Báo cáo Y học: Myristyl and palmityl acylation of pI 5.1 carboxylesterase from porcine intestine and liver Tissue and subcellular distribution potx

Myristyl and palmityl acylation of pI 5.1 carboxylesterase from porcine intestine and liver Tissue and subcellular distribution Sylvie Smialowski-Fle  ter, Andre  Moulin, Josette Perrier and Antoine Puigserver Institut Me  diterrane  en de Recherche en Nutrition, UMR-INRA, Faculte  des Sciences et Techniques de St-Je  ro à me, Marseille, France Immunoblotting analyses revealed the presence o f carb- oxylesterase in the porcine small intestine, liver, submaxillary and parotid glands, kidney cortex, lungs and cerebral cortex. In the intestinal mucosa, the pI 5.1 enzyme was d etected in several subcellular fractions including the microvillar frac- tion. Both fatty monoacylated and diacylated monomeric (F1), trimeric (F3) and tetrameric (F4) forms of the intestinal protein were puri®ed here for t he ®rst time by performing hydrophobic chromatography and gel ®ltration. The molecular mass of t hese three e nzymatic forms w as estimate d to be 60, 180 and 240 kDa, respectively, based on size- exclusion chromatography and SDS/PAGE analysis. The existence of a covalent attachment linking palmitate and myristate to porcine intestinal carboxylesterase (PICE), which was suggested by the results of gas-liquid chroma- tography (GLC) experiments in which the fatty acids resulting from alkali treatment of the protein forms were isolated, was con®rmed here by the fact that [ 3 H]palmitic and [ 3 H]myristic acids were incorporated into porcine enterocytes and hepatocytes in cell primary cultures. Besides these two main fatty a cids, the p resence of oleic, ste aric, and arachidonic acids was also detected by GLC and further con®rmed by performing radioactivity counts on the 3 H- labelled PICE forms after an immunoprecipitation proce- dure using speci®c polyclonal antibodies, followed by a SDS/ PAGE separation step. Unlike the F1 and F4 forms, w hich were both myristoylated and palmitoylated, the F3 form was only palmitoylated. The monomeric, trimeric and tetrameric forms o f PICE w ere all able to hydrolyse short chain f atty acids containing glycerides, as well as phorbol esters. The broad speci®city of fatty acylated carboxylesterase is dis- cussed in terms of its possible involvement in the metabolism of ester-containing xenobiotics and signal transduction. Keywords: carboxylesterase; fatty acylation; gas-liquid chromatography; porcine enterocytes; porcine hepatocytes. Carboxylesterases (EC 3.1.1.1), which are f ound in many vertebrates, insects, plants and mycobacteria, have been reported to be involved in xenobiotic metabolism due to their ability to hydrolyse a number o f substrates co ntaining ester, thioester and amide bonds [1±3]. As some carboxy- lesterase (CE) isoenzymes display lipase -like activity, it has been suggested that they might play a part in lipid metabolism [4]. Moreover, the t wo CE forms with p I values of 5.2 and 5.6, which have be en isolated from rat liver [5], are known to deacylate the s tructural analog of diacylglyc- erol 4- b-phorbol-12-b-myristate-13-a-acetate (PMA). It has therefore been suggested that these enzymes may have activating ef fects on protein kinase C [5,6]. A porcine intestinal carboxylesterase (PICE) was re- cently puri®ed to homogeneity and found to consist of a single isoform with a pI of 5.1, based on isoelectric focusing data [7]. The amino-acid sequence deduced from the cloned cDNA consisted of 565 residues and showed 97% identity with that of porcine liver carboxyle sterase (PLCE) [8], a protein which belongs to the GXSXG family of serine proteases, and more than 50% identity with those of other CE from various mammalian species [9±11]. The molecular mass of the porcine intestinal mucosa enzyme was estimated t o be 240 kDa by size- exclusion chromatography, and 60 kDa using S DS/PAGE under both reducing and nonreducing conditions [7], which strongly suggests that the protein consisted of four apparently identical and active polypeptide subunits, unlike other mammalian CE which are known to be monomeric or t rimeric e nzymes [12]. The two disul®de bridges present in PICE were recently assigned to Cys70± Cys99 ( loop A) and C ys256±Cys267 (loop B), whereas the ®fth Cys71 r esidu e was thought to be blocked rather t han being p resent in the free f orm, from the lack o f a lkylation with iodoacetamide [13]. In the present study, we report on the tissue and subcellular distribution of PICE using speci®c polyclonal antibodies and by purifying three active molecular f orms of the enzyme, and show for the ®rst time that all these forms are both m yristoylated and p almitoylated. Correspondence to A. Puigserver, Institut Me  diterrane  en de Recherche en Nutrition, UMR-INRA 1111, Faculte  des Sciences et Techniques de St-Je  roà me, Av enue Escadrille Normandie Niemen, F-1 3397 Marseille cedex 20, Franc e. Fax: + 33 4 91 28 84 40, Tel.:+33491288838, E-mail: antoine.puigserver@lbbn.u-3mrs.fr Abbreviations: CE, carboxylesterase; DEAE, diethylaminoethyl; FA, fatty acid; GLC, gas-liquid chromatography; KLH, keyhole limpet haemocyanin; P ICE, porcine intestinal c arboxylesterase; PLCE, porcine liver carboxylesterase; PMA, 4-b-phorbol-12-b-myristate- 13-a-acetate; pN PA, p-nitrophenylacetate; PVDF, poly(vinylidene di¯uoride). Enzyme: carboxylesterase (EC 3.1.1.1). (Received 3 August 2001, revised 22 N ovember 2001 , accepted 2 7 November 2001) Eur. J. Biochem. 