Chem. Res. Toxicol. 1995,8, 103-110
103
Mercurials and Dimercaptans: Synergism in the Induction of Chemoprotective Enzymes Ryan R. Putzer, Yuesheng Zhang, Tory Prestera, W. David Holtzclaw, Kristina L. Wade, and Paul Talalay” Department of Pharmacology and Molecular Sciences, The Johns Hopkins University School Medicine, Baltimore, Maryland 21205
of
Received August 16,1994@
The induction of NAD(P)H:quinone reductase (EC 1.6.99.2; QR) in Hepa lclc7 murine hepatoma cells provides a versatile quantitative model for measuring the potencies of inducers for Phase 2 detoxication enzymes. Since many inducers of these enzymes also protect animals and their cells against the toxic and neoplastic effects of carcinogens, understanding the mechanisms of induction of Phase 2 enzymes is important. Both HgClz and 2,3-dimercaptopropanol (BAL) are inducers of QR in these cells, and paradoxically BAL (which is about 30 times less potent than HgC12) enhances the inducer potency of HgClz substantially. This synergism depends on the presence of two thiol groups on adjacent carbon atoms. Since nonchelated mercury(I1)-thiol compounds did not show synergism, the formation of very high affinity bidentate chelates appears to be essential for such synergism. A major mechanism for the augmentation of the inducer potency of mercury(I1) by BAL is the more rapid cellular uptake and the accumulation of higher intracellular concentrations of mercury. It is also possible t h a t BAL-mercury chelates are intrinsically more potent a s inducers. Although equimolar mixtures of BAL and HgClz and the synthetic chelate isolated from such mixtures were more potent inducers than HgClz alone, the presence of excess BAL increased this inducer synergism even further. By chromatography we showed the reversible formation of higher order complexes between BAL and mercury(I1). Such complexes are transported into cells more efficiently and appear to be more potent than free HgClz or the chelate obtained from equimolar mixtures of BAL and HgC12. Depletion of intracellular glutathione levels by treatment of cells with buthionine sulfoximine (a glutathione synthetase inhibitor) enhanced the inducer potencies of HgClZ, BAL, and their chelates, whereas elevation of intracellular glutathione levels by treatment with the ethyl ester of glutathione lowered the inducer potency of a BAL-mercury(I1) chelate. Induction by BAL and by HgClz and the synergism of inducer potency of their mixtures depend on the activation of the same genetic enhancer element (ARE or EpRE) that is part of the inducer mechanism of all other monofunctional inducers.
Introduction Exposure of animals and their cells to a variety of dissimilar chemical agents results in substantial inductions of Phase 2 detoxication enzymes’ [e.g., glutathione transferases, quinone reductase (QR),2 UDPglucuronosyltransferases, epoxide hydrolase] (2-5). In-
* Corresponding author, at the Department of Pharmacology and Molecular Sciences,Johns Hopkins University School of Medicine, 725 N. Wolfe St., Baltimore, MD 21205. Voice: 410-955-3499; FAX: 410550-6818. @Abstractpublished in Advance ACS Abstracts, November 15,1994. Phase 1and Phase 2 enzymes. Enzymes involved in the metabolism of xenobiotics have been classified intotwo broad categories. Phase 1 enzymes (principally cytochromes P-450) functionalize compounds largely by oxidative or reductive reactions. Phase 2 enzymes cany out the conjugations of such functionalired compounds with endogenous ligands (e.g., glutathione and glucuronic acid). Quinone reductase is classified as a Phase 2 enzyme because it serves protective functions (I), is induced coordinately with other Phase 2 enzymes, and is regulated by enhancer elements that are similar to those controlling other Phase 2 enzymes (2). Abbreviations: QR, quinone reductase, NAD(P)H:(quinone acceptor) oxidoreductase (EC 1.6.99.2); BAL (British anti-Lewisite), 2,3dimercaptopropanol; BSO, DL-buthionine (RS)-sulfoximine; DMSO, dimethyl sulfoxide; CD or CD-, the concentration of an inducer required to double the QR specific activity in Hepa lclc7 murine hepatoma cells under specified conditions; CDQH,the concentration of an inducer required to ,doublethe growth hormone production in Hepa lclc7 murine hepatoma cells transfected with the plasmid p41-284GH, which contains a 41-bp enhancer element and the minimal promoter region of the mouse glutathione transferase Ya gene (8),linked to a human growth hormone reporter gene (2);GSH, glutathione.
ducers of these enzymes include oxidizable diphenols and phenylenediamines, Michael reaction acceptors (i.e., olefins or acetylenes conjugated to electron-withdrawing groups), isothiocyanates, peroxides, mercury(I1)salts and organomercurials, trivalent arsenicals, l,2-dithiole-3thiones, and vicinal dimercaptans (1,2-dithiols) ( 2 , 5 , 6 ) . With the exception of vicinal dimercaptans, all known inducers contain electrophilic centers (or may acquire such centers by metabolism). More specifically, the majority of inducers are soft electrophiles, and their inducer potencies generally parallel their affinities for thiol groups, and their reactivities as substrates for glutathione (GSH) transferases (6, 7). Our interest in the induction of Phase 2 detoxication enzymes stems from the widespread recognition that elevation of these enzymes protects animals and their cells against the toxic and neoplastic effects of carcinogens and other electrophiles (4). The generally coordinate inductions of these enzymes (and the accompanying elevations of GSH levels) establish a protective state against electrophiles that has been termed the “electrophile counterattack response” (5). An understanding of the chemistry of inducers and the molecular mechanisms of induction therefore is very important for devising safe and effective strategies for chemical protection (chemoprotection) against the development of malignancy.
