Inhibition of yeast lipase (CRL1) and cholesterol esterase (CRL3) by6-chloro-2-pyrones: comparison with porcine cholesterol esterase

Mary Stoddard Hatch a, William M. Brown a, Jason A. Deck b, Lucy A. Hunsaker a,Lorraine M. Deck b;*, David L. Vander Jagt a;1

a Department of Biochemistry and Molecular Biology, University of New Mexico School of Medicine, Albuquerque, NM 87131, USAb Department of Chemistry, University of New Mexico, Albuquerque, NM 87131, USA

Received 11 July 2001; received in revised form 18 September 2001; accepted 18 October 2001

Abstract

Previously, it was demonstrated that pancreatic cholesterol esterase is selectively inhibited by 6-chloro-2-pyrones withcyclic aliphatic substituents in the 3-position. Inhibition is reversible and is competitive with substrate. Pancreatic cholesterolesterase is a potential target for treatment of hypercholesterolemia. In the present study, yeast cholesterol esterase fromCandida cylindracea (also called C. rugosa CRL3) was compared to porcine pancreatic cholesterol esterase for inhibition by aseries of 3-alkyl- or 5-alkyl-6-chloro-2-pyrones. In addition, CRL3 was compared with the related yeast lipase CRL1.Inhibition of CRL3 by substituted 6-chloro-2-pyrones was competitive with binding of the substrate p-nitrophenyl butyrate.Inhibition constants ranged from 0.2 WM to s 90 WM. Small changes in the alkyl group had profound effects on binding. Thepattern of inhibition of CRL3 is quite distinct from that observed with porcine cholesterol esterase. Molecular modelingstudies suggest that the orientation of binding of these inhibitors at the active site of CRL3 can vary but that the pyrone ringconsistently occupies a position close to the active site serine. CRL1 is highly homologous to CRL3. Nevertheless, patterns ofinhibition of CRL1 by substituted 6-chloro-2-pyrones differ markedly from patterns observed with CRL3. The substituted 6-chloro-2-pyrones are slowly hydrolyzed in the presence of CRL1 and are pseudosubstrates of CRL3, but are simple reversibleinhibitors of pancreatic cholesterol esterase ß 2001 Elsevier Science B.V. All rights reserved.

1. Introduction

The K/L hydrolase fold is an example of a tertiaryfold that many proteins adopt despite showing nosequence similarity. This fold belongs to the doublywound K/L superfold [1,2] and is one of the foldsrecognized in the SCOP classi¢cation of mostly par-allel K/L structures [3]. The K/L hydrolase fold wasinitially identi¢ed based upon comparison of the

three-dimensional structures of dienelactone hydro-lase, haloalkane dehalogenase, wheat serine carbox-ypeptidase II, acetylcholine esterase and Geotrichumcandidum lipase [4]. Members of the superfamily ofK/L hydrolases [5] have catalytic triads similar tothose ¢rst identi¢ed in serine proteases, except thatthe linear order in the K/L hydrolases is nucleophile,acidic residue, histidine, which is the mirror image ofthe order in serine proteases, namely, histidine, as-partate, serine. The nucleophilic residue in K/L hy-drolases, which can be serine, aspartate or cysteine,resides in a conserved pentapeptide located in a ‘nu-cleophilic elbow’ [6]. Many lipases including bile saltactivated lipases are members of the K/L hydrolases.

0167-4838 / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved.PII: S 0 1 6 7 - 4 8 3 8 ( 0 1 ) 0 0 3 0 4 - 1

* Corresponding author. Fax: +1-505-277-5438.E-mail addresses: ldeck@unm.edu (L.M. Deck),

dlvanderjagt@salud.unm.edu (D.L. Vander Jagt).1 Also corresponding author. Fax: +1-505-272-5788

BBAPRO 36527 1-5-02

Biochimica et Biophysica Acta 1596 (2002) 381^391www.bba-direct.com

These generally have serine as the nucleophile andthus are serine K/L hydrolases and can be consideredcounterparts of the serine proteases.

Bile salt activated lipases in vertebrates are pan-creatic enzymes that are secreted into the intestine toaid in lipid digestion. In some species, includingman, the same enzyme is secreted into milk. Thesubstrate speci¢city of mammalian bile salt activatedlipases is broad, including triacylglycerol, cholesterolesters, esters of fat soluble vitamins and phospholip-ids. Various names have been used for this hydrolaseincluding pancreatic cholesterol esterase (CEase),pancreatic lysophospholipase, nonspeci¢c lipase,pancreatic carboxylester lipase, as well as bile saltactivated lipase. Similar enzymes have been isolatedfrom yeast, where they are presumed to function infat utilization as a source of energy. G. candidumlipases and Candida rugosa lipase-1 (CRL1) are pri-marily triacylglycerol hydrolases while C. rugosa li-pase-3 (CRL3, also called Candida cylindraceaCEase) is speci¢cally a CEase [7^9]. CRL1 andCRL3 share 89% identity.

The crystal structures of a number of CEases andCRLs are available including CRL1 and CRL3[8,9], bovine CEase [10,11] and human CEase[12,13]. The catalytic triads (S194, D320, H435 forhuman and bovine; S209, E341, H449 for CRL1and CRL3) are similar with the active site serineoccupying an energetically unfavorable conforma-tion [9,10,13], which is commonly observed with ser-ine K/L hydrolases. The crystal structure of CRL3has cholesteryl linoleate bound at the active site.There are as yet no CEase crystal structures withbound inhibitors. It has been suggested that CRL3may be a good model for mammalian CEase [9].There are several crystal structures of CRL1^inhib-itor complexes [8].