269, 1109±1117 (2002) Ó FEBS 2002 MATERIALS AND METHODS Tissues and reagents All the pig organs used here w ere obtained from the local slaughterhouse. EAH±Sepharose 4B, Octyl±Sepharose, DEAE±Sepharose Fast Flow, Sephacryl S-200 (allyl dextran and N,N-methylene-bisacrylamide matrix, 2.6 ´ 60 cm column), Superdex 200 HR (dextran/agarose matrix, 1.0 ´ 30 cm column), [9,10(n)- 3 H]myristic acid (speci®c activity 54 Ciámmol A1 ) and [9,10(n)- 3 H]palmitic acid (speci®c activity 54 Ciámmol A1 ) were purchased from Pharmacia Biotech. Substrates, BSA, 5-bromo-4-chloro-3- iodo-phosphate, Ponceau S , c arbodiimide, protein A± Sepharose, prednisolone, glucagon, insulin and SDS were from Sigma Chemical Co. Chloroform and butanol were provided by SDS (Peypin, France), while methanol was purchased from Carlo Erba. Williams E medium, fetal bovine serum, penicillin and streptomycin were obtained from Gibco-BRL. The nitrocellulose sheets (0.2 lm) were from Schleicher and Schuell and the IgG fraction of goat anti-(rabbit IgG) serum conjugated with peroxidase was from Organon Teknika Corporation-Cappel. Enzyme and protein assays Enzyme activity determin ations were performed t itrimetri- cally on tributyrin (65 m M ), butyrylcholine (132 m M ), a-naphthylacetate (26 m M ) and phorbol diester (1.3 m M ) at 37 °C using a Metrohm (Herisau, S witzerland) pH-stat (718 STAT Titrino, Radiometer) and 0.01 M NaOH as described p reviously [7]. One unit of enzymatic activity corresponds to 1 lmol fatty a cid released per min. All the activities were measured at pH 8.0, including that towards p-nitrophenylacetate [14]. In order to determine the activities of aminopeptidase N (a microvillous membrane marker) (Na + /K + )-ATPase (a basolateral plasma membrane marker), NADPH-cytochrome c reductase (a microsomal contamination marker), cytochrome c oxydase (a mito- chondrial marker) and acid phosphatase (a lysosome marker) during subcellular fractionation experiments, we used methods which have been described in previous papers [15±19]. Protein concentrations were determined as described by B radford [20]. Peptide synthesis and puri®cation The peptide KMKFLTLDLHGDPRE, corresponding to the amino-acid sequence from positions 281±296 on the PICE polypeptide chain, was s ynthesized by the M arseille CNRS-CIML laboratory using an Applied Biosystem Peptide Synthesizer 431 A. The peptide was puri®ed by RP-HPLC on a Kontron apparatus equipped with an ALLTIMA C18 column (4.6 mm ´ 25 cm), and its molec- ular mass was determined on an Applied Biosystems MALDI-TOF Voyager D E RP mass s pectrometer. Preparation of polyclonal antiserum The peptide was covalently a ttached to keyho le limpet haemocyanin (KLH) from the mollusc Concholepas con- cholepas with 2% glutaraldehyde. The resulting peptide± KLH conjugate was dialyzed, lyophilized and further used (1 mg dissolved in 0.5 m L N aCl/P i with complete Freund's adjuvant) to immunize New Zealand INRA 1077 male white rabbits by subcutaneous injection. Three weeks later, the same a mount of peptide±KLH conjugate emulsi®ed with incomplete F reund's adjuvant w as injected intramus- cularly. After a further 10-day p eriod, 0.5 mg antigen was injected subcutaneously, and the same amount of antigen was then injected intravenously on the following day. Finally, 10 days later, blood was collected from the marginal ear v e in, allowed to clot f or 1 h and s uccessively centrifuged at 3000 g for 10 min and at 15 000 g for 15 min. The immune serum was then c ollected , ®ltered and stored at 4 °C. Puri®cation of polyclonal antibodies and immunoprecipitation experiments The PICE K281±E296 peptide (30 mg) was covalently linked to EAH-Sepharose 4B (30 m gámL A1 gel) using 0.1 M carbodiimide according t o Pharmacia Biotech instru ctions. The antibodies speci®cally bound to the immobilized peptide w ere elute d w ith a 0.5 M NaCl containing 0.1 M acetate buffer ( pH 3.5), immediately neutralized with 1 .5 M Tris/HCl buffer (pH 9.3) in the presence of 0 .5 M NaCl to prevent protein denaturation, and ®nally stored at 4 °C. Antibody titration and speci®city determination were per- formed using a conventional ELISA assay [21]. The puri®ed anti-PICE Ig were then covalently linked to protein A±Sepharose gel as previously described [22], and used to precipitate the tritiated protein. Molecular mass determination and immunoblot analyses SDS/PAGE was carried out using Laemmli's m ethod [23]. Proteins were electrotransferred overnight o nto a nitrocel- lulose sheet at 50 V in 20 m M Tris/HCl buffer (pH 8.5) containing 150 m M glycine and 20% ethanol. The nitrocel- lulose membrane was subsequently saturated with 10% BSA before being incubated with 1: 100 (v/v) diluted rabbit anti-PICE immune serum, and the reacting antibodies were further detected with 800-fold diluted peroxidase-conjugat- ed goat anti-(rabbit IgG) Ig. Both radiolabelled proteins and immune protein p recipitates were separated by p erforming electrophoresis on a 12% polyacrylamide gel in t he presence of SDS, stained with Coomassie Blue for gel-slicing and scintillation counting and/or subjected to immunoblot analysis. Preparatory to the radioactivity assays, the sliced gels and i mmunoblots were s olubilized in 0.5 mL of a 30% (w/v) hydrogen peroxide solution for 5 h at 95 °C, and the count was performed in 5 mL scintillation ¯uid on a Packard-Tri-Carb Model 2100 TR liquid s cintillation spectrometer as p reviously described [24]. Isoelectric focus- ing (IEF) was p erformed as described b y R obertson et al . [25]. Tissue distribution and intestinal mucosal subcellular fractionation of carboxylesterase Pig organs were dissected out and immediately homogen- ized in 20 m M Tris/HCl buffer, pH 7.3, containing 0.25 M sucrose, 10 m M KCl, 1 m M MgCl 2 ,1l M phenylmethane- sulfonyl ¯uoride and 1 m M benzamidine. The homogenates were subsequently centrifuged at 10 000 g for 2 0 m in, the 1110 S. Smialowski-Fle  ter et al. (Eur. J. Biochem. 269) Ó FEBS 2002 supernatant w as again centrifuged a t 105 000 g and a t 4 °C for an additional 45-min period, and the resulting superna- tant was ®nally used for the anti-PICE Ig staining proce- dure. Crude brush-border membrane preparations were obtained from the subcellular fraction of the intestinal mucosa as previously described [26,27]. Brie¯y, pig intestinal mucosal scrapings were homogenized in four times their mass of a 5-m M Tris/HCl buffer ( pH 7.3) containing 0.25 M sucrose, 10 m M KCl and 1 m M MgCl 2 , ®ltered through a gauze and further subje cted to d ifferential centrifugation to obtain the membrane fraction. PICE puri®cation Fresh porcine intestine was s craped off and the muco sa was either i mmediately used o r frozen at A80 °C until use. The lipids w ere p artially extracted from about 200 g of the mucosa by placing them in a chloroform/butanol mixture (9 : 1, v /v). After a homogenization step in 500 mL of 20 m M Tris/HCl, 0.35 M NaCl at pH 8.0, centrifugation was performed at 10 000 g for1handproteinsfromthe resulting s upernatant ( S1) w ere precipitated by adding solid (NH 4 ) 2 SO 4 to the solution (0.7 M ®nal concentration) under gentle stirring at 4 °C for 2 h . A fter a ®rst centrifugation at 10 000 g for 30 min, the pellet was dissolved in 100 mL o f 20 m M Tris/HCl pH 8.0 c ontaining 0.7 M (NH 4 ) 2 SO 4 and then dialysed against the buffer. A second centrifugation took place under t he same experimental conditions and t he resulting s upernatant was applie d to a n octyl±Sepharose gel equilibrated with the same buffer and the proteins were eluted with a 20-m M Tris/HCl buffer, pH 8.0, containing 0.4 M NaCl (buffer A). The active proteins eluted were successively precipitated with 80% (w/v) ammonium sulfate at 4 °C, and after be ing centrifuged at 10 000 g for 30 m in, theyweredissolvedin10mLofbufferAanddialysed overnight against the same buffer. The dialysate was then applied to a DEAE-Sepharose Fast Flow column (1.5 ´ 14 cm) equilibrated w ith buffer A , an d the p roteins were eluted with a linear 0.1±0.3 M NaCl gradient. The active fractions on tributyrin were ®nally applied to a Superdex 200 HR gel c olumn (1.0 ´ 30 cm) and eluted with a20-m M Tris/HCl buffer pH 8 .0 containing 0.35 M NaCl, at a ¯ow rate of 0.5 mLámin A1 . Molecular mass determination This was achieved by performing gel ®ltration on a Sephacryl S -200 column (2.6 ´ 60 cm) and the proteins were eluted with 0.35 M NaCl in 20 m M Tris/HCl, pH 8.0, and by SDS/PAGE on a 12% (w/v) gel as previously described [23]. The electrophoretic molecular-mass markers (14.4±97 k Da) and isoelectric focusing calibration kit (pH 4 .5±9.6) were obtained from B io-Rad Laboratories. The MW-GF 1 000 kit ( 29±2000 kDa) f rom Sigma Chem- ical Co, was used for the g el ®ltration procedure. Amino-acid composition and sequence determination The amino-acid composition of t he puri®ed PICE was determined using a Waters chromatography system as previously described [13,28], after 24 h hydrolysis in 6 M HCl at 110 °C. The amino-acid sequence of the Ponceau red-stained proteins was determined by performing Edman degradation on an Applied Biosystems Model 470 A protein gas-phase sequencer [29]. Lipid extraction and fatty acid identi®cation The lipids present in the puri®ed PICE and porcine serum albumin u sed a s the control substance were completely removed with chloroform/methanol/water (2 : 2 : 1 .8, v /v/ v) as described by Bligh & Dyer [30]. The covalently b ound fatty acids were released from the protein under alkaline conditions. E thanol containing 1 M KOH w as used and t he protein solution was i ncubated a t 8 0 °C for 1 h , a nd then dried under a stream of nitrogen. After adding the same amount of water, the aqueous layer was acidi®ed with HCl and the free fatty acids were extracted with hexane and dried before performing methanolysis at 100 °C for 1 h using 1 4% BF 3 in methanol [31]. After the methylation, the fatty acids were identi®ed on a PerkinElmer gas-liquid chromatogra- phy autosystem XL equ ipped with P erkinElmer integrator 1022S, using n-heptadecanoic acid as an internal standard. Cell cultures and labelling Mature porcine enterocytes (16 ´ 10 6 cellsámL A1 )and hepatocytes (3.4 ´ 10 6 cellsámL A1 )wereisolatedas described by Bader et al. [32] and by Seglen [33], respec- tively. Prior to the labelling experiments, enteroc ytes and hepatocytes were i ncubated for 4 h at 37 °C in W illiams E medium supplemented with 5% (v/v) fetal bovine serum, prednisolone (5 lmoláL A1 ), glucagon (0.014 lgámL A1 ), insulin (0.16 Uá mL A1 ), penic illin ( 200 U ámL A1 ), streptomy- cin (200 lgámL A1 )and63lCiámL A1 of [9,10(n)- 3 H]myristic acid or [9,10(n)- 3 H]palmitic acid (speci®c a ctivity 54 Ci ámmol A1 ). Prior to use, the fetal bovine serum was delipidated using 1,2,2-trichloro-1,2,2-tri¯uoroethane [34]. At the e nd of the labelling period, cells were aspirated from the dishes and centrifuged at 900 g for 5 min Cell pellets were then washed extensively in NaCl/P i , homogenized, centrifugated at 10 000 g for 10 min at 4 °C, and the supernatant was sampled for analysis. RESULTS Tissue distribution of pI 5.1 carboxylesterase The presence of the pI 5.1 CE isoform i n 1 1 homogenates from various porcine tissues was checked by performing immunoblot analysis on the soluble e xtracts u sing puri®ed polyclonal antibodies directed against a synthetic amino- acid peptide corresponding to the amino-acid sequence located between residues 281 and 296 in the PICE polypeptide chain. Figure 1 shows t hat these antibodies speci®cally revealed a 60-kDa band corresponding to the pI 5.1 CE in the soluble extracts from small intestine, parotid and submaxillary glands, liver, kidney cortex, lung and brain cortex. The highest level of expression of pI 5.1 CE was observed in the liver, followed by the small intestine, but it is worth noting t hat the enzym e was not detected in the soluble extracts of homogenates from colon, stomach, pancreas and kidney medulla, or in those from skin, bladder, tongue, trachea, b rain medulla and cerebellum, heart, pharynx and suprarenals (data not shown). Esterase activity on tributyrin was observed only in t he so luble Ó FEBS 2002 Carboxylesterase fatty acylation (Eur. J. Biochem. 