0893-228x/95/2708-0103$09.00/00 1995 American Chemical Society
104 Chem. Res. Toxicol., Vol. 8, No. 1, 1995
HgC12 is one of the more potent inducers that have been described and displays a limited range of chemical reactivities. Consequently, mercury(11) appeared to be an attractive probe for analyzing the mechanism of Phase 2 enzyme induction. The unexpected finding that vicinal dithiols, e.g., 2,3-dimercaptopropanol(BAL), which are powerful nucleophiles and extremely avid chelators for mercury (91,were also inducers of Phase 2 enzymes suggested the need for a closer examination of the relation between induction by mercury(I1) and vicinal dithiols. We hoped that such information might shed light on the mechanism of the induction process. We expected that high affinity dimercaptan-mercury chelates would be inactive as inducers. To our surprise, mixtures of HgCl2 and BAL were even more potent inducers than either compound alone, i.e., the two agents acted synergistically. To elucidate the mechanism of this unusual interaction, we have measured the effects of BAL and HgClz on the induction of quinone reductase (a representative Phase 2 enzyme) in Hepa lclc7 murine hepatoma cells grown in 96-well microtiter plates (10, 111. This system provides highly quantitative information on the potency of inducers. Parallel measurements also were obtained on the intracellular accumulation of radioactive 203HgC12. Our experiments show that the inducer synergism of mixtures of BAL and HgCl2 depends in major part on more rapid cellular uptake and higher levels of accumulation of chelated mercury(I1). Higher intrinsic inducer potencies of BAL-mercury chelates may also play a role in the observed synergisms.
Experimental Procedures Cell Culture. For the QR assays, Hepa lclc7 murine hepatoma cells were grown in a minimum essential medium containing 10% fetal calf serum that was heated a t 55 "C for 90 min with 1%charcoal and was filtered. The human growth hormone gene expression assays were also carried out on Hepa lclc7 cells grown in Eagle's minimum essential medium containing 10% untreated fetal calf serum. All cells were maintained a t 37 "C in a humidified atmosphere of 5% COz. Chemical Compounds. 2,3-Dimercaptopropanol was supplied by Aldrich (Milwaukee WI). m-Buthionine (RS)-sulfoximine (BSO) was obtained from Sigma (St. Louis, MO). Ethyl glutathione was the gift of B. A. Teicher (Dana-Farber Cancer Institute, Boston, MA). 203HgC12(specific activity: 2.72 mCi/ mg) was obtained from the Buffalo Materials Research Center (SUNY a t Buffalo, Buffalo, NY).Radioactivity was measured in a liquid scintillation counter on samples dissolved in Ecolite Plus (ICN, Costa Mesa, CAI. 203HgC12(obtained in 1 N HC1 solution) was neutralized with 1 N KOH to pH 7. Preparation of Chelates. BAL-mercury chelates were prepared by incubating 1 : l and 20:l molar ratios of [BALY [203HgC12]for 20 min in aqueous methanol solution (50% v/v) at 25 "C. 203HgC12(2.72 mCi/mg of mercury, 2.04 mCi/mL) was diluted 1000-fold with HzO, and nonradioactive HgCl2 was added to give a specific activity of 1223 cpdpmol. Ten micromoles of this diluted zo3HgC12was then mixed with 10 or 200 pmol BAL in 4 mL of 50% aqueous methanol solution. The resulting precipitates (which formed immediately) were processed in two ways. They were either (i) washed successively with water and methanol, dried in a vacuum, and then dissolved in DMSO, or (ii)dissolved in a small volume of DMSO directly, and then chromatographed with acetone on a preparative silica gel plate (2 mm thick Analtech, Newark, DE). The appropriate bands (determined by radioactivity and fluorescence quenching) were collected, the chelates were eluted with DMSO, and the solutions were filtered through 0.5 pm Millipore filters. The mercury concentrations of the solutions were adjusted as desired on the basis of radioactivity measurements. Analytical TLC of
Putzer et al. BAL-mercury chelates was carried out on silica-coated plastic sheets (40 x 80 mm, 0.25 mm thick; Polygram Si1 G / W , Brinkmann, Westbury, NY)with acetone as the developing solvent. Quinone Reductase Assay. The inducer potencies of compounds were determined by measuring QR specific activities in Hepa lclc7 murine hepatoma cells grown in 96-well microtiter plates (10,11). The cells (10 000 per well) were grown for 24 h and then exposed to serial dilutions of inducers in growth medium for 48 h, unless otherwise indicated. The inducers were dissolved in DMSO, and the final concentrations of DMSO were 0.1% by volume. All wells were analyzed 72 h after plating of cells. Duplicate plates were treated identically: one plate was used for determination of QR activity, and the other was stained with Crystal Violet (Sigma, St. Louis, MO), to