A New Ultrasensitive Bioluminescence-Based Method for Assaying Monoacylglycerol Lipase

International Journal of Molecular Sciences

Article A New Ultrasensitive Bioluminescence-Based Method for Assaying Monoacylglycerol Lipase

Matteo Miceli 1, Silvana Casati 1 , Pietro Allevi 2, Silvia Berra 1 , Roberta Ottria 1 , Paola Rota 2 , Bruce R. Branchini 3 and Pierangela Ciuffreda 1,*

1 Dipartimento di Scienze Biomediche e Cliniche “Luigi Sacco”, Università degli Studi di Milano, Via G.B. Grassi 74, 20157 Milano, Italy; [email protected] (M.M.); [email protected] (S.C.); [email protected] (S.B.); [email protected] (R.O.) 2 Dipartimento di Scienze Biomediche, Chirurgiche e Odontoiatriche, Università degli Studi di Milano, Via della Commenda 10, 20122 Milano, Italy; [email protected] (P.A.); [email protected] (P.R.) 3 Department of Chemistry, Connecticut College, New London, CT 06320, USA; [email protected] * Correspondence: [email protected]; Tel.: +39-02-5031-9195

Abstract: A novel bioluminescent Monoacylglycerol lipase (MAGL) substrate 6-O-arachidonoylluciferin, a D-luciferin derivative, was synthesized, physico-chemically characterized, and used as highly sensitive substrate for MAGL in an assay developed for this purpose. We present here a new method based on the enzymatic cleavage of arachidonic acid with luciferin release using human Monoacylglycerol lipase (hMAGL) followed by its reaction with a chimeric luciferase, PLG2, to produce bioluminescence. Enzymatic cleavage of the new substrate by MAGL was demonstrated, and kinetic constants Km and Vmax were determined. 6-O-arachidonoylluciferin has proved to be a highly sensitive substrate for MAGL. The bioluminescence assay (LOD 90 pM, LOQ 300 pM) is Citation: Miceli, M.; Casati, S.; much more sensitive and should suffer fewer biological interferences in cells lysate applications than Allevi, P.; Berra, S.; Ottria, R.; Rota, P.; typical ﬂuorometric methods. The assay was validated for the identiﬁcation and characterization Branchini, B.R.; Ciuffreda, P. A New Ultrasensitive Bioluminescence-Based of MAGL modulators using the well-known MAGL inhibitor JZL184. The use of PLG2 displaying Method for Assaying distinct bioluminescence color and kinetics may offer a highly desirable opportunity to extend the Monoacylglycerol Lipase. Int. J. Mol. range of applications to cell-based assays. Sci. 2021, 22, 6148. https://doi.org/ 10.3390/ijms22116148 Keywords: bioluminescence; monoacylglycerol lipase; kinetic assay; PLG2

Academic Editor: Christophe Morisseau 1. Introduction Received: 4 May 2021 In the last decade, the endocannabinoid system (ECS) has stimulated renewed in- Accepted: 5 June 2021 terest, thanks to medical research results based on modern technologies highlighting its Published: 7 June 2021 large distribution in human tissues [1,2] and its involvement in several physiological and pathological processes [3]. Indeed, deeper investigations in this field have demon- Publisher’s Note: MDPI stays neutral strated how the dysregulation of the ECS can induce pathological conditions such as pain with regard to jurisdictional claims in and inflammation [4–6], neurological disorders [7], and cancer [8–10]. Although cannabi- published maps and institutional affil- noid receptor stimulation [11,12] and endocannabinoid turnover modulation [13–15] have iations. been investigated, targeting the ECS through enzyme modulation remains an ambitious objective [16]. In this regard, several studies have shown the involvement of the monoacylglycerol lipase (MAGL) as a playmaker in simultaneously coordinating multiple lipid signaling pathways in both physiological and disease contexts [17], thus suggesting a role Copyright: © 2021 by the authors. as a therapeutic target. Licensee MDPI, Basel, Switzerland. MAGL inhibitors are potential antiproliferative and cytostatic drugs for cancer chem- This article is an open access article otherapy not only for enhancing endocannabinoid signaling but also controlling fatty acid distributed under the terms and release for synthesis of protumorigenic signaling lipids. Moreover, these inhibitors are also conditions of the Creative Commons Attribution (CC BY) license (https:// antineurodegenerative compounds that induce an anti-inflammatory activity by reducing creativecommons.org/licenses/by/ the production of eicosanoids. The role of MAGL as a therapeutic target has increased 4.0/). interest not only in the development of selective inhibitors but also in the set-up of efficient

Int. J. Mol. Sci. 2021, 22, 6148. https://doi.org/10.3390/ijms22116148 https://www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 2 of 14

ducing the production of eicosanoids. The role of MAGL as a therapeutic target has increased interest not only in the development of selective inhibitors but also in the set-up of efficient protocols amenable to high-throughput screening (HTS) formats. In this context, different analytical methods to determine MAGL activity, generally based on its enzymatic action on fatty acid compounds, have been developed using various techniques. The procedures most used to screen MAGL inhibitors quantify the arachidonic acid released from 2-AG hydrolysis using high-performance liquid chromatography (HPLC) coupled with UV [18] or mass spectrometric detector [19]. Furthermore, a fluorescence- based assay employing 7-hydroxycoumarinyl arachidonate (7-HCA) substrate [20] and Int. J. Mol. Sci.new2021, 22procedures, 6148 for MAGL continuous assay based on long-wavelength fluorogenic sub- 2 of 14 strates, 7-hydroxyresorufinyl arachidonate [21] and 7-hydroxyresorufinyl octanoate [22] have been reported. Bioluminescenceprotocols (BL) amenable approach toes high-throughput offer several advantages screening (HTS) such formats. as high In sensitivity, this context, different lowering of the background,analytical methods wide dynamic to determine range, MAGL and activity,relative generallyease of performance. based on its enzymatic Beetle action luciferases have beenon fatty extensively acid compounds, investigated have been and developedare being usingwidely various used techniques.for bioanalytical The procedures purposes, and manymost of used these to screen applications MAGL inhibitors can be used quantify in a the high-throughput arachidonic acid screening released from 2-AG format in supporthydrolysis of drug development using high-performance [23]. Luciferases liquid chromatography catalyze the ATP-dependent (HPLC) coupled ox- with UV [18] or mass spectrometric detector [19]. Furthermore, a fluorescence-based assay employing idation of firefly D-luciferin (LH2), a benzothiazole-containing compound, in a two-step 7-hydroxycoumarinyl arachidonate (7-HCA) substrate [20] and new procedures for MAGL reaction [24]: (i) adenylation of LH2 with ATP, producing luciferyl adenylate with the re- continuous assay based on long-wavelength fluorogenic substrates, 7-hydroxyresorufinyl lease of pyrophosphate, and (ii) oxidation of the luciferyl adenylate by molecular oxygen, arachidonate [21] and 7-hydroxyresorufinyl octanoate [22] have been reported. producing a dioxetanoneBioluminescence that rapidly (BL) cleaves, approaches generating offer severalCO2 and advantages electronically such as excited high sensitivity, singlet oxyluciferin,lowering which of thedecays background, by emitting wide dynamica photon range, of light and in relative the green ease of to performance. the red Beetle range with a quantumluciferases yield have of 15–60%, been extensively depending investigated on the andluciferase are being [25,26]. widely Biolumines- used for bioanalytical cent imaging (BLI)purposes, has been and employed many of thesefor im applicationsaging various can bebiological used in aphenomena, high-throughput such screening as protein–proteinformat interaction in support [27] ofand drug gene development regulation [23[28].]. LuciferasesThe firefly catalyzeluciferase the system ATP-dependent oxidation of firefly D-luciferin (LH ), a benzothiazole-containing compound, in a two-step is widely used for BLI systems, which generally2 employ LH2 or aminoluciferin as a sub- reaction [24]: (i) adenylation of LH2 with ATP, producing luciferyl adenylate with the strate in the presence of ATP, Mg2+, and O2 to produce detectable photons. Based on this release of pyrophosphate, and (ii) oxidation of the luciferyl adenylate by molecular oxygen, strategy, several bioluminescent probes have been applied to the in vivo imaging of H2S, producing a dioxetanone that rapidly cleaves, generating CO2 and electronically excited - F , H2O2, nitroreductase,singlet oxyluciferin, aminopeptidase, which and decays so on by [29–33]. emitting a photon of light in the green to the red Given our continuedrange with interest a quantum in the yield development of 15–60%, depending of selective on the probes luciferase for [MAGL25,26]. Bioluminescentsuit- able for in vitro andimaging in vivo (BLI) assays, has been we employedreport here for a imagingstraightforward various biological assay in which phenomena, ara- such as chidonoyl luciferinprotein–protein (ArLuc-1) is interactionhydrolyzed [27 by] and MAGL, gene regulation and the released [28]. The LH firefly2 reacts luciferase with system is PLG2 to generatewidely light usedin a forsingle BLI systems,reaction which mixture generally (Scheme employ 1). LHLH22or was aminoluciferin chosen as asthe a substrate 2+ in the presence of ATP, Mg , and O2 to produce detectable photons. Based on this strategy, substrate to be modified through an ester linkage with the arachidonoyl group, because it − produces readilyseveral detected bioluminescent (with PMTs) probes 562 nm have maximum been applied BL. to the in vivo imaging of H2S, F ,H2O2, nitroreductase, aminopeptidase, and so on [29–33]. Indeed, it has beenGiven found our continued that theinterest 6′-hydroxy in the developmentgroup of LH of2 selectiveis crucial probes for forBL MAGL[34]; suitable therefore, the blockingfor in vitroof thisand functionin vivo assays,essentially we report quenches here BL. a straightforward Once the caged assay substrate in which arachi- is cleaved by MAGL,donoyl BL luciferin can be (ArLuc-1)produced is via hydrolyzed the reaction by MAGL, with andluciferase the released (Scheme LH2 1).reacts with PLG2 PLG2 was usedto generate instead light of the in awild-type single reaction enzy mixtureme, because (Scheme it is1 ).a LHthermostable2 was chosen lucifer- as the substrate ase that producesto a be bright modified BL throughsignal and an ester is resistant linkage withto low the arachidonoylpH shifting, group,making because it well it produces suited for cell-basedreadily assays. detected (with PMTs) 562 nm maximum BL.

