Suppression of nerve growth factor-induced neuronal differentiation of PC12 cells. N-acetylcysteine uncouples the signal transduction from ras to the mitogen-activated protein kinase cascade.

The cellular redox state is thought to play an important role in a wide variety cellular signaling pathways. Here, we investigated the involvement of redox regulation in the nerve growth factor (NGF) signaling pathway and neuronal differentiation in PC12 cells. N-acetyl-L-cysteine (NAC), which acts as a reductant in cells both by its direct reducing activity and by increasing the synthesis of the cellular antioxidant glutathione, inhibited neuronal differentiation induced by NGF or by the expression of oncogenic ras in PC12 cells. NAC suppressed NGF-induced c-fos gene expression and AP-1 activation. These results suggest that neuronal differentiation and NGF signaling are subject to regulation by the cellular redox state. NAC also suppressed the NGF-induced activation of mitogen-activated protein kinases (MAPKs) and decreased the amount of tyrosine phosphorylation of MAPKs. The suppression of MAPK by NAC was independent of glutathione synthesis. In parallel with the suppression of MAPK, the activation of MAPK kinase kinase activity was also suppressed in the presence of NAC. In contrast, NGF-induced activation of Ras was not inhibited by NAC. The inhibitory effect of NAC on the MAPK cascade was independent of transcription and translation. Thus, NAC suppresses NGF-induced neuronal differentiation by uncoupling the signal transduction from Ras to the MAP kinase cascade in PC12 cells.

in cells by several pathways, including electron transfer reactions, xanthine oxygenase, NADPH oxidase, and ␥-ray and UV light irradiation. The increased level of ROIs, referred to as oxidative radical stress, is generally cytotoxic, and hence cells possess several antioxidant systems, such as the antioxidant glutathione (GSH), thioredoxin/adult T cell leukemia-derived factor, and antioxidant enzymes, such as catalase, which operate to eliminate ROIs. These regulatory systems of the cellular redox state are thus essential for cell survival. When either excessive oxidative radical stress occurs or cellular antioxidant enzymes, such as superoxide dismutase, are down-regulated, the cells die by necrosis or apoptosis. In neuronal systems, oxidative radical stress is postulated to be the cause of several neuronal degenerative diseases, such as Parkinson's disease (1,2) and a familial form of amyotrophic lateral sclerosis (3)(4)(5).
Several lines of evidence have indicated that cellular redox plays an essential role not only in cell survival but also in cell signaling systems and cell growth. First, thioredoxin/adult T cell leukemia-derived factor has been shown to promote interleukin 2 receptor expression in T cells (6). Second, the activity of the transcription factors, NFB and AP-1, is regulated by the redox state, i.e. redox regulation. AP-1 is subjected to redox regulation through its conserved cysteine residue (7) and is activated by antioxidants (8,9). The activity of NFB is enhanced by antioxidants in vitro (10,11) but is potently and rapidly activated from its inactive cytoplasmic form by treatment of cells with H 2 O 2 (8,12). In addition, the antioxidants N-acetyl-L-cysteine (NAC) and thioredoxin prevent the activation of NFB both by H 2 O 2 and by several other extracellular stimuli, such as tumor necrosis factor ␣, phorbol ester, interleukin 1, and lipopolysaccharide (8,9,12). Furthermore, oxidative stress induces the expression of WAF1/CIP1, a cyclindependent kinase inhibitor, and arrests cell cycle progression (13). Thus, it is plausible that redox widely regulates many cellular functions.
The rat pheochromocytoma PC12 cell line is one of the most widely used cell lines in the study of the cellular signaling system, and these studies have revealed a common pathway linked to many growth factor stimuli, namely, the mitogenactivated protein kinase (MAPK) cascade (for review, see Refs. 14 -17). The striking feature of the PC12 cell line is its ability to differentiate into neuronal cells in response to nerve growth factor (NGF). The binding of NGF to its receptor TrkA on the plasma membrane triggers the activation of Ras, the GDPbound form of which changes to the GTP-bound form. The activated Ras then stimulates the signaling pathway known as the MAPK cascade, which consists of three kinases forming a chain reaction in a linear alignment. In the MAPK cascade, MAP kinase kinase kinases (MAPKKKs) are directly activated by Ras, which in turn control the dual-specificity kinases, MAP kinase kinases (MAPKKs). Finally, MAPKs are activated by MAPKK and then activate several transcription factors to induce neuronal differentiation. The importance of this cascade has been verified by the fact that the activation of Ras and MAPK is sufficient and necessary for neuronal differentiation of PC12 cells (18 -21). Furthermore, this cascade has been revealed to be linked not only to many receptor tyrosine kinases but also to trimeric G proteins, which transduce signals from seven-helix transmembrane receptors (for review, see Ref. 22).