assess compound toxicity. When mixtures of inducers were used (e.g., BAL and HgC12) the compounds were mixed in DMSO at the appropriate concentration ratio, allowed to incubate a t 25 "C for about 30 min, and then diluted into the growth medium. When two inducers were used sequentially, the medium containing the first inducer was discarded after the indicated interval, the cells were washed with 200 pL per well of Dulbecco's phosphatebuffered saline, and the second inducer was added in growth medium. The ratio of specific activities of QR of treated cells t o those receiving solvent only is defined a s the induction ratio (treateduntreated). A semilogarithmic plot of induction ratio with respect to inducer concentration was then used to determine (by interpolation) the CD values, i.e., the concentration required to double the QR specific activity. The reproducibility of CD values is discussed in the Results section. In experiments involving prior treatment with BSO, or with ethyl GSH,cells were grown for 24 h, and all wells were then treated (including controls) with identical concentrations of one of the indicated agents for a n additional 24 h. Finally, cells were exposed for a further 24 h to serial dilutions of the inducer under study combined with a second treatment with BSO or ethyl glutathione a t the same concentrations used during the first treatment. The cells were then assayed for QR activity. Cellular Uptake and Induction by aosHgClz.Induction experiments with z03HgC12 followed the standard induction protocol for Hepa lclc7 cells in 96-well microtiter plates. Duplicate pairs of microtiter plates were used for the experiments shown in Figures 4 and 7. The following protocols were used. Two solutions (60% DMSO in water, v/v) containing 0.747 mM undiluted 203HgC12alone or mixed with 3.735 mM BAL were added to growth medium so that the final concentrations of 203HgC12and BAL in the medium were 1.25 and 6.25 pM, respectively. Cells were exposed to serial dilutions of these two solutions for 2, 4, and 24 h, as well as for the standard 48-h induction period, and exposure was terminated by replacing the inducer-containing medium with an equal volume of fresh medium. All plates were analyzed 48 h aRer the initial exposure to inducers, as follows: one plate from each duplicate set was washed with growth medium (200 &/well, then 150 pUwell), and half of the wells in these plates were assayed for QR activity, whereas the other half of the wells were assayed for intracellular mercury content. The second set of plates was washed twice with Hanks' solution (200 pUwell each time) and assayed for protein content. For the QR measurements, cells were lysed for 10 min at 37 "C with 50 pUwell of a digitonin solution (0.8% digitonin and 2 mM EDTA, pH 7.8). For radioactivity measurements, cells were lysed with 150 pL of digitonin solutiodwell, and 100 pL of the lysate was assayed for radioactivity. For protein measurements, cells were lysed with 150 p L of digitonin solutiodwell, and 20 pL of the lysate was analyzed for protein content by the bicinchoninic acid method (12). The intracellular concentration of mercury (expressed a s pmol of mercury/mg of protein) then was related to the specific activity of QR. Mercury uptake rates were further explored by the following experiment (see Figure 5). Solutions (50% DMSO in water, v/v) contained 0.536 mM undiluted 203HgC12and different concentration ratios of BAL (1.25, 2.5, 5, and 10). Each of these
Znduction of Quinone Reductase by HgClz and Thiols
solutions was mixed with growth medium to provide final concentrations of 1.34 pM 203HgC12 and the corresponding concentrations of BAL. Cells were exposed to these solutions for 30 min, 6 h, and 24 h, and exposure was stopped by replacing the inducer-containing medium with an equal volume of fresh medium. At 24 h, all plates were washed once with medium (200puwell) and twice with Hanks’ solution (200 puwell each time). Cells were then lysed with 150 pL of digitonin solution per well. Intracellular mercury and protein contents were measured by methods described above. Transient Gene Expression Assays. Human growth hormone was used as a reporter gene (13). The experiments were carried out as described (2), except that the plasmid construct was transfected into Hepa lclc7 cells grown in Eagle’s minimum essential medium containing 10%fetal calf serum for 48 h in the presence of various concentrations of inducers. The plasmid p41-284GH was constructed to contain the 41-bp enhancer element of the mouse GSH transferase Ya gene and the minimal promoter region of that gene (81, linked to the human growth hormone reporter gene (2). The plasmid p284GH is identical except that the 41-bp enhancer element is absent (2).