SchemeScheme 1. Bioluminescence 1. Bioluminescence reactions ofreactions compound of 6compound0-O-arachidonoylluciferin 6′-O-arachidonoylluciferin (ArLuc-1). The substrate(ArLuc-1 chemical). The sub- structure and biochemicalstrate chemical reactions structure leading toand 562 biochemical nm peak light reac emissiontions leading (λ max) areto 562 shown. nm peak light emission (λ max) are shown. 0 Indeed, it has been found that the 6 -hydroxy group of LH2 is crucial for BL [34]; therefore, the blocking of this function essentially quenches BL. Once the caged substrate is cleaved by MAGL, BL can be produced via the reaction with luciferase (Scheme1). PLG2 was used instead of the wild-type enzyme, because it is a thermostable luciferase that produces a bright BL signal and is resistant to low pH shifting, making it well suited for cell-based assays. Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 3 of 14

Int. J. Mol. Sci.2.2021 Results, 22, 6148 and Discussion 3 of 14 Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 3 of 14 2.1. Design and Synthesis of the Probe In this study, a novel BL assay for MAGL was established. To resolve the optimum BL assay conditions2.2. Results Results for andMAGL and Discussion Discussion using the luciferase–luciferin reaction, the substrate Ar- 2.1. Design and Synthesis of the Probe Luc-1 was designed,2.1. Design and andvarious Synthesis conditions of the Probe for the assay were examined. The instability In this study, a novel BL assay for MAGL was established. To resolve the optimum BL of LH2 toward acid, Inbase, this light,study, and a novel oxygen BL assay [35], for together MAGL withwas established. the presence To ofresolve the carboxyl the optimum assay conditions for MAGL using the luciferase–luciferin reaction, the substrate ArLuc-1 group, precludedBL directassay conditions esterification for MAGL of LH using2 at the the 6 luciferase–luciferin′-position. However, reaction, LH2 theis stablesubstrate in Ar- was designed, and various conditions for the assay were examined. The instability of solution at neutralLuc- pH,1 was whereas designed, in and acid various solution conditions (pH < for2), thea rapid assay degradationwere examined. to TheD-cyste- instability LH2 toward acid, base, light, and oxygen [35], together with the presence of the carboxyl ine, and a benzothiazolof LH2 towarde derivative acid, base, occurs. light, and In oxygenalkaline [35], solution, together0 LHwith2 theis oxidized presence ofto the dehy- carboxyl group, precluded direct esterification of LH2 at the 6 -position. However, LH2 is stable group, precluded direct esterification of LH2 at the 6′-position. However, LH2 is stable in droluciferin, whichin solution is a potent at neutral inhibitor pH, whereas of luciferase in acid [36]. solution Consequently, (pH < 2), a rapid the ester degradation of LH2 to D- solution at neutral pH, whereas in acid solution (pH < 2), a rapid degradation to D-cyste- is prepared, as incysteine, the synthesis and a benzothiazole of other analog derivativeues, by occurs. construction In alkaline of solution,the labile LH thiazoline2 is oxidized to ine, and a benzothiazole derivative occurs. In alkaline solution, LH2 is oxidized to dehy- ring at a late stagedehydroluciferin, in the synthesis. which The issynthesis a potent inhibitorof ArLuc- of1 luciferase (Scheme[ 2)36 ].follows Consequently, a variation the ester droluciferin, which is a potent inhibitor of luciferase [36]. Consequently, the ester of LH2 of LH2 is prepared, as in the synthesis of other analogues, by construction of the labile of the method applicable to the syntheses of LH2 derivatives modified at the 6′-position isthiazoline prepared, ring as in at the a late synthesis stage in of the other synthesis. analog Theues, synthesisby construction of ArLuc- of 1the(Scheme labile thiazoline2) follows [37]. N,N′-dicyclohexylcarbodiimidering at a late stage in the sy (DCC)nthesis. and The 4-dimethylaminopyridine synthesis of ArLuc-1 (Scheme 2)(DMAP) follows awere variation a variation of the method applicable to the syntheses of LH2 derivatives modified at the used to directlyof 6couple0 -positionthe method arachidonic [37 applicable]. N,N0-dicyclohexylcarbodiimide acid to the to synthesesthe commercially of LH2 derivatives (DCC) available and modified 4-dimethylaminopyridine 6-hydroxy-2-cya- at the 6′-position nobenzothiazol,[37].(DMAP) producing N,N′-dicyclohexylcarbodiimide were compound used to directly 2 in couplenearly (DCC) arachidonicquantitative and 4-dimethylaminopyridine acid yield. to the Compound commercially (DMAP) 2 was available were then reacted withused6-hydroxy-2-cyanobenzothiazol, D-cysteine to directly hydrochloridecouple arachidonic producingunder acid toconditions compoundthe commercially that 2 in nearlydid available not quantitative hydrolyze 6-hydroxy-2-cya- yield. the Com- nobenzothiazol, producing compound 2 in nearly quantitative yield. Compound 2 was ester yielding ArLuc-pound1 2in was 52% then overall reacted yield. with ArLuc- D-cysteine1 was hydrochloride completely under characterized conditions by that 1H– did not then reacted with D-cysteine hydrochloride under conditions that did not hydrolyze the 13 hydrolyze the ester yielding ArLuc-1 in 52% overall yield. ArLuc-1 was completely char- and C–NMR and 2D–COSY, HMQC, and HMBC correlation spectra. The purity, as- 1 esteracterized yielding by 1ArLuc-H– and1 in13C–NMR 52% overall and yield. 2D–COSY, ArLuc- HMQC,1 was completely and HMBC characterized correlation spectra.by H– sessed by H1–NMR,13 was greater than 98%, and no signals for free LH2 were detected. The andThe purity,C–NMR assessed and 2D–COSY, by H1–NMR, HMQC, was greater and HMBC than 98%, correlation and no signalsspectra. for The free purity, LH2 were as- purity of the compoundsesseddetected. by H1–NMR, Thewas purity also was ofverified the greater compound by than thin 98%, was layer alsoand chromatography no verified signals by for thin free layer LH and chromatography2 were by lumino-detected. The and metric assay. purityby luminometric of the compound assay. was also verified by thin layer chromatography and by luminometric assay.

Scheme 2. Chemical synthesis of ArLuc-1. Scheme 2. Chemical synthesisScheme of ArLuc- 2. Chemical1. synthesis of ArLuc-1. Because of the high sensitivity of the luminometric assay [38], the luciferin derivative Because of the high sensitivity of the luminometric assay [38], the luciferin derivative Because of hadthe tohigh be free sensitivity of LH2, since of the traces luminometric of it in substrate assay preparations [38], the luciferin would result derivative in a signifi- had to be free of LH , since traces of it in substrate preparations would result in a signiﬁcant cant increase of background2 luminescence and lower the sensitivity of the method. To had to be free ofincrease LH2, since of background traces of it luminescence in substrate and preparations lower the sensitivity would ofresult the method. in a signifi- To ensure ensure that the observed BL signal was due only to the consecutive hydrolysis and PLG2 cant increase ofthat background the observed luminescence BL signal was dueand only lower to the the consecutive sensitivity hydrolysis of the andmethod. PLG2 reactionsTo reactions and assess the stability of the substrate in reaction condition, the enzymatic as- ensure that the observedand assess BL the signal stability was of thedue substrate only to inthe reaction consecutive condition, hydrolysis the enzymatic and PLG2 assay was say was further investigated by systematic elimination of MAGL from the assay. The re- reactions and assessfurther the investigated stability of by the systematic substrate elimination in reaction of condition, MAGL from the the enzymatic assay. The as- results sults (Figure 1) show that LH2 production is not observed unless ArLuc-1 is hydrolyzed say was further (investigatedFigure1 ) show by that systematic LH2 production elimination is not observedof MAGL unless from ArLuc- the assay.1 is hydrolyzed The re- by byMAGL, MAGL, demonstrating demonstrating the the stability stability of theof the substrate substrate under under assay assay condition. condition. sults (Figure 1) show that LH2 production is not observed unless ArLuc-1 is hydrolyzed by MAGL, demonstrating the stability of the substrate under assay condition.

FigureFigure 1. 1. ComparisonComparison of ofluminescence luminescence produced produced by ArLuc- by ArLuc-1 in 1presencein presence (●) or ( •absence) or absence (▪) of Mon- (I) of oacylglycerolMonoacylglycerol lipase lipase (MAGL). (MAGL). Data Datarepresent represent mean mean ± standard± standard deviation deviation (SD). (SD).