Inhibitors of pancreatic CEase have been shown tolower absorption of dietary cholesterol esters in ani-mal studies [14^16]. In addition, CEase gene knock-out mice have impaired ability to absorb cholesteroladministered as cholesterol esters [17]. It appears,therefore, that hydrolysis of cholesterol esters in theintestine, catalyzed by CEase, is required for absorp-tion of dietary cholesterol esters and that inhibitorsof CEase may be useful therapeutics. Previously wereported that substituted 6-chloro-2-pyrones with ali-cyclic groups at the 3-position were selective inhibi-

tors of porcine CEase compared to serine proteases[18]. These pyrones are reversible inhibitors of por-cine CEase, although they have the potential to func-tion as suicide inhibitors of serine hydrolases andserine proteases [14,19]. In the present study wehave extended this set of substituted 6-chloro-2-py-rones to include substituents in the 3-position and 5-position and have compared the patterns of inhibi-tion of CRL1 and CRL3 and with porcine CEase totest whether CRL3 is a good model for mammalianCEase. In addition, molecular modeling studies havebeen carried out to gain insight into the orientationof pyrone inhibitors complexed to CRL3.

2. Materials and methods

2.1. Materials

CRL3 (sold as C. cylindracea CEase) was fromBoehringer Mannheim. CRL1, porcine CEase, p-ni-trophenylbutyrate, taurocholate and bu¡ers werefrom Sigma.

2.2. Synthesis of substituted 6-chloro-2-pyrones

The structures of the substituted 6-chloro-2-py-rones are shown in Table 2. Pyrones 1^7 were syn-thesized previously [18] by a procedure similar tothat reported by Boulanger and Katzenellenbogen[20]. Pyrones 8 and 9 were synthesized and com-pared to previous reports [19,20]. Pyrones 10 and11, which are new compounds, were prepared ac-cording to [18]. Brie£y, for 10 (6-chloro-3-methylcy-clopentyl-2-pyranone) and 11 (6-chloro-5-methylcy-clopentyl-2-pyranone): in the same manner asdescribed for the syntheses of 3-alkyl-6-chloro-2-py-rones [18] diethyl methylcyclopentylmalonate wasadded to sodium hydride in tetrahydrofuran fol-lowed by methyl cis 2-chloroacrylate. Saponi¢cationand decarboxylation of the resulting triester gave amixture of 2-methylcyclopentyl glutaconic acids asan oily solid. Trituration with ethyl acetate gavethe pure E-isomer. Assignment of the isomer wasbased upon previous reports of NMR data. 1HNMR (DMSO-d6) N 0.85^1.70 (m, 9H), 2.15 (d,2H), 3.20 (d, 2H), 6.82 (t, 1H), 12.4 (s, 2H). Reac-tion with acetyl chloride and chromatography of the

BBAPRO 36527 1-5-02

M. Stoddard Hatch et al. / Biochimica et Biophysica Acta 1596 (2002) 381^391382

resulting oil gave 20% of the 3-isomer (10) as a clearoil. 1H NMR N 1.08^1.80 (m, 9H), 2.44 (d, 2H), 6.15(d, 1H, J = 6.75 Hz), 7.05 (d, 1H, J = 6.75 Hz). The5-isomer (11), eluted second from the column, wasobtained as an oil (10%): 1H NMR N 1.10^1.80 (m,9H), 2.40 (d, 2H), 6.23 (d, 1H, J = 9.45 Hz), 7.25 (d,1H, J = 9.30 Hz). Assignment of 3- and 5-isomerswas based upon previous reports of NMR data[18]. Anal. (C11H13ClO2) C, H.

2.3. Enzyme assays

CRL1 and CRL3 activity was monitored in 25mM Tris, pH 7.4, containing 6 mM taurocholate,by following the hydrolysis of p-nitrophenylbutyrate,1 mM, at 405 nm, 25‡C. Porcine CEase was analyzedsimilarly except in 0.1 M HEPES [18]. Michaelis con-stants and kcat values were determined by nonlinearregression analysis of initial rate data using the Enz-¢tter program (Elsevier Biosoft). p-Nitrophenylbuty-rate was selected as the substrate in this study be-cause it is a convenient substrate with which toanalyze the kinetic properties of CRL1, CRL3 andporcine CEase.

2.4. Kinetic studies

Kinetic studies to determine inhibition constants,Ki, of the substituted 6-chloro-2-pyrones were carriedout in the assay bu¡ers listed above with and withoutaddition of inhibitor by measuring initial rates withp-nitrophenylbutyrate as the variable. Reactions wereinitiated by adding CRL or CEase. Ki values wereobtained from linear regression analysis of doublereciprocal plots.

2.5. Product identi¢cation

To determine whether the substituted 6-chloro-2-pyrones were simple reversible inhibitors or weresubstrates, pseudosubstrates or suicide inhibitorsof CRL1 and CRL3, assay mixtures containing in-hibitors and enzyme were incubated overnight. Thesamples were extracted with methylene chloride,acidi¢ed with 6 N HCl, and extracted again withmethylene chloride. The two extracts were evapo-rated, and the material was analyzed by TLC (silica)with 90% methanol, 10% ethyl acetate. Mixtureswithout CRL1 or CRL3 were used as controls. Forporcine CEase, it was demonstrated previously thatsubstituted 6-chloro-2-pyrones are simple reversiblecompetitive inhibitors [18].

