ARCHIVES
OF BIOCHEMISTRY
Vol. 293, No. 1, February
AND BIOPHYSICS
14, pp. 122-129,1992
Characterization and Partial Purification of Squalene2,3-Oxide Cyclase from Saccharomyces cerev&jae’ Gianni Balliano,
Franca Viola, Maurizio
Ceruti, and Luigi Catte12
Istituto di Chimica Farmaceutica Applicata, Corso Raffaello 31, 10125 Torino, Italy
Received July 3, 1991, and in revised form September 27, 1991
The membrane nature of squalene oxide cyclase from Saccharomyces cerevisiae was investigated by comparing properties of the enzyme recovered from both microsomes and the soluble fraction of the yeast homogenate. The “apparent soluble” form and microsomal form of the enzyme were both stimulated by the presence of mammalian soluble cytoplasm and corresponded to one another in response to detergents Triton X- 100 and Triton X-114. The observed strong dependence of the enzyme activity on the presence of detergents and the behavior of the enzyme after Triton X-l 14 phase separation were peculiar to a lipophilic membrane-bound enzyme. A study of the conditions required to extract the enzyme from microsomes confirmed the lipophilic character of the enzyme. Microsomes, exposed to ipotonic conditions to remove peripheral membrane proteins, retained most of the enzyme activity within the integral protein fraction. Quantitative dissociation of the enzyme from membranes occurred only if microsomes were treated with detergents (Triton X-100 or octylglucoside) at concentrations which alter membrane integrity. The squalene oxide cyclase was purified 140 times from yeast microsomes by (a) removal of peripheral proteins, (b) extraction of the enzyme from the integral protein fraction with octylglucoside, and (c) separation of the solubilized proteins by DEAE Bio-Gel A chromatography. Removal of the peripheral proteins seemed to be a key step necessary for obtaining high yields. o 1992 Academic PWS, I~C.
The central role played by squalene-2,3-oxide cyclases (EC 5.4.99.7) in sterol biosynthesis has often been described (1, 2). These enzymes catalyze the cyclization of the substrate into different cyclic triterpenoids in higher 1 This work was supported by grants from Minister0 e della Ricerca Scientifica e Tecnologica-Italy. * To whom correspondence should be addressed.
122
dell’universita
plants, whereas in animals, yeast, and fungi the main cyclization product is lanosterol (1). The complex catalytic mechanism of these enzymes has attracted much research, mainly with mammalian squalene oxide cyclase, on the cyclization process and backbone rearrangement of squalene oxide (3-7). Recently, important steps, with the mammalian enzymes, were taken by two groups who solubilized and purified squalene oxide cyclase from hog and rat liver (8,9). As we were interested in comparing squalene oxide cyclase from liver and yeast, as well as testing the pharmacological activity of new classes of inhibitors of these enzymes (lo-15), we thought it would be of particular interest to solubilize, purify, and characterize the squalene oxide cyclase from Saccharomyces cerevisiae. In a previous comparison of mammalian and yeast squalene oxide cyclase (16), it was concluded that, while the microsomal nature of the mammalian enzyme was clearly established, the yeast enzyme appeared to be soluble. However, all attempts to purify the yeast enzyme failed, resulting in the nearly complete loss of activity. This appears to have been the only attempt to characterize and purify squalene oxide cyclase from yeast. In the present paper we describe the characterization of yeast squalene oxide cyclase as a membrane-bound enzyme and the partial purification from microsomes. The hydrophobic character of the enzyme was investigated by treatments with different detergents and particularly by using the Triton X-114 phase separation method (17). MATERIALS
AND
METHODS
Chemicals. Squalene-2,3-oxide and [3-3H]squalene-2,3-oxide (44 X 10s dpm/pmol) were synthesized as reported (10). Triton X-100, Triton X-114, n-octylglucoside, Emulphogene (polyoxyethylene lo-tridecyl), dithiothreitol, and zymolase were obtained from Sigma Chemical Co. DEAE Bio-Gel A agarose and Bio-Beads SM 2 beads were purchased from Bio-Rad. Yeast strain, growth conditions, and microsome preparation. Saccharomyces cereuisiae (ATCC 12341) was cultured as previously reported (12). For microsome preparation, an earlier method (12) was
0003-9S61/92 53.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
CHARACTERIZATION
OF YEAST
slightly modified. The yeast homogenate, obtained by sonication of washed cells in 1 vol of 0.1 M Na/K phosphate buffer (pH 7) containing 0.5 mM DTT: was clarified by low-speed centrifugation (10,OOOgfor 15 min) and then centrifuged at 150,000g for 60 min. The resulting supernatant (Slw or “soluble fraction”) was stored at -8O’C. The pellet was resuspended in 0.1 M phosphate buffer (pH 7) and recentrifuged at 150,OOOgfor 60 min to give a washed “microsomal fraction.” This fraction, suspended in a minimum of buffer, was rapidly frozen at -70°C in a cooled acetone bath and then stored at -80°C. Both frozen fractions, &, and microsomes, retained squalene oxide cyclase activity for several months. No appreciable difference was observed between frozen fractions that were thawed and freshly prepared fractions. When not otherwise indicated, all squalene oxide cyclase preparations were obtained from cells harvested at the stage of the maximum production of the enzyme, that is in late logarithmic phase, and homogenized by ultrasound (12). In particular cases, yeast cells were homogenized either by two passages through an X-Press (BIOX X-Press, Jarfalla, Sweden) at 2000 kg/cm’ or by zymolase digestion followed by homogenization of spheroplasts in a Potter-Helvehjem homogenizer (18). Triton X-114 phase separation procedure. Commercial Triton X-114 was condensed three times in 0.1 M Na/K phosphate buffer (PH 7), following the procedure described in Ref. (17). The detergent-rich phase recovered after the third condensation had a concentration of 25% (w/ v) Triton X-114 and was used as a source of detergent for all phaseseparation experiments described here. Microsomes were mixed with condensed Triton X-114 in 0.1 M Na/K phosphate buffer (pH 7), so that the final concentrations of detergent and protein were 2 and 0.8%, respectively. The mixture (10 ml) was left at 0°C under gentle stirring for 1 h and then centrifuged at 100,OOOgfor 1 h to remove unsolubilized material. The supernatant was kept and used for the temperature-induced phase partitioning. Proteins from the soluble fraction of the yeast homogenate were prepared in the same manner except for the centrifugation at 100,OOOgwhich was omitted. Phase partition was carried out essentially as described by Bordier (17). Briefly, the Triton X-114 solution of protein (8 ml) was layered over a cushion of 6% (w/v) sucrose, 0.1 M Na/K phosphate buffer (pH 7), and 0.06% Triton X-114 (2 ml) in a 15-ml polystyrene tube. The tube was warmed to 30°C for 5 min and then centrifuged for 5 min at 9OOg. The upper phase was removed, mixed with 0.5% fresh Triton X-114 at O”C, and overlain again on the sucrose cushion. After incubation at 30°C for 5 min and centrifugation at 900g for 5 min, the upper layer was recovered as the aqueous phase and the oily pellet as the detergent-rich phase. The squalene oxide cyclase activity of the aqueous phase could be directly tested, as the detergent concentration of the phase (0.05-0.1%) did not inhibit the enzyme. For testing the detergent, phase, the sample had to be treated with Bio-Beads to lower detergent concentration to beneath the inhibitory level (0.5%). Removal of peripheral proteins by ipotonic treatment. Washed microsomes were diluted with 0.1 M Na/K phosphate buffer (pH 7) to a protein concentration of 10 mg/ml. They were then dialyzed against distilled water containing 0.5 mM DTT at 4’C for 24 to 48 h and then centrifuged at 150,OOOgfor 60 min, giving a supernatant fraction and a pellet termed, respectively, “peripheral proteins” and “integral proteins.” Peripheral proteins were restored to usual ionic strength conditions by dialysis against 0.1 M Na/K phosphate buffer, pH 7. Purification of yeast squukne oxide cyclase. All steps were carried out at 0 to 4°C. Microsomes (762 mg protein), suspended in 0.1 M Na/ K phosphate buffer (pH 7) containing 0.5 mM DTT (buffer A), were diluted to a protein concentration of 15 mg/ml with distilled water (+0.5 mM DTT). Diluted microsomes (50 ml) were dialyzed twice (each 24 h) against 1 liter of distilled water (f0.5 mM DTT). Dialyzed microsomes