Figure 1. Comparison of luminescence produced by ArLuc-1 in presence (●) or absence (▪) of Mon- oacylglycerol lipase (MAGL). Data represent mean ± standard deviation (SD).

Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 4 of 14

Int. J. Mol. Sci. 2021, 22, 6148 4 of 14

2.2. PLG2 Expression and Purification 2.2.We PLG2 expressed Expression PLG2, and Purification a thermostable and specific activity enhanced green (λmax = 559 nm)We light-emitting expressed PLG2, luciferase a thermostable that was and en specificgineered activity from enhanced a chimeric green protein (λmax consisting = 559 nm ) oflight-emitting the large N-terminal luciferase domain that was of engineered P. pyralis luciferase from a chimeric fused proteinto the small consisting C-terminal of the largedo- mainN-terminal of Luciola domain italica of luciferaseP. pyralis [39].luciferase Because fused PLG2 to the is a small thermostable C-terminal luciferase domain that of Luciola pro- ducesitalica aluciferase bright BL [signal39]. Because and is resistant PLG2 is to a thermostablelow pH shifting, luciferase it is well that suited produces for cell-based a bright assays.BL signal E. coli and JM109 is resistant competent to low cells pH were shifting, transforme it is welld with suited the for pQE-30 cell-based vector assays. containingE. coli theJM109 PLG2 competent cDNA sequence cells were fused transformed to an N-term withinal the 6xHis pQE-30 tag. After vector expression, containing PLG2 the PLG2 was purifiedcDNA sequence by affinity fused chromatography to an N-terminal and 6xHis obtained tag. in After 28.7 expression, mg/L yield PLG2 [24]. was purified by affinity chromatography and obtained in 28.7 mg/L yield [24]. 2.3. PLG2 Activity: Effects of DMSO and ArLuc-1 2.3. PLG2 Activity: Effects of DMSO and ArLuc-1 To develop the bioluminescence assay, several preliminary controls were performed. The effectTo develop of DMSO the was bioluminescence investigated to assay, determine several whether preliminary the solvent controls would were performed.influence theThe assay effect readout. of DMSO DMSO was investigatedis commonly to used determine to solubilize whether substrates the solvent and would inhibitors influence that arethe poorly assay readout.soluble in DMSO aqueous is commonlysolutions in used enzymatic to solubilize assays, substrates but it can and inhibit inhibitors some en- that zymesare poorly in a concentration-dependent soluble in aqueous solutions manner. in In enzymatic fact, while assays, luciferases but itexhibit can inhibit high selec- some tivityenzymes and sensitivity, in a concentration-dependent practical applications manner. can be Inlimited fact, whileby low luciferases resistance exhibit to organic high solvents.selectivity Several and sensitivity, concentrations practical of DMSO applications (0.5%, can 1%, be 2.5%, limited and by 5%) low were resistance assessed to organicto de- terminesolvents. sufficient Several ArLuc- concentrations1 solubility of levels DMSO that (0.5%, did not 1%, inhibit 2.5%, andPLG2 5%) activity. were assessedAs shown to determine sufficient ArLuc-1 solubility levels that did not inhibit PLG2 activity. As shown in Figure 2, 2.5% (v/v) DMSO is a good compromise between the complete solubilization in Figure2 , 2.5% (v/v) DMSO is a good compromise between the complete solubilization of the substrate and a tolerable reduction in PLG2 activity. of the substrate and a tolerable reduction in PLG2 activity.

Figure 2. PLG2 luminescence emission in the presence of varying concentrations from 0.5 to 5% of FigureDMSO. 2. ErrorPLG2 bars luminescence represent meanemission± SD. in the presence of varying concentrations from 0.5 to 5% of DMSO. Error bars represent mean ± SD. Next, the assay required a standard curve relating relative light units (RLU) to the Next, the assay required a standard curve relating relative light units (RLU) to the concentration of LH2 produced (Figure3). Calibration curves with LH 2 standards (0 to concentration250 nM) were of made LH2 produced for PLG2, (Figure as described 3). Calibration in Materials curves and with Methods, LH2 standards and linearity (0 to 250 was nM)observed were made over the for entirePLG2, concentration as described rangein Materials tested. and Methods, and linearity was ob- servedThe over limit theof entire detection concentration (LOD) and range limit tested. of quantiﬁcation (LOQ) of the assay were calculated from calibration curves and blank values according to the formulas LOD = 3.3σ/s and LOQ = 10σ/s, where “σ” is the averaged standard deviation of the blanks and “s” is the slope of the calibration curve [40,41]. The resulting values were 0.09 nM for LOD and 0.3 nM for LOQ. Subsequently, ArLuc-1 was evaluated as a potential inhibitor of PLG2, because the absence of this interference is essential for the sensitivity of the assay. To address this issue, luminescence experiments were performed with PLG2 and LH2 in the presence of increasing concentrations of ArLuc-1 (0.1, 1, 5 µM). Reassuringly, PLG2 activity is essentially independent of ArLuc-1 at the concentrations tested (Figure4).

Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 5 of 14

Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 5 of 14 Int. J. Mol. Sci. 2021, 22, 6148 5 of 14

Figure 3. Luminescence (RLU) of PLG2 as a linear function of LH2 concentration in the presence of 2.5% DMSO. All points are expressed as mean ± SD.

The limit of detection (LOD) and limit of quantification (LOQ) of the assay were calculated from calibration curves and blank values according to the formulas LOD = 3.3σ/s and LOQ = 10σ/s, where “σ” is the averaged standard deviation of the blanks and “s” is the slope of the calibration curve [40,41]. The resulting values were 0.09 nM for LOD and 0.3 nM for LOQ. Subsequently, ArLuc-1 was evaluated as a potential inhibitor of PLG2, because the absence of this interference is essential for the sensitivity of the assay. To address this issue, luminescence experiments were performed with PLG2 and LH2 in the presence of increasing concentrations of ArLuc-1 (0.1, 1, 5 µM). Reassuringly, PLG2 activ- ityFigure is essentially 3. Luminescence independent (RLU) ofof PLG2 ArLuc- as a1 linearat the function concentrations of LH2 concentration tested (Figure in the4). presence of Figure 3. Luminescence (RLU) of PLG2 as a linear function of LH2 concentration in the presence of 2.5% DMSO. All points are expressed as mean ± SD. 2.5% DMSO. All points are expressed as mean ± SD.

The limit of detection (LOD) and limit of quantification (LOQ) of the assay were calculated from calibration curves and blank values according to the formulas LOD = 3.3σ/s and LOQ = 10σ/s, where “σ” is the averaged standard deviation of the blanks and “s” is the slope of the calibration curve [40,41]. The resulting values were 0.09 nM for LOD and 0.3 nM for LOQ. Subsequently, ArLuc-1 was evaluated as a potential inhibitor of PLG2, because the absence of this interference is essential for the sensitivity of the assay. To address this issue, luminescence experiments were performed with PLG2 and LH2 in the presence of increasing concentrations of ArLuc-1 (0.1, 1, 5 µM). Reassuringly, PLG2 activity is essentially independent of ArLuc-1 at the concentrations tested (Figure 4).

FigureFigure 4. 4. PLG2PLG2 activity activity in in presence presence of of the the indicated indicated co concentrationsncentrations of of ArLuc-1. ArLuc-1. Error Error bars bars represent represent meanmean ± ±SD.SD. 2.4. ArLuc-1 as MAGL Substrate: Assay Development and Validation 2.4. ArLuc-1 as MAGL Substrate: Assay Development and Validation After demonstrating excellent sensitivity and linearity for LH2 detection and estab- After demonstrating excellent sensitivity and linearity for LH2 detection and estab- lishing that MAGL could hydrolyze ArLuc-1 and release free LH2 (Figure1), the two-step lishing that MAGL could hydrolyze ArLuc-1 and release free LH2 (Figure 1), the two-step MAGL assay was further investigated by steady state kinetics analysis. Brieﬂy, 10 µL of MAGL assay was further investigated by steady state kinetics analysis. Briefly, 10 µL of hMAGL enzyme solution (25 ng) were pipetted into the wells of a white 96-well plate, and hMAGL enzyme solution (25 ng) were pipetted into the wells of a white 96-well plate, and 90 µL of PLG2 assay mix, containing varying amounts of ArLuc-1 (ﬁnal concentrations 90 µL of PLG2 assay mix, containing varying amounts of ArLuc-1 (final concentrations 0.5, 1, 5, 8, 10, and 20 µM) were quickly added by automatic injector. Luminescence was 0.5, 1, 5, 8, 10, and 20 µM) were quickly added by automatic injector. Luminescence was read immediately and at 3 min intervals for 10 min. Control experiments yielded a very

