ATP depletion affects the phosphorylation state, ligand binding, and nuclear transport of the 4 S polycyclic aromatic hydrocarbon-binding protein in rat hepatoma cells.

In the rat, cytochrome P4501A1 gene expression is thought to be regulated by several trans-acting factors including the 4 S polycyclic aromatic hydrocarbon (PAH)-binding protein. Phosphorylation and dephosphorylation have been suggested to influence the function of many cytosolic receptors and transcription factors. The ATP level within H4IIE rat hepatoma cells could be depleted by treatment with sodium azide or 2,4-dinitrophenol; restoration of the original ATP levels occurred with addition of glucose to the cell culture. ATP depletion reduced the phosphate content of the 4 S protein by ~25-30%, which lowered the binding of benzo[a]pyrene (B[a]P) to the 4 S protein by >60%. This effect could not be reversed by the addition of ATP to the binding reaction mixtures. Alkaline phosphatase treatment of the purified 4 S protein in a cell-free system also reduced the B[a]P binding to the protein. Cells treated with a protein phosphatase inhibitor, okadaic acid, and a protein kinase inhibitor, staurosporin, affected the B[a]P binding of the 4 S protein positively and negatively, respectively,. These data suggested that phosphorylation is involved in the interaction of the 4 S protein with the PAH. The nuclear translocation of the predominantly cytosolic binding protein has been investigated after ligand binding. Western blots with the immunopurified 4 S PAH-binding protein from cytosolic and nuclear lysates showed significant differences in the distribution of the 4 S receptor between cytosolic and nuclear compartments in control and ATP-depleted cells. Ligand binding stimulated the movement of the receptor into the nucleus, which was completely blocked by reducing the intracellular ATP concentration. These findings provide new information on the role of ATP and phosphorylation on the interaction of B[a]P with 4 S PAH-binding protein and its nuclear translocation.

The cytochrome P450 monooxygenases catalyze biotransformation reactions involving both endogenous and exogenous substrates (1,2). One of the members of the cytochrome P450 family, CYP1A1, 1 is responsible for the the bioactivation of PAHs such as B[a]P to metabolites that may play a role in environmental carcinogenesis (3). The expression of CYP1A1 is significantly elevated in the livers of rats that have been exposed to PAHs such as 3-methylcholanthrene and B[a]P. Evidence exists that such induction may be facilitated at least in part through a cytosolic receptor, the 4 S PAH-binding protein (4 -10). The 4 S protein specifically binds certain PAHs with high affinity and in a saturable manner (4,(11)(12) and undergoes a nuclear translocation (4,12,13). The activity of the 4 S protein in cultured rat liver cells is positively correlated with the extent of induction of aryl hydrocarbon hydroxylase (13). The cytosolic 4 S protein has recently been identified as glycine N-methyl transferase (GNMT) (14), an enzyme that catalyzes the synthesis of sarcosine from S-adenosylmethionine and glycine and serves as a folate carrier as well. It has also been demonstrated that the cytosolic 4 S PAH-binding protein serves as a transcriptional activator of CYP1A1 expression (15). Thus, the 4 S protein belongs to a class of substances that exhibit multiple functions, as an enzyme and as an activator of transcription.
The activity of many factors is modulated by posttranscriptional events, with phosphorylation representing a major mechanism for regulating gene expression in eukaryotic cells (16,17). Almost every eukaryotic transcription factor analyzed in detail has been shown to undergo phosphorylation. The state of phosphorylation of a number of transcription factors, including yeast GAL4 (18), HSF (19), STE 12 (20), and mammalian Sp1 (21) and Oct-2 (22) has been correlated with the extent of their activation. In many cases, however, the functional consequence of such phosphorylation, if any, remains largely unknown.
Phosphorylation may affect gene regulation in a number of ways. Phosphorylation may increase the affinity of the transactivation domain for a component of the basal transcriptional machinery or for a protein that mediates its interaction with the initiation complex (23). Phosphorylation may repress gene activity by causing the dissociation of the activation domain from an inhibitory protein. It has been suggested that cAMPdependent protein kinase-mediated phosphorylation inactivates ADR-1 by preventing its ability to interact with the general transcriptional machinery while bound to DNA (24). The involvement of phosphorylation in trans-repression of c-fos has been suggested (25,26). Phosphorylation may affect the specificity of binding of the transcription factors to their targets (27).
