Toxic Heavy Metal Ions Activate the Heme-regulated Eukaryotic Initiation Factor-2a Kinase by Inhibiting the Capacity of Hemin-supplemented Reticulocyte Lysates to Reduce Disulfide Bonds*

Addition of toxic heavy metal ions (Cd2+, Hg2+, and Pb2+) to hemin-supplemented rabbit reticulocyte lysate brings about the activation of the heme-regulated eu- karyotic initiation factor 2a kinase (HRI) and the inhibition of protein chain initiation. In this report we examined the effects of monothiol and dithiol com- pounds, metal ion-chelating agents, and metallothi- oneins (MT) on metal ion-induced inhibition of protein synthesis. The dithiol compounds dithiothreitol and 2,3-dimercaptopropane sulfonic acid prevented and relieved the inhibition of protein synthesis caused by Cd2+ and Hg2+ in hemin-supplemented lysates, but the monothiol compounds 2-mercaptoethanol, cysteamine, D-(-)penicillamine, and glutathione had no effect. The inhibition of protein synthesis caused by Cd2+ was reversed by the addition of excess EDTA but not by the addition of excess nitrilotriacetic acid. Toxic heavy metal ions inhibited the capacity of hemin-supple- mented lysate to reduce disulfide bonds. Addition of excess EDTA to Cd2+-inhibited lysates restored the capacity of the lysate to reduce disulfide bonds and inhibited the phosphorylation of eukaryotic initiation factor eIF-2. MTs and their apoproteins

Addition of toxic heavy metal ions (Cd2+, Hg2+, and Pb2+) to hemin-supplemented rabbit reticulocyte lysate brings about the activation of the heme-regulated eukaryotic initiation factor 2a kinase (HRI) and the inhibition of protein chain initiation. In this report we examined the effects of monothiol and dithiol compounds, metal ion-chelating agents, and metallothioneins (MT) on metal ion-induced inhibition of protein synthesis. The dithiol compounds dithiothreitol and 2,3-dimercaptopropane sulfonic acid prevented and relieved the inhibition of protein synthesis caused by Cd2+ and Hg2+ in hemin-supplemented lysates, but the monothiol compounds 2-mercaptoethanol, cysteamine, D-(-)penicillamine, and glutathione had no effect. The inhibition of protein synthesis caused by Cd2+ was reversed by the addition of excess EDTA but not by the addition of excess nitrilotriacetic acid. Toxic heavy metal ions inhibited the capacity of hemin-supplemented lysate to reduce disulfide bonds. Addition of excess EDTA to Cd2+-inhibited lysates restored the capacity of the lysate to reduce disulfide bonds and inhibited the phosphorylation of eukaryotic initiation factor eIF-2. MTs and their apoproteins (apoMTs) inhibited the activation of HRI and protected protein synthesis from inhibition by Cd2+, Hg2+, and Pb2+. Addition of apoMTs to heavy metal ion-inhibited lysates restored the capacity of lysates to reduce disulfide bonds. The restoration of the lysate's thioredoxin/thioredoxin reductase activity was accompanied by the inactivation of HRI and the resumption of protein synthesis, indicating that apoMTs can "detoxify" metal ions already bound to proteins. Several observations presented in this report suggest that the binding of metal ions to the a-domain of MT is responsible for the ability of MT to sequester bound metal in a non-toxic form. Addition of glucose 6-phosphate or NADPH had no effect on protein synthesis in metal ion-inhibited lysates, and NADPH concentrations in Cd2+-inhibited and hemin-supplemented control lysates were equivalent. The data suggest that the metal ions cause the inhibition of protein synthesis by binding to vicinal sulfhydryl groups present in some critical protein(s), possibly the dithiols present in the active site of thio-* This work was supported by National Institutes of Health Grant R29ES-04299 and the Oklahoma State University Agricultural Experiment Station. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertkement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. This is Oklahoma State University Agricultural Experiment Station journal article .
$ To whom correspondence should be addressed 454 PS 11, Oklahoma State University, Stillwater OK 74078-0454. redoxin and (or) thioredoxin reductase, which leads to the activation of HRI.
Toxic heavy metal ions ( i e . Cd2+, Hg2+, and Pb2+) have been found to inhibit protein synthesis in hemin-supplemented reticulocyte lysates with biphasic kinetics (1). The shut off of protein synthesis was due to an inhibition of protein chain initiation and occurred in conjunction with the phosphorylation of the a-subunit of eukaryotic initiation factor (eIF)' 2, the loss of eIF-2B activity (also currently designated RF, reversing factor (2), and GEF, guanine nucleotide exchange factor (3)), and the disaggregation of polyribosomes. Inhibition of protein synthesis initiation in hemin-supplemented lysates also occurs in response to a variety of other conditions (reviewed in Refs. 4-6), which include addition of oxidants or sulfhydryl reactive agents, and heat stress (7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20). The data indicate that the inhibition of protein synthesis, which is observed under these conditions, apparently occurs due to the activation of the heme-regulated eIF-a kinase (HRI) or a protein kinase with similar antigenic and physical properties (1,8,11,21). The mechanism by which HRI becomes activated in the presence of hemin is not yet well understood, but maintenance of protein synthesis in heminsupplemented reticulocyte lysates has been shown to require the presence of certain sugar phosphates (e.g. Glc-6-P), a NADPH-regenerating system, and a functional thioredoxinl thioredoxin reductase system (7)(8)(9)(10)(11).
