Identification of Two Tamoxifen Target Proteins by Photolabeling with

INSERM U397 Institut Claudius Regaud 20-24 rue du pont Saint Pierre 31052 Toulouse Cedex, France, and Laboratoire de biochimie des protéines, Sanofi ...
0 downloads 0 Views 117KB Size
Download PDF
766

Bioconjugate Chem. 2002, 13, 766−772

Identification of Two Tamoxifen Target Proteins by Photolabeling with 4-(2-Morpholinoethoxy)benzophenone Fabienne Me´sange,† Mohamed Sebbar,† Joe¨l Capdevielle,‡ Jean-Claude Guillemot,‡ Pascual Ferrara,‡ Francis Bayard,† Marc Poirot,† and Jean-Charles Faye*,† INSERM U397 Institut Claudius Regaud 20-24 rue du pont Saint Pierre 31052 Toulouse Cedex, France, and Laboratoire de biochimie des prote´ines, Sanofi Synthe´labo, Labe`ge Innopole, BP 137, 31676 Labe`ge Cedex, France. Received December 3, 2001; Revised Manuscript Received April 9, 2002

Our quest to identify target proteins involved in the activity of tamoxifen led to the design of photoaffinity ligand analogues of tamoxifen able to cross-link such proteins. A new tritiated photoprobe, 4-(2-morpholinoethoxy)benzophenone (MBoPE), was synthesized and used to identify proteins involved in tamoxifen binding in rat liver. MBoPE, which has structural features in common with the potential antagonist of the intracellular histamine receptor (N,N-diethyl-2-[(4-phenylmethyl)phenoxy]ethanamine HCl: DPPE) is unable to bind the estrogen receptor although it does compete with tamoxifen for an antiestrogen binding site (AEBS). This tritiated benzophenone derivative was obtained by metalcatalyzed halogen-tritium replacement reaction. Because of its high specific activity, four target proteins could be photolabeled, three of which were identified with Mr of 60 000, 49 500, and 14 000, while the fourth at 27 500 was in too low an amount and could not be sequenced. The 49.5 kDa protein corresponded by mass spectrometry to the microsomal epoxide hydrolase already identified with an aryl azide photoprobe [Mesange, F., et al. (1998) Biochem. J. 334, 107-112]. The 60 and 14 kDa proteins were identified as the carboxylesterase (ES10) and the liver fatty acid binding protein (L-FABP), respectively. The inhibitory effect of tamoxifen on carboxylesterase activity and the competitive efficacy of oleic acid on [3H]tamoxifen binding suggest that both proteins are AEBS subunits. Moreover, treatment of hepatocytes with antisense mRNA directed against ES10 or L-FABP abolished both tamoxifen and MBoPE binding. On the basis of previous pharmacological arguments, the 27.5 kDa protein might correspond to the sigma I receptor. Altogether, these results confirm that the microsomal epoxide hydrolase is a target for tamoxifen and provide evidence of two new target proteins implicated in cell lipid metabolism.

INTRODUCTION

The antiestrogen tamoxifen has few known noxious side effects and has become the world’s second most widely used antitumor drug. It was initially designed as an antagonist of binding of the steroid hormone estradiol to the estrogen receptor (ER). Given the ubiquitous presence of antiestrogen binding sites (AEBS) that specifically bind triphenylethylenic antiestrogens such as tamoxifen with high affinity (KD ) 10-9 M), several group (1-5) have focused attention on the potential existence of a protein or proteins that are different from ER and mediate several of the biological activities of tamoxifen (6, 7). By developing specific AEBS-targeted ligands based on the tamoxifen backbone but without ER affinity, we (8, 9) and others (5, 10) have hypothesized that AEBS mediates estrogen receptor-independent biological activities of tamoxifen. Moreover, the AEBS prototypical ligand DPPE (N,N-diethyl-2-[(4-phenylmethyl)phenoxy]ethanamine HCl), which was expected to be a specific ligand of a new intracellular histamine receptor (11) has been shown to potentiate the chemotherapeutic index of cytotoxic drugs in vitro (12, 13). Clinical trials have shown * To whom correspondence shoud be sent: Jean-Charles Faye, INSERM U397 Institut Claudius Regaud 20-24 rue du pont Saint Pierre 31052 Toulouse Cedex, France. Phone: 33561424646. Fax: 33561424631. E-mail: [email protected]. † INSERM U397 Institut Claudius Regaud. ‡ Sanofi Synthe ´ labo, Labe`ge Innopole.

that DPPE, in association with cyclophosphamide or doxorubicin, is successful in the treatment of metastatic prostate cancer (14) and metastatic breast cancer (15). Further, this product has been shown, by acting in the liver by an estrogen receptor-independent pathway, to modulate the serum lipids in a manner similar to tamoxifen (16) Identification and characterization of the proteins is important to understand the side effects of tamoxifen and to design new drugs. For this reason, our group has worked to identify and purify AEBS using photolabeling strategies (9). On the basis of the structure of DPPE described by Brandes (17), an azido-photoaffinity tritiated ligand was used to identify microsomal epoxide hydrolase as a component of AEBS (18). Because of the low specific activity (2 Ci/mmol) obtained with this ligand, a tritiated photoactivable ligand, 4-(2-morpholinoethoxy)benzophenone, also named morpholino-benzoylphenoxy-ethanamine (MBoPE), was prepared having high specific activity (40 Ci/mmol). This photoactivable aryl ketone (19) was chemically more stable than the aryl azide previously used, was activated at 360 nm (with less protein damage), and inserted into unreactive C-H bonds in the presence of water. This tamoxifen-competing ligand could be employed after electrofocusing prepurification. By these techniques, specific photolabeling (competed by either tamoxifen or unlabeled MBoPE) of microsomal epoxide hydrolase was obtained, as previously reported (18), as well as the fatty acid binding protein L-FABP and the carboxylesterase ES10.