readlow immediately background (Figureand at 31 )min signal intervals with ArLuc- for 10 min.1 in theControl absence experiments of MAGL. yielded Measurement a very Figurelowof thebackground 4. PLG2 net luminescent activity (Figure in presence signal1) signal of at thewith varying indicated ArLuc- concentrations co1 ncentrationsin the absence of of substrateArLuc-1. of MAGL. Error ArLuc- Measurement bars1 representwere con- of meanthe net ± SD. luminescent signal at varying concentrations of substrate ArLuc-1 were converted verted into amounts of LH2 produced using calibration curves. The data were ﬁt to the intoMichaelis–Menten amounts of LH2 equationproduced to using obtain calibration kinetic parameters curves. The Km data and were Vmax fit to (Figure the Michaelis–5). 2.4.Menten ArLuc-The equation1 ﬁnal as MAGL step to obtainin Substrate: the validationkinetic Assay parameters Development of the assay Km and wasand Validation Vmax undertaken (Figure with 5). JZL184, a known MAGLAfter inhibitor demonstrating [42], to excellent substantiate sensitivity that ArLuc- and linearity1 is a true for substrate LH2 detection for MAGL. and estab- DMSO lishingsolutions that ofMAGL JZL184 could (100, hydrolyze 1, 0.5, 0.1, ArLuc- and 0.011 andµM) release were incubated free LH2 (Figure with hMAGL 1), the two-step for 20 min, MAGLand then assay the was results further of subsequent investigated two-step by steady assays state were kinetics plotted analysis. as a dose-response Briefly, 10 µL curve of hMAGL(Figure enzyme6). The activitysolution of(25h ng)MAGL were was pipetted calculated into the as describedwells of a white for the 96-well kinetic plate, assay. and The 90IC50 µL of for PLG2 JZL184 assay was mix, found containing to be 260 varying nM, whichamounts agrees of ArLuc- well with1 (final values concentrations reported by 0.5,Lauria 1, 5, 8, et 10, al. (217and 20 nM) µM) and were by Savinainen quickly added et al. by (209 automatic nM) [21 ,injector.43]. Luminescence was read immediately and at 3 min intervals for 10 min. Control experiments yielded a very low background (Figure 1) signal with ArLuc-1 in the absence of MAGL. Measurement of the net luminescent signal at varying concentrations of substrate ArLuc-1 were converted into amounts of LH2 produced using calibration curves. The data were fit to the Michaelis– Menten equation to obtain kinetic parameters Km and Vmax (Figure 5).

Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 6 of 14

Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 6 of 14 Int. J. Mol. Sci. 2021, 22, 6148 6 of 14

Figure 5. Kinetic study of MAGL activity with ArLuc-1. Data are reported as mean ± SD. The curve model is that of the Michaelis–Menten.

The final step in the validation of the assay was undertaken with JZL184, a known MAGL inhibitor [42], to substantiate that ArLuc-1 is a true substrate for MAGL. DMSO solutions of JZL184 (100, 1, 0.5, 0.1, and 0.01 µM) were incubated with hMAGL for 20 min, and then the results of subsequent two-step assays were plotted as a dose-response curve (Figure 6). The activity of hMAGL was calculated as described for the kinetic assay. The IC50 for JZL184 was found to be 260 nM, which agrees well with values reported by Lauria FigureFigure 5. 5. KineticKinetic study study of of MAGL MAGL activity activity with with ArLuc- ArLuc-1.. Data Data are are reported reported as mean ± SD.SD. The The curve curve modeletmodel al. (217is is that that nM) of of the theand Michaelis–Menten. Michaelis–Menten. by Savinainen et al. (209 nM) [21,43].

Figure 6. Concentration-dependent inhibi inhibitiontion of JZL184 on MAGL. Data, reportedreported asas meanmean ±± SD, are expressed as the percentage of controls (without(without JZL184).JZL184).

2.5. Docking Studies We also performed a dockingdocking study of ArLuc-ArLuc-1 usingusing aa crystallographiccrystallographic model the structure of human monoacylglycerol lipase (hMAGL) co-crystallized in complex with a structure of human monoacylglycerol lipase (hMAGL) co-crystallized in complex with a noncovalently bound inhibitor (PDB ID: 5ZUN, resolution 1.35 Å). This structure recently noncovalently bound inhibitor (PDB ID: 5ZUN, resolution 1.35 Å). This structure recently was effectively used in a structure-based drug design of a novel series of reversible MAGL was effectively used in a structure-based drug design of a novel series of reversible MAGL inhibitors [44]. The docking procedure (see experimental) was validated by docking inhibitors [44]. The docking procedure (see experimental) was validated by docking the Figurethe cognate 6. Concentration-dependent ligand (1-(2-chlorobiphenyl-3-yl)-4-[4-(1,3-thiazol-2-ylcarbonyl)-piperazin-1- inhibition of JZL184 on MAGL. Data, reported as mean ± SD, arecognate expressed ligand as the percentage(1-(2-chlorobiphenyl-3-yl)-4-[4-(1,3-thiazol-2-ylcarbonyl)-piperazin-1- of controls (without JZL184). yl]pyrrolidin-2-one) of 5ZUN. This ligand was built through the Schrödinger Maestro Build yl]pyrrolidin-2-one) of 5ZUN. This ligand was built through the Schrödinger Maestro Toolbar, prepared with Schrödinger Ligand Preparation, and docked into the apo form 2.5.Build Docking Toolbar, Studies prepared with Schrödinger Ligand Preparation, and docked into the apo of 5ZUN protein. The obtained top pose (XP docking score −12.37 kcal mol-1) showed form of 5ZUN protein. The obtained top pose (XP docking score -12.37 kcal mol-1) showed an excellentWe also performed superimposition a docking to its study native of crystallographicArLuc-1 using a posecrystallographic in 5ZUN (RMSD model 0.16,the an excellent superimposition to its native crystallographic pose in 5ZUN (RMSD 0.16, structuremaximum of atomichuman displacement: monoacylglycerol 0.33 Å),lipase conﬁrming (hMAGL) the co-crystallized accuracy of this in approach.complex with a maximum atomic displacement: 0.33 Å), confirming the accuracy of this approach. noncovalentlyThe top pose bound obtained inhibitor by (PDB Glide ID: docking 5ZUN, of re ArLuc-solution1 shows 1.35 Å). a G This score structure and MM–GBSA recently wasbinding effectivelyThe energy top pose used values obtained in a (− structure-based10.33 by Glide and − docking78.68 drug kcal ofdesign mol-1,ArLuc- of respectively) a1 novelshows series a G thatscore of reversible conﬁrm and MM–GBSA its MAGL ability inhibitorsbindingto bind toenergy [44]. MAGL The values in docking the ( catalytic−10.33 procedure and site. −78.68 In (see particular, kcal experimental) mol-1, (Figure respectively)7 was), the validated carbonyl that confirm oxygenby docking its atom ability the of cognatetothe bind substrate to ligandMAGL ester in moiety(1-(2-chlorobiphenyl-3-yl)-4-[4-(1,3-thiazol-2-ylcarbonyl)-piperazin-1- the catalytic is H-bonded site. In to particular, main chain (Figure nitrogen 7), atoms the carbonyl of Ala51 oxygen and Met123. atom yl]pyrrolidin-2-one)During catalysis, these of residues5ZUN. This constitute ligandthe was oxyanion built through hole involved the Schrödinger in the stabilization Maestro Buildof the Toolbar, developing prepared negative with charge Schrödinger on the carbonylLigand Preparation, in the transition and docked state en into route the to apo the form of 5ZUN protein. The obtained top pose (XP docking score -12.37 kcal mol-1) showed tetrahedral intermediate [45]. Moreover, the distance from the hydroxyl oxygen of the ancatalytic excellent serine superimposition (Ser122) to the to reactiveits native carbonyl crystallographic carbon of pose the substrate in 5ZUN ester (RMSD is 3.09 0.16, Å, maximumwhich is the atomic ideal displacement: position for the 0.33 initial Å), nucleophilicconfirming the attack accuracy to take of place.this approach. The top pose obtained by Glide docking of ArLuc-1 shows a G score and MM–GBSA binding energy values (−10.33 and −78.68 kcal mol-1, respectively) that confirm its ability to bind to MAGL in the catalytic site. In particular, (Figure 7), the carbonyl oxygen atom

Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 7 of 14 Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 7 of 14

of the substrate ester moiety is H-bonded to main chain nitrogen atoms of Ala51 and ofMet123. the substrate During catalysis,ester moiety these is residues H-bonded consti to tutemain the chain oxyanion nitrogen hole atoms involved of Ala51 in the andsta- Met123.bilization During of the developingcatalysis, these negative residues charge consti on tutethe carbonyl the oxyanion in the hole transition involved state in en the route stabilization of the developing negative charge on the carbonyl in the transition state en route Int. J. Mol. Sci. 2021, 22, 6148 to the tetrahedral intermediate [45]. Moreover, the distance from the hydroxyl oxygen7 of 14 of tothe the catalytic tetrahedral serine intermediate (Ser122) to the [45]. reactive Moreover carb,onyl the distancecarbon of from the substhe hydroxyltrate ester oxygen is 3.09 Å,of thewhich catalytic is the serineideal position (Ser122) for to thethe reactiveinitial nucleophilic carbonyl carbon attack of to the take subs place.trate ester is 3.09 Å, which is the ideal position for the initial nucleophilic attack to take place.

