Thioltransferase (Glutaredoxin) Reactivates the DNA-binding Activity of Oxidation-inactivated Nuclear Factor I*

The reversible oxidative inactivation of transcription factors has been proposed to be important in cellular responses to oxidant stress and in several signal transduction pathways. The nuclear factor I (NFI) family of transcription factors is sensitive to oxidative inactivation due to the presence of a conserved, oxidation-sensitive cysteine residue within the NFI DNA-binding domain. Here we show that restoration of the DNA-binding activity of oxidized NFI-C can be catalyzed in vitro by the cellular enzyme thioltransferase (glutaredoxin) coupled to GSH and GSSG reductase. To test whether GSH-dependent pathways play a role in the maintenance of NFI activity in vivo, we used buthionine sulfoximine, an agent that inhibits GSH synthesis, andN-acetylcysteine, an agent that can replenish intracellular GSH. Pretreatment of HeLa cells with buthionine sulfoximine greatly potentiated the inactivation of NFI by the oxidizing agent diamide. Inclusion of N-acetylcysteine in the culture medium during the recovery period following diamide treatment increased the extent of restoration of NFI activity. These results suggest that maintenance of the DNA-binding activity of NFI proteins during oxidant stress in vivo requires a GSH-dependent pathway, likely involving thioltransferase-catalyzed reduction of the oxidation-sensitive cysteine residue on NFI.

The reversible oxidative inactivation of transcription factors has been proposed to be important in cellular responses to oxidant stress and in several signal transduction pathways. The nuclear factor I (NFI) family of transcription factors is sensitive to oxidative inactivation due to the presence of a conserved, oxidation-sensitive cysteine residue within the NFI DNA-binding domain. Here we show that restoration of the DNA-binding activity of oxidized NFI-C can be catalyzed in vitro by the cellular enzyme thioltransferase (glutaredoxin) coupled to GSH and GSSG reductase. To test whether GSHdependent pathways play a role in the maintenance of NFI activity in vivo, we used buthionine sulfoximine, an agent that inhibits GSH synthesis, and N-acetylcysteine, an agent that can replenish intracellular GSH. Pretreatment of HeLa cells with buthionine sulfoximine greatly potentiated the inactivation of NFI by the oxidizing agent diamide. Inclusion of N-acetylcysteine in the culture medium during the recovery period following diamide treatment increased the extent of restoration of NFI activity. These results suggest that maintenance of the DNA-binding activity of NFI proteins during oxidant stress in vivo requires a GSH-dependent pathway, likely involving thioltransferase-catalyzed reduction of the oxidation-sensitive cysteine residue on NFI.
Modulation of DNA-binding activity by oxidation and reduction reactions (redox regulation) has been demonstrated for a number of transcription factors, including NF-B/Rel proteins, Fos and Jun proteins, and others (1)(2)(3)(4)(5)(6). The mechanism of oxidative inactivation of DNA binding is through the oxidation of specific cysteine residues close to or within the DNA-binding domains of these proteins. This modulation of the DNA-binding activity of transcription factors by redox state has been proposed to be important both in cellular response to oxidant stress and in a number of signal transduction pathways including those mediated by tumor necrosis factor ␣ (7,8). While the process of oxidative inactivation of DNA binding has been studied extensively in vitro, the biochemical pathways that mediate the reversible oxidation of oxidation-sensitive cysteine residues are only poorly understood. We are using the NFI 1 family of transcription factors as a model system to characterize the pathways for reversible oxidative inactivation of DNAbinding proteins in vitro and in vivo.