2.6. Molecular modeling

Flexible docking of the inhibitors to CRL3 wasperformed using the AutoDock 3.0 software suitefrom the Scripps Research Institute [21,22]. The crys-tal structure of CRL3 (1CLE.pdb) solved by D.Ghosh et al. to a 2.0 Aî resolution [9] was modi¢edto accommodate the docking. The coordinates ofpolar hydrogens were added as predicted by Sybyl6.6 from Tripos. Partial charges were assigned fromunited Kollman dictionary charges and all substrateand ordered water atoms were removed. Inhibitorring structure was predicted from BFGS minimiza-tion in Sybyl and partial charges were assigned ac-cording to the Gasteiger^Huckel method. Van derWaals parameters for chlorine were derived fromthe Merck FF force ¢eld and scaled to match theAutoDock 3.0 energy function. Pictures were gener-

Table 1E¡ects of bu¡er and taurocholate on the kinetic parameters of CRL1, CRL3 and porcine CEase with p-nitrophenyl butyrate as sub-strate

CRL1 CRL3 CEase

kcat (min31) Km (WM) kcat (min31) Km (WM) kcat (min31) Km (WM)

Tris3taurocholate 3.8U106 219 4.5U102 61 3.9U104 731+taurocholate 3.6U106 206 7.1U102 70 2.4U104 371HEPES3taurocholate 3.5U106 212 4.2U102 73 4.0U104 244+taurocholate 2.8U106 249 6.1U102 51 3.9U104 188

BBAPRO 36527 1-5-02

M. Stoddard Hatch et al. / Biochimica et Biophysica Acta 1596 (2002) 381^391 383

ated using Visual Molecular Dynamics from the Uni-versity of Illinois [23].

3. Results

3.1. E¡ects of bu¡er and bile salts on the kineticparameters of CRL1, CRL3 and porcine CEase

p-Nitrophenylbutyrate was used as substrate in allof the kinetic studies of CRL1, CRL3 and porcineCEase. Kinetic parameters are summarized in Table1. CRL1 e⁄ciently utilizes p-nitrophenylbutyrate,kcat/Km = 1.7U1010 M31 min31, which approaches adi¡usion limited value. For CRL1, there are no sig-ni¢cant bu¡er e¡ects nor is the rate a¡ected by ad-dition of taurocholate. CRL3 utilizes p-nitrophenyl-butyrate less e⁄ciently than CRL1, kcat/Km = 1.2U107 M31 min31 in the presence of tauro-cholate. There is a modest bu¡er e¡ect on the Mi-

chaelis constant. The presence of taurocholate in-creased kcat both in Tris bu¡er and in HEPESbu¡er but had opposite e¡ects on the Michaelis con-stants in these two bu¡er systems. Porcine CEasewas the most sensitive to bu¡er and to taurocholate,especially the Michaelis constant. For CEase, kcat/Km = 2.1U108 M31 min31 in HEPES bu¡er in thepresence of taurocholate. The generally small e¡ectof taurocholate on the rate of hydrolysis of p-nitro-phenylbutyrate by CRL1, CRL3 or porcine CEasesuggests that all three of these hydrolases can accom-modate small substrates e⁄ciently without the needfor a bile salt.

3.2. Inhibition of CRL1 and CRL3 (C. cylindraceacholesterol esterase) by substituted 6-chloro-2-pyrones

A series of 11 3-substituted or 5-substituted 6-chloro-2-pyrones was tested for inhibition of CRL1and CRL3 using p-nitrophenylbutyrate as substrate.The results are shown in Table 2. CRL3 is muchmore sensitive to inhibition by substituted 6-chloro-2-pyrones than is CRL1, in spite of the 89% sequenceidentity. For CRL3, dissociation constants rangedfrom 0.2 to 96 WM. The most potent inhibitorswere 3-substituted derivatives in which a cyclohexanering was tethered to the 3-position of the pyrone witha three-carbon or four-carbon tether (pyrones 6 and7). For CRL1, dissociation constants ranged from 4to s 900 WM. For both CRL1 and CRL3, inhibitionwas consistently competitive as shown in Fig. 1 forinhibition of CRL3 by pyrone 4.

Table 26-Chloro-2-pyrones as inhibitors of CEase, CRL1 and CRL3

Data for pyrones 1^7 from [18].

Fig. 1. Inhibition of CRL3 by pyrone 4 is competitive with thebinding of substrate. This plot is representative of all of thedata in Table 2. Open circles: with inhibitor; closed circles:without inhibitor.

BBAPRO 36527 1-5-02

M. Stoddard Hatch et al. / Biochimica et Biophysica Acta 1596 (2002) 381^391384

3.3. Product analysis

Porcine CEase is reversibly inhibited by substi-tuted 6-chloro-2-pyrones, as demonstrated previously[18]. CRL1 and CRL3 by comparison slowly hydro-lyze the substituted 6-chloro-2-pyrones. CRL1 orCRL3 was incubated with su⁄cient inhibitor to com-pletely inhibit activity, followed by removal of inhib-itor on a PD10 desalting column. This resulted inimmediate recovery of activity of CRL1, suggestingthat the pyrones are not suicide inhibitors or irrevers-ible inactivators of CRL1. Treatment of CRL1 withinhibitor for up to 18 h, followed by extraction andanalysis of products by TLC, revealed the slow ap-pearance of a single product that had a mobilitycorresponding to the expected glutaconic acid pro-duced by hydrolysis of the pyrone lactone ring andhydrolysis of the resulting acyl chloride. Thus, sub-stituted 6-chloro-2-pyrones are slowly hydrolyzed inthe presence of CRL1. For CRL3, initial treatmentwith a pyrone inhibitor followed by desalting af-forded inactive CRL3 which recovered activity overa period of hours, suggesting that these pyrones arepseudosubstrates of CRL3 in which the acylenzymecan accumulate.