3 Abbreviations used: DTT, dithiothreitol; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis.
SQUALENE
OXIDE
123
CYCLASE
were then centrifuged at 150,OOOgfor 1 h and the resulting supernatant was discarded. To the sedimented membranes (integral proteins, 442 mg protein), resuspended in 12 ml of buffer A at 4”C, a solution of octylglucoside in buffer A was added dropwise to a final concentration of 8.5 mg/ml and a detergent-to-protein ratio of 1:l (w/w). The solution was gently stirred for 30 min at 4°C and then centrifuged for 90 min at 150,OOOg.The solubilized proteins (94 mg) were dialyzed against 0.01 M citrate/phosphate buffer (pH 7) containing 0.5 mM DTT to remove octylglucoside. The sample was concentrated to 25 ml on an ultrafiltration cell (Amicon) and Emulphogene was then added to a final concentration of 1 mg/ml. A solubilized sample (25 ml, 88 mg) was applied to a DEAE Bio-Gel A column (1.5 X 11.5 cm) equilibrated with 0.01 M citrate/ phosphate buffer (pH 7) containing 0.5 mM DTT and 1 mg/ml Emulphogene (buffer B). After washing with 2 vol of buffer B, the column was developed with a linear gradient of 0.05 to 0.5 M NaCl in this buffer (200 ml) at a rate of 15 ml/h. Five-milliliter fractions were collected. Pooled fractions (25 ml, 25 mg protein) containing squalene oxide cyclase activity were dialyzed against 1 liter of buffer B and then applied to a second column of DEAE Bio-Gel A (1.5 X 11.5 cm) equilibrated with the same buffer. The column was washed with 4 vol of buffer B. The enzyme was then eluted with a linear gradient of 0 to 0.3 M NaCl in the same buffer (240 ml). Three-milliliter fractions were collected. Separation of proteins during the chromatographic processes was monitored by continuous recording at 280 nm of the eluted fractions. Fractions, combined on the basis of the Azso reading, were then tested for squalene oxide cyclase activity in the presence of 1 mg/ml Triton X-100, using 0.2 ml aliquots, according to the procedure described below. Purified fractions remained active for more than 2 months if stored at 0-4”C, after addition of 0.03% NaN,. Activity was, however, rapidly lost when samples were frozen. Squahe oxide cychse assay. The enzyme assay was almost the same as that reported earlier (12). The reaction mixture contained, in 0.5 ml of incubation buffer, (R, S)-[3-3H]squalene-2,3-oxide (100,000 dpm, 25 FM); Tween 80,0.01% (w/v); and 0.5 mM DTT. Protein concentration depended on the sample to be tested: generally, microsomes and the soluble fraction (S,,,) were tested at protein concentrations of 6 and 10
6
12
I8
24
growth
38
30
time
42
48
(h)
FIG. 1. Change of the ratio of “apparent soluble” to microsomal enzyme with the growth phase of the culture. Yeast cells, harvested at different growth phases, were homogenized in 1 vol of 0.1 M Na/K phosphate buffer (pH 7) and treated as described under Materials and Methods to give the soluble fraction and microsomes. Squalene oxide cyclase activity of both fractions was then tested under the same conditions. All values represent the means of duplicate cultures and the experiment was repeated with essentially identical results. Standard deviations were less than 10%. (- - -) Yeast cell concentration.
124
BALLIANO
IO
m
30
terrperoture
40
50
m
(T 1
ET AL.
4
6
6
7
8
a
PH
FIG. 2. Influence of temperature (A) and pH (B) on the sedimentable (microsomes) and nonsedimentable (Si,) form of squalene oxide cyclase. buffer. (0) Microsomes; Assay conditions are described under Materials and Methods. Buffer used in experiment B was 0.1 M citrate/phosphate (A) S150fraction.
mg/ml, respectively, whereas solubilized and purified fractions were tested at lower concentrations. The reaction mixture was incubated for 30 min at 35°C in a shaking water bath. Boiled samples were used as a blank. The nonsaponifiable extract was analyzed on silica gel plates (Merck) developed with CHxClz. Authentic samples of squalene oxide and lanosterol were cochromatographed as references. The identity of the [3H]lanosterol formed was established following a procedure previously reported (19, 20): the band comigrating with lanosterol, diluted with authentic sample, was acetylated and chromatographed on silvernitrate-impregnated silica gel plates by migration in cyclohexane:benzene, 70:30, on continuous TLC for 14 h or CHC& free from ethanol as developing solvents; plates were scanned in a Packard radiochromatogram reader and the only radioactive peak (190% of the starting radioactive band) was to be associated with lanosteryl acetate. Isotope counting and activity calculation were done as already described (12). Other methods. Dialysis was always done with Spectra/Par molecular porous membranes (Spectrum Medical Industries) with a molecular weight cutoff of 12,000 to 14,000 Da. The protein content of all samples (from microsomes to the purest fractions) was measured using the BioRad reagent, according to the supplier’s instructions, with -y-globulin as a standard. Octylglucoside was removed from samples by dialysis, whereas Triton X-100, Triton X-114, and Emulphogene were removed by batch chromatography on Bio-Beads SM 2 (21). The soluble fraction from rat liver was prepared according to a procedure described elsewhere (22).
RESULTS
Partition of squalene oxide cyclase activity between miThe supernatant fraccrosomes and the soluble fraction. tion (S15,, fraction) obtained after pelleting microsomes
at 150,OOOggenerally gave half the specific activity of the microsomal fraction. All the assayed disruption systems (sonication, X-Press, zymolase digestion) gave similar results. This apparent partition of the enzyme between microsomes and the soluble fraction led us to investigate the properties of the apparent soluble form (nonsedimentable form) of the enzyme. The supernatant obtained in a standard preparation of microsomes was subjected to repeated and prolonged centrifugations at 200,OOOg. This treatment did not remove enzyme activity from the soluble fraction: after 10 h centrifugation the starting level of specific activity was reduced by only 20%. Cell culture age seemed to influence distribution of the enzyme activity between pellet and supernatant: in cells harvested in the logarithmic phase, activity was mostly associated with microsomes, whereas in older cultures a greater proportion of activity was nonsedimentable (Fig. 1). Comparative properties of the microsomal and “apparent soluble” enzyme. The nonsedimentable form of the enzyme had the same optimum pH (7.4) and temperature (35°C) as the microsomal form (Fig. 2). The requirement of a cytoplasmic factor was evaluated by testing the activity of both forms in the presence of a soluble fraction obtained from rat liver homogenate (5 mg/ml protein) which, unlike the yeast soluble fraction, was thoroughly devoid of its own squalene oxide cyclase activity. The sol-
CHARACTERIZATION
OF YEAST
uble cytoplasm from rat liver had a stimulatory effect on either the nonsedimentable or the sedimentable fraction of the yeast homogenate (Fig. 3). The specific activity of both forms was doubled by Triton X-100, 1 mg/ml, which was the optimum detergent concentration. The activation of the nonsedimentable form occurred within a narrower range of detergent concentration. The dependence of the enzyme activity on the presence of detergent was more carefully evaluated through reactivation experiments. Sedimentable and nonsedimentable forms of the enzyme were separately treated at O’C for 1 h with Triton X-100 at a detergent-to-protein ratio of 1: 1 (w/w) and, after exhaustive removal of the solubilizing agent, their activity was tested at increasing amounts of the same detergent, added as a reactivating agent (Fig. 4). Solubilized enzyme displayed a strong dependence on the presence of detergent which, at a concentration of 0.5-l mg/ml, sharply reactivated the enzyme derived from both sources. Evidence on the similarity between the sedimentable and the nonsedimentable forms of the enzyme was also provided by the Triton X-114 phase-separation experiments. The Triton X-114 phase-separation method, reported to be able to separate soluble and membrane proteins into different classes (17, 23), was applied either to the soluble or to the microsomal fraction of the yeast homogenate. While proteins from different sources partitioned in a different way, after phase separation in Triton X-114 solutions (less than 10% of ,!& protein was recovered in the detergent phase, whereas microsomal
SQUALENE
OXIDE
125
CYCLASE
Triton
X-100
(XI
FIG. 4. Reactivation experiments with squalene oxide cyclase solubilized by Triton X-100. Microsomes (63 mg protein), suspended in 3 ml of 0.1 M Na/K phosphate buffer, pH 7.2, were solubilized for 30 min at 0°C with 2% Triton X-100. The solubilized fraction was treated with Bio-Beads to remove the solubilizing agent. The activity of treated samples (A) (0.5 mg/ml protein) was assayed at increasing amounts of Triton X-100 in 0.1 M K/Na phosphate buffer, pH 7. A similar procedure was followed for the nonsedimentable fraction of the homogenate starting from 8 mg/ml protein treated with 0.8% Triton X-100. The activity of these treated samples (0) was tested at a protein concentration of 5 mg/ml.
IO
06' 0 protein
5 of
enzyme
IO preparation
15 (m#ml)
FIG. 3. Influence of mammalian soluble cytoplasm on squalene oxide cyclase from Sacchuromyces cereuisiae. Microsomes and the soluble fraction prepared from yeast cells harvested at the late logarithmic phase were tested in the presence of the soluble fraction obtained from rat liver homogenate (5 mg/ml protein). (0) Yeast microsomes; (A) yeast soluble fraction; (-) with mammalian soluble fraction; ( * * * ) without mammalian soluble fraction.
protein partitioned between the aqueous and the detergent-rich phases at a ratio near 2:3), partition of squalene oxide cyclase obtained from both sources was almost the same (65-70% of enzyme activity partitioned in the detergent phase) (Fig. 5). Moreover, the activity recovered in the aqueous phase was strictly dependent on the residual detergent: its exhaustive removal caused a complete inactivation of the enzyme and the activity of the phase could be restored by adding a suitable amount of Triton x-100. Extraction of squalene oxide cychse from microsomes. The investigation on the interaction between squalene oxide cyclase and yeast membranes was further carried on by studying the conditions required to obtain a complete removal of the enzyme from microsomes. Conditions able to remove peripheral membrane proteins, such as short exposure to high or low pH or treatment with extremely high or low ionic strength (24-27), were first tested. Treatment with diluted acid (pH 2.5) or alkali (pH 10.2) resulted in a complete loss of enzyme activity.
126
BALLIANO
‘O°C A
B
C
0
E
F
FIG. 6. Phase separation of yeast squalene oxide cyclase from microsomes and the soluble fraction. Microsomes (30 mg protein) and the soluble fraction (70 mg protein) were separately treated with Triton X114 as described under Materials and Methods. After phase partition the volume of the resulting two phases was measured. Protein content and squalene oxide cyclase activity in the phases were detected. The concentration of Triton X-114 was measured photometrically at 275 nm in a control experiment. Values represent the means of two independent determinations, which did not differ by more than 10%. (A) Volume; (B) Triton X-114; (C) proteins from the soluble fraction; (D) proteins from microsomes; (E) squalene oxide cyclase from the soluble fraction; (F) squalene oxide cyclase from microsomes.
The most efficient and conservative method, among those based on the change of ionic strength, was dialysis of microsomes against distilled water. This treatment, described under Materials and Methods, caused a release of TABLE
ET AL.
45 to 50% of microsomal protein (peripheral proteins), whereas more than 85% of the enzyme activity was retained by the membrane fraction. Quantitative extraction of squalene oxide cyclase from microsomes could only be obtained by solubilization with detergents. Solubilization was monitored by treating microsomes with detergents at different concentrations for 60 min at 4°C under stirring, then pelleting the unsolubilized material, and finally measuring the squalene oxide cyclase activity in the supernatant. Since the activity of the solubilized fractions could be influenced by different concentrations of the detergent used in solubilization, all fractions were treated to remove the solubilizing agent, and then they were all tested with Triton X-100 at the same concentration (1 mg/ml). The results obtained with Triton X-100 as a solubilizing agent (Table I) showed no preferential solubilization: the specific activity of the solubilized fractions did not change over a range of detergentto-protein ratios from 0.1 to 1 (w/w). Solubilization with octylglucoside appeared to be more critical, as this detergent, at concentrations higher than 15-20 mg/ml, caused an irreversible inactivation, while at a detergent-to-protein ratio lower than 1 the solubilizing ability strongly decreased. Both Triton X-100 and octylglucoside most effectively extracted the enzyme at a detergent-to-protein ratio of about l:l. The starting protein concentration did not seem to influence the result, at least within the range of 8 to 15 mg protein/ml. Partial purification of squalene oxide cyclase. In our first attempt to purify squalene oxide cyclase from yeast we solubilized microsomes by Triton X-100 and then chromatographed the solubilized material on a DEAE BioGel A column. This procedure caused a 60-70% loss of I
Activity of SqualeneOxide CyclaseSolubilized from Intact Microsomesby Detergents Concentration of solubilizing agent bdml)
Detergent/protein (w/w)
Solubilized proteins (md
(nmol/h/mg)
Total activity (nmol/h)
Specific activity
Solubilization with Triton X-100 1 5 10 20
0.12 0.58 1.17 2.35
16.56 28 31 33.3
11.38 13.7 13.9 8.11
188.4 383.6 430.9 270
Solubilization with octylglucoside 5 10 20 30
0.58 1.17 2.35 3.53
13.44 23.4 22.8 23.16
7.13 15.64 0.62 0.00
95.8 367 14 0.00
Note. Microsomes (100 mg protein; specific activity, 1.55 nmol/h/mg protein) were solubilized with increasing amounts of Triton X-100 or octylglucoside as described in the text. Solubilized fractions, recovered after pelleting the insolubilized material, were freed from the solubilizing agent. The activity of all fractions was then tested in the presence of Triton X-100 (1 mg/ml) as a reactivating agent.
CHARACTERIZATION
OF YEAST TABLE
SQUALENE
OXIDE
127
CYCLASE
II
Partial Purification of Squalene Oxide Cyclase from Saccharonyces cereuisiue
Purification
Fraction
Protein (mg)
Microsomes Integral proteins Solubilized proteins Combined fractions Combined fractions
762 442 94 24 3.8
step
Hypotonic treatment Solubilization with octylglucoside First DEAE passage Second DEAE passage
Specific activity bmol/h/md 0.59 1.09 10 17.5 a4
Purification factor 1 1.85 17 29.7 142
Note. The enzyme was assayed as described under Materials and Methods. Pelleted fractions (microsomes and integral without reactivating agent. Solubilized fractions were tested with Triton X-100 at a final concentration of 1 mg/ml.
total activity and gave a preparation enriched by no more than 12- to 20-fold in comparison with microsomes. We therefore thought it better to enrich the material to be solubilized by treatment to remove the peripheral protein fraction (Table II, step 1). Solubilization of such enriched samples by detergent (octylglucoside) resulted in an overall 17-fold purification, with no loss of activity (Table II). Octylglucoside was preferred to Triton X-100 as a solubilizing agent because of its more specific ability to solubilize material depleted of extrinsic membrane proteins. After the solubilization step, octylglucoside was replaced by Emulphogene, a detergent which combines good reactivation ability with optical transparency in the uv region. Chromatography on DEAE Bio-Gel A in the presence of detergent further purified the enzyme (Fig. 6, Table II). The procedure described under Materials and Methods, i.e., two chromatographic passages performed at two different slopes of the NaCl gradient, yielded an overall 140fold purification with a very high recovery of total activity.
0.5r
;
1
-0.5 -0.4
;
-0.3
g z
-0.2 -0.1
o.oL
0L.-.-.L,&.' 0 50
loo Elude
VOIWIB
150
m
'0.0 250
(ml)
FIG. 6. Second chromatography of soiubilized squalene oxide cyclase on DEAE Bio-Gel A. Pooled active fractions eluted from the first chromatographic step (25 mg protein) were applied to a 20-ml DEAE BioGel A column and proteins were eluted. (-) AQao;( * . * ) squalene oxide cyclase activity; (- -) NaCl gradient.