The NFI family of site-specific DNA-binding proteins is required both for adenovirus DNA replication and for the expression of a large number of cellular and viral genes (9 -11). In vertebrates, this family of proteins is encoded by four distinct genes (NFI-A, NFI-B, NFI-C, and NFI-X), and the expression patterns of the four genes are regulated during cell differentiation and development (12,13). We had shown previously that the DNA-binding activity of the NFI family of transcription factors was inactivated by oxidation in vitro (14). This sensitivity to oxidation requires the presence of a single oxidationsensitive cysteine residue that is conserved within the DNAbinding domains of all NFI proteins from Caenorhabditis elegans to humans. Mutation of this oxidation-sensitive cysteine residue creates NFI proteins with apparently normal DNA-binding activity, but the DNA binding of the mutant proteins is not sensitive to oxidative inactivation in vitro. The dispensability of this residue for DNA binding, together with its phylogenetic conservation in the DNA-binding domain of all NFI proteins, suggests a potential regulatory function for this oxidation-sensitive cysteine residue. An analogous situation for regulation by cysteine modification exists for human immunodeficiency virus type 1 protease, which has two highly conserved cysteine residues that are not required for basal activity; however, modification of these cysteines, e.g. by glutathionylation, alters the activity of the protease in a reversible fashion (15,16).
If the modulation of DNA-binding activity by redox state is a regulated physiological process, we would anticipate that specific cellular proteins would mediate the oxidation and reduction of oxidation-sensitive cysteine residues on DNA-binding proteins. Indeed, there are two classes of proteins in mammalian cells that have been shown to mediate the reduction of oxidation-sensitive cysteine residues on DNA-binding proteins in vitro. Studies from several laboratories have demonstrated that the bifunctional HAP1/Ref-1 protein can reactivate oxidized Fos/Jun proteins in vitro (17,18). This redox activity is present in the N-terminal domain of HAP1/Ref-1 and is separate from the C-terminal domain of HAP1/Ref-1 that mediates repair of DNA damage (19). Also, the cellular reducing protein thioredoxin has been shown to reactivate oxidized NF-B in vitro (2,3). Thioredoxin was first discovered as the protein that provided reducing equivalents to ribonucleotide reductase and has been shown to mediate the reduction of a number of intracellular proteins (20). Likewise, thioltransferase (glutaredoxin) has also been shown to catalyze the reactivation of many oxidatively modified proteins. In particular, we found that thioltransferase (TTase) is specific for protein-glutathione mixed disulfides (protein-SSG) (21,22), and the catalytic intermediate TTase-SSG is selectively recycled to the reduced enzyme by GSH (23), with formation of GSSG and regeneration of GSH by coupling with NADPH and GSSG reductase. These characteristic interactions of thioltransferase with glutathione distinguish it from thioredoxin, which favors intramolecular disulfide substrates and is turned over by NADPH and thioredoxin reductase, independently of glutathione.
Since oxidation of the single cysteine residue on NFI in the presence of GSH likely forms NFI-SSG in vitro or in vivo, we tested the ability of thioltransferase to restore the DNA-binding activity of NFI that was oxidized by diamide in the presence of GSH. Here we report that thioltransferase can reactivate oxidized NFI protein in vitro in a GSH-dependent manner. In addition, we demonstrate that treatment with buthionine sulfoximine, a known inhibitor of GSH synthesis (24), potentiates the oxidative inactivation of NFI in cultured HeLa cells and that inclusion of N-acetylcysteine (precursor of intracellular GSH) in the culture medium during the recovery period following diamide treatment increases the extent of restoration of the DNA-binding activity of NFI. These findings suggest an important role for thioltransferase and intracellular GSH in the reduction of oxidation-sensitive cysteine residues in the NFI family of transcription factors and potentially other site-specific DNA-binding proteins.

