cAMP-dependent negative regulation of rat aldehyde dehydrogenase class 3 gene expression.

We investigated the inhibitory effects of intracellular cyclic adenosine monophosphate (cAMP) levels in regulating class 3 aldehyde dehydrogenase (aldh3) gene expression using cultures of primary rat hepatocytes and transient transfection experiments with HepG2 cells. In addition to regulation by an Ah receptor-dependent mechanism, expression of many members of the Ah gene battery have been shown to be negatively regulated. As was seen for the cytochrome P450 (cyp1A1) gene, aldh3 is transcriptionally inducible by polycyclic aromatic hydrocarbons (PAH), and this induction involving function of the arylhydrocarbon (Ah) receptor is inhibited by the protein kinase C (PKC) inhibitors, 1-(5-isoquinolinesulfonyl)-2-methylpiperazine di-HCl (H7) and staurosporine. However, PAH induction of ALDH-3 activity, protein, and mRNA was potentiated 2-4-fold by addition of the protein kinase A (PKA) inhibitors, N-(2-(methylamino)ethyl)-5-isoquinolinesulfonamide di-HCl (H8) and N-(2-guanidinoethyl)-5-isoquinolinesulfonamide HCl (HA1004). These PKA inhibitors had no effect on the PAH induction of the cyp1A1 Protein kinase A activity of cultured hepatocytes was specifically inhibited by H8 and HA1004 in a concentration-dependent manner, but not by H7, and there was an inverse correlation observed between potentiation of PAH-induced aldh3 gene expression and inhibition of specific PKA activity by the PKA inhibitors. The cAMP analog dibutyryl cAMP, the adenylate cyclase activator forskolin, and the protein phosphatase 1 and 2A inhibitor okadaic acid all dramatically inhibited both PAH induction and H8 potentiation of PAH induction of aldh3 expression but had no effect on induction of cyp1A1 expression in cultured hepatocytes. Both basal and PAH-dependent expression of a chloramphenicol acetyltransferase expression plasmid containing approximately 3.5 kilobase pairs of the 5′-flanking region of aldh3 (pALDH3.5CAT) were enhanced 3-4-fold by the PKA inhibitor H8 but not by the PKC inhibitor H7 (>20 μM). cAMP analogs, activators of PKA activity, or protein phosphatase inhibitors diminished expression of the reporter gene in a manner identical to the native gene in cultured rat hepatocytes. Using deletion analysis of the pALDH3.5CAT construct, we demonstrated the existence of a negative regulatory region in the 5′-flanking region between −1057 and −991 base pairs which appears to be responsible for the cAMP-dependent regulation of this gene under both basal and PAH-induced conditions. At least two apparently independent mechanisms which involve protein phosphorylation regulate aldh3 expression. One involves function of the Ah receptor which requires PKC protein phosphorylation to positively regulate both aldh3 and cyp1A1 gene expression and the other a cAMP-responsive process which allows PKA activity to negatively regulate expression of aldh3 under either basal or inducible conditions.

