1276
Biochemistry 1987, 26, 1276-1 283
Detergents as Probes of Reconstituted Rat Liver Cytochrome P-450 Function? Laurence S . Kaminsky,*it Deborah Dunbar,' F. Peter Guengerich,s and Jung J a Lee' Wadsworth Center for Laboratories and Research, New York State Department of Health, Albany, New York 12201, and Department of Biochemistry and Center in Molecular Toxicology, Vanderbilt University School of Medicine, Nashville, Tennessee 37232 Received September 26, 1986
ABSTRACT: A series of 16 ionic, zwitterionic, and nonionic detergents have been used to perturb the catalytic activities of major cytochrome P-450 (P-450) forms from untreated (UT-A), phenobarbital-treated (PB-B), and P-naphthoflavone-treated (BNF-B) rats in reconstituted systems with NADPH-P-450 reductase. Detergent effects on R warfarin hydroxylase activities were correlated with detergent effects on the quaternary structures of P-450 and reductase, and on their 1:1 complexes as determined by gel exclusion chromatography using sodium cholate as a prototype detergent. The detergent concentrations used did not in most cases affect rates of NADPH-dependent reduction of cytochrome c by the reductase. With P-450 BNF-B, ionic and zwitterionic detergents enhanced warfarin hydroxylase activities a t low concentrations and produced marked inhibition a t higher concentrations, while nonionic detergents only inhibited. With P-450 UT-A, some nonionic and zwitterionic detergents increased rates a t low concentrations and inhibited a t higher concentrations. P-450 PB-B was inhibited by detergents of all three classes at low and high Concentrations. The concentrations of a detergent required to affect 50% inhibition differed for the three P-450s, suggesting, together with the differential susceptibilities to detergent-mediated rate enhancing effects, that the reductase interacts functionally differently with the three P-450s. Chromatographic studies demonstrated that concentrations of sodium cholate which optimally enhanced metabolic rates with P-450 BNF-B facilitated the uptake of the P-450 into the functional reductase/P-450 complex, and higher concentrations of cholate, which completely inhibited activity, produced profound disruptions of the complex. The data have provided insight into the functional interactions required for monooxygenase activity.
x e first resolution of the hepatic microsomal monooxygenase system into solubilized fractions comprising rat liver cytochrome P-450 (P-450),I NADPH-P-450 reductase, and phospholipid provided a major impetus for the elucidation of the mechanism of action of this system (Lu & Coon, 1968). These components were reconstituted into systems capable of catalyzing the metabolism of a wide variety of substrates (Lu & Levin, 1974; Coon, 1978). We subsequently demonstrated that such reconstituted systems maintain the metabolic regioand stereoselectivity of hepatic microsomes, at least with R and S warfarin as substrates (Kaminsky et al., 1983). The role of the phospholipid in maintaining optimal activities of reconstituted systems has not been fully resolved, but it can be substituted for by preincubating the P-450 isozyme and reductase at high concentrations (Muller-Enoch et al., 1984). It was consequently concluded that the major effect of phospholipids is to facilitate the formation of an active P450/NADPH-P-450 reductase complex in the reconstituted system. Nonionic detergents such as Emulgen 9 11, Triton N- 101, and Triton X-100 have been shown to substitute for the phospholipid in reconstituted systems of partially purified P-450 from PB-treated rats catalyzing benzphetamine N-demethylation (Lu et al., 1974). It was also observed that these detergents only functioned to increase activities when they were present at low concentrations; at higher concentrations, they strongly inhibited reconstituted system metabolism. The nonionic detergent octyl P-D-glucopyranoside exhibited a 'This work was supported in part by US. Public Health Service Grants ES 03516 (L.S.K.) and ES 01590 and ES 00267 (F.P.G.). *New York State Department of Health. Vanderbilt University School of Medicine.
0006-2960/87/0426-1276$01.50/0
similar profile of enhancement and inhibition of benzphetamine N-demethylase activity with reconstituted P-45oLM2 from PB-treated rabbits, as the detergent concentration was increased (Dean & Gray, 1982). From sedimentation equilibrium experiments, it was concluded that the inhibitory effects of the detergent coincided with a disaggregation of the P-450, although the ultracentrifugation conditions differed from those in the reconstituted metabolic studies. These detergent-mediated effects were not confined to nonionic detergents since the zwitterionic detergent CHAPS behaved similarly with cyclohexane hydroxylase activity in a rabbit P-450LM2 reconstituted system (Wagner et al., 1984). These studies were interpreted to mean that the CHAPS blocked the functional interaction between reductase and the P-450, possibly by altering the state of aggregation of the P-450. The diversity of available detergents and their effects on the monooxygenase system (Denk, 1982) provide an approach for probing the functional interactions between the components of this system. We previously developed an approach for investigating P-450 function using warfarin metabolism, which complements the use of detergent probes (Kaminsky et al., 1980, 1981, 1983, 1985; Guengerich et al., 1982). The sodium salt of R warfarin, which has been used throughout our studies,
'
Abbreviations: PB, phenobarbital; BNF, (3-naphthoflavone; P-450, rat liver cytochrome P-450; CHAPS, 3- [(3-~holamidopropyl)dimethyIammoniol-I-propanesulfonate; CHAPSO, 3-[(3-cholamidopropyI)dimethylammonio] -2-hydroxy-1-propanesulfonate; HPLC, high-performance liquid chromatography; FPLC, fast protein liquid chromatography; P-450 UT-A, P-450 PB-B, and P-450 BNF-B, three forms of rat liver P-450, the properties of which have been reported elsewhere (Guengerich et al., 1982) [for comparison to preparations made in other laboratories, see Guengerich (1987) and Waxman (1986)l; EDTA, ethylenediaminetetraacetic acid; Tris-HCI, tris(hydroxymethy1)aminomethane hydrochloride; kDa, kilodalton(s); BSA, bovine serum albumin.