Recombinant NFI Expression and
Purification-Construction of the pET220 vectors that express the NH 2 -terminal 220 amino acids of wild-type or mutant human NFI-C from a T7 promoter was described previously (25). All of the NFI proteins expressed contain a hexahistidine tag fused to the N terminus of the human NFI-C protein. Expression of wild-type and mutant NFI proteins in Escherichia coli cells and preparation of bacterial extracts were as described (14). The NFI proteins were purified by absorption to phosphocellulose equilibrated with 25 mM Hepes, 10% sucrose, 350 mM NaCl, 0.1% Nonidet P-40, and 2 mM DTT and eluted with the same buffer containing 1.0 -1.5 M NaCl. The final purification step was achieved by Ni 2ϩ -nitrilotriacetic acid-agarose chromatography as described previously (14). Protein concentration was determined by the method of Bradford (26). Protein preparations were analyzed by electrophoresis on 10% SDS-polyacrylamide gels, determined to be Ͼ90% pure by silver staining (Bio-Rad), and quantified using bovine serum albumin as a standard.
Gel Mobility Shift Assay (GMSA)-NFI DNA-binding activity was determined by GMSAs with purified recombinant proteins or extracts of nuclei using the 32 P-labeled FIB-2.6 oligonucleotide or a control mutant oligonucleotide (FIB-2.6C2) as described (25). The FIB-2.6 oligonucleotide contains a well characterized NFI-binding site cloned previously from human genomic DNA, whereas the FIB-2.6C2 oligonucleotide has a single point mutation that reduces NFI binding Ͼ100-fold in vitro. The inactivation of NFI DNA-binding activity by the oxidizing agent diamide and the reactivation by reducing agents or the TTase-reducing system were performed as indicated in the figure legends. NFI⅐DNA complexes were quantified using a Molecular Dynamics Model 400 PhosphorImager.
Preparations of Thioltransferase-Thioltransferase was purified from human red blood cells as described (27), and recombinant human thioltransferase was purified from E. coli (28); both sources of the enzyme display the same catalytic characteristics (28) and are used interchangeably. One unit of thioltransferase is defined as the amount of enzyme required to catalyze the glutathione-dependent reduction of 1 mol of cysteinylglutathione disulfide/min under standard assay conditions in which GSSG formation is coupled to NADPH oxidation by GSSG reductase (21,28). The reactivation of the DNA-binding activity of oxidized NFI was demonstrated with purified human thioltransferase isolated in both ways.
Cell Culture and Preparation of Nuclear Extracts-HeLa cells were cultured in ␣-minimum essential medium supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 g/ml streptomycin, and 100 units/ml penicillin at 37°C in 5% CO 2 . To induce oxidative stress, cells were treated with 10 mM diamide at 37°C for 15-30 min and washed with phosphate-buffered saline as described in the figure legends. Subsequent recovery of the cells was performed for 2 h at 37°C in the presence or absence of the protein synthesis inhibitor cycloheximide (25 g/ml). To deplete the cells of glutathione, cells were incubated with 100 M buthionine sulfoximine (BSO) for 48 h. BSO blocks the synthesis of glutathione by inhibiting the enzyme ␥-glutamylcysteine synthetase (24). More than 90% of the cells were viable under this condition. Cells were treated with diamide as described above and allowed to recover from oxidant stress in the presence and absence of 25 mM N-acetylcysteine. For the preparation of nuclear extracts, the cells were scraped into cold phosphate-buffered saline and harvested. The cells were lysed, and nuclear extracts were prepared in 25 mM Hepes, pH 7.5, 10% sucrose, 0.01% Nonidet P-40, and 1 mM phenylmethylsulfonyl fluoride containing 0.35 M NaCl as described previously (30).