10.1021/bc015588t CCC: $22.00 © 2002 American Chemical Society Published on Web 06/19/2002

Protein Targets of Tamoxifen

Bioconjugate Chem., Vol. 13, No. 4, 2002 767 1H

Figure 1. Outline of synthesis of [3H] MBoBPE. MATERIALS AND METHODS

Chemicals. [3H]-Tamoxifen (specific activity 84 Ci/ mmol) was from Amersham, and unlabeled steroids were from Steraloı¨d. Tamoxifen was a gift from Imperial Chemical Industries. Tris, KCl, sodium azide, glycerol, EDTA, dimethylformamide, and other reagents were obtained from Sigma. Synthesis of Photosensitive Ligand MBoPE (Figure 1). Synthesis of 2,4,6-Trichloro- 4′-methoxybenzophenone. AlCl3 (4 g, 30 mmol) was progressively added to a solution of anisole (2.2 g, 20 mmol) and 2,4,6-trichlorobenzoyl chloride (5 g, 20 mmol) in 50 mL of anhydrous dichloromethane at -80 °C in an inert atmosphere, and the reaction was monitored by TLC. The reaction mixture was gradually brought to ambient temperature and poured into a mixture of crushed ice and 0.1 N HCl (150 mL). The organic layer was collected, washed with water, dried (Na2SO4), and concentrated to yield 5.5 g of a white solid (87% yield, mp ) 110-111 °C). Mass spectrum (70 eV) m/e (relative intensity) 315.6. 1H NMR (100 MHz, DMSO-d6), δ; 3.83 (s, 3H, OCH3); 6.87 (d, 2H, arom); 7.26 (d, 2H, arom), 7.59 (d, 2H, arom). Anal. (C14H9Cl3O2) C, H, N, Cl. Synthesis of 4-Hydroxyphenyl 2,4,6-Trichlorophenyl Ketone. One equivalent of 4-methoxyphenyl 2,4,6-trichlorophenyl ketone was heated to melting point at 250 °C with 10 equiv of pyridinium chloride. After 4 h, 30 mL of water were added followed by 30 mL of ether. The ether phase was washed with a 0.1 N solution of HCl, dried over Na2SO4, and evaporated. The phenolic product was recrystallized in 2-propanol (yield 70%). Mass spectrum (70 eV) m/e (relative intensity) 302.6. 1H NMR (100 MHz, DMSO-d6) δ; 6.82-7.59 (m, 6H, arom); 12.4 (s, 1H, OH phenolic, exchangeable with 2H2O). Anal. (C14H9ClO2) C, H, N, Cl. Synthesis of 4-(2-Morpholinoethoxy)phenyl 2,4,6Trichlorophenyl Ketone. 4-Hydroxyphenyl 2,4,6-trichlorophenyl ketone was added (0.5 g, 1.6 mmol) to a solution of NaOH (200 mg, 4.8 mmol) in 15 mL of water. The mixture was boiled for 1 h. A total of 1.1 equiv of 2-morpholinoethyl chloride was added gradually to the mixture. After the sample was refluxed for 12 h, the reaction mixture was cooled and extracted with 30 mL of diethyl ether. The organic layer was washed with 3 × 10 mL of 0.1 N NaOH, and then with 3 × 10 mL of brine. The organic layer was extracted with 1.5 mL of 6 N HCl. The aqueous layer was evaporated, dried over P2O5, and recrystallized from ethanol (75% yield, mp ) 110-111 °C). Mass spectrum (70 eV) m/e (relative intensity) 415.9.

