Biochimica et Biophysica A cta, 1038 (1990) 114-118

11 4

Elsevier BBAPRO 33571

Binding of inhibitors to the major glutathione S-transferase from bovine brain. Competitive binding between bilirubin and glutathione P a u l R. Y o u n g a n d A n i t a V. B r i e d i s Department of Chemistry, University of lllinois at Chicago, Chicago, IL (U.S.A.)

(Received 10 August 1989)

Key words: Glutathione S-transferase; Ligand binding; Inhibitor binding

The binding of non-substrate ligands to the glutathione S-transferase (RX: glutathione R-transferase, EC 2.5.1.18) from bovine brain has been investigated kinetically by monitoring the inhibition of the enzyme-catalyzed reaction between glutathione and 1-chloro-2,4-dinitrobenzene. Bilirubin, thyroxine, lithocholic acid, retinoic acid and retinol are competitive inhibitors with respect to glutathione. Cooperative binding effects are observed with lithocholic acid, retinoic acid and retinol while cooperative binding is not observed with thyroxine or bilirubin. Bilirubin is the most potent inhibitor with constants of 0.1 and 110 ~tM. 50% of the total activity is lost upon binding to the high-affinity site and the remainder is lost at higher bilirubin concentrations. In spite of the apparently favorable binding for bilirubin, it is estimated that the high intracellular concentrations of reduced glutathione will saturate the enzyme and allow only a small fraction of the bilirubin in brain to bind to the enzyme. It is concluded that the binding of these iigands may be of minor importance in vivo.

Introduction In addition to their presumed capacity in xenobiotic detoxification, the glutathione S-transferase ( R X : glutathione R-transferase, EC 2.5.1.18) have been suggested to have roles in ligand binding and transport [1-3]. The enzymes have been noted to bind heme derivatives [4], bilirubin [5], steroids and bile acids, leukotrienes [6,7], etc. With some of these ligands, however, specific carrier proteins [8] have been found which bind the ligands more efficiently and in several cases, ligands have been found to be displaced from the enzyme by physiological concentrations of glutathione [9,101. We have recently described the purification and kinetic mechanism of the major glutathione S-transferase from bovine brain [11]. The fact that the enzyme is kinetically simple at pH 6.5, following a random, rapid equilibrium mechanism, means that straightforward inhibition analysis can be used to distinguish

Abbreviations: CDNB, 1-chloro-2,4-dinitrobenzene;GST, glutathione S-transferase. Correspondence: P.R. Young, Department of Chemistry, University of Illinois of Chicago, P.O. Box 4348, Chicago, IL 60680, U.S.A.

modes and constants for ligand binding. In order to more clearly define the in vivo role of this enzyme, we have examined the inhibition of the enzyme-catalyzed glutathione-(1-chloro-2,4-dinitrobenzene) conjugation reaction by bilirubin and a variety of other hydrophobic compounds. We have found that blirubin and several other inhibitors bind competitively to the glutathione site raising serious questions regarding the physiological role of these binding reactions.

Materials and Methods Materials

All chemicals used were readily available commercial products and were used without further purification unless otherwise noted. The isolation of the enzyme and the assay methods have been described previously [11]. Inhibition studies

Stock solutions of 0.01-0.1 M arachidonic aicd, corticosterone, thyroxine, lithocholic acid and retinoic acid were prepared in ethanol. Bilirubin stock solutions were prepared in 0.02 M NaOH. Stock solutions were diluted as needed, but the ethanol concentrations in the enzyme assay were typically less than 1% and never exceeded 4%. The amount of base added in the bilirubin

0167-4838/90/$03.50 © 1990 ElsevierSciencePublishers B.V, (BiomedicalDivision)

115 experiments was always small and the p H of the cell was monitored after addition and adjusted if necessary. Solutions of bilirubin and retinoic acid were protected from light and enzyme assays using these inhibitors were carried out in a darkened room. Assays were carried out as previously described [11] using 0.25 #g of enzyme in 1.0 ml phosphate buffer (0.1 M, p H 6.5, 25 o C). The rates were corrected for background reaction and the enzyme was added last to start the reaction. Average values for kinetic constants (common intercepts, etc.) were obtained by standard non-linear least-squares analyses of the appropriate data using the computer program EK-Analyst (Pewter Scientific, Chicago, IL) and representative plots of the data are constructed using these average values. Inhibition constants for ligands showing more complex binding behavior were determined by iterative analysis using the equations given in the text. Error limits on individual kinetic constants are estimated to be less than +4% unless otherwise stated. Determination of bilirubin concentrations in brain One fresh bovine brain was immersed in buffer containing 1 m M reduced glutathione immeidately after removal from the animal and protected from light at 2°C. Within 1 h, the brain was homogenized and centrifuged as described above. The bilirubin concentration in the supernatant was determined by standard methods [12]. In the procedure, bilirubin was coupled with diazotized sulfanilic acid to yield and azobilirubin; the absorbance at 600 nm was measured and the concentration calculated using (log e = 4.86). Total and conjugated bilirubin concentrations were determined in the presence or absence, respectively, of a caffeine-benzoate-acetate accelerating agent [12]. Non-conjugated bilirubin was taken as the difference between total and conjugated.