Figure 7. Top scoring docking pose of compound ArLuc-1 (green stick) into the MAGL binding site. Residues are num- Figure 7. Top scoring docking pose of compound ArLuc-1 (green stick) into the MAGL binding site. Residues are numbered Figurebered according 7. Top scoring to the docking PDB crystal pose of(5ZUN) compound convention. ArLuc- 1The (green triad stick) catalytic into theresidues MAGL are binding shown site.in orange Residues sticks, are num-those accordingberedforming according tothe the oxyanion PDB to crystalthe hole PDB (5ZUN)in crystalmagenta convention. (5ZUN) sticks. convention.H-bonds The triad and catalytic The the tr πiad– residuesπ stackingcatalytic are interactionresidues shown in are orange are sh representedown sticks, in orange those with forming sticks, yellow those the and oxyanionformingblue dashed holethe oxyanionlines, in magenta respectively. hole sticks. in magenta H-bondsThe black sticks. and arrow theH-bonds representsπ–π stacking and thethe interaction πdistance–π stacking between are interaction represented the oxygen are with represented atom yellow of the and with hydroxyl blue yellow dashed group and lines,blueof Ser122 respectively. dashed and lines, the The carbon respectively. black of arrowthe carbonylThe represents black group arrow the of distancerepresents ArLuc- between1. the distance the oxygen between atom the of oxygen the hydroxyl atom of group the hydroxyl of Ser122 group and theof carbonSer122 ofand the the carbonyl carbon groupof the ofcarbonyl ArLuc- group1. of ArLuc-1. The luciferin moiety of ArLuc-1 falls into the narrow amphiphilic pocket [44] of MAGL,TheThe luciferinand luciferin its benzothiazole moiety moiety of ArLuc-of ArLuc-component1 falls1 intofalls forms theinto narrow athe π– πnarrow amphiphilicstacking amphiphilic interaction pocket [44 pocketwith] of MAGL,the [44] histi- of andMAGL,dine its of benzothiazole the and catalytic its benzothiazole triad component (Ser122, component forms His269, a π –Asp239) formsπ stacking a [44],π–π interaction whilestacking its nitrogeninteraction with the histidineatom with is involvedthe of histi- the catalyticdinein a water-mediatedof triadthe catalytic (Ser122, triad His269,H-bonding (Ser122, Asp239) network His269, [44], Asp239)with while the its [44], Tyr194 nitrogen while hydroxyl atomits nitrogen is involvedgroup atom and in is the ainvolved water- Ala51 mediatedinmain-chain a water-mediated H-bonding carbonyl network H-bondingoxygen. withAdditionally, network the Tyr194 with the hydroxyl arachidonoylthe Tyr194 group hydroxyl chain and theis group well-accommodated Ala51 and main-chain the Ala51 carbonylmain-chainin the large, oxygen. carbonyl hydrophobic Additionally, oxygen. pocket Additionally, the arachidonoylof MAGL the (Figure arachidonoyl chain 8) is well-accommodatedsurrounded chain is by well-accommodated hydrophobic/aro- in the large, hydrophobicinmatic the residueslarge, pocket hydrophobic as Phe159, of MAGL Phe209, pocket (Figure Ala151,of 8MAGL) surrounded Al a156,(Figure Ile179, by 8) hydrophobic/aromaticsurrounded Leu148, Leu205, by hydrophobic/aro- Leu213, residues Leu214, asmaticand Phe159, Leu241. residues Phe209, The as extensivePhe159, Ala151, Phe209, Ala156, hydrophobic Ala151, Ile179, interactio Leu148,Ala156, Ile179, Leu205,ns can Leu148,play Leu213, a significant Leu205, Leu214, Leu213, role and in Leu241. Leu214,the sta- The extensive hydrophobic interactions can play a signiﬁcant role in the stabilization of andbilization Leu241. of theThe complex. extensive hydrophobic interactions can play a significant role in the sta- thebilization complex. of the complex.

Figure 8. 2D representation of interactions between MAGL binding site and ArLuc-1.

3. Materials and Methods 3.1. Chemicals and Reagents Reactions progress was monitored by analytical thin-layer chromatography (TLC) on precoated aluminum foil (Silica Gel 60 F254-plate, Sigma–Aldrich, St. Louis, MO, USA), Int. J. Mol. Sci. 2021, 22, 6148 8 of 14

and the products were visualized by UV light. Purity of all compounds (>98%) was verified by thin layer chromatography and NMR measurements [46,47]. The chemicals required for validation, all of analytical grade, and all reagents required for bacterial cell culture and the kit for plasmid extraction were purchased from Sigma–Aldrich (St. Louis, MO). E. coli JM109 competent cells were from Promega, (Fitchburg, WI, USA). Ni-NTA agarose and 3–12 mL 10kDa Slide-A-Lyzer™ Dialysis Cassette used for protein purification were purchased from Thermo Fisher Scientific (Waltham, MA, USA). SDS-PAGE experiments were performed using Mini-PROTEAN® II handcast systems (Bio-Rad, Hercules, CA, USA) kits for both equipment and reagents, unless specified. Monoacylglycerol lipase (human recombinant, 50 µg, hMAGL) was purchased from Cayman Chemical (Ann Arbor, MI, USA). For bioluminescence experiments, LH2 potassium salt was purchased from Promega Italia Srl, while ATP magnesium salt was from Sigma–Aldrich.

3.2. Instruments 1H–NMR spectra were recorded in CDCl3 (isotopic enrichment 99.95%) solutions at 300 K using a Bruker AVANCE 500 instrument (500.13 MHz for 1H, 125.76 MHz for 13C) using 5 mm inverse detection broadband probes and deuterium lock. Chemical shifts (δ) are given as parts per million relative to the residual solvent peak (7.26 ppm for 1H and 77.0, central line, for 13C) and coupling constants (J) are in Hertz. For two-dimensional experiments (COSY-45, HSQC, and HMBC), standard Bruker microprograms using gradient selection (gs) were applied (for more information, see electronic Supplemental Data). The quantiﬁcation of the cDNA plasmid was done with a NanoDrop™ 2000 Microvol- ume Spectrophotometer (Thermo Fisher Scientiﬁc, Waltham, MA, USA) and the sonicator was a Sonopuls HD2070 (Bandelin, Berlin, Germany) at ~50% power (~35 W). Lumines- cence signals (λmax = 559 nm) were recorded by a GloMax®-Multi+ Microplate Multimode Reader (Promega, Fitchburg, WI, USA). The GloMax reagent lines and injector were cleaned before and after use with 70% ethanol, then abundantly rinsed with ddH2O water.

3.3. Synthesis of 6-O-Arachidonoylluciferin (ArLuc-1)

To a stirred solution of arachidonic acid (0.18 mmol) in anhydrous CH2Cl2 (3 mL), under an argon atmosphere and cooled in an ice bath, was added DMAP (2 mg, 0.01 mmol) and 6-hydroxy-2 cyanobenzothiazole (30 mg, 0.18 mmol). DCC (72 mg, 0.35 mmol) was added to the reaction mixture at 0 ◦C. Stirring was continued overnight at room tempera- ture. Product formation was conﬁrmed by TLC (petroleum ether/ethyl acetate, 95:5). The CH2Cl2 solution was washed twice with saturated NaHCO3 (2.5 mL), dried on anhydrous Na2SO4, and concentrated to give the crude product as a brown oil. Filtration on column chromatography on silica gel (petroleum ether/ethyl acetate, 95:5) gave 80 mg (0.17 mmol) of the nitrile product as a colorless oil. The nitrile and D-cysteine hydrochloride monohy- drate (0.19 mmol) were dissolved in a mixed solvent of 1.5 mL MeOH and 400 µL CH2Cl2 in an argon atmosphere. A total of 500 µLH2O and K2CO3 (0.19 mmol) were added, and the solution was kept stirring at RT for 1 h. HCl 1M was added until the solution reached a 2–3 pH value. The precipitate was collected and washed with water to give compound 1 ArLuc-1 as a white solid (30 mg overall yield 30%). H–NMR (CDCl3, 500 MHz): δ (ppm) 00 00 00 0.90 (3H, t, J = 7.6, 20 -CH3), 1.28-1.40 (6 H, m, 17 –19 CH2), 1.88 (2H, tt, J = 7.0, 7.6 Hz, 00 00 00 3 -CH2), 2.06 (2H, dt, J = 7.6, 7.6 Hz, 16 -CH2), 2.24 (2H, dt, J = 7.0, 7.4 Hz, 4 -CH2), 00 00 00 00 2.64 (2H, t, J = 7.0 Hz, 2 -CH2), 2.82-2.88 (6H, m, 7 –10 - and 13 -CH2), 3.80-3.87 (2H, AB part of ABX system, 5a- and 5b-H), 5.32-5.50 (9H, m, 500–600–800–900–1100–1200–1400–1500-CH and 4-H), 7.29 (1H, dd, J = 2.3, 8.9 Hz, 50-H), 7.73 (1H, d, J = 2.3 Hz, 70-H), 8.16 (1H, d, 0 13 00 00 00 J = 8.9 Hz, 4 -H). C–NMR (CDCl3, 500 MHz): δ (ppm) 14.1 (20 ), 22.6 (19 ), 24.7 (3 ), 25.7 (700, 1000, 1300), 26.5 (400), 27.2 (1600), 29.3 (1700), 31.5 (1800), 33.7 (200), 35.1 (5), 78.2 (4), 114.7 (70), 121.6 (50), 125.2 (40), 127.5, 127.8, 128.0, 128.4, 128.6, 128.6, 129.3, 130.5 (500, 600, 800, 900, 1100, 1200, 1400, 1500), 136.7 (80), 149.7 (60), 150.8 (90), 158.0 (2), 160.3 (20), 167.1 (100), 171.9 (COOH). IUPAC numbering of 6-O-arachidonoylluciferin, 1H and 13C NMR spectra of 6-O-arachidonoylluciferin (1) are reported in Supplementary Materials. Int. J. Mol. Sci. 2021, 22, 6148 9 of 14

3.4. Stability of the Substrate The stability of ArLuc-1 was assessed in PLG2 working solution. A total of 10 µL of the ArLuc-1 working solution were pipetted into the wells of a white 96-well plate; the injection system of the instrument automatically injected 100 µL of PLG2 mix into the well to be analyzed, and the luminescence was read every 30 min for 150 min.

