Screening for Enzyme Inhibitors by Surface Plasmon Resonance

Anal. Chem. 2004, 76, 5243-5248

Accelerated Articles

Screening for Enzyme Inhibitors by Surface Plasmon Resonance Combined with Mass Spectrometry Jonas Borch and Peter Roepstorff*

Department of Biochemistry and Molecular Biology, University of Southern Denmark, Campusvej 55, DK 5230 Odense M, Denmark

We have developed a novel strategy to identify enzyme inhibitors that interact directly with their enzyme targets. In the approach, an enzyme is immobilized on a sensor chip, and it is determined whether the immobilized enzyme is still active by incubation with model substrates and mass spectrometric analysis of the products. Putative inhibitors or mixtures containing putative inhibitors are then injected over the sensor chip for binding analysis with surface plasmon resonance. It is then tested whether the bound compounds inhibit the enzymatic activity by subsequent incubation with the model substrate and mass spectrometric analysis. If the bound compound inhibits the enzyme, the inhibitor is eluted from the enzyme and characterized by mass spectrometry. To test the strategy, it has been applied to the well-characterized interaction between trypsin and pure bovine pancreas trypsin inhibitor. Furthermore, fractions of plant extracts were screened for binding to and inhibition of carboxypeptidase B.

Enzyme inhibitors are used widely in the biological sciences, e.g., to inhibit enzymatic activity during purification of biological molecules and in classification and characterization of enzymes. The activities of some medical drugs are based on their actions as inhibitors of specific enzymes. Important examples are the anticancer drug Imatinib (Glivec), which inhibits tyrosine kinases,1 HIV protease inhibitors acting against the HIV virus,2 and an * To whom correspondence should be addressed. E-mail. [email protected]. Fax: +45 65 93 26 61. (1) Capdeville, R.; Buchdunger, E.; Zimmermann, J.; Matter, A. Nat. Rev. Drug Discovery 2002, 1, 493-502. (2) Molla, A.; Granneman, G. R.; Sun, E.; Kempf, D. J. Antiviral Res. 1998, 39, 1-23. 10.1021/ac049335f CCC: $27.50 Published on Web 08/20/2004

© 2004 American Chemical Society

inhibitor of angiogenin converting enzyme (a protease) used against renal disease.3 Due to the potential use of enzyme inhibitors as therapeutical agents, there is a general interest for development of screening procedures to identify novel enzyme inhibitors. Enzyme inhibitors have several modes of action. Some interact directly with the active site of the enzyme; for example, bovine pancreas trypsin inhibitor (BPTI) inhibits its endoproteolytic enzyme target by specific interaction with the active site, hence blocking entry of substrate.4 Other inhibitors interact with cofactors. EDTA, for example, removes divalent cations needed for enzymatic activity of several enzymes by chelation. Traditionally, screening procedures for enzyme inhibitors are based on measurements of activity of a purified enzyme, to which purified fractions of biological material or compounds of chemical libraries are added. Other strategies are based on identification of physical interactions between the enzyme and potential inhibitors and assaying for inhibitory action later. Examples are affinity chromatography,5,6 phage display,7,8 and the yeast two-hybrid system.9,10 Mass spectrometric analysis has been employed in screening for enzyme inhibitors in a number of cases. These (3) Thurman, J. M.; Schrier, R. W. Am. J. Med. 2003, 114, 588-598. (4) Ascenzi, P.; Bocedi, A.; Bolognesi, M.; Spallarossa, A.; Coletta, M.; De Cristofaro, R.; Menegatti, E. Curr. Protein Pept. Sci. 2003, 4, 231-251. (5) Zhang, B.; Palcic, M. M.; Schriemer, D. C.; Alvarez-Manilla, G.; Pierce, M.; Hindsgaul, O. Anal. Biochem. 2001, 299, 173-182. (6) Hamdaoui, A.; Schoofs, L.; Wateleb, S.; VandenBosch, L.; Verhaert, P.; Waelkens, E.; DeLoof, A. Biochem. Biophys. Res. Commun. 1997, 238, 357360. (7) Krook, M.; Lindbladh, C.; Eriksen, J. A.; Mosbach, K. Mol. Diversity 1998, 3, 149-159. (8) Tanaka, A. S.; Silva, M. M.; Torquato, R. J. S.; Noguti, M. A. E.; Sampaio, C. A. M.; Fritz, H.; Auerswald, E. A. FEBS Lett. 1999, 458, 11-16. (9) Jaffrey, S. R.; Snyder, S. H. Science 1996, 274, 774-777. (10) Zhang, J.; Zhang, L. F.; Zhao, S. M.; Lee, E. Y. C. Biochemistry 1998, 37, 16728-16734.