269) 1111 fractions from small intestine, colon, liver and pancreas homogenates. In the latter homogenates, the activity observed was presumably that of lipase, although the presence of some contaminating activity due to the presence of microorganisms i n the colon c ould not be ruled out. Subcellular distribution of porcine intestinal carboxylesterase The distribution o f PICE activity on t ributyrin and that o f marker enzymes on their speci®c s ubstrates in a number of subcellular f ractions from pig i ntestinal mucosa is g iven in Table 1 . Esterase activity on t ributyrin was detected in four subcellular f ractions, a nd in all these fractions, immunoblot analysis using the puri®ed polyclonal a nti-PICE Ig yielded a single band at 60 kDa. The highest level of activity was observed in the microsomal and soluble fractions, which yielded 41% and 32% of the total enzyme activity, respectively. As the microvillar frac tion accounted for as much as 18% of t he overall esterase a ctivity on t ributyrin, it was suggested that some of the PICE might correspond to an enterocytic b rush border membrane protein. Puri®cation of the molecular forms of porcine intestinal CE Figure 2A shows the PICE elution pro®le systematically obtained with a Sephacryl S-200 gel ®ltration column, whether the puri®cation procedure was that used in the present study or that described by David et al.[7].Asingle molecular form (F4) was obtained, which showed the presence of a single 60-kDa band with a pI value of 5.1 upon SDS/PAGE analysis under reducing and nonreducing conditions and isoelectric focusing. When the F4 molecular form was further puri®ed using a Superdex 200 HR gel ®ltration c olumn, two distinct molecular forms (F3 and F1) were separated (Fig. 2B ). Based on the elution pro®les of standard proteins, the apparent molecular mass of t hese forms w as found to be 180 k Da and 60 kDa, respectively. Surprisingly, the dimeric molecular form F2 was not observed. Again, SDS/PAGE and isoelectric focusing analysis showed that F3 and F1 corresponded to a single 60-kDa band with a pI of 5 .1 (Fig. 3). Our results and those obtained by David et al.[7] strongly suggested the existence o f a single polypeptide chain c orresponding to the monomeric form of P ICE (F1) and giving r ise to t he noncovalent association of three and four apparently identical subunits corresponding to the F3 and F4 molecular forms of PICE, r espectively. PH- and substrate-dependent activity of PICE molecular forms At pH 8.0, which was used for running both the gel ®ltration experiments and esterase activity assays on trib- utyrin, F4 a nd F1 were found to have similar speci®c activity values on tributyrin as substrate (% 290 Uámg A1 protein), whereas F3 was about threefold less active (Table 2). At pH 6 .5, however, F4 and F3 were found to have almost equal levels of esterase activity on tributyrin, whereas F1 was slightly less active. Overall, at the more acidic pH value, the three forms were 30±40% less active than at the more basic pH value. A number of ester- containing compounds including p-nitrophenylacetate, a-naphthylacetate and butyrylcholine were also t ested as possible substrates at pH 8.0 (Table 2 ). The tetrameric, Fig. 1. Immunoblotting an d esterase a ctivity on tributyrin o f pI 5.1 CE from porcine tissues. Esterase activity was measured as indicated in the Materials and Methods section. Total proteins (30 lg) presen t in homogenates from 11 p orcine tissues were electrop ho resed in a 12% SDS/PAGE, transferred o nto a nitrocellulose memb rane, and reacted with polyclonal antibodies raised against the syn thetic peptide corre- sponding to the amino-acid sequence extending from K281 to E296 of the PICE polypeptide chain. Lane 1, small intestine; lane 2, colon; lane 3, stomach; la ne 4, parotid; lane 5, submaxil lary; lane 6, liver; lane 7, pancreas; lane 8, kidney cortex; lan e 9, kidney m edulla; lane 10, lung; lane 11, brain cortex. Table 1. Subcellular localization of CE activity in porcine intes tinal mucosa. At each st ep of the s ubcellular fractio nation procedure, t he enzyme activities were measured in the pellet and the s up ernatant and expressed as a percentage of t he total activity. The values are means based on three separate subcellular fractionations. Fraction Enzymatic marker activities (%) Subcellular fraction Esterase activity on tributyrin (%) Immunoblot analysis a 10 000 g pellet Cytochrome c oxidase (70  5) Mitochondrial 10  3 + CaCl 2 pellet NADPH/H + Cytochrome c reductase (80  3); Microsomal and 40  5 + Na + /K + ATPase (73  7) basolateral membranes 105 000 g pellet Aminopeptidase N (78  3) Microvillar 18  6 + Final supernatant Acid phosphatase (81  7) Soluble 32  7 + a Presence (+) of immunoreactive PICE detected with polyclonal antibodies directed against the PICE K281±E296 peptide. 1112 S. Smialowski-Fle  ter et al. (Eur. J. Biochem. 