Total activity (nmol/h) 450 482 940 425 319 protein)
were tested
SDS-PAGE of the material eluted from the DEAE column revealed the presence of different components in each of the chromatographic fractions. Unlike squalene oxide cyclase from hog and rat liver (8, 9), the yeast enzyme did not tolerate passage on a hydroxylapatite column: all attempts to purify the DEAE fractions using this technique resulted in an almost complete loss of the enzyme activity. Kinetic analysis of the purified enzyme activity gave an apparent KM (35 k 4 PM) comparable with the value obtained in the microsome preparation (12). DISCUSSION It has not hitherto been possible to classify squalene oxide cyclase from S. cerevisiae as a membrane-bound enzyme with certainty. Shechter et al. (16), in an attempt to characterize the yeast enzyme, concluded that squalene oxide cyclase may be a particle-bound enzyme, solubilized by some molecules with detergent properties released during the preparation of the homogenate. Our first attempt to classify squalene oxide cyclase from S. cerevisiae came up against considerable difficulties, due to the partition of the enzyme between the cytoplasmic and the membrane fractions of the homogenate. The soluble fraction (5&J retained much of the total activity, especially when cells were harvested in the stationary phase of growth. Moreover, the activity of the soluble fraction was no longer sedimentable, not even after prolonged centrifugation at 200,OOOg.Sedimentable and nonsedimentable forms of the enzyme, however, corresponded to one another not only in properties such as the response to pH and temperature, but also in properties which are peculiar to membrane-bound enzymes. Both forms showed, for example, the same response to Triton X-100 either when the enzyme was exposed to the detergent action just for the incubation time or when it was reactivated after the solubilization from the native form: in this case, the activity of the enzyme was extremely dependent on the presence of the detergent. The similarity of soluble and microsomal forms of yeast squalene oxide cyclase was also
128
BALLIANO
well demonstrated by the results obtained with the Triton X-114 phase partitioning, a method often applied to separate integral membrane proteins from soluble and peripheral proteins (17,23). Enzyme derived from both sedimentable and nonsedimentable fractions of the yeast homogenate was largely recovered, after phase separation in Triton X-114 solution, in the detergent-rich phase. Besides, as reported for some integral membrane glicoproteins (23), the activity recovered in the aqueous phase was totally dependent on the residual detergent present in the phase. If the enzyme present in the nonsedimentable fraction of the homogenate was a true soluble enzyme, its partition between the two Triton X-114 phases should overlap partition of the other proteins of the fraction, and its stability within the aqueous phase should be independent of the residual detergent. Noteworthy evidence for the similarity of the two forms of the enzyme was the sensitivity of both to mammalian cytosolic factors, which are known to stimulate other membrane-bound enzymes of sterol biosynthesis (28). The strong lipophilic character of the squalene oxide cyclase was further confirmed by the study of the conditions required to extract the enzyme from microsomes. None of the conditions described as able to remove extrinsic membrane proteins caused a significant dissociation of squalene oxide cyclase from microsomes. Indeed, to quantitatively solubilize the squalene oxide cyclase from microsomes, preparations had to be exposed to detergent concentrations equal to or higher than those reported to disrupt the membrane structure (29). The binding of the enzyme with membranes must therefore be considered rather tight, certainly tighter than the binding of another cyclizing enzyme, squalene hopene cyclase from Bacillus acidocaldarious, which dissociates at a detergent concentration which does not alter membrane integrity (30). All these results agree with the view that yeast squalene oxide cyclase is a membrane-bound enzyme belonging to the integral proteins. Recovery of part of the enzyme activity in the nonsedimentable fraction of the homogenate is probably not due to the existence of a true soluble form of the enzyme but rather to the formation of low-density microparticulate material during the homogenization step. The increase of the apparent soluble enzyme activity with the ageing of the yeast culture might depend on the release from lysing cells of some substance with detergent properties, which could facilitate the formation of microparticulate material. Recently, the occurrence of membraneassociated sterol carrier protein factors in yeast cells has been suggested (31). These factors, released during the homogenization step, could influence the apparent partitioning of the enzyme between membrane and soluble fractions. Taking the above conclusions into account, we tried to purify squalene oxide cyclase from yeast microsomes by
ET AL.
a procedure which guaranteed high recovery of the initial activity. In order to optimize the selectivity of the enzyme extraction from membranes, a two-phase procedure was designed, comprising (a) removal of the poorly active peripheral proteins and (b) solubilization of integral proteins with detergents. This substantially enriched the preparation (15- to 30-fold compared to microsomes) and seemed to protect the enzyme from partial inactivation caused by the chromatographic steps: the enzyme solubilized from intact microsomes (then in the presence of peripheral proteins) lost 65 to 75% of the initial activity after the first chromatographic passage on DEAE BioGel A, whereas, when solubilization was preceded by the removal of the peripheral protein fraction, the preparation eluted from the column maintained a higher level of total activity. Explanation of this fact shall be sought within the possible interaction between peripheral proteins and the squalene oxide cyclase. The most active sample recovered after purification (purification factor, 140) did not appear homogeneous in SDS-PAGE. Further investigations will be required to gain a more selective separation of yeast squalene oxide cyclase from other membrane proteins. The Triton X114 phase-separation method, which proved to be very useful in investigating the nature of this membrane-bound enzyme, might be a powerful tool for realizing the above purpose. REFERENCES 1. Dean, P. D. G. (1971) Steroiablogiu
2,143-157.
2. Schroepfer, G. J., Jr. (1982) Annu. Reu. Biochem. 51,555-585. 3. Corey, E. J., Ortiz de Montellano, P. R., Lin, K., and Dean, P. D. G. (1967) J. Am. Chem. Sot. 89,2797-2798.
4. van Tamelen, E. E., and James, D. R. (1977) J. Am. Chem. Sot. 99, 950-952.
5. Crosby, L. O., van Tamelen, E. E., and Clayton,
R. B. (1969) J.
Chem. Sot. Chem. Commun., 532-533. 6. van Tamelen, E. E., PedIai, A. D., Li, E., and James, D. R. (1977) J. Am. Chem. Sot. 99,6778-6780. 7. Johnson, W. S., Telfer, S. J., Cheng, S., and Schubert, U. (1987) J. Am. Chem. Sot. 109,2517-2518.
8. Duriatti, A., and Schuber, F. (1988) Biochem. Btiphys. Res. Commun. 151,1378-1385. 9. Kusano, M., Abe, I., Sankawa, Pharm. Bull. 39,239-241.
U., and Ebizuka,
Y. (1991) C&m.
10. Duriatti, A., Bouvier-Nave, P., Benveniste, P., Schuber, F., Delprino, L., Balliano, G., and Cattel, L. (1985) Biochem. Pharmacol. 34,
2765-2777. 11. Cattel, L., Ceruti, M., Viola, F., Delprino, L., Balliano, A., and Bouvier-Nave, P. (1986) Lipids 21, 31-38.
G. Duriatti,
12. Balliano, G., Viola, F., Ceruti, M., and Cattel, L. (1988) Biochim. Biophys. Acta 959,9-19. 13. Gerst, N., Schuber, F., Viola, F., and Cattel, Pharmacol. 35.4243-4250. 14. Viola, F., Ceruti, M., Balliano, I1 Farmaco 45,965-978.
L. (1986) Biochem.
G., Caputo, O., and Cattel, L. (1990)
CHARACTERIZATION
OF YEAST
15. Ceruti, M., BaIIiano, G., Viola, F., Cattel, L., Gerst, N., and Schuber, F. (1987) Eur. J. Med. Ckem. 22, 199-208. 16. Shechter, I., Sweat, F. W., and Block, K. (1970) B&him. Biophys. Acta 220,463-468. 17. Bordier, C. (1981) J. Biol. Chem. 256, 1604-1607. 18. Osumi, T., Nishino, T., and Katsuki, H. (1979) J. Biockem. 85, 819-826. 19. Cattel, L., Balliano, G., Caputo, O., and Viola, F. (1981) Planta Med. 41,326-336. 20. BalIiano, G., Caputo, O., Viola, F., Delprino, L., and Cattel, L. (1983) Pkytockemistry 22,909-913. 21. Holloway, P. W. (1973) Anal. Biochem. 53, 304-308. 22. Ceruti, M., Viola, F., Dosio, F., Cattel, L., Bouvier-Nave, P., and Ugliengo, P. (1988) J. Ckem. Sot. Perkin Trans. 1, 461-469. 23. Pryde, J. G. (1986) TIBS 11, 160-163.
SQUALENE
OXIDE
129
CYCLASE
24. Singer, S. J. (1974) Annu. Reu. Biockm. 25. Kuriyama,
3,805-833.
Y. (1972) J. Biol. Chem. 247,2979-2988.
26. Kramer, R. M., Wuthrich, C., Bollier, C., Allegrini, P. (1978) Biochim. Biophys. Acta 607,381-394.
P., and Zahler,
27. Capaldi, R. A., and Green, D. E. (1972) FEBS L.&t. 26.205-209. 28. Dempsey, M. E. (1985) in Methods in Enzymology (Law, J. H., and Rilling, H. C., Eds.), Vol. III, pp. 293-303, Academic Press, San Diego. 29. Prado, A., Arrondo, J. L. R., Villena, A., Goni, F. M., and Macarulla, J. M. (1983) B&him. Biophys. Acta 733, 163-171. 30. Seckler, B., and Poralla,
K. (1986) B&him.
Biophys. Acta 881,
356-363. 31. Bujons, J., Guajardo,
110,604-606.
