Thiol-modifying inhibitors for understanding squalene cyclase function Paola Milla 1 , Alexander Lenhart 2 , Giorgio Grosa 3 , Franca Viola 1 , Wilhelm A. Weihofen 2 , Georg E. Schulz 2 and Gianni Balliano 1 1 Universita ` degli Studi di Torino, Dipartimento di Scienza e Tecnologia del Farmaco, Torino, Italy; 2 Universita ¨ t Freiburg, Institut fu ¨ r Organische Chemie und Biochemie, Freiburg, Germany; 3 Universita ` degli Studi del Piemonte Orientale ÔA. Avogadro, Dipartimento di Scienze Chimiche, Alimentari, Farmaceutiche e Farmacologiche, Novara, Italy The function of squalene-hopene cyclase from Alicycloba- cillus acidocaldarius was studied by labelling critical cysteine residues of the enzyme, either native or inserted by site- directed mutagenesis, with different thiol-reacting molecules. The access of the substrate to the active centre cavity through a nonpolar channel that contains a narrow constriction harbouring a cysteine residue (C435) was probed by labelling experiments on both a C435S mutant, lacking C435 of the channel constriction, and a C25S/C50S/C455S/C537S mu- tant, bearing C435 as the only cysteine residue. Labelling experiments with tritiated 3-carboxy-4-nitrophenyl-dithio- 1,1¢,2-trisnorsqualene (CNDT-squalene) showed that the cysteine residue at the channel constriction was covalently modified by the squalene-like inhibitor. Time-dependent inactivation of the C25S/C50S/C455S/C537S mutant by a number of squalene analogues and other agents with thiol- modifying activity suggested that modifying C435 caused the obstruction of the channel constriction thus blocking access of the substrate to the active site. The tryptic fragment comprising C435 of the quadruple mutant labelled with the most effective inhibitor had the expected altered molecular mass, as determined by LC-ESI-MS measurements. The arrangement of the substrate in the active site cavity was studied by using thiol reagents as probes in labelling experiments with the double mutant D376C/C435S in which D376, supposedly the substrate-protonating residue, was substituted by cysteine. The inhibitory effect was evaluated in terms of the reduced ability to cyclize oxidosqualene, as the mutant is unable to catalyse the reaction of squalene to hopene. Among the inhibitors tested, the substrate analogue squalene-maleimide proved to be a very effective time- dependent inhibitor. Keywords: Alicyclobacillus acidocaldarius;membrane protein; site-directed mutagenesis; squalene cyclase; thiol reagents. Oxidosqualene cyclases (OSCs) and squalene-hopene cyclases (SHCs) are key enzymes in triterpenoid biosynthesis: they transform acyclic isoprenoid precursors into tetra- and pentacyclic compounds [1]. OSCs can be considered taxonomic markers, as they catalyse the conversion of 2,3- oxidosqualene into lanosterol in nonphotosynthetic organ- isms (fungi and mammals), and into cycloartenol and other tetra- and pentacyclic triterpenes in plants [2]. In prokary- otes, SHCs convert squalene into hopene or diplopterol (Fig. 1), pentacyclic triterpene precursors of hopanoids. These compounds are thought to have functions similar to those of sterols in eukaryotic membranes [3]. An important contribution to the understanding of the catalytic mechanisms controlled by OSCs and SHCs came from the crystal structure of SHC from Alicyclobacillus acidocaldarious [4,5]. X-ray analysis revealed a membrane protein with membrane-binding characteristics similar to those of two prostaglandin-H 2 synthase isoenzymes [6,7]. These membrane proteins are called monotopic as they are shaped so as to submerge from one side of the membrane into the nonpolar part of the phospholipid bilayer without protruding through it [8]. The enzyme has a hydrophobic plateau plunging into the lipophilic centre of the membrane. A nonpolar channel connects the plateau and the active centre through a narrow constriction formed by four amino-acid residues, which appear to act as a gate that permits substrate passage (Fig. 2). The cavity hosting the active site is lined by nonpolar residues, but has a highly polar patch at the top. It seems to be shaped so as to bind the substrate in a specific product-like conformation, to trigger cyclization by protonating a terminal double bond, to assist ring-closures by stabilizing the cationic intermedi- ates and, finally, to deprotonate the hopanyl cation to form hopene or, in a side reaction, to hydroxylate the cation to form hopan-22-ol (diplopterol). We report here a new approach for studying the access of the substrate to the active site cavity of A. acidocaldarius SHC (E.C. 5.4.99.x) and its arrangement in it. To this aim, Correspondence to G. Balliano, Dipartimento di Scienza e Tecnologia del Farmaco, Via P. Giuria 9, I-10125 Torino, Italy. Fax: +39 011 6707695, Tel.: +39 011 6707698, E-mail: gianni.balliano@unito.it Abbreviations: OSC, oxidosqualene cyclase; SHC, squalene-hopene cyclase; CNDT-squalene, 3-carboxy-4-nitrophenyl-dithio-1,1¢, 2-trisnorsqualene; U14266A, (U14), 3b-(2-dimethylaminoethoxy)- androst-5-en-17-one; CPTO, 2-(4-chlorophenyl)-D 2 -thiazoline-1- oxide; DTS, (dimeric thiolsulfinate), 2(4-chlorobenzamido) ethane- thiosulfinic acid S-2(4-chlorobenzamido) ethyl ester; NaB 3 H 4 ,sodium borotritiure (Ph) 3 P, triphenylphosphine, Enzymes: oxidosqualene cyclase (EC 5.4.99.7); squalene-hopene cyclase (EC 5.4.99.x). Note: P. Milla and A. Lenhart contributed equally to this work. Note: a web site is available at http://hal9000.cisi.unito.it/wf/DIPARTIMEN/Scienza_e_/index.htm (Received 1 November 2001, revised 18 February 2002, accepted 25 February 2002) Eur. J. Biochem. 269, 2108–2116 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02861.x critically located Cys residues, either present in native protein or inserted by site-directed mutagenesis, were labelled with different thiol-reacting molecules, designed and synthesized in our laboratories. MATERIALS AND METHODS NMR and MS of chemical products 1 H-NMR spectra were recorded on a Jeol EX-400 or Jeol GX-270, with SiMe 4 as internal standard. Mass spectra were obtained on a VG Analytical 7070 EQ-HF spectro- meter by electron impact ionization. IR and UV spectra were recorded, respectively, on Perkin-Elmer 781 and Beckman DU 70 spectrophotometers. Chemicals Light petroleum refers to the fractions of bp 40–60 °C. Tetrahydrofuran was distilled under sodium benzophenone ketyl. Silica gel was 70–230 mesh. Squalene was from Merck, polyoxyethylene 9 lauryl ether (polidocanol) was from Sigma-Aldrich, Italy), 2,3-oxidosqualene was prepared as described in [10]. 3-Carboxy-4-nitrophenyl-dithio-1,1¢,2- trisnorsqualene (CNDT-squalene) (1), dodecyl- (3)and squalene-maleimide (2) were synthetized as described else- where [11]. U14266A (U14; 3b-(2-dimethylaminoethoxy)- androst-5-en-17-one) [12] was provided by Upjohn Company. 2-(4-chlorophenyl)-D 2 -thiazoline was synthetized as reported previously [13]. m-chloroperbenzoic acid, cystamine dihydrochloride, 4-chlorobenzoyl chloride and triethylamine were purchased from Sigma-Aldrich. 2-(4-Chlorophenyl)-D 2 -thiazoline-1-oxide (CPTO) (4). To an ice-cold and well stirred solution of 2-(4-chlorophe- nyl)-D 2 -thiazoline (2 g, 0.0101 mol) in CH 2 Cl 2 (20 mL), m-chloroperbenzoic acid (2.05 g, 0.0101 mol) dissolved in CH 2 Cl 2 (40 mL) was slowly added. The mixture was stirred for 3 h on ice while a precipitate appeared. CHCl 3 (60 mL) was then added and the solution was washed with 5% NaHCO 3 (2 · 120 mL) and saturated brine (40 ng NaCl in 100 mL H 2 O; 1 · 100 mL). The organic phase was dried over anhydrous sodium sulfate and evaporated in vacuo. The crude product was purified by flash-chromatography using CHCl 3 as eluant to give 1.36 g CPTO (63% yield). ESI-MS m/z:213(M + , 11), 195 (100), 185 (8), 137 (43); 1 H- NMR (CDCl 3 ) d:3.14(m,1H,5-H a ), 3.36 (m, 1H, 5-H b ), 4.68 (m, 1H, 4-H a ), 4.91 (m, 1H, 4-H b ), 7.49 (d, 2H, aromatic protons), 8.04 (d, 2H, aromatic protons); IR (KBr) m max : 3375, 3040, 2980, 2920, 1612, 1590, 1485, 1400Æcm )1 ; UV (CH 3 OH) k max : 203, 260. N,N¢-(Dithiodi-2,1-ethanediyl)bis 4-chlorobenzamide. To an ice-cold and well stirred suspension of cystamine dihydrochloride (0.77 g, 0.00343 mol) and triethylamine (12 mL, great excess) in CHCl 3 (10 mL), 4-chlorobenzoyl chloride (1.5 g, 0.00857 mol) dissolved in chloroform (4 mL) was slowly added under argon atmosphere. The mixture was stirred overnight at room temperature. CHCl 3 (10 mL) was then added and the mixture was extracted with chloroform (2 · 50 mL) and chloroform/ethyl acetate 50/ 50 (1 · 50 mL). The pooled organic phases were washed with 5% NaHCO 3 (2 · 50 mL) and saturated brine (1 · 50 mL). After anhydrification with anhydrous sodium sulfate the solvent was evaporated under reduced pressure. The crude product was purified by column chromatography on silica with chloroform as eluant to give 1.05 g N,N¢- (dithiodi-2,1-ethanediyl)bis 4-chlorobenzamide (68% yield). ESI-MS m/z: 428 (M + , < 1), 215 (33), 182 (60), 139 (100), 111 (39); 1 H-NMR (CDCl 3 /CD 3 OD) d:2.81(t,4H,-CH 2 - S), 3.57 (t, 4H, -CH 2 -N), 7.23 (d, 4H, aromatic protons), 7.61 (d, 4H, aromatic protons); IR (KBr) m max : 3302, 3236, 1638, 1628, 1597, 1541, 1489Æcm )1 ;UV(CH 3 OH) k max : 204, 235. 2(4-Chlorobenzamido) ethanethiosulfinic acid S-2(4- chlorobenzamido) ethyl ester (DTS ¼ dimeric thiolsulf- inate) (5). A solution of N,N¢-(dithiodi-2,1-ethanediyl)bis 4-chlorobenzamide (400 mg, 0.932 mmol) dissolved in CH 2 Cl 2 (25 mL) was stirred while m-chloroperbenzoic acid (85% purity, 147 mg, 0.932 mmol) was slowly added; it was then allowed to react for a further 3 h with continuous stirring. CH 2 Cl 2 (20 mL) was then added and the mixture was washed with 5% NaHCO 3 (2 · 30 mL) and saturated brine (1 · 50 mL). After anhydrification with anhydrous Fig. 1. Reactions catalysed by (A) prokaryotic SHC and (B) eukaryotic OSC. Fig. 2. Surface representation of SHC sliced in the middle of the molecule with nonpolar (yellow), positively charged (blue) and negatively charged (red) areas. The large internal cavity is connected with the hydrophobic plateau on the right by a nonpolar channel. A detergent molecule LDAO that has been found in a crystal structure of SHC [4] is shown as ball-and-stick model. The catalytic acid D376 and the residues forming the channel constriction are indicated (V174 lying in front of C435 was omitted for clarity). The figure was produced with GRASP [9]. Ó FEBS 2002 Effects of SH-modifying agents on squalene cyclase (Eur. J. Biochem. 