The compounds tested as inhibitors of the bovine brain glutathione S-transferase are listed in Table I along with the mode of inhibition and the observed inhibition constant(s). All compounds tested were competitive with one substrate and non-competitive with the other. The fact that linear double-reciprocal plots were observed for all of the inhibitors in the table at concentrations at which micellar aggregates may exist, means that the microscopic rate constants for monomer-micelle dissociation are significantly larger than the microscopic rate constants involved in E1 complex formation. Of the compounds examined, only arachidonic acid and corticosterone exhibit linear competitive inhibition with respect to 1-chloro-2,4-dinitrobenzene (CDNB). With glutathione as the variable substrate a non-competitive pattern was observed with the common intersection point above the x-axis. A common intersection such as this is consistent with a mechanism [14] in which the binding of inhibitor alters the binding constant for glutathione by a factor fl; in the case of corticosterone, fl = 0.8. Replots of the slopes of the double-reciprocal plots of velocity vs. variable [CDNB] are linear for both arachidonic acid and corticosterone and the inhibition constants in Table I are taken from the negative x-intercepts of these replots. Lithocholic acid, retinoic acid and retinol exhibit simple competitive inhibition with respect to glutathione and non-competitive inhibition with respect to CDNB, with an intersection point on the x-axis. Replots of the slopes of the double-reciprocal plots (Fig. 1), however, show significant upward curvature at high inhibitor concentrations. Upward curvature such as this cannot be produced by differential binding to the two subunits unless that binding is coopertive in nature; the binding of the first molecule must change the geometry at the second binding site to make binding of the inhibitor more favorable. The equation given below describes the binding behavior of an inhibitor that is

Results and Discussion TABLE I

Using 1-chloro-2,4-dinitrobenzene as substrate, we have previously suggested that the kinetic mechanism of the bovine brain glutathione S-transferase can be described by random binding, with all binding steps in rapid equilibrium with respect to turnover. Further, we observed inhibition by a variety of compounds, including products, that suggested that inhibitors can bind to free enzyme and both binary complexes. Although the enzyme consists of non-identical subunits, the kinetic data suggest that for the ligands and substrates examined, the subunits are catalytically equivalent and that binding is non-cooperative. This is consistent with previous report for the rat liver heterodimer isozymes which were also found to display no evidence of cooperativity [13].

Inhibition of glutathione S-transferase from bovine brain a KI

K(

cb

Competitive with CDNB: (KinCDNB= 0.41 mM) Arachidonic acid 2.5 mM Corticosterone 3.3 mM Competitive with GSH: ( g GSH = 0.06 mM) Bilirubin 0.1/~M 100 jaM Thyroxine 6.6/x M Lithocholic acid 41 jaM 0.28/.tM Retinoie acid 30 ~M 12/,tM Retinol 10 ~M 0.3/LM

0.007 0.4 0.03

a Inhibition constants for the conjugation of 1-chloro-2,4-dinitrobenzene by glutathione catalyzed by bovine brain glutathione Stransferase; 0.1 M phosphate buffer (pH 6.5) 25 o C. b Interaction coefficient for cooperative binding; see Eqn. 1.