Although the importance of both the signaling system involving the MAPK cascade and the cellular redox regulation system has been established, the relationship between them so far remains unclear. However, several studies have suggested the existence of an intimate cross-talk between cellular redox and the signal transduction of NGF. First, NGF increases the level of cellular GSH or the activity of the antioxidant enzymes, GSH peroxidase and catalase, in PC12 cells (23)(24)(25). Second, NGF effectively protects cells from injury by oxidative radical stress (26), and the activation of Ras is also effective in suppression of cell death caused by H 2 O 2 (25). Third, the expression of a protein phosphatase CL100, which is capable of dephosphorylating MAPK, is potently induced by oxidative stress (27), and the essential signaling molecule Ras is reported to be regulated by redox stress (28). Finally, exogenously added H 2 O 2 or other oxidizing agents induce the activation of MAPK in several cell types (29,30).
In this study, we investigated the involvement of the cellular redox state in neuronal differentiation and the cellular signaling pathway of NGF in PC12 cells.
Cell Cultures and Cell Lines-PC12 cells and PCMTras21 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated horse serum, 5% fetal bovine serum, and 50 g/ml kanamycin. PCMTras21 (25) was cloned from PC12 cells transfected with the plasmid pMTIDrasneo, which was constructed by insertion of the neomycin-resistant gene into the plasmid pMTID-ras encoding the oncogenic T24 Ha-ras gene under influence of the metallothionein promoter (MTID Ϫ113/Ϫ85 promoter) (31). For NGF stimulation, 50 g/ml NGF was added to the medium.
GSH Assay-GSH was measured by a modification of a method originally described by Anderson (32). Cells (1 ϫ 10 6 ) on a 60-mm dish were collected and suspended in 200 l of ice-cold PBS. An equal volume of 10% trichloroloacetic acid was then added to the cell suspension and the mixture was centrifuged (20,000 ϫ g for 10 min) at 4°C. The supernatant was extracted with 6 volumes of diethylether six times to remove the trichloroloacetic acid. The resulting extract was incubated with 200 M NADPH and 1 unit/ml glutathione reductase for 10 min at 37°C, and then the GSH level was measured by adding 5,5-dithiobis (2nitrobenzoic acid) to the mixture at a final concentration of 1 mM. The reaction was monitored by the absorbance change at 415 nm. Protein concentrations were determined using the BCA protein assay reagent system (Pierce).
Electrophoretic Mobility Shift Assays-Nuclear extract was prepared as described by Dignam et al. (33). A 32 P-labeled DNA probe of the AP-1 binding site (AGCTTTGATGACTCAGTCGA) was prepared by annealing the oligonucleotide AGCTTTGATGACTCAG with the oligonucleotide TCGACTGAGTCATCAA and labeling it with Klenow fragments and [ 32 P]dCTP. Nuclear extracts (10 g) were mixed with the 32 Plabeled DNA probe (1 ϫ 10 4 cpm) in a reaction buffer containing 10 mM Tris-Cl, pH 7.6, 5% glycerol, 50 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 1 mM dithiothreitol (DTT), 10 mg/ml BSA, and 100 g/ml poly(dI⅐dC) and then were incubated at room temperature for 30 min. Following the incubation, the samples were fractionated through 8% acrylamide gel in a buffer containing 40 mM Tris acetate, pH 8.5, 1 mM EDTA.
Chloramphenicol Acetyltransferase (CAT) Assay-For transfection with plasmids, PC12 cells (1 ϫ 10 6 ) on a 6-cm dish were incubated with 20 l of Lipofectin (Life Technologies, Inc.), together with 2 g of plasmid (pAPCAT) in 2 ml of Opti-MEM medium (Life Technologies, Inc.) for 6 h. pAPCAT encodes the triple AP-1 binding sites between the HindIII and SalI sites upstream of the CAT gene of the plasmid pBLCAT5 (34). After transfection, the culture was divided equally between four dishes and cultured further in 4 ml of Dulbecco's modified Eagle's medium supplemented with 10% horse serum and 5% fetal bovine serum for 2 days. The cells were collected and disrupted by three freeze-thaw treatments in 200 l of 0.25 M Tris-Cl, pH 7.8, and then centrifuged at 15,000 rpm for 15 min. 90 l of supernatant was incubated with 0.1 Ci [ 14 C]chloramphenicol and 1 mM acetyl-CoA at 37°C overnight. The samples were developed by thin-layer chromatography (TLC) using a silica gel in chloroform/methanol (114:6 by volume). CAT activity was normalized by the protein content in each sample.