0 1987 American Chemical Society
D E T E R G E N T E F F E C T S O N C Y T O C H R O M E P-450 F U N C T I O N
is freely soluble in water and is thus less likely than many other insoluble or slightly soluble P-450 substrates to be sequestered into detergent micelles, thus becoming unavailable for metabolism. Furthermore, warfarin is a substrate for many forms of P-450 (Guengerich et al., 1982) and, in particular, yields different metabolite profiles with the three major forms from untreated and PB- or BNF-treated rats. Thus, a single substrate can be used to investigate a number of P-450s without introducing the potentially confounding factor of multiple substrates. In the present investigation, we have utilized a total of 16 detergents to probe the function of P-450s UT-A, PB-B, and BNF-B (the major hepatic isozymes present in untreated, PB-treated, and BNF-treated rats, respectively) using R warfarin as a substrate. The influence of the detergents on the metabolism of R warfarin in reconstituted systems of the three isozymes has been compared to assess the generality of detergent-mediated effects and to evaluate the effect of detergent type (ionic, nonionic, or zwitterionic). From these studies, sodium cholate was selected as a prototype to investigate the mechanism of the detergent-mediated effects by investigating its effects on the aggregation states of P-450 BNF-B, the NADPH-P-450 reductase, and a 1:1 mixture of these proteins. EXPERIMENTAL PROCEDURES Materials. P-450 forms UT-A, PB-B, and BNF-B were purified to electrophoretic homogeneity from liver microsomes of PB- and BNF-treated male Sprague-Dawley rats as previously described in detail (Guengerich et al., 1982). The specific contents of the preparations used here were 12.8, 15.2, and 14.0 nmol of P-450 (mg of protein)-’, respectively. Detergent was removed from the P-450 preparations by sequential adsorption to a column of hydroxylapatite, removal of detergent (monitored at 280 nm) with 10 mM potassium phosphate buffer (pH 7.4) containing 0.05 mM EDTA and 20% glycerol, and elution with the same buffer in which the phosphate concentration had been raised to 0.5 M. Limits for residual levels of detergents and phospholipids have been reported elsewhere (Muller-Enoch et al., 1984). NADPH-P-450 reductase was purified to electrophoretic homogeneity from hepatic microsomes isolated from PB-induced rats by octylamino-Sepharose 4B (Imai, 1976) and 2’,5’-ADP-agarose chromatography (Yasukochi & Masters, 1976) as subsequently modified (Guengerich et al., 1982). Cytochrome c was type I11 from horse heart and was purchased from Sigma (St. Louis, MO). Racemic warfarin was resolved by the method of West et al. (1961) into optically pure R and S warfarin. The warfarin metabolites 6-, 8-, and 4‘-hydroxywarfarin were synthesized by modifications (Fasco et al., 1977) of previously reported syntheses (Buckle et al., 1975; Pohl et al., 1975; Hermodson et al., 1971). The metabolites were characterized by their UV spectra and melting points. All detergents were purchased in the purest available grades and used without further treatment. Tween 80, Brij 35, Nonidet P-40, sodium taurochenodeoxycholate, sodium deoxycholate, cetylpyridinium chloride, and Lubrol PX were purchased from Sigma (St Louis, MO), sodium cholate, octyl @-D-glucopyranoside,Zwittergent, CHAPS, and CHAPSO were from CalBiochem (La Jolla, CA), Triton X-100 was from Aldrich (Milwaukee, WI), sodium dodecyl sulfate was from Bio-Rad (Richmond, CA), and Emulgens 91 1 and 913 were from Kao-Atlas (Tokyo, Japan). Methods. P-450 concentrations were determined from reduced C O vs. reduced difference spectra (Omura & Sato,
VOL. 26, NO. 5 , 1987
1277
1964). Protein concentrations were estimated by using the method of Bradford (1976) and Lowry et al. (1951). The concentration of NADPH-P-450 reductase solutions was determined spectrophotometrically at 455 nm using an extinction coefficient of 21.2 mM-’ cm-’ (Vermilion & Coon, 1974). Cytochrome c was reduced with sodium dithionite, and the excess dithionite was removed by passage through a column of Sephadex G-25 (Yonetani & Ray, 1965). The rates of detergent-induced autoxidation of cytochrome c were determined by monitoring the decrease in the 550-nm absorbance band (Kaminsky et al., 1971). Effects of detergents on NADPH-P-450 reductase activity were monitored by their effects on the rate of reduction of cytochrome c determined as previously reported (Phillips & Langdon, 1962). Warfarin Hydroxylation Assays. The effects of detergents on the activity of the P-450s were determined in reconstituted systems of each P-4S0 and NADPH-P-450 reductase as previously described in detail (Kaminsky et al., 1981). In these studies, the phospholipid was excluded and replaced by varying concentrations of detergent. Detergent was added to the reaction mixture after the reductase (1.0 pM) was added to P-450 (1 .O pM) but before the addition of the substrate, R warfarin. The reaction mixture was incubated at 37 OC for 1 min prior to initiation of the reaction with NADPH. Control mixtures were made up in the same manner but without detergent. In one study, the system was reconstituted by adding the P-450 BNF-B to the reductase slowly while mixing thoroughly with a vortex mixer. Reaction mixtures were analyzed by HPLC for the rates of warfarin metabolite formation (Kaminsky et al., 1981). All of the metabolic data presented are the average of two determinations, which varied by less than 10%. Gel Filtration Studies. Gel filtration studies of P-450 BNF-B, the