Results Reproducibility of Inducer Potency Measurements. The interpretation of the experiments to be described requires validation of the reproducibility of measurements of the inducer potencies of compounds. Whereas some variability in potency measurements was observed in different experiments, within a single experiment these values were highly reproducible. Thus when the CD values for HgClz and BAL were measured on eight microtiter plates for each compound in a single experiment, the following values (fSEM) were obtained: for HgCl2 ( n = 81, CD = 0.569 f 0.016pM (range 0.496-0.636); and for BAL ( n = 8),CD = 24.9 f 0.56 pM (range 21.7-27.1).3 The larger inter-experiment variations in CD values (see below) probably result from uncontrollable differences in experimental conditions. Since the conclusions drawn below are invariably based on comparisons made within single experiments and have been confirmed by several repeat experiments, the interexperimental variations in potency do not affect the conclusions. Induction of Quinone Reductase by Heavy Metals and Vicinal Dithiols. Divalent metal cations of Group IIB induce quinone reductase activity in murine hepatoma cells with markedly different potencies that parallel the affinities of these metals for sulfhydryl POUPS (5, 14). Thus, HgClz (CD = 0.61 f 0.028 pM [SEMI; n = 31 experiments) is much more potent than CdClz (CD = 12 pM), whereas ZnClz (CD = 225 pM) is only a very weak inducer. The suggestion that interaction of inducers with sulfhydryl groups is a central event in the mechanism of induction of QR (and by inference of other Phase 2 enzymes) is further strengthened by the high inducer potency of phenylmercuric chloride (CD = 0.04-0.10 pM), which is widely recognized as a very potent and relatively specific sulfhydryl reagent (15).Prior studies have also implicated the participation of sulfhydryl groups in the induction mechanism (2,5-7). The differences between the CD values for BAL given here (24.9 that recorded below (19.6 f 1.0 pM;n = 27) arise from the fact that the first set of values were means obtained from 8 microtiter plates in a single experiment, whereas the second set of values were mean values from 27 different experiments. These results further emphasize the higher degree of consistency of intraexperimental than inter-experimental CD value determinations. & 0.56 p M n = 8) and
0.01
0.1
1
10
100
CONCENTRATION (1M)
Figure 1. Induction of quinone reductase in murine hepatoma cells by a range of concentrations of BAL, HgC12, equimolar mixture of BAL and HgC12, and a synthetic BAL-mercury chelate. The CD values obtained were as follows: for HgC12, CD = 0.72 p M ; for BAL, CD = 24 pM;for an equimolar mixture of BAL and HgC12, CD = 0.28 p M ; and for BAL-mercury chelates isolated by TLC from equimolar mixtures of BAL and HgClz, CD = 0.24 pM. The induction ratios, expressed as the ratio of specific activities of QR in cells treated with inducer to untreated cells, are shown as a function of the logarithm of the
inducer concentration.
Furthermore, the extremely high inducer potency of trivalent arsenical compounds suggests that two vicinal sulfhydryl groups may mediate induction (2, 5). In light of these findings, it seemed surprising that powerful nucleophiles and metal chelators such as BAL (CD = 19.6 f 1.0 pM [SEMI; n = 27)3 and other vicinal dithiols, e.g., 1,2-ethanedithiol(CD = 34 f4.0 pM [SEMI; n = 6) (5) and 2,2’-biphenyldithiol (CD = 30 pM), were also QR inducers. In contrast, monothiols (e.g., 2-mercaptoethanol: CD = 112-220 pM) were only very weak inducers (present work and ref 5). These observations naturally raised the intriguing question of how combinations of HgClz and BAL, which form a bidentate chelate Le., Hg(BAL)l with enormous avidity (KD= lo-& M) (91, might behave as inducers if cells were exposed to mixtures of both agents simultaneously. Inducer Potency of Equimolar Mixtures of H&lz with Thiols. Contrary to the expectation that chelation of HgClz with BAL would abolish inducer activity, we found that an equimolar mixture of these substances was strikingly more potent (CD = 0.28 pM for each inducer) than either HgClz alone (CD = 0.72 pM) or BAL alone (CD = 24 pM). Thus, the potency of HgClz was raised more than 2.5-fold in the equimolar mixture (Figure 1); Le., the two components behaved synergistically. A synthetic BAL-mercury chelate [presumably Hg(BAL), although not of unequivocally established structure], which was isolated by TLC from an equimolar mixture of BAL and HgClZ, had a CD value of 0.24 pM, which is equivalent to that of the solution (CD = 0.28 pM) from which it was derived. In contrast, equimolar mixtures of HgC12 with the intrinsically weak monothiol inducers 2-mercaptoethanol (CD = 112 pM, in this experiment) or 3-mercapto-l,2propanediol (CD = 140pM) were equipotent (mixture CD values were 0.27 and 0.31 pM, respectively) to HgClz alone (CD = 0.31 pM in this experiment). Mercury(I1) combines with monothiols with considerably lower affinities (KD= 10-35-10-38 M) than with vicinal dithiols (KD= 10-42-10-45 M for dicoordinate chelation) (9, 16). Furthermore, an equimolar mixture of phenylmercuric chloride and BAL had a CD value of 0.053 pM, which was similar to the CD value of phenylmercuric chloride alone (CD = 0.060 pM). Notably, organomercurials do
106 Chem. Res. Toxicol., Vol. 8, No. 1, 1995 "I
Putzer et al. CD lulh
I
0.17
I 0.01
1
0.1
1
10
0.13
100
CONCENTRATION (pM)
Figure 2. Induction of quinone reductase in murine hepatoma cells by a range of concentrations of BAL, HgClZ, and mixtures of BAL and HgC12. The effect of increasing the [BALMHgClz] ratio is shown. The CD values obtained were as follows: for HgC12, CD = 0.48 yM; for BAL, CD = 13yM;and for mixtures of BAL and HgClz at molar concentration ratios of 1,5,and 50, CD = 0.10,0.04,and 0.02pM, respectively. The induction ratios are shown as a function of the logarithm of inducer concentration. The concentrationsfor the mixtures are expressed as those of the limiting component (HgC12), and the CD values are expressed with respect to this component.
