Mechanism-based Inhibition of iPLA₂β Demonstrates a Highly Reactive Cysteine Residue (C651) That Interacts with the Active Site: Mass Spectrometric Elucidation of the Mechanisms of Underlying Inhibition

Abstract

The multi-faceted roles of calcium-independent phospholipase A₂β (iPLA₂β) in numerous cellular processes have been extensively examined through utilization of the iPLA₂-selective inhibitor (E)-6-(bromomethylene)-3-(1-naphthalenyl)-2H-tetrahydropyran-2-one (BEL). Herein, we employed accurate mass/high resolution mass spectrometry to demonstrate that the active site serine (S465) and C651 of iPLA₂β are covalently cross-linked during incubations with BEL demonstrating their close spatial proximity. This crosslink results in macroscopic alterations in enzyme molecular geometry evidenced by anomalous migration of the cross-linked enzyme by SDS-PAGE. Molecular models of iPLA₂β constructed from the crystal structure of iPLA₂α (patatin) indicate that the distance between S465 and C651 is approximately 10 Å within the active site of iPLA₂β. Kinetic analysis of the formation of the 75 kDa iPLA₂β-BEL species with the (R) and (S) enantiomers of BEL demonstrated that the reaction of (S)-BEL with iPLA₂β was more rapid than for (R)-BEL paralleling the enantioselectivity for the inhibition of catalysis by each inhibitor with iPLA₂β. Moreover, we demonstrate that the previously identified selective acylation of iPLA₂β by oleoyl-CoA occurs at C651 thereby indicating the importance of active site architecture for acylation of this enzyme. Collectively, these results identify C651 as a highly reactive nucleophilic residue within the active site of iPLA₂β which is thioesterified by BEL, acylated by oleoyl-CoA and located in close spatial proximity to the catalytic serine thereby providing important chemical insights on the mechanisms through which BEL inhibits iPLA₂β and the topology of the active site.

Phospholipases A₂ (PLA₂s) catalyze the hydrolysis of the sn-2 carboxylic ester bond of glycerophospholipids to generate free fatty acids and lysophospholipids, which, together with their downstream metabolites, are involved in lipid metabolic and signaling pathways ^1–5. Eukaryotic PLA₂s are grouped into three main families: secretory, cytosolic, and calcium-independent phospholipases A₂ ⁶. Calcium-independent PLA₂s (iPLA₂s) or patatin-like phospholipases (PNPLA in HUGO nomenclature) are intracellular enzymes that do not require calcium ion for membrane association or catalysis and consist of nine family members (PNPLA1 through PNPLA9) also known as iPLA₂ (β, γ, δ, ε, ζ, η, φ, ι κ) ^7–9. Each of these calcium-independent PLA₂s contain conserved nucleotide-binding (GXGXXG) and lipase (GXSXG) consensus motifs separated by a 10–50 amino acid spacer arm which collectively form a catalytic domain highly homologous to the plant phospholipase/lipase patatin. Previously published studies have demonstrated that calcium-independent phospholipase A₂β (iPLA₂β or PNPLA9) plays an important role in agonist-induced arachidonic acid release, receptor mediated signaling, insulin secretion, apoptosis, cellular proliferation and migration, vascular inflammation, and lysolipid production mediating calcium influx ^10–20.

The mechanism-based inhibitor (E)-6-(bromomethylene)-3-(1-naphthalenyl)-2H-tetrahydropyran-2-one (BEL), was originally developed for studying the catalytic mechanism of chymotrypsin ²¹. In later studies, BEL was demonstrated to be a selective and potent mechanism-based inhibitor of iPLA₂β ^{22, 23}, and has been widely used to distinguish iPLA₂ activities from those of other phospholipases A₂ ^{13, 24–30}. Two plausible mechanisms for iPLA₂ inhibition were considered based on the proposed mechanism for the inhibition of chymotrypsin by BEL ^{1, 21}. In both mechanisms, the inactivation of the enzyme is initiated by nucleophilic attack of the serine hydroxyl on the lactone carbonyl thereby opening of the bromoenol lactone ring of BEL to generate a tethered α-bromomethyl ketone (Scheme 1, Step I) within the active site of the enzyme. This tethered α-bromomethylketone is a powerful alkylating agent that has several possible downstream fates. The α-bromomethyl ketone while tethered to the catalytic serine (S465 in iPLA₂β) is subject to nucleophilic attack by one or more spatially adjacent nucleophilic residues (Step II). Covalent modification of these residues would impair the catalytic activity of the enzyme even after subsequent deacylation of S465 (Step III) due to steric hindrance at the active site. This mechanism is supported by multiple lines of evidence for iPLA₂β, including kinetic assays and autoradiography ^{22, 23}, but the alkylated residue(s) in iPLA₂β were not identified. A second mechanistic possibility is that the enzyme undergoes deacylation immediately after hydrolyzing the bromoenol lactone ring of BEL followed by regeneration of the active site serine to generate a diffusible α-bromomethyl keto carboxylic acid (Step IV). This α-bromomethyl keto carboxylic acid could potentially alkylate any available cysteine residues (including those on the protein surface distal to the active site) that could also potentially result in inactivation of the enzyme. This mechanism was proposed based on the mass spectrometric observation that multiple cysteine residues were covalently modified by the keto carboxylic acid formed by iPLA₂β after BEL treatment ³¹, but it fails to address the fact that the ring-opened diffusible BEL (i.e., the α-bromomethyl keto carboxylic acid) does not inhibit the activity of the enzyme when added exogenously in comparable concentrations in the presence of reducing agents ²³. In early work, we calculated the capture ratio of the ring opened compound using a variety of methods to demonstrate that the majority of inhibition was mediated by covalently bound BEL at the active site in the presence of a reducing agent ²². The reactive α-haloketones that were released by hydrolysis of the acyl enzyme were rapidly inactivated by reducing agents such as DTT in the incubation medium in vitro or by GSH in vivo²².

Scheme 1. Proposed mechanisms for the inactivation of the iPLA₂β by BEL.

I.. Nucleophilic attack of the carbonyl carbon of the bromoenol lactone ring of BEL by the catalytic serine (S465) of iPLA₂β generates the electrophilic α-bromomethyl ketone.

Pathway A:

II. The α-bromomethyl ketone remains tethered to the active site through an acyl link-age. An appropriately positioned nucleophilic residue (within the active site) then attacks the α-carbon of the ring-opened BEL intermediate to eliminate bromide resulting in the stable alkylation of the neighboring nucleophilic residue in iPLA₂β. Hydroxide ion or water can also be the attacking nucleophile (II′).

III. Following deacylation of the active site serine by iPLA₂β catalyzed hydrolysis, the stable α-ketone remains covalently linked to the active site of the enzyme through the adjacent nucleophilic amino acid resulting in inhibition of enzyme activity by the active site.

Pathway B:

IV. The acyl-enzyme intermediate undergoes hydrolysis to release the α-bromomethyl ketone carboxylic acid.

V. The diffusible α-bromomethyl ketone carboxylic acid reacts with cysteine residues on the enzyme, resulting in the covalent modification of the enzyme at sites distant to the active site.