The aldehyde dehydrogenases (ALDH, 1 aldehyde NAD(P) ϩ oxidoreductase EC 1.2.1.3) are a family of NAD(P) ϩ -dependent enzymes that oxidize a broad class of aldehydes to their carboxylic acids. The family is divided into at least three classes based on sequence similarity, and the class 3 aldehyde dehydrogenase gene is expressed in a tissue-specific manner in microsomal and cytosolic fractions (1,2). The highest level of basal expression of aldh3 2 is seen in corneal epithelium, stomach, and heart, whereas PAH-induced expression is seen principally in liver, lung, bladder, colon, spleen, and thymus of rodents. This gene is also expressed at high levels in neoplastic tissue and some cell lines. Takimoto et al. (5) and Xie et al. (6) have characterized the 5Ј-flanking region of the aldh3 gene and demonstrated that it contains at least three major functional domains: a strong promoter proximal to the transcription start site, an inhibitory region just upstream of the promoter, and a PAH-responsive enhancer region. The transcription of the aldh3 gene appears to be controlled by cooperation of at least these three functional domains.
Expression of this gene in liver is mediated by the arylhydrocarbon (AhR) receptor, a cytosolic protein capable of binding PAH as ligands (7). After ligand binding, the AhR-ligand complex is translocated into the nucleus, forms a heterodimeric complex with the arylhydrocarbon nuclear transporter (ARNT), and they interact with specific DNA sequences, designated xenobiotic responsive elements (XREs), to alter the transcription of specific genes. The AhR mediates induction of a number of xenobiotic metabolizing enzymes (termed the Ah gene battery), including cytochrome P4501A1, cytochrome P4501A2, glutathione S-transferase Ya 1 (GST1), NAD(P)H: quinone oxidoreductase, (QOR), UDP-glucuronosyltransferase 1.6 (UGT1A6), and class 3 aldehyde dehydrogenase (8). The first gene whose activation was shown to be directly mediated by the AhR was cyp1A1 (7,9); multiple XREs have been identified in the 5Ј-upstream regulatory region of the cyp1A1 gene. Although there are sequences of the aldh3 PAH-responsive enhancer nearly identical to the XREs in the cyp1A1 flanking sequences, one functional aldh3 XRE is located much farther upstream than those seen in either cyp1A1, glutathione Stransferase Ya 1 , or NADPH-quinone oxidoreductase genes (5,10,11). A second feature of genes in the Ah gene battery is the low levels of constitutive expression that appears to be due to negative control mechanisms (12). Several different mechanisms have been described for the negative regulation of cyp1A1, including putative regulatory genes on mouse chromosome 7 (12), regulation by nucleosome structure (13), and specific negative regulatory transcription factors (14).
Several studies have suggested that protein kinase C (PKC) plays an important role in the regulation of cyp1A1 gene expression. Carrier et al. (15) and Reiners et al. (16) have provided evidence that some step in transactivation by the AhR is dependent upon phosphorylation by PKC. Our studies (17) with cultured rat hepatocytes demonstrated that the PKC inhibitors H7 and staurosporine concomitantly inhibited PAH induction of all five genes we tested, including aldh3, cyp1A1, gst1, qor, and ugt1. 6 (ugt1A6). This result suggests that PKC-dependent phosphorylation of the AhR is required for PAH induction of all five xenobiotic metabolizing enzymes. The protein kinase A inhibitors H8 and HA1004 were without effect on all of these genes at equivalent concentrations, except PAH induction of ALDH-3 mRNA and protein which was stimulated 2-4-fold (17). In this study, we sought to characterize the synergistic effects of the PAH, 1,2-benzanthracene, and protein kinase A inhibitors, H8 and HA1004, on aldh3 gene expression and the mechanism whereby PAH regulates levels of the ALDH-3 enzyme in cultured adult rat hepatocytes. Utilizing reporter genes containing the 5Ј-flanking region of the aldh3 gene transfected into HepG2 tumor cells, we demonstrate the existence of a cAMP-dependent negative cis-acting element.

Methods
Primary Hepatocyte Cell Culture-Hepatocytes were routinely prepared from male adult Sprague-Dawley rats (180 -250 g, (Hsd:Sprague-Dawley SD), Harlan Sprague-Dawley, Indianapolis, IN) by in situ liver collagenase perfusion (18,19) with modification (17). At 24 h in culture, inducing agents were added to the fresh media. H7, H8, HA1004, dibutyryl cAMP, forskolin, and okadaic acid were added 1 h prior to addition of BA. Equivalent amounts (Յ1.0% v/v) of solvent were added to control cells. At the desired times, the media were removed from the dishes by aspiration, the cells were washed with Dulbecco's phosphatebuffered saline (2 ϫ 1 ml), and cells were harvested (17). Protein concentrations were determined utilizing bicinchoninic acid with bovine serum albumin as a protein standard (20). The samples were stored at Ϫ70°C and analyzed within 2-4 days for enzyme activity. The proteins were stable for over 6 months at Ϫ70°C when analyzed by Western immunoblot analyses.
Enzyme Assays-Cytochrome P4501A1 activity was determined using the specific substrate 7-ethoxyresorufin (21). Aldehyde dehydrogenase 3 activity was measured by monitoring the increase in absorbance at 340 nm caused by NADPH production during the oxidation of benzaldehyde as a substrate (22). Protein kinase A activity was measured as described by using the protein kinase A assay system obtained from Life Technologies, Inc.
Northern Analysis-Total RNA was isolated by modification of the method of Chomczynski and Sacchi (23) as described previously (17). Northern blot experiments were performed after size-fractionation of the denatured RNA (25 g) on formaldehyde-containing 1% agarose gels and transfer of the RNAs to Zetaprobe membranes by diffusion (24). Hybridization was carried out at 43°C overnight in 0.25 M sodium phosphate buffer, pH 7.2, containing 0.25 M sodium chloride, 50% formamide, 7% SDS, 1 mM EDTA, and 32 P-labeled cDNA probe. The hybridized membranes were washed three times for 15 min with 2 ϫ SSC and 0.1% SDS, 0.5 ϫ SSC, and 0.1% SDS or 0.1 ϫ SSC and 0.1% SDS at room temperature. The washed membrane was used to expose x-ray film at Ϫ70°C with an intensifying screen (DuPont NEN). The exposed x-ray film was developed, and the optical density of the relevant bands was quantitated by densitometry using a Bio-Rad Model 620 video densitometer (La Jolla, CA).
Probes Used for mRNA Analysis-Cytochrome P4501A1 mRNA was measured using a 635-bp PstI fragment of the plasmid pA8 for cyp1A1 provided by R. N. Hines, Department of Pharmacology and Toxicology, Wayne State University (25). Aldehyde dehydrogenase 3 mRNA was measured using a EcoRI/HