profiling. Cell pellets were lysed in a hypotonic buffer and subjected to 3 freeze-thaw cycles. The sample was centrifuged (20 000g; 15 min) and the supernatant removed. The protein concentrations for all conditions of each cell line were normalized. Typically, 1 µL of a 0.5-5.0 mg/mL sample was applied to a spot on a SELDI (Bio-Rad, Hercules, CA) reversed phase target (H50). For MALDI-MS acquisition, the sample was air-dried and 0.75 µL of saturated sinapinic acid (50% acetonitrile/ddH20, 0.1% TFA) was applied. For the SELDI-MS experiments, the air-dried samples were washed with dilute HEPES buffer (10 mM) prior to the addition of sinapinic acid. Data were collected on a PBS-IIc SELDI-TOF mass spectrometer equipped with a ProteinChip autoloader. Mass spectra were recorded in linear positive ion mode at source voltage of 20 kV and a 1 GHz digitizer rate. Each spot was sampled using the following protocol: position 30-70 (0-100 scale), with 2 warming laser shots (not included in the average) and 8 laser shots (included in the average). The laser intensity was fixed at 220 (0-300 scale) and the detector sensitivity at 6 (0-10 scale) for all lysate profiling. Chromatographically enriched fractions were surveyed at a laser intensity range of 190-240. The PBSIIc TOF MS was externally calibrated using the “All-In-One” peptide mass standard (Bio-Rad, Hercules CA). Spectra were normalized to total ion current using ProteinChip Software (Version 3.1). Chromatography. The protein lysates, typically 10-15 mg, were adjusted to 50 mM Tris (1 M stock) and 3 mM DTT. The samples were loaded on a Mono Q HR 5/5 column (0.98 mL column volume (CV)) using an Ettan LC chromatography system from GE Healthcare (Piscataway, NJ). The initial buffer was 50 mM Tris, pH 8.0 and the eluting buffer was 50 mM Tris, pH 8.0, with 1 M NaCl. The chromatographic separation was at a flow-rate of 0.5 mL/min and monitored at 280 nm. A flowthrough fraction during loading was collected for further analyses. After the absorption returned to baseline, separation of bound proteins was accomplished with a NaCl gradient from 0 to 500 mM over 10 CVs of buffer. The AEX flow-through and selected AEX fractions were further analyzed on the same system with reversed-phase chromatography using a µRPC C2/C18 ST (4.6 /100) column

Discovery of MetAP2-specific Substrates (1.66 mL CV) from GE HealthCare (Piscataway, NJ). Mobile phase A was 0.1% TFA in water and mobile phase B was 0.08% TFA in 80% acetonitrile (AcN). A flow rate of 0.5 mL/min was used throughout the chromatography. Prior to loading, all samples were adjusted to pH 3 with 10% TFA. Initially, mobile phase B was 5% AcN. Chromatography was accomplished with a gradient from 5 to 70% B over 10 CVs. Elutions were monitored by UV-absorption at 215 nm and 0.5 mL fractions were collected. Analysis of Reversed-Phase Fractions by Fourier Transform Mass Spectrometry. Reversed-phase chromatographic fractions harboring CypA were lyophilized to dryness and stored at -20 °C prior to mass spectral characterization. Each fraction was initially resuspended in 50 µL of 50:50 H2O/CH3OH with 0.1% formic acid. Ten microliters of these stock solutions was further diluted into 90 µL of the electrospray solution, followed by centrifugation at 14 000 rpm for 5 min. Ten microliters of these solutions was injected into a Bruker Apollo I nanoelectrospray ion source, terminated with a New Objective 15 µm ID PicoTip Emitter at 200-250 nL/min (capillary inlet set at 1200 V and drying gas at 275 °C/20 psi). Ions were accumulated in the source hexapole for 200 ms prior to transmission through a quadrupole that serves as both a simple ion guide and a mass-selecting filter. Ions were permitted to further accumulate for 500 ms in the collision cell prior to injection into the Infinity Cell (12 ms time-of-flight) on a Bruker 7T Fourier-Transform Mass Spectrometer. A total of 512k data sets were obtained for each sample, with approximately 10-50 spectra coaveraged for broadband, mass selection and collisionally induced dissociation (CID) experiments, while 150-200 scans were co-averaged for electron capture dissociation (ECD) experiments. CID in the collision cell was accomplished by applying a -12 V DC bias at approximately 4.6 × 10-6 T. ECD in the Infinity cell was accomplished by gating electrons for 750 ms from a hollow cathode held at 1.75 A with a 50 mV DC bias. Data Reduction and BigMascot Identification and Characterization. Data files were imported into MIDAS Analysis (NHMFL, Florida State University) and processed with one zerofill and without apodization. Peak lists for each spectrum were generated manually using the NewMatch Tool in the MIDAS software. The ECD data obtained for each sample were further processed into a Mascot searchable text file and queried in BigMascot (Matrix Science, Boston, MA) against the NCBInr database, using N-terminal acetylation as a variable modification. The error tolerances for the searches were set at (0.5 Da for the precursor ion mass and (0.75 Da for the product ion masses. In-Gel Digestion and Bottom-Up Protein Identification/ Characterization. Excised gel slices were reduced, alkylated, and digested with proteases, including Asp-N (Roche, Indianapolis, IN), Lys-C (Roche, Indianapolis, IN), and trypsin (Promega, Madison, WI), according to the manufacturer’s suggested protocols using a MassPrep System (Perkin-Elmer, Waltham, MA). The extracted samples were collected in a 96well plate, concentrated to dryness and solubilized in 7 µL of buffer. LC/MS/MS of the extracts was performed as described previously.27,28 A quartenary HP 1100 pump (Hewlett-Packard, Palo Alto, CA) was interfaced with a Finnigan LCQ-Deca XP ion trap mass spectrometer (ThermoFinnigan, San Jose, CA). The tip at the end of the 100 × 365 µm fused silica capillary was pulled to 5 µm with a P-2000 laser (Sutter Instruments Co., Novato, CA). The capillary was packed with 10 cm of 5-µm C-18