reductase, and a mixture of the reductase and the P-450 were performed with a Pharmacia Superose 6 column at room temperature. A Pharmacia FPLC system, which included gradient capability, was used for the chromatography. Peaks were detected at 214 nm with a Waters Model 481 variable-wavelength absorbance detector. Areas under the peaks were determined with a Spectra Physics S P 4270 integrator. For studies of steady-state aggregation equilibria of the monooxygenase protein components, solutions were prepared identically with those for the metabolic incubations. Mixtures were incubated for 30 min at room temperature and were then clarified by centrifugation at 5g for 5 min. The injection volume was 100 pL (at protein concentrationsof 1 pM for each protein) onto the Superose 6 column. The flow rate of the eluting buffer [0.05 M Tris-HC1 (pH 7.4), containing 0.05 M NaCl] was 0.3 mL m i d . Sodium cholate was added to the buffer to final concentrationsof 0.32 and 3.0 g L-I. Studies were run in sequence from low to high cholate concentration, and changes were achieved by using a gradient of increasing sodium cholate concentrations. The column was equilibrated for 24 h at each cholate concentration prior to use. Multiple chromatographic runs were conducted with each sample. The system was calibrated for linearity of response to protein concentration with bovine serum albumin. Gel filtration standards (Bio-Rad, Richmond, CA) were used for molecular weight calibrations, which were performed at the various sodium cholate concentrations. RESULTS Effects of Detergents on Cytochrome c. At concentrations of 8 g L-I, the detergents Brij 35, Emulgens 911 and 913, Tween 80, Nonidet P-40, Triton X-100, CHAPS, CHAPSO,
1278
K A M I N S K Y ET A L .
BIOCHEMISTRY
Table I: Effect of Detergents on Rates of NADPH-P-450 Reductase Catalyzed Reduction of Cytochrome c detergent
Tween 80 octyl 0-D-glucopyranoside Emulgen 91 1 Emulgen 9 13 Brij 35 Triton X-100 Nonidet P-40 cetylpyridinium chloride sodium deoxycholate
sodium cholate sodium taurochenodeoxycholate
M, 1310 294
concn (g L-’)
8.0 8.0 8.0 8.0
8.0
1200 628 602 340 415
8.0 2.0 0.01
431 522
0.2
0.5
1.o 8.0 2.0
sodium dodecyl sulfate
288
Zwittergent
364
CHAPS0 CHAPS Lubrol PX
658 651 604
0.01 0.05 0.05 0.1 0.5 8.0
8.0 8.0
% of control rate 1I9 96 120 118 113
LUBROL PX INonionic)
109 108
0 86 64 93 108 83 65 19 16 46 39 92 90 115
sodium cholate, sodium deoxycholate, octyl p-D-glUC0pyranoside, sodium taurochenodeoxycholate, Zwittergent, Lubrol PX, and cetylpyridinium chloride did not increase normal rates of ferrocytochrome c (0.5 mM) autoxidation at pH 7.4. In contrast, sodium dodecyl sulfate at 1 g L-’ produced a slight increase in the rates of autoxidation, which was alleviated when the detergent concentration was decreased to 0.5 g L-I. Effects of Detergents on NADPH-P-450 Reductase. Of the 16 detergents, 10 either increased the rate of NADPHP-450 reductase catalyzed reduction of cytochrome c by up to 20%, when added at a concentration of 8 g L-I, or decreased the rate by less than 10% (Table I). Of the remaining detergents, Nonidet P-40 and sodium taurochenodeoxycholate produced slightly increased rates at 2 and 0.2 g L-I, respectively, while cetylpyridinium chloride completely inhibited reduction at 0.01 g L-I, sodium deoxycholate produced 14% inhibition at 0.5 g L-I, Zwittergent produced 24% inhibition at 0.05 g L-l, and sodium dodecyl sulfate produced 35% inhibition at 0.01 g L-’ (Table I). Since the detergents did not potentiate the autoxidation of cytochrome c at the concentrations used here, these results indicate the detergent concentration ranges which -produce marked decreasesin reductase activity. Effects of Detergents on P-450-Mediated Warfarin Hydroxylation in General. The effects of the detergents on the activities of the P-450s were monitored by their metabolism of R warfarin. P-450 UT-A was monitored by the rates of formation of 4‘- and 6-hydroxywarfarin, P-450 PB-B by 4’hydroxywarfarin formation, and P-450 BNF-B by 6- and 8-hydroxywarfarin formation. The effects of the detergents on the rates of formation of these metabolites in reconstituted systems of the purified P-450s and the reductase devoid of phospholipid are shown in Figures 1-5. The concentrations of the detergents which decreased the rates of metabolism to 50%of the rates determined in the absence of detergent (ED,,) are presented in Table 11. Effects of Detergents on P-4.50 BNF-B Catalyzed Metabolism of R Warfarin. For the reconstituted system with isozyme P-450 BNF-B, all of the ionic and zwitterionic detergents enhanced the rates of warfarin metabolism at low concentrations, while at increasing concentrations the rates were
I
0
08
0
0.8
g L‘’ FIGURE 1: Effects of sodium taurochenodeoxycholate,sodium dodecyl sulfate, Zwittergent, and Lubrol PX on the rates of formation of 6-hydroxywarfarin ( 0 )and 8-hydroxywarfarin (0)from R warfarin catalyzed by purified and reconstituted P-450 BNF-B. Rates are presented as a percentage of the rate determined in the absence of detergent. Catalytic conditions are provided under Experimental Procedures. DETERGENT CONCENTRATION
I
0 120 -
a
I
NONIDET INOnionici
CETYLPYRIDINIUM CHLORIDE !lonlci
1
0
t
DETERGENT CONCENTRATION
g L-’
FIGURE 2:
€ $ f a of Emulgens 91 1 and 913, Brij 35, Triton X-100, Nonidet P-40, and cetylpyridinium chloride on the rates of formation of 6-hydroxywarfarin ( 0 )and 8-hydroxywarfarin (0)from R warfarin
catalyzed by purified and reconstituted P-450 BNF-B. Details are as in Figure 1 and under Experimental Procedures.