R., and Kyler,
S. (1988) J. Am. Ckem. Sot.
OF BIOCHEMISTRY
Vol. 293, No. 1, February
AND BIOPHYSICS
14, pp. 122-129,1992
Characterization and Partial Purification of Squalene2,3-Oxide Cyclase from Saccharomyces cerev&jae’ Gianni Balliano,
Franca Viola, Maurizio
Ceruti, and Luigi Catte12
Istituto di Chimica Farmaceutica Applicata, Corso Raffaello 31, 10125 Torino, Italy
Received July 3, 1991, and in revised form September 27, 1991
The membrane nature of squalene oxide cyclase from Saccharomyces cerevisiae was investigated by comparing properties of the enzyme recovered from both microsomes and the soluble fraction of the yeast homogenate. The “apparent soluble” form and microsomal form of the enzyme were both stimulated by the presence of mammalian soluble cytoplasm and corresponded to one another in response to detergents Triton X- 100 and Triton X-114. The observed strong dependence of the enzyme activity on the presence of detergents and the behavior of the enzyme after Triton X-l 14 phase separation were peculiar to a lipophilic membrane-bound enzyme. A study of the conditions required to extract the enzyme from microsomes confirmed the lipophilic character of the enzyme. Microsomes, exposed to ipotonic conditions to remove peripheral membrane proteins, retained most of the enzyme activity within the integral protein fraction. Quantitative dissociation of the enzyme from membranes occurred only if microsomes were treated with detergents (Triton X-100 or octylglucoside) at concentrations which alter membrane integrity. The squalene oxide cyclase was purified 140 times from yeast microsomes by (a) removal of peripheral proteins, (b) extraction of the enzyme from the integral protein fraction with octylglucoside, and (c) separation of the solubilized proteins by DEAE Bio-Gel A chromatography. Removal of the peripheral proteins seemed to be a key step necessary for obtaining high yields. o 1992 Academic PWS, I~C.
The central role played by squalene-2,3-oxide cyclases (EC 5.4.99.7) in sterol biosynthesis has often been described (1, 2). These enzymes catalyze the cyclization of the substrate into different cyclic triterpenoids in higher 1 This work was supported by grants from Minister0 e della Ricerca Scientifica e Tecnologica-Italy. * To whom correspondence should be addressed.
122
dell’universita
plants, whereas in animals, yeast, and fungi the main cyclization product is lanosterol (1). The complex catalytic mechanism of these enzymes has attracted much research, mainly with mammalian squalene oxide cyclase, on the cyclization process and backbone rearrangement of squalene oxide (3-7). Recently, important steps, with the mammalian enzymes, were taken by two groups who solubilized and purified squalene oxide cyclase from hog and rat liver (8,9). As we were interested in comparing squalene oxide cyclase from liver and yeast, as well as testing the pharmacological activity of new classes of inhibitors of these enzymes (lo-15), we thought it would be of particular interest to solubilize, purify, and characterize the squalene oxide cyclase from Saccharomyces cerevisiae. In a previous comparison of mammalian and yeast squalene oxide cyclase (16), it was concluded that, while the microsomal nature of the mammalian enzyme was clearly established, the yeast enzyme appeared to be soluble. However, all attempts to purify the yeast enzyme failed, resulting in the nearly complete loss of activity. This appears to have been the only attempt to characterize and purify squalene oxide cyclase from yeast. In the present paper we describe the characterization of yeast squalene oxide cyclase as a membrane-bound enzyme and the partial purification from microsomes. The hydrophobic character of the enzyme was investigated by treatments with different detergents and particularly by using the Triton X-114 phase separation method (17). MATERIALS
AND
METHODS
Chemicals. Squalene-2,3-oxide and [3-3H]squalene-2,3-oxide (44 X 10s dpm/pmol) were synthesized as reported (10). Triton X-100, Triton X-114, n-octylglucoside, Emulphogene (polyoxyethylene lo-tridecyl), dithiothreitol, and zymolase were obtained from Sigma Chemical Co. DEAE Bio-Gel A agarose and Bio-Beads SM 2 beads were purchased from Bio-Rad. Yeast strain, growth conditions, and microsome preparation. Saccharomyces cereuisiae (ATCC 12341) was cultured as previously reported (12). For microsome preparation, an earlier method (12) was
0003-9S61/92 53.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
CHARACTERIZATION
OF YEAST
slightly modified. The yeast homogenate, obtained by sonication of washed cells in 1 vol of 0.1 M Na/K phosphate buffer (pH 7) containing 0.5 mM DTT: was clarified by low-speed centrifugation (10,OOOgfor 15 min) and then centrifuged at 150,000g for 60 min. The resulting supernatant (Slw or “soluble fraction”) was stored at -8O’C. The pellet was resuspended in 0.1 M phosphate buffer (pH 7) and recentrifuged at 150,OOOgfor 60 min to give a washed “microsomal fraction.” This fraction, suspended in a minimum of buffer, was rapidly frozen at -70°C in a cooled acetone bath and then stored at -80°C. Both frozen fractions, &, and microsomes, retained squalene oxide cyclase activity for several months. No appreciable difference was observed between frozen fractions that were thawed and freshly prepared fractions. When not otherwise indicated, all squalene oxide cyclase preparations were obtained from cells harvested at the stage of the maximum production of the enzyme, that is in late logarithmic phase, and homogenized by ultrasound (12). In particular cases, yeast cells were homogenized either by two passages through an X-Press (BIOX X-Press, Jarfalla, Sweden) at 2000 kg/cm’ or by zymolase digestion followed by homogenization of spheroplasts in a Potter-Helvehjem homogenizer (18). Triton X-114 phase separation procedure. Commercial Triton X-114 was condensed three times in 0.1 M Na/K phosphate buffer (PH 7), following the procedure described in Ref. (17). The detergent-rich phase recovered after the third condensation had a concentration of 25% (w/ v) Triton X-114 and was used as a source of detergent for all phaseseparation experiments described here. Microsomes were mixed with condensed Triton X-114 in 0.1 M Na/K phosphate buffer (pH 7), so that the final concentrations of detergent and protein were 2 and 0.8%, respectively. The mixture (10 ml) was left at 0°C under gentle stirring for 1 h and then centrifuged at 100,OOOgfor 1 h to remove unsolubilized material. The supernatant was kept and used for the temperature-induced phase partitioning. Proteins from the soluble fraction of the yeast homogenate were prepared in the same manner except for the centrifugation at 100,OOOgwhich was omitted. Phase partition was carried out essentially as described by Bordier (17). Briefly, the Triton X-114 solution of protein (8 ml) was layered over a cushion of 6% (w/v) sucrose, 0.1 M Na/K phosphate buffer (pH 7), and 0.06% Triton X-114 (2 ml) in a 15-ml polystyrene tube. The tube was warmed to 30°C for 5 min and then centrifuged for 5 min at 9OOg. The upper phase was removed, mixed with 0.5% fresh Triton X-114 at O”C, and overlain again on the sucrose cushion. After incubation at 30°C for 5 min and centrifugation at 900g for 5 min, the upper layer was recovered as the aqueous phase and the oily pellet as the detergent-rich phase. The squalene oxide cyclase activity of the aqueous phase could be directly tested, as the detergent concentration of the phase (0.05-0.1%) did not inhibit the enzyme. For testing the detergent, phase, the sample had to be treated with Bio-Beads to lower detergent concentration to beneath the inhibitory level (0.5%). Removal of peripheral proteins by ipotonic treatment. Washed microsomes were diluted with 0.1 M Na/K phosphate buffer (pH 7) to a protein concentration of 10 mg/ml. They were then dialyzed against distilled water containing 0.5 mM DTT at 4’C for 24 to 48 h and then centrifuged at 150,OOOgfor 60 min, giving a supernatant fraction and a pellet termed, respectively, “peripheral proteins” and “integral proteins.” Peripheral proteins were restored to usual ionic strength conditions by dialysis against 0.1 M Na/K phosphate buffer, pH 7. Purification of yeast squukne oxide cyclase. All steps were carried out at 0 to 4°C. Microsomes (762 mg protein), suspended in 0.1 M Na/ K phosphate buffer (pH 7) containing 0.5 mM DTT (buffer A), were diluted to a protein concentration of 15 mg/ml with distilled water (+0.5 mM DTT). Diluted microsomes (50 ml) were dialyzed twice (each 24 h) against 1 liter of distilled water (f0.5 mM DTT). Dialyzed microsomes