NMR (100 MHz, DMSO-d6) δ; 3.47 (m, 4H, -N(CH2CH2)2O); 4.01 (m, 4H, -N(CH2CH2)2O); 3.65-3.69 (t, 2H, J ) 6.8 Hz) Ar-O-CH2-CH2-N; 4.40-4.53 (t, 2H, J ) 6.05 Hz, Ar-O-CH2-CH2-N-); 6.93-6.96 (m, 2H, arom); 6.99-7.05 (m, 2H, arom), 7.23-7.41 (m, 2H, arom). Anal. (C19H19Cl4NO3) C, H, N, Cl. Synthesis of 4-(2-Morpholinoethoxy)phenyl 2,4,6-3H Phenyl Ketone. The tritiation was performed at the Commissariat a` l’Energie Atomique (Saclay, France) by catalytic tritium exchange with 3H2 over Pd/C (5%) in ethanol. The chemical and radiochemical purity of 4-(2morpholinoethoxy) phenyl (2,4,6-3H phenyl) ketone ([3H]MBoPE) was determined by HPLC under the following conditions: isocratic elution on a C18 RP column (254 × 4.6) licrosorb (Bishof, Germany) coupled to UV detection at 220 nm using 60% acetonitrile in ammonium acetate, pH 8. The specific activity of this compound, determined by quantitative HPLC, was 40 Ci/mmol. The mass of an injected sample was determined by comparing the integrated area of the eluted peak with an area calibration curve generated by injecting known weights of unlabeled MBoPE, and the activity of the eluting peak was determined by scintillation counting. Preparation and Solubilization of Rat Liver Microsomes. Sprague-Dawley rats were ovarectomized by dorsal route at least three weeks before sacrifice by cervical dislocation. Livers were removed and the microsomes prepared and solubilized as previously described (20). Preparation and Solubilization of Microsomes from Cultured Cells. All procedures were performed at 4 °C; 5 × 106 cells were pelleted by centrifugation at 1500 rpm for 15 min and washed with PBS. Homogenization was carried out by sonication in TTE buffer pH 7.4 (Tris-HCl 50 mM, EDTA 1.5 mM, thioglycerol 12 mM). The homogenate was centrifuged at 10000g for 20 min, and AEBS measurements obtained from the supernatant as previously described (18). The protein concentration was determined by the Bradford method (21). Binding Studies. Binding studies were routinely performed with 200 µL aliquots of solubilized material. For the KI determination, [3H]tamoxifen or [3H]MBoPE were incubated in the presence or absence of various concentrations of unlabeled compounds and in the presence of 1 µM estradiol to block ligand binding to estrogen receptors. After incubation (16 h at 4 °C), the bound and free radioligands were separated on Sephadex LH-20 columns (20). No-specific binding was determined in the presence of 1 µM tamoxifen, and its value was corrected according to the method of Rosenthal (22). The equilibrium dissociation constant (KD) was determined with increasing tritiated ligand and calculated with the EBDA Ligand computer program (23). Photoinactivation of the Tamoxifen Binding Capacity. After incubation of the soluble preparation for 18 h at 4 °C with 150 nM MboPE into 800 µL aliquots, the samples were UV-irradiated for various periods of time (8 W) at 365 nm. The residual [3H]tamoxifen binding capacity was measured as described above, after elimination of the noncovalent bound ligand by passage through a Sephadex G25 column (Pharmacia). A control experiment was carried out in the absence of MBoPE to determine the UV induced protein degradation. Electrofocusing. Solubilized microsomal proteins were subjected to liquid electrofocusing in a Rotofor Cell (Biorad). The 20 mM CHAPS supernatant of the microsomal extraction was mixed with 3 mL of pH 3-10 ampholytes and 2 mL of pH 5-8 ampholytes. Electrofocusing was performed for 6 h at 4 °C with power 12 W,

768 Bioconjugate Chem., Vol. 13, No. 4, 2002

and the solution of each of 20 compartments was collected. After the pH was measured and adjusted to 7.4 with 100 mM Tris, each fraction was checked for [3H]tamoxifen and [3H]MBoPE binding. Photolabeling. The electrofocused fraction pH 6.7 was incubated for 18 h at 4 °C with [3H]MBoPE (150 nM) in the presence or absence of tamoxifen (1 µM). UV irradiation was carried out for 45 min and the photolinked proteins precipitated with 10% TCA. After centrifugation at 12000g, the pellet was washed twice with ethanol/ ether (v/v: 1/2), dried, and dissolved in electrophoresis sample buffer (50 mM Tris, pH 6.8, 100 mM dithiotreitol, 8% glycerol, 2% SDS). The radioactivity was determined before loading onto a polyacrylamide gel. SDS-PAGE and Autoradiography. Samples were subjected to electrophoresis as described by Laemmli (24). Following protein migration, the gel was either rapidly stained (25) and cut in 5-mm slices, which were then crushed and proteolyzed with proteinase k (the radioactivity contained in 10000g supernatant was determined), or dried in the presence of Amplify (Amersham) and autoradiographed. Proteolyses, Peptide Purifications, and Sequences. Ingel trypsinisation of the radioactive bands revealed by the autoradiogram was performed as previously described by Rosenfeld et al. (25). The peptides eluted from the polyacrylamide matrix were purified on a Brownlee Labs microgradient system on a C18 Ultrasphere column (250 × 2.1 mm). Elution was performed with a linear gradient of 5 to 70% acetonitrile, 0.1% TFA at a flow rate of 300 µL/min, and the peaks detected at 218 nm were manually collected. Sequence analysis was performed on a gas-phase sequencer (470A, Applied Biosystems). ORF Subcloning of Human Liver Fatty Acid Binding Protein. Total RNA from the hepato-carcinoma cell line SKHep-1 was obtained by a rapid guanidinium thiocyanate procedure (26). The cDNA encoding for the human L-FABP was acquired by RT-PCR of total RNA using the superscript preamplification system (GIBCO-BRL) with random hexamers for the reverse transcription step. Oligonucleotides with EcoRI and BamHI restriction sites, matching the first 15 and last 16 bases of the ORF cDNA encoding the human L-FABP, respectively (27), were used for PCR. The amplification product was cloned in the EcoRI, BamHI restriction sites of pSG5 vector (28) and sequenced by the dideoxy chain-termination technique (29). Rat Liver Carboxylesterase. (ES10) cDNA (30) was a generous gift from Dr. M. Robbi (University of Brussels, Belgium). It should be subcloned in the pSG5 vector at the EcoRI restriction site. Overproduction of Human L-FABP and ES10. Plasmids (1 µg) were transiently transfected in COS-7 cells using the DEAE-Dextran methodology, as previously described (18). Cells were harvested by scraping 48 h after transfection and sonicated in TTE buffer. Enzyme ES10 Assays. Carboxylesterase activity was measured colorimetrically at 30 °C, using p-nitrophenylacetate as substrate, according to the procedure previously reported by Hosokawa (31). Assays were carried out on the solubilized microsomal subfraction of transiently transfected COS-7 cells. Increasing concentrations, up to 10-4 M, of tamoxifen were added and the concentration that inhibited 50% of the enzyme activity (IC50) was determined. Inhibition of Specific Protein Syntheses in SKHep-1 Cells. Inhibition of the synthesis of specific proteins was obtained by using an oligonucleotide antisense strategy

Me´sange et al.