Analytical Chemistry, Vol. 76, No. 18, September 15, 2004 5243

include identification of inhibitors isolated by affinity chromatography and detection of affinity-dependent depletion of putative inhibitors from chemical libraries incubated with immobilized enzymes.11 Another approach includes mixing of putative inhibitors with the enzyme followed by electrospray ionization mass spectrometry (MS), which can monitor intact complexes in the gas phase.12 Jankowski and co-workers13 immobilized proteins from an extract and incubated the derivatized beads with model substrates to monitor selected enzymatic reactions by MS. They demonstrated that this approach could be used to screen for inhibitors by mixing potential inhibitors into the model substrate before incubation with the beads and then detect changes in enzymatic activity. The optical surface plasmon resonance (SPR) phenomenon can detect binding in the vicinity of a surface due to local changes in the refractive index. In the case of Biacore instruments, ligands are immobilized in flow cells (fc) on a dextran matrix attached to the gold-coated sensor chip. Soluble, unlabeled analytes are then passed over the surface in a continuous stream. If they bind the ligand, an increase proportional to the amount of bound material is detected by the SPR response. The flow may then be changed to buffer without analyte, which allows dissociation to be monitored. Since the response is plotted against time, kinetic information can be derived from the experiments. Concomitant measurement of binding to a control fc allows determination of whether binding is specific. Recently, methods have been developed for elution of analyte bound to the ligand immobilized on the senorchip. The eluate can then be analyzed by MS (reviewed in ref 14). In the present study, a new screen for enzyme inhibitors based on surface plasmon resonance (SPR) combined with mass spectrometry (MS) is presented. The SPR-MS-based approach is designed to specifically find inhibitors that interact directly with the enzyme target. First, direct interactions in mixtures of compounds with a known, immobilized enzyme is detected with SPR. Second, inhibition of the enzymatic action of the immobilized enzyme is detected with MS, and finally, the inhibitor is recovered from the immobilized enzyme and characterized with MS. Strategy. The strategy is outlined in Figure 1. It is designed for observation of inhibitors that interact directly with their enzyme-target. SPR detection on a biacore 3000 instrument was chosen as the first step due to the instrument’s ability to perform automated binding experiments and recovery of bound analytes. Mass spectrometry was chosen to monitor enzymatic activity, inhibition, and molecular mass of the recovered inhibitors, because of its superior sensitivity and specificity. After immobilization of an enzyme of interest on the sensor chip, the following strategy was employed: (1) The activity of the immobilized enzyme was confirmed by incubation of model substrate on the enzymederivatized sensor chip, and the enzymatic reaction was monitored by mass spectrometry. (2) Binding tests of compounds from crudely fractioned extracts of biological material was then per(11) Cancilla, M. T.; Leavell, M. D.; Chow, J.; Leary, J. A. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 12008-12013. (12) Gao, J. M.; Cheng, X. H.; Chen, R. D.; Sigal, G. B.; Bruce, J. E.; Schwartz, B. L.; Hofstadler, S. A.; Anderson, G. A.; Smith, R. D.; Whitesides, G. M. J. Med. Chem. 1996, 39, 1949-1955. (13) Jankowski, J.; Stephan, N.; Knobloch, M.; Fischer, S.; Schmaltz, D.; Zidek, W.; Schluter, H. Anal. Biochem. 2001, 290, 324-329. (14) Mattei, B.; Borch, J.; Roepstorff, P. Anal. Chem. 2004, 76, 18A-25A.