3.5. PLG2 Expression and Purification The pQE-30 vector containing the PLG2 cDNA sequence (GenBank: KY486507) fused with N-terminal 6xHis tag was used to transform E. coli JM109 competent cells subsequently grown on LB plates containing ampicillin (100 µg/mL) for selection. PLG2 expression was induced in liquid cultures with 0.1 mM IPTG, overnight at 300 rpm and 37 ◦C, to avoid protein precipitation. The cells were harvested by ultracentrifugation at 5000× g for 10 min and frozen at −80 ◦C for subsequent purification. Frozen cell pellets were lysed by sonication, and 6xHis-PLG2 was affinity purified on Ni-NTA agarose according to the manufacturer’s instructions. Positive elution fractions were pooled and dialyzed against PCB buffer (50 mM Tris, 150 mM NaCl, 1 mM DTT, 1 mM EDTA, pH 7) using a 3–12 mL-10kDa Slide-A-Lyzer™ Dialysis Cassette. The protein concentrate was divided into small aliquots, flash-frozen into liquid nitrogen, and stored at −80 ◦C. The purity of purified PLG2 was checked by SDS-PAGE, followed by Coomassie staining.

3.6. Stock, Working, and Enzymatic Assay Solutions

LH2: LH2 working solution was prepared at 0.5 mM concentration by dissolving LH2 potassium salt in Tris-HCl 50 mM, pH 7.8. The concentration was adjusted by dilution based on UV using a molar absorptivity coefficient of 7600 M/cm at 266 nm. LH2 calibration curve was obtained pipetting appropriate amounts of the 0.5 mM LH2 working solution into test tubes containing DMSO and adding 50 mM Tris-HCl pH 7.8 buffer with 1 mM EDTA (DMSO 5%). ATP: ATP solution was prepared by dissolving 9 mM ATP in 50 mM Tris-HCl, pH 7.8. The concentration was adjusted by dilution based on UV using a molar absorptivity coefficient of 15,400 M/cm at 259 nm. ArLuc-1: The substrate ArLuc-1 was dissolved in DMSO at 1 mM concentration to obtain stock solution. To prepare working solution an aliquot of stock, solution was diluted 1:20 in DMSO. To prepare ArLuc-1 testing solutions, ArLuc-1 stock solution was diluted in DMSO to final concentrations of 1, 10, and 50 µM. To prepare ArLuc-1 reagent solution, ArLuc-1 working solution was pipetted into an Eppendorf tube, then 50 mM Tris-HCl buffer, pH 7.8, was slowly added, vortexing the tube to obtain homogeneous solution at 2.8 µM final concentration. PLG2: PLG2 working solution was prepared from an aliquot of purified PLG2, ice- thawed, by dilution to a 3.3 µM concentration with 50 mM Tris-HCl, pH 7.8. MgSO4 and ATP solution were added to a final concentration of 10 mM and 0.4 µM, respectively. PLG2 calibration solution for the LH2 calibration curve was prepared as PLG2 working solution and DMSO were added to reach a final concentration of 2.5%. PLG2 assay mix solution for the enzymatic assay was prepared from an aliquot of purified PLG2, ice-thawed, by dilution to a 3.3 µM concentration with 50 mM Tris-HCl, pH 7.8. MgSO4, ATP, and ArLuc-1 were added at final concentration of 10 mM, 0.4 µM, and 1.4 µM, respectively. hMAGL: hMAGL was diluted to 250 ng/mL in 50 mM Tris-HCl buffer pH 7.8 containing 1mM EDTA. JZL184: JZL184 stock solution was prepared by dissolving JZL184 powder into DMSO at 4 mM concentration. JZL184 working solutions were prepared at four different concentrations, (40, 20, 4, 0.4 µM) by diluting an aliquot of stock solution with DMSO. Any leftover mix could be stored at 4 ◦C and used within 48 h with no loss in performance. Int. J. Mol. Sci. 2021, 22, 6148 10 of 14

3.7. PLG2 DMSO Trials The DMSO concentration effects on the enzyme activity was determined by adding different amounts of DMSO to PLG2 working solution (0.5%, 1%, 2,5%, and 5% (v/v)). A sample of PLG2 working solution without DMSO was treated as control. The enzyme stability was expressed as the remaining activity assayed relative to the control value. All measurements were performed as follows: 100 µL of solution containing LH2 were pipetted into the wells of a white 96-well plate. The injection system of the instrument would then automatically inject 100 µL of PLG2 mix into the ﬁrst well to be analyzed. After the injection, the luminometer had a lag time of 1 s, and then integrated the luminescence signal for 1 s. The luminometer would then pass to the following well to be analyzed. The reagent lines and injector were cleaned before and after use with 70% ethanol.

3.8. PLG2 ArLuc-1 Trials The effects of ArLuc-1 concentration on PLG2 enzymatic activity were assessed at 3 different concentrations of ArLuc-1 (0.1, 1, 5 µM). In a typical experiment, 100 µL of the LH2 working solution were pipetted in a well of a white 96-well plate, and 20 µL of the appropriate ArLuc-1 testing solution were added. Then, the injection system of the instrument automatically injected 100 µL of PLG2 working solution into the ﬁrst well to be analyzed, and the plate was read every 3 min for 10 min. The time course analyses were linearly regressed, and the slope obtained used to compare PLG2 activity considering as 100% activity the experiment performed without the addition of ArLuc-1.

3.9. Linearity Tests and Sensitivity of the Method

Calibration curves with LH2 standards at different concentrations were built for PLG2. All measurements were performed as described, above (paragraph 3.8.). The tested LH2 concentrations were 0.5, 2.5, 5, 12.5, 25, 50, 125, 250 nM. The LOD and LOQ for the assay were calculated on calibration curves. Six blank values were averaged (50 mM Tris-HCl pH 7.8 buffer with 1 mM EDTA, DMSO 2.5%), and the resulting SD was used to calculate LOD and LOQ.

3.10. MAGL Bioluminescent Assays and Kinetic Tests MAGL assays were set up as follows: 10 µL of the hMAGL solution (enzyme 25 ng) were pipetted at the bottom of the appropriate well; then, 90 µL of the PLG2 assay mix were quickly added to the enzyme, and the plate was read immediately and every 3 min for 10 min. The kinetic values were determined for hMAGL with ArLuc-1 as substrate. Experiments were set up in the same fashion as the ones above. This time, six reagent solutions were prepared for ArLuc-1 in-well ﬁnal concentrations of 0.5, 1, 5, 8, 10, and 20 µM. For negative controls, 50 mM Tris-HCl, pH 7.8 was used, while for blank samples, 50 mM Tris-HCl, pH 7.8 containing 2.5% DMSO was used. The signal coming from negative control solutions was subtracted to that of the corresponding concentration and time point, in order to consider only the signal due to enzymatic hydrolysis. The resulting values were converted into LH2 µmoles in accordance with the calibration curves, and the data points were regressed to a linear curve. The slope of each curve was converted into µmol/min/mg produced by the enzyme. The values between the various experiments were averaged and ﬁt to a Michaelis–Menten equation.

3.11. MAGL Assay Validation JZL184 was chosen for the method validation due to its well-known MAGL inhibitory activity. In a typical experiment, 10 µL of the appropriate JZL184 working solution, or of the stock solution for the highest inhibitor concentration or of DMSO for 100% hMAGL activity curve, were diluted in an Eppendorf tube with PLG2 working solution that was added dropwise and vortexed to obtain a homogeneous solution, avoiding JZL184 precipitation. hMAGL solution was then added and incubated for 30 min. Then, 10 µL of Arluc-1 working solution were pipetted into a well of a white 96-well plate, and 90 µL of the Int. J. Mol. Sci. 2021, 22, 6148 11 of 14

solution prepared before and incubated were added, and the luminescence assay was run. Final concentrations for JZL184 were 100, 1, 0.5, 0.1, and 0.01 µM. For all the other reagents (hMAGL, PLG2, Arluc-1), ﬁnal concentrations resemble that of bioluminescence assays. Blank and negative controls were also analyzed. Control experiment samples were performed as validation experiment without the addition of hMAGL. For blank experiments, 10µL of the appropriate JZL184 working solution, or of the stock solution for the highest inhibitor concentration, were diluted in an Eppendorf tube with PLG2 assay buffer (50 mM Tris-HCl, pH 7.8, 10 mM MgSO4, 0.4 µM ATP) that was added dropwise. The tube was vortexed to obtain a homogeneous solution avoiding JZL184 precipitation. After 30 min, 10 µL of Arluc-1 working solution were pipetted in a well, 90 µL of the appropriate blank solution were added, and the assay was run. In a typical time, course analysis luminescence was read every 3 min for 10 min. The resulting points were regressed to a linear curve (r2 ≥ 0.98 for every curve). The slopes were normalized, and then that of the curve relative to hMAGL without inhibitors was taken as the 100% value, and the others were related to this one.