In this study, we analyzed BEL-inactivated iPLA₂β by mass spectrometry, SDS-PAGE and auto-radiography. Remarkably, incubation of iPLA₂β with BEL results in the generation of a rapidly migrating covalently modified species of iPLA₂β on SDS-PAGE. Tandem mass spectrometric analyses of the rapidly migrating band demonstrated a BEL-bridged internal crosslink between the active site serine (S465) and C651 indicating an internally tethered protein with a decreased Stokes radius. Moreover, the BEL crosslink between S465 and C651 demonstrates the close spatial proximity of the active site in iPLA₂β to C651. Importantly, C651 is between the IQ (residues 622–635) and “1-9-14” (residues 694–705) motifs of the calmodulin binding domain in iPLA₂β previously identified ³². Substantiating these results was the marked enantioselectivity and kinetics of iPLA₂β inhibition by (S)-BEL vs. (R)-BEL demonstrated through the extent of covalent modification of iPLA₂β residues following treatment with each enantiomer. Collectively, these results provide important chemical insights into the mechanisms by which BEL inhibits iPLA₂β and the interaction of the enzyme active site with C651 which is present in the calmodulinbinding domain of iPLA₂β.

EXPERIMENTAL PROCEDURES

Materials

(E)-6-(bromomethylene)-3-(1-naphthalenyl)-2H-tetrahydropyran-2-one (BEL) was purchased from Cayman Chemical (Ann Arbor, MI). [³H]-BEL was synthesized and purified as previously described ³³. Antibodies against iPLA₂β epitopes were generated in rabbits prior to purification by immunoaffinity chromatography as described previously ³². Simply Blue^™ Safe Stain and BenchMark^™ Protein Ladder were purchased from Invitrogen (Carlsbad, CA). ECL™ Western Blotting Detection Reagents were from GE Healthcare Bio-Sciences Corp. (Piscataway, NJ). Sequence grade modified trypsin was from Promega (Madison, WI). Most solvents were purchased from Burdick & Jackson (Muskegon, MI). Nano-HPLC columns and pre-columns were obtained from Dionex (Sunnyvale, CA). Most other chemicals were obtained from Sigma (St. Louis, MO).

Recombinant His-tagged iPLA₂β

Recombinant His-tagged iPLA₂β was expressed and purified from Sf9 cells as described previously ³⁴.

Measurement of iPLA₂ β Catalytic Activity

Inhibition of the phospholipase activity of purified recombinant iPLA₂β by BEL was measured through radiolabeled substrate assays (using 1-palmitoyl-2-[1-¹⁴C]arachidonoyl-sn-glycero-3-phosphocholine ³⁵), real-time fluorescence assays (using 2-decanoyl-1-(O-(11-(4,4-difluoro-5,7-dimethyl-4-bora-3a,4a-diaza-s-indacene-3-pro-pionyl)amino)undecyl)-sn-glycero-3-phosphocholine ³⁶), or by electrospray ionization mass spectrometry (using 1-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphocholine ³⁷) as previously described.

Separation of BEL Enantiomers

BEL enantiomers were separated by HPLC using a chiral column as described previously ³⁵. Briefly, a Chirex column comprised of 3,5-dinitrobenzoyl-(R)-phenylglycine attached to a silica matrix was used as the stationary chiral phase. After equilibration with hexane/dichloroethane/ethanol (150:15:1), the BEL enantiomers were injected and eluted isocratically at a flow rate of 2 mL/min. The elution of BEL was monitored by UV absorbance at 280 nm. Peaks corresponding to (S)- or (R)-BEL retention times 19 and 21 min, respectively) were collected, dried under a nitrogen stream and stored at −20°C. The concentrations of (R)- and (S)-BEL were determined by its UV absorbance at 280 nm in acetonitrile (ε₂₈₀ = 6130 cm⁻¹ M⁻¹)³⁵.

Polyacrylamide Gel Electrophoresis of Covalent BEL- iPLA₂β Adducts

Racemic BEL, (R)-BEL and (S)-BEL were stored as solids and redissolved in ethanol immediately before use. After dissolving in ethanol, BEL was injected into 25 mM imidazole buffer, pH 7.8 containing 1 mM DTT and 50 mM NaCl and incubated with purified iPLA₂β for the indicated time periods at room temperature. The ratio of BEL to protein in the reaction varied from 0 to 10. The reactions were terminated by addition of 2 × SDS-PAGE loading buffer (0.125 M Tris, pH 6.8, 4 % SDS (w/v), 20 % (v/v) glycerol, 0.001 % bromphenol blue (w/v), 0.2 M DTT (DL-dithiothreitol)) and the mixture was boiled for 5 min. BEL-protein adducts were electrophoretically separated on 7% SDS-PAGE gels and stained with Coomassie Blue. The resultant gels were scanned with a Kodak Image Station 440 to quantify the protein band intensities using Kodak 1D software.

[³H]-BEL Labeling and Autoradiographic Analysis

Prior to use, [³H]-BEL was purified by HPLC on a reverse phase column (C18, 4.6 mm × 250 mm, Beckman ODS) using a mobile phase of 60% acetonitrile in water at a flow rate of 1 mL/min. The elution of BEL was monitored by UV absorbance at 280 nm. Chloroform and methanol were added to the collected BEL fraction to obtain a final ratio of 1:1:1 (chloroform : methanol : eluted fraction (v/v/v)). Optically pure [³H]-BEL was extracted into chloroform, dried under a nitrogen stream, and redissolved in ethanol. The concentration of BEL was determined by its UV absorbance at 280 nm in acetonitrile (ε₂₈₀ = 6130 cm⁻¹ M⁻¹). Purified [³H]-BEL was incubated with iPLA₂β in a solution of 25 mM imidazole, pH 7.8, 1 mM DTT and 50 mM NaCl at a molar ratio of 3:1 (BEL : protein) at room temperature for 5 min. The reaction was terminated by adding an equal volume of 2× SDS-PAGE buffer. Proteins were separated on 7% SDS-PAGE gels and visualized by Coomassie Blue staining. Following fixation with 40% methanol containing 10% acetic acid, the gels were incubated with a fluorographic reagent (Amplify, Amersham) for 30 min, dried under vacuum and exposed to film.

Covalent Acylation of iPLA₂β with Oleoyl-CoA

Purified iPLA₂β (5 μM) was incubated in the presence or absence of oleoyl-CoA (50 μM) for 60 min at 30°C. Protein was precipitated with CHCl₃/CH₃OH as previously described ³⁸, dried using a Speed-Vac, and solubilized in 25 mM NH₄HCO₃ containing 0.1% RapiGest^™ (Waters, Milford, MA). Trypsin was added to the samples at a ratio of 1:30 (w/w) followed by incubation at 37°C for two hours. After termination of the trypsinolysis by addition of trifluoroacetic acid (0.5% final concentration), the samples were further incubated at 37°C for 45 min and centrifuged at 13,000 rpm for 15 min. The supernatant was removed and the pellet was resuspended in 70% isopropanol by a combination of vortexing and sonication. Both the supernatant and the pellet fractions were analyzed by nano-LC/MS/MS. Processed samples were applied to a PepMap C4 precolumn (300 μm × 1 mm, Dionex, Sunnyvale, CA) using a Surveyor autosampler (ThermoFisher Scientific, Waltham, MA) and washed with 0.1% formic acid for 5 min. A nonlinear gradient from 90% mobile phase A (0.1% formic acid in water) and 10% mobile phase B (50% acetonitrile/50% isopropanol, 0.1% formic acid) to 10% A and 90% B was applied to elute the tryptic peptides from the precolumn onto a reverse-phase C4 analytical column (75 μm × 15 cm, Dionex, Sunnyvale, CA) at a flow rate of 250 nL/min. Peptides eluting from the C4 column were introduced into an LTQ-Orbitrap mass spectrometer (ThermoFisher Scientific, Waltham, MA) using a TriVersa^™ Nanomate system (Advion, Ithaca, NY) with a spray voltage of 1.7 kV. The mass spectrometer was operated in a data-dependent acquisition mode with dynamic exclusion enabled allowing the five most intense ion peaks from the full mass scan to undergo product ion mass analysis. The instrument was calibrated externally. Full mass scans were performed by the Orbitrap mass analyzer with a mass resolution of r = 30,000 at m/z 400 Thomson. The lock mass option was enabled to ensure a mass accuracy of < 5 ppm (utilizing the m/z value of the protonated diisooctyl phthalate (391.2843) as the lock mass). Product ion mass spectra were acquired either by the ion trap or the Orbitrap mass analyzer with a mass resolution of 15,000 at m/z 400 Thomson. For the product ion analyses, a normalized collision energy of 30% was applied, the activation time was set at 30 ms with the activation parameter q = 0.250, and precursor ions were isolated within the range of 2 to 4 Thomson. The acquired data was searched against a customized protein sequence database using both SEQUEST (Bioworks™, ThermoFisher Scientific, Waltham, MA) and MASCOT algorithms (Matrix Science, London, UK). Positive search results were manually confirmed.