Metallothioneins (MT) are low molecular weight, cysteinerich, high affinity metal-binding proteins (reviewed in Refs. 22 and 23). The observations that MT have the ability to bind to heavy metal ions, and that their synthesis is induced by the ions, suggest that MTs play a role in cellular metal metabolism, homeostasis, or detoxification. MT synthesis is also induced in certain tissues and cell types in response to a variety of other chemical and physical stresses, suggesting that MT may have additional biological roles. Regulation of intracellular oxidation-reduction potential or activated oxygen detoxification are some of the other possible functions which have been proposed for MT (22,23). Gene transfer studies have directly demonstrated the ability of MT to protect cells from the toxic effect of heavy metal ions. However, The abbreviations used are: eIF, eukaryotic initiation factor; MTI or 11, metallothionein I or 11; apoMTI or 11, apometallothionein I or 11; eIF-201, 38,000-dalton a-subunit of eIF-2; HRI, heme-regulated eIF-2a kinase; DMPS, 2, 3-dimercaptopropane sulfonic acid DTT, dithiothreitol; DTPA, diethylenetriamine pentaacetic acid NTA, nitrilotriacetic acid Glc-6-P, glucose 6-phosphate; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; hsp, heat shock protein.
the observation that the accumulation of excessive amounts of Cd2+-containing MT appears to be the cause of renal damage in chronic Cd2+ poisoning has cast some doubt on the importance of MT as a specific and effective defense mechanism against Cd2+ exposure in animals (24).
The observation that the relative order of potency of the metal ions in bringing about the inhibition of protein synthesis (Hg2+>Pb2+>Cd2+>>>Zn2+) (1) is similar to the established association constants for the metals with cysteine and with sulfides (H$+>Pb >Cd2+>Zn2+) (25)(26)(27), suggests that the metals are acting through binding to sulfhydryls. Furthermore, the inhibition of protein synthesis by arsenite and the inhibition of protein synthesis by Cd2+ at concentrations 30-50 times lower than those required for Zn2+ (l), imply that this inhibition is brought about through the interaction of the metal ions with spatially adjacent (vicinal) sulfhydryls (28,29). In this report we have compared the ability of monothiol compounds, dithiol compounds, other metal-chelating agents, and MTs to prevent or relieve the inhibition of protein synthesis by heavy metal ions in hemin-supplemented reticulocyte lysates. The data presented here fulfill all the additional basic criteria suggested for demonstrating the presence of functional dithiols in the active sites of proteins (30-32). The observations that toxic heavy metal ions have no effect on NADPH levels, yet inhibit the capacity of hemin-supplemented lysates to reduce disulfide bonds, suggests that the site of action of the metal ions may be the vicinal sulfhydryl groups present in the active site of thioredoxin and (or) thioredoxin reductase (33, 34). Addition of MT and their apoproteins (apoMT) to hemin-supplemented lysates protected protein synthesis from inhibition by toxic heavy metals. Furthermore, apoMTs were observed to have the ability to "detoxify" the metal ions, restoring the capacity to reduce disulfide bonds, and protein synthesis in previously inhibited lysates.
Protein Synthesis in the Reticulocyte Lysate-Reticulocyte lysates were prepared from anemic rabbits as described (35), using buffered saline containing 5 mM glucose to wash the reticulocytes prior to their lysis. The inclusion of glucose in the wash buffer yields reticulocytes replete in glucose. Reticulocyte lysates prepared in this manner are capable of maintaining high constant levels of NADPH relative to NADP in the absence of added Glc-6-P, and show no additional increase in NADPH levels upon the addition of Glc-6-P (see text). This is in contrast to the properties of lysate prepared without glucose in the wash buffer, as reported by Ernst et al. (7). Protein synthesis was measured by the incorporation of ['4C]leucine into protein at 30 "C in standard reticulocyte reaction mixtures containing 20 p~ hemin.HC1 as described (7,35). Where indicated protein synthesis was inhibited by the addition of the metal ions Cd(OAc)p, Pb(OAc)z, Zn(OAc)n, and HgClz prepared as solutions in deionized water. Tritrations to determine the concentrations of H e , Pb", Cd2+ and Zn2+, which were required to inhibit protein synthesis by 50% in 30 min compared to a hemin-supplemented control (I&), have been carried out in a number of different lysates. As noted previously (I), different lysate preparations were found to vary some-what in their sensitivity to inhibition by the various ions, but the variation occurred for all the ions in the same direction (more or less sensitive) and the relative order of potency always remained the same: HgZf (1-4 pM) > Pb2+ (5-20 pM) > Cd" (8-25 pM) > Zn2+ (0.25-1.0 mM). Equivalent results were obtained in experiments using the different lysates, when metal ion concentrations were used which gave equivalent degrees of inhibition. For all the protection assays, chelating agents, monothiol or dithiol compounds, and MTs or apoMTs were added to the protein synthesis mixes just prior to the beginning of the incubation, after the addition of the metal ion. The apoMTs, as prepared, were sufficiently concentrated, such that addition of an equivalent volume of 0.008 N HCl to the hemin-supplemented control had no effect on protein synthesis. Therefore, to prevent oxidation, the apoMT solutions were not neutralized prior to their addition to protein synthesis mixtures. Changes in values (AICs0), where reported for comparative purposes, were determined using the same reticulocyte lysate preparation.