269) 2109 sodium sulfate the solvent was evaporated under reduced pressure. The crude product was purified by column chromatography on silica with CHCl 3 andthenCHCl 3 / CH 3 OH 98 : 2 to give DTS (5) (220 mg, 53% yield). ESI- MS m/z: 429 (M ± 16, 54), 214 (29), 182 (44), 156 (35), 139 (100), 111 (54); 1 H-NMR (CD 3 OD) d:3.57(m,4H,-CH 2 - SO and -CH 2 -S), 3.82–3.93 (m, 4H, -CH 2 -N), 7.53 (m, 4H, aromatic protons), 7.88 (m, 4H, aromatic protons); IR (KBr) m max : 3680, 3412, 2920, 1664, 1597, 1541, 1480Æcm )1 ; UV (CH 3 OH) k max : 203, 236. Radiochemicals [2- 14 C]-mevalonate (50 mCiÆmmol )1 ) was from NEN. [ 14 C]- squalene and [ 14 C]-3S-2,3-oxidosqualene were prepared by incubating an S 10 supernatant (25 mg proteins) of a pig liver homogenate with 1 lCi [ 14 C]-mevalonate in the presence of the OSC inhibitor U14266A (U14) [12], essentially as described by Popjak [14]. The nonsaponifiable lipids were separated by two-step TLC on silica gel plates (Merck) (20 · 20 cm, 0.5 mm layer). The plates were first developed in light petroleum to a height of about 10 cm above the origin. After drying, the plates were developed to 15 cm above the origin with n-hexane/ethyl acetate (90 : 10; v/v). Radioactive areas corresponding to squalene and 2,3-oxido- squalene were scraped off and eluted with dichloromethane. The solvent was dried under N 2 and [ 14 C]-squalene and [ 14 C]-3S-2,3-oxidosqualene were dissolved in benzene. The radiochemical purity of products was evaluated by scanning TLC plates with a System 2000 Imaging Scanner (Packard). Radioactivity was measured by Liquid Scintillation Count- ing (Beckman). All of the radiolabelled compounds were compared chromatographically with authentic radio-inert samples. Determination of the radioactive substances and isotope counting were carried out as already described [15,16]. Radiolabelled CNDT-squalene (1) was synthesized via the following steps (Fig. 3): (i) synthesis of [1- 3 H]trisnor- squalene alcohol; (ii) synthesis of [1- 3 H]trisnorsqualene thioacetate; (iii) synthesis of [1- 3 H]trisnorsqualene thiol; (iv) transformation of [1- 3 H]trisnorsqualene thiol into [ 3 H]- CNDT-squalene (1). (i) [1- 3 H]Trisnorsqualene alcohol: [1- 3 H]-(4E,8E,12E, 16E)-4,8,13,17,21-pentamethyl-4,8,12,16,20-docosapentaen- 1-ol. Pure trisnorsqualene aldehyde, obtained as described by Ceruti et al. [17] (20.5 mg, 0.053 mmol) was dissolved in methanol (0.5 mL) and added to the NaB 3 H 4 -containing phial (total activity 25.0 mCi, specific activity 500 mCiÆ mmol )1 , total amount 0.05 mmol). After 3 h, NaBH 4 (excess, 6.0 mg, 0.159 mmol) was added to complete the reaction. After an additional hour, the methanol was evaporated under nitrogen and the reaction mixture was dissolved in dichloromethane, transferred to a single necked flask, and evaporated to dryness under reduced pressure to give crude [1- 3 H]trisnorsqualene alcohol. The product was purified, after dissolution with light petroleum, by column chromatography on silica gel with 100% light petroleum to remove impurities, then light petroleum/diethylether 90 : 10 to give 18 mg (0.046 mmol) of pure [1- 3 H]trisnorsqualene alcohol. The radiochemical purity of the alcohol was determined by radiochromatogram with light petroleum/ diethylether 80 : 20 and then revealed with iodine vapour. Total activity: 4.2 mCi; specific activity: 92 mCiÆmmol )1 ; chemical yield: 88%. (ii) [1- 3 H]Trisnorsqualene thioacetate: [1- 3 H]-(4E,8E, 12E,16E) S-[4,8,13,17,21-pentamethyl-4,8,12,16,20-docos- apentaenyl] thioacetate. A solution of diisopropyl azodi- carboxylate (71.6 mg, 0.354 mmol) in 0.5 mL anhydrous tetrahydrofuran was added to a well-stirred solution of triphenylphosphine (92.5 mg, 0.350 mmol) in 3 mL anhy- drous tetrahydrofuran at 0 °C. The mixture was stirred at 0 °C for 30 min and produced a white precipitate. A solu- tion of [1- 3 H] trisnorsqualene alcohol (18 mg, 0.046 mmol) and thiolacetic acid (37.2 mg, 0.488 mmol) in 0.5 mL anhydrous tetrahydrofuran was added dropwise under nitrogen and the mixture stirred for 1 h at 0 °Cand1h at room temperature. The mixture was evaporated under nitrogen and the crude product triturated in light petroleum; the suspension was purified by column chromatography on silica gel with 100% light petroleum to remove impurities, then light petroleum/diethylether 99.5 : 0.5 to give pure [1- 3 H] trisnorsqualene thioacetate (16.5 mg, 0.037 mmol). The radiochemical purity of the thioacetate was determined by radiochromatogram with light petroleum/diethylether 98 : 2 and then revealed with iodine vapour. Total activity: 3.6 mCi; specific activity: 92 mCi mmol )1 ; chemical yield: 80%. (iii) [1- 3 H]Trisnorsqualene thiol: [1- 3 H]-(4E,8E,12E,16E) 4,8,13,17,21-pentamethyl-4,8,12,16,20-docosapentaenyl 1-thiol. [1- 3 H]trisnorsqualene thioacetate (16.5 mg, 0.037 mmol) was dissolved in anhydrous diethylether (2 mL) and added dropwise to a suspension of LiAlH 4 (11.6 mg, 0.31 mmol) in anhydrous diethylether (5 mL) under nitrogen. The mixture was stirred 25 min at room temperature then 25 min under reflux. LiAlH 4 excess was destroyed by the careful addition of 7 mL 1 M HCl solution. The ether layer was separated and the aqueous phase extracted with dichloromethane. The combined organic phases were dried over sodium sulfate and the solvent evaporated under reduced pressure. The crude product was used without purification for the following step. (iv) [1- 3 H]Trisnorsqualene nitrobenzoic acid (1): [1- 3 H]- 6-nitro-3-[(4E,8E,12E,16E)-4,8,13,17,21-pentamethyl-4,8,12, 16,20-docosapentaenyldisulfamyl] benzoic acid; 5,5¢-dithio- bis(2-nitrobenzoic acid) (41 mg, 0.103 mmol) was dissolved Fig. 3. Scheme for the synthesis of [1- 3 H] CNDT-squalene. The asterisk indicates position of 3 H-label: (i) NaB 3 H 4 (ii) (Ph) 3 P, CH 3 COSH, diisopropyl azodicarboxylate; (iii) LiAlH 4 (iv) 5,5¢-dithiobis(2-nitro- benzoic acid). 2110 P. Milla et al. (Eur. J. Biochem. 269) Ó FEBS 2002 in ethanol (9 mL). To this solution [1- 3 H] trisnorsqualene thiol dissolved in ethanol (10 mL) was added under nitrogen. The mixture was stirred overnight at room temperature and then the solvent evaporated under reduced pressure. The resulting yellow oil was purified by column chromatography on silica gel with dichloromethane/meth- anol 92 : 8 as eluent to give [1- 3 H] trisnorsqualene nitro- benzoic acid (1) (11.1 mg, 0.0185 mmol). The radiochemical purity of the disulfide was determined by radiochromato- gram with dichloromethane/methanol 85 : 15 and then revealed with iodine vapour. Total activity: 0.91 mCi; specific activity: 92 mCiÆmmol )1 ;chemicalyield:50%. Expression and purification of recombinant SHC Wild-type SHC was kindly provided by Prof. K. Poralla (Universita ¨ tTu ¨ bingen) [18]. For production of recombinant SHC, the overexpression system described elsewhere [19] was used. Mutants D376C/ C435S, C455S and C50S had been generated for structure analysis [5] utilizing the phosphorothioate method (Sculp- tor TM , Amersham). To generate the quadruple mutant, gene fragments containing mutations C455S and C50S were introduced into expression plasmid pKSHC1 using restriction sites SacI/HindIII and EcoRI/ApaI, respectively. Mutations C25S and C537S were created with the megaprimer method [20,21]. The first round of PCR amplification was per- formed with Pwo polymerase (Peqlab, Erlangen, Germany) and primers MP-C25S and MP-ApaI or MP-C537S and MP-HindIII, respectively. The PCR products were purified by agarose gel electrophoresis and gel extraction (QIAquick ` Gelextraction-Kit, Qiagen) and were used as ÔmegaprimersÕ in the second round of PCR with the additional primers MP-EcoRI or MP-SacI. Sequences of the synthetic oligonucleotides (MWG, Ebersberg, Germany) were as follows (with changes from the wild-type sequence under- lined): Mp-C25S, 5¢-CTCCTCTCC AGCCAAAAGG-3¢; MP-C537S, 5¢-GGCGAGGAC AGCCGCTCGTAC-3¢; MP-ApaI, 5¢-GTACAGGGCCCACGTGCCG-3¢;MP- EcoRI, 5¢-AACAGAATTCATGGCTGAGCAGTTGG TG-3¢; MP-HindIII, 5¢-CAGCCAAGCTTGCATGCCTG -3¢;MP-SacI,5¢-CATGCAGAGCTCGAAC GGCG-3¢. The purified PCR products were digested with EcoRI/ ApaIandSacI/HindIII, respectively, purified by agarose gel electrophoresis and gel extraction, and were ligated into the vector fragments obtained in an analogous manner. Ligation products were transformed into Escherichia coli JM105 cells according to standard protocols [22]. Mutagenesis results were validated by DNA sequencing (SeqLab, Go ¨ ttingen, Germany). Expression and purification of the mutant SHC was performed as described by Wendt et al.[19]. Enzyme assay Cyclization of squalene. Purified SHC (3–5 lg) was incubated at 55 °C for 30 min in 1 mL 0.1 M Na citrate buffer (pH 6.0) containing 1.5 mgÆmL )1 polidocanol and 10 l M [ 14 C]squalene (3000 c.p.m.). The reaction was stopped by adding 1 mL 10% KOH in MeOH, and the nonsaponifiable lipids were extracted twice with 1 mL petroleum ether. The extract was chromatographed on silica gel plates developed in petroleum ether. The radioactivities of squalene, hopene and diplopterol ( 15% of products formed) were evaluated by a System 2000 Imaging Scanner (Packard). Cyclization of oxidosqualene. Purified SHC (3–80 lg) was incubated at 55 °Cfor30minin1mL0.1 M Na citrate buffer (pH 6.0) containing 1.5 mgÆmL )1 polidocanol and 10 l M [ 14 C]oxidosqualene (3000 c.p.m.). The reaction was stopped with KOH and the nonsaponifiable lipids were extracted as described for squalene. The extract was chromatographed on silica gel plates developed in CH 2 Cl 2 . The radioactivities of chromatographic bands (oxidosqua- lene and hop-22(29)-en-3-ol) [23] were evaluated as des- cribed for squalene. Time-dependent inactivation SHC (0.12–3.2 mgÆmL )1 ) was preincubated at 55 °Cin0.1 M Na citrate buffer (pH 6.0) containing 1.5 mgÆmL )1 polido- canol and different concentrations of reagents. After the preincubation period, the inhibitor concentration was de- creased by 40-fold dilution, and the squalene or oxidosqua- lene cyclizing activity was determined as described above. Labelling with radioactive CNDT-squalene (1) SHC wild-type and mutants (40 lg) were incubated at 55 °C for1hin0.1 M Na citrate buffer (pH 6.0) containing 1.5 mgÆmL )1 polidocanol and 2 m M [ 3 H]-CNDT-squalene (1)(10 · 10 6 d.p.m., 5.26 · 10 10 d.p.m./mmol). The enzyme was precipitated with 5 vol. cold acetone. The precipitate was washed twice with cold acetone and resuspended in sample buffer for SDS/PAGE with or without 2-mercaptoethanol. The samples were then analysed by 10% SDS/PAGE, the gel was stained, enhanced with 20% 2,5-diphenyloxazole in acetic acid, washed in water, dried and exposed to a Kodak X-OMAT XAR-5 film at )80 °C for 7–10 days. Digestion and MS analysis of labelled protein For tryptic proteolysis, labelled protein (1 mg) dissolved in buffer-B (20 lL, 20 m M Tris/HCl, pH 8.0; 0.6% C 8 E 4 ; 200 m M NaCl) was precipitated with 80 lLEtOHat0°C to eliminate any detergent present. Precipitated protein was resuspended in 50 lL buffer-D (10 m M N-ethyl-morpho- line-HOAc, pH 7.8, 2 m M CaCl 2 ) and digested with 10 lg trypsin (Promega) at 37 °C and occasional agitation for 2 h. Trypsin addition and incubation were repeated, the protease was heat inactivated (5 min at 98 °C) and fragments were analysed by LC-ESI-MS. Ten lLdigestedSHCwere applied onto a Phenomenex RP column (Jupiter 5 l C4300 A, 150 mm · 2.0 mm), the fragments were eluted in a gradient 0–50% CH 3 CN in 0.1% formic acid and mapped by ESI-MS (Finnigan TSQ7000, scanzone 200–2300 amu/z). Deconvolution and data analysis were performed with the program BIOWORKS 8.2 (Finnigan). RESULTS Blocking substrate access to the active centre The idea of inhibiting the enzyme by blocking substrate access to the active centre stemmed from the observation Ó FEBS 2002 Effects of SH-modifying agents on squalene cyclase (Eur. J. Biochem. 269) 2111 that one of the four amino-acid residues present at the constriction of the hypothetical access channel is a cysteine (C435), an amino-acid residue that is easy to modify (Fig. 2). The possible inhibitory effect of labelling the C435 residue with a series of substrate analogues or other thiol modifyingagents(Fig.4)wasthenstudied.AsSHC possesses five cysteine residues and no disulfide bridges, labelling experiments were performed not only on the native protein but also on two mutants obtained by site-directed mutagenesis: C435S, lacking C435 at the channel constric- tion, and the C25S/C50S/C455S/C537S-mutant (Ôquadruple mutantÕ), bearing C435 as its only cysteine residue. The mutants were first characterized for their optimal tempera- ture and kinetic parameters (Table 1). All of the mutants showed the optimal temperature at 60 °C like wild-type SHC (data not shown). The quadruple mutant proved to be less active than either the wild-type or the single mutant, as indicated by the values of k cat and k cat /K M . A radioactive thiol modifying squalene-analogue [(CNDT-squalene (1)] was first used to test the ability of a squalenoid molecule to reach the channel constriction. Native SHC and quadruple mutant (40 lgprotein)were incubated separately in 0.1 M Na citrate buffer (pH 6) containing 1.5 mgÆmL )1 polidocanol, for 60 min at 55 °C with 0.2 m M [1- 3 H]-CNDT-squalene (1)(10· 10 6 d.p.m., 5.26 · 10 10 d.p.m.Æmmol )1 ). SDS/PAGE of modified enzyme followed by fluorography were then used to verify labelling of the protein by the inhibitor. An aliquot of incubated protein was treated with 2-mercaptoethanol followed by electrophoretic analysis to verify formation of disulfide bridges between thiol residues of the protein and the squalenoid radioactive moiety of the inhibitor. As shown by the results reported in Fig. 5, both wild-type protein and mutant were covalently labelled by the radio- active inhibitor. The unambiguous labelling of the quad- ruple mutant, bearing only C435, indicates that inhibitor has reached and modified the cysteine residue at the channel constriction. Inactivation experiments carried out at con- centrations of inhibitors up to 0.5 m M showed that CNDT- squalene (1) behaves as a poorly irreversible inhibitor of the quadruple mutant. Fig. 4. Structures of synthesized inhibitors. Table 1. Kinetic parameters for the wild-type and mutant SHCs with squalene or oxidosqualene as substrates. The initial rates were measured at 55 °C as described in Material and methods. The kinetic values were determined from double-reciprocal plots. NA, not applicable. Protein Substrate: squalene Substrate: oxidosqualene K M (l M ) k cat (min )1 ) k cat /K M (min )1 l M )1 ) K M (l M ) k cat (min )1 ) k cat /K M (min )1 l M )1 ) Wild-type 13.2 ± 2.5 2.08 ± 0.28 0.16 1.5 ± 0.6 0.547 ± 0.07 0.304 C435S 9.5 ± 2.9 1.43 ± 0.19 0.15 1.4 ± 0.5 0.232 ± 0.02 0.166 C25S/C50S/C455S/C537S 13.6 ± 5.4 0.39 ± 0.16 0.03 1.0 ± 0.5 0.042 ± 0.01 0.042 D376C/C435S NA NA NA 1.3 ± 0.2 0.020 ± 0.002 0.015 Fig. 5. SDS/PAGE and fluorography of SHC wild-type and quadruple mutant incubated with [1- 3 H] CNDT-squalene (1). Coomassie stained SDS/PAGE (lanes 1 and 2) and fluorography (lanes 3–6). Samples without 2-mercaptoethanol treatment (lanes 3 and 5) and with 2-mercaptoethanol treatment (lanes 4 and 6). M, markers. 2112 P. Milla et al. (Eur. J. Biochem. 269) Ó FEBS 2002 The effect of obstructing the channel constriction bearing C435 was then evaluated by treating the quadruple mutant, having C435 as the only thiol residue, with other thiol- modifying agents of different shapes and sizes (Fig. 4). Some of them [squalene-maleimide (2) and dodecyl-malei- mide (3)] were designed to link a cysteine residue with a maleimide ring bearing a flexible lipophilic chain, others [CPTO (4)andDTS(5)] to modify the thiol residue with an arm bearing the bulky and rigid chlorophenylketone group at its ends. The effect of the inhibitors on the quadruple mutant was compared with that observed on the C435S mutant, lacking only the cysteine residue of the channel constriction (Table 2). Inhibitors proved to have little effect on the mutant lacking the cysteine at the channel constriction, whereas they were all good inactivating agents for the mutant bearing only the cysteine residue of the channel constriction. DTS (5) was the most effective inactivating agent of the quadruple mutant, as indicated by t ½ inactivation values (Table 2 and Fig. 6). Covalent modification of C435 by DTS (5) was confirmed by tryptic digestion and MS analysis of the quadruple mutant treated with the inhibitor. The DTS-labelled tryptic peptide comprising C435 has a theor- etical molecular mass of 6895.8 Da. Deconvoluted mass spectra (Fig. 7B) indicated a molecular mass of 6893.0 Da and refined data analysis a molecular mass of 6894.2 Da for the expected peptide recorded by scans nos 1481–1486. The corresponding peak with elution time 47.85 min was the most prominent peak of the chromatogram (Fig. 7A). A difference between the theoretical and the experimental values of 1.6 Da can be tolerated due to the systematic error of the mass spectrometer (0.01–0.05%). Even if an incom- plete tryptic digest of SHC with up to two missed