increasingly and markedly inhibited (Figures 1-3). Sodium cholate was the most effective of the detergents used, increasing the rates of formation of 6- and 8-hydroxywarfarin from R warfarin by 57%over rates in the absence of detergent. In contrast, the nonionic detergents did not produce increases in the metabolic rates but only inhibition even at low concentrations. The two detergents, which on a weight basis were most effective in inhibiting P-450 BNF-B, cetylpyridinium chloride, and sodium dodecyl sulfate, probably acted by inhibiting the reductase (vide supra). Sodium deoxycholate and Zwittergent also apparently inhibited reconstituted P-450 BNF-B activity by inhibiting the reductase. All of the other detergents inhibited P-450 BNF-B activity at concentrations which did not affect reductase activity. These detergents probably inhibited the warfarin hydroxylase activity either by
TWEEN 80 iNonionic!
120
O C M L B D GLUCOPYRANOSIDE INonionic)
\.
TRITON X - l W ;Nonion!cl
izwltfarlonlc)
I20
80 LL
0 4040
w I-
? SODIUM DEOXYCHOMTE
,
,
LLkLJ CHAPS
TWEEN-80
izwlfterlonlc!
iNO"l0"lCi
1.0
SODIUM CHOMTE ilonlc)
2.0
3.0
4.0
2.0
4.0
DETERGENT CONTENTRATION
>
I
1279
V O L . 2 6 , N O . 5, 1 9 8 7
DETERGENT EFFECTS O N CYTOCHROME P-450 FUNCTION
6.0
8.0
g L-'
FIGURE 5: Effects of sodium cholate, Triton X-100, Tween-80, and CHAPS on the rates of formation of 4'-hydroxywarfarin from R warfarin catalyzed by purified and reconstituted P-450 PB-B. Details are as in Figure 1 and under Experimental Procedures.
*\
0
4
8
12
160
4
12
8
DETERGENT CONCENTRATION g L-' FIGURE3: Effects of Tween 80, octyl P-Dglucopyranoside, CHAPSO,
CHAPS, sodium deoxycholate, and sodium cholate on the rates of formation of 6-hydroxywarfarin ( 0 )and 8-hydroxywarfarin (0)from R warfarin catalyzed by purified and reconstituted P-450 BNF-B. Details are as in Figure 1 and under Experimental Procedures. 120 -
120-O
W E E N 80 (Nonionic!
BD
1
OCTYL GLUCOPYRANOSIDE
3
5
ZWITTERGENT izwilterlonlc)
DETERGENT CONCENTRATION g L-I 4: Effects of Tween 80, octyl P-D-glucopyranoside, Zwittergent, and Triton X-100 on the rates of formation of 4'-hydroxywarfarin ( 0 )and 6-hydroxywarfarin (0)from R warfarin catalyzed by purified and reconstituted P-450 UT-A. Details are as in Figure 1 and under Experimental Procedures. FIGURE
inhibiting the P-450or by disrupting the interaction of the reductase with the P-450. The most effective of these detergents as inhibitors, on a weight basis, was Nonidet P-40 (Table 11). In all cases, detergent-mediated enhancement and inhibition of P-450BNF-B catalyzed metabolism of R warfarin equivalently affected t h e rates of formation of both warfarin me-
Table 11: Inhibitory Effects of Detergents on Purified and Reconstituted P-450 Warfarin Hydroxylase Activities P-450 BNF-B 0.15 2.6 1.7
ED,,' (g L-') P-450
P-450 detergent type PB-B UT-A Lubrol PX nonionic Tween 80 nonionic 1 .o 0.10 octyl nonionic 1.8 P-O-glUCOpyranoside Emulgen 9 1 1 nonionic 0.11 Emulgen 9 13 nonionic 0.22 Brij 35 nonionic 0.18 Triton X-100 nonionic 0.16 0.08 0.20 Nonidet P-40 nonionic 0.09 cetylpyridinium ionic 0.01 chloride sodium dodecyl ionic 0.01 sulfate sodium ionic 0.40 taurochenodeoxycholate sodium ionic 2.6 deoxycholate 1.7 1.4 sodium cholate ionic CHAPSO zwitterionic 4.2 CHAPS zwitterionic 4.6 3.3 Zwittergent zwitterionic 0.08 0.08 'The concentration of detergent producing 50% inhibition of warfarin hydroxylation by the reconstituted P-450s. Activities are compared with activities in the absence of detergent. tabolites. There was no conversion of any of the P-450sto P-420at the detergent concentrations used here. When the P-450BNF-B system was reconstituted with the more vigorous mixing of the P-450and reductase, the rate of warfarin hydroxylation was increased by 10% relative to that in a normally mixed system.
Effects of Detergents on P-450 UT-A Catalyzed Metabolism of R Warfarin. With reconstituted systems of P-450 UT-A and reductase, low concentrations of zwitterionic and nonionic detergents produced increases in rates of R warfarin metabolism (Figure 4). A t higher detergent concentrations, metabolism was markedly inhibited. In contrast to the situation with P-450BNF-B, a single class of detergents (nonionic) produced variable effects with respect to enhancing metabolic rates. Tween 80, a nonionic detergent, did not enhance rates of metabolism, while Triton X-100 and octyl p-D-glUC0-
1280 B I O C H E M I S T R Y pyranoside, which are also nonionic, did. Sodium cholate, an ionic detergent, did not enhance metabolic rates at low concentrations, The inhibitory effects of Zwittergent on R warfarin metabolism were probably a consequence of its inhibition of reductase activity. The other detergents used either disrupted the interaction between the reductase and the P-450 or inhibited the P-450. All of the detergents used equivalently affected the rates of formation of 4'- and 6-hydroxywarfarin from R warfarin (Figure 4). Effects of Detergents on P-450PB-B Catalyzed Metabolism of R Warfarin. The warfarin hydroxylase activity of the reconstituted system containing P-450 PB-B was not enhanced by the addition of an ionic, zwitterionic, or either of two nonionic detergents (Figure 5 ) . At all detergent concentrations, metabolic rates were inhibited, and inhibition increased with increasing detergent concentrations. Triton X- 100 was the most efficient of the detergents used in inhibiting the warfarin hydroxylase activity of P-450 PB-B (Table 11). The effects of the detergents CHAPS and sodium cholate on the R warfarin hydroxylase activity of reconstituted P-450 PB-B were investigated by using a 3-fold excess of the substrate R warfarin relative to the concentration used in all other studies. In no case did the increased warfarin concentration alter the detergent-mediated inhibitory effects. Steady-Stare Aggregation Equilibria. Sodium cholate was selected to represent the effects of detergents on the aggregation equilibria of P-450 BNF-B, the reductase, and the mixture of the P-450 and reductase, because it did not affect reductase activity (cytochrome c reduction), it both increased and inhibited P-450 BNF-B activity, it did not interfere with the detection of proteins eluting from the columns, and it is widely used in enzyme studies. The concentrations of sodium cholate selected for these studies corresponded to those producing maximal enhancement of P-450BNF-B dependent warfarin hydroxylase activity and maximal inhibition, respectively. The Superose 6 column has an optimal molecular weight separation range of 5000 to 5 X lo6. Semi-log plots of the molecular weights of the standards vs. their retention times on the Superose 6 column were linear, but y-globulin deviated from the line (Figure 6). Sodium cholate at the concentrations used in these studies did not affect the calibration plots. With bovine serum albumin as the standard, the peak areas varied linearly with concentration of protein injected onto the column (Figure 6). Chromatography of P-450 BNF-B on the Superose 6 column yielded a small peak corresponding to a dimer (1 10 kDa) and a number of smaller peaks corresponding to high molecular weight aggregates (Figure 7). The peak areas were considerably less than expected, and chromatography of a second preparation of the P-450 yielded no peak at all in the absence of detergent (Figure 8). The probable explanation for these effects is that the P-450 precipitated on the column. The elution profiles of the P-450 in the presence of 0.32 or 3.0 g of sodium cholate L-' showed larger molecular weight aggregates, apparently solubilized by cholate. In the absence of cholate, the reductase eluted primarily as a single peak of 460 kDa, equivalent to a hexamer, with a minor monomer component (Figure 7). The 1:l mixture of P-450 BNF-B and the reductase eluted as two peaks; the major peak was broad and asymmetric and was centered at 560 kDa, and the minor peak was at 110 kDa (Figure 7). The area under the major peak was 2.3-fold greater than that of the reductase alone under the same conditions. In the presence of 0.32 g of sodium cholate L-I, no changes occurred in the
KAMINSKY ET AL.
I
\--"*
30
40
50
60
70
MINUTES
FIGURE6: (I) Semilog plot of apparent molecular weight vs. retention time on a Superose 6 column. The protein standards are indicated by (+) and are (1) thyroglobulin, (2) y-globulin, (3) ovalbumin, (4) myoglobin, (5) vitamin B-12, and BSA, bovine serum albumin, and oligomers. The unknown proteins or complexes are indicated by (0) and are (A,) P-450 dimer, (B,) NADPH-P-450 reductase monomer, (B6) NADPH-P-450 reductase hexamer, and (A6:B6)P-450 hexamer/NADPH-P-450 reductase hexamer complex. (11) Chromatographic peak areas of bovine serum albumin vs. quantity injected onto Superose 6 column. The bar indicates the range of protein concentrations used in this study. The numbers represent retention times in minutes. Experimental details are reported under Experimental Procedures.