Preparation of Apornetallothioneins-Metallothioneins were dissolved in deionized H20 (pH 2, adjusted with the addition of HC1) at a concentration of -10 mg/ml. The solutions were dialyzed against deionized HzO (pH 2) (36) in Spectropor 7 dialysis tubing (Spectrum Medical Supplies, Inc.) for 24 h at 4 "C. The solutions (-900 pl) were then loaded onto Sephadex G-25 fine columns (37) (4.5 ml) previously equilibrated with 0.008 N HCl, and fractions (-1 ml) were eluted with 0.008 N HCl and monitored for absorbance at 215 nm. Removal of the metal ions from metallothioneins was confirmed by the loss from the absorption spectrum of a broad shoulder at 250 nm (36). The concentration of apometallothionein was estimated by saturating a portion of each fraction (diluted in 50 mM Tris-HCl, pH 6.8) with Cd(OAc)z while monitoring the absorbance at 250 nm as described by Kage and Vallee (36). The MTI preparation used in the experiments contained 64.8 pg of Cd and 7.5 pg of Zn/mg protein and the MTII preparation contained 64.4 pg of Cd and 18 pg of Zn/mg protein. Two MTI preparations from Sigma, containing 77.7 and 76.4 pg of Cd/mg of protein, were found to have little ability to protect protein synthesis from metal inhibition. These MTI preparations were more inhibitory to protein synthesis when added alone to incubation mixtures than was the MTI preparation with the lower Cd content. Removal of the metal ion from these preparations, as described above, yielded an active apoMTI preparation.
Phosphoprotein Profiles-Protein synthesizing lysates were pulsed with [y-32P]ATP (0.5 mCi/pl protein synthesis mix final concentration) at 10-14 or 25-29 min (during and after shut off). Aliquots, containing an equivalent of 2.5 pl of protein synthesis mix, were denatured in SDS sample buffer and analyzed by SDS-PAGE on 8% gels (37.5:l acrylamide/bis) followed by autoradiography as described (38).
Assay for the Capacity of Lysates to Reduce Disulfide Bonds-The ability of lysates to reduce the disulfide bonds present in insulin was measured by a modification of the method described by Jackson et al. (11). Lysates were incubated under protein synthesis conditions at 30 "C. At the times indicated in the figure legends, 10-pl samples were added to 7 pl of a 1.2 mg/ml insulin solution. After incubation for 15 min at 30 "C, reaction mixtures were diluted with 32 pl of buffer containing 20 mM Tris-HCl, pH 7.5, 1 mM Mg(OAc)?, and 50 mM NaCl, and 1 p1 of a 50 mM Thiolyte MB solution (prepared in acetonitrile) was then added. Assay mixtures were incubated for another 5 min at 30 "C, and the reactions were then terminated by the addition of 20 p1 of 4 times concentrated SDS sample buffer. Aliquots (25 pl) of each sample were analyzed by SDS-PAGE on 20% gels. After electrophoresis, gels were fixed by soaking for 30 min in a solution containing 40% methanol and 10% acetic acid. Gels were then illuminated with a long wavelength ultraviolet light box (Chromato-vue, Ultra-Violet Products, Inc.) and fluorescent bands were photographed with Kodak Panatomic X film through a yellow wratten gelatin filter No. 8 (cut off = 460 nm) using an exposure time of 2 min (39).
Measurement of NADPH and NADP-The levels of NADPH and NADP+ in hemin-supplemented and Cd2+-inhibited lysates were measured by the method described by Ernst et al. (7). Flourescence of samples were measured in a 1 X 1-cm quartz cuvette using a Perkin-Elmer Cetus 650-40 Fluorescence Spectrophotometer with an excitation wavelength of 340 nm and an emission wavelength of 460 nm, and slit widths of 5 and 20 nm, respectively. Determinations were done in triplicate, and the experiments were repeated at least twice.