269) Ó FEBS 2002 trimeric and monomeric forms of PICE were found to display different enzymatic activities on these substrates. Although F4 a nd F1 were equally active on tributyrin, as already i ndicated i n Table 2, the latter form was roughly 2±3 times more active than the former on the other three substrates, namely a-naphthylacetate, p-nitrophenylacetate and butyrylcholine. It is worth noting that F 3 w as the most active on butyrylcholine and the least active on tributyrin, and that p-nitrophenylacetate is apparently the most ef®cient substrate f or PICE molecular f o rms in general. N-Terminal amino-acid sequence and fatty acid content of PICE molecular forms Table 3 gives t he fatty acid co ntent of t he F4, F3 and F1 molecular forms o f PICE, as well as the N-terminal amino- acid sequence of F3, in addition to that of F4 previously reported by David et al. [7]. The nine ®rst amino acids of the F4 and F3 polypeptide chains were found to be identical, whereas in F1, no N-terminal amino acid could be d etected, which strongly suggests that the polypeptide chain was Fig. 3. Polyacrylamide gel electrophoresis of PICE molecular forms. (A) The electro phoresis was carried out on a 12% polyacrylamide gel in the presence of SDS under reducing conditions. (B) IEF was per- formed at pH 4±9 with a calibration kit (pH 4.46±9.6) and silver staining. Fig. 2. Gel ®ltration of porcine intestinal CE molecular forms. (A) Porcine i ntestinal CE, puri® ed as indicated in Materials and methods or as described by David et al. [ 7], was applied to a Sephacryl-S200 column (2.6 ´ 60 cm) and eluted with 20 m M Tris/HCl buer con- taining 0.35 M NaCl, pH 8.0. (B) The F4 molecular form w as then applied to a Superdex 200-HR column (1.0 ´ 30 cm) and eluted with the above-mentioned buer. Esterase activity on tributyrin was assayed as indicated in Materials and methods. Solid and d otted lines represent the protein absorban ce at 280 nm and the esterase activity on tributyrin, respectively. Table 2. Substrate-dependent activity of PICE molecular forms. 100 % speci®c activity on tributyrin at pH 8.0 corresponds to 290 U ámg protein A1 for both the F4 and F1 f orms. All the results are m eans on three enzymatic determinations. Substrate Activity determination pH Relative speci®c activity (%) F4 F3 F1 Tributyrin 8.0 100 30 100 6.5 40 40 27.5 a-Naphthylacetate 8.0 42 74 85 p-Nitrophenylacetate 8.0 103 123 170 Butyrylcholine 8.0 15 60 52 Ó FEBS 2002 Carboxylesterase fatty acylation (Eur. J. Biochem. 269) 1113 blocked. As the amino-acid composition of the three molecular f orms of PICE was found to have remained unchanged, these data are not shown. As far as fatty acylation is concerned, both the F4 and F1 forms of P ICE were found to have fairly similar FA p ro®les in sharp c ontrast with the F3 form ( Table 3). Myristic and palmitic acids were the predominantly linked FA, while a number of minor fatty acids including stearic, oleic and arachidonic acids could also be detected. It is worth noting that myristic acid was not released from the F3 form after alkaline hydrolysis, contrary to what was observed in the case of the F4 and F1 forms. The quantitative determination of fatty acids released from a given molecular form of P ICE relative to the amount of protein deduced from its amino- acid composition indicated t hat s toichiometric amounts of myristic and palmitic acids were present in F1 (1 mol FA per mol of F1). B y contrast, less m yristic acid than palmitic acid was detected in F4 (0.2±0.4 mol of myristic acid as compared to 1 mol palmitic acid per mol of F4). PICE acylation in enterocyte and hepatocyte cell cultures Figure 4A shows the SDS/PAGE protein bands and the immunoblot pro®le obtained u sing the polyclonal antibod- ies raised against the K281±E296 amino-acid sequence of PICE, with the soluble proteins from enterocyte primary cell cultures in the presence of labelled [ 3 H]palmitic acid and [ 3 H]myristic acid. A single band corresponding to a protein with a m olecular m ass o f % 60 kDa was observed in b oth cases in t he enterocytic c ells. The patte rn o f r adioactivity in the gel slices showed a good correlation with the relative mobility of the immunoreactive PICE (Fig. 4A). About a four-fold higher level of 3 H radioactivity was counted in the Table 3. N-terminal amino-acid sequence and fatty acid c ontent o f PICE molecular forms. Molecular forms N-Terminal sequence Fatty acid content Major c Minor d F4 NH2-GQPASPPVV a C14:0 ; C16:0 C18:0 ; C18:1 ; C20:4(n-6) F3 NH2-GQPASPPVV b C16:0 C18:0 ; C18:1 ; C20:4(n-6) F1 Blocked C14:0 ; C16:0 C18:0 ; C18:1 ; C20:4(n-6) a From David et al. [7], and with an amino-acid sequence yield of about 0.5 mol glycine per mol of protein. b Yield of about 0.9 mol glycine per mol of protein. c About 1 mol fatty acid per mol of protein, except for F4 (0.2±0.4 mol myristic acid per mol of protein). d Less than 0.1 mol fatty acid per mol of protein. Fig. 4. SDS/PAGE, immunoblotting and pattern of radioactivity obtained with s oluble proteins from enterocytes (A) and hepatocytes (B) incubated with [ 3 H]fatty acid. 1, Ponceau red protein staining; 2, immunoblotting with polyclonal antibodies dire cted against the PICE peptide K 281 to E 296 (the arrow indi- cates the position of immunoreactive PICE). The r elative m obilities o f molecular mass markersinSDS/PAGEareindicatedbelow the b lot: (a) phospho rylase b (97 kDa); ( b ) albumin (66 kDa); (c) ovalbumin ( 45 kDa); (d) carbonic anhydrase (30 kDa); (e) trypsin inhibitor ( 20.1 kDa); and (f) a-lactalbumin (14.4 kDa). 1114 S. Smialowski-Fle  ter et al. (Eur. J. Biochem. 