A computer search performed on the primary amino acid sequence of the 4 S protein did indicate some sites for potential posttranslational modification, which could have biological sig-nificance. Putative phosphorylation sites for casein kinase II in the 4 S protein exist at threonines 68 and 83 and at serines 147 and 250. The serine at position 88 of the 4 S protein may serve as a substrate for protein kinase C. It has already been shown that GNMT is an excellent substrate for the catalytic subunit of cAMP-dependent protein kinase in an in vitro phosphorylating system (28), which results in an increase of 2-fold in enzyme activity. However, it is not known whether phosphorylation has any measurable effect on the PAH binding activity of the 4 S protein.
Here, we report that the depletion of ATP in intact cells dephosphorylated the protein by ϳ25%, caused a loss in ligand binding activity, and retarded its nuclear import; restoration of cellular ATP levels reversed these events.

EXPERIMENTAL PROCEDURES
Materials and Cells- 35  (v/v) fetal bovine serum and gentamycin sulfate/penicillin-streptomycin-ampicillin (1 mg/ml). They were passaged every 2 days after dissociation of the cells by treatment with 0.25% trypsin. For the ATP depletion experiments, cells were grown to about 70% confluency in ␣-MEM containing 10% fetal bovine serum and gentamycin sulfate/ penicillin-streptomycin-ampicillin. The cells were washed with phosphate-free Dulbecco's modified Eagle's medium and incubated in the same medium for 12 h at 37°C. The cells (in some experiments dually labeled with 35 S and 32 P) were harvested by scraping, washed three times at 37°C with 50 ml of phosphate-free KRBH (25 mM Hepes, pH 7.4, 25 mM NaHCO 3 , 120 mM NaCl, 4.95 mM KCl, 2.54 mM CaCl 2 , 1.2 mM MgCl 2 , 1.2 mM potassium phosphate monobasic, and bovine serum albumin (2 mg/ml)) each for 20 min, divided into two equal aliquots, and resuspended in the same buffer. The control cells were incubated in KRBH and 22 mM glucose (glucose was added to maintain the ATP level), whereas the treated group was placed in KRBH and 20 mM sodium azide or 1 mM 2,4-dinitrophenol. After this treatment, one-half of the cells from the control and treated groups were transferred to a conical tube, washed twice with KRBH at 37°C to remove the azide, resuspended in KRBH and glucose, and incubated again for 15-20 min. In some experiments, the protein phosphorylation inhibitor staurosporin and the phosphatase inhibitor okadaic acid were added to the culture medium; these inhibitors remained in the reaction mixtures throughout the incubation time. All the samples were incubated at 37°C for 90 min in a humidified atmosphere of 5% CO 2 and 95% O 2 . After centrifugation, the cells were suspended in 5 packed cell volumes of ice-cold TEDGP (40 mM Tris, pH 7.4, 1 mM disodium EDTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, and 10% glycerol) and homogenized using 15-20 strokes of a Dounce homogenizer. Cell lysis was checked periodically by light microscopy. The cell lysates were centrifuged for 1 h at 100,000 ϫ g at 4°C. ATP analysis was conducted by the luciferase method using an assay kit, as detailed by the manufacturer (Sigma). Cell viability under different experimental conditions was checked by a fluorometric assay using a Live/Dead Eukolight Viability/Cytotoxicity Kit from Molecular Probes Inc. (Eugene, OR) as well as by the trypan blue exclusion method.
Assay for Specific Binding of B[a]P-Specific binding activity was assessed by sucrose gradient analysis as described previously (14). The 100,000 ϫ g supernatant fraction (0.5-1 mg of protein) was incubated for 1 h at 4°C with 10 nM [ 3 H]B[a]P with or without a 200-fold excess of unlabeled B[a]P. In most of these experiments, 95% ethanol was used as the solubilizing vehicle for the PAH. The final concentration of ethanol did not exceed 2%, a level that had no effect on the specific binding activity (29). In some experiments, the 4 S protein was immunopurified from control and azide-treated cells and analyzed for B[a]P binding activity. Addition of glucose to either cells or reaction mixtures did not affect the B[a]P binding activity of the 4 S protein. In some cases, immunopurified 4 S protein from control samples was treated with alkaline phosphatase as described by Nielsen et al. (30) and was analyzed for B[a]P binding activity.