3.4. Inhibition of porcine CEase by 6-chloro-2-pyrones: comparison with CRL1 and CRL3

Inhibition of porcine CEase by the series of py-rones in summarized in Table 2. Dissociation con-stants ranged from 25 nM to 2 WM. Thus porcineCEase is much more sensitive to substituted 6-

chloro-2-pyrones than is CRL1 or CRL3. SinceCRL3 has been suggested to be a model for mam-malian CEase, comparison of porcine CEase andCRL3 was of special interest. Patterns of inhibitionwere markedly di¡erent. For example, CRL3 isthree-fold more sensitive than CEase to pyrone 7.By comparison, porcine CEase is 1450 and 740 timesmore sensitive than CRL3 to inhibition by pyrones 1and 8, respectively.

3.5. Molecular modeling of the interaction of CRL3with 6-chloro-2-pyrones

Molecular docking and energy minimization stud-ies were undertaken to compare the predicted rankorder and experimental rank order of inhibition ofCRL3 by the 11 pyrones. The results are summarizedin Table 3. There was generally good agreement inthe rank orders. Speci¢cally, pyrones 7, 6 and 4 werepredicted to be the most potent inhibitors, in agree-ment with experimental observations. Modeling ofall 11 pyrones consistently placed the pyrone ringnear the active site serine, consistent with the exper-imental observation that the pyrones are slowly hy-drolyzed. However, the predicted lowest energy ori-entation of the bound pyrones varied. For pyrones 6,7, and 8, the substituent in the 3-position was ori-ented into the cholesterol binding site as de¢ned inthe crystal structure of the CRL3^cholesteryl lino-leate complex [9]. For the other pyrones, the sub-stituent in the 3- or 5-position was oriented intothe fatty acid binding site. Representative resultsare shown in Fig. 2 for the docked complexes ofCRL3 and pyrones 4 and 6.

4. Discussion

4.1. Physiological roles of mammalian cholesterolesterases

Pancreatic CEase (EC 3.1.1.13) is best understoodas an enzyme that is secreted into the intestine to aidin lipid absorption. Its role in hydrolysis of choles-terol esters as an essential step in determining thebioavailability of dietary cholesterol esters is sup-ported by in vivo studies of CEase inhibitors [14^16] and by studies with gene knockout mice where

Table 3Predicted rank vs. observed rank of CLR3

Pyrone # Predicted rank Observed rank

1 11 92 10 103 6 44 3 35 4 56 2 27 1 18 9 119 8 710 5 611 7 8

BBAPRO 36527 1-5-02

M. Stoddard Hatch et al. / Biochimica et Biophysica Acta 1596 (2002) 381^391 385

impaired absorption of cholesterol ester was ob-served [17]. Gene knockout mice, however, did notexhibit impaired absorption of unesteri¢ed cholester-ol which suggests that CEase does not have a choles-terol transport function. This conclusion di¡ers fromthe results of a number of studies with rats in whichunesteri¢ed cholesterol absorption was decreased inanimals made CEase de¢cient by removing pancre-atic juice [52] or by treating pancreatic juice withantibody to CEase [53]. CEase increased the absorp-tion of free cholesterol into cultured Caco-2 cells [54]and into gut sacs [24], further suggesting a transportrole for CEase. Other studies, however, did not sup-port a role for CEase in transport of unesteri¢edcholesterol to enterocytes; uptake of cholesterol byCaco-2 cells has been reported to be independent ofadded CEase or of transfection with CEase cDNA[25,26]. These con£icting reports regarding a trans-port function for CEase may re£ect structural di¡er-ences in the various CEase used in these studies. TheC-termini of CEase possess a consensus proline-rich

11 amino acid repeat sequence (PVPPTGDSGAP)n,which di¡ers among species in the number of repeats,followed by an extreme C-terminal conserved nineamino acid consensus sequence (AQMPVVIGF).The number of repeats varies from two to four inmost animals to 16 in human. These proline-richrepeats allow CEase to bind to membrane associatedheparin, and it has been suggested that this mayfacilitate cholesterol transport [27]. However, a re-cent clinical trial with CVT-1 (sodium cellulose sul-fate), a potent and speci¢c inhibitor of human CEasethat binds to the heparin binding site on CEase, didnot block absorption of unesteri¢ed cholesterol [28].

Pancreatic CEase may have a major role in rees-teri¢cation of cholesterol within intestinal mucosalcells [54]. CEase is associated with chaperoneGrp94 during transit through the endoplasmic retic-ulum/Golgi secretory pathway in pancreatic cells [29]and remains associated with Grp94 as a complexduring transit to the intestinal lumen where someof the complex is internalized by enterocytes [30].

Fig. 2. Molecular docking and energy minimization of the binding of pyrones 4 and 6 to CRL3 predict di¡erent orientations of the 3-substituents. For pyrone 4, the 3-substituent is oriented into the fatty acid site. For pyrone 6, the 3-substituent is oriented into thecholesterol site. For reference, cholesteryl linoleate in superimposed (from [9]).

BBAPRO 36527 1-5-02

M. Stoddard Hatch et al. / Biochimica et Biophysica Acta 1596 (2002) 381^391386

The detection of internalized CEase and Grp94 isconsistent with earlier immunocytochemistry studiesof CEase [31]. Therefore, pancreatic CEase mayfunction within enterocytes to esterify cholesterol pri-or to formation of chylomicron particles.