269) Ó FEBS 2002 PICE labelled with [ 3 H]myristic acid than in that labelled with [ 3 H]palmitic acid, in agreement with the a bove ®nding. In order to check whether the labelling was really due to PICE and not to another protein with the same molecular mass, the enzyme from enterocyte homogenates was immunoprecipitated with the puri®ed polyclonal antibodies. A single band at % 60 kDa which contained [ 3 H]palmitic (2.6 ´ 10 3 d.p.m.) or [ 3 H]myristic acids (2.1 ´ 10 3 d.p.m.) was revealed in the blot (data not shown). As the CE from porcine intestine and liver show 97% amino-acid sequence identity, and as the puri®ed speci®c polyclonal antibodies raised against PICE c ross-react w ith PLCE, w e e xtended the fatty acylation analysis to hepato- cytes.AsshowninFig.4B,asinglebandwithanapparent molecular mass of 60 kDa, corresponding to PLCE, was revealed by the speci®c polyclonal antibodies and the protein was l abe lled by m yristic or palmitic acids. An alysis of the radioactivity patterns in the blot showed the presence of a similar 3 H content in PLCE, whether the protein was l abelled with [ 3 H]palmitic acid or [ 3 H]myristic acid. DISCUSSION The results of immunoblot analysis performed on soluble extracts from porcine tissue homogenates showed that the pI 5.1 CE was present mainly in the liver, but also to a lesser extent in the small intestine, submaxillary and parotid glands, kidney cortex, lungs, and brain cortex. This CE isoform was not detected, however, in the other two main parts of t he digestive tract, namely the s tomach and colon, or in the pancreas and kidney medulla. The fact that the highest expression of the protein isoform was recorded in the liver might be due to the presence of several CE isoenzymes in this tissue [35,36] and to some lack of speci®city of the polyclonal antibodies used for t he analysis. However, the peptide extending from K 281 to E 296 in PICE was chosen as a speci®c antigen site because it showed more than 86% sequence identity with those from porcine liver [8], human liver [37] and human b rain [38] CE. A s esterase activity on tributyrin was detected only in the small intestine, liver, p ancreas, where lipase activity is known t o exist, and in the colon, it is suggested that there was no direct relationship between the presence of esterase activity on tributyrin in a given tissue and that of the pI 5.1 CE. Subcellular fractionation of the intestinal mucosa showed that PICE was unevenly distributed among the various fractions co rresponding to mitochondria, microsomes, microvilli and c ytosol. Although t he enzyme has previously been found to contain the tetrapeptide HAEL at the C-terminus of the polypeptide chain [7], w hich is thought to serve as a retention signal f or proteins on the luminal side of the ER, it is apparently not retained in the ER. Both the immunoblot analysis and e sterase activity o n tributyrin determinations showed that PICE was present in the cytosol fraction as well as in the mitochondrial and microvillar fractions, although the possible occurrence of some non- speci®c adsorption of PICE to subcellular membranes during the fractionation procedure cannot be ruled out. A tetrameric form of the porcine intestinal CE was recently puri®ed from a soluble protein fraction (105 000 g supernatant) and characterized [7,13]. In the present study, a separate puri®cation procedure was carried out from the total protein fraction (10 000 g supernatant) in order to isolate the membrane-bound enzyme. Both monomeric (60 kDa) and trimeric forms (180 kDa) could t herefore be isolated using a Superdex column, while the tetrameric form (240 kDa) which was isolated b y g el ®ltration on Sephacryl S-200 column corresponded to t hat previously described by David et al. [7]. Hydrophobic interactions may contribute signi®cantly to the polymerization of PICE monomers, due to the p resence of covalently bound fatty a cids, as suggested in Fig. 5. The interactions between the monomers in the tetrameric form F4 were apparently stronger than those occurring in the trimeric form F3, as no monomeric form F1 was observed in the elution pro®le f rom the Sephacryl S-200 column, in con trast to the pro®le of t he Superdex column (Fig. 2 ). The possibility that there may have been a difference in af®nity between the molecu lar forms depend- ing on the type of polysaccharide matrix used for gel ®ltration purposes cannot be ruled out. Similar results have been observed, for example, in the case of galectins and ricins, two groups of proteins known t o have lipolytic activities [39,40]. Whatever arguments may be put forward to explain the existence of several mo lecular forms in PICE, the behaviour of this protein on Sephacryl S-200 is comparable to that o f liver CE [8] but different from that of rat i ntestinal CE [41]. PICE was found to be more active on tributyrin at pH 8.0 than at pH 6.5, which is not surprising for a serine enzyme on account of the s tate of protonation of th e histidine residue from the catalytic triad. In addition, most of the F4 esterase activity on tributyrin at pH 8.0 was due to F 1, and as F3 was found to be threefold less active than both F 4 and F1, the interactions between monomers in F4 and F3 were probably d ifferent, l eading to distinct conformational states that did not display t he same catalytic a ctivity on tributyrin. Fig. 5. A possible scheme for explaining the existence of PICE mono- meric and pol ymeric f orms. M and P s tand for activated myristic acid and palmitic acid, respectively. NMT, N-myritoyltransferase; PAT, palmitoylacyltransferase; a nd MPT, myritoylproteinth ioesterase. Ó FEBS 2002 Carboxylesterase fatty acylation (Eur. J. Biochem. 