Effects of Monothiol and Dithiol Compounds and Metal
Chelating Agents on Protein Synthesis in Heavy Metal Ioninhibited Reticulocyte Lysates-DTT was found to protect protein synthesis in hemin-supplemented lysates from inhibition by Cd" and Hg' (Fig. 1, A and B ) . Furthermore, addition of DTT to lysates containing Cd2+ or H P , after the shutoff of protein synthesis, was found to restore protein synthesis to near control rates. We subsequently examined the ability of a variety of monothiol and dithiol compounds to protect protein synthesis from inhibition by Cd2+ or H P . The monothiol compounds cysteamine, 2-mercaptoethanol, D(-)-penicillamine and glutathione at concentrations of 1 mM did not protect protein synthesis in hemin-supplemented lysates from inhibition by Cd" or Hg+ (data not shown). However, protein synthesis was partially or fully protected from inhibition by Cd2+ when 50 p~ DTT, DMPS, or diethyl dithiocarbamic acid was added to the lysate ( Table I). Furthermore, the delayed addition of these dithiol compounds to Cd2+-inhibited lysates at 10 min was observed to stimulate the rate of protein synthesis, indicating that these compounds have the ability to partially or fully reverse Cd2+-induced inhibition of protein synthesis (Table I). DTT and DMPS, but not diethyl dithiocarbamic acid, were found to have similar effects on protein synthesis when added to hemin-supplemented lysates containing H g + (Table I). Monothiol compounds similarly added at 10 min to Cd2+-or Hg2+-inhibited lysates had no stimilatory effect on protein synthesis (data not shown). Concentrations of DMPS and DCCA higher than 50 pM were found to be inhibitory to protein synthesis, such that the ability of higher concentrations of these compounds to fully restore protein synthesis in Cd2+-or Hg+-inhibited lysates could not be accessed. The reversal of Cd2+ inhibition by dithiols, but not by monothiols, has been traditionally accepted as a criterion for the presence of an active-site dithiol within a protein (30, 31). However, since thiols can readily Aliquots (5 pl) were taken at the times indicated in the figure and ["Clleucine incorporated into acid-precipitable protein was determined as described under "Experimental Procedures."

TABLE I
Effect of dithiol compounds on protein synthesis in Cd2+-and HP-inhibited reticulocyte lysates Protein synthesis mixture (40 pl) containing 20 p~ hemin. HC1 were incubated at 30 "C as follows: with no additions (control); with 12.5 p~ Cd(OAc),, (Cd"); with 2.5 pM HgCl,, ( H e ) ; with 50 p~ of the dithiol compound added alone (DTT, DMPS, or diethyl dithiocarbonic acid (DDCA)), or with the metal ion, plus the addition of the dithiol compound at the beginning of the incubation (0 min) or at the time when protein synthesis begins to shut off (at 10 min for Cdz+-inhibited lysates and at 5 min for H$+-inhibited lysates). Aliquots (5 pl) of incubation mixtures were taken at 5, 10, 15, 20, and 30 min. The data are reported as counts/min (cpm X of ["C] leucine incorporated into acid precipitable protein in 30 min from a 5-pl aliquot of the incubation mixture (0 min addition of dithiol compound), or as the rate of ['4C]leucine incorporation into acidprecipitable protein (cpm/min) between 20 and 30 min of incubation (10 or 5 min addition of dithiol compound). reduce protein disulfides and exert independent effects on proteins, Gaber and Fluharty (32) have suggested an additional criterion (the reversal of Cd" inhibition by a 10-fold excess of EDTA but not NTA) be met to establish the presence of an active-site dithiol. The binding affinity of EDTA for Cd2+ is close to the estimated binding affinity of dithiols to Cd", while the binding affinity of NTA for Cd2+ is approximately the same as the binding affinity of monothiols to Cd2+ (25,27,40). EDTA and DTPA, but not NTA, effectively protected hemin-supplemented lysates from inhibition of protein synthesis by Cd2+ (Table 11). In addition, EDTA and DTPA were able to reverse the inhibition of protein synthesis upon their addition to Cd2+-inhibited lysates at 10 min (Table  11). Little effect of EDTA, DTPA, or NTA was observed upon their addition to lysates in which protein synthesis was inhibited by the addition of H g + (Fig. 1, other data not shown). At very low Hg2+ concentrations (0.5 p M ) slight protection by DTPA and EDTA was observed (data not shown). Addition of EDTA or DTPA (but not NTA) to Pb2+-inhibited lysates was also observed to protect and restore protein synthesis (data not shown). DTPA, which has a 100 times higher binding affinity than EDTA for Cd2+ (40), was also more effective than EDTA in protecting protein synthesis from inhibition by Cd". Only a 2-fold excess of DPTA over Cd2+ was required to protect hemin-supplemented lysates from inhibition of protein synthesis; 50 p~ DTPA was observed to be capable of fully protecting protein synthesis from inhibition by 25 p~ Cd2+ (data not shown).