116

/

04

~02

cant conformational change occurs upon inhibitor binding. Bilirubin is a very powerful inhibitor of the enzymecatalyzed reaction. At low concentrations of bilirubin (below 0.5 /~M), the rapid loss of about 50% of the activity is observed followed by a gradual decline in the remaining activity as the concentration of bilirubin is increased. Reciprocal plots of fraction activity inhibited vs. bilirubin concentration are linear at very low concentrations of bilirubin (Fig. 2), with a non-linear dependence at higher concentrations. In Fig. 2, the inhibition by low concentrations of bilirubin ( < 0.5 /~M), extrapolated to the y-axis, yields a value corresponding to 50% inhibition (an intercept of 2.0), providing more evidence that the subunits, although different, have very similar values of Vmax [11]. The inhibition at low concentrations of bilirubin is competitive with respect to glutathione, as demonstrated by the reciprocal plot of fraction inhibited vs. bilirubin concentration at various concentrations of glutathione and constant [CDNB] (Fig. 3A). The negative x-intercepts of these plots give the apparent inhibition constant at each glutathione concentration and K~ is calculated using

Q

E,,.o0.o,,o°¢,03... /

"7 ~

3.C

0.0

I

0.0

I

10.0

I

I

I

20.0

I

300 1

I

I

I

p~

40.0

,

.0

600

, mM-1

rGlutathione~

Fig. 1. Double-reciprocal plots of 1 / v o vs. 1/[GSH] for the reaction between CDNB and GSH catalyzed by bovine brain glutathione S-transferase at the following fixed concentrations of lithocholic acid: $, zero; o, 5.0 ~M; A, 10.0 /~M; A, 15.0/tM; It, 20.0/~M; and ra, 30.0 laM. Assay mixtures contained 0.25 /~g enzyme in 1.0 ml phosphate buffer (0.10 M, pH 6.5), 25 o C; rates were corrected for background reaction. Inset: replot of (slope) vs. [lithocholic acid]; the solid line is drawn using Eqn. 1 and the constants given in Table I.

gapp KI

1 + [GSH] ]

KA ] 3.0

2.0

competitive with one substrate in a random, rapid equilibrium mechanism and binds cooperatively to two non-identical binding sites (while substrate does not) [14].

( [A][B] ~/_ 130

+

0//~'~

--.S-~ 1 . 0 ( 10.C

00I 0.0 , 0i2 , 04~ ~ 0.6 / 1

{A} + [I]]

, /jM -1

[Bi,i,.~io]

K,j

2[I]

1

8.(3

(1) >o

Glutathione is A, C D N B is B and inhibitor is I. The interaction factor is c and it becomes significant when c < 1. The best fit of the observed data for inhibition by lithocholic acid (Fig. 1) is given by K l ---41 jaM and c = 0.007, making the effective binding constant for the second mole of inhibitor approx. 0.3 ~tM. A similar mode of inhibition is observed for retinoic acid and for retinol. Thyroxine also shows competitive inhibition with respect to glutathione, but slope replots are linear, giving a simple inhibition constant of 6.6 /~M. Inhibition constants for the compounds examined and values for the interaction factors are shown in Table I. The statistical fit to the multi-parameter equation generates significant uncertainty but the numbers in the table are probably accurate to within ~ 10%. In all cases, the interaction factor is quite small suggesting that a signifi-

"

/

6.0

#

/

./

4.0

,> 2.0 (

0.0 0.0

.0

40 1

.0

8.0

10.0

, )jM -1

F__Bilirubin-I

Fig. 2. Double-reciprocal plots of 1/(fraction activity inhibited) vs. 1/[bilirubin] for the reaction between CDNB and GSH catalyzed by

bovine brain glutathione S-transferase at the following fixed concentrations of reactants: [GSH] = 0.20 mM and [CDNB] = 1.0 mM. Assay mixtures contained 0.25 ~tg enzyme in 1.0 ml phosphate buffer (0.10 M, pH 6.5), 25°C; rates were corrected for background reaction. Inset: Expanded scale to show data for high bilirubin concentrations. The solid lines are drawn using Eqn. 2 and the data shown in Table I.

117 Expressing Eqn. 2 for the case where k~ >> K I and [I] < K ( (low bilirubin concentrations), the reciprocal of the fraction inactivation is given by Eqn. 3 (v i and v0 are the velocities in the presence and absence of inhibitor, respectively).

A

12.0

>o >-

8.G

4x

0.0

-4.0

3.G

g

1/(l_v~)

4.0

8.0

Vo

O ~ i

I

10.0

I

I

20.0

i

(~

3 .0

t

40.0

Fig. 3. (A) Double-reciprocal plots of 1/(fraction activity inhibited) vs. 1/[bilinJbinl for the reaction between CDNB and GSH catalyzed by bovine brain glutathione S-transferase at the following fixed concentrations of glutathione: o, 0.05 mM; o, 0.10 mM; HI, 0.20 mM; and 12, 0.30 raM. (B) Double-reciprocal plots of 1/o 0 vs. 1/[GSH] for the reaction between CDNB and GSH catalyzed by bovine brain glutathione S-transferase at the following fixed concentrations of bilirubin: o; o, 10.0/~M; El, 20.0 #M; and (m, 30.0 #M. For both A and B, assay mixtures contained 0.25 #g enzyme in 1.0 ml phosphate buffer (0.10 M, pH 6.5), 25°C; [CDNB] =1.0 mM; rates were corrected for background reaction.