elution profiles of the reductase or the 1:l mixture, except that the area under the 1IO-kDa peak decreased by approximately 20% and that under the 560-kDa peak increased by approximately 20%. With 3.0 g of sodium cholate L-' in the sample and elution buffer, there was a significant shift in the pattern of the peaks of the reductase and the protein mixture (Figure 7 ) . The reductase disaggregated principally to a 140-kDa species (dimer) with the original hexamer peak remaining as a minor shoulder. The chromatogram of the mixture also exhibited a disaggregation with a major peak at 140 kDa and a higher molecular weight shoulder. When the 1:1 mixture of P-450 and reductase was prepared, by slowly adding the P-450 to the reductase while being mixed with a vortex device, and chromatographed, the elution profiles were altered from those previously described (Figure 7). A major peak was obtained at 700 kDa with a very minor peak at lower molecular weight (Figure 8). The same preparation and chromatography of the mixture but in the presence of 3.0
D E T E R G E N T E F F E C T S O N C Y T O C H R O M E P-450 F U N C T I O N 0.05 1
9
I
VOL. 26, NO. 5 , 1987
1281
I
I1
8
c i
I
ooob-----
&A,'-'\'&
----
j
0 00
MINUTES
Chromatograms of P-450 BNF-B ( 1 pM) NADPH-P-450 reductase (1 pM) (- - -), and a 1:1 mixture of P-450 BNF-B (1 wM) and the reductase (1 wM) (-) prepared by gentle mixing, on a Superose 6 column. (I) Elution buffer was 0.05 M Tris-HCI (pH 7.4) containing 0.05 M NaCl at a flow rate of 0.3 mL mi&. (11) Elution buffer was as above but with the addition of 3.0 g of sodium cholate L-I. The numbers represent retention times in minutes. FIGURE
7:
(-a-),
g of sodium cholate L-' yielded two major peaks at 450 and 120 kDa with several minor high molecular weight aggregates (Figure 8). DISCUSSION The systematic assessment of the effects of 16 detergents (including ionic, zwitterionic, and nonionic categories) on the reconstituted activity of P-450 BNF-B has provided a number of insights into the function of this P-450. If we assume that detergents simulate the action of phospholipids in facilitating functional complex formation between the reductase and P-450 BNF-B (Muller-Enoch et al., 1984), then our results indicate that ionic, including zwitterionic, detergents facilitate complex formation at low concentrations and are disruptive at higher concentrations. In contrast, nonionic detergents are disruptive over all of the concentration ranges used. With respect to their ability to inhibit P-450 BNF-B dependent warfarin hydroxylase activity, the 16 detergents fall into 4 groups on a weight basis: Zwittergent, Nonidet P-40, Triton X-100, Brij 35, Emulgens 9 11 and 9 13, Lubrol PX, and sodium taurochenodeoxycholate inhibited the activity by 50% at concentrations between approximately 0.1 and 0.4 g L-l; Tween 80, octyl P-D-glucopyranoside,sodium deoxycholate, and sodium cholate at approximately 1.5-2.5 gL-'; CHAPS and C H A P S 0 at approximately 4.0-4.5 g L-I; and cetylpyridinium chloride and sodium dodecyl sulfate at 0.01 g L-]. In the last group, the potent inhibitory effects were probably a consequence of the
Q N
U
I
MINUTES 8: Chromatograms of P-450 BNF-B ( 1 pM) (---), NADPH-P-450 reductase (---),and a 1:l mixture of P-450 BNF-B (1 pM) and the reductase (1 pM) (-) prepared by slow addition of the P-450 to a solution of the reductase during mixing with a vortex device, on a Superose 6 column. (I) Elution buffer was 0.05 M Tris-HC1 (pH 7.4) containing 0.05 M NaCl at a flow rate of 0.3 mL mi&. (11) Elution buffer was as above but with the addition of 3.0 g of sodium cholate L-l. The numbers represent retention times in FIGURE
minutes.
inhibition of reductase activity. The relationship of detergent charge to its ability to enhance reconstituted P-450 BNF-B activity is not generally applicable to all P-450s. Thus, with P-450 UT-A, relationships based on detergent charge were less clearly differentiated. In this case, two of the three nonionic detergents used enhanced warfarin hydroxylase activity at low concentrations while the third inhibited at all concentrations used. These rate-enhancing effects of nonionic detergents did not indicate a reversal of detergent charge requirements of the two P-450s since with P-450 UT-A an ionic detergent produced enhanced warfarin hydroxylase rates while another did not. Detergent-mediated effects on reconstituted metabolic rates were clearly dependent on the specific P-450 present, and this was reinforced by the results obtained with P-450 PB-B. With this form, ionic, zwitterionic, and nonionic detergents all failed to enhance warfarin hydroxylase activity, and only inhibitory effects were observed. In contrast, an early study on a partially purified P-450 preparation, which was probably identical with P-450 PB-B, demonstrated enhanced benzphetamine N-demethylase activity in the presence of the nonionic detergents Emulgen 911, Triton N-101, and Triton X-100 (Lu et al., 1974). Apparently, the substrate plays a role in determining the detergent-mediated effects on reconstituted P-450 activities. Overall, the results discussed above suggest that there are profound differences between the functional interactions of
1282 B I O C H E M I S T R Y the reductase with the various forms of P-450. Since detergents only enhance these facilitative interactions at low concentrations, the detergent effects leading to the increased rates are probably subtle. All of the detergents used at higher concentrations inhibited all three P-450s, but the large variations in the EC5+ of a single detergent with different isozymes support the concept of different types and affinities of interactions between the reductase and the various P-450s. The most marked difference in the inhibitory effects of a detergent on the various P-450s was that with Tween 80 where 2.6 g L-' is required to inhibit P-450 BNF-B by 50% while only 0.1 g L-' has a comparable effect with P-450 UT-A. Such differences suggest that the detergent-mediated inhibition of P-450 activity is not a consequence of the sequestration of the substrate by the detergent since such an apparent decrease in substrate concentration should equivalently affect all P-450 activities. This is supported by the fact that a detergent concentration-rate profile of P-450 BNF-B warfarin hydroxylase activity performed with a 3-fold excess of substrate over that normally used was identical with a control performed at the substrate concentration regularly used. Two factors suggest that the detergents inhibit P-450 function without denaturing the secondary