Figure 2. [3H]MBoPE saturation binding curve on solubilized microsomes. Increasing concentrations of [3H]MBoPE were incubated with 100 µL of solubilized microsomes as described in Materials and Methods (9) total binding ()) unspecific binding in the presence of a 100-fold excess of either MBoPE or Tx (0) specific binding. The graph shown is representative of five independent experiments. The corresponding Scatchard representation of specific binding is shown in the inset.

with 18-mer oligonucleotides matching with the first 5′ translated codons of a protein encoding mRNA (32, 33). The following oligodeoxynucleotides (ODN) were made: 3′-TACTCAAAGAGGCCGTT-5′ and 3′-TGCTACAC CGAGGCAG-5′ for human L-FABP and human carboxylesterase, respectively. Controls were carried out with randomized bases contained in the antisense ODN and with the 5′/3′ base inverted sequence of the antisense ODN. The transfections of SKHep-1 with oligonucleotides (final concentration ) 20 µM) were carried out using the cationic liposome lipofectin system according to the manufacturer’s recommendations (Gibco BRL). RESULTS

Binding Studies and Photoinactivation. As previously reported (19), MBoPE is able to compete with [3H]tamoxifen on AEBS of solubilized microsomes. The inhibitory efficacy of MBoPE on tamoxifen binding (KI ) 1.3 10-7 M) was determined by Dixon plots and was, in good agreement with the affinity of [3H]MBoPE (KD ) 1.1 10-7 M), calculated from the results in Figure 2, with the EBDA ligand computer program (23). The ability of MBoPE to inactivate photochemically tamoxifen binding was investigated on solubilized microsomes of rat liver as previously described (9). After as little as 5 min, MBoPE produces a covalent irreversible linkage to AEBS that inhibits [3H]tamoxifen binding (Figure 3). In the absence of MBoPE, 100% of the tamoxifen binding is conserved for up to 30 min, showing that UV irradiation at 365 nm does not destroy the proteins but activates the photoprobe since only 50% of the sites remained accessible in its presence. The progressive decrease in residual tamoxifen binding indicates that the photolinkage of MBoPE is time dependent, thus enhancing the usefulness of this photoprobe for tamoxifen target labeling. Purification of the Photolabeled Proteins. Because of its high specific activity, MBoPE has an advantage over the previous azido photoprobe (9) in that it can be used in the presence of free ligands to provide a nonspecific estimation and in binding studies after solubilization and

Protein Targets of Tamoxifen

Bioconjugate Chem., Vol. 13, No. 4, 2002 769

Figure 3. Photoinactivation of solubilized AEBS. Unlabeled MBoPE (500 nM) were incubated with solubilized AEBS extracts and subjected (grey bars) or not (black bars) to UV irradiation for 45 min. The residual [3H]tamoxifen binding was determined after removal of free ligand on a PD10 column.

Figure 5. Photolabeling with [3H]MBoPE after electrofocusing. The pH 6.7 fraction was incubated for 18 h at 4 °C with [3H]MBoPE (100 nM) in the absence (lane 1) or presence of unlabeled (lane 2) 10 µM MBoPE, (lane 3) 1 µM Tx. After UV irradiation at 360 nm, samples were TCA precipitated, solubilized with sample buffer, and subjected to SDS-PAGE on 10% gel. Bottom: specific radioactivity (lane 1 minus lane 3) determined by counting. Upper: the gel was fixed, incubated with Amplify, dried, and exposed for 23 days to the obtained autoradiogram.

Figure 4. Specific binding studies after iso-electrofocusing. Isoelectrofocusing was performed with a rotofore on solubilized microsomal protein as described in Materials and Methods. Binding was carried out either with (b) [3H]Tx (3 nM) or (() [3H]MBoPE (100 nM) in the presence or absence of 1 µM unlabeled Tx. Inset: Scatchard analysis of the fraction 12.