RESULTS
Our earlier study showed that Cys-3 present in the DNAbinding domain of all NFI proteins was sensitive to oxidation and chemical modification in vitro (14). Oxidation inactivates the DNA-binding activity of NFI, and subsequent treatment with the nonphysiological reducing agent DTT restores the binding activity. Since GSH is a major intracellular redox buffer and is believed to play a role in the maintenance of reduced sulfhydryl groups on proteins in vivo (31), we tested whether reduced GSH alone could restore the DNA-binding activity of the oxidized NFI-C220 protein in vitro in comparison with the nonphysiological reducing agents DTT and ␤-mercaptoethanol. As shown previously, purified NFI-C220 binds specifically to the FIB-2.6 oligonucleotide, which contains a consensus NFI-binding site (Fig. 1, lane 2). Treatment of NFI-C220 in vitro with 1 mM diamide abolished the DNA-binding activity (Fig. 1, lane 3). As expected, this inactivation was reversed when the oxidized NFI protein was further incubated with a 5-fold molar excess of the nonphysiological reducing agent DTT (Fig. 1, lane 4) or ␤-mercaptoethanol (lane 5). However, a 5-fold molar excess of GSH failed to reactivate oxidized NFI under these conditions (Fig. 1, lane 6), and GSSG also had no effect, as expected (lane 7). The lack of an effect of GSH alone in reactivating oxidized NFI-C220 is similar to observations with other sulfhydryl proteins whose activities are inhibited by oxidation and not restored by GSH alone. For example, oxidatively inactivated phosphofructokinase and glutathione S-transferase are not reactivated by GSH unless thioltransferase is also present (22), indicating that rapid reactivation requires enzyme catalysis.
Thioltransferases have long been implicated in the maintenance of sensitive intracellular thiols and the reduction of protein disulfides (22). To determine whether TTase can regenerate the DNA-binding activity of oxidized NFI, NFI protein was treated with diamide, diluted in the presence of GSH to quench the diamide, and then incubated in the absence or presence of thioltransferase. As shown previously, the DNAbinding activity was abolished when NFI was oxidized with diamide ( Fig. 2, lane 3 versus lane 1). With subsequent incubation in the presence of a 4-fold molar excess of GSH alone or GSH plus NADPH and GSSG reductase, a low level of binding activity was restored (Fig. 2, lanes 4 and 7, respectively). This limited reactivation by GSH at 30°C was not observed when the incubation was performed at 4°C (Fig. 1, lane 6). A much greater restoration of DNA-binding activity was seen when oxidized NFI was incubated with thioltransferase in the presence of GSH and the GSH-regenerating system (Fig. 2, lane 6). This increased reactivation required both TTase and the regenerating system since incubation with either component alone produced levels of DNA-binding activity similar to that seen with GSH alone (Fig. 2, lanes 7 and 8). The extent of NFI reactivation by TTase was only slightly less than that obtained with the nonphysiological reducing agent DTT (Fig. 2, lane 5).
In the absence of oxidation, GSH had no effect on the binding activity of NFI (Fig. 2, lane 2), and as expected TTase, did not affect the DNA-binding activity of either the non-oxidized wild type or the oxidation-resistant Cys-3 mutants of NFI (data not shown). These data indicate that TTase, in the presence of GSH and a GSH-regenerating system, can effectively reactivate the DNA-binding activity of diamide-oxidized NFI. On a molar basis, thioltransferase (0.7 M) is at least 3000-fold more efficient than GSH alone (2 mM) and 1200-fold better than DTT (1 mM) in effecting NFI reactivation under the same conditions.
For a more detailed analysis of TTase-mediated reactivation of oxidized NFI, we measured the time course and TTase concentration dependence of NFI reactivation (Fig. 3). TTase greatly accelerated the rate of reactivation of oxidized NFI, reaching an apparent plateau within 5 min, whereas the reactivation of NFI by GSH either alone or with a GSH-regenerating system was much slower and was incomplete even after 30 min (Fig. 3 and data not shown). During a brief 5-min incubation, increasing the amount of TTase increased the extent of NFI reactivation (Fig. 3, inset). These results indicate that TTase, in the presence of GSH and a GSH-regenerating system, increases the rate of reactivation of oxidized NFI compared with GSH either alone or with a GSH-regenerating system.