CEase is synthesized in lactating mammary glandsand is secreted into milk in a number of species in-cluding human. Pancreatic and milk CEase are de-rived from the same gene [32]. It appears that thesame enzyme is present in serum [33] and is synthe-sized by human macrophages [34], eosinophils [35],endothelial cells [36], and liver [37]. The physiologicalroles of CEase that is present in these various loca-tions are not clear, nor is it known whether inhibitorsdesigned to block hydrolysis of dietary cholesterolesters through inhibition or inactivation of CEasein the intestine will have additional biological e¡ects.It is noteworthy that the lipid lowering e¡ect ofWAY-121,898, a novel carbamate inhibitor ofCEase, was observed both by oral and by parenteraladministration of the drug [15]. This apparent broaddistribution of CEase expression, however, has beenquestioned in a recent report of tissue levels of CEasemRNA in which only pancreatic and intestinal tis-sues were positive [38].

4.2. Comparison of mammalian cholesterol esterases,CRL1 and CRL3

CRL1 and CRL3 are two of ¢ve C. rugosa lipases,all of which are 534 amino acid proteins with 85^90% identity [39]. Nevertheless, CRL1 is a triglycer-ide hydrolase and CRL3 is a CEase. CRL1 andCRL3 like other lipases have hinge-like surface loopsthat block substrate access to the active site when theloops are closed [40]. Crystal structures of CRL1 inthe open and closed conformations and structures ofCLR1^inhibitor complexes reveal the movement ofthe 26 residue surface loop. In the open conforma-tion, a long tunnel becomes available to accommo-date a fatty acid. Modeling studies suggested that thesn-2 fatty acid of triacylglycerol substrates will insertinto the tunnel and that the remaining two fattyacids of the substrate will occupy positions in a de-pression created by the opening of the surface loop[39]. The structure of CRL3 complexed with choles-teryl linoleate revealed the fatty acid in the tunneland the cholesterol moiety residing in a hydrophobic

interfacial cavity formed where the two monomersface each other in the dimeric structure [9]. Of the55 residues that di¡er between CRL1 and CRL3, 23are located around the active site and dimer inter-face; these are presumed to account for the change insubstrate speci¢city from a triglyceride hydrolase(CRL1) to a CEase (CRL3). It is noteworthy, how-ever, that both CRL1 and CRL3 can accommodatep-nitrophenylbutyrate e⁄ciently (Table 1) withoutthe requirement of taurocholate, suggesting that thebile salt is not required for movement of the surfaceloop.

The structures of bovine CEase di¡er in some fun-damental ways from CRL3. The surface loop thatde¢nes lipases including CRL3 is truncated in bovineCEase, revealing a relatively open active site [10,11],in agreement with modeling predictions [41]. Thecrystal structure of the catalytic domain of humanCEase is similar to bovine CEase [13]. The two re-ported structures of bovine CEase di¡er. CEase crys-tallized in the absence of bile salts had the conservedC-terminus peptide lodged in the active site with dis-tortion of the oxyanion hole [11]. Crystallization inthe presence of bile salts produced a conformationwith an open active site, a productive orientation ofthe oxyanion hole, and two molecules of bound bilesalt, one near the active site and one at a remote site[10]. Human CEase, however, was crystallized with-out bound bile salt but with a functional oxyanionhole [13]. Bovine CEase crystals isolated in the pres-ence of bile salts formed a dimer with the active sitesfacing each other as observed with CRL1 and CRL3.Bovine CEase crystallized without bound bile saltsand human CEase each had one monomer per asym-metric unit. Although it has been proposed that bilesalt-induced dimerization is an essential part of theactivation process [42], it has been demonstrated instudies with full length and truncated recombinanthuman CEase that dimerization is not required forbile salt activation [43].

4.3. Inhibitors of cholesterol esterases

Numerous types of chemicals have been studied aseither mechanism based or reversible inhibitors ofCEase, including boronic and borinic acids, arylha-loketones, aryl phosphates and phosphonates, andaryl and cholesteryl carbamates [44^46]. Application

BBAPRO 36527 1-5-02

M. Stoddard Hatch et al. / Biochimica et Biophysica Acta 1596 (2002) 381^391 387

of carbamate chemistry has been especially usefulfor development of structure^activity relationships(SAR). Carbamates are transient inhibitors (pseudo-substrate inhibitors) of CEase in which the acylation(carbamoylation) step is followed by a slower deacy-lation step. SAR can be developed for the individualsteps of binding, acylation and deacylation. It hasbeen suggested from studies of carbamates that bilesalts increase the hydrophobicity of both the fattyacid and cholesterol binding sites of CEase [45]. Inaddition, carbamates have been used to develop SARfor each of these sites [46].

6-Chloro-2-pyrones, which were developed as in-hibitors of serine proteases, have the potential(Scheme 1) to function as (1) substrates, (2) pseudo-substrate inhibitors, (3) suicide inhibitors, (4) revers-ible inhibitors, or (5) inactivators. Initial studies ofthe inhibition of chymotrypsin by 3-benzyl-6-chloro-2-pyrone were interpreted as demonstrating suicideinhibition [19]. Subsequent studies indicated thatthis pyrone is a pseudosubstrate inhibitor in which

acylation is much faster than deacylation such thatthe acylenzyme accumulates and rearranges to theacid chloride, which hydrolyzes while tethered tothe active site serine, followed by slow hydrolysisand release of product [20,47]. This interpretationwas supported by crystallographic analysis of theacylenzyme [48]. By comparison, treatment of chy-motrypsin with 5-benzyl-6-chloropyrone led to inac-tivation, in which the active site serine was covalentlyattached to the 6-position of an intact pyrone, withloss of chloride [49]. Other related studies of inhibi-tion of serine proteases by haloenol lactones such as3-chloroisocoumarins and halomethylene furanoneshave demonstrated mechanism based (suicide) inacti-vation [50,51]. The results from the present studysuggest that 3- and 5-substituted pyrones inhibit por-cine CEase as simple reversible competitive inhibitorswhile inhibition of CRL1 and CRL3 is more com-plex. For CRL1, the pyrones are slowly hydrolyzedto the expected product without accumulation of in-active enzyme, suggesting that these are simple, albeit