electrofocusing. Similar binding patterns were found for tritiated tamoxifen and MBoPE (Figure 4), and the affinity parameters for tamoxifen and MBoPE (Figure 4; inset) were determined on fraction numbers 3 (pH ) 3.2) and 12 (pH ) 6.7). We were not able to obtain a saturation curve of binding with fraction 3, although KD values similar to those observed in the crude extract were found for [3H]Tx (KD ) 3 × 10-9 M) and [3H]MBoPE (KD ) 1 × 10-7 M) with fraction 12 (Figure 4). This result is in good agreement with previous observations (9), which gave a pI of about 6.4 for AEBS when photolabeling, with another azido ligand, was carried out before chromatofocusing. The fraction with a pI ) 6.7 was therefore used

for photolabeling and further purification steps. Acidprecipitable proteins were subjected to SDS-PAGE, and photolabeling was evaluated either by autoradiography or counting (Figure 5). Radioactive peaks were extracted and loaded onto polyacrylamide gels of 12 or 18% for bands corresponding to Mr of 60 000 and 49 500, or 14 000, respectively. The proteins were lightly fixed, stained, and directly trypsinized in the gel as described in Materials and Methods and in ref 27. The radioactive peak at Mr 49 500 corresponds, according to mass spectroscopy analyses, to the microsomal epoxide hydrolase previously reported (18). Eluted peptides, corresponding to the protein with apparent Mr of 60 000, were simultaneously analyzed by mass spectroscopy and purified by C18 reverse phase column HPLC. Two distinct peptides were sequenced according to the Edman method. The detected amino acid sequences were NLFHR, DAGAPTFMYEFEYR, corresponding to amino acids 220224 and 418-431 of the rat carboxylesterase (ES-10) identified by Robbi et al. (30). The theoretical molecular masses of these amino acid sequences correspond to two peaks observed in mass spectroscopy. These amino acid series exhibited 100% homology with sequences from the rat carboxylesterase ES 10. The same experimental procedures were performed for the Mr 14 000 protein and two peptides were sequenced, namely, the MVTTFK and AMGLPEDLIQK amino acid sequences which exhibit 100% homology with sequences 91-96 and 21-31 of the rat liver fatty acid protein (L-

770 Bioconjugate Chem., Vol. 13, No. 4, 2002

Figure 6. FABP and ES10 overexpressions: [3H]Tx binding analyses of COS-7 transfected cells. FABP and ES10 corresponding cDNA were transiently transfected and microsomal membranes incubated at 4 °C with increasing concentrations of [3H]tamoxifen (0.05-10 nM). (1) control cells (O) FABP, (b) ES10. The equilibrium dissociation constant (KD) was calculated with the EBDA Ligand computer program (24). Western blot analyses of solubilized microsomes (20 µg of protein loaded) with antibodies directed against ES10 (lanes 1 and 2) or FABP (lanes 3 and 4). Lane 1: transfected cells with ES10 cDNA (5 µg). Lanes 2 and 3 mock transfected cells. Lane 4: transfected cells with liver FABP cDNA (5 µg).

FABP), the full-length cDNA of which was isolated by Gordon et al. (34). The radioactive peak seen at 27 500 (Figure 5) could not be further purified. We have previously shown that AEBS ligands exhibit structural similarities with sigma receptor ligands and bind with high affinity to the sigma I receptor (35) for which the reported molecular mass in the liver is 28 kDa (36). Protein Overexpression. Binding Studies. To ensure that carboxylesterase and FABP belonged to the AEBS complex, their cDNA open reading frames were subcloned in the pSG5 vector and transiently transfected in COS-7 cells. The specific increase of AEBS was determined by [3H]Tx Scatchard analyses and by binding studies of [3H]MBoPE at the single concentration of 150 × 10-9 M. As seen in Figure 6, although the protein overexpressions determined by Western blots were increased at least 10fold, the endogenous proteins were not detected by antibodies under our conditions, and the increase in Tx binding to solubilized microsomes of transfected cells with carboxylesterase overexpression was 1.85-fold, and 1.4-fold in the case of FABP. As shown by Scatchard analysis, the overexpresion of each protein does not change the KD of tamoxifen binding. Similar increases in binding site numbers were observed with tritiated MBoPE: 2.5-5.9 pmol/mg of protein for carboxylesterase and 2.5-5 pmol/mg of protein for FABP. Competition for [3H]Tx Binding by the p-Nitrophenylacetate and Oleic Acid Subtrates of Carboxylesterase and FABP, Respectively. Increasing concentrations of [3H]Tx in the presence or absence of p-nitrophenylacetate or oleic acid were used to determined the competitive activities of these products. The solubilized microsomal fractions of the respective transfected cells were used as biological supports (rat liver and SKHep1 cells were also used

Me´sange et al.

Figure 7. [3H]Tx binding displacement with oleic acid. Competitions were performed on liver microsomal membrane proteins incubated with increasing concentrations of [3H]Tx in the (b) absence and in the (0) presence of oleic acid (30 µM).

without any modification of the results). Although the carboxylesterase substrate (p-nitrophenylacetate) did not compete with tamoxifen up to 10-3 M, the oleic acid at 3 × 10-8 M was able to significantly displace tamoxifen. As seen in Figure 7 the Bmax decreased, whereas the KD was unchanged by the competitor evidencing a inhibitory effect of the oleic acid on the tamoxifen binding. However, p-nitrophenylacetate, at concentrations exceeding 10-3 M, induced various problems of solubility (formation of aggregates in the test tube), binding saturability and reproducibility. To ensure that carboxylesterase was a tamoxifen target, its catalytic activity was therefore estimated in the presence of tamoxifen or MBoPE. Tamoxifen and MBoPE were apparently able to inhibit the catalysis of p-nitrophenylacetate by carboxylesterase with IC50 ) 20 × 10-6 and 60 × 10-6 M, respectively. Inhibition of Protein Expression by Antisense Oligonucleotides. Following our previous studies with the human hepatoma cell line SKHep1 for which the halflife (26 ( 2 h) and concentration of AEBS had previously been estimated (18), an antisense oligonucleotides strategy, directed against the corresponding human liver mRNA of the characterized proteins was chosen. Figure 8 shows the Scatchard plots of [3H]Tx binding following various cell treatments. Controls were carried out with the corresponding scrambled and 5′/3′ inverted sequence oligonucleotides at the same concentration (20 µM). Antisense ODNs against L-FABP and ES10 were able to inhibit [3H]Tx binding to a similar extent (50%) without any effect on the KD, whereas the scrambled and inverted oligonucleotides had no effect, thus demonstrating the specificity of the inhibition. Binding of [3H]MBoPE at a single concentration (150 × 10-9 M) gave similar inhibition results (data not shown).