or tertiary structures of P-450. At the detergent concentrations used in this study to inhibit the P-450s, there was no conversion of P-450 to cytochrome P-420. With P-450s BNF-B and UT-A, which each catalyze the formation of two metabolites from warfarin, none of the detergents differentially affected the rates of formation of the two products. This suggests that detergent-mediated rate changes were not a consequence of detergent-mediated conformational changes in the catalytic sites of the P-450s since such changes would be expected to differentially affect the formation rates of the two metabolites. This conclusion contradicts that of Ingelman-Sundberg (1977), who suggested that nonionic detergents enhanced the catalytic activity of microsomal-bound P-450 by modifying its conformation. Since detergent-mediated effects on P-450 activities were apparently not a consequence of secondary or tertiary conformational changes, the sequestration of substrate, or of spin-state changes (Wagner & Grey, 1985), we sought an explanation in quaternary structure changes. Commercially packed molecular sieve HPLC columns are well-suited to assessing the role of such aggregation state changes in monooxygenase activities since they resolve aggregates of components of the reconstituted system based on molecular weight, they can be used under conditions similar to those used for metabolic studies, and when used with the FPLC system they yield highly reproducible results. Our studies to assess quaternary structure changes were designed to determine whether the sodium cholate concentrations which optimally enhanced and fully inhibited warfarin hydroxylase activities of P-450 BNF-B would produce changes in the aggregation states of P-450, the reductase, or the mixture of the reductase and P-450. Such correlations would imply an association between aggregation state and activity. Such an association has previously been invoked to explain the effects of octyl fi-D-ghcopyranoside and CHAPS on reconstituted P-450LM2from PB-treated rabbits (Dean & Gray, 1982; Wagner et al., 1984). Although ultracentrifugation studies have been interpreted to indicate that complex formation between P-450 and reductase is not a prerequisite for monooxygenase activity (Dean & Gray, 1982; Wagner et al., 1984), many other studies have indicated the necessity of complex formation for such activity
K A M I N S K Y ET AL.
to occur (Miwa et al., 1979, 1981; Guengerich & Holladay, 1979; Muller-Enoch et al., 1984; Ruckpaul et al., 1980; Guengerich & Davidson, 1982; French et al., 1980). P-450LM2 and rabbit liver reductase on gel exclusion chromatography each eluted as hexamers, which on mixing formed a hexamer/hexamer complex of 800 kDa (French et al., 1980). The rat liver reductase in our studies also eluted as a hexamer while the P-450 BNF-B was apparently dimeric, although problems with precipitation added some uncertainty to the latter result. When the mixture of the reductase and P-450, formed by slow addition of the P-450 to the reductase while mixing, was chromatographed, a single peak at 700 kDa was observed, which was apparently a hexamer/hexamer complex equivalent to that reported for the rabbit P-450LM2 system (French et al., 1980). When the reductase and P-450 were mixed together less vigorously corresponding to the method typically used to reconstitute in the metabolic studies, a complex was also observed on chromatography but centered at a lower molecular weight, and uncomplexed dimeric P-450 was also present. The width of the complex peak suggests that multiple aggregation equilibria were occurring in the complexes. This would explain the relatively low molecular weight of this complex relative to that of the hexamer/hexamer complex described above. Low concentrations of sodium cholate apparently facilitated the uptake of the uncomplexed P-450 into the complex as demonstrated by the changes in the relative areas of the two chromatographic peaks. This effect possibly explains the observed increased warfarin hydroxylase activity of the reconstituted system at low cholate concentrations. The fact that vigorous mixing of the protein components of the reconstituted system, which leads to a single complex, resulted in an increase in catalytic rates supports this conclusion. The high cholate concentration disaggregated the reductase to a dimer and with the mixture, when prepared by both methods, caused major disruption of the reductase/P-450 complex. Thus, the detergent-mediated inhibitions of reconstituted monooxygenase activities are probably due to disruption of the reductase/P-450 complexes, possibly a consequence of the disruption of the quaternary structures of one or both of the constituent proteins. The sodium cholate mediated disaggregation of the reductase from a hexamer to a dimer, while it apparently was capable of profoundly inhibiting reconstituted P-450 activities, did not affect the NADPH-dependent electron transfer from the reductase to cytochrome c. The quaternary structure of the reductase thus does not inherently affect its electrontransfer properties but more likely controls its interaction with the electron acceptor, P-450. While all of the gel filtration studies were performed with only sodium cholate as detergent, we have no reason to expect that the other detergents used in this study will function to enhance and inhibit P-450-catalyzed rates via different mechanisms. In summary, detergents differentially affect catalytic activities of individual forms of P-450, indicating that the functional interactions of the reductase with various P-450s vary. Some detergents at low concentrations increase rates of reconstituted monooxygenase activity, and with sodium cholate as an example, this is apparently due to increased functional complex formation. All of the detergents used at higher concentrationsinhibit due to disruption of the functional complex. Registry No. PB, 50-06-6; BNF, 6051-87-2; P-450, 9035-51-2; CHAPS, 75621-03-3; CHAPSO, 82473-24-3; NADPH-cytochrome P-450 reductase, 9039-06-9; Zwittergent, 71 833-44-8; Lubrol PX, 9002-92-0; Tween 80, 9005-65-6; octyl 0-D-glucopyranoside,
D E T E R G E N T E F F E C T S O N C Y T O C H R O M E P-450 F U N C T I O N
29836-26-8; Emulgen 91 1, 9016-45-9; Triton X-100, 9002-93-1; Nonidet P-40, 9036-19-5; cetylpyridinium chloride, 123-03-5;sodium dodecyl sulfate, 151-21-3;sodium taurochenodeoxycholate,6009-98-9; sodium deoxycholate, 302-95-4; sodium cholate, 361-09-1;warfarin 4’-hydroxylase, 87843-69-4; warfarin 6-hydroxylase, 87843-70-7; warfarin 8-hydroxylase, 104520-85-6.