cleavages is taken into account, it can be ruled out that other peptides caused the observed MS signals, because potential tryptic peptides are residues 403–466 (one missed cleavage, 7042.6 Da instead of the observed value 6894.2 Da), the unlabelled peptide containing C435 (6682.2 Da) and resi- dues 274–332 (two missed cleavages, 6500.4 Da). Further- more, due to the high intensity of the chromatogram peak assigned to the labelled peptide it is unlikely that peptides originating from unspecific cleavage of the labelled SHC and having a molecular mass close to that of the labelled peptide contributed to the deconvolution peak. Therefore the identification of the labelled peptide could be interpreted as unambiguous. Binding of a squalene analogue inside the active centre Labelling a specific amino-acid residue at the active centre with a substrate analogue is a crucial step in studying the Table 2. Inactivating effect t ½ of thiol-modifying inhibitors on enzy- matic activity of mutant C435S and quadruple mutant C25S/C50S/ C455S/C537S with squalene as substrate. Experimental conditions were as described in Materials and methods. Data are mean values from three independent experiments with a mean deviation of ± 10%. Inhibitor Inhibitor concentration (l M ) t ½ (min) C435S C25S/C50S/ C455S/C537S Dodecyl-maleimide (3) 500 > 60 30 Squalene-maleimide (2) 200 > 60 45–60 CPTO (4) 200 > 60 23 DTS (5) 100 45 4.5 DTS (5) 20 > 60 22 Fig. 6. Half-logarithmic plot of time-dependent inactivation of mutant C25S/C50S/C455S/C537S by DTS (5). Data are shown for DTS (5) concentrations of 100 l M (r)and20l M (j). Fig. 7. LC-ESI-MS results: RP-HPLC separation of tryptic digested mutant C25S/C50S/C455S/C537S labelled with DTS (5). Elution time [min] is denoted above the peaks (A). Deconvoluted spectra of ESI-MS scans nos 1481–1486 corresponding to the elution time 47.85 min (B). Ó FEBS 2002 Effects of SH-modifying agents on squalene cyclase (Eur. J. Biochem. 269) 2113 complex interaction between substrate and enzyme. This approach was successfully adopted with yeast oxidosqua- lene cyclase, which was irreversibly inhibited by CNDT- squalene (1), a thiol-reacting squalene-like molecule [24]. SHCs, unlike OSCs, bear no cysteine residues at the active centre, as they lack the cysteine residue present in the conserved active site motif DCTA of eukaryotic OSCs (in prokaryotic cyclases the motif is DDTA). No other cysteine residues appear at the active centre cavity of SHC. The goal of binding a squalene analogue inside the active site of SHC covalently may thus be pursued by introducing a critical cysteine residue into the active centre cavity by site-directed mutagenesis, able to serve as a Ôsticky pointÕ for squalene analogues with thiol-modifying activity. The D376C/C435S mutant, already used in crystallization experiments of SHC [5], seemed to fit the above purpose. In this mutant, the D376 at the reaction initiation site of the central cavity has been replaced by a cysteine residue, and C435 at the channel constriction has been replaced by serine. The latter substi- tution was required to allow thiol-reacting molecules to move to their target located in the active site cavity, without interacting with the cysteine residue at the channel constric- tion. A series of thiol-reacting agents were tested as time- dependent inhibitors of the D376C/C435S mutant. The SHC mutant C435S was used as control. Mutant D376C/ C435S lost the ability to cyclize squalene due to the absence of one of the two aspartate residues in the crucial DDTA motif [25], but it still shared the ability to cyclize oxido- squalene with wild-type SHC [23] and with the control C435S mutant. Therefore, the inhibitory properties of thiol reagents were assayed towards the oxidosqualene cyclizing activity. First, enzymatic activity of the mutants with oxidosqualene as a substrate, was characterized and compared with wild-type enzyme. Both the D376C/C435S and the C435S mutants showed the same temperature profile (data not shown) for oxidosqualene cyclizing activ- ity, with maximum activity at 60 °C, coincident with the optimal temperature for squalene cyclization by wild-type SHC [3]. Comparison of kinetic parameters showed that the double mutant was less efficient in cyclizing oxidosqualene than either the wild-type or the mutants bearing the native DDTA motif (Table 1). All thiol-reacting agents tested had a stronger inhibitory effect on the D376C/C435S mutant, which was time- dependent, than they did on the C435S mutant with intact initiation site motif DDTA (Table 3). The reactive sub- strate-analogue squalene-maleimide appeared to be the most effective inhibitor of the D376C/C435S mutant bearing a cysteine residue at the active site. DISCUSSION A series of thiol modifying agents were used as tools to elucidate some structural/functional features of squalene- hopene cyclase. The study started from the observation that C435 of the enzyme, is located on the putative path of the substrate from the membrane interior to the active centre. C435 is in fact located at a constriction formed by four amino-acid residues, which separates the large central cavity containing residues critical for catalysis from a lipophilic channel open towards the inner surface of the protein (Fig. 2). This constriction appears to be sufficiently mobile to act as a gate for substrate passage, due to the flexibility of a loop bearing two of the four amino-acid residues as indicated by the higher crystallographic B-factors [5]. The ability of the inhibitor CNDT-squalene (1)tomodifyC435 covalently provided the first evidence that a substrate analogue can move along the lipophilic channel and reach the enzyme’s putative gate-constriction, establishing that this is in fact the entrance to the active centre. Even stronger support for the position of the entrance gate came from the inactivating experiments with the C25S/ C50S/C455S/C537S mutant, which bears C435 as the only cysteine of the protein. When this SHC-mutant was exposed to thiol-modifying agents, especially to the inhibitor DTS (5), rapid time-dependent inhibition was observed (Fig. 6). Such inhibition can be explained as a consequence of an obstruction of the channel constriction. Other explanations, such as a direct influence of the inhibitor on the catalytic process, may be ruled out since C435 is not involved in the reaction mechanism as shown by the essentially unchanged activity of the C435S mutant compared to wild-type (Table 1). Moreover, when the inhibitors could in principle overcome the barrier of the channel constriction and act directly in the active site cavity, as in the case of the C435S mutant, poor inhibition effect was observed (Table 2). Interestingly, the inhibitors squalene- and dodecyl-malei- mide, which modify thiol residues with a group bearing a mobile and flexible chain, proved to be less active than CPTO (4)andDTS(5).Thereasonforthedifferent effectiveness of the inhibitors could, of course, simply result from their different reactivity or, as in the case of DTS (5), from some enhancement effect due to the formation of reactive intermediates during the reaction between the intact inhibitor and thiol groups [26,27]. Moreover, a difference depending on size and rigidness of groups modifying C435 may occur. The series of thiol-reacting agents used to study access of the substrate to the active site of SHC was also found to be a useful tool for studying the possibility of forming an enzyme–inhibitor complex with the inhibitor covalently bound inside the active centre cavity. The mutant used for the study was the D376C/C435S mutant bearing a cysteine residue at the reaction initiation site in the central cavity and a serine substituting the cysteine at the channel constriction. This mutant is unable to cyclize squalene [25], but recognizes 2,3-oxidosqualene as a substrate, an ability shared with wild-type SHC and with mutants bearing at least one of the two aspartate residues of the conserved sequence DDTA Table 3. Inactivating effect t ½ of thiol-modifying inhibitors on enzy- matic activity of mutant C435S and mutant D376C/C435S with oxidosqualene as substrate. Experimental conditions were as described in Materials and methods. Data are mean values from three inde- pendent experiments with a mean deviation of ± 10%. Inhibitor Inhibitor concentration (l M ) t ½ (min) C435S D376C/C435S Dodecyl-maleimide (3) 200 > 60 12 Squalene-maleimide (2) 200 > 60 1.5 CPTO (4) 200 > 60 36 DTS (5) 200 45 7.5 2114 P. Milla et al. (Eur. J. Biochem. 269) Ó FEBS 2002 [28]. Recently, the preference for oxidosqualene of such mutants was confirmed with a SHC mutant bearing the eukaryotic DCTAEA OSC-motif instead of the DDTAVV SHC-motif [28]. The exchange of the prokaryotic by the eukaryotic motif was carried out in [23]. The strong time-dependent inactivation of inhibitors on the D376C/C435S mutant, and the poor inhibition of the C435S mutant, indicate that inactivation occurs at C376. Interestingly, squalene-maleimide (2) [11] which should have a higher ability to deliver a reactive group to the squalene- hosting cavity of the enzyme, proved to be the most powerful inactivating agent. It appears possible that, for the first time, a covalent SHC–inhibitor complex has been formed, in which a squalene analogue in a noncyclized form is bound to the squalene-hosting cavity covalently. These results open the fascinating prospect of solving the structure of a SHC with a complexed squalenoid molecule, and determining in detail the interactions between residues at the active site and the squalene skeleton in its not-yet cyclized conformation. In summary, rationally designed mutants of SHC, C25S/ C50S/C455S/C537S (Ôquadruple mutantÕ)andD376C/ C435S, obtained by site-directed mutagenesis, have been effectively inhibited by thiol-reacting molecules DTS (5)and squalene-maleimide (2) in a time-dependent manner. Ex- perimental evidence suggests that DTS (5) inhibits the quadruple mutant by obstructing the substrate access to the active site, while squalene-maleimide (2) inactivates the D376C/C435S mutant by irreversible occupation of the active centre cavity. These results, in particular those relating to the D376C/C435S mutant, may be regarded as the first step towards the goal of preparing stable covalent complexes for structural analysis. ACKNOWLEDGEMENTS This work was supported by the Ministero dell’Universita ` edella Ricerca Scientifica e Tecnologica (MURST), Italy (ex 60%), and by the Deutsche Forschungsgemeinschaft under SFB-388. The authors thank Prof. K. Poralla (Universita ¨ tTu ¨ bingen) for providing wild-type SHC, D. Kessler and B. Fu ¨ llgrabe for help with mutagenesis and C. Warth for ESI-MS measurements. REFERENCES 1. Abe, I., Rohmer, M. & Prestwich, G.D. (1993) Enzymatic cycli- zation of squalene and oxidosqualene to sterols and triterpenes. Chem. Rev. 93, 2189–2206. 2. Nes, W.D. (1990) Control of sterol biosynthesis and its import- ance to developmental regulation and evolution. Recent Adv. Phytochem. 