Whole Cell Protein Labeling-For the amino acid labeling experiments, approximately 1 ϫ 10 5 cells were incubated for 20 h in a 75-cm 2 flask (Costar) with 50 ml of methionine-free RPMI 1640 medium containing 25 mM Hepes, pH 7.4, and L-[ 35 S]methionine (0.3 mg, 400 Ci). The cells were washed three times at 37°C with 40 ml of KRBH, resuspended in KRBH, and divided into two aliquots: (a) one treated with a final concentration of 20 mM sodium azide for 90 min to generate dephosphorylated 4 S PAH-binding protein; and (b) one that received glucose as a control. Cells were collected by centrifugation and lysed by two rounds of freezing and thawing. The cytosolic binding activity was measured in the supernatant obtained after centrifugation for 5 min at 400 ϫ g. The washed pellet was then resuspended in 2 ml of 0.4 M NaCl in TEDGP and extracted for 25 min by gentle tumbling. After centrifugation for 10 min at 400 ϫ g, the resultant pellet was reextracted by the same procedure, and the supernatants were pooled. These supernatants (4 ml) were dialyzed against TEDGP at 4°C for 30 min. Nuclear B[a]P binding activity was determined. Each sample was also immunoprecipitated with an antibody against GNMT as described below.
For the dual labeling experiments, cells were preincubated for 1 h at 37°C in phosphate-free Dulbecco's modified Eagle's medium containing 10% dialyzed calf serum. Approximately 10 7 cells were incubated for an additional 18 h in phosphate-free Dulbecco's modified Eagle's medium containing 10% dialyzed calf serum, 20 Ci/ml [ 32 P]orthophosphoric acid, and 5 Ci/ml [ 35 S]methionine. In some experiments, B[a]P was added during the last 30 min of the incubation period.
Immunoprecipitation-Cytosolic and nuclear extracts (ϳ0.5 mg) were treated by addition of 50 l of a 1:1 slurry of protein G-Agarose Plus in 1 ml of TNET (1% Triton X-100, 150 mM NaCl, 2 mM EDTA, 20 mM Tris, pH 8.0, and bovine serum albumin (2 mg/ml)). After 30 min at 4°C, the samples were centrifuged for 5 min at 15,000 ϫ g, the supernatant was transferred to a fresh tube, and affinity-purified antibodies (ϳ1.0 g) against the 4 S protein were added. The antibody concentration was optimized by titration to ensure complete removal of the 4 S protein from the lysates. After incubation at 4°C for 1 h, 25 l of protein G-Agarose Plus was added, and incubation continued overnight. Immunoprecipitates were washed five times in TNET, and 1:1 SDS loading sample buffer was added. Samples were heated in a boiling water bath for 5 min and centrifuged, and supernatants were analyzed by SDSpolyacrylamide gel electrophoresis. In some experiments in which B[a]P binding activity of the protein was to be determined, the pH of the solution was raised to 11-12, which resulted in the elution of the protein into the supernatant. Before analyzing the protein for B[a]P binding, the pH was restored to 7.4.

SDS-Polyacrylamide Gel Electrophoresis and Western Analysis-
The products of immunoprecipitation were analyzed by SDS-polyacrylamide gel electrophoresis on 10% polyacrylamide gels. For the 35 S-labeling experiments, the polyacrylamide gels were exposed to phosphorimaging screens, which were then developed on a PhosphorImager after 24 h of exposure. In some experiments, portions of the gels that contained 4 S protein were sliced into 2-mm sections, and the radioactivity was extracted in Solvable TM ; the radioactivity was counted in a liquid scintillation spectrometer. For the Western blotting analysis, the polyacrylamide gels were electrolytically transferred to Immobilon P membranes that had been presoaked for 30 s in methanol. The electroblotted proteins were probed with affinity-purified antibodies against the 4 S receptor. The protein was visualized by chemiluminescence as detailed by the manufacturer of the BM Chemiluminescence Western blotting kit.

Effect of Sodium Azide and 2,4-Dinitrophenol Treatment on the ATP Level and Survival of the H4IIE Cells-
The ATP level in H4IIE cells that were treated with sodium azide was rapidly depleted (Fig. 1) without affecting the cell viability. 2,4-Dinitrophenol treatment resulted in a similar reduction in intracellular ATP (data not shown). Incubation of either control or treated cells for up to 90 min in KRBH medium resulted in Ͼ70% viability as determined by the fluorescent assay and trypan blue exclusion methods. During this time, the ATP concentration in control cells remained at approximately 80% of the 0 time value. The ATP concentration in azide-treated cells, however, fell to 45% and Ͻ5% of the 0 time value by 15 and 90 min, respectively. The kinetics of the reduction in intracellular ATP in the H4IIE cells as a result of the azide or 2,4dinitrophenol treatment was similar to that reported for lymph-oid cells (31). Based on our results, 90 min of treatment with sodium azide was chosen for all the subsequent experiments.

Generation of Dephosphorylated 4 S Protein and the Effects on the Specific Binding of B[a]P-
The purpose of these experiments was to ascertain the effects of the dephosphorylation of the 4 S protein on the high affinity and saturable binding of B[a]P to the 4 S protein. The H4IIE cells, dually labeled with 35 S and 32 P, were treated with sodium azide for 90 min, and 32 P content (normalized with 35 S) of the immunopurified 4 S protein was determined in control and sodium azide-treated cells (Fig. 2). Approximately a 25-30% reduction in the phosphate content of the 4 S protein was observed in the azide-treated cells. In these studies, we confirmed that complete immunodepletion of the 4 S protein had occurred in the treated lysate by probing the supernatants obtained after removing the immunoprecipitates with antibodies against the 4 S protein and analysis by Western blotting; no 4 S protein was detected (data not shown).  B[a]P binding to the 4 S protein (100,000 ϫ g supernatant) in control and sodium azide-and 2,4-dinitrophenol-treated H411E cells was analyzed by sucrose density sedimentation techniques (Fig. 3A). The specific binding of B[a]P to the control cytosol was approximately 7,500 cpm/mg protein. The specific interaction of B[a]P to the 4 S protein in cytosol obtained from sodium azide-or 2,4-dinitrophenol-treated cells was reduced by approximately 65%. When the 4 S protein was immunopurified and assayed for B[a]P binding activity from the above samples, similar results were obtained (Fig. 3B). Alkaline phosphatase treatment of the control samples also reduced the B[a]P binding activity of the protein (Fig. 4). Addition of sodium arsenate, a specific inhibitor of alkaline phosphatase, reversed this effect. Restoration of the ATP level to the original control value by incubation of the H4IIE cells in glucose-containing medium for 30 min resulted in a prompt return of binding activity (Fig. 5). Similar results have been obtained when the human hepatoma HepG2 cells were used as a source of the 4 S PAH-binding protein (data not shown).

Effect of Staurosporin and Okadaic Acid on B[a]P Binding to the 4 S Protein-The previous results suggested that phosphorylation of the 4 S protein is involved in the specific interaction of B[a]P to the 4 S PAH-binding protein.
Since phosphorylation and dephosphorylation probably occur continuously within these cells, we determined the effects of a protein kinase inhibitor, staurosporin, and a phosphatase inhibitor, okadaic acid, on the specific binding. Staurosporin reduced the binding activity of the protein by ϳ50%, whereas okadaic acid, a selective phosphatase type 1 and 2A inhibitor, increased it by approximately 2-fold (Fig. 6). A titration showed that only 10 nM okadaic acid was needed to increase the specific binding activity with B[a]P (data not shown), suggesting a role for the type 2 and 2A kinases in the phosphorylation of the 4 S PAHbinding protein.