Protein Targets of Tamoxifen

Figure 8. Inhibition of [3H]Tx binding with oligonucleotides. Scatchard analyses of [3H]Tx binding in SKHep-1 cells were carried out as in Figure 6. (1) Control cells, (O) FABP antisense, (*) FABP scrambled ODN, (0) FABP inverted ODN, (b) ES10 antisense, (9) ES10 scrambled ODN, and (3) ES10 inverted ODN. DISCUSSION

Identifying the intracellular target proteins of tamoxifen constitutes an important step in understanding the mechanism of action of this antiestrogen. We have for several years been using a specific photolabeling strategy based on the highest affinity site for tamoxifen binding (AEBS). Microsomal epoxide hydrolase was the first protein shown to be a member of the AEBS (18). However, the low specific activity of the previous photoprobe has prevented the determination of other potential protein targets of tamoxifen in the protein complex with molecular mass between 100 and 700 kDa (6, 37, 38). Using products of the intestinal metabolism of triarylethylene antiestrogen (39), we synthesized a new radioactive photoprobe, named MBoPE, with a specific activity 20-fold higher than the former. This photoprobe has, in addition to epoxide hydrolase, led to the determination of two new proteins involved in tamoxifen binding. Although the overexpression of these proteins obtained by transient transfection of their ORF does not substantially increase the maximal binding (Bmax) of tritiated tamoxifen, their inhibition with the antisense strategy suggests that the liver fatty acid binding protein (LFABP) and liver carboxylesterase ES10 also belong to the heteropolypeptide AEBS. Moreover, while in both experiments only Bmax was modified and the affinity (KD) did not change, a tamoxifen interaction with the same binding site could be observed in each case. L-FAPB is widely expressed in liver, and photolabeling of this protein might be nonspecific. However, L-FABP is a cytosolic protein, and its concentration in the microsomal subcellular fraction is low. The involvement of FABP in AEBS has already been suggested by Hwang (40). We show here that the competition of [3H]Tx with oleic acid in binding studies is noncompetitive, which would suggest different binding domains for tamoxifen and fatty acid. Altogether, the results obtained here show that L-FABP is a protein target of tamoxifen. Although its precise function remains unknown, it is assumed to be involved in the uptake, intracellular transport, compartmentalization, and metabolism of fatty acids. It is also involved in peroxisome proliferation and has been described as a specific mediator of the mitogenesis induced by two classes of carcinogenic peroxisome proliferators (41).

Bioconjugate Chem., Vol. 13, No. 4, 2002 771

The microsomal carboxylesterase ES 10 is the second new protein we have shown to be a tamoxifen target. Its catalytic inhibition by tamoxifen and MBoPE accord with its belonging to AEBS. This protein is included in a family of at least four enzyme isoforms. All are glycoproteins of the high-mannose type, which are able to cleave simple aliphatic and aromatic esters (42). It has been suggested hat the enzyme might correspond to the activity of acyl coenzyme A cholesterol acyltransferase (ACAT) (43). In agreement with this hypothesis, we recently observed that tamoxifen was an inhibitor of ACAT in vitro and in vivo (C. Danthier, unpublished data), and the enzymatic activity of ES10 might require FABP as a cofactor. Taken together, these data indicate that FABP and ES 10 are possible components of the antiestrogen binding site and suggest that one mode of action of tamoxifen would be to interfere with lipid metabolism in an estrogen receptor-independent mechanism. Moreover, it has recently been reported in yeast that sigma I receptor corresponds to the C7/C8 sterol isomerase implied in tamoxifen binding and activity (36, 44), suggesting that this enzyme would be a part of AEBS. All these results suggest that tamoxifen interacts with a large heteroprotein complex, the molecular mass of which is probably about 700 kDa. Most of the proteins in this complex interfere with cholesterol biosynthesis, since we have shown that they are inhibited by tamoxifen. These results could explain the cholesterol downregulation reported in various experimental and therapeutic treatments (45). ACKNOWLEDGMENT