not form bidentate chelates with BAL (17,181. The lack of synergism between monothiols and HgC12, and between BAL and phenylmercuric chloride, suggests that formation of a bidentate (or possibly higher order) chelate between mercury(I1) and vicinal dithiols is required for synergism of inducer potency. Furthermore, no appreciable synergism was observed between BAL and CdClz or ZnCl2 (results not shown). Significantly, both Zn(I1) and Cd(I1) are much weaker ligands for thiols than is mercury(I1) (14). Effect of the Concentration Ratio of BAL to HgC12 on Inducer Potency. The above experiments did not clarify whether an equimolar mixture of HgC12 and BAL produced maximal synergism of these two inducers. Consequently, we measured the inducer potencies of a series of BAL-HgC12 mixtures containing concentration ratios of [BALY[HgC121 ranging from 1 to 400. Surprisingly, excess BAL strikingly enhanced the already synergistic potency of the equimolar mixture of BAL and HgC12 (Figure 2). This potency enhancement was found to increase with the concentration of excess BAL and to approach a limiting value. As shown in Figure 2, the CD value for HgClz alone (0.48 pM) was decreased 5-fold (CD = 0.1 pM),12-fold (CD = 0.04 p M ) , and 24-fold (CD = 0.02 pM) for [BALHHgC121 mixtures with ratios of 1, 5, and 50, respectively. In these experiments with excess BAL, no correction was required for the intrinsic inducer activity contributed by BAL acting alone, since the concentration of BAL at the CD end point for all mixtures produced negligible inductions (cf. Figure 1). Hence, the presence of excess BAL clearly enhanced the potency of BAL-HgC12 mixtures to a far greater degree than would have been expected on the basis of the intrinsic inducer activity of BALs4 Effect of Sequential Treatment with HgC12 and BAL on Inducer Potency of HgClZ. Sequential rather than simultaneous treatment with HgClz and BAL has provided further insight into the mechanism by which BAL enhances the inducer potency of HgC12. When cells were treated first with HgC12 for 4 or 24 h, washed with phosphate-buffered saline to remove residual extra4Aa expected, excess BAZ, enhanced the inducer potency of the synthetic BAL-mercury chelate,to the same extent as the corresponding BAL-HgClz mixtures (results not shown).
4
24
48
TREATMENT TIME (h)
Figure 3. Effect of sequential treatment of murine hepatoma cells, first with HgClz and then with BAL, on the potency of HgClz as an inducer of QR. The cells were washed between treatments.The inducer potencies were determined on all wells of 96-well microtiter plates after a total treatment period of 48 h. Plates 1 and 2 were treated with HgClz for the first 4 or 24 h, respectively. Plates 9 and 10 were treated with BAL for the last 24 or 44 h, respectively. Plates 3,4, and 5 were treated for the first 4 h with HgClz and for the remaining 44 h with BAL at concentrations that were 1,10, or 25 times those of the prior HgClz treatment (as indicated).Plates 6,7, and 8 were exposed to HgClz for the first 24 h, and for the remaining 24 h were treated with BAL at concentrations that were 1,10, or 25 times those of the prior HgClz treatment. The CD values shown were obtained by plotting the induction ratios for QR as a function of the logarithm of inducer concentration (HgC12 in the case of mixtures).