Western Blotting

Proteins were separated on 7% SDS-PAGE gels and transferred to immobilon-P membranes in 10 mM CAPS buffer (pH 11) containing 10% methanol. Powdered milk (5% (w/v)) was used to block the membranes prior to incubation with primary antibody. Horseradish peroxidase-linked secondary antibody was used in combination with an ECL detection system to visualize immunoreactive bands.

In-gel Digestion of Proteins for Mass Spectrometric Analyses

Protein bands visualized with Coomassie Blue were excised from the gel, cut into small pieces (~ 1 mm³), and destained twice by vortexing for 10 min in 100 μL of 25 mM ammonium bicarbonate containing 50% acetonitrile. The gel pieces were dried utilizing a Speed Vac and then rehydrated in 25 mM ammonium bicarbonate, pH 8 containing trypsin (12.5 μg/mL) on ice for 45 min. Excess trypsin solution was then removed and a small amount of 25 mM ammonium bicarbonate, pH 8 was added to cover the gel pieces. Following overnight incubation at 37°C, 100 μL of 50% acetonitrile containing 5% formic acid were then added and the resultant tryptic peptides were extracted by vigorous vortexing and sonication. The extraction step was repeated once and the peptide extracts were combined and evaporated to ~ 10 μL using a Speed Vac. The peptide extract was diluted in 10% acetonitrile containing 0.1% formic acid before analysis by either MALDI-TOF or ESI-LC/MS/MS.

In-solution Digestion of Protein for Mass Spectrometric Analyses

Native or BEL-treated protein was precipitated by chloroform and methanol as described previously ³⁸. Briefly, four volumes of methanol and one volume of chloroform were added to the protein solution. After vortexing, three volumes of water were then added and the mixture was centrifuged at 14,000 × g for 15 min. The upper phase was discarded and replaced with three volumes of methanol. Proteins were recovered as a pellet following centrifugation for 15 min at 14,000 × g, and re-constituted into 0.1% RapiGest^™ in 25 mM NH₄HCO₃. Trypsin was added at a ratio of 1:30 (enzyme to protein, w/w) and the samples were incubated at 37°C for two hours. Trifluoroacetic acid was then added to a final concentration of 0.5%. The samples were further incubated at 37°C for 45 min and centrifuged at 13,000 rpm for 15 min. The supernatant was transferred to a new tube, and the pellet was resuspended into 70% isopropanol by vortexing and intermittent sonication. Both the supernatant and the pellet fractions were analyzed by mass spectrometry.

Analysis of the Intact Protein by MALDI-TOF

iPLA₂β was incubated with BEL as described above. Reactions were terminated by addition of 0.1% TFA (final pH=2) followed by overnight dialysis (Slide-A-Lyzer MINI Units, MWCO 7000, Pierce, Rockford, IL) against one liter of water at 4°C. Samples were then mixed with a matrix solution (5 mg/ml α-CHCA in 50% acetonitrile and 0.1% TFA) and spotted onto a MALDI target plate. Mass spectra were acquired in the positive-ion linear mode on a 4700 Proteomics Analyzer (MALDI-TOF/TOF) with a Nd:YAG laser at a 200 Hz repetition rate. BSA was added as an internal standard for calibration.

ESI-LC/MS/MS Analysis

Tryptic peptides were loaded onto a PepMap C4 precolumn (300 μm × 1 mm, Dionex, Sunnyvale, CA) by a Surveyor autosampler (ThermoFisher Scientific, Waltham, MA). The precolumn was then washed with 0.1% formic acid for 5 min. A nonlinear gradient from 90% of mobile phase A (0.1% formic acid in water) and 10% of mobile phase B (50% acetonitrile/50% isopropanol, 0.1% formic acid) to 10% of A and 90% of B was applied to elute the peptides from the precolumn onto a reverse-phase C4 analytical column (75 μm × 15 cm, Dionex, Sunnyvale, CA) at a flow rate of 250 nL/min. The eluted peptides were introduced into an LTQ-Orbitrap mass spectrometer (ThermoFisher Scientific, Waltham, MA) by a TriVersa^™ Nanomate system (Advion, Ithaca, NY) at a constant voltage of 1.7 kV. The mass spectrometer was operated in a data-dependent acquisition mode with dynamic exclusion enabled, i.e. each full mass scan in the Orbitrap was followed by tandem mass scans (product ion analysis) of the five most intense peaks from the full mass spectrum in the ion trap. In selected cases, the most intense ion peak from the peptide fragmentation was selected for further fragmentation to provide additional sequence information (i.e., MS³ analyses). The instrument was calibrated externally following the manufacturer’s instructions. Full mass spectra were acquired in the Orbitrap mass analyzer with a target mass resolution of r = 30,000 at m/z 400 Th. MS² spectra were acquired either in the ion trap or in the Orbitrap mass analyzer with a target mass resolution of r = 15,000 at m/z 400 Th. For MS² and MS³ analyses, the normalized collision energy of 25% was applied, the activation time was set at 30 ms with an activation parameter q = 0.25, and precursor ions were isolated within the range of 2 and 4 Th. AGC target settings for FT full mass scan, FT MSⁿ scan and ion trap MSⁿ scan were 5 × 10⁵, 2 × 10⁵ and 1 × 10⁴, respectively. The maximum injection time for MSⁿ scan (FT and ion trap) and FT full mass scan were 1 s and 500 ms, respectively. The acquired data were searched against a customized protein sequence database using both SEQUEST and Mascot algorithms ^{39, 40}. Positive search results were individually confirmed.

RESULTS

Incubation of BEL with iPLA₂β Results in the Production of Two Resolvable Bands by SDS-PAGE that Each Contain BEL-linked Covalent Adducts

Intriguingly, analysis of iPLA₂β incubated with racemic BEL by SDS-PAGE revealed the presence of two bands visualized by Coomassie Blue staining, whereas only one band was present in the untreated iPLA₂β sample (Fig. 1A). The upper band of the BEL-treated enzyme migrated similarly as the native iPLA₂β displaying a molecular weight of ~ 80 kDa. The lower band migrated more rapidly exhibiting an apparent molecular weight of 75 kDa (i.e., approximately 5 kDa less than the upper band).

Figure 1. SDS-PAGE autoradiographic analyses of BEL-treated iPLA₂ β.

A, Resolution of BEL-treated iPLA₂β into two bands by SDS-PAGE. iPLA₂β was incubated with racemic BEL at 22°C for 10 min and electrophoresed on a 7% SDS-PAGE gel. The upper band has an apparent molecular weight similar to native iPLA₂β (80 kDa) while the lower band has an apparent molecular weight of approximately 75 kDa. B, Radiolabeling of iPLA₂β with racemic [³H]-BEL. iPLA₂β was incubated with racemic [³H]-BEL at room temperature for 5 min followed by separation on a 7% SDS-PAGE gel stained with Coomassie Blue (left) followed by visualization of the radiolabeled bands by autoradiography (right). Two bands are present in the lane containing [³H]-BEL-treated iPLA₂β indicating that both bands are modified by BEL. “−“ indicates protein treated with ethanol vehicle alone; “+” indicates protein treated with [³H]-BEL.