3 Abbreviations used: DTT, dithiothreitol; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis.
SQUALENE
OXIDE
123
CYCLASE
were then centrifuged at 150,OOOgfor 1 h and the resulting supernatant was discarded. To the sedimented membranes (integral proteins, 442 mg protein), resuspended in 12 ml of buffer A at 4”C, a solution of octylglucoside in buffer A was added dropwise to a final concentration of 8.5 mg/ml and a detergent-to-protein ratio of 1:l (w/w). The solution was gently stirred for 30 min at 4°C and then centrifuged for 90 min at 150,OOOg.The solubilized proteins (94 mg) were dialyzed against 0.01 M citrate/phosphate buffer (pH 7) containing 0.5 mM DTT to remove octylglucoside. The sample was concentrated to 25 ml on an ultrafiltration cell (Amicon) and Emulphogene was then added to a final concentration of 1 mg/ml. A solubilized sample (25 ml, 88 mg) was applied to a DEAE Bio-Gel A column (1.5 X 11.5 cm) equilibrated with 0.01 M citrate/ phosphate buffer (pH 7) containing 0.5 mM DTT and 1 mg/ml Emulphogene (buffer B). After washing with 2 vol of buffer B, the column was developed with a linear gradient of 0.05 to 0.5 M NaCl in this buffer (200 ml) at a rate of 15 ml/h. Five-milliliter fractions were collected. Pooled fractions (25 ml, 25 mg protein) containing squalene oxide cyclase activity were dialyzed against 1 liter of buffer B and then applied to a second column of DEAE Bio-Gel A (1.5 X 11.5 cm) equilibrated with the same buffer. The column was washed with 4 vol of buffer B. The enzyme was then eluted with a linear gradient of 0 to 0.3 M NaCl in the same buffer (240 ml). Three-milliliter fractions were collected. Separation of proteins during the chromatographic processes was monitored by continuous recording at 280 nm of the eluted fractions. Fractions, combined on the basis of the Azso reading, were then tested for squalene oxide cyclase activity in the presence of 1 mg/ml Triton X-100, using 0.2 ml aliquots, according to the procedure described below. Purified fractions remained active for more than 2 months if stored at 0-4”C, after addition of 0.03% NaN,. Activity was, however, rapidly lost when samples were frozen. Squahe oxide cychse assay. The enzyme assay was almost the same as that reported earlier (12). The reaction mixture contained, in 0.5 ml of incubation buffer, (R, S)-[3-3H]squalene-2,3-oxide (100,000 dpm, 25 FM); Tween 80,0.01% (w/v); and 0.5 mM DTT. Protein concentration depended on the sample to be tested: generally, microsomes and the soluble fraction (S,,,) were tested at protein concentrations of 6 and 10
6
12
I8
24
growth
38
30
time
42
48
(h)
FIG. 1. Change of the ratio of “apparent soluble” to microsomal enzyme with the growth phase of the culture. Yeast cells, harvested at different growth phases, were homogenized in 1 vol of 0.1 M Na/K phosphate buffer (pH 7) and treated as described under Materials and Methods to give the soluble fraction and microsomes. Squalene oxide cyclase activity of both fractions was then tested under the same conditions. All values represent the means of duplicate cultures and the experiment was repeated with essentially identical results. Standard deviations were less than 10%. (- - -) Yeast cell concentration.
124
BALLIANO
IO
m
30
terrperoture
40
50
m
(T 1
ET AL.
4
6
6
7
8
a
PH
FIG. 2. Influence of temperature (A) and pH (B) on the sedimentable (microsomes) and nonsedimentable (Si,) form of squalene oxide cyclase. buffer. (0) Microsomes; Assay conditions are described under Materials and Methods. Buffer used in experiment B was 0.1 M citrate/phosphate (A) S150fraction.
mg/ml, respectively, whereas solubilized and purified fractions were tested at lower concentrations. The reaction mixture was incubated for 30 min at 35°C in a shaking water bath. Boiled samples were used as a blank. The nonsaponifiable extract was analyzed on silica gel plates (Merck) developed with CHxClz. Authentic samples of squalene oxide and lanosterol were cochromatographed as references. The identity of the [3H]lanosterol formed was established following a procedure previously reported (19, 20): the band comigrating with lanosterol, diluted with authentic sample, was acetylated and chromatographed on silvernitrate-impregnated silica gel plates by migration in cyclohexane:benzene, 70:30, on continuous TLC for 14 h or CHC& free from ethanol as developing solvents; plates were scanned in a Packard radiochromatogram reader and the only radioactive peak (190% of the starting radioactive band) was to be associated with lanosteryl acetate. Isotope counting and activity calculation were done as already described (12). Other methods. Dialysis was always done with Spectra/Par molecular porous membranes (Spectrum Medical Industries) with a molecular weight cutoff of 12,000 to 14,000 Da. The protein content of all samples (from microsomes to the purest fractions) was measured using the BioRad reagent, according to the supplier’s instructions, with -y-globulin as a standard. Octylglucoside was removed from samples by dialysis, whereas Triton X-100, Triton X-114, and Emulphogene were removed by batch chromatography on Bio-Beads SM 2 (21). The soluble fraction from rat liver was prepared according to a procedure described elsewhere (22).
RESULTS
Partition of squalene oxide cyclase activity between miThe supernatant fraccrosomes and the soluble fraction. tion (S15,, fraction) obtained after pelleting microsomes
at 150,OOOggenerally gave half the specific activity of the microsomal fraction. All the assayed disruption systems (sonication, X-Press, zymolase digestion) gave similar results. This apparent partition of the enzyme between microsomes and the soluble fraction led us to investigate the properties of the apparent soluble form (nonsedimentable form) of the enzyme. The supernatant obtained in a standard preparation of microsomes was subjected to repeated and prolonged centrifugations at 200,OOOg. This treatment did not remove enzyme activity from the soluble fraction: after 10 h centrifugation the starting level of specific activity was reduced by only 20%. Cell culture age seemed to influence distribution of the enzyme activity between pellet and supernatant: in cells harvested in the logarithmic phase, activity was mostly associated with microsomes, whereas in older cultures a greater proportion of activity was nonsedimentable (Fig. 1). Comparative properties of the microsomal and “apparent soluble” enzyme. The nonsedimentable form of the enzyme had the same optimum pH (7.4) and temperature (35°C) as the microsomal form (Fig. 2). The requirement of a cytoplasmic factor was evaluated by testing the activity of both forms in the presence of a soluble fraction obtained from rat liver homogenate (5 mg/ml protein) which, unlike the yeast soluble fraction, was thoroughly devoid of its own squalene oxide cyclase activity. The sol-
CHARACTERIZATION
OF YEAST
uble cytoplasm from rat liver had a stimulatory effect on either the nonsedimentable or the sedimentable fraction of the yeast homogenate (Fig. 3). The specific activity of both forms was doubled by Triton X-100, 1 mg/ml, which was the optimum detergent concentration. The activation of the nonsedimentable form occurred within a narrower range of detergent concentration. The dependence of the enzyme activity on the presence of detergent was more carefully evaluated through reactivation experiments. Sedimentable and nonsedimentable forms of the enzyme were separately treated at O’C for 1 h with Triton X-100 at a detergent-to-protein ratio of 1: 1 (w/w) and, after exhaustive removal of the solubilizing agent, their activity was tested at increasing amounts of the same detergent, added as a reactivating agent (Fig. 4). Solubilized enzyme displayed a strong dependence on the presence of detergent which, at a concentration of 0.5-l mg/ml, sharply reactivated the enzyme derived from both sources. Evidence on the similarity between the sedimentable and the nonsedimentable forms of the enzyme was also provided by the Triton X-114 phase-separation experiments. The Triton X-114 phase-separation method, reported to be able to separate soluble and membrane proteins into different classes (17, 23), was applied either to the soluble or to the microsomal fraction of the yeast homogenate. While proteins from different sources partitioned in a different way, after phase separation in Triton X-114 solutions (less than 10% of ,!& protein was recovered in the detergent phase, whereas microsomal
SQUALENE
OXIDE
125
CYCLASE
Triton
X-100
(XI
FIG. 4. Reactivation experiments with squalene oxide cyclase solubilized by Triton X-100. Microsomes (63 mg protein), suspended in 3 ml of 0.1 M Na/K phosphate buffer, pH 7.2, were solubilized for 30 min at 0°C with 2% Triton X-100. The solubilized fraction was treated with Bio-Beads to remove the solubilizing agent. The activity of treated samples (A) (0.5 mg/ml protein) was assayed at increasing amounts of Triton X-100 in 0.1 M K/Na phosphate buffer, pH 7. A similar procedure was followed for the nonsedimentable fraction of the homogenate starting from 8 mg/ml protein treated with 0.8% Triton X-100. The activity of these treated samples (0) was tested at a protein concentration of 5 mg/ml.