cellular HgC12, and then treated with BAL for 44 or 24 h, respectively, a clear-cut synergism of potency was also observed (Figure 3). For instance, when cells were treated for 24 h with HgC12, washed, and then exposed to BAL for 24 h, under conditions where the concentration of BAL was 10 times that of HgC12, the inducer potency of HgC12 was increased nearly 3-fold (i.e., the CD value decreased from 0.47 to 0.17 pM). If the initial exposure interval to HgClz was shortened to 4 h and the treatment period with BAL was extended to 44 h, the inducer potency measured at 48 h of induction was also markedly increased. Sigmficantly, the magnitude of the potency enhancement in the sequential treatment experiments also increased as the concentration of the excess BAL (the second component) was raised, and also approached a limiting value. These findings suggest that the mechanism by which BAL enhances the activity of mercury may be similar when the two agents are administered simultaneously or sequentially. Nevertheless, these experiments imply that the enhancing effect of BAL on the inducer potency of HgClz may not result solely from the increased uptake of mercury from the medium (as is shown below with 203HgC12),but that interactions between these components also can occur after the HgC12 enters the cells and that the resulting product(s), presumably mercury-BAL chelate(s1,may be intrinsically more potent as inducers than HgC12 alone. BAL Enhances the Cellular Uptake of Mercury. Experiments with 203HgC12provided further insight into the mechanism by which BAL enhanced the inducer potency of mercury(I1). Murine hepatoma cells, grown for the standard 24-h period, were exposed to a series of concentrations of z0sHgC12alone or in mixtures with BAL ([BALY[HgC12] = 5) for periods of 2,4,24, and 48 h. The medium was replaced, and incubation was continued for an additional 46, 44, 24, or 0 h, respectively. The intracellular mercury concentrations were then mea-
Chem. Res. Toxicol., Vol. 8, No. 1, 1995 107
Induction of Quinone Reductase by HgClz and Thiols
l4 !I
c
30
E
cn I
3
g>
20
Pa
10
E
:
100
'
z
0.15
0'
0.15
0.3
EXTRACELLULAR [Hg] (pM)
0.3
Figure 4. Effect of BAL on the concentration-dependent intracellular uptake of z03HgC12by murine hepatoma cells. The intracellular radioactive mercury concentration (expressed as pmol of mercurylmg of cellular protein) was measured 48 h after the addition of the inducers. The intracellular mercury concentration is shown as a function of external HgClz concentration aRer 2,4, 24,or 48 h of exposure to HgClz alone or a BALHgCl2 mixture in which the molar ratio of [BALy[HgC12]was 5. 800
1
lo 5
2.5
./-
1.25 HgCl2 alone
0
10
20
30
EXPOSURE TIME (h)
Figure 5. Effect of increasing concentrations of BAL on the cellular uptake of z03HgC12by murine hepatoma cells. All cells were exposed to 0.67pM z03HgC12for 30 min, 6 or 24 h, either alone or in the presence of BAL at molar concentration ratios of [BALy[HgC121 of 1.25,2.5, 5, or 10. The cells were then washed with medium once and incubated for a total of 24 h. Intracellularradioactive mercury concentrations, expressed as pmol of mercury/mg of cellular protein, were measured at 24 h. The intracellular Hg(I1)concentration for HgClz alone was 5.70, 37.3, and 116 pmovmg of protein after 30 min, 6 h, or 24 h exposure to HgC12, respectively.
sured. Figure 4 shows that the intracellular mercury accumulation at each time point increased linearly with the external concentration of HgC12. The intracellular mercury concentration also increased with time and reached a plateau between 24 and 48 h. At each time point, the presence of a 5-fold excess of BAL greatly augmented (5-8 times) the intracellular accumulation of mercury. The stimulatory effect of increasing the BAL concentration on the accumulation of mercury from a constant external HgCl2 concentration pool (0.67 pM) is shown in Figure 5. The cellular uptake of mercury rises and approaches a plateau as the concentration ratio of [BALy[HgC12] is increased. For example, after 6 h exposure, mercury uptake is 10 times greater for a 5:l mixture than for HgC12 alone, and almost 20 times greater for a 1 O : l mixture. This enhancement is presumably mediated, a t least in part, by formation of the dicoordinate chelate species, which would be expected to facilitate the cellular uptake of mercury by increasing
2
4
6
8
I 0 1 2 1 4
STRIP NUMBER
Figure 6. Analytical TLC profiles showing the formation of two BAL-mercury chelates. Mixtures of BAL and 203HgC12at concentration ratios of 1, 5, 10,and 20 were prepared in 50% aqueous DMSO and chromatographed on plastic silica gel TLC plates (40 x 80 mm) with acetone. The plates were cut horizontally into sixteen 5-mm strips, and the radioactivity of each strip was determined. The fraction (%) of the total radioactivity found in each strip is shown. Note the presence of two radioactive regions with Rf values of 0.4(strips 6-9) and 0.8 (strips 11-14], respectively. lipophilicity. However, since mercury uptake continued to increase steadily when the concentration ratio ([BALV [Hgcl~])was raised from 5 to 10, dicoordinate chelation alone (which probably occurs nearly quantitatively in equimolar mixtures) is insufficient to account fully for these results, and additional interactions between BAL and HgCl2 are probably involved.