To determine whether one or both bands of the doublet observed by SDS-PAGE were covalently modified by BEL, we incubated iPLA₂β with racemic [³H]-BEL, subjected the treated enzyme to SDS-PAGE, and analyzed the gel by autoradiography. An identical doublet pattern was observed after visualization by autoradiography as that present on the Coomassie blue-stained SDS-PAGE gel (Fig. 1B) indicating that [³H]-BEL was incorporated into both bands in similar amounts. Since the incorporation of radiolabeled BEL survives boiling under denaturing conditions and subsequent SDS gel electrophoresis, each band likely contains covalently bound BEL. These results suggest that the incubation of iPLA₂β with BEL yields at least two covalently modified products which can be resolved by SDS-PAGE.

Western Blotting and MALDI-TOF Mass Spectrometry Demonstrate that the 75 kDa Band is Intact iPLA₂β

Since SDS-PAGE and autoradiographic analyses showed that the 75 kDa band had an apparent molecular weight which was 5 kDa less than the native iPLA₂β protein and was covalently modified by BEL, we postulated that the molecular weight difference could potentially result from cleavage of the iPLA₂β protein backbone, perhaps through an unusual proteolytic or protein splicing event that occurred following reaction of the enzyme with BEL. To determine whether the 75 kDa band was the product of a BEL-mediated cleavage reaction, we performed Western blotting using antibodies directed against the N-terminus, middle and C-terminus of iPLA₂β and measured the masses of the protein products by MALDI-TOF mass spectrometry. Western blotting demonstrated that the 75 kDa band contained the all identifiable epitopes present in the 80 kDa native protein (Fig. 2B). Thus, the de novo appearance of the 75 kDa band after incubation with BEL did not result from proteolytic cleavage and contains the C-terminus, the N-terminus as well as the internal residues 277–295 and 506–525.

Figure 2. Western blot analysis of BEL-treated iPLA₂β utilizing antibodies directed against four different regions of the iPLA₂β protein.

Epitopes for the four antibodies are indicated in Panel A. iPLA₂β was incubated with (S)-BEL at 22°C for 3 min. Following separation on a 7% SDS-PAGE gel, the protein bands were transferred to Immobilon-P PVDF membranes and probed separately with the four different antibodies against iPLA₂β in Panel B. Both native and BEL-treated iPLA₂β are immunoreactive to all four antibodies. “−“ indicates the control sample and “+” indicates the BEL-treated sample.

To confirm the structural integrity of iPLA₂β, racemic BEL-treated iPLA₂β was directly analyzed by MALDI-TOF mass spectrometry in the linear mode. A single broad peak with a mass of 82,931 Th is observed in the mass spectrum of the native iPLA₂β, which agrees well with its predicted mass (83,860 Th) (Fig. 3A). The peak is broad because the resolution of MALDI-TOF MS in the linear mode using this instrument is not sufficient to resolve the isotopologues of an 80 kDa protein or multiple protein species with small mass shifts caused by common protein modifications (e.g., deamidation, acetylation, phosphorylation, etc.). A peak with a similar shape and average mass is observed in the mass spectrum of the racemic BEL-treated iPLA₂β (Fig. 3B). Close examination of these mass spectra revealed that the peak of BEL-treated iPLA₂β is 318 Da greater in mass than the native protein, suggesting the covalent linkage of a single BEL-derived moiety to iPLA₂β. However, since the observed peak represents a mixture of native and the BEL-modified iPLA₂β and is broad due to the relatively low mass resolving power of the instrument for the small mass shift (318 Da) of the native protein, the observed mass shift does not necessarily reflect the exact mass difference caused by the modification. However, clearly, no peak with a mass ≈ 5 kDa less than the predicted peak is present in the mass spectrum. Since the mass resolution of MALDI-TOF MS in the linear mode should be sufficient to distinguish two molecular species with a low kDa mass difference, the absence of a peak with a 5 kDa lower mass suggests that the more rapidly migrating band is due to alterations in the conformation of BEL-treated iPLA₂β that either modifies its interaction with the polyacrylamide matrix or changes its effective Stokes radius and does not result from cleavage of the peptide backbone of the protein.

Figure 3. MALDI-TOF mass spectrometric analyses of native and BEL-treated iPLA₂β holoproteins.

Native (A) and BEL-treated iPLA₂β (B) were mixed with a MALDI matrix solution (5 mg of α-CHCA in 50% acetonitrile/0.1% TFA), and spotted onto a MALDI target plate. Bovine serum albumin was added as a molecular mass marker and the acquired mass spectra were calibrated using its monomeric and dimeric masses. A single peak corresponding to iPLA₂β is observed in the BEL-treated sample and there is no indication of the presence of a peak possessing a 5 kDa mass shift. These results demonstrate that the two bands observed by SDS-PAGE of the BEL-treated iPLA₂β have similar molecular weights.

Both S465 and C651 of iPLA₂β Are Covalently Modified by BEL

To identify the residues in iPLA₂β which are covalently modified by the mechanism-based suicide inhibitor BEL, we first incubated racemic BEL with purified recombinant iPLA₂β and analyzed the tryptic peptides by LC-MS/MS. Importantly, these experiments were performed in the presence of DTT to scavenge the reactive α-bromomethyl ketone derivative of BEL which upon release from the catalytic serine may diffuse out of the active site and non-specifically modify distant iPLA₂β nucleophilic residues. In the BEL-treated sample, two unique tryptic peptides were detected as triply-charged ions at m/z 880.4606 Th and 908.4530 Th, respectively. Their monoisotopic masses are 254.0939 Da greater than the expected values for the iPLA₂β peptides ⁴⁵⁶DLFDWVAGTSTGGILALAILHSK⁴⁷⁸ (D456-K478) and ⁶⁴⁴SPQVPVTCVDVFRPSNPWELAK⁶⁶⁵ (S644-K665) at a mass accuracy of 3.5 ppm (Fig. 4, Panels A and D). Collision-induced dissociation (CID) of each peptide (generating MS² spectra) revealed the presence of ions that closely correlated with those predicted for the unmodified tryptic peptides as well as the presence of unique peaks having masses 254 Da greater than anticipated.

Figure 4. Mass spectral analysis of the tryptic peptides of BEL-modified iPLA₂β identifying S465 and C651 as sites of modification.

Purified recombinant iPLA₂β was incubated with racemic BEL at a molar ratio of 1:1 at 22°C for 3 min. The reaction was terminated by addition of CHCl₃/CH₃OH, vortexing, and centrifugation to pellet the precipitated protein. The protein pellet was solubilized in buffer containing RapiGest^™, digested with trypsin, and analyzed by LC/MS/MS as described in “Experimental Procedures”. A and D, Full mass spectra showing the presence of unique peaks in the BEL-treated sample in the mass range of m/z 878–887 and 905–915, respectively. B, CID mass spectrum (MS²) of the ion at m/z 880.36 (3+) which corresponds to the BEL-modified tryptic peptide, ⁴⁵⁶DLFDWVAGTSTGGILALAILHSK⁴⁷⁸. C, CID mass spectrum (MS³) of the ion at m/z 913. E, CID mass spectrum (MS²) of the ion at m/z 908.36 (3+) which corresponds to the BEL-modified tryptic peptide, ⁶⁴⁴SPQVPVTCVDVFRPSNPWELAK⁶⁶⁵. F, CID mass spectrum (MS³) of the ion at m/z 1156. MS² and MS³ mass spectra in combination with high mass accuracy (< 5 ppm) identify that S465 and C651 are modified by an α-keto-substituted carboxylic acid leading to a mass increase of 254 kDa.