3.12. In silico Molecular Docking Simulations Protein coordinates were downloaded from the Protein Data Bank, accession code 5ZUN. The Schrödinger suite 2017 (Schrödinger, Cambridge, MA, USA) was used for all the computational procedures, and, unless otherwise stated, default settings were applied. The protein was prepared using Protein Preparation Wizard; tetraethylene glycol (PG4), 1,2 ethanediol (EDO), chloride ion (CL), and water molecules (except water molecules HOH542 and HOH585) were removed. Cognate ligand (1-(2-chlorobiphenyl-3-yl)-4-[4-(1,3- thiazol-2-ylcarbonyl)-piperazin-1-yl]pyrrolidin-2-one) and ArLuc-1 were built using the Maestro Build Toolbar and prepared for docking by Schrödinger Ligand Preparation. The receptor grid was created by the Receptor Grid Generation, the box was centered on the cognate ligand of 5ZUN, and the size of the ligand diameter midpoint cubic box was set to 14 Å. The docking was run with Glide Docking using the extra precision (XP) mode, keeping at most 20 poses per ligand and 200 poses to include in the minimization. The top poses were refined using MM–GBSA (molecular mechanics with generalized Born surface area), as implemented in Prime. Residues at a distance of 12 Å from the ligand were considered as a flexible region in the refinement.

3.13. Statistical Analysis All experiments were performed in triplicate and independently replicated at least once. A calibration curve was constructed to convert luminescence data in to LH2 produced. The values of negative controls were subtracted from the enzymatic curve. The initial velocity was determined from the linear portion of the resulting curve. All data were elaborated using GraphPad Prism 6.0c (GraphPad Software, La Jolla, CA, USA), and kinetic parameters Km and Vmax were calculated applying a nonlinear regression analysis (Michaelis–Menten). The quantitative data were calculated as means ± standard errors. The IC50 was calculated with Prism 6 (GraphPad software) using a log (inhibitor) vs. response–Variable slope (four parameters) function.

4. Conclusions Fervent research on the ECS in the last decade has focused on the application of new techniques to deeply investigate the biochemical mechanisms of action of this ubiquitous system. The modulation of the activity of the ECS enzymes remains an important challenge for mechanistic and pharmacological purpose. For this reason, having a range of useful tools for varied research interests such as HTS methods for enzymes inhibitors or activators and development of methods for ex vivo and in vivo assessments is of fundamental importance. Our work, focused on the enzyme MAGL, aimed to add a new useful and very sensitive tool for ECS research. The efﬁcacy of the newly developed luminescence- based assay for measuring the activity of MAGL has been demonstrated. D-luciferin Int. J. Mol. Sci. 2021, 22, 6148 12 of 14

esterified with arachidonic acid, ArLuc-1, has been synthesized and was found to enable highly sensitive (LOD 90 pM, LOQ 300 pM) bioluminescent detection of MAGL activity in vitro for potential use in the study of MAGL inhibitors. Recombinant PLG2 was also produced and purified from bacterial cells in good yield (28.7 mg/L). This luciferase variant functioned well in the presence of 2.5% DMSO, confirming its stability and resistance to potentially denaturing conditions. The compatibility of the BL-based MAGL assay with DMSO, which is essential for solubilizing ArLuc-1, suggests that additional applications of PLG2 can be developed with reagents that are poorly soluble in aqueous buffers. The BL-assay described here will very likely expand the scope of MAGL inhibitor screening, which often involves solubilization of inhibitors in DMSO. The major advantages of this new bioassay for MAGL activity are high sensitivity and rapidity making it very suitable in applications when the amount of biological sample is limited. ArLuc-1 represents a valuable addition to the available toolkit of substrates useful for determining the activity of MAGL and can allow the development of efficient HTS protocols for the screening of potential selective inhibitors of this endocannabinoid metabolic enzyme. Furthermore, the very high sensitivity of this method together with the lower biological interference associated with bioluminescence compared to fluorescence-based methods suggests likely applicability to ex vivo or in vivo research.

Supplementary Materials: Supplementary Materials are available online at https://www.mdpi. com/article/10.3390/ijms22116148/s1. Author Contributions: Conceptualization P.C.; Investigation, M.M., S.C., R.O., P.A., P.R. and S.B.; Validation, M.M.; Formal Analysis, M.M., S.C., R.O., P.A., P.R. and S.B.; Resources P.C.; Writing— Original Draft Preparation, M.M., R.O., S.C., P.A. and P.C.; Writing—Review & Editing, R.O., S.C. and P.C.; Funding Acquisition, P.C. and B.R.B. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by Università degli Studi di Milano, Finanziamento di Ateneo— Linea 3—Bando SEED—PSR 2019 funding to P.C. and by the Air Force Office of Scientific Research grant FA9550-18-1-0017 to B.R.B. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. Conflicts of Interest: The authors declare no conflict of interest.

References 1. Joshi, N.; Onaivi, E.S. Endocannabinoid System Components: Overview and Tissue Distribution. Adv. Exp. Med. Biol. 2019, 1162, 1–12. [CrossRef] 2. Meccariello, R. Endocannabinoid System in Health and Disease: Current Situation and Future Perspectives. Int. J. Mol. Sci. 2019, 1162, 3549. [CrossRef] 3. Freitas, H.R.; Isaac, A.R.; Malcher-Lopes, R.; Diaz, B.L.; Trevenzoli, I.H.; de Melo Reis, R.A. Polyunsaturated fatty acids and endocannabinoids in health and disease. Nutr. Neurosci. 2018, 21, 695–714. [CrossRef] 4. Ottria, R.; Cappelletti, L.; Ravelli, A.; Mariotti, M.; Gigli, F.; Romagnoli, S.; Ciuffreda, P.; Banﬁ, G.; Drago, L. Plasma endocannabinoid behaviour in total knee and hip arthroplasty. J. Biol. Regul. Homeost. Agents 2016, 30, 1147–1152. [PubMed] 5. Pezzilli, R.; Ciuffreda, P.; Ottria, R.; Ravelli, A.; Melzi d’Eril, G.; Barassi, A. Serum endocannabinoids in assessing pain in patients with chronic pancreatitis and in those with pancreatic ductal adenocarcinoma. Scand. J. Gastroenterol. 2017, 52, 1133–1139. [CrossRef][PubMed] 6. Woodhams, S.G.; Sagar, D.R.; Burston, J.J.; Chapman, V. The role of the endocannabinoid system in pain. In Pain Control. Handbook of Experimental Pharmacology; Schaible, H.G., Ed.; Springer: Berlin/Heidelberg, Germany, 2015; Volume 227, pp. 119–143. [CrossRef] 7. Cristino, L.; Bisogno, T.; Di Marzo, V. Cannabinoids and the expanded endocannabinoid system in neurological disorders. Nat. Rev. Neurol. 2020, 16, 9–29. [CrossRef][PubMed] 8. Laezza, C.; Pagano, C.; Navarra, G.; Pastorino, O.; Proto, M.C.; Fiore, D.; Piscopo, C.; Gazzerro, P.; Bifulco, M. The Endocannabi- noid System: A Target for Cancer Treatment. Int. J. Mol. Sci. 2020, 21, 747. [CrossRef][PubMed] Int. J. Mol. Sci. 2021, 22, 6148 13 of 14