Previously, we had shown that the DNA-binding activity of NFI is inactivated by oxidation in vitro (14). To determine whether the DNA-binding activity of NFI is sensitive to oxidative inactivation in intact cells, we measured NFI DNA-binding activity in extracts of HeLa cells treated with diamide (Fig. 4). HeLa cells (Ͻ50% confluent) were cultured in the presence of 10 mM diamide and harvested, and nuclear extracts were prepared as described under "Materials and Methods." As shown previously (30), nuclear extracts from untreated HeLa cells yielded multiple specific NFI⅐DNA complexes (Fig. 4, lane 4). NFI DNA-binding activity was reduced by Ͼ75% after diamide treatment of HeLa cells for 30 min (Fig. 4, lane 5). When the diamide-treated HeLa cells were washed and cultured for a further 2-h recovery period in the absence of diamide, there was an ϳ2-fold increase in NFI DNA-binding activity (Fig. 4,  lane 6), but the level of binding activity remained below that seen in control cells. This partial restoration of NFI DNAbinding activity appears to be independent of protein synthesis since incubation of HeLa cells with the protein synthesis in-

FIG. 2. Reactivation of diamide-oxidized NFI-C220 by thioltransferase.
Diamide oxidation of NFI-C220 and GMSA analysis of NFI binding activity were performed as described in the legend to Fig.  1. Components were added sequentially to each reaction in the order listed below. NFI-C220 (ϳ10 ng) was incubated with 0.5 mM diamide at 4°C for 5 min, followed by the addition of 2 mM GSH. Oxidized NFI-C220 was then diluted 10-fold and incubated in the presence of the reactants listed below at 30°C for 15 min. 1 fmol of 32 P-labeled FIB-2.6 oligonucleotide was added, and incubations were continued at 4°C for hibitor cycloheximide during the recovery period had little or no effect (Fig. 4, lane 7). As expected, incubation of HeLa cells with cycloheximide in the absence of diamide had little or no effect on NFI DNA-binding activity (Fig. 4, lane 8).
Since intracellular glutathione is a major redox buffer in cells and a known substrate for thioltransferase (21)(22)(23) and may protect proteins from oxidative inactivation (31,32), we investigated the potential role of glutathione in the oxidative inactivation and subsequent recovery of NFI DNA-binding activity. HeLa cells were cultured with two agents known to affect the levels of intracellular glutathione, BSO and N-acetylcysteine. BSO reduces intracellular GSH levels by inhibition of ␥-glutamylcysteine synthetase, whereas N-acetylcysteine can increase GSH levels by providing a high intracellular concentration of cysteine, a precursor of GSH (24). HeLa cells were cultured in the presence and absence of BSO or N-acetylcysteine as described under "Materials and Methods." BSO treatment alone had no affect on the level of NFI DNA-binding activity present in the nuclear extracts from HeLa cells (Fig. 5,  lanes 1 and 5). However, upon treatment with diamide, cells cultured in the presence of BSO had a greatly potentiated inactivation of NFI DNA-binding activity compared with control cells (Fig. 5, lane 6 versus lane 2). As shown previously, a partial restoration of NFI DNA-binding activity occurred in the BSO-and diamide-treated cells during a subsequent 2-h recovery period (Fig. 5, lane 7). This partial restoration of NFI DNA-binding activity was more pronounced when N-acetylcysteine was present in the culture medium during the recovery period (Fig. 5, lane 8). In this experiment, diamide treatment of the control HeLa cells caused only a 40% reduction in NFI DNA-binding activity, which showed little or no recovery during subsequent incubation in the presence or absence of Nacetylcysteine (Fig. 5, lanes 2-4). We have observed that the extent of NFI inactivation by diamide and the subsequent recovery of activity appear to be cell density-dependent and that little or no reactivation of binding occurs unless Ͼ70% inactivation of NFI is achieved during the initial diamide treatment (data not shown). This apparent cell density dependence may be due to previously observed differences in GSH levels in low versus high density cells in actively growing cultures (33,34). These results indicate that intracellular GSH levels may affect both the sensitivity of NFI to oxidative inactivation and the subsequent recovery of NFI DNA-binding activity after oxidant stress. DISCUSSION While our previous study showed the existence of a highly conserved, oxidation-sensitive cysteine residue in the NFI family of transcription factors (14), the physiological significance of the residue remained unclear. Here we