In the MS² spectrum of the putative modified peptide D456-K478, all of the observed y ions (y₁₅ – y₂₁) exhibited increased masses (i.e., 254 Da greater than the expected mass values) in comparison to those anticipated for the unmodified tryptic peptide. In contrast, the masses of all observed b ions (b₄ - b₇, and b₉) matched well with the expected values. These results indicate that the modified residue is located in the sequence of ⁴⁶⁵STGGILALAILHSK⁴⁷⁸ (Fig. 4B). To determine which residue within this sequence was covalently modified, MS³ experiments were performed. Fragmentation of the ion at m/z 933.01 Th (present in the MS² spectrum and assigned as the doubly-charged y₁₇ ion) demonstrated the anticipated masses in the MS³ spectrum indicating that the modification site is present in the sequence ⁴⁶²AGTS⁴⁶⁵ (Fig. 4C). Collectively, the MS² and MS³ spectra identify S465, the only overlapping residue between the sequences ⁴⁶²AGTS⁴⁶⁵ and ⁴⁶⁵STGGILALAILHSK⁴⁷⁸, as the site of covalent modification by BEL.

Next, the ion at m/z 908.4530 Th representing the putative modified peptide S644-K665 was investigated further by MS² and MS³ fragmentation studies. Analysis of the MS² spectrum revealed the presence of y₁₆ – y₂₀ ions that each possessed a mass increase of 254 Da relative to the expected values for the unmodified peptide. These results indicate that the addition of BEL occurred in the sequence ⁶⁵⁰TCVDVFRPSNPWELAK⁶⁶⁵. Notably, the b₃, b₄, b₆, b₇, y₁₁, y₁₂ and y₁₄ ions do not exhibit any shifts in mass. Collectively, these results demonstrate that C651 is the modified residue (Fig. 4). To substantiate this conclusion, we acquired the MS³ spectrum of the ion at m/z 1157.24 Th (Fig. 4). The b* and y ions observed in this MS³ spectrum unambiguously identify C651 as the site of modification. Collectively, LC-MS/MS analysis of the tryptic digest from BEL-treated iPLA₂β demonstrated that both S465 and C651 were modified upon BEL treatment leading to a mass shift of 254 Da.

Our finding that C651 of iPLA₂β is modified after treatment with BEL is in accordance with the results of Song et al., which demonstrated that several cysteine residues were modified by BEL in the absence of DTT and that C651 was the only modified cysteine residue in the presence of 1 mM DTT ³¹. In contrast to the present results, no modification of the lipase consensus sequence motif, GTSTG, was detected in those experiments in either the absence or presence of DTT.

S465 and C651 Are Covalently Linked by a Bidentate BEL Adduct

Since both S465 and C651 of iPLA₂β are covalently modified by BEL and at least one product of the reaction of iPLA₂β with BEL migrates more rapidly than the native protein on an SDS-PAGE gel, we hypothesized that S465 and C651 were cross-linked by BEL as a covalent bridge as proposed in the inhibition mechanism (Scheme 1, Pathway A). To investigate whether S465 and C651 were connected through a BEL-derived covalent adduct, we analyzed the tryptic digest of BEL-treated iPLA₂β by LC-MS/MS specifically searching for the species corresponding to the cross-linked product of the S465-containing tryptic peptide (D456-K478, designated as peptide α) and the C651-containing tryptic peptide (S644-K665, designated as peptide β) through C₁₆H₁₂O₂ (the formula of the BEL-derived product formed after the lactone of BEL has been hydrolyzed with subsequent elimination of the bromide). A candidate peptide was thus identified with a mass which matches that of the putative cross-linked species. This peptide was detected as an ion either with a +5 charge (the peak at m/z 1018.7330 Th) or a +6 charge (the peak at m/z 849.1084 Th) at a mass accuracy of < 3.5 ppm (Fig. 5A). The tandem mass spectra (MS² spectra) of these two ions were then acquired in the Orbitrap mass analyzer at a resolution of 15,000 at m/z 400 Th to verify the peptide sequence and to determine which residues were covalently modified (Fig. 5B & 5C). Since the high mass accuracy product ion spectra were acquired in the Orbitrap, the charge states and the m/z values of the product ions were determined with high confidence levels ensured by the high resolution and high mass accuracy of the FT data. In the MS² spectra, y and b ions corresponding to the fragments from peptide α (y_3a through y_6a, y_8a, y_9a, y_11a through y_13a ions and b_3a through b_9a ions) and peptide β (y_11β through y_14β ions, b_4β, b_6β and b_7β ions) were observed, indicating that these two tryptic peptides are present in the putative cross-linked adduct. Peaks assigned as y_14a through y_20a ions all displayed a mass shift corresponding to the addition of peptide β and a fragment with a mass of 236. Correspondingly, y_16β, and y_18β through y_20β ions all display a mass shift corresponding to the addition of peptide α and a fragment with a mass of 236 (Fig. 5B & 5C). Collectively, the assigned fragments from the MS² mass spectra (Table I) demonstrate that S465 and C651 are cross-linked through a covalent derivative of BEL with the formula C₁₆H₁₂O₂. The proposed structure of the cross-linked peptides and the identified fragments are illustrated in Fig. 5D. Based on the structure of the cross-linker, we estimate that the maximum distance between the sulfur of C651 and the oxygen of S465 is 9 Å.

Figure 5. Mass spectrometric evidence for the cross-linking of S465 and C651.

Trypsinized iPLA₂β treated with (S)-BEL was analyzed by ESI-LC/MS/MS as described in “Experimental Procedures”. A, The survey full mass spectrum acquired using the Orbitrap mass analyzer with a resolution r = 30,000 at m/z 400 shows that the cross-linked peptide is detected as ions in two different charge states. The peaks at m/z 849.1113 and 1018.7330 represent the monoisotopic ions of the +6- or +5- charged cross-linked peptide, respectively. Each of the ions were fragmented by CID in the ion trap, and the fragments were analyzed in the Orbitrap mass analyzer with a resolution r = 15,000 at m/z 400. B, the tandem mass spectrum (MS² spectrum) of the ion at m/z 1018.7330 Th. C, the tandem mass spectrum (MS² spectrum) of the ion at m/z 849.1113 Th. D, the structure and the identified fragments of the cross-linked peptides. The S465-containing tryptic peptide (α) and C651-containing tryptic peptide (β) are cross-linked by a link-age between S465 and C651 via a BEL hydrolytic product.

Table 1.