This work was supported in part by grants from INSERM, ARC, Ligue nationale contre le cancer and Institut Claudius Regaud. (ES10) cDNA and antibodies are from Dr. M. Robbi (University of Brussels, Belgium). LITERATURE CITED (1) Faye, J. C., Lasserre, B., and Bayard, F. (1980) Antiestrogen specific, high affinity saturable binding sites in rat uterine cytosol. Biochem. Biophys. Res. Commun. 93, 1225-31. (2) O’Brian, C. A., Liskamp, R. M., Solomon, D. H., and Weinstein, I. B. (1985) Inhibition of protein kinase C by tamoxifen. Cancer Res. 45, 2462-5. (3) Lam, H. Y. (1984) Tamoxifen is a calmodulin antagonist in the activation of cAMP phosphodiesterase. Biochem. Biophys. Res. Commun. 118, 27-32. (4) Sutherland, R. L., Murphy, L. C., San Foo, M., Green, M. D., Whybourne, A. M., and Krozowski, Z. S. (1980) Highaffinity anti-oestrogen binding site distinct from the oestrogen receptor. Nature 288, 273-5. (5) Brandes, L. J., and Hermonat, M. W. (1984) A diphenylmethane derivative specific for the antiestrogen binding site found in rat liver microsomes. Biochem. Biophys. Res. Commun. 123, 724-8. (6) Faye, J. C., Jozan, S., Redeuilh, G., Baulieu, E. E., and Bayard, F. (1983) Physicochemical and genetic evidence for specific antiestrogen binding sites. Proc. Natl. Acad. Sci. U.S.A. 80, 3158-62. (7) Sudo, K., Monsma, F. J., Jr., and Katzenellenbogen, B. S. (1983) Antiestrogen-binding sites distinct from the estrogen receptor: subcellular localization, ligand specificity, and distribution in tissues of the rat. Endocrinology 112, 425-34. (8) Faye, J. C., Fargin, A., Valette, A., and Bayard, F. (1987) Antiestrogens, different sites of action than the estrogen receptor? Horm. Res. 28, 202-11. (9) Poirot, M., Chailleux, C., Fargin, A., Bayard, F., and Faye, J. C. (1990) A potent and selective photoaffinity probe for the anti-estrogen binding site of rat liver. J. Biol. Chem. 265, 17039-43. (10) Taylor, F. R., Saucier, S. E., Shown, E. P., Parish, E. J., and Kandutsch, A. A. (1984) Correlation between oxysterol

772 Bioconjugate Chem., Vol. 13, No. 4, 2002 binding to a cytosolic binding protein and potency in the repression of hydroxymethylglutaryl coenzyme A reductase. J. Biol. Chem. 259, 12382-7. (11) Brandes, L. J., Macdonald, L. M., and Bogdanovic, R. P. (1985) Evidence that the antiestrogen binding site is a histamine or histamine-like receptor. Biochem. Biophys. Res. Commun. 126, 905-10. (12) Brandes, L. J., LaBella, F. S., and Warrington, R. C. (1991) Increased therapeutic index of antineoplastic drugs in combination with intracellular histamine antagonists. J. Natl. Cancer Inst. 83, 1329-36. (13) Jones, J. A., Albright, K. D., Christen, R. D., Howell, S. B., and McClay, E. F. (1997) Synergy between tamoxifen and cisplatin in human melanoma cells is dependent on the presence of antiestrogen-binding sites. Cancer Res. 57, 265760. (14) Brandes, L. J., Bracken, S. P., and Ramsey, E. W. (1995) N,N-Diethyl-2-[4-(phenylmethyl)phenoxy]ethanamine in combination with cyclophosphamide: an active, low-toxicity regimen for metastatic hormonally unresponsive prostate cancer. J. Clin. Oncol. 13, 1398-403. (15) Brandes, L. J., and Bracken, S. P. (1998) The intracellular histamine antagonist, N,N-diethyl-2-[4-(phenylmethyl)phenoxy] ethamine HCL, may potentiate doxorubicin in the treatment of metastatic breast cancer: Results of a pilot study. Breast Cancer Res. Treat. 49, 61-8. (16) Lazier, C. B., and Breckenridge, W. C. (1990) Comparison of the effects of tamoxifen and of a tamoxifen analogue that does not bind the estrogen receptor on serum lipid profiles in the cockerel. Biochem. Cell Biol. 68, 210-7. (17) Brandes, L. J. (1984) A diphenylmethane derivative selective for the anti-estrogen binding site may help define its biological role. Biochem. Biophys. Res. Commun. 124, 2449. (18) Me´sange, F., Sebbar, M., Kedjouar, B., Capdevielle, J., Guillemot, J. C., Ferrara, P., Bayard, F., Delarue, F., Faye, J. C., and Poirot, M. (1998) Microsomal epoxide hydrolase of rat liver is a subunit of the anti-oestrogen-binding site. Biochem J. 334, 107-12. (19) Delarue, F., Kedjouar, B., Mesange, F., Bayard, F., Faye, J. C., and Poirot, M. (1999) Modifications of benzylphenoxy ethanamine antiestrogen molecules: influence affinity for antiestrogen binding site (AEBS) and cell cytotoxicity. Biochem. Pharmacol. 57, 657-61. (20) Chailleux, C., Poirot, M., Mesange, F., Bayard, F., and Faye, J. C. (1994) Characterization of the membranous antiestrogen binding protein: I. Partial purification of the protein in its active state. J. Recept. Res. 14, 23-35. (21) 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, 24854. (22) Rosenthal, R. E. (1967) A graphic method for the determination and presentation of binding parameters in a complexe system. Anal. Biochem. 20, 522-532. (23) McPherson, G. A. (1985) Analysis of radioligand binding experiments. A collection of computer programs for the IBM PC. J. Pharmacol. Methods 14, 213-28. (24) Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-5. (25) Rosenfeld, J., Capdevielle, J., Guillemot, J. C., and Ferrara, P. (1992) In-gel digestion of proteins for internal sequence analysis after one- or two-dimensional gel electrophoresis. Anal. Biochem 203, 173-9. (26) Chomczynski, P., and Sacchi, N. (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate-phenolchloroform extraction. Anal. Biochem 162, 156-9. (27) Lowe, J. B., Boguski, M. S., Sweetser, D. A., Elshourbagy, N. A., Taylor, J. M., and Gordon, J. I. (1985) Human liver fatty acid binding protein. Isolation of a full length cDNA and comparative sequence analyses of orthologous and paralogous proteins. J. Biol. Chem. 260, 3413-7. (28) Green, S., Issemann, I., and Sheer, E. (1988) A versatile in vivo and in vitro eukaryotic expression vector for protein engineering. Nucleic Acids Res. 16, 369.