269) 1115 Covalent changes in proteins with myristate have been observed i n s everal eukaryotic proteins [ 42,43]. The preva- lent type, myristoylation, which has been thoroughly characterized, seems to occur cotranslationally at the a-amino group of the N-terminal glycine, included in the Gly-XXX-Ser/Thr c onsensus s equence, whereas palmitic acid is thought to be added post-translationally at the sulfhydryl group of cystein via a thioester bond [43]. As far as we know, n o fatty acylation of CE h as been reported to occur so f ar. The results of the present study clearly indicate that PICE contained covalently bound fatty acids, and the fact that acylation o f t he enzyme occurred was f urther con®rmed u sing enterocyte and hepatocyte cell primary cultures in the presence of the two main corresponding radiolabelled fatty acids. F1 contained the same amount of myristic acid and palmitic acid, close to stoichiometry, while F3 contained only palmitic acid. The amount of myristic acid present in F4 was only a bout a quarter of that recorded in palmitic acid. The resistance of F1 to Edman d egradation might therefore be due to the myristoylation of the N-terminal G-Q-P-A-S- consensus sequence [7], as the monomeric form of PICE was found to contain a stoichio- metric amount of the f atty acid. As we r ecen tly established that Cys71 in t he PICE amino-acid sequence could not be alkylated with iodoacetamide, except in t he presence o f 100 m M dithiothreitol in the medium, this residue was thought to be a good candidate for palmitoylation of the PICE F1 form via a thioester linkage [13]. This a ssumption is consistent with the well-known sensitivity of thioester- type fatty acid linkages to alkaline methanolysis and the effects of reducing agents [44]. A question therefore arises about the ®nding that F1 apparently has an N-terminal myristoylated glycine, whereas F3 has a f ree amino group containing an N-terminal glycine residue. To answer the question as t o whether F1 is cotranslationally myristoylated and then deacylated before undergoing trimerization, or whether the formation of the trimer occurs competitively with N-terminal blocking of the monomer, further e xperi- ments are certainly required. As mentioned above, Fig. 5 gives a possible scheme for the formation of PICE multimers. As the F1 and F4 molecular forms of PICE are both myristoylated and palm itoylated, the functional s igni®cance of this twofold fatty acylation of the intestinal CE is still unclear. The increase in the af®nity with membranes resulting from the presence of covalently linked palmitic and myristic acids in PICE should facilitate the possible targeting, anc horing, and crossing of the cellular m em- branes, a s s uggested by the subcellular pattern of distribu- tion of the enzyme observed here. Moreover, PICE was found in the present study to deacylate PMA (data n ot shown), a structural analog of diacylglycerol, and to be variably active on a number o f ester containing xenobiotics. The high speci®city of PICE towards exogenous ester containing substrates along with the presence of the enzyme in the microsomal and cytosolic fractions suggests that it may be involved in the xenobiotic metabolism. This hypothesis needs to be con®rmed by further experimental data, as does the suggestion that the enzyme may be involved in cell signal transduction via diacylglycerol and protein kinase C [5,6]. ACKNOWLEDGEMENTS We are grateful to Dr E. H. Ajandouz for his helpful advice. We thank Mrs D. Moinier and Mr J. Bonicell for their contribution to the automatic sequencing and m ass spectrometry determinations, r espec- tively, Dr G. Pieroni for fatty acids analysis, Dr V. Girod for dissection of pigs, and Dr J. Blanc f or revising the English manuscript. REFERENCES 1. Krisch, K. (1971) Carboxylic ester hydrolases. In The Enzymes V (Boyer, P.D., ed.), pp. 43±49. Academic Press, New York and London. 2. Heymann, E. (1980) Carboxylesterases and amidases. In Enzy- matic Basis Detoxi®cation, Vol. II (J akoby, W.B., ed.), pp. 29 1± 323. Academic Press, New York. 3. Satoh, T. (1987) R ole of carboxylesterases in xenobiotic metabo- lism. Rev. Biochem. Toxico l. 8 , 155±181. 4. Mentlein, R., Schumann, M. & Heymann, E. (1984) Comparative chemical and immunological characterization of ®ve lipolytic enzymes (carboxylesterases) from rat liver microsomes. Arch. Biochem. Biophys. 234 , 612±621. 5. Mentlein, R. (1986) The tumor promoter 12-O-tetradecanoyl phorbol 13-acetate and regulatory d iacylglycerols are substrates for t he same ca rbo xylestera se. J. Biol. Chem. 26 1 , 7816 ±7818. 6. Maki, T., Hosokawa, M., Sato h, T. & Satoh, K. (1991) C hang es in carboxylesterase isoenzymes of rat liver microsomes during hepatocarcinogenesis. Jpn J. Cancer Res. 82, 8 00±806. 7. David, L., Guo, X J., Vi llard, C., Moulin, A. & Puigs erver, A. (1998) Puri®cation and molecular cloning of porcine intestinal glycerol-ester hydrolase. Evidence for its identity with c arboxy- lesterase. Eur. J. Bioc hem. 257, 142±148. 8. Matsushima, M ., Inoue, H., Ichinose, M., Tsukada, S ., Miki, K., Kurokawa, K., Takahashi, T. & Takahashi, K . (1991) The nucleotide and deduced amino acid sequences of porcine liver proline-b-naphthylamidase. Evidence for the identity with car- boxylesterase. FEB S Lett. 293, 37±41. 9. Robbi, M., Beaufay, H. & Octave, J N. ( 1990) Nucleotide sequence o f cD NA cod ing for rat liver pI 6.1 esteras e (ES -10), a carboxylesterase located in the lumen of the endoplasmic reticu- lum. Biochem. J. 269, 451±458. 10. Robbi, M. & B eaufay, H. (1994) Cloni ng and sequencing of rat liver carboxylesterase ES-3 (egasyn). Biochem. Biophys. Res. Comm. 20 3, 1404±1411. 11. Yan, B., Yang, D., Brady, M . & Parkinson, A. (1994) Rat kidne y carboxylesterase. Cloning, sequencing, cellular localization, and relationship to rat liver hydrolase. J. Biol. Chem. 269, 29688± 29696. 12. Satoh, T. & Hosokawa, M. (1995) Molecular aspects of carb- oxylesterase isoforms in comparison with other esterases. Toxicol. Lett. 82/ 83, 439±445. 13. Smialowski-Fle  ter, S., M oulin, A., Villard, C. & Puigserver, A. (2000) Structure-function relationships in the carboxylic-ester- hydrolase superfamily. Disul®de b ridge a rrangement i n porcine intestinal glycerol-ester hydrolase. Eur. J. Biochem. 267, 2227± 2234. 14. Serrero, G., Ne  grel, R. & Ailhaud, G. (1975) Characterization and partial p uri®cation of an intestin al lipase. Biochem. Biop hys. R es. Comm. 65 , 89±99. 