Effects of MTI and II, and ApoMTI and II on Heavy Metal Ion-induced Inhibition of Protein Synthesis in Hemin-supple-of Disulfide Bond Reduction rnented Lysates-Titrations of metal ions into hemin-supplemented reticulocyte lysate were carried out to determine the concentrations H P , Pb", and Cd2+, which inhibit protein synthesis by 50% of the hemin-supplemented control (I&) in a 30-min incubation (Fig. 2). Rabbit liver MTI (20 p~) and MTII (30 p~) added at the beginning of the incubation were found to protect hemin-supplemented lysates from inhibition by low levels of H$+ and Pb", and somewhat higher levels of Cd2+ (Fig. 2, A-C). Based on their metal content, 20 pM MTI and 30 p~ MTII were estimated to contain equivalent amounts of unoccupied metal-binding sites (-56 p~) . The change in the ICso (AICsO) for the inhibition of protein synthesis by H$+, Pb2+, and Cd2+ in the presence of 20 p~ MTI were calculated to be approximately 7, 8, and 15 pM, respectively. In the presence of 30 p~ MTII, the AICso for H P , Pb2+, and Cd2+ inhibition were approximately 3,6, and 13 p~, respectively. Two other MTI preparations obtained from   Sigma, containing approximately 20% more bound Cd2+ (77.7 and 76.4 pg of Cd/mg of protein) were found to have essentially no capacity to protect protein synthesis from inhibition by Pb2+ and had a greatly reduced capacity to protect protein synthesis from inhibition in the presence of Hg+ (AICso -1 pm) and Cd2+ (AICso -3 pm) (data not shown).
Bound Cd2+ and Zn2+ were removed from the MTs by dialysis of the proteins against water at pH 2 followed by gel filtration. Five p~ apoMTI increased the ICs0 for the inhibition of protein synthesis by H P , Pb2+, and Cd2+ by 22, 17, and 18 p~, respectively, and apoMTII (5 pm) increased the ICs0 for H e , Pb2+, and Cd2+ inhibition by 17, 24, and 22 p~, respectively (Fig. 1, A-C). Protein synthesis was completely protected from the inhibitory effects of 5 p~ H F , 10 p~ Pb2+, and 12.5 p~ Cd+' in the presence of 2 p~ apoMTI or 11, while protein synthesis was inhibited by 25-50% by these concentrations of metal ions in the presence of 1 pM apoMT (data not shown). ApoMTI and I1 (5 p~) alone were observed to inhibit protein synthesis by approximately 5-15%. This inhibition was observed to be reversed in the presence of low concentrations of each metal ion. The capacity of metal ions to stimulate protein synthesis to levels above those present in the apoMT controls (Fig. 2) probably reflects the ability of bound metal ions to protect cysteines present in apoMT from oxidation. Concentrations of apoMT higher than 5 pM were found to be more inhibitory to protein synthesis in the absence of added metal ions (data not shown).
The Capacity of ApoMTI and 1 1 to Restore Protein Synthesis in Heavy Metal Ion-inhibited Lysates-The addition at 10 min of either 5 p~ apoMT1 or apoMTII to lysates incubated in the presence of 10 p~ H P restored protein synthesis to near the rate of proteins synthesis in lysates to which the apoMT was added at the beginning of the incubation (data not shown). Restoration was not immediate and occurred after a lag period of 5-10 min. Similarly, the addition of apoMTI or I1 to lysates after the shut off of protein synthesis had occurred in the presence of 10 p~ Pb2+ or 12.5 pM Cd" stimulated protein synthesis to rates near those present in heminsupplemented control lysates (data not shown). MTs were in general ineffective in restoring protein synthesis in metal ioninhibited lysates, giving only marginal stimulation in protein synthesis rates, at very low concentration of the inhibitory ions.
The Effects of Heavy Metal Ions on the Capacity of Hemin- supplemented Lysates to Reduce Disulfide Bonds-The capacity of lysates to reduce the disulfide bonds present in insulin was used as an assay to determine the overall ability of lysates t o maintain sulfhydryl groups in proteins in a reduced state. The assay presumably reflects the activity of the thioredoxinl thioredoxin reductase system present in the lysate (9-11). Toxic heavy metal ions (Cd2+, Hg+, and Pb2+), which were all previously found to inhibit protein synthesis in heminsupplemented lysates by bringing about the activation of HRI (l), were all found to inhibit the capacity of lysates to reduce disulfide bonds (Fig. 3). A low level of Zn2+ (50 p~) , which is only slightly inhibitory to protein synthesis, was observed to inhibit the lysate's capacity to reduce disulfides by 80-90%, while a higher, inhibitory level of Zn2+ (1 mM) completely inhibited the lysate's capacity to reduce disulfides. The capacity to reduce disulfides was unaffected (if not stimulated) in lysates where protein synthesis was inhibited under nonoxidative conditions due to the activation of HRI by hemedeficiency, or due to the activation of the double-stranded RNA-activated eIF-2a kinase by the addition of doublestranded RNA (poly(I)*poly (C)) (Fig. 3).