ATP Depletion and Nuclear Translocation of the 4 S PAHbinding Protein-After interaction with B[a]P, the 4 S protein has been reported to translocate into the nucleus (4,12,13). In addition, we (14) and the laboratory of Wagner et al. (28) have demonstrated the presence of the 4 S protein and GNMT, respectively, in liver nuclei. We therefore explored the role of ATP in the translocation mechanism of the receptor. Fig. 7 shows the kinetics of the specific binding of B[a]P to the 4 S protein.
In the cytosol, B[a]P interaction with the binding protein reached a maximum by 60 min after addition to the cell culture. An additional 100 min were required before achieving peak specific binding of B[a}P. We have previously shown that a steady state level of CYP1A1 mRNA in H4IIE cells after B[a]P induction gradually increased during this period, reaching a maximum by 6 -10 h (32).
The effects of ATP depletion on B[a]P-protein translocation were examined by treating the cells with sodium azide in presence or absence of B[a]P. The 4 S PAH-binding protein was immunoprecipitated from the cytosolic and nuclear compartments, and the precipitates were solubilized and analyzed by Western blotting on SDS denaturing gels (Fig. 8). In azidetreated cells induced with B[a]P, translocation of the 4 S protein into the nucleus was completely inhibited (Fig. 8, compare  and d). In control cells (no B[a]P treatment), the 4 S protein was located predominantly in the cytosol (Fig. 8, lanes e and f) independent of the ATP status. To confirm these observations, the cells were labeled with [ 35 S]methionine and depleted of ATP by azide treatment in the presence or absence of B[a]P. Immunoprecipitated 4 S protein from both the cytosolic and nuclear compartments showed a similar distribution pattern (data not shown).

DISCUSSION
Polycyclic aromatic hydrocarbons specifically bind to the 4 S protein, and the ligand-bound "receptor" translocates into the nucleus, wherein it modulates CYP1A1 expression as a transactivator (15). The 4 S protein, like a number of other proteins, fulfills multiple functions within tissues. It may: (a) catalyze sarcosine formation from glycine and S-adenosylmethionine, fulfilling its enzymatic role as a methyltransferase (33); (b) serve as a folate carrier (34); and (c) bind PAHs such as B[a]P in a traditional ligand-receptor manner (14) in carrying out a transcriptional activation mission. However, the enzymatic and activator functions are conducted by different subunit configurations. Methyltransferase activity requires the tetramerization of the protomeric 33-kDa subunit (34), whereas the role of a transcriptional activator requires either the dimeric or monomeric form (14); the tetramer is inactive in this regard. Consequently, the conversion of monomer (or dimer) to tetramer is important in determining the function of the resultant molecule.
In this article, we have shown that ATP depletion partly dephosphorylates the 4 S protein, which affects its interaction with B[a]P. Addition of ATP to the reaction mixtures containing dephosphorylated 4 S protein and [ 3 H]B[a]P did not restore the binding activity of the protein. This observation indicates that ATP possibly acts through the phosphorylation of the 4 S protein. One possible explanation for selective dephosphorylation is that the phosphates at those sites turn over more rapidly than at moieties at other sites and thus are more rapidly lost in ATP-depleted cells. Alternatively, this dephosphorylation may also be the result of activation of specific phosphatases in ATP-depleted cells. The use of sodium azide and 2,4-dinitrophenol as ATP-depleting agents in eukaryotic cells is widely reported (31,(35)(36)(37)(38). In the case of the glucocorticoid receptor, azide treatment of the host cells selectively dephosphorylates serines 220 and 234 (37). Decreased binding activity of the immunopurified 4 S protein from azide-treated cells eliminates the involvement of the nonspecific effects of azide treatment. Reverse effects of staurosporin, a protein phosphorylation inhibitor, and okadaic acid, an inhibitor of protein phosphatase on the B[a]P binding of the 4 S protein, reinforce the involvement of phosphorylation in the ligandreceptor interaction. Dephosphorylation of the purified 4 S protein from the rat liver by alkaline phosphatase treatment and its effect on the B[a]P binding also indicates the involvement of the phosphorylation in regulating the activity of this receptor. The process of rephosphorylation of the 4 S protein appears to occur quickly in the H4IIE cells, since introduction of glucose into the medium resulted in a rapid restoration of the ATP level to the formation of phosphorylated protein and to the recovery of B[a]P binding activity, all within 30 min.