IO
06' 0 protein
5 of
enzyme
IO preparation
15 (m#ml)
FIG. 3. Influence of mammalian soluble cytoplasm on squalene oxide cyclase from Sacchuromyces cereuisiae. Microsomes and the soluble fraction prepared from yeast cells harvested at the late logarithmic phase were tested in the presence of the soluble fraction obtained from rat liver homogenate (5 mg/ml protein). (0) Yeast microsomes; (A) yeast soluble fraction; (-) with mammalian soluble fraction; ( * * * ) without mammalian soluble fraction.
protein partitioned between the aqueous and the detergent-rich phases at a ratio near 2:3), partition of squalene oxide cyclase obtained from both sources was almost the same (65-70% of enzyme activity partitioned in the detergent phase) (Fig. 5). Moreover, the activity recovered in the aqueous phase was strictly dependent on the residual detergent: its exhaustive removal caused a complete inactivation of the enzyme and the activity of the phase could be restored by adding a suitable amount of Triton x-100. Extraction of squalene oxide cychse from microsomes. The investigation on the interaction between squalene oxide cyclase and yeast membranes was further carried on by studying the conditions required to obtain a complete removal of the enzyme from microsomes. Conditions able to remove peripheral membrane proteins, such as short exposure to high or low pH or treatment with extremely high or low ionic strength (24-27), were first tested. Treatment with diluted acid (pH 2.5) or alkali (pH 10.2) resulted in a complete loss of enzyme activity.
126
BALLIANO
‘O°C A
B
C
0
E
F
FIG. 6. Phase separation of yeast squalene oxide cyclase from microsomes and the soluble fraction. Microsomes (30 mg protein) and the soluble fraction (70 mg protein) were separately treated with Triton X114 as described under Materials and Methods. After phase partition the volume of the resulting two phases was measured. Protein content and squalene oxide cyclase activity in the phases were detected. The concentration of Triton X-114 was measured photometrically at 275 nm in a control experiment. Values represent the means of two independent determinations, which did not differ by more than 10%. (A) Volume; (B) Triton X-114; (C) proteins from the soluble fraction; (D) proteins from microsomes; (E) squalene oxide cyclase from the soluble fraction; (F) squalene oxide cyclase from microsomes.
The most efficient and conservative method, among those based on the change of ionic strength, was dialysis of microsomes against distilled water. This treatment, described under Materials and Methods, caused a release of TABLE
ET AL.
45 to 50% of microsomal protein (peripheral proteins), whereas more than 85% of the enzyme activity was retained by the membrane fraction. Quantitative extraction of squalene oxide cyclase from microsomes could only be obtained by solubilization with detergents. Solubilization was monitored by treating microsomes with detergents at different concentrations for 60 min at 4°C under stirring, then pelleting the unsolubilized material, and finally measuring the squalene oxide cyclase activity in the supernatant. Since the activity of the solubilized fractions could be influenced by different concentrations of the detergent used in solubilization, all fractions were treated to remove the solubilizing agent, and then they were all tested with Triton X-100 at the same concentration (1 mg/ml). The results obtained with Triton X-100 as a solubilizing agent (Table I) showed no preferential solubilization: the specific activity of the solubilized fractions did not change over a range of detergentto-protein ratios from 0.1 to 1 (w/w). Solubilization with octylglucoside appeared to be more critical, as this detergent, at concentrations higher than 15-20 mg/ml, caused an irreversible inactivation, while at a detergent-to-protein ratio lower than 1 the solubilizing ability strongly decreased. Both Triton X-100 and octylglucoside most effectively extracted the enzyme at a detergent-to-protein ratio of about l:l. The starting protein concentration did not seem to influence the result, at least within the range of 8 to 15 mg protein/ml. Partial purification of squalene oxide cyclase. In our first attempt to purify squalene oxide cyclase from yeast we solubilized microsomes by Triton X-100 and then chromatographed the solubilized material on a DEAE BioGel A column. This procedure caused a 60-70% loss of I
Activity of SqualeneOxide CyclaseSolubilized from Intact Microsomesby Detergents Concentration of solubilizing agent bdml)
Detergent/protein (w/w)
Solubilized proteins (md
(nmol/h/mg)
Total activity (nmol/h)
Specific activity
Solubilization with Triton X-100 1 5 10 20
0.12 0.58 1.17 2.35
16.56 28 31 33.3
11.38 13.7 13.9 8.11
188.4 383.6 430.9 270
Solubilization with octylglucoside 5 10 20 30
0.58 1.17 2.35 3.53
13.44 23.4 22.8 23.16
7.13 15.64 0.62 0.00
95.8 367 14 0.00
Note. Microsomes (100 mg protein; specific activity, 1.55 nmol/h/mg protein) were solubilized with increasing amounts of Triton X-100 or octylglucoside as described in the text. Solubilized fractions, recovered after pelleting the insolubilized material, were freed from the solubilizing agent. The activity of all fractions was then tested in the presence of Triton X-100 (1 mg/ml) as a reactivating agent.
CHARACTERIZATION
OF YEAST TABLE
SQUALENE
OXIDE
127
CYCLASE
II
Partial Purification of Squalene Oxide Cyclase from Saccharonyces cereuisiue
Purification
Fraction
Protein (mg)
Microsomes Integral proteins Solubilized proteins Combined fractions Combined fractions
762 442 94 24 3.8
step
Hypotonic treatment Solubilization with octylglucoside First DEAE passage Second DEAE passage
Specific activity bmol/h/md 0.59 1.09 10 17.5 a4
Purification factor 1 1.85 17 29.7 142
Note. The enzyme was assayed as described under Materials and Methods. Pelleted fractions (microsomes and integral without reactivating agent. Solubilized fractions were tested with Triton X-100 at a final concentration of 1 mg/ml.
total activity and gave a preparation enriched by no more than 12- to 20-fold in comparison with microsomes. We therefore thought it better to enrich the material to be solubilized by treatment to remove the peripheral protein fraction (Table II, step 1). Solubilization of such enriched samples by detergent (octylglucoside) resulted in an overall 17-fold purification, with no loss of activity (Table II). Octylglucoside was preferred to Triton X-100 as a solubilizing agent because of its more specific ability to solubilize material depleted of extrinsic membrane proteins. After the solubilization step, octylglucoside was replaced by Emulphogene, a detergent which combines good reactivation ability with optical transparency in the uv region. Chromatography on DEAE Bio-Gel A in the presence of detergent further purified the enzyme (Fig. 6, Table II). The procedure described under Materials and Methods, i.e., two chromatographic passages performed at two different slopes of the NaCl gradient, yielded an overall 140fold purification with a very high recovery of total activity.
0.5r
;
1
-0.5 -0.4
;
-0.3
g z
-0.2 -0.1
o.oL
0L.-.-.L,&.' 0 50
loo Elude
VOIWIB
150
m
'0.0 250
(ml)
FIG. 6. Second chromatography of soiubilized squalene oxide cyclase on DEAE Bio-Gel A. Pooled active fractions eluted from the first chromatographic step (25 mg protein) were applied to a 20-ml DEAE BioGel A column and proteins were eluted. (-) AQao;( * . * ) squalene oxide cyclase activity; (- -) NaCl gradient.