Reversible Formation of a Second Chelate with Excess BAL. The increase in cellular mercury uptake in the presence of excess BAL suggests the formation of a higher order chelate that is transported more readily than that formed in the equimolar mixture [presumably Hg(BAL)I. Direct TLC analysis of DMSO solutions of the precipitates formed in mixtures of BAL and 203HgC12 revealed the formation of two chelates in DMSO solution: the first had a higher R f value (0.8) and predominated in the 1:lBAL:HgC12 mixture. Upon raising the concentration of BAL, the first species progressively disappeared while the concentration of a slower migrating species ( R f = 0.4) increased proportionately (Figure 6). A 20:l mixture ([BALY[HgC12]), for example, when applied to a TLC plate, contained the second ( R f = 0.4) species exclusively. In an effort to isolate these chelates, the precipitates obtained from a 20:l mixture ([BALY [HgCld were washed extensively with water and methanol, and the residues were dissolved in DMSO and analyzed by TLC. Only the more rapidly migrating (Rf = 0.8) species was detected in such washed chelate preparations. Thus, removal of excess BAL converts the second species into the first; furthermore, subsequent addition of BAL (approximately a 20-fold molar excess) to this isolated material reconverted the rapidly migrating chelate once again to the slower migrating species. The reversible formation of the second species suggests that it involves additional but weaker BAL-mercury interaction( s). These findings are consistent with observations of Casas and Jones (91, who found that the first two dissociation constants for BAL-mercury chelation were KD= M for dissociation of Hg(BAL) and KD= lop7M for dissociation of Hg(BAL)22-(association rather than dissociation constants were reported). The first complex (Rf= 0.8) is clearly the water-insoluble Hg(BAL), which has been the only species previously isolated from mixtures of HgCl2 and BAL (17-21 ). It is possible that the second species (Rf = 0.4) which is refractory to
108 Chem. Res. Toxicol., Vol. 8, No. 1, 1995
6-
4-
2-
INTRACELLULAR Hg (pmollmg protein)
Figure 7. Induction of QR in murine hepatoma cells by zosHgClzalone (lower graph) and by a mixture of BAL and z03HgC12(concentrationratio 5:l)(upper graph), after standard 48-h induction periods. The preparation of inducer-containing medium is described in the text. Plates for measuring radioactive mercury and QR activity were washed with medium (200 pUwell, then 150 pUwell). Plates for measuring protein content were washed twice with Hanks’ solution (200pUwell). All cells were lysed with digitonin and analyzed for specific activity of QR and mercury content. Note that at each concentration of intracellular mercury, the chelated species produce greater induction.
isolation is a higher order chelate, e.g., Hg(BAL)z2-. These results raise the possibility that the presence of excess BAL in vivo might promote the formation of a more permeable chelate(s), thereby accounting for the enhanced mercury uptake observed in the presence of excess BAL (cf. Figures 4 and 5). If the higher order chelate is charged [e.g., Hg(BAL)z2-1, as suggested (9) it is not obvious why it would enter cells more readily than HgC12, which is apparently transported into cells as an uncharged species (22). Intracellular Inducer Potencies of Chelated and Unchelated Mercury. The striking enhancement of cellular mercury uptake from mixtures containing excess BAL (relative to that of the equimolar BAL-HgC12 mixture or the isolated chelate) is roughly comparable to the enhancement of inducer activity by excess BAL. For example, after 24-h exposure the cellular uptake of mercury from a 5:l mixture of BAL and HgClz is about 5 times greater than that of HgC12 alone (Figure 51, whereas the inducer potency of this mixture is about 12fold higher. In the following experiment, we attempted to relate the intracellular mercury concentration at the time of measurement of the specific activity of QR to the degree of induction of the enzyme. Murine hepatoma cells were exposed to a series of concentrations of either 203HgC12or of a mixture of 203HgC12with BAL at a molar concentration ratio of 5 ([BALy[HgC121. After the standard 48-h induction period, the intracellular mercury concentration (pmol of mercury/mg of protein) and the QR induction ratio were determined (Figure 7). The intracellular concentration of chelated mercury required to double QR specific activity after 48-h exposure, i.e., the “intracellular CD value”, is one-half that of the intracellular concentration of unchelated HgCl2 required to exert the same effect (CD = 15 and 30 pmoYmg of protein, respectively) (Figure 7). It is tempting to conclude from these experiments that chelated mercury is intrinsically a more potent inducer than unchelated mercury. However, this is not necessarily a correct conclusion since the kinetics of mercury uptake and the intracellular concentrations are quite different in the presence and absence of BAL. It is possible that the apparently higher potency of chelated mercury can be
Putzer et al.