The ability of Thiolyte to derivatize insulin (reduced by prior incubation with a hemin-supplemented lysate) in the presence of added H e , demonstrated that the metal ion itself is not directly affecting the ability of Thiolyte to react with the thiols in reduced insulin. In addition, the heavy metal ion-induced inhibition of the lysates capacity to reduce disulfide bonds was found to be very rapid, occurring within the first 15-min incubation step of the lysate with insulin, even without any prior preincubation of the lysate in the presence of the metal ion (data not shown). On 10% SDS-polyacrylamide gels, the metal ions were found to have little effect on the overall derivatization of lysate proteins by the Thiolyte, demonstrating that these metal ions are not present at a concentration sufficient to undergo a generalized reaction with free thiol groups present in proteins within the lysate (data not shown).

Effects of EDTA and ApoMTs on Disulfide Bond Reduction and eIF-Sa Phosphorylation in Metal Ion-containing Ly-
sates-The ability of lysate to reduce disulfide bonds or phosphorylate eIF-2 was determined under conditions where the addition of EDTA, apoMT, or M T was previously found to preserve (addition at 0 min) or restore (addition at 10 min) protein synthesis in metal ion-containing lysates (Figs. 1 and 2). Addition of EDTA to Cd2+-containing lysates or apoMT (I or 11) to He-containing lysates was observed to maintain and restore the capacity of these lysates to reduce disulfide bonds, in conjunction with the inhibition of HRI activation and/or activity (Fig. 4, apoMTII data not shown). ApoMTs were found to have similar effects on insulin reduction and eIF-2 phosphorylation in lysates incubated in the presence of Cd2+ (data not shown). MTI and 11, also inhibited HRI activation and prevented the phosphorylation of eIF-2 in Cd"-containing lysates (data not shown). Cyclic AMP inhibited eIF-2a phosphorylation (Fig. 5 B ) in the absence of any restorative effects on the capacity to reduce disulfide bonds (Fig. 5A), under conditions where cAMP was previously found to maintain protein synthesis in He-containing lysates (1).
In some experiments, we noted that the sulfhydryl groups of a 24-kDa protein became oxidized in hemin-supplemented lysates in the presence of toxic heavy metals. These sulfhydryls were observed to be restored to the reduced state upon restoration of the lysates capacity to reduce disulfide bonds (Fig. 4). The oxidation of the 24-kDa protein occurred at or after the time of protein synthesis shut off, suggesting that its oxidation is not involved in bringing about the activation of HRI. Jackson and co-workers (11) have similarly noted the oxidation of a 24-kDa protein in gel-filtered (glucose deprived) lysates, and have previously reported that cAMP stimulated protein synthesis, but did not promote disulfide bond reduction in these lysates. Similar results were obtained in protection experiments where EDTA or apoMTI were added to metal containing lysates at 0 min, and the ability of the lysates to reduce disulfide bonds or phosphorylate eIF-2 were measured a t 10 min. The addition of 5 p~ apoMTII gave the same results as the addition of apoMTI (data not shown). In control experiments, addition of apoMTI or I1 alone was found to have no effect on the capacity of lysate to reduce disulfide bonds or the phorphorylation of eIF-2 (data not shown).

Effect of Cd2+ on NADPH and N A D P Concentrations in
Hemin-supplemented Lysates-Addition of Glc-6-P (1) or NADPH (0.1-1 mM, data not shown) was found to have no effect on the inhibition of protein synthesis by Cd2+, H e , or Pb2+ in hemin-supplemented lysates. This suggests that the presence of heavy metal ions do not deplete hemin-supplemented lysates of Glc-6-P or NADPH, thus bringing about the inhibition of protein synthesis. To verify this assumption, NADPH and NADP+ concentrations were measured by enzymatic cycling. NADPH concentrations were not found to vary significantly in lysates over a 20-min incubation, with NADPH being maintained at concentrations of 8.8 & 0.6,9.2 f 0.6, and 9.1 f 0.6 p~ in hemin-supplemented, Cd2+-inhibited (75 p~) , and heme-deficient lysates, respectively. NADP+ concentrations were found to be 2.1 & 0.2 and 2.2 f 0.6 p~, respectively, in hemin-supplemented and Cd2+-inhibited lysates. The lysate preparation used in these experiments was apparently replete with Glc-6-P, since addition of 1 mM Glc-6-P to a hemin-supplemented lysate did not significantly increase the concentration of NADPH present ([NADPHIcl,. 6.p = 8.5 f 0.3 PM).