242, 283–327. 3. Ourisson, G., Rohmer, M. & Poralla, K. (1987) Characterization and partial purification of squalene-hopene cyclase from Bacillus acidocaldarius. Biochim. Biophys. Acta 881, 356–363. 4. Wendt, K.U., Poralla, K. & Schulz, G.E. (1997) Structure and function of squalene cyclase. Science 277, 1811–1815. 5. Wendt, K.U., Lenhart, A. & Schulz, G.E. (1999) The structure of the membrane protein squalene-hopene cyclase at 2 A ˚ resolu- tion. J. Mol. Biol. 286, 175–187. 6. Picot, D., Loll, P.J. & Garavito, R.M. (1994) The X-ray crystal structure of the membrane protein prostaglandin H 2 synthase-1. Nature 367, 243–249. 7. Luong, C., Miller, A., Barnett, J., Chow, J., Ramesha, C. & Browner, M.F. (1996) Flexibility of the NSAID binding site in the structure of human cyclooxygenase-2. Nature Struct. Biol. 3, 927– 933. 8. Blobel, G. (1980) Intracellular protein topogenesis. Proc. Natl Acad.Sci.USA77, 1496–1500. 9. Nicholls, A., Sharp, K.A. & Honig, B. (1991) Protein folding and association: insight from the interfacial and thermodynamic properties of hydrocarbons. Proteins: Struct. Funct. Genet. 11, 281–296. 10. Duriatti, A., Bouvier-Nave ´ , P., Benveniste, P., Schuber, F., Delprino, L., Balliano, G. & Cattel, L. (1985) In vitro inhibition of animal and higher plants, 2,3-oxidosqualene-sterol cyclases by 2-aza-2,3-dihydrosqualene and derivatives and by other ammonium-containing molecules. Biochem. Pharmacol. 34, 2765– 2777. 11. Grosa, G., Viola, F., Ceruti, M., Brusa, P., Delprino, L., Dosio, F. & Cattel, L. (1994) Synthesis and biological activity of a squale- noid maleimide and other classes of squalene derivatives as irre- versible inhibitors of 2,3-oxidosqualene cyclase. Eur. J. Med. Chem. 29, 17–23. 12. Field, R.B., Holmlund, C.E. & Whittaker, N.F. (1979) The effects of the hypocholesteremic compound 3 b-(b-dimethylaminoeth- oxy) -androst-5-en-17-one on the sterol and steryl ester composi- tion of Saccharomyces cerevisiae. Lipids 14, 741–747. 13. Caujolle, R., Favrot, J.D., Loiseau, P.R., Payard, M., Amarouch, H., Lazrek, H., Seguela, J.P., Loiseau, P.M., Bories, C. & Gayral, P. (1991) Synthe ` se, activite ´ antiparasitaire et antifongique d’aryl-2- thiazolines. Pharm. Acta Helv. 66, 237–240. 14. Popjak, G. (1969) Enzymes of sterol biosynthesis in liver and intermediates of sterol biosynthesis. Methods Enzymol. XV, 393– 454. 15. Ceruti, M., Grosa, G., Rocco, F., Dosio, F. & Cattel, L. (1994) A convenient synthesis of [3- 3 H]-squalene and [3- 3 H]-oxidosqualene. J. Labelled Compounds Radiopharmaceuticals 34, 577–585. 16. Balliano, G., Viola, F., Ceruti, M. & Cattel, L. (1992) Char- acterization and partial purification of squalene-2,3-oxide cyclase from Saccharomyces cerevisiae. Arch. Biochem. Biophys. 293, 122– 129. 17.Ceruti,M.,Balliano,G.,Viola,F.,Cattel,L.,Gerst,N.& Schuber, F. (1987) Synthesis and biological activity of azasqua- lenes, bis-azasqualenes and derivatives. Eur. J. Med. Chem. 22, 199–208. 18. Ochs,D.,Kaletta,C.,Entian,K.D.,Beck-Sickinger,A.&Poralla, K. (1992) Cloning, expression, and sequencing of squalene-hopene cyclase, a key enzyme in triterpenoid metabolism. J. Bacteriol. 174, 298–302. 19. Wendt, K.U., Feil, C., Lenhart, A., Poralla, K. & Schulz, G.E. (1997) Crystallization and preliminary X-ray crystallographic analysis of squalene-hopene cyclase from Alicyclobacillus acid- ocaldarius. Protein Sci. 6, 722–724. 20. Barik, S. (1993) Site-directed mutagenesis by double polymerase chain reaction. In PCR Protocols; Current Methods and Applica- tions (White, B.A., ed.), pp. 277–286. Humana Press, Totowa, New Jersey. 21. Sakar, G. & Sommer, S.S. (1990) The ÔmegaprimerÕ method of site- directed mutagenesis. Biotechniques 8, 404–407. 22. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring. Harbor Laboratory Press, Cold Spring Harbor, New York. 23. Dang, T. & Prestwich, G.D. (2000) Site-directed mutagenesis of squalene-hopene cyclase: altered substrate specificity and product distribution. Chem. Biol. 7, 463–469. 24. Balliano, G., Grosa, G., Milla, P., Viola, F. & Cattel. L. (1993) 3-Carboxy-4-nitrophenyl-dithio-1,1¢,2-trisnorsqualene: a site-direc- ted inactivator of yeast oxidosqualene cyclase. Lipids 28, 903–906. 25. Feil, C., Su ¨ ssmuth, R., Jung, G. & Poralla, K. (1996) Site-directed mutagenesis of the putative active-site residues in squalene-hopene cyclase. Eur. J. Biochem. 242, 51–55. Ó FEBS 2002 Effects of SH-modifying agents on squalene cyclase (Eur. J. Biochem. 269) 2115 26. Kice, J.L. & Cleveland, J.P. (1973) Nucleophilic substitution reaction involving sulfenic acids and sulfenyl derivatives. The nucleophile- and acid-catalyzed oxygen-18 exchange of phenyl benzenthiosulfinate. J. Am. Chem. Soc. 95, 104–109. 27. Oae, S. & Okuyama, T. (1992) Organic Sulfur Chemistry: Bio- chemical Aspects. CRC Press Inc. Boca Raton, Florida. 28. Sato, T. & Hoshino, T. (1999) Functional analysis of the DXDDTA motif in squalene-hopene cyclase by site-directed mutagenesis experiments: initiation site of the polycyclization reaction and stabilization site of the carbocation intermediate of the initially cyclized A-ring. Biosci.Biotechnol.Biochem.63, 2189– 2198. 2116 P. Milla et al. (Eur. J. Biochem. 269) Ó FEBS 2002 . and squalene- hopene cyclases (SHCs) are key enzymes in triterpenoid biosynthesis: they transform acyclic isoprenoid precursors into tetra- and pentacyclic. Italy The function of squalene- hopene cyclase from Alicycloba- cillus acidocaldarius was studied by labelling critical cysteine residues of the enzyme,