Northern Blot Analysis-Total RNA was prepared following the acid guanidium thiocyanate-phenol-chloroform extraction method (35). 20 g of RNA was electrophoresed through 1% agarose containing 17% formaldehyde, and the gels were then stained with 0.05 g/ml ethidium bromide. The fractionated RNA was transferred to a nylon filter (Hybond-N ϩ , Amersham Corp.) by capillary blotting. c-fos probe DNA was prepared by labeling with [␣-32 P]dCTP using a DNA labeling kit (Pharmacia) following elution of the c-fos DNA fragment of PstI-digested pfos-1 from an agarose gel. Hybridization was performed using Rapid hybridization buffer (Amersham Corp.) following the manufacturer's instructions.
The radioactivity corresponding to the phosphorylation of myelin basic protein or of KN-MAPK was then estimated using an Imaging plate (Fuji film) following SDS-polyacrylamide gel electrophoresis.
Western Blotting-Following SDS-polyacrylamide gel electrophoresis, the proteins were transferred to an ECL membrane (Amersham Corp.) using electroblotting apparatus (AE-6675P, ATTO). The filter was stained with the antibody using an ECL Western blotting system (Amersham Corp.) following the manufacturer's instructions.
Mono-Q Column Chromatography-Fractionation of proteins by column chromatography was performed using the Smart System (Pharmacia). 5 ϫ 10 6 cells were solubilized in 0.5 ml of buffer A and then were centrifuged for 10 min at 100,000 rpm to obtain the supernatant. The cell extract (100 l) was subjected to a Mono-Q PC 1.6/5 column equilibrated with buffer A. The column was washed with 300 l of buffer A, and then adsorbed proteins were eluted with a 1.1-ml linear gradient of 0 -0.4 M NaCl, followed by a 0.1-ml linear gradient 0.4 -1.0 M NaCl. The eluate was recovered in 100 l fractions. All procedures were carried out at 4°C.
Analysis of GTP/GDP Bound Forms of Ras-GTP/GDP binding to Ras was analyzed as described by Muroya et. al. (36). Briefly, 1 ϫ 10 6 cells on a 6-cm dish were incubated with 50 Ci/ml [ 32 P]phosphorus in phosphate-free Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 1% horse serum and 0.5% fetal bovine serum.
The cells were washed with ice-cold buffer containing 20 mM Tris-Cl, pH 7.4, 150 mM NaCl, and 1 mM Na 3 VO 4 and then solubilized in 1 ml of buffer containing 20 mM Tris-Cl, pH 7.4, 150 mM NaCl, 20 mM MgCl 2 , 0.5% Triton X-100, 1 g/ml leupeptin, 1 g/ml antipain, and 1 mM Na 3 VO 4 . The cell extract was incubated with 1 l of anti-Ras antibody, 2 l of anti-Rat IgG antibody, and 2 l of protein A-Sepharose at 4°C for 1 h, and then the immune complex was washed with 1 ml of ice-cold buffer four times. Guanine nucleotides bound to Ras were analyzed by TLC using a polyethylene imine sheet (Marchery-Nagel) in 0.8 M KH 2 PO 4 (pH 3.4).

RESULTS
Treatment with NGF for 3 days induced neuronal differentiation of PC12 cells with the appearance of neurite extensions (Fig. 1A). This NGF-induced morphological change in PC12 cells was completely abolished in the presence of NAC in concentrations higher than 10 mM.
Ras is an essential signal mediator of NGF (18,19), and its activation mimics the effects of NGF in inducing neuronal differentiation in PC12 cells (37,38). The PCMTras21 cell line, which carries an oncogenic Ha-ras gene under the control of the metallothionein promoter, is induced to differentiate, with extension of neurite processes, by ZnCl 2 . We found that the neuronal differentiation of PCMTras21 cells triggered by ZnCl 2 was also completely inhibited by NAC (Fig. 1A). Thus, NAC is suspected to abolish NGF-induced neuronal differentiation by suppressing signal transduction downstream of Ras activation.