Me´sange et al. (29) Sanger, F., Nicklen, S., and Coulson, A. R. (1977) DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. U.S.A. 74, 5463-7. (30) Robbi, M., Beaufay, H., and Octave, J. N. (1990) Nucleotide sequence of cDNA coding for rat liver pI 6.1 esterase (ES10), a carboxylesterase located in the lumen of the endoplasmic reticulum. Biochem. J. 269, 451-8. (31) Hosokawa, M., Hirata, K., Nakata, F., Suga, T., and Satoh, T. (1994) Species differences in the induction of hepatic microsomal carboxylesterases caused by dietary exposure to di(2-ethylhexyl)phthalate, a peroxisome proliferator. Drug Metab. Dispos. 22, 889-94. (32) Thierry, A. R., and Dritschilo, A. (1992) Intracellular availability of unmodified, phosphorothioated and liposomally encapsulated oligodeoxynucleotides for antisense activity. Nucleic Acids Res. 20, 5691-8. (33) Erickson, R. P. (1993) The use of antisense approaches to study development. Dev. Genet. 14, 251-7. (34) Gordon, D. A., Shelness, G. S., Nicosia, M., and Williams, D. L. (1988) Estrogen-induced destabilization of yolk precursor protein mRNAs in avian liver. J. Biol. Chem. 263, 262531. (35) Kedjouar, B., Daunes, S., Vilner, B. J., Bowen, W. D., Klaebe, A., Faye, J. C., and Poirot, M. (1999) Structural similitudes between cytotoxic antiestrogen-binding site (AEBS) ligands and cytotoxic sigma receptor ligands. Evidence for a relationship between cytotoxicity and affinity for AEBS or sigma-2 receptor but not for sigma-1 receptor. Biochem. Pharmacol. 58, 1927-39. (36) Hanner, M., Moebius, F. F., Flandorfer, A., Knaus, H. G., Striessnig, J., Kempner, E., and Glossmann, H. (1996) Purification, molecular cloning, and expression of the mammalian sigma 1 binding site. Proc. Natl. Acad. Sci. U.S.A. 93, 8072-7. (37) Fargin, A., Faye, J. C., le Maire, M., Bayard, F., Potier, M., and Beauregard, G. (1988) Solubilization of a tamoxifenbinding protein. Assessment of its molecular mass. Biochem. J. 256, 229-36. (38) Matin, A., Hwang, P. L., and Kon, O. L. (1987) Murine antiestrogen-binding protein: characterization, solubilization and modulation by lipids. Biochim. Biophys. Acta 931, 36475. (39) Ruenitz, P. C., and Bagley, J. R. (1986) Metabolism of nitromiphene (CI 628) in the immature female rat: role of gastrointestinal microflora in the biotransformation of a triarylethylene antiestrogen. Cancer Res. 46, 6255-9. (40) Hwang, P. L. (1990) High-affinity binding sites for oxygenated sterols in rat liver microsomes: possible identity with antiestrogen binding sites. Biochim. Biophys. Acta 1033, 15461. (41) Khan, S. H., and Sorof, S. (1994) Liver fatty acid-binding protein: specific mediator of the mitogenesis induced by two classes of carcinogenic peroxisome proliferators. Proc. Natl. Acad. Sci. U.S.A. 91, 848-52. (42) Mentlein, R., Suttorp, M., and Heymann, E. (1984) Specificity of purified monoacylglycerol lipase, palmitoyl-CoA hydrolase, palmitoyl-carnitine hydrolase, and nonspecific carboxylesterase from rat liver microsomes. Arch. Biochem. Biophys. 228, 230-46. (43) Becker, A., Bottcher, A., Lackner, K. J., Fehringer, P., Notka, F., Aslanidis, C., and Schmitz, G. (1994) Purification, cloning, and expression of a human enzyme with acyl coenzyme A: cholesterol acyltransferase activity, which is identical to liver carboxylesterase. Arterioscler. Thromb. 14, 1346-55. (44) Labit-Le Bouteiller, C., Jamme, M. F., David, M., Silve, S., Lanau, C., Dhers, C., Picard, C., Rahier, A., Taton, M., Loison, G., Caput, D., Ferrara, P., and Lupker, J. (1998) Antiproliferative effects of SR31747A in animal cell lines are mediated by inhibition of cholesterol biosynthesis at the sterol isomerase step. Eur. J. Biochem. 256, 342-9. (45) Jordan, V. C., and Morrow, M. (1999) Tamoxifen, raloxifene, and the prevention of breast cancer. Endocr. Rev. 20, 253-78.

BC015588T