Fragment Ions Identified in the Tandem Mass Spectra of the Cross-linked Peptide

Fragments from Peptide α (456 – 478)			Fragments from Peptide β (644 – 665)
Ion Type	Charge State	m/z	Ion Type	Charge State	m/z
b3α	1	376.1868	b4β	1	412.2192
b4α	1	491.2147	b6β	1	608.3403
b5α	1	677.2937	b7β	1	709.3886
b6α	1	776.3622	b10β	4	912.4721
b7α	1	847.3981	y11β	2	672.8599
b8α	1	904.4179	y12β	2	722.3905
b9α	1	1005.4678	y13β	2	779.9047
y3α	1	371.2044	y14β	2	829.4113
y4α	1	484.2885	y16β	4	1121.3323
y5α	1	597.3713	y16β	5	897.2646
y6α	1	668.4098	y18β	4	1170.3571
y8α	1	852.5314	y18β	5	936.4869
y9α	1	965.6143	y19β	4	1195.1229
y11α	1	1135.7209	y20β	5	981.9154
y12α	1	1192.7418
y12α	2	596.8738
y13α	1	1293.7811
y13α	2	647.3988
y14α	5	817.8368
y16α	4	1061.5631
y17α	4	1079.3209
y17α	5	863.6592
y18α	4	1104.0878
y18α	5	883.4738
y19α	4	1150.6109
y19α	5	920.6881
y20α	4	1179.6131
y20α	5	943.6931
y21α	5	973.1121
y21α	6	811.0901

Peptide α: ⁴⁵⁶DLFDWVAGTSTGGILALAILHSK⁴⁷⁸

Peptide β: ⁶⁴⁴SPQVPVTCVDVFRPSNPWELAK⁶⁶⁵

Differential Kinetics of Covalent Modification of S465 and C651 by (S)-BEL and (R)-BEL

Previously, it was demonstrated that BEL could be chromatographically resolved into its (S)- and (R)-enantiomers by chiral HPLC, and that (S)-BEL selectively inhibits iPLA₂β activity in comparison to (R)-BEL (17). To investigate whether (S)-BEL may be selective in the covalent modification of S465 and C651, iPLA₂β was treated with either (S)- or (R)-BEL, and analyzed by SDS-PAGE and ESI-LC/MS/MS. The results demonstrated that treatment of iPLA₂β with either (S)- or (R)-BEL produced similar changes in iPLA₂β electrophoretic behavior as racemic BEL and that both are able to modify S465 and C651.

To further scrutinize differences in enantioselectivity, we performed detailed kinetic analyses by incubating iPLA₂β with (S)- or (R)-BEL for various time intervals and comparing the ratio of the two bands formed by Western analysis and correlating this to covalent alterations determined by mass spectrometry. As shown in Fig. 6A and 6B, incubations with (S)- BEL or (R)-BEL both result in a similar kinetic profile of the formation of the doublet pattern following SDS-PAGE. Notably, the intensity of the upper band decreases whereas the intensity of the lower band increases with incubation time. However, the rate at which the intensity of the lower band increases is markedly faster for (S)-BEL-treated iPLA₂β than for the (R)-BEL-treated sample. For (S)-BEL, the intensity of the 75 kDa band was nearly equal to that of the 80 kDa band after only two minutes incubation (Fig. 6A), but for (R)-BEL, the intensity of the 75 kDa band increases at a slower rate. The 80 kDa band was still dominant after 60 min incubation in the presence of (R)-BEL (Fig. 6B). To quantitatively determine the time-dependence of enantioselective alterations, we measured the net intensity of each band by densitometry, normalized that value to the total net intensity of both bands in the same lane (relative intensity) and plotted the relative intensity of each band as a function of incubation time. As shown in Fig. 6C, the intensity of the 75 kDa band increased rapidly in the presence of (S)-BEL, reached a similar intensity as the 80 kDa band within two minutes, and remained at this level for the remaining incubation time examined. In contrast, for (R)-BEL, the intensity of the 75 kDa band increased slowly, and even after 60 min incubation represented only 35% of the total intensity present in the 80 kDa band.

Figure 6. Time course assay analysis of the reaction of (S)-BEL and (R)-BEL with iPLA₂β as determined by SDS-PAGE and densitometry.

Purified recombinant iPLA₂β was incubated with (S)-BEL (A) and (R)-BEL (B) at 22°C for various time intervals as indicated. The reactions were terminated by adding an equal amount of SDS-PAGE sample loading buffer (2×) and the iPLA₂β adducts were separated on a 7% SDS-PAGE gel prior to visualization by Coomassie Blue staining. C, the net intensity of each protein band from A and B were measured utilizing a Kodak Image Station, normalized to the total intensity of the protein bands present in the same lane, and expressed as a percentage of the total intensity. The relative intensity of each protein band was then plotted against the incubation time. Solid circles, 75 kDa band of iPLA₂β following incubation with (S)-BEL; Open circles, 80 kDa band of iPLA₂β following incubation with (S)-BEL; Solid squares, 75 kDa band of iPLA₂β following incubation with (R)-BEL, Open squares, 80 kDa band of iPLA₂β following incubation with (R)-BEL. All experiments were performed in duplicate.

In addition, we also assessed the levels of the two modified tryptic peptides and the cross-linked peptide in (S)-BEL- and (R)-BEL-treated samples by mass spectrometry through a semi-quantitative method using the relative peak intensities of known peptides. In this method, the peak area of a target peptide is normalized against the average peak area of a group of representative iPLA₂β peptides that serve as endogenous internal standards, and the resultant number (normalized relative abundance) represents the level of the target peptide in a sample. Comparison of the normalized relative abundance of a target peptide in different samples would provide insight into how different reaction conditions affect the level of this peptide. The normalized peak areas of the two modified peptides and the cross-linked peptide from (S)-BEL- or (R)-BEL-treated samples were calculated as described in “Experimental Procedures” and depicted in Fig. 7.

Figure 7. Comparison of the extent of iPLA₂β modification by (S)-BEL and (R)-BEL.

An equal amount of purified iPLA₂β was incubated with either (S)-BEL or (R)-BEL at a molar ratio of 1:1 at 22°C for 3 min. The reaction was terminated by precipitating the protein with chloroform/methanol. The protein pellet was solubilized in buffer containing RapiGest^™, digested with trypsin, and analyzed by LC/MS/MS as described in “Experimental Procedures”. The normalized relative abundance of a target peptide was calculated as described in “Experimental Procedures”. A, the normalized relative abundance of the BEL-modified S465-containing tryptic peptide (D456-K478) in (R)-BEL- or (S)-BEL-treated samples. B, the normalized relative abundance of the BEL modified C651-containing tryptic peptide (S644-K665) in (R)-BEL- or (S)-BEL-treated samples. C, the normalized relative abundance of the cross-linked peptide in (R)-BEL or (S)-BEL-treated samples. The results indicate that treatment with (S)-BEL results in a higher level of protein modification. By examining the changes in band intensity following SDS-PAGE in the time-course experiment and comparing the levels of modified S465 and C651 as determined by ESI-LC/MS/MS, we found that (S)-BEL is more efficient at covalent modification of iPLA₂β than (R)-BEL. These observations are in good agreement with the fact that (S)-BEL is a selective inhibitor of iPLA₂β, and demonstrate the enantioselectivity of (S)- vs. (R)-BEL in the kinetics of covalent modification of iPLA₂β leading to inhibition of enzymic activity.

As shown in the figure, the normalized peak area of the adducted S465-containing peptide is ten times higher in the (S)-BEL-treated sample than in the (R)-BEL treated sample (Fig. 7A, 0.35 vs. 0.035). Similarly, a nine-fold difference was observed for the modified C651-containing peptide and four-fold difference for the cross-linked peptide from the (S)-BEL-treated sample relative to the (R)-BEL-treated iPLA₂β (Fig. 7B, 0.009 vs. 0.001 and 0.30 vs. 0.08). This difference likely reflects the stereoelectronic characteristics and interactions of (S)-BEL and (R)-BEL at the active site and their differential ability to be catalytically cleaved to modify iPLA₂β.

C651 is Acylated in vitro by Oleoyl-CoA

Previously, we demonstrated that purified recombinant iPLA₂β possesses palmitoyl-CoA hydrolase activity ³⁴ in addition to its well-characterized lysophospholipase and phospholipase A₂ activities ^{41, 42}. Importantly, the acyl-CoA thioesterase activity of iPLA₂β is completely dependent upon the presence of the catalytic serine (S465) of the enzyme ³⁴. Moreover, iPLA₂β is covalently autoacylated in a highly substrate-specific manner (by oleoyl- but not palmitoyl-CoA), which occurs at a second site distinct from the hydrolytic lipase site (GXSXG) ³⁴. Therefore, iPLA₂β modulates multiple metabolic alterations resulting from the production of lipid 2nd messengers (e.g., arachidonic acid and lysolipids) or alternatively through hydrolysis or transacylation of saturated acyl-CoAs from specific cellular membrane compartments.