To elucidate the mechanism of the action of NAC in suppressing the differentiation, Northern blot analysis was performed (Fig. 2). Neuronal differentiation by NGF is thought to be mediated by the expression of several genes, such as c-fos, expression of which is regulated by several transcription factors linked to the cellular signaling pathway of NGF (for review, see Ref. 39). Northern blot analysis revealed that NGF rapidly induced c-fos expression in PC12 cells as reported by Greenberg et al. (40). NGF-induced expression of the c-fos gene was completely suppressed when PC12 cells were cultured in the presence of NAC. Furthermore, we investigated the activation of AP-1, which is one of the first transcription factors activated by NGF (for review, see Ref. 39) using the electrophoretic mobility shift assay method (Fig. 3A). AP-1 has been reported previously to be subjected to redox regulation (7) in which NAC increases its binding activity by an increase in AP-1 expression (8). Consistent with this, electrophoretic mobility shift assay showed a slight increase in the DNA binding activity of AP-1 in cells treated with NAC. However, the stimulatory effect of NAC on AP-1 was minimal, and it was not possible to detect an NAC-induced increase in AP-1 activity when using the CAT assay with a reporter plasmid encoding AP-1 binding sites (Fig. 3B). In contrast, the most striking effect of NAC on AP-1 was the suppression of its activation induced by NGF. The DNA binding activity of AP-1 increased rapidly within 30 min after NGF stimulation and reached maximal levels after 60 min in control cells. This AP-1 activation was apparently suppressed in cells treated with NAC (Fig.  3A). The inhibitory effect of NAC on NGF-induced AP-1 activation was confirmed by CAT assay (Fig. 3B). These results indicate that the step(s) inhibited by NAC occurs upstream of the activation of the transcription factors in the NGF signaling pathway.
Activation of MAPK is thought to be occur upstream of the activation of transcription factors in the nucleus in response to stimulation by growth factors (for review, see Ref. 15). MAPK was rapidly activated within 5 min after NGF stimulation, and the activated levels were sustained for a long period in PC12 cells (Fig. 4A), consistent with the previous study (41). MAPK activation was markedly suppressed in cells treated with NAC. Western blotting analysis revealed that the treatment with NAC did not alter the quantity of MAPKs with molecular masses of 41 and 43 kDa in PC12 cells (Fig. 4B, left), whereas the electrophoretically retarded bands of MAPKs (activated MAPKs) induced by NGF were markedly reduced by NAC, indicating that NAC inhibited NGF-induced phosphorylation of MAPKs. The reduction of phosphorylation of MAPKs by NAC was confirmed by blot analysis using anti-phosphotyrosine antibody (Fig. 4B, right). Thus, a step(s) upstream of MAPK in the NGF signaling pathway was thought to be suppressed by NAC.
NAC is a unique compound that acts as an antioxidant both

FIG. 3. Inhibitory effects of NAC on NGF-induced activation of AP-1.
A, DNA binding activity of AP-1 was analyzed by electrophoretic mobility shift assay. PC12 cells were incubated in the presence or absence of 20 mM NAC for 4 h, and then NGF was added to the medium. Nuclear extracts prepared at the appropriate time after NGF stimulation were mixed with 32 P-labeled oligonucleotide probes of the AP-1 binding site and then subjected to electrophoresis through an 8% acrylamide gel. B, transcriptional activation capacity of AP-1 was analyzed by CAT assay. PC12 cells (1 ϫ 10 6 ) on a 6-cm dish were transfected with 2 g of the reporter plasmid pAPCAT, which encodes AP-1 binding sites, using Lipofectin. After transfection, the cells were divided equally between four dishes, and each dish was further incubated for 2 days in the absence or presence of NGF or 30 mM NAC. The upper panel shows one result of TLC in three independent experiments. The lower graph presents the CAT activity as mean Ϯ S.D.

FIG. 4. The effects of NAC on the activation and phosphorylation of MAPK and the effects of other reducing reagents.