15. Louvard,D.,Maroux,S.,Baratti,J.,Desnuelle,P.&Mutafts- chiev, S. (1973) On the p reparation and some properties of c losed membrane vesicles from hog duodenal and jejunal brush bo rder. Biochim. Biophys. Acta 29 1 , 747±763. 16. Quigley, J.P. & Gotterer, G.S. (1969) Distribution of (Na + -K + )- stimulated ATPase activity in rat intestinal mucosa. Biochim. Biophys. Acta 173, 456±468. 1116 S. Smialowski-Fle  ter et al. (Eur. J. Biochem. 269) Ó FEBS 2002 17. Sottocasa, G.L., Kuylenstierna, B., Ernster, L. & Berg strand, A. (1967) An electron -transport system associated with the outer membrane of liver mitochondria. A biochemical and morpho- logical s tudy. J. Cell Biol. 32, 415±438. 18. Cooperstein, S.J. & Lazarow, A. (1951) A microspectrophoto- metric method for the determination of cytochrome oxidase. J. Biol. Chem. 189, 665±670. 19. Murer, H., Amman, E., Biber, J. & Hopper, U. (1976) The surfac e membrane of the s mall intestinal epithelial cell. I. Localization of adenyl cyclase. Biochim. Biophys. Acta 433, 509±519. 20. Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248±254. 21. Engvall, A. & Perlmann, P. (1971) Enzyme-linked immunosorbent assay ( ELISA). Qu antitative assay of i mmunoglobulin G. Immu- nochemistry 8, 871±874. 22. Werner, S. & Machleidt, W. (1978) Isolation of precursors of cytochrome oxidase from Neurospora crassa: A pplication of sub- unit-speci®c antibodies and protein A from Staphylococcus aureus. Eur. J. Bioche m. 90, 99±105. 23. Laemmli, U. (1970) Cleavage of structural proteins during the assembly of t he h ead o f bacteriophage T4. Nature 227, 680±685. 24. Aspbury, R.A., Fischer, M.J. & Rees, H.H. (1998) Fatty acylation of polypeptides in the nemato de Caenorhabditis elegans. Bioc him. Biophys. Acta 13 82, 111± 119. 25. Robertson, E.F., Danelly, H.K., Mallpy, P.J. & Reeves, H.C. (1987) Rapid isoelectric focusing in a vertical polyacrylamide minigel system. Anal. Biochem. 16 7, 290±294. 26. Schmitz, J., P reiser, H., Maestracci, D., Ghosh, B.K., C erda, J.J. & Crane, R.K. (1973) Puri®cation of the h uman intestinal brush border membrane. Biochim. Biop hys. A cta 323, 98± 112. 27. Maury, J., Nicoletti, C., Guzzo-Chambraud, L. & Maroux, S. (1995) The ®lamentous brush border glycocalyx, a mucin-like marker of enterocyte hyper-polarization. Eur. J. Biochem. 228, 323±331. 28. Bidlingmeyer, B.A., Cohen, S.A. & Tarvin, T.L. (1984) Rapid analysis of amino acids using pre-column derivatization. J. Chromatogr. 336, 93±104. 29. Hewick, R.M. , Hunkapiller, M.W., H ood, L.E. & Dreyer, W.J. (1981) A gas-liquid s olid phase peptide and protein s equenator. J. Biol. Chem. 256, 7990±7994. 30. Bligh, E .G. & Dyer, W.J. ( 1959) A rapid method o f t otal lipid extraction and puri®cation. Can. J. Biochem. Physiol. 37, 911±917. 31. Christie, W.W. (19 82) Lipid A nalysis, 2nd edn. Perg amon Press, Oxford. 32. Bader, A., Hansen, T., Kirchner, G ., Allmeling, C., Haverich, A. & B orlak, J.T. (2000) Primary porcine enterocyte and hepatocyte cultures to study d rug oxidation r eactions. Br. J. Pharmacol. 129, 331±342. 33. Seglen, O. ( 1976) Preparation of isolated rat liver cell s. Methods Cell Biol. 13, 29±83. 34. Agenese, S .T., Spierto, F.W. & Hannon, W.H. (1983) Evaluation of four reagents for delipidation o f serum. Clin. Biochem. 16, 98± 100. 35. Heymann, E. & Junge, W. (1979) Characterization of the i soen- zymes of pig-liver esterase. 1. Chemical studies. Eur. J. Biochem. 95, 509±518. 36. Junge, W. & Heymann, E. (197 9) Characterization of the isoen- zymes of pig-liver esterase. 2. K inetic studies. Eur. J. Biochem. 95, 519±525. 37. Shibata, F., Takagi, Y., Kitajima, M., Kuroda, T. & Omura, T. (1993) Molecular Cloning and characterization of a human carboxylesterase gene. Genomics 17, 76± 82. 38. Mori, M., Hosokawa, M., Ogasawara, Y., Tsukada, E. & Chiba, K. (1999) cDNA cloning, characterization and stable expression of novel human brain carboxylesterase. FEBS Lett. 458, 17±22. 39. Bassen, R., Brichory, F., Caulet-Maugendre, S., Delaval, P. & Dazord, L. (2000) Vertebrate galectins: structure and function, role i n tumoral process. Bull. Cancer 87 , 703±707. 40. Moulin, A. & Pieroni, G. (1993) Demonstration of a lipolytic activity associated with the ricin B chain. C. R. Acad. Sci. Paris, Serie III (316), 7±12. 41. Fernandez-Lopez, V., Serrero, G., Ne  grel, R. & Ailhaud, G. (1976) Esterolytic activities of rat intestinal mucosa. 2. Puri®cation and properties of a glycerol-ester hydrolase. Eur. J. Biochem. 71, 259±270. 42. Towler, D.A. & Gordon, J.I . (1988) The biology and enzymology of eukarytic p rotein acylation. Ann. Rev. Biochem. 57, 69±99. 43. Resh, M.D. (1999) Fatty acylation of proteins: new insights into membrane targeting of myristoylated and palmitoylated proteins . Biochim. Biophys. A cta 145 1, 1 ±16. 44. Veit, M., Nu È rnberg, B., Spicher, K ., Harteneck, C., Ponimaskin, E., Schultz, G. & Schmidt, M.F.G. (1994) The a-subunits of G-protein G12 and G13 are palmitoylated, but not amidically myristoylated. FEBS Lett. 339 , 160±164. Ó FEBS 2002 Carboxylesterase fatty acylation (Eur. J. Biochem. 269) 1117 . Myristyl and palmityl acylation of pI 5. 1 carboxylesterase from porcine intestine and liver Tissue and subcellular distribution Sylvie Smialowski-Fle  ter, Andre  Moulin, Josette Perrier and. 4 51 ± 458 . 10 . Robbi, M. & B eaufay, H. (19 94) Cloni ng and sequencing of rat liver carboxylesterase ES-3 (egasyn). Biochem. Biophys. Res. Comm. 20 3, 14 04 14 11. 11 . Yan, B., Yang, D., Brady,. for 10 min at 4 °C, and the supernatant was sampled for analysis. RESULTS Tissue distribution of pI 5. 1 carboxylesterase The presence of the pI 5. 1 CE isoform i n 1 1 homogenates from various porcine