Scheme 1. Substituted 6-chloro-2-pyrones can function as (1) pseudosubstrates of serine hydrolases in which the acylenzyme intermedi-ate B accumulates; (2) suicide inhibitors in which formation of B is followed by attack of an active site nucleophile to form C andthen deacylation to form D; (3) simple substrates in which the reactive acyl chloride B hydrolyzes to acylenzyme E followed by deacy-lation to regenerate active enzyme F with liberation of the substituted glutaconic acid product. Alternatively, substituted 6-chloro-2-pyrones can be simple reversible inhibitors or irreversible active site inhibitors where an active site nucleophile attacks the 6-positionof the pyrone with release of chloride.

BBAPRO 36527 1-5-02

M. Stoddard Hatch et al. / Biochimica et Biophysica Acta 1596 (2002) 381^391388

poor, substrates in which deacylation is faster thanacylation. By comparison, treatment of CRL3 pro-duces inactive enzyme that slowly recovers its activ-ity, suggesting that these pyrones are pseudosub-strates of CRL3 where acylation is faster thandeacylation. For both CRL1 and CRL3, the pyronesare su⁄ciently poor substrates that they can betreated as reversible competitive inhibitors to obtaininhibition constants.

4.4. Modeling the active sites of CRL3^pyronecomplexes

The modeling results (Fig. 2) raise the questionwhether 3- and 5-substituted 6-chloro-2-pyronesform complexes with CRL3 in which the substituentcan orient either towards the fatty acid binding siteor towards the cholesterol binding site. Generallyenergy minimization of the CRL3^pyrone complexessuggested that there were several structures of similarenergy. The rank order data in Table 3 represent thelowest energy complexes. Since other complexes withonly slightly higher energy may be predicted to orientdi¡erently, the docking results may not give a clearindication of preferred orientations. Crystallographymay provide clari¢cation of this issue.

Acknowledgements

This work was supported by grants from theAmerican Heart Association (New Mexico a⁄liate),the National Science Foundation REU program andby US Army/DOD Predoctoral Fellowships (W.M.B.and J.A.D.).

References

[1] C.A. Orengo, D.T. Jones, J.M. Thornton, Protein superfa-milies and domain superfolds, Nature 372 (1994) 631^634.

[2] J.S. Richardson, The anatomy and taxonomy of proteinstructure, Adv. Protein Chem. 34 (1981) 167^339.

[3] A.G. Murzin, S.E. Brenner, T. Hubbard, C. Chothia, SCOP:a structural classi¢cation of proteins database for the inves-tigation of sequences and structures, J. Mol. Biol. 247 (1995)536^540.

[4] D.L. Ollis, E. Cheah, M. Cygler, B. Dijkstra, F. Frolow,S.M. Franken, M. Harel, S.J. Remington, I. Silman, J.

Schrag, J.L. Sussman, K.H.G. Verschueren, A. Goldman,The alpha/beta hydrolase fold, Protein Eng. 5 (1992) 197^211.

[5] M. Cygler, J.D. Schrag, J.L. Sussman, M. Harel, I. Silman,M.K. Gentry, B.P. Doctor, Relationship between sequenceconservation and three-dimensional structure in a large fam-ily of esterases, lipases, and related proteins, Protein Sci. 2(1993) 366^382.

[6] J.D. Schrag, M. Cygler, Lipases and K/L hydrolase fold,Methods Enzymol. 284 (1997) 85^107.

[7] J.D. Schrag, Y. Li, S. Wu, M. Cygler, Ser-His-Glu triadforms the catalytic site of the lipase from Geotrichum candi-dum, Nature 351 (1991) 761^764.

[8] P. Grochulski, F. Bouthillier, R.J. Kazlauskas, A.N. Serreqi,J.D. Schrag, E. Ziomek, M. Cygler, Analogs of reactionintermediates identify a unique substrate binding site in Can-dida rugosa lipase, Biochemistry 33 (1994) 3494^3500.

[9] D. Ghosh, Z. Wawrzak, V.Z. Pletnev, N. Li, R. Kaiser, W.Pangborn, H. Jornvall, M. Erman, W.L. Duax, Structure ofuncomplexed and linoleate-bound Candida cylindracea cho-lesterol esterase, Structure 3 (1995) 279^288.

[10] X. Wang, C.-S. Wang, J. Tang, F. Dyda, X.C. Zhang, Thecrystal structure of bile salt activated lipase: insights intothe bile salt activation mechanism, Structure 5 (1997)1209^1218.

[11] J.C.-H. Chen, L.J.W. Miercke, J. Krucinski, J.R. Starr, G.Saenz, X. Wang, C.A. Spilburg, L.G. Lange, J.L. Ellsworth,R.M. Stroud, Structure of bovine pancreatic cholesterol es-terase at 1.6 Aî : novel structural features involved in lipaseactivation, Biochemistry 37 (1998) 5107^5117.