Total activity (nmol/h) 450 482 940 425 319 protein)
were tested
SDS-PAGE of the material eluted from the DEAE column revealed the presence of different components in each of the chromatographic fractions. Unlike squalene oxide cyclase from hog and rat liver (8, 9), the yeast enzyme did not tolerate passage on a hydroxylapatite column: all attempts to purify the DEAE fractions using this technique resulted in an almost complete loss of the enzyme activity. Kinetic analysis of the purified enzyme activity gave an apparent KM (35 k 4 PM) comparable with the value obtained in the microsome preparation (12). DISCUSSION It has not hitherto been possible to classify squalene oxide cyclase from S. cerevisiae as a membrane-bound enzyme with certainty. Shechter et al. (16), in an attempt to characterize the yeast enzyme, concluded that squalene oxide cyclase may be a particle-bound enzyme, solubilized by some molecules with detergent properties released during the preparation of the homogenate. Our first attempt to classify squalene oxide cyclase from S. cerevisiae came up against considerable difficulties, due to the partition of the enzyme between the cytoplasmic and the membrane fractions of the homogenate. The soluble fraction (5&J retained much of the total activity, especially when cells were harvested in the stationary phase of growth. Moreover, the activity of the soluble fraction was no longer sedimentable, not even after prolonged centrifugation at 200,OOOg.Sedimentable and nonsedimentable forms of the enzyme, however, corresponded to one another not only in properties such as the response to pH and temperature, but also in properties which are peculiar to membrane-bound enzymes. Both forms showed, for example, the same response to Triton X-100 either when the enzyme was exposed to the detergent action just for the incubation time or when it was reactivated after the solubilization from the native form: in this case, the activity of the enzyme was extremely dependent on the presence of the detergent. The similarity of soluble and microsomal forms of yeast squalene oxide cyclase was also
128
BALLIANO
well demonstrated by the results obtained with the Triton X-114 phase partitioning, a method often applied to separate integral membrane proteins from soluble and peripheral proteins (17,23). Enzyme derived from both sedimentable and nonsedimentable fractions of the yeast homogenate was largely recovered, after phase separation in Triton X-114 solution, in the detergent-rich phase. Besides, as reported for some integral membrane glicoproteins (23), the activity recovered in the aqueous phase was totally dependent on the residual detergent present in the phase. If the enzyme present in the nonsedimentable fraction of the homogenate was a true soluble enzyme, its partition between the two Triton X-114 phases should overlap partition of the other proteins of the fraction, and its stability within the aqueous phase should be independent of the residual detergent. Noteworthy evidence for the similarity of the two forms of the enzyme was the sensitivity of both to mammalian cytosolic factors, which are known to stimulate other membrane-bound enzymes of sterol biosynthesis (28). The strong lipophilic character of the squalene oxide cyclase was further confirmed by the study of the conditions required to extract the enzyme from microsomes. None of the conditions described as able to remove extrinsic membrane proteins caused a significant dissociation of squalene oxide cyclase from microsomes. Indeed, to quantitatively solubilize the squalene oxide cyclase from microsomes, preparations had to be exposed to detergent concentrations equal to or higher than those reported to disrupt the membrane structure (29). The binding of the enzyme with membranes must therefore be considered rather tight, certainly tighter than the binding of another cyclizing enzyme, squalene hopene cyclase from Bacillus acidocaldarious, which dissociates at a detergent concentration which does not alter membrane integrity (30). All these results agree with the view that yeast squalene oxide cyclase is a membrane-bound enzyme belonging to the integral proteins. Recovery of part of the enzyme activity in the nonsedimentable fraction of the homogenate is probably not due to the existence of a true soluble form of the enzyme but rather to the formation of low-density microparticulate material during the homogenization step. The increase of the apparent soluble enzyme activity with the ageing of the yeast culture might depend on the release from lysing cells of some substance with detergent properties, which could facilitate the formation of microparticulate material. Recently, the occurrence of membraneassociated sterol carrier protein factors in yeast cells has been suggested (31). These factors, released during the homogenization step, could influence the apparent partitioning of the enzyme between membrane and soluble fractions. Taking the above conclusions into account, we tried to purify squalene oxide cyclase from yeast microsomes by
ET AL.
a procedure which guaranteed high recovery of the initial activity. In order to optimize the selectivity of the enzyme extraction from membranes, a two-phase procedure was designed, comprising (a) removal of the poorly active peripheral proteins and (b) solubilization of integral proteins with detergents. This substantially enriched the preparation (15- to 30-fold compared to microsomes) and seemed to protect the enzyme from partial inactivation caused by the chromatographic steps: the enzyme solubilized from intact microsomes (then in the presence of peripheral proteins) lost 65 to 75% of the initial activity after the first chromatographic passage on DEAE BioGel A, whereas, when solubilization was preceded by the removal of the peripheral protein fraction, the preparation eluted from the column maintained a higher level of total activity. Explanation of this fact shall be sought within the possible interaction between peripheral proteins and the squalene oxide cyclase. The most active sample recovered after purification (purification factor, 140) did not appear homogeneous in SDS-PAGE. Further investigations will be required to gain a more selective separation of yeast squalene oxide cyclase from other membrane proteins. The Triton X114 phase-separation method, which proved to be very useful in investigating the nature of this membrane-bound enzyme, might be a powerful tool for realizing the above purpose. REFERENCES 1. Dean, P. D. G. (1971) Steroiablogiu
2,143-157.
2. Schroepfer, G. J., Jr. (1982) Annu. Reu. Biochem. 51,555-585. 3. Corey, E. J., Ortiz de Montellano, P. R., Lin, K., and Dean, P. D. G. (1967) J. Am. Chem. Sot. 89,2797-2798.
4. van Tamelen, E. E., and James, D. R. (1977) J. Am. Chem. Sot. 99, 950-952.
5. Crosby, L. O., van Tamelen, E. E., and Clayton,
R. B. (1969) J.
Chem. Sot. Chem. Commun., 532-533. 6. van Tamelen, E. E., PedIai, A. D., Li, E., and James, D. R. (1977) J. Am. Chem. Sot. 99,6778-6780. 7. Johnson, W. S., Telfer, S. J., Cheng, S., and Schubert, U. (1987) J. Am. Chem. Sot. 109,2517-2518.
8. Duriatti, A., and Schuber, F. (1988) Biochem. Btiphys. Res. Commun. 151,1378-1385. 9. Kusano, M., Abe, I., Sankawa, Pharm. Bull. 39,239-241.
U., and Ebizuka,
Y. (1991) C&m.
10. Duriatti, A., Bouvier-Nave, P., Benveniste, P., Schuber, F., Delprino, L., Balliano, G., and Cattel, L. (1985) Biochem. Pharmacol. 34,
2765-2777. 11. Cattel, L., Ceruti, M., Viola, F., Delprino, L., Balliano, A., and Bouvier-Nave, P. (1986) Lipids 21, 31-38.
G. Duriatti,
12. Balliano, G., Viola, F., Ceruti, M., and Cattel, L. (1988) Biochim. Biophys. Acta 959,9-19. 13. Gerst, N., Schuber, F., Viola, F., and Cattel, Pharmacol. 35.4243-4250. 14. Viola, F., Ceruti, M., Balliano, I1 Farmaco 45,965-978.
L. (1986) Biochem.
G., Caputo, O., and Cattel, L. (1990)
CHARACTERIZATION
OF YEAST
15. Ceruti, M., BaIIiano, G., Viola, F., Cattel, L., Gerst, N., and Schuber, F. (1987) Eur. J. Med. Ckem. 22, 199-208. 16. Shechter, I., Sweat, F. W., and Block, K. (1970) B&him. Biophys. Acta 220,463-468. 17. Bordier, C. (1981) J. Biol. Chem. 256, 1604-1607. 18. Osumi, T., Nishino, T., and Katsuki, H. (1979) J. Biockem. 85, 819-826. 19. Cattel, L., Balliano, G., Caputo, O., and Viola, F. (1981) Planta Med. 41,326-336. 20. BalIiano, G., Caputo, O., Viola, F., Delprino, L., and Cattel, L. (1983) Pkytockemistry 22,909-913. 21. Holloway, P. W. (1973) Anal. Biochem. 53, 304-308. 22. Ceruti, M., Viola, F., Dosio, F., Cattel, L., Bouvier-Nave, P., and Ugliengo, P. (1988) J. Ckem. Sot. Perkin Trans. 1, 461-469. 23. Pryde, J. G. (1986) TIBS 11, 160-163.
SQUALENE
OXIDE
129
CYCLASE
24. Singer, S. J. (1974) Annu. Reu. Biockm. 25. Kuriyama,
3,805-833.
Y. (1972) J. Biol. Chem. 247,2979-2988.
26. Kramer, R. M., Wuthrich, C., Bollier, C., Allegrini, P. (1978) Biochim. Biophys. Acta 607,381-394.
P., and Zahler,
27. Capaldi, R. A., and Green, D. E. (1972) FEBS L.&t. 26.205-209. 28. Dempsey, M. E. (1985) in Methods in Enzymology (Law, J. H., and Rilling, H. C., Eds.), Vol. III, pp. 293-303, Academic Press, San Diego. 29. Prado, A., Arrondo, J. L. R., Villena, A., Goni, F. M., and Macarulla, J. M. (1983) B&him. Biophys. Acta 733, 163-171. 30. Seckler, B., and Poralla,
K. (1986) B&him.
Biophys. Acta 881,
356-363. 31. Bujons, J., Guajardo,
110,604-606.
R., and Kyler,
S. (1988) J. Am. Ckem. Sot.