A, time course of MAPK activation after NGF stimulation. PC12 cells were incubated in the absence or presence of 20 mM NAC for 4 h and then NGF was added to each culture. At the indicated time after NGF stimulation, cell lysates were prepared and the activity of MAPK was estimated. B, Western blotting analysis of MAPK to detect phosphorylation. Lysates were prepared from cells treated with NGF for 10 min or cells left untreated. Proteins were fractionated through a 12.5% acrylamide gel, and then the blots were screened with anti-MAPK antibody (left) and anti-phosphotyrosine antibody (right), respectively. The phosphorylated MAPKs with molecular masses of 41 and 43 kDa are marked by asterisks. C, GSH-independent inhibition of MAPK by NAC. PC12 cells were cultured for 12 h with or without 20 mM NAC and/or 100 M buthionine sulfoximine. MAPK activity was estimated following incubation with NGF for 10 min (top), and cellular GSH level was measured (bottom). D, comparison of the effect on MAPK activity of the reductants, NAC, ␤-mercaptoethanol and DTT. PC12 cells were incubated with 30 mM NAC, 10 mM ␤-mercaptoethanol, or 10 mM DTT for 4 h, and then the cells were or were not further incubated with or without NGF for 10 min. by its direct reducing activity and by increasing the biosynthesis of the major cellular antioxidant GSH. This gives rise to the question of whether the inhibitory effects of NAC on MAPK are mediated by an increase in the level of GSH. With respect to the contribution of GSH, the effects of NAC were analyzed in cells in which GSH levels had been reduced by treatment with buthionine sulfoximine, a specific inhibitor of ␥-glutamylcysteine synthetase, the rate-limiting enzyme of GSH synthesis. As shown in Fig. 4C, incubation with 20 mM NAC for 12 h increased cellular GSH more than 10-fold, whereas 100 M buthionine sulfoximine completely abolished the increase of cellular GSH and, furthermore, reduced it to less than 10% of the level in normal cells. In these GSH-depleted cells, NAC retained the ability to inhibit NGF-induced activation of MAPK. Thus, the effect of NAC is independent of the increase in GSH levels in the cells and is mediated by its direct reducing activity. We further analyzed the effects of other reductants (␤-mercaptoethanol and DTT), on the activation of MAPK (Fig.  4D). The effects of these compounds were less prominent than the effect of NAC, but it was clear that they also suppressed the activation of MAPK. The major signaling system of NGF has been revealed to be mediated by the MAPK cascade consisting of MAPK and its upstream kinases, MAPKKs and MAPKKKs. The activity of MAPKKK was analyzed as total MAPKK activating activity by the newly developed kinase coupling assay system. In this analysis, GST-MAPKK was incubated with cell extracts in the presence of ATP; the incubation was followed by collection of the activated GST-MAPKK by using GSH Sepharose. Finally, the activity of recovered GST-MAPKK was analyzed using KN-MAPK as a substrate. In this assay system, the amount of phosphorylation of KN-MAPK represents the total cellular MAPKKK activities. 2 This analysis revealed that NAC suppressed the NGF-induced activation of MAPKKK in a dose-dependent manner in the range of concentrations that had inhibitory effects on NGF-induced neurite extension, consistent with the suppression of MAPK (compare Fig. 5 with Fig. 1B). It should be noted that the activity of MAPKKK was potently inhibited below the basal level when the cells were treated with NAC above 10 mM. These results, taken together, suggest that NAC inhibits NGF-induced neuronal differentiation by suppressing the kinases comprising the MAPK cascade.
NAC was able to suppress both MAPK and MAPKKK activity, even in the presence of actinomycin-D or cycloheximide, indicating that the inhibitory effects of NAC on the MAPK cascade were independent of both transcription and translation (Fig. 6). In fact, the action of NAC was so rapid that NGFinduced MAPK activation was apparently decreased in cells incubated with NAC for only 10 min, and treatment with NAC for 1 h was sufficient to develop the full inhibitory effects on MAPK (Fig. 7).