Furthermore, the reduction in intracellular ATP exerts a marked effect on the cytosol-to-nucleus trafficking. Entry into the nucleus is blocked under these conditions. As ATP depletion decreases the ligand binding by approximately 60%, whereas nuclear transport is completely inhibited, this block in the nuclear transport may be partly the result of decreased binding of the dephosphorylated protein and/or of the involvement of ATP in some energy-dependent translocation of the ligand and receptor from the cytosol to the nucleus. Alternatively, unknown phosphorylation events participating in other cellular processes may be involved in the regulation of ligandreceptor trafficking from the cytosol to the nucleus. Phosphorylation-dependent redistribution of receptors between the cytosolic and nuclear compartments has been reported previously. Modification of either the protein itself or a cytoplasmic  anchor protein would allow energy-dependent translocation through the nuclear pore complex (39). For example, in the case of the T antigen, phosphorylation of serines 111 and 112 by casein kinase II accelerated nuclear import (40). It has been reported that stimulus-dependent phosphorylation may cause the nuclear translocation and therefore activation of transcriptional factors such as NF-AT (41), NFIL-6 (42), and ISGF3 (43,44). In addition, stimulus-dependent translocation (and activation) of the protein kinases themselves into the nucleus, wherein phosphorylation of certain factors would then occur, has been reported to regulate activity of those transcriptional proactivators that are normally resident in this compartment (45)(46).
Cytosolic receptors, such as the glucocorticoid receptor, also participate in an ATP-dependent cycle that involves phosphorylation (37). These hormone receptors exist in a phosphorylated state in the absence of ligand but can be hyperphosphorylated in the presence of the specific hormone or agonist (47). The phosphorylation took place on predominantly serine moieties, although some threonines were also involved (48). These phosphorylation sites were found predominantly in the N-terminal domain.
It is of interest that the enzymatic function of GNMT is also enhanced by phosphorylation in vitro with cAMP-dependent protein kinase (28). This implies that phosphorylation must occur at the tetramer level as well. Whether the phosphorylation of the tetramer is accompanied by a translocation from the cytosol to the nucleus is not yet known. Furthermore, it is not known whether the phosphorylation of the tetramer occurs at the same site(s) as the modification of the monomer or dimer. It is tempting to speculate that phosphorylation of the monomer (or dimer), as affects PAH transport, or of the tetramer, which may influence enzymatic activity, is required for the flow of traffic from the cytosol to the nucleus. However, interaction of the protein with B[a]P may occur at the level of the tetramer, which then stimulates disaggregation to either the monomer or dimer state, i.e. the transcriptional activator state, or may take place at the monomer or dimer form, which then prevents oligomerization to the tetramer. In either case, the rate-limiting factor in transcriptional activation of CYP1A1 expression is not the amount of protein per se but the concentration of the monomer (or dimer), which appears to be dependent on the concentration of B[a]P.
In addition to the ease of interaction with other regulatory proteins, the extent of dimerization, and subcellular localization, the posttranslational phosphorylation of the 4 S protein may also affect its stability or the extent of DNA binding, all of which could influence the overall transcriptional activation property of this and other trans-factors (49). In the case of c-jun, phosphorylation at one set of sites negatively regulated the interaction with DNA, whereas modification at another sequence increased the trans-activation (50,51). The phosphorylation may also influence the biological property of the protein by inducing allosteric changes as well as electrostatic repulsion, both of which had been reported previously with other regulators (52,53).
In conclusion, we have demonstrated that the intracellular level of ATP and the phosphorylated state of the 4 S protein appear to govern the ligand-receptor interaction and entry of this transcriptional activator complex from the cytosol to the nucleus, wherein it can interact with appropriate cis-elements within CYP1A1. Which sites in the 4 S protein are phosphorylated, what effect phosphorylation and ligand interaction have on the state of oligomerization, and how phosphorylation stimulates translocation and the subsequent interaction with responsive genes remain to be established.