E
1.4
BAL
HgCI,
BAUHgCI, (1:l)
BAUHgCI, (2:l)
BAUHgCI, (5:l)
Figure 8. Effect of treatment with buthionine sulfoximine on the potencies of BAL, HgC12, and BAL-HgC12 mixtures (concentration ratios 1, 2, or 5) as inducers of QR in murine hepatoma cells. The cells were treated with 100pM BSO during the entire 48-h induction period and with the inducers only during the last 24 h of that period. The CD values are shown (in pM concentrations above each bar) for cells treated (0) or untreated (W with BSO. Note the logarithmic scale. Treatment with 100 pM BSO alone raised the QR specific activity to 1.48 times the untreated values. The induction ratios have been related to control cells treated with BSO. Note that treatment with BSO increased the potency of each inducer and of the mixtures by a factor of 2-3. ascribed to a more rapid uptake and a higher earlier intracellular level. The intracellular concentration- time curves are quite different in the presence and absence of BAL, and how this might affect enzyme induction kinetics is not known (cf. Figure 5). Effects of Manipulation of Intracellular Glutathione Levels on QR Induction. GSH is the principal intracellular nonprotein thiol (23). Manipulation of intracellular GSH levels contributed further insight into our findings. BSO rapidly depresses cellular GSH levels by inhibiting its synthesis (23). For example, a 24-h treatment of murine hepatoma cells with 100 pM BSO depressed GSH levels by about When these cells were treated with 100 pM BSO during the entire 48-h induction period and were exposed to HgC12, BAL, or their mixtures during the last 24 h, the potencies of these inducers increased (Figure 8). Thus the CD values for BAL, HgC12, and mixtures of these components in molar ratios of 1,2, or 5 ([BALy[HgC121) all decreased by a factor of 2-3. In contrast, treatment of Hepa lclc7 cells with the ethyl ester of GSH, which elevates cellular GSH levels (231, depressed the inducer potency of the BALmercury chelate (Figure 9). In the last experiment, the inducer potency of the chelate was studied, rather than that of mixtures of BAL and HgC12, in order to obviate the problems of direct interactions between HgC12 and ethyl glutathione. The potency of the BAL-mercury chelate decreased as the concentration of ethyl GSH to which the cells were exposed was raised. These observations suggest that the thiol group of glutathione, and perhaps other cellular thiol groups, attenuate the activity of these inducers, perhaps by nonspecific scavenging reactions, thereby lowering the effective inducer concentrations available for signaling of enzyme induction. Bergelson et al. (24)have recently shown that GSH levels exercise control over chemical induction of chloramphenicol acetyltransferase activity from the EpRE element of the mouse GSH transferase Ya gene. S. R. Spencer and P. Talalay, unpublished observations.
Chem. Res. Toxicol., Vol. 8, No. 1, 1995 109
Induction of Quinone Reductase by HgClz and Thiols 4
/
HgC$ (QRf
g =
5 3E2 3 Z 0 3
F $
2n 3o 2
ze 8 OO
0.5
1
8
1.5
[ETHYL-GLUTATHIONE] (mM)
Figure 9. Effect of treatment of murine hepatoma cells with ethyl glutathione on the potency of a synthetic BAL-mercury chelate as inducer of QR. The cells were treated with 0, 0.5,
1.0, and 1.5 mM ethyl glutathione during the entire 48-h induction period and with the BAL-Hg chelate only during the last 24 h of that period. Ethyl glutathione raised the QR specific activities slightly in a concentration-dependent manner (1.3fold at 1.5 mM). The inductions have been related to control cells treated with ethyl glutathione only. Note that when the intracellular GSH levels were raised, the inducer potency of the BAL-mercury complex decreased.
Molecular Regulation of Induction Synergism. The inductions of QR and of GSH transferases are regulated by a common enhancer sequence (ARE or EpRE elements) that resides upstream of the genes of these enzymes and responds to inducers (2, 5, 25, 26). This regulatory element (contained within a 41-bp sequence of the 5'-upstream region of the mouse glutathione transferase Ya gene) together with the promoter region of this gene (8)was inserted into a plasmid (p41284GH) expressing human growth hormone as a reporter gene (2). The potency of a wide variety of inducers in stimulating growth hormone production in hepatoma cells transfected with the above construct correlated very closely with the potency of the same inducers in raising QR activity in Hepa lclc7 cells (2). In no case was the plasmid lacking the 41-bp enhancer (p284GH) inducible. Among the many inducers examined, this correlation was observed for both HgC12 and BAL,suggesting that all inducers, irrespective of chemical type, regulate enzyme induction through a common genetic enhancer element. In the present experiments, we also observed synergism when the above construct was exposed to an equimolar mixture of BAL and HgClZ, confirming that the mechanism of synergism indeed involves events preceding transcriptional activation, rather than the transcriptional process itself. Figure 10 compares the induction ratios (treatedhntreated) as a function of concentration for QR in Hepa lclc7 cells with growth hormone production in the same cells transfected with the aforementioned construct. The potencies in these systems were as follows: for HgC12 (CDQR= 0.60 pM;CDGH= 0.59 pM), for BAL (CDQR= 13 pM; CDGH= 22 pM), and for an equimolar mixture of BAL and HgClz (CDQR= 0.1 pM; CDGH= 0.06 pM,for each component). On the basis of the relatively close agreement in inducer potencies, we conclude that the synergism of the inducer response is mediated through the same enhancer element involved in the intrinsic inducer response to both HgC12 and BAL. Discussion and Conclusions Our experiments provide insight into two aspects of the mechanisms of the monofunctional induction of Phase 2 enzymes: the participation of thiol groups in this process, and the potent synergism between mercury(I1)
1:1
(GH,q++~ /
(Qj
i
-1
/?
7