DISCUSSION
H e , Pb2+, and Cd2+ bind avidly to thiols, forming mercaptides with the sulfhydryl group of cysteine, which are more stable than complexes that these metal ions form with other amino acid side chains (41). Their relative order of inhibitory potency for protein synthesis (Hg2+>Pb2+>Cd2+>ZnZ+) is similar to the order of their binding affinities to sulfides and cysteine (25-27). The observation that H e inhibits the activity of partially purified HRI in vitro: suggests that the metal ions are not activating HRI in situ by binding directly to the kinase. Because of its high binding affinity, H e is generally thought to bind specifically to sulfhydryl groups at low concentrations. The observation that low concentrations Pb2+ (5-10 p~) inhibit protein synthesis to a greater extent ' R. L. Matts, unpublished observations. in hemin-supplemented lysates than similar concentrations of Cd2+, is more consistent with the observed interaction of Pb2+ with enzymes containing active-site dithiols than with its interaction with enzymes containing single functional sulfhydryls (41). Arsenite was also observed previously to be a potent inhibititor of protein synthesis (1). Both arsenite and Cd2+ are known to have a high affinity and selectivity for vicinal sulfhydryl groups (28, 42-47). Cd2+ was also found to be a much more potent inhibitor of protein synthesis than Zn2+ (30-50 times). These observations support the hypothesis that toxic heavy metal ions are binding to some protein(s) containing a vicinal sulfhydryl group, whose function is critical for maintaining protein synthesis in hemin-supplemented reticulocyte lysates. The data presented in this study further support this hypothesis. The ready reversal of Cd2+-inhibition by dithiols, but not by monothiols, is a commonly accepted criterion for demonstrating the presence of active site dithiols in enzymes (30, 31). The reversal of Cd2+-induced inhibition of protein synthesis by EDTA, but not by NTA, meets an additional criterion for establishing the presence of a functional dithiol in a protein (32) that is not compromised by the problem attendent with the ability of thiols to readily reduce protein disulfides and thus exert independent effects on proteins. Comparable results in experiments examining the effects of mono or dithiol compounds, and metal-chelating agents on H$+-and Pb2+-induced inhibition of protein synthesis suggest that these ions are also acting through a similar mechanism. The ability of EDTA and DTPA to reverse the inhibition of protein synthesis induced by Cd2+ or Pb2+, but not H e , is consistent with the fact that only Hg2' has a stability constant for the binding to cysteine (Hg(cys)Z) that is much greater than the stability constant of its binding to EDTA and DTPA (26,40,41).
Maintenance of protein synthesis initiation in rabbit reticulocyte lysates has been shown to require the presence of hemin, a sugar phosphate (i.e. Glc-6-P) and a reducing system capable of reducing disulfide bonds (7-11). Sugar phosphates are required as a stimulatory "cofactor" affecting the rate of protein chain initiation and for NADPH generation by way of their metabolism through the pentose phosphate shunt. The requirement for reducing power is thought to be met by this NADPH generation, together with an active thioredoxin/ thioredoxin reductase system (10,ll). The data in this report indicate that toxic heavy metal ions inhibit the lysate's capacity to reduce disulfide bonds. The ability of CAMP, a competitive inhibitor of HRI (6, 48), to stimulate protein synthesis in lysates whose reducing capacity remains inhibited, suggests that the capacity to reduce disulfide bonds is required for maintaining HRI in an inactive state, but it is not required for the maintenance of protein synthesis in general. The data indicate that heavy metal ions neither deplete lysates of Glc-6-P or NADPH, nor inhibit the lysate's capacity to generate NADPH. Cd2+ has been reported to be a potent inhibitor of partially purified thioredoxin/thioredoxin reductase activity (49) and of protein disulfide isomerase (50, 51). Heavy metal ions may then inhibit the lysate's reducing capacity directly, through binding to the vicinal sulfhydryl groups present in the active sites of either thioredoxin or thioredoxin reductase (33,34).
MTs and apoMTs were observed to maintain protein synthesis in hemin-supplemented lysates in the presence of Cd", Pb", or Hg+'* by preventing the inhibition of the lysate's thioredoxin/thioredoxin reductase activity, thus inhibiting the activation of HRI. Presumably this is the result of the apoMTs capacity to bind and sequester these toxic heavy metal ions (22, 23). The ability of apoMTs to restore the capacity of metal ion-inhibited lysates to reduce disulfide bonds, in conjunction with the inactivation of HRI and the resumption of protein synthesis, indicates that apoMTs also have the ability to detoxify metal ions already bound to proteins. The affinity constant for the binding of H P to cysteine (Hg(cys),) has been reported to be M -~ (52). Given the very slow off rates expected for Hg' from such a complex, the relatively rapid restoration of protein synthesis that is observed upon the addition of apoMTs to Hg2"inhibited lysate suggests that a direct interaction of apoMT with the protein-bound metal may be occurring. MTs have been shown to be capable of donating bound Cu or Zn to metalrequiring enzymes, activating their catalytic activity (53-57). The kinetics of the metal ion-exchange between MT and these enzymes has led to the suggestion that a direct interaction between MT and other proteins may occur and that MT contains unusually reactive metal-binding sites which are capable of donating metal to apometalloproteins (57). The data presented in this report indicate that apoMT contains metal-binding sites that are also capable of extracting tightly bound metal ions from proteins.