Since localization of the acylation site would provide insight into the mechanism of acyl-CoA hydrolysis and iPLA₂β structure, we incubated iPLA₂β in the presence and absence of oleoyl-CoA followed by precipitation, solubilization, and trypsinolysis of the protein prior to analysis of the resultant peptides by LC/MS/MS as described in “Experimental Procedures.” A peptide with a mass 264 Da greater than the expected mass of ⁶⁴⁴SPQVPVTCVDVFRPSNPWELAK⁶⁶⁵ (Fig. 8A) was identified indicative of the addition of an oleoylate moiety to the peptide. The CID mass spectrum (MS²) of this peptide revealed several fragments whose m/z values correspond well with the b₃ and b₄ ions predicted for the unmodified tryptic peptides as well as those whose m/z values exhibit a mass increase of 264 Th greater than the predicted values for the unmodified peptides (y₁₅, y₁₆, y₁₈, y₁₉ and y₂₀ ions). The MS² spectrum of the S644-K665 peptide confirms its sequence identity, but does not provide sufficient information for us to determine the modification site (Fig. 8B). Therefore, we performed MS³ analysis of the ion at m/z 1161.8 from the MS² spectrum to identify the residue modified by oleic acid. Because the m/z values of all the y ions (y₉ – y₁₄) match the predicted values for the unmodified peptide while the ion at m/z 1161.8 represents a modified fragment with a mass increase of 264 Th greater than the control fragment, we concluded that C651 is modified by an oleate group (Fig. 8C). Collectively, these results identify C651 as a highly reactive nucleophilic residue which is in close spatial proximity with the catalytic serine of iPLA₂β (S465) and is a target for both alkylation by the α-bromomethyl ketone of BEL as well as acylation by oleoyl-CoA.

Figure 8. Identification of Covalent Oleoylation of C651 in iPLA₂β by Oleoyl-CoA.

A, Full scan mass spectrum (m/z 350 – 1600) at 112.4 min of the LC-MS/MS analysis of the tryptic digest from iPLA₂β treated with oleoyl-CoA in vitro. A triply-charged ion of m/z 911.8425 was observed with a resolution power of 30000 at m/z 400. The calculated mass of this ion (2732.5041 Th) matches the mass of the oleoylated ⁶⁴⁴SPQVPVTCVDVFRPSNPWELAK⁶⁶⁵ with a mass accuracy of 3 ppm. B, the ion of m/z 911.8425 was selected for CID (collision-induced dissociation) fragmentation. The resulting fragments (y and b ions as labeled) confirm the sequence of the peptide as ⁶⁴⁴SPQVPVTCVDVFRPSNPWELAK⁶⁶⁵ and localize the modification site as C651. C, the ion of m/z 1161.8 (the base peak in panel B) from the fragment of 911.8425 was selected for further fragmentation by CID (MS³). The fragmentation pattern of this ion substantiates the identity of the oleoylated peptide and the modification site.

DISCUSSION

Since its initial discovery as a selective and potent mechanism-based inhibitor for iPLA₂β, BEL has been widely used to explore the multiple biologic functions of iPLA₂β-mediated lipid 2nd messenger generation, calcium signaling, and cellular growth, proliferation and differentiation mediated by this enzyme ^{2–5, 8, 10, 15, 35, 43, 44}. Thus, a detailed understanding of the mechanism through which iPLA₂β is inactivated by BEL, the covalent adducts which are formed, and the fate of the highly reactive α-bromomethylketone under different redox conditions is central to defining the specificity and selectivity of this inhibitor. Accordingly, we utilized high mass accuracy mass spectrometry in conjunction with complimentary protein analytic techniques to identify three distinct BEL adducts of iPLA₂β including a covalently cross-linked product demonstrating the close spatial proximity of the catalytic active site serine (S465) to C651 through a BEL-tethered adduct linking these residues. Examination of the enantioselective mechanism-based inhibition of iPLA₂β with (R)-BEL and (S)-BEL in the current study revealed that the kinetics of covalent modification of iPLA₂β with each of the BEL enantiomers was consistent with our previous observations demonstrating an order of magnitude difference in the potency of (S)-BEL to inhibit iPLA₂β catalytic activity vs. (R)-BEL ³⁵. The nucleophilic reactivity of C651 within the active site was further established by the identification of this residue as a site of covalent fatty acylation of iPLA₂β by oleoyl-CoA as we previously reported ³⁴.

Although the mechanism of inhibition of iPLA₂s by BEL has been proposed to be similar to that previously elaborated by Katzenellenbogen et al. for inhibition of chymotrypsin by haloenol lactone suicide substrate inhibitors ^{21, 45–48}, important questions remained regarding the chemical identity of reaction intermediates and the modified residues of iPLA₂β. Nucleophilic attack of the carbonyl of the BEL bromoenol lactone ring by the catalytic serine (S465) of iPLA₂β results in the formation of a highly reactive α-bromomethyl ketone adduct (Step I of Scheme 1). This highly reactive intermediate can either 1) react with a nucleophilic residue within the active site to form a tethered intermediate (Step II of Scheme 1) which can undergo hydrolysis at S465 (Step III of Scheme 1); 2) undergo elimination of the bromide (with water as the attacking nucleophile)(Step II′ of Scheme 1); or 3) undergo hydrolysis to release the α-bromomethyl ketone carboxylic acid which diffuses out of the active site to react with distal cysteine residues on the surface of the enzyme (Step V of Scheme 1). In our early work, we calculated that the average enzyme turnover number for inactivation of cytosolic myocardial iPLA₂ by BEL is approximately 10 mol of BEL for every mole of phospholipase A₂ and that the stoichiometry of BEL to fully inhibited iPLA₂β is approximately 1.2 to 1 (BEL:enzyme ratio) ²². These data indicate that a substantial fraction of the α-bromomethyl keto carboxylic acid derivative of BEL leaves the catalytic site and diffuses into the surrounding solution. Thus, strong nucleophilic agents such as DTT or glutathione are required to rapidly scavenge the released α-bromomethyl ketone derivative of BEL to prevent extensive non-specific modification of iPLA₂ surface cysteine residues and/or the inhibition of other enzymes. In initial in vitro studies, we utilized DTT to rapidly inactivate the released α-bromomethyl keto carboxylic acid derivative of BEL which diffused out of the active site ^{23, 49}. We specifically point out that DTT does not completely attenuate BEL inhibition of iPLA₂β, presumably due to its inability to effectively interact with the α-bromomethyl keto carboxylic acid derivatives of BEL generated and tethered to the active site.

Previously, Song and coworkers ³¹ reported that most cysteine residues of iPLA₂β were modified upon treatment with BEL in the absence of DTT and believed that neither the catalytic site, S465, nor any adjacent residues were modified. These observations led to their proposal that inactivation occurs by release of ring opened BEL by diffusion of the α-bromomethyl keto carboxylic acid through alkylating cysteine residues (mechanism B in Scheme 1) many of which are distal from the active site. Furthermore, Song et al. reported that C651 was the only cysteine residue modified by BEL in the presence of DTT (1 mM), but its alkylation by BEL was not sufficient to completely inactivate iPLA₂β activity. From this, they concluded that DTT could protect the enzyme from being inactivated by BEL through decreasing cysteine alkylation at sites distant to the active site. However, the direct demonstration of BEL attached to the active site serine and a cross-linked product with C651 identify a different mechanism responsible for iPLA₂β inactivation by BEL.