5244 Analytical Chemistry, Vol. 76, No. 18, September 15, 2004

Figure 1. Strategy for inhibitor screening: After confirmation of activity of the immobilized enzyme by mass spectrometric monitoring of the action on a model substrate (step 1), binding of putative inhibitors is detected with SPR (step 2). Next, inhibition of the enzyme activity on the model substrate is assessed by mass spectrometry (step 3). and finally, the inhibitor is eluted from the sensor chip and characterized by mass spectrometry (step 4).

formed by monitoring binding of compounds to the immobilized enzyme by SPR. (3) Inhibition of the immobilized enzyme after binding of putative inhibitors was tested by incubation of the resulting complex with substrate followed by analysis of the reaction products by MS. Finally, (4) compounds binding to and inhibiting the enzyme was characterized by mass spectrometry after recovery from the sensor chip. EXPERIMENTAL SECTION Protein Extraction and Fractionation. Extraction. Intact Solanum tuberosum (potato) tubers, Capsicum annuum var. annuum (red pepper), Solanum melongena (egg plant), and Lycopersicon esculentum var. cerasiforme (cherry tomato) fruits were boiled for 10 min in water followed by homogenization with 50 mM 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES), pH 7.4, containing 16 mg/mL ascorbic acid in a domestic blender (Braun). Next, debris was sedimented by centrifugation at 40000g for 30 min at 4 °C, and the cleared supernatant was stored at -20 °C until use. The centrifugation step was repeated if precipitation occurred after thawing. Size Fractionation. To fractionate potato tuber proteins, extracts were spun through Microcon membranes (Millipore, Billerica, MA) with a 10 000 molecular weight cutoff at 10.000 g until the extract was concentrated ∼20 times. Ammonium Sulfate Precipitation. Proteins were precipitated from daily pepper fruit, egg plant, and tomato extracts by adding solid ammonium sulfate to 70% saturation on ice and incubating for 30 min. The 0-70% insoluble fraction was separated from the supernatant by cenrtifugation at 20000g for 30 min. The pellet was washed briefly with water, resuspended to 1:20 of the original extract volume in running buffer, and used for binding and inhibition experiments. Surface Plasmon Resonance. A Biacore 3000 instrument (Biacore AB, Uppsala, Sweden) was used for all experiments. For

running and sample buffer 10 mM HEPES, 150 mM NaCl, pH 7.4 (HBS) was used for experiments with immobilized trypsin, and HBS containing 100 µM ZnCl2 (HBSZ) was used for carboxypeptidase B experiments. For binding studies, the buffers contained 0.005% (v/v) of the nonionic detergent polysorbate 20 (P20) (Biacore AB). For studies where MS was used downstream, detergent was omitted. Immobilization. Immobilization was performed on sensor chips with a carboxymethyl dextran matrix (CM5 sensor chips, Biacore AB) by amine coupling according to the manufacturer’s recommendations. A level of 3000 RU per flow cell was typically reached for trypsin (Promega, Madison, WI) and 5000 RU per flow cell for carboxypeptidase B (CPB) and bovine serum albumin (BSA) (both from Sigma, St. Louis, MO). Activity Assays. Trypsin. Model peptides were exposed to immobilized enzymes in the Biacore instrument with the sensor chip docked in the integrated flow cartridge (IFC) with SPR monitoring. The model peptide calcitonin gene-related peptide (CGRP 8-37) (see Figure 2a for sequence and cleavage sites) (Sigma) with a monoisotopic mass of 3124.73, was incubated on a trypsin-derivatized flow cell in 4 µL of HBS at a concentration of 2.5 µM for 10 min, after which the reaction products were collected for mass spectrometric analysis. The method command microrecover was used to control the incubation. CPB. Activity assays were performed in a Surface Prep Unit (Biacore AB) in the SP_2 configuration. To allow for a higher amount of immobilized enzyme, the Surface Prep Unit employs 16 mm2 of the sensor chip as a single flow cell in the SP_2 configuration. It is not possible to measure binding to the sensor chip in the Surface Prep Unit. Model peptides were incubated on the CPB-derivatized sensor chip in a 2-µL droplet controlled by the instrument program MSPRECOVER. The model peptide was the m/z 1175.6 (monoisotopic mass) peptide of a tryptic digest of CGRP 8-37 (see Figure 3c for sequence and cleavage sites) at a concentration of 7 µM. For tests of inhibition, the sample was injected for 8 min over the sensor chip prior to peptide incubation. Fractions of plant extracts were diluted 1:20 prior to inhibition, while the concentration of pure, commercially available carboxypeptidase inhibitor was 0.2 µM. Binding Studies. Trypsin. The 3.3 µM BPTI was injected over the myoglobin (reference) and trypsin (enzyme) derivatized flow cells on the sensor chip. To monitor dissociation, the flow was automatically changed to running buffer after sample injection. The surface was regenerated with 5% formic acid. CPB. Binding assays were performed with 2-min injections of the sample over the BSA (reference) and CPB (enzyme) flow cells in series at a flow rate of 10 µL/min. To monitor dissociation, the flow was automatically changed to running buffer after sample injection. The surface was regenerated between cycles by consecutive injections of 8 M urea and 0.05% SDS. For reproducible measurements, it was necessary to introduce a 20-min period between each cycle, presumably to allow for refolding of the enzyme. Recovery. Trypsin. To control the binding, wash, and recovery of BPTI to immobilized trypsin, the method commands INJECT, MICRORECOVER, and BYPASSWASH were used. Inhibitor binding was performed for 8 min. Then the IFC, except the flow cells, was washed with 5% formic acid and ammonium bicarbonate,