where Kap p is the intercept and K^ is the Michaelis constant for glutathione. The inhibition constant determined in this manner is 0.11 #M. The inhibition at high concentrations of bilirubin is also competitive with respect to glutathione (Fig. 3B); the apparent Vm~x, however, is only half of the maximum velocity observed in the absence of bilirubin (open circles). Non-linear least-squares fit of the data in Fig. 2 using Eqn. 2 generate values of 0.11 and 110 # M for the two inhibition constants for bilirubin binding; the solid lines in Fig. 2 are calculated using these values. [A][B] ~/

o0 Vm,~

[A]

[I]

[I] ]

~KAKBj~I+'-~A + ' ~ I + 2 K [ ] (I+[B]\/_

[A]+[I]~{I+[A]. [I]X

(3)

_

As evident from this equation, a linear dependence is predicted at low inhibitor concentrations with a y-intercept of 2 (Figs. 2 and 3A). Eqn. 2 can also be arranged to demonstrate the competitive inhibition that is observed at high inhibitor concentrations ( K I' >> K I and [I] > KI' ). With glutathione (A) as the variable substrate:

12.0

B

O'%.O

= 2 K I ( I + [A]~ 1

(2)

[A] V.~x

+

Vm.~

(4)

Similar, biphasic binding of bilirubin to the rat liver glutathione transferase 1-2 (B, Y,Yc, 'ligandin') has been reported with dissociation constants of 0.02 and 3 # M [5]. As with the brain enzyme, the liver enzyme transferase B is a heterodimer. The subunit M, values for the two enzymes are different, however, 22000-24000 for the brain and 25 000-28 000 for the liver) as are their isoelectric points (9.4 for the liver enzyme and 7.4 for the enzyme from obvine brain), leaving little doubt that the two are very different proteins. An interesting feature of the inhibition data in Table I is that the majority of the compounds tested are competitive with glutathione and non-competitive with respect to CDNB. This is somewhat surprising since 'hydrophobic' compounds such as lithocholic acid, thyroxine, vitamin A metabolites and bilirubin would, a priori, be expected to bind to the 'hydrophobic substrate' site and therefore, be competitive with CDNB. This general trend of tight inhibitor binding to the glutathione site has also been observed for inhibition of the brain transferase by N A D H and a variety of nucelotides and their analogs [11]. In the present case the only inhibitors that appear competitive with CDNB, and are therefore potential substrates or substrate analogs at low concentrations, are arachidonic acid and corticosterone. The observation that arachidonic acid binds in the 'second substrate site' is consistent with our previous observation [11] that the long-chain fatty acids are excellent competitive inhibitors with respect to CDNB, suggesting a 'fatty acyl'-character for this site. Binding at this site is consistent with the observation that GST isozymes have been reported to catalyze the conjugation of glutathione to epoxyeicosatrienoic acids [15], to catalyze the transformation of P G H 2 to P G F 2 [16] and the conversion of leukotriene A 4 to C 4 [17,18].

118 The inhibition constant for arachidonate, 2.5 raM, is large however, suggesting that arachidonate binding in vivo by the brain transferase is of minor importance. Clearly, the observed binding constant for bilirubin at the high affinity site (0.11 gM) is low enough to possibly be of physiological significance and it is often suggested that bilirubin binding is an important function of these enzymes. The observation, however, that bihrubin and glutathione bind competitively to the brain enzyme, raises the important question of the effect of the large pool of reduced glutathione on the ability of the available enzyme to bind bilirubin. Using the approximation that 1 g of tissue < 1 ml of cytoplasmic volume, the concentration of reduced giutathione in brain is > 1 mM and the concentration of total bilirubin (conjugated and free) is approx. 1.2 gM. Assuming the Michaelis and inhibition constants in vivo are about the same as we have measured in vitro, the fraction activity remaining can be estimated by Eqn. 3. If all of the available bilirubin can bind to the enzyme, approx. 30% of the total activity would be inhibited, while if only free bilirubin binds (0.4 /~M), about 10% of the activity would be lost. Thus, in the presence of physiological concentrations of substrates and inhibitor (bilirubin) the brain transferase would still retain most of its catalytic activity. The fraction of the total bilirubin that would be complexed by the enzyme can also be calculated if the concentration of enzyme in brain can be estimated. Based on the activity of crude supernatant, there are approx. 3 mg of transferase per 100 g of tissue, giving a concentration of about 0.2 #M. Again, using the observed kinetic constants, only 4% of the total bilirubin would be bound by the enzyme. If only free bilirubin exhibits the high-affinity binding, less than 0.5% would be bound to the enzyme. Similar calculations can be made for other ligands that bind competitively with glutathione. Since none of those that we have examined bind as tightly as bilirubin, the high intracellular concentrations of reduced glutathione will always dominate the binding equilibrium and only a very small fraction of the ligand will be bound to the enzyme. These results are similar to those recently reported by a number of workers who have found that physiological concentrations of giutathione displace ligands bound to GST isozymes [9,10]. Interestingly, however, glutathione was found to inhibit lithocholate binding to rat liver YbYb, but did not affect binding to YaYc [10]. Also, cholic acid and lithocholic acid, both high affinity binders to isozyme YaYa and structurally very similar, apparently do not compete for the same