REFERENCES Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. Buckle, D. R., Cantello, B. C. C., Smith, H., & Spicer, B. A. (1975) J. Med. Chem. 18, 391-394. Coon, M. J. (1978) Methods Enzymol. 52, 200-206. Dean, W. L., & Gray, R. D. (1982) J. Biol. Chem. 257, 14679- 14685. Denk, H. (1982) in Hepatic Cytochrome P-450 Monooxygenase Systems (Schenkman, J. B., & Kupfer, D., Eds.) pp 469-495, Pergamon Press, New York. Fasco, M. J., Piper, L. J., & Kaminsky, L. S. (1977) J . Chromatogr. 131, 365-313. French, J. S., Guengerich, F. P., & Coon, M. J. (1980) J. Biol. Chem. 255, 41 12-41 18. Guengerich, F. P. (1987) in Mammalian Cytochromes P-450 (Guengerich, F. P., Ed.) Vol. 1, pp 1-54, CRC Press, Boca Raton, FL. Guengerich, F. P., & Holladay, L. A. (1979) Biochemistry 18, 5442-5449. Guengerich, F. P., & Davidson, N. K. (1982) Arch. Biochem. Biophys. 215, 462-417. Guengerich, F. P., Dannan, G. A,, Wright, S . T., Martin, M. V., & Kaminsky, L. S. (1982) Biochemistry 21,6019-6030. Hermodson, M. A,, Barker, W. M., & Link, K. P. (1971) J . Med. Chem. 14, 167-169. Imai, Y. (1976) J. Biochem. (Tokyo) 80, 267-276. Ingelman-Sundberg, M. (1977) Biochim. Biophys. Acta 488, 225-234. Kaminsky, L. S., & Guengerich, F. P. (1985) Eur. J. Biochem. 149, 479-489. Kaminsky, L. S., Wright, R. L., & Davison, A. J. (1971) Biochemistry IO, 458-462. Kaminsky, L. S., Fasco, M. J., & Guengerich, F. P. (1980) J . Biol. Chem. 255, 85-91.
VOL. 26, NO. 5 , 1987
1283
Kaminsky, L. S . , Fasco, M. J., & Guengerich, F. P. (1981) Methods Enzymol. 74, 262-212. Kaminsky, L. S . , Guengerich, F. P., Dannan, G. A., & Aust, S. D. (1983) Arch. Biochem. Biophys. 225, 398-404. Lowry, 0. H., Rosebrough, N . J., Farr, A. L., & Randall, R. J. (1981) J. Biol. Chem. 193, 265-275. Lu, A. Y. H., & Coon, M. J. (1968) J. Biol. Chem. 243, 1331-1332. Lu, A. Y. H., & Levin, W. (1974) Biochim. Biophys. Acta 344, 205-240. Lu, A. Y. H., Levin, W., & Kuntzman, R. (1974) Biochem. Biophys. Res. Commun. 60, 266-272. Miwa, G. T., & Lu, A. Y. H. (198 1) Arch. Biochem. Biophys. 21 I, 454-458. Miwa, G. T., West, S . B., Huang, M.-T., & Lu, A. Y. H. (1979) J. Bioi. Chem. 254, 5695-5700. Muller-Enoch, D., Churchill, P., Fleisher, S., & Guengerich, F. P. (1984) J. Biol. Chem. 259, 8174-8182. Omura, T., & Sato, R. (1964) J. Biol. Chem. 239,2370-2378. Phillips, A. H., & Langdon, R. G. (1962) J. Biol. Chem. 237, 2652-2660. Pohl, L. R., Haddock, R., Garland, W. H., & Trager, W. F. (1975) J. Med. Chem. 18, 513-519. Ruckpaul, K., Rein, H., Ballou, D., & Coon, M. J. (1980) Biochim. Biophys. Acta 626, 41-56. Vermilion, J. L., & Coon, M. J. (1974) Biochem. Biophys. Res. Commun. 60, 1315-1322. Wagner, S . L., & Gray, R. D. (1985) Biochemistry 24, 3809-38 14. Wagner, S . L., Dean, W. L., & Gray, R. D. (1984) J. Biol. Chem. 259, 2390-2395. Waxman, D. J. (1986) in Cytochrome P-450 Structure, Mechanism, and Biochemistry (Ortiz de Montellano, P. R., Ed.) pp 525-539, Plenum Press, New York. West, B. D., Preis, S., Schroeder, C. H., & Link, K. P. (1961) J. Am. Chem. SOC.83, 2676-2619. Yasukochi, Y., & Masters, B. S. S. (1916) J. Biol. Chem. 251, 5 337-5 344. Yonetani, T., & Ray, G. S. (1965) J. Biol. Chem. 240, 3392-3 398.