Figure 2. Assay for binding and inhibition of trypsin with BPTI: (a) mass spectrum of the model peptide CGRP 8-37 after 10 min of incubation on the trypsin-derivatized sensor chip in the integrated flow cartridge. The sequence of the model peptide with theoretical and observed tryptic cleavage sites is shown to the right. The theoretical tryptic cleavage sites are indicated with vertical arrows, while the observed tryptic fragments are depicted with horizontal bars and corresponding monoisotopic masses below the sequence. (b) SPR sensorgram of 3.3 µM BPTI binding to immobilized trypsin (full line) and control flow cell (dotted line). (c) (left) SPR sensorgram of 3.3 µM BPTI binding to immobilized trypsin and subsequent incubation of the model peptide CGRP 8-37; (right) mass spectrum of the recovered model peptide. (d) (left) SPR sensorgram of 3.3 µM BPTI binding to immobilized trypsin and subsequent recovery; (right) mass spectrum of the recovered inhibitor.

and finally, BPTI was recovered by adding 5% formic acid to the flow cell for 2 min with stopped flow. CPB. Inhibitor binding was performed for 8 min with the method command MS_INJECT using the same dilutions of the samples as in the screen. To wash off unbound protein, the command MS-WASH was used to flush solutions of 8 M urea and HBS through the flow system, but omitting the flow cells. Finally, elution of the bound and washed protein was obtained by injection of 8 M urea over the flow cells and collecting the eluent as controlled by the MS_RECOVER procedure. The commands MS_INJECT, MS_WASH, and MS_RECOVER are optimized versions of the standard commands INJECT, BYPASSWASH, and MICRORECOVER. Mass Spectrometry. Sample Preparation. The peptides recovered from the sensor chip were upconcentrated on a Poros 20 R2 reversed-phase column packed in a Eppendorf gel-loader Analytical Chemistry, Vol. 76, No. 18, September 15, 2004

5245

A Perseptive DE STR mass spectrometer from Applied Biosystems was used for MALDI-TOF MS. Spectra of inhibitors were recorded in linear positive ion mode, with delayed extraction, while reflector positive ion mode was used for model substrates and products.

Figure 3. Screen for inhibitors of CPB in plant extracts. (a) Mass spectrum of a model peptide after 30-min incubation on the CPBderivatized sensor chip in the Surface Prep Unit. In the shown sequence the theoretical CPB cleavage site is indicated with a vertical arrow, and the resulting fragment is depicted with a horizontal bar and the corresponding mass below the sequence. The cleavage product is observed as [M + Na]+ at m/z 1070.5 and as [M + H]+ at m/z 1048.5. (b) (left panel) Overlaid SPR sensorgrams of binding activities in fractions from plant extracts to immobilized CPB. Fractionation: Potato (red trace) filtrate from ultrafiltration with a molecular weight cutoff at 10 000. Egg plant (green trace) pepper (blue trace), and tomato (yellow trace) ammonium sulfate precipitate (0-70% saturation). CI (black trace) is the commercial carboxypeptidase inhibitor from Sigma. (c) Representative mass spectra of a model peptide incubated for 10 min on immobilized CPB after injection of the