research articles (RP) packing material (single-phase columns). The capillary column was aligned in line with the in-let of the massspectrometer and peptides eluting off the columns were sprayed directly into the inlet of mass spectrometer. The LCQ was set to acquire a full scan between 400 and 1400 m/z followed by full MS/MS scans between 400 and 2000 m/z of the top three most intense ions from the preceding MS scan. The data files were searched using MASCOT Version 2.2 (Matrix Science, Boston, MA) against the NCBInr database (NCBI nr 2008.01.16) using N-terminal acetylation, oxidation of Met, and alkylation of Cys as variable modifications. Peptide Synthesis, Purification, and Characterization. Purified peptides were purchased from SynPep (Dublin, CA) or synthesized internally. Briefly, the peptides were assembled on a 433A automated synthesizer (Applied Biosystems, Foster City, CA) using standard Fastmoc deprotection/coupling cycles with preloaded NR-Fmoc-amino acid Wang resins. The peptides were cleaved and deprotected with reagent K (trifluoroacetic acid/water/thioanisole/phenol/ethanedithiol/triisopropylsilane, 80:5:5:5:2.5:2.5),29 recovered by precipitation, purified by RP-HPLC, eluted with a linear gradient of 0.1% trifluoroacetic acid/water-acetonitrile and lyophilized. The final products were >95% pure by RP-HPLC and electrospray ionization massspectrometry (ESI-MS). N-Terminal Peptide MetAP1/MetAP2Activity Assay. Recombinant human MetAP1 and MetAP2 were prepared as described previously, and a coupled-enzyme chromogenic assay was used to measure MetAP enzyme activity by monitoring the production of free methionine with L-amino acid oxidase and horseradish peroxidase.16 The lyophilized peptide samples were dissolved in H2O to make 50 mM stock. To determine Kcat and Km, peptide solutions (1 × 10-2 to 5 mM final concentration) were incubated with MetAP1 or MetAP2 (1 × 10-2 to 1 µM final concentration) in assay solutions at room temperature. The rate (in mOD/min) was determined from each enzyme reaction and was converted to nM/min. Concentration response curves of enzyme activity rate versus peptide concentration were then generated for each peptide. Kinetic constants (Kcat and Km) were calculated from nonlinear regression analysis of the plot using GraphPad Prism (San Diego, CA). Preparation of White Blood Cells. Terminal blood was obtained by cardiac puncture from mice with heparin charged syringe into EDTA tubes. Blood was centrifuged at 2000g for 10 min to separate plasma. After removal of 50 µL plasma samples for compound quantification, the remainder of sample was diluted 1:1 in dextran (3%) and incubated at room temperature (30 min). After separation of phases, the top layer containing leukocyte-rich fraction was transferred to a new tube and centrifuged at 200g for 10 min. Pellets were resuspended gently in 1 mL of red blood cell lysis buffer (10 mM Tris/HCl, pH 7.5, 0.8% NH4Cl, 0.05 mM Na2EDTA) and incubated for 3 min at 37 °C to eliminate red cell contamination of the sample. The isolated white blood cells were then washed once with 10 mL of PBS, pelleted, and lysed in the IEF sample buffer. 1D/2D Gel and Western Analysis. 1D/2D gels and Western analyses were carried out using Novex gels (4-12%, MES; MOPS) and the 2D Zoom system according to the maufacturer’s protocol (Invitrogen, Carlsbad, CA), with minor changes. Protein lysates (20-30 µg) were loaded in each lane or hydrated into a first-dimension IEF Zoom strip. CypA was focused on a 6-10 pI range Zoom strip with a modified IEF protocol as follows: 200 V/30 min, 450 V/15 min, 750 V/30 min, and 2000 V/45min. Thioredoxin-1 was focused over a 3-10 pI range Journal of Proteome Research • Vol. 7, No. 11, 2008 4809

research articles

Warder et al.

Figure 1. Strategy for the discovery, identification and verification of treatment dependent peptidome and small protein (