[12] R.L. Kingston, H.M. Baker, K.M. Loomes, L. Blackberg,O. Hernell, E.N. Baker, Crystallization and preliminary X-ray analysis of native and recombinant human bile-salt de-pendent lipase, strategies for improvement of di¡ractionquality, Acta Crystallogr. D 56 (2000) 478^480.

[13] S. Terzyan, C-S. Wang, D. Downs, B. Hunter, X. Zhang,Crystal structure of the catalytic domain of human bile saltactivated lipase, Protein Sci. 9 (2000) 1783^1790.

[14] J.M. Bailey, L.L. Gallo, J. Gillespie, Inhibition of dietarycholesterol ester absorption by 3-BCP, a suicide inhibitorof cholesterol-esterase, Biochem. Soc. Trans. 23 (1995) 408S.

[15] B.R. Krause, D.R. Sliskovic, M. Anderson, R. Homan, Lip-id-lowering e¡ects of WAY-121,898, an inhibitor of pancre-atic cholesteryl ester hydrolase, Lipids 33 (1998) 489^498.

[16] S.M. Jeon, H.S. Kim, T.G. Lee, S.H. Ryu, P.G. Suh, S.J.Byun, Y.B. Park, M.S. Choi, Lower absorption of choles-teryl oleate in rats supplemented with Areca catechu L. ex-tract, Ann. Nutr. Metab. 44 (2000) 170^176.

[17] P.N. Howles, C.P. Carter, D.Y. Hui, Dietary free and esteri-¢ed cholesterol absorption in cholesterol esterase (bile salt-stimulated lipase) gene-targeted mice, J. Biol. Chem. 271(1996) 7196^7202.

[18] L.M. Deck, M.L. Baca, S.L. Salas, L.A. Hunsaker, D.L.Vander Jagt, 3-Alkyl-6-chloro-2-pyrones: selective inhibitorsof pancreatic cholesterol esterase, J. Med. Chem. 42 (1999)4250^4256.

BBAPRO 36527 1-5-02

M. Stoddard Hatch et al. / Biochimica et Biophysica Acta 1596 (2002) 381^391 389

[19] R.B. Westkaemper, R.H. Abeles, Novel inactivators of ser-ine proteases based on 6-chloro-2-pyrone, Biochemistry 22(1983) 3256^3264.

[20] W.A. Boulanger, J.A. Katzenellenbogen, Structure-activitystudy of 6-substituted 2-pyranones as inactivators of K-chy-motrypsin, J. Med. Chem. 29 (1986) 1159^1163.

[21] G.M. Morris, D.S. Goodsell, R.S. Halliday, R. Huey, W.E.Hart, R.K. Belew, A.J. Olson, Automated docking using aLamarckian genetic algorithm and empirical binding freeenergy function, J. Comp. Chem. 19 (1998) 1639^1662.

[22] G.M. Morris, D.S. Goodsell, R. Huey, A.J. Olsen, Distrib-uted automated docking of £exible ligands to proteins: par-allel applications of AutoDock 2.4, J. Comput.-Aided Mol.Des. 10 (1996) 293^304.

[23] W. Humphrey, A. Dalke, K. Schulten, VMD-Visual Molec-ular Dynamics, J. Mol. Graphics 14 (1996) 33^38.

[24] S.G. Bhat, H.L. Brockman, The role of cholesteryl ester hy-drolysis and synthesis in cholesterol transport across rat in-testinal mucosal membrane. A new concept, Biochem. Bio-phys. Res. Commun. 109 (1982) 486^492.

[25] Y. Huang, D.Y. Hui, Metabolic fate of pancreas-derivedcholesterol esterase in intestine: An in vivo study usingCaco-2 cells, J. Lipid Res. 31 (1990) 2029^2037.

[26] R. Shamir, W.J. Johnson, R. Zolfaghari, H.S. Lee, E.A.Fisher, Role of bile salt-dependent cholesteryl ester hydro-lase in the uptake of micellar cholesterol by intestinal cells,Biochemistry 34 (1995) 6351^6358.

[27] C.A. Spilburg, D.G. Cox, X. Wang, B.A. Bernat, M.S. Bos-ner, L.G. Lange, Identi¢cation of a species speci¢c regula-tory site in human pancreatic cholesterol esterase, Biochem-istry 34 (1995) 3942^3947.

[28] M.S. Bosner, A.A. Wol¡, R.E. Ostlund Jr., Lack of e¡ect ofcholesterol esterase inhibitor CVT-1 on cholesterol absorp-tion and LDL cholesterol in humans, Cardiovasc. DrugsTher. 13 (1999) 449^454.

[29] N. Brueau, D. Lombardo, M. Bendayan, Participation ofGRP94-related protein in secretion of pancreatic bile salt-dependent lipase and in its internalization by the intestinalepithelium, J. Cell Sci. 111 (1998) 2665^2679.

[30] N. Brueau, D. Lombardo, M. Bendayan, The a⁄nity bind-ing sites of pancreatic bile salt-dependent lipase in pancreaticand intestinal tissues, J. Histochem. Cytochem. 48 (2000)267^276.

[31] L.L. Gallo, Y. Chiang, G.V. Vahouny, C.R. Treadwell, Lo-calization and origin of rat intestinal cholesterol esterasedetermined by immunocytochemistry, J. Lipid Res. 21(1980) 537^545.

[32] J. Nilsson, L. Blackberg, P. Carlsson, S. Enerback, O. Her-nell, G. Bjursell, cDNA cloning of human-milk bile-salt-stimulated lipase and evidence for its identity to pancreaticcarboxylic ester hydrolase, Eur. J. Biochem. 192 (1990) 543^550.