In PC12 cells, three kinases, including Raf-1, B-Raf, and MEKK, are known to act as MAPKKKs (42)(43)(44). Among these MAPKKK species, B-Raf is refractory to inhibition by cAMP, whereas Raf-1 is effectively inhibited in PC12 cells cultured in serum-containing medium, suggesting a difference in the regulation of enzyme activity in these kinases (45). This prompted us to analyze the effects of NAC on each of these MAPKKK species. For this purpose, we attempted to separate the three kinases by fractionation of cell lysates using Mono-Q column chromatography (Fig. 8). Western blotting revealed that these MAPKKK species eluted separately from the column: B-Raf, with a molecular mass of about 97 kDa, was enriched in fractions 2-11; Raf-1, about 70 kDa, was in fractions 2-4 and 7-13; and MEKKs of about 66 and 97 kDa were in fractions 8 -9 and 10 -11, respectively. Faint bands between 97 and 66 kDa on the MEKK immunoblot were nonspecific. Kinase assay revealed that NGF enhanced the MAPKKK activity to twice basal levels in these Mono-Q fractions. This activation was completely suppressed by NAC in all of the fractions, suggesting that Raf-1, B-Raf, and MEKK were equally subject to suppression by NAC. To confirm this, immune complex kinase assays were performed for Raf-1 and MEKK (Fig. 9). This revealed that NGFinduced activation of these two MAPKKKs was suppressed by NAC. We failed, however, to evaluate the effect of NAC on B-Raf because the anti-B-Raf antibody that we used resulted in the activation of the kinase by itself (data not shown). However, it was most plausible that B-Raf was also suppressed by NAC. These results may imply a possibility that the common step(s) for the activation of MAPKKK is sensitive to NAC. Although 2 S. Matsuda, unpublished observations. the precise mechanism is unknown, the exchange of the GDPbound form of Ras to the GTP-bound form (i.e. Ras activation) triggers the activation of these MAPKKKs in the NGF signaling pathway. Therefore, we analyzed GTP/GDP binding to Ras in cell lysates labeled with [ 32 P]phosphorus by using thin-layer chromatography (Fig. 10). In control cells, the amount of GTP binding was less than 5% of the total guanine nucleotides bound to Ras. NGF rapidly activates Ras; the GTP-bound form increased to 20% within 5 min. NAC had no effect on Ras activation by NGF. Furthermore, we could not detect any effect of NAC on the activation of Ras 30 min after NGF stimulation. In accordance with this, NAC inhibited neuronal differentiation induced by the activation of Ras as described above and shown in Fig. 1A. Taken together, these results may indicate that the signal transduction from Ras to MAPKKK is specifically inhibited by NAC. DISCUSSION An antioxidant, NAC, was reported to modulate the activity of several redox-sensitive cellular signal transduction components by its direct reducing activity via a thiol base on the molecule and/or by an increase in cellular GSH levels. In addition, we have demonstrated in this report that NAC suppresses NGF-induced activation of AP-1 and the kinases of the MAPK cascade in accordance with the suppression of neuronal differentiation.
We could not rule out the possibility that a pharmacological effect of NAC unrelated to its reducing activity, if any, would responsible for these suppression; however, it would be most plausible that NAC inhibits MAPK activation by its direct reducing activity. This suggests that NGF signaling and neuronal differentiation of PC12 cells is subject to redox regulation. In fact, high concentrations of ␤-mercaptoethanol and DTT also significantly suppress MAPK activation, although less prominently than NAC. However, of the reducing reagents, only NAC is effective in suppressing differentiation in PC12 cells; the other reductants are cytotoxic when used over a long period. The mechanism by which NAC suppresses NGF signaling is unclear, but there are some possible explanations. One is that NGF stimulates the production of ROIs, which may act as a signal mediator, and the effects of these ROIs are suppressed by NAC acting as an antioxidant or free radical scavenger. Recently, several lines of evidence have been reported showing that ROIs are generated by extracellular stimuli, such as TGF-␤ (46), tumor necrosis factor ␣, ␤FGF (47), CD25 (48), and platelet-derived growth factor (30). In these cases, the blockage of the shift in cellular redox state induced by reducing agents or antioxidant enzymes results in the attenuation of cellular responses to these stimuli. In the case of vascular smooth muscle cells, H 2 O 2 is generated by platelet-derived growth factor stimulation, and both NAC and catalase block the protein phosphorylation and the activation of MAPK elicited by platelet-derived growth factor (30). In addition, it was reported recently that another radical molecule, nitric oxide (NO) triggers a switch to growth arrest during neuronal differentiation in PC12 cells (49). Thus, it can be speculated that the production of ROIs plays an essential role in NGF signaling. To clarify this, we assessed the cellular redox state using a redox-sensitive fluorescent dye, 2Ј7Ј-dichlorofluorescin diacetate, which is oxidized to fluorescent 2Ј7Ј-dichlorofluorescein, as described by Ohba et al. (46). We observed the shifts in cellular redox to the reduced state by 12 h of incubation with NAC, but so far we have been unable to detect any significant changes in redox state after NGF stimulation (data not shown). Thus, the inhibitory effects of NAC in PC12 cells are unlikely to be due to the suppression of ROIs.