demonstrate both that the in vitro oxidative inactivation of NFI-C can be reversed by treatment with the cellular enzyme thioltransferase (Figs. 2 and 3) and that NFI DNA-binding activity is subject to oxidative inactivation and reactivation in intact cells (Figs. 4 and 5). The inefficiency of GSH alone in reactivating oxidized NFI (Figs. 2 and 3), together with the potentiation of the oxidative FIG. 3. Time course of reactivation of oxidized NFI-C220 by thioltransferase. The oxidation of recombinant NFI-C220 by diamide and subsequent incubation in the presence and absence of TTase were performed as described under "Results." To measure the time course of reactivation, diamide-oxidized NFI was incubated at 30°C for 5, 10, 15, or 30 min with identical amounts of appropriate reagents including TTase (0.016 units) and GSH (0.2 mM) or the GSH-regenerating system as described in the legend to Fig. 2. The samples were then incubated with 1 fmol of 32 P-labeled FIB-2.6 oligonucleotide for 20 min at 4°C and analyzed by electrophoresis. NFI⅐DNA complexes were quantified on a PhosphorImager and are expressed in arbitrary relative density units. The amount of DNA-binding activity is plotted versus the time of incubation. The inset shows the dependence of NFI DNA-binding activity reactivation on TTase concentration in the presence of GSH and the GSH-regenerating system for a 5-min reactivation time. The percent of NFI DNA-binding activity recovered is relative to DTT-reactivated NFI (100%). inactivation of NFI in intact cells by treatment with an agent shown to deplete intracellular GSH (BSO) (Fig. 5), suggests that GSH may be necessary for the protection and/or restoration of NFI DNA-binding activity in vivo, but it is probably not sufficient. Hence, the ability of TTase to catalyze GSH-dependent reactivation of NFI DNA-binding activity in vitro suggests that TTase may play an important role in the regulation of NFI function in vivo. Although the physiological functions of thioltransferase are not established, its catalytic activity implicates its involvement in homeostatic maintenance of intracellular thiols and regulation of receptors and enzymes (22,(35)(36)(37). In this context, pyruvate kinase, hemoglobin, phosphofructokinase, and glutathione S-transferase are examples of proteins that are inactivated under oxidizing conditions and reactivated by TTase in conjunction with GSH and GSSG reductase.
We are uncertain of the nature of the oxidized cysteine residue on NFI that prevents DNA binding. Our previous study demonstrated that while the oxidation-sensitive cysteine residue could be replaced by a number of amino acid residues (Gly, Ala, Ser, Thr, Leu, Ile, Val, and His) with retention of DNAbinding activity, replacement with the larger residues trypto-phan and arginine abolished binding (14). These findings suggest that either the addition of a large adduct due to mixeddisulfide formation with GSH or the formation of an intramolecular disulfide with one of the other essential cysteine residues on NFI could impair DNA binding. Since TTase has been shown to be highly selective for GSH-dependent reduction of protein-SSG substrates (21,22), its catalytic effectiveness in reducing oxidized NFI in the presence of GSH is consistent with NFI-SSG being the actual substrate for TTase. This interpretation is based on extensive studies of TTase catalysis of GSH-dependent reduction of model protein mixed disulfides, including cysteinyl, cysteaminyl, and glutathionyl mixed disulfides of hemoglobin, papain, and bovine serum albumin (21,22). In addition to the generation of mixed and intramolecular disulfides, the formation of stable sulfenic, sulfinic, or sulfonic acids by monothiol oxidation has also been proposed as a mechanism for the oxidative inactivation of a number of proteins (1,17,38). In this context, in vitro treatments of glyceraldehyde-3-phosphate dehydrogenase with H 2 O 2 and tetrathionate and of phosphofructokinase with GSSG were designed to give glyceraldehyde-3-phosphate dehydrogenase-S-OH, glyceraldehyde-3-phosphate dehydrogenase-S-SO 3 Ϫ , and phosphofructokinase-SSG as the corresponding inactivated forms of the enzymes (39). Comparison of equimolar amounts of TTase and thioredoxin near their physiological concentrations indicated that TTase was much more effective than thioredoxin for reactivating phosphofructokinase-SSG, as would be predicted from the substrate selectivity of TTase that we delineated (21,22). Thioredoxin, however, was better for reactivating the two types (sulfenate and thiosulfate) of modified glyceraldehyde-3-phosphate dehydrogenase (39).