[33] J. Brodt-Eppley, P. White, S. Jenkins, D.Y. Hui, Plasmacholesterol esterase level is a determinant for an atherogeniclipoprotein pro¢le in normolipidemic human subjects, Bio-chim. Biophys. Acta 1272 (1995) 69^72.

[34] F. Li, D.Y. Hui, Modi¢ed low density lipoprotein enhancesthe secretion of bile salt-stimulated cholesterol esterase byhuman monocyte-macrophages: species-speci¢c di¡erencesin macrophage cholesteryl ester hydrolase, J. Biol. Chem.272 (1997) 28666^28671.

[35] F.W. Holtsberg, L.E. Ozgur, D.E. Garsetti, J. Myers, R.W.Egan, M.A. Clark, Presence in human eosinophils of a lyso-phospholipase similar to that found in the pancreas, Bio-chem. J. 309 (1995) 141^144.

[36] F. Li, D.Y. Hui, Synthesis and secretion of the pancreatic-type carboxyl ester lipase by human endothelial cells, Bio-chem. J. 329 (1998) 675^679.

[37] K.E. Winkler, E.H. Harrison, J.B. Marsh, J.M. Glick, A.C.Ross, Characterization of a bile salt-dependent cholesterylester hydrolase activity secreted from HepG2 cells, Biochim.Biophys. Acta 1126 (1992) 151^158.

[38] S. Lindquist, L. Blackberg, O. Hernell, Distribution of thebile-salt-stimulated lipase in human tissues, FASEB J. 15(2001) A193.

[39] M. Lotti, F. Grandori, F. Fusetti, S. Longhi, S. Brocca,A. Tramontano, L. Alberghina, Cloning and analysis ofCandida cylindracea lipase sequences, Gene 124 (1993) 45^55.

[40] M. Cygler, J.D. Schrag, Structure and conformational £ex-ibility of Candida rugosa lipases, Biochim. Biophys. Acta1441 (1999) 205^214.

[41] S.R. Feaster, D.M. Quinn, B.L. Barnett, Molecular model-ing of the structures of human and rat pancreatic cholesterolesterases, Protein Sci. 6 (1997) 73^79.

[42] D. Campese, D. Lombardo, L. Multigner, H. Lafont, A. DeCaro, Implications of a tyrosine residue in the unspeci¢c bilesalt binding site of human pancreatic carboxyl ester hydro-lase, Biochim. Biophys. Acta 784 (1984) 147^157.

[43] K.M. Loomes, H.E.J. Senior, Bile salt activation of humancholesterol esterase does not require protein dimerisation,FEBS Lett. 405 (1997) 369^372.

[44] S.R. Feaster, D.M. Quinn, Mechanism-based inhibitors ofmammalian cholesterol esterase, Methods Enzymol. 286(1997) 231^252.

[45] S.R. Feaster, K. Lee, N. Baker, D.Y. Hui, D.M. Quinn,Molecular recognition by cholesterol esterase of active siteligands: structure-activity e¡ects for inhibition by aryl car-bamates and subsequent carbamylenzyme turnover, Bio-chemistry 35 (1996) 16723^16734.

[46] G. Lin, C.-T. Shieh, H.-C. Ho, J.-Y. Chouhwang, W.-Y.Lin, C.-P. Lu, Structure-reactivity relationships for the inhi-bition mechanism at the second alkyl-chain-binding site ofcholesterol esterase and lipase, Biochemistry 38 (1999) 9971^9981.

[47] M.H. Gelb, R.H. Abeles, Mechanism of inactivation of chy-motrypsin by 3-benzyl-6-chloro-2-pyrone, Biochemistry 23(1984) 6596^6604.

[48] D. Ringe, J.M. Mottonen, M.H. Gelb, R.H. Abeles, X-raydi¡raction analysis of the inactivation of chymotrypsin by 3-benzyl-6-chloro-2-pyrone, Biochemistry 25 (1986) 5633^5638.

BBAPRO 36527 1-5-02

M. Stoddard Hatch et al. / Biochimica et Biophysica Acta 1596 (2002) 381^391390

[49] D. Ringe, B.A. Seaton, M.H. Gelb, R.H. Abeles, Inactiva-tion of chymotrypsin by 5-benzyl-6-chloro-2-pyrones: 13CNMR and X-ray di¡raction analysis of the inactivator-en-zyme complex, Biochemistry 24 (1985) 64^68.

[50] J.C. Powers, C.M. Kam, Isocoumarin inhibitors of serinepeptidases, Methods Enzymol. 244 (1994) 442^457.

[51] S.B. Daniels, J.A. Katzenellenbogen, Halo enol lactones:studies on the mechanism of inactivation of alpha-chymo-trypsin, Biochemistry 25 (1986) 1436^1444.

[52] C.R. Borja, G.V. Vahouny, C.R. Treadwell, Role of bile and

pancreatic juice in cholesterol absorption and esteri¢cation,Am. J. Physiol. 206 (1964) 223^228.

[53] L.L. Gallo, S.B. Clark, S. Myers, G.V. Vahouny, Cholester-ol absorption in rat intestine: Role of cholesterol esteraseand acyl coenzyme A:cholesterol acyltransferase, J. LipidRes. 25 (1984) 604^612.

[54] A. Lopez-Candales, M.S. Bosner, C.A. Spilburg, L.G.Lange, Cholesterol transport function of pancreatic choles-terol esterase: Directed sterol uptake and esteri¢cation inenterocytes, Biochemistry 32 (1993) 12085^12089.

BBAPRO 36527 1-5-02

M. Stoddard Hatch et al. / Biochimica et Biophysica Acta 1596 (2002) 381^391 391