NAC partially activates the transcription factor AP-1 (Ref. 8 and see Fig. 3A). Furthermore, the survival-promoting effects of NAC in suppression of apoptosis in PC12 cells are dependent on specific gene expression (50). This leads to another possible explanation that NAC causes the expression of specific gene(s), resulting in the suppression of NGF signaling and neuronal differentiation. However, this possibility is ruled out by the result that the inhibition of MAPK by NAC is not blocked by either cycloheximide or actinomycin-D.
Thus, it is most plausible that the NGF signaling pathway consists of a redox-sensitive step(s) that may be regulated by NAC. The results that NAC suppressed MAPKKKs without affecting Ras activation indicate that the signal transduction from Ras to MAPKKK is the primary step sensitive to NAC. This idea is supported by the observation that NAC is capable of suppressing neuronal differentiation induced by the expression of oncogenic Ras in PC12 cells (see Fig. 1). The uncoupling of all types of MAPKKK species (Raf-1, B-Raf, and MEKK) from Ras may result in suppression of the downstream events in the NGF signaling system.
It has been elucidated that Ras activates Raf, but it still remains to be elucidated how MAPKKK molecules are activated by Ras (for review, see Refs. 51 and 52). So far, it has been revealed that Raf kinase binds to activated Ras (53)(54)(55), and then this kinase is translocated to the plasma membrane and finally activated (56,57). This suggests that the interaction with an unknown component(s) on the membrane is essential for the activation of Raf.
The most important property of NAC is to function as a reducing agent capable of mediating thiodisulfide exchange reactions (58), suggesting that cysteine residues on the target proteins are modulated by NAC. An interesting characteristic of the molecular structure of Raf kinases is the evolutionarily conserved regions, namely CR1, CR2, and CR3. CR1 has a composite structure consisting of a Ras-binding domain followed by a zinc-finger motif at the N-terminal regulatory domain (for review, see Refs. 51 and 52). CR2 is another element involved in the regulatory domain, and CR3 is the C-terminal catalytic domain of the kinase. Several cysteine residues on the zinc-finger motif, Ras-binding domain, and CR3 are highly conserved; thus, these residues are possible sites related to redox regulation. In this respect, the importance of the cysteine residue in the zinc-finger motif for regulation of Raf kinase is demonstrated by the fact that cysteine mutation results in activation of the kinase and abolishment of the Ras-Raf interaction (59). Interestingly, it has been reported that the activity of AP-1 is redox-regulated via the conserved cysteine residue in the DNA-binding domain adjacent to the leucine-zipper motif (7). If this is also the case in Raf kinase, it is plausible that the cysteine residues on the zinc-finger motif of Raf are regulated by redox.
Recently, direct evidence of the redox regulation of Ras has been presented (60). It has shown that NO, a redox-active molecule, binds to a cysteine residue (Cys-118), which is exposed on the surface of the Ras molecule, leading its guanine nucleotide exchange followed by activating the downstream MAPK cascade. Thus, Ras may act as a molecular switch that transfers a redox signal to the nucleus. In this context, it is also possible that NAC reacts with the Cys residue on the Ras molecule, which in turn uncouples the signal transduction to the target molecule. To clarify the redox regulation of NGF signaling, it is necessary to identify which step is inhibited by NAC in the signaling from Ras to Raf.
In this study, we revealed that the NGF signaling pathway is subject to a complicated redox regulation in which NAC activates AP-1 while it suppresses MAPKKK. A similar composite system that is modulated by redox has also been reported in the signaling to NFB. NFB is a redox-regulated transcription factor, the DNA binding of which is abolished by oxidation, whereas it is fully activated by reduction (10,11). However, several reports have revealed that antioxidants (including NAC) suppress the activation of NFB elicited by extracellular stimuli, such as tumor necrosis factor ␣, phorbol ester, interleukin 1, and lipopolysaccharide (8,9,12), suggesting that NAC suppresses an unknown redox-sensitive kinase(s) upstream of NFB (61)(62)(63)(64). This suggests that NFB signaling may be subject to a dual redox regulation in which the reductant(s) directly activates NFB while it suppresses upstream kinase(s). Thus, the signaling pathways to AP-1 and NFB are both modulated by a dual redox regulation system. Interestingly, many transcription factors, such as Myb (65), are also reported to be regulated by redox. It might be that most signaling pathways linked to transcription factors are subject to dual redox regulation.