site on this isozyme. Although it is possible that the diversity of GST isozymes allows a certain fraction of the total GST enzymes in any one tissue to bind bile acids, etc., in the presence of high levels of glutathione, the situation is certainly not simple and it is becoming increasingly apparent that many specific binding and transport proteins exist in cells which possess high GST levels and that these ligands bind to their specific carrier proteins in preference to GST isozymes [6,8]. In light of these observations, the binding of ligands to the active site of GST isozymes may play a minor role in intracellular transport.

Acknowledgements This work was supported by a grant from the National Institutes of Health, NS-17094, and a Research Career Development Award (to P.R.Y.), NS-00775.

References 1 Jokoby, W.B. (1978) Adv. Enzymol. Relat. Areas Mol. Biol. 46, 46, 383-414. 2 Mannervik, B. (1985) Adv. Enzymol. Relat. Areas Mol. Biol. 57, 357-417. 3 Douglas, K.T. (1987) Adv. Enzymol. Relat. Areas Mol. Biol. 59, 103-167. 4 Senjo, M., Ishibashi, T. and Imai, Y. (1985) J. Biol. Chem. 260, 9191-9196. 5 Arias, I.M., Ohmi, N., Bhargrave, M. and Listowsky, I. (1980) Mol. Cell. Biochem. 29, 71-80. 6 Sun, F.F., Chau, L.Y., Spur, B., Corey, E.J., Lewis, R.A. and Austen, K.F. (1986) J. Biol. Chem. 261, 8540-8546. 7 Sun, F.F., Chau, L.Y. and Austen, K.F. (1987) Fed. Proc. 46, 204-207. 8 Takikawa, H., Sugiyama, Y. and Kopowitz, N. (1986) J. Lipid Res. 27, 955-966. 9 Clark, A.G. and Carrol, N. (1986) Biochem. J. 233, 325-331. 10 Takikawa, H., Stolz, A., Sugimoto, M., Sugiyama, Y. and Kaplowitz, N. (1986) J. Lipid Res. 27, 652-657. 11 Young, P.R. and Briedis, A.V. (1989) Biochem. J. 257, 541-548. 12 Michaelsson, M. (1961) Scand. J. Clin. Invest. 13, 1-9. 13 Danielson, U.H. and Mannervik, B. (1985) Biochem. J. 231, 263-267. 14 Segal, I.H. (1975) in Enzyme Kinetics, pp. 347-404, John Wiley & Sons, New York. 15 Spearman, M.E., Prough, R.A., Estabrook, R.W., Falck, J.R., Manna, S., Leibman, K.C., Murphy, R.C. and Capdevila, J. (1985) Arch. Biochem. Biophys. 242, 225-230. 16 Burgess, J.R., Yang, H., Chang, M., Rao, M.K., Tu, C.P. and Reddy, C.C. (1987) Biochem. Biophys. Res. Commun. 142, 441-447. 17 Soderstrom, M., Mannervik, B., Orning, B., Hanunarstrom, S. (1985) Biochem. Biophys. Res. Commun. 128, 265-270. 18 Wu, C. (1986) Biochem. Biophys. Res. Commun. 134, 85-92.

Binding of inhibitors to the major glutathione S-transferase from bovine brain. Competitive binding between bilirubin and glutathione.

The binding of non-substrate ligands to the glutathione S-transferase (RX:glutathione R-transferase, EC 2.5.1.18) from bovine brain has been investiga...
463KB Sizes 0 Downloads 0 Views