The log of the average apparent binding constant of MT at neutral pH has been calculated to be 15.7 for Cd2+ binding and 11.7 for the binding of Zn2+ (22, 23). MTs are composed of two binding domains which vary in their metal binding specificity and affinity (22, 23, 58-63). The a-domain is capable of binding four divalent metal ions. The initial binding of Cd2+ to apometallothionein occurs preferentially to the a-domain (58-63), with an apparent affinity constant which is greater than the binding constant of Cd2+ to EDTA[log k > 16.61 (63,64). The p-domain, which can bind three divalent metal ions, has a lower average metal ion binding affinity and a greater tendency to lose metal ions. The overall binding affinity of Cd2+ to the P-domain has been estimated to be lower than the binding affinity of Cd2+ to EDTA, and the lability of Cd2+ bound in the @-domain has been suggested to be similar to that of Zn2+ (63, 64). After its initial binding to MT, Cd2+ redistributes preferentially to the a-domain displacing bound Zn2+ (60, 63). This observation has led to the proposal that the a-domain of MT may function in the sequestration of toxic metal ions, while the &domain may function in the storage of essential metals, such as Zn and Cu, and in the donation of these ions to apometalloproteins (60). The observation that only chelating agents with a binding affinity for Cd2+ equal to or greater than EDTA (e.g. DPTA, but not NTA) have the capacity to maintain or restore protein synthesis in the presence of Cd2+ suggests that only the four metal-binding sites present in the a-domain of MT would have high enough binding affinity to protect protein synthesis from Cd2+ inhibition. This proposal is supported by several other observations. The AIc50 (13 p~) for the inhibition of protein synthesis by Cd2+ in the presence of MTII correlated best with the concentration of displacable Zn2+ calculated to be bound to the a-domain (16 p~) .
MTI preparations, whose a-domains were estimated to be saturated with bound Cd2+, were observed to have little or no capacity to protect protein synthesis from inhibition by Pb2', and only marginally changed the IC6o values for the inhibition of protein synthesis by H%+ or Cd2+. Furthermore, the average value for the inhibition of protein synthesis by Cd", H P , or Pb2+ in the presence of 5 p~ apoMTI or I1 was 21 pM. A AIc5o of 35 PM would be expected if the chelation of metal ions by all seven binding sites played a role in protecting protein synthesis from inhibition. Thioredoxin maintains the sulfhydryls in a variety of cytoplasmic proteins in a reduced state, and it has been sug-gested that thioredoxin plays a direct role in the regulation of a variety of protein functions and enzymatic activities through reversible oxidation of dithiols to disulfides (33,34). Previous studies with partially purified HRI in vitro have demonstrated that sulfhydryl groups play an important role in the regulation of HRI activation and activity (5, 6, 65, 66), as does its phosphorylation state (6, 67, 68). In the absence of a functional thioredoxin/thioredoxin reductase system, a change in the redox state of certain critical sulfhydryl groups in HRI could occur, either through direct oxidation or through thioldisulfide exchange (5, 6, ll), leading to its activation. An alternate possibility, discussed by Jackson et al. ( l l ) , is that the real target of oxidation may not be HRI, but a protein which plays a role in regulating the activation or inactivation of HRI. HRI has recently been demonstrated to interact with the 90-kDa heat shock protein (hsp 90) in hemin-supplemented lysates in situ (69) and hsp 90 has been reported to affect the activity of HRI in uitro (70,71).
Conditions or agents which bring about the activation of HRI in hemin-supplemented reticulocyte lysates are also known to bring about the heat shock or stress response in eukaryotic cells (72-75). Phosphorylation of eIF-2 occurs in cultured cells in response to a variety of stresses (76,77), and activation of an eIF-Sa kinase with properties similar to HRI has been reported to occur in HeLa cells in response of heat shock (78). It has been suggested that accumulation of improperly folded proteins within a cell may be the signal that actually initiates the heat shock response (79). Several of the major heat shock proteins (hsps) are members of multigene families. These proteins are thought to play a role in the folding and/or unfolding of incompletely folded or improperly folded polypeptides (75). Thioredoxin has been shown to greatly accelerate the rate at which denatured proteins, which contain incorrect disulfide bonds, refold and recover activity (80). Inhibition of thioredoxin activity could lead to the accumulation of oxidized, improperly folded, or denatured proteins, which then might affect the association of hsp 90 with HRI, and therefore, the activity of HRI. We are currently investigating how toxic heavy metal ions affect the sulfhydryl status of HRI and hsp 90, and the association of these proteins with one another in the reticulocyte lysate.