Specifically, the results of the current study clearly demonstrate that S465 is modified as a BEL adduct, leading to a mass shift of 254 Da. This is of particular relevance since site directed mutagenesis of S465 to alanine abolishes the phospholipase⁴¹ and acyl-CoA hydrolase³⁶ activities of iPLA₂β. In addition, we provide evidence that reaction of iPLA₂β with BEL results in the formation of a tethered adduct between S465 and C651. The absence of the BEL modified S465 containing peptide or the S465-BEL-C651 species in the work of Song et al.³¹ may have been due to their loss during workup using different isolation conditions or due to the low specific activity (~ 600 pmol fatty acid•min⁻¹•mg⁻¹ protein) of iPLA₂β used in their study vs. ~ 1 μmol fatty acid•min⁻¹•mg⁻¹ protein used in the current study. The BEL-modified S465 and the cross-linked S465-BEL-C651 species are stable acylenzyme intermediates in the inactivation pathway as evidenced by their identification by tandem mass spectrometry, MALDI-TOF and a doublet present on SDS-PAGE after high temperature denaturation. The stability of these intermediates may be attributed to a slow or disabled deacylation step as has been demonstrated from studies on serine proteases whose structural and catalytic features are well established and have been shown to exhibit a high degree of structural homology with the patatin family of phospholipases/lipases ^50–53. In a study on the inactivation mechanism of chymotrypsin by haloenol lactones, the deacylation of the active site serine was demonstrated to be the rate-limiting step in the pathway, and the half life of a stable acyl enzyme could be as long as 12.5 hrs ⁴⁶. An acyl-enzyme intermediate formed between porcine pancreatic elastase and its natural substrate, β-casomorphin-7, was detected by mass spectrometry, crystallized and characterized for greater understanding of the catalytic mechanism ⁵⁴. These studies suggested that the deacylation of the acyl enzyme intermediate requires the participation of a “hydrolytic water” molecule, which is appropriately positioned and activated by a nearby residue to generate a hydroxide ion which attacks the carbonyl carbon of the ester linkage. Following the hydrolysis of the bromoenol lactone by S465, the ring opened product allows greater mobility and flexibility of the BEL adduct within the active site. The position of this tethered adduct may displace hydrolytic water molecules or block their entrance or exit from the active site. As a result of these conformational changes, the carbonyl carbon of the ester linkage may not be susceptible to hydrolysis. The observation of an acyl enzyme intermediate formed with oleoyl-CoA that is stable on a catalytic timescale suggests that intramolecular transacylation could influence iPLA₂β substrate specificity, kinetics, intermolecular interactions and membrane localization each resulting in novel modes of enzyme regulation.

The results presented in the current study are therefore consistent with mechanism A (Scheme 1) (i.e., a large fraction is cross-linked internally while tethered) while other reactive sites are also modified without departure from the active site. Specifically, the BEL derivative remains tethered to the active site serine (S465) through an ester linkage after the hydrolysis of the lactone ring, and this tethered species is poised to alkylate an accessible nucleophile within the active site, forming an irreversibly inactivated enzyme (Scheme 1, Step II). This cross-linked intermediate can also undergo a subsequent deacylation step leading to regeneration of the active site serine (Scheme 1, Step III). The three modification products identified in the present study implicate the thiol of C651 (-SH) and water as the reactive nucleophiles. The reaction with the thiol generates a modified C651 and the cross-linked products (Scheme 1, Steps II and III) while the reaction with water facilitated by H-bonding leads to debromination of the α-bromomethyl ketone derivative of BEL which remains esterified to S465 as a stable adduct (Scheme 1, Step II′). Formation of any one of the three modified products results in an inactive enzyme.

The modification of C651 by BEL and the presence of a BEL adduct cross-linking C651 to S465 indicates that C651 is in close spatial proximity to the active site serine. Based upon the mass shift and the proposed inactivation mechanism, we were able to determine the structure of the cross-link and the maximum distance between C651 and S465 as approximately 9 Å. Considering its close spatial proximity to the active site serine and its strong nucleophilicty, this cysteine residue may play an important role in regulating iPLA₂ activity. Moreover, C651 is localized between two calmodulin binding motifs in the primary sequence of iPLA₂β, i.e. a “1-9-14” motif (⁶²²IRKGQGNKVKKLSI⁶³⁵) and an IQ motif (⁷⁰¹IQYFRLNPQLGSDI⁷¹⁴), indicating that it may have a role in regulating inhibition of phospholipase activity by calmodulin ³².

Previous studies have demonstrated that (S)-BEL is a more potent inhibitor of iPLA₂β than (R)-BEL ³⁵. To investigate potential correlations between their inhibitory potencies toward iPLA₂β and their abilities to covalently modify the enzyme, we compared the levels of the modified peptides in (S)-BEL and (R)-BEL-treated iPLA₂β samples through a mass spectrometric-based semi-quantitative approach. The results showed that the levels of the modified peptides are significantly higher in (S)-BEL-treated iPLA₂β than in the (R)-BEL-treated enzyme (i.e., the BEL-modified S465-containing tryptic peptide, the BEL-modified C651-containing tryptic peptide, and the cross-linked peptide are ten, six, and four times more abundant, respectively). The data indicate that the high potency of (S)-BEL against iPLA₂β positively correlates with its ability to generate stable BEL modified peptides. The ability of (S)-BEL and (R)-BEL to covalently modify iPLA₂β likely reflects differences in their diastereotopic interactions with the enzyme active site, which would affect: 1) binding affinity of BEL or phospholipid substrate; 2) the rate of formation of the acylenzyme intermediate; 3) the dwell time of the hydrolyzed inhibitor in the active site; and 4) the rate of deacylation. Future studies utilizing molecular models of iPLA₂β based upon the crystal structures of patatin-like enzymes in combination with computational simulations using BEL enantiomers as probes will likely provide further insight into the enantioselectivity of BEL for other members of the iPLA₂ family of enzymes.

In conclusion, this study has evaluated two distinct mechanisms for the inactivation of iPLA₂β by its mechanism-based inhibitor, BEL. The results from mass spectrometric and SDS-PAGE analyses support a mechanism by which a stable acyl-enzyme intermediate is formed (after the lactone of BEL is hydrolyzed) by the enzyme and subsequent reactions of this intermediate with nearby nucleophilic residues in the active site resulting in inhibition of iPLA₂β activity. The identification of nearly 50% cross-linked peptides by gel electrophoresis not only demonstrates the existence of the acyl-enzyme intermediate, but also reveals the spatial proximity of C651 to the active site S465 and that this interaction accounts for a substantive portion of BEL-mediated inactivation. Moreover, the inhibitory effects of (S)-BEL and (R)-BEL on iPLA₂β have been shown to correlate with their abilities to covalently modify the enzyme. The current studies facilitate our understanding of the catalytic mechanism of PLA₂β, demonstrate the active site connectivity with C651 and confirm the utility of BEL in identifying iPLA₂β-mediated alterations in lipid metabolism that can be substantiated by complimentary genetic loss of function experiments as we have previously demonstrated ^{17, 44, 55}.

ABBREVIATIONS

BEL

(E)-6-(bromomethylene)-3-(1-naphthalenyl)-2H-tetrahydropyran-2-one

α-CHCA

α-cyano-4-hydroxycinnamic acid

DTT

dithiothreitol

ESI LC-MS

electrospray ionization liquid chromatography-mass spectrometry

iPLA₂

calcium-independent phospholipase A₂

MALDI-TOF

matrix-assisted laser desorption/ionization time-of-flight

TFA

trifluoroacetic acid