9. Vago, R.; Ravelli, A.; Bettiga, A.; Casati, S.; Lavorgna, G.; Benigni, F.; Salonia, A.; Montorsi, F.; Orioli, M.; Ciuffreda, P.; et al. Urine Endocannabinoids as Novel Non-Invasive Biomarkers for Bladder Cancer at Early Stage. Cancers 2020, 12, 870. [CrossRef] [PubMed] 10. Casati, S.; Giannasi, C.; Minoli, M.; Niada, S.; Ravelli, A.; Angeli, I.; Mergenthaler, V.; Ottria, R.; Ciuffreda, P.; Orioli, M.; et al. Quantitative Lipidomic Analysis of Osteosarcoma Cell-Derived Products by UHPLC-MS/MS. Biomolecules 2020, 10, 1302. [CrossRef][PubMed] 11. Lu, D.; Immadi, S.S.; Wu, Z.; Kendall, D.A. Translational potential of allosteric modulators targeting the cannabinoid CB1 receptor. Acta Pharmacol. Sin. 2019, 40, 324–335. [CrossRef][PubMed] 12. Morales, P.; Goya, P.; Jagerovic, N. Emerging strategies targeting CB2 cannabinoid receptor: Biased agonism and allosterism. Biochem. Pharmacol. 2018, 157, 8–17. [CrossRef] 13. Vago, R.; Bettiga, A.; Salonia, A.; Ciuffreda, P.; Ottria, R. Development of new inhibitors for N-acylethanolamine-hydrolyzing acid amidase as promising tool against bladder cancer. Bioorg. Med. Chem. 2017, 25, 1242–1249. [CrossRef] 14. Malamas, M.S.; Farah, S.I.; Lamani, M.; Pelekoudas, D.N.; Perry, N.T.; Rajarshi, G.; Miyabe, C.Y.; Chandrashekhar, H.; West, J.; Pavlopoulos, S.; et al. Design and synthesis of cyanamides as potent and selective N-acylethanolamine acid amidase inhibitors. Bioorg. Med. Chem. 2020, 28, 115195. [CrossRef] 15. van Egmond, N.; Straub, V.M.; van der Stelt, M. Targeting Endocannabinoid Signaling: FAAH and MAG Lipase Inhibitors. Annu. Rev. Pharmacol. Toxicol. 2021, 61, 441–463. [CrossRef] 16. Di Marzo, V. New approaches and challenges to targeting the endocannabinoid system. Nat. Rev. Drug Discov. 2018, 17, 623–639. [CrossRef] 17. Mulvihill, M.M.; Nomura, D.K. Therapeutic potential of monoacylglycerol lipase inhibitors. Life Sci. 2013, 92, 492–497. [CrossRef] 18. Saario, S.M.; Savinainen, J.R.; Laitinen, J.T.; Järvinen, T.; Niemi, R. Monoglyceride lipase-like enzymatic activity is responsible for hydrolysis of 2-arachidonoylglycerol in rat cerebellar membranes. Biochem. Pharmacol. 2004, 67, 1381–1387. [CrossRef][PubMed] 19. King, A.R.; Lodola, A.; Carmi, C.; Fu, J.; Mor, M.; Piomelli, D. A critical cysteine residue in monoacylglycerol lipase is targeted by a new class of isothiazolinone-based enzyme inhibitors. Br. J. Pharmacol. 2009, 157, 974–983. [CrossRef][PubMed] 20. Wang, Y.; Chanda, P.; Jones, P.G.; Kennedy, J.D. A fluorescence-based assay for monoacylglycerol lipase compatible with inhibitor screening. Assay Drug Dev. Technol. 2008, 6, 387–393. [CrossRef][PubMed] 21. Lauria, S.; Casati, S.; Ciuffreda, P. Synthesis and characterization of a new fluorogenic substrate for monoacylglycerol lipase and application to inhibition studies. Anal. Bioanal. Chem. 2015, 407, 8163–8167. [CrossRef] 22. Miceli, M.; Casati, S.; Ottria, R.; Di Leo, S.; Eberini, I.; Palazzolo, L.; Parravicini, C.; Ciuffreda, P. Set-Up and Validation of a High Throughput Screening Method for Human Monoacylglycerol Lipase (MAGL) Based on a New Red Fluorescent Probe. Molecules 2019, 24, 2241. [CrossRef][PubMed] 23. McCaffrey, A.; Kay, M.A.; Contag, C.H. Advancing molecular therapies through in vivo bioluminescent imaging. Mol. Imaging 2003, 2, 75–86. [CrossRef][PubMed] 24. Branchini, B.R.; Southworth, T.L. A Highly Sensitive Biosensor for ATP Using a Chimeric Firefly Luciferase. Methods Enzymol. 2017, 589, 351–364. [CrossRef] 25. Niwa, K.; Ichino, Y.; Kumata, S.; Nakajima, Y.; Hiraishi, Y.; Kato, D.-I.; Viviani, V.R.; Ohmiya, Y. Quantum yields and kinetics of the firefly bioluminescence reaction of beetle luciferases. Photochem. Photobiol. 2010, 86, 1046–1049. [CrossRef] 26. Ando, Y.; Niwa, K.; Yamada, N.; Enomoto, T.; Irie, T.; Kubota, H.; Ohmiya, Y.; Akiyama, H. Firefly bioluminescence quantum yield and colour change by pH-sensitive green emission. Nat. Photon 2008, 2, 44–47. [CrossRef] 27. Luker, K.E.; Smith, M.C.P.; Luker, G.D.; Gammon, S.T.; Piwnica-Worms, H.; Piwnica-Worms, D. Kinetics of regulated protein- protein interactions revealed with firefly luciferase complementation imaging in cells and living animals. Proc. Natl. Acad. Sci. USA 2004, 101, 12288–12293. [CrossRef][PubMed] 28. Badr, C.E.; Tannous, B.A. Bioluminescence imaging: Progress and applications. Trends Biotechnol. 2011, 29, 624–633. [CrossRef] [PubMed] 29. Yuan, M.; Ma, X.; Jiang, T.; Zhang, C.; Chen, H.; Gao, Y.; Yang, X.; Du, L.; Li, M. A novel coelenterate luciferin-based luminescent probe for selective and sensitive detection of thiophenols. Org. Biomol. Chem. 2016, 14, 10267–10274. [CrossRef] 30. van de Bittner, G.C.; Bertozzi, C.R.; Chang, C.J. Strategy for dual-analyte luciferin imaging: In vivo bioluminescence detection of hydrogen peroxide and caspase activity in a murine model of acute inflammation. J. Am. Chem. Soc. 2013, 135, 1783–1795. [CrossRef] 31. Li, J.; Chen, L.; Wu, W.; Zhang, W.; Ma, Z.; Cheng, Y.; Du, L.; Li, M. Discovery of bioluminogenic probes for aminopeptidase N imaging. Anal. Chem. 2014, 86, 2747–2751. [CrossRef][PubMed] 32. Feng, P.; Zhang, H.; Deng, Q.; Liu, W.; Yang, L.; Li, G.; Chen, G.; Du, L.; Ke, B.; Li, M. Real-Time Bioluminescence Imaging of Nitroreductase in Mouse Model. Anal. Chem. 2016, 88, 5610–5614. [CrossRef][PubMed] 33. Ke, B.; Wu, W.; Liu, W.; Liang, H.; Gong, D.; Hu, X.; Li, M. Bioluminescence Probe for Detecting Hydrogen Sulfide in Vivo. Anal. Chem. 2016, 88, 592–595. [CrossRef][PubMed] 34. White, E.H.; Wörther, H.; Seliger, H.H.; McElroy, W.D. Amino Analogs of Firefly Luciferin and Biological Activity Thereof1. J. Am. Chem. Soc. 1966, 88, 2015–2019. [CrossRef] 35. White, E.H.; McCapra, F.; Field, G.F.; McElroy, W.D. The structure and synthesis of firefly luciferin. J. Am. Chem. Soc. 1961, 83, 2402–2403. [CrossRef] Int. J. Mol. Sci. 2021, 22, 6148 14 of 14

36. Bowie, L.J. Synthesis of firefly luciferin and structural analogs. In Bioluminescence and Chemiluminescence; Kaplan, N., Colowick, N., Eds.; Elsevier: Amsterdam, The Netherlands, 1978; Volume 57, pp. 15–28. ISBN 9780121819576. 37. Toya, Y.; Takagi, M.; Kondo, T.; Nakata, H.; Isobe, M.; Goto, T. Improved Synthetic Methods of Firefly Luciferin Derivatives for Use in Bioluminescent Analysis of Hydrolytic Enzymes; Carboxylic Esterase and Alkaline Phosphatase. BCSJ 1992, 65, 2604–2610. [CrossRef] 38. Lundin, A.; Rickardsson, A.; Thore, A. Continuous monitoring of ATP-converting reactions by purified firefly luciferase. Anal. Biochem. 1976, 75, 611–620. [CrossRef] 39. Branchini, B.R.; Southworth, T.L.; Fontaine, D.M.; Kohrt, D.; Talukder, M.; Michelini, E.; Cevenini, L.; Roda, A.; Grossel, M.J. An enhanced chimeric firefly luciferase-inspired enzyme for ATP detection and bioluminescence reporter and imaging applications. Anal. Biochem. 2015, 484, 148–153. [CrossRef] 40. ICH Q2B Validation of Analytical Procedures: Methodology; International Commission on Harmonisation: London, UK, 1996. 41. Shrivastava, A.; Gupta, V. Methods for the determination of limit of detection and limit of quantitation of the analytical methods. Chron Young Sci. 2011, 2, 21. [CrossRef] 42. Long, J.Z.; Li, W.; Booker, L.; Burston, J.J.; Kinsey, S.G.; Schlosburg, J.E.; Pavón, F.J.; Serrano, A.M.; Selley, D.E.; Parsons, L.H.; et al. Selective blockade of 2-arachidonoylglycerol hydrolysis produces cannabinoid behavioral effects. Nat. Chem. Biol. 2009, 5, 37–44. [CrossRef][PubMed] 43. Savinainen, J.R.; Yoshino, M.; Minkkilä, A.; Nevalainen, T.; Laitinen, J.T. Characterization of binding properties of monoglyceride lipase inhibitors by a versatile fluorescence-based technique. Anal. Biochem. 2010, 399, 132–134. [CrossRef] 44. Aida, J.; Fushimi, M.; Kusumoto, T.; Sugiyama, H.; Arimura, N.; Ikeda, S.; Sasaki, M.; Sogabe, S.; Aoyama, K.; Koike, T. Design, Synthesis, and Evaluation of Piperazinyl Pyrrolidin-2-ones as a Novel Series of Reversible Monoacylglycerol Lipase Inhibitors. J. Med. Chem. 2018, 61, 9205–9217. [CrossRef] 45. Labar, G.; Bauvois, C.; Borel, F.; Ferrer, J.-L.; Wouters, J.; Lambert, D.M. Crystal structure of the human monoacylglycerol lipase, a key actor in endocannabinoid signaling. Chembiochem 2010, 11, 218–227. [CrossRef][PubMed] 46. Casati, S.; Manzocchi, A.; Ottria, R.; Ciuffreda, P. 1H, 13C and 15N NMR spectral assignments of adenosine derivatives with different amino substituents at C6-position. Magn. Reson. Chem. 2011, 49, 279–283. [CrossRef][PubMed] 47. Casati, S.; Manzocchi, A.; Ottria, R.; Ciuffreda, P. 1H, 13C and 15N NMR assignments for N6-isopentenyladenosine/inosine analogues. Magn. Reson. Chem. 2010, 48, 745–748. [CrossRef][PubMed]