By analogy to this comparison, it appears less likely that monothiol oxygenation plays a role in the oxidative inactivation of NFI. In other studies in which monothiol oxygenation has been proposed as a means of inactivation, the redox proteins thioredoxin (2, 3) and HAP1/Ref-1 (16 -18) were reported to effectively reactivate the oxidized proteins. In preliminary experiments, we did not observe reactivation of diamide-oxidized NFI with typical concentrations of the reduced form of either HAP1 or thioredoxin; however, broader studies are necessary to define the relative effectiveness of TTase and these other redox proteins (data not shown; proteins generously provided by Drs. I. Hickson and A. Gronenborn, respectively). The apparent differences in the protein requirements for the reductive reactivation of different intracellular proteins may reflect distinct biochemical pathways used to prevent or recover from oxidative stress. For example, E. coli encodes at least three different proteins capable of donating reducing equivalents to ribonucleotide reductase, each of which may also function in the reduction of oxidized intracellular proteins (40). It will be necessary in the future to selectively deplete each of the known potential redox regulators in eukaryotic cells to assess their different functions and their potential for synergism in vivo.
Our data demonstrate that NFI DNA-binding activity is sensitive in vivo to oxidative inactivation particularly under conditions of glutathione depletion (Fig. 5). However, it remains to be determined how glutathione participates in this process. Besides reacting directly with diamide or other oxidative mediators, GSH is likely to be incorporated into NFI molecules during oxidative stress to form NFI-SSG, and then the GSH would be removed by dethiolation by TTase. In fact, we have observed incorporation of [ 35 S]GSH into NFI in the presence of diamide (data not shown). This observation is consistent with the fact that oxidant treatment of a broad variety of cells (hepatocytes, neutrophils, gastric mucosal cells, etc.) generally has produced 35  it has been suggested that S-thiolation during oxidative stress might protect particular proteins from irreversible damage (46,47).
The oxidative inactivation of the DNA-binding activity of NFI in intact cells and the effects of the GSH modulators BSO and N-acetylcysteine (Figs. 4 and 5) suggest that inactivation of NFI can occur during oxidative stress in vivo, and the extent and duration of the changes in NFI activity would be dependent on the intracellular GSH status. For example, oxidative stress occurs in various physiological and pathological states, including human immunodeficiency virus infection, where it is believed to be the result of chronic exposure of the infected cells to inflammatory cytokines (48). The oxidative situation associated with the infection may foster S-glutathionylation and concomitant modulation of the activity of viral proteins like the human immunodeficiency virus type 1 protease. In this regard, thioltransferase has been shown to display site selectivity in deglutathionylation of Cys-95 and Cys-67 of the protease (16). Whether analogous conditions of oxidative stress can lead to the inactivation of NFI DNA-binding activity in situ is under investigation. In addition, it would be of interest to see whether the oxidative inactivation of NFI family members occurs early in apoptosis since oxidative stress is believed to be an important event during apoptotic cell death (49,50). Future experiments involving expression of transfected cDNA for TTase may provide more direct evidence for a role of this enzyme in regenerating oxidatively damaged proteins in situ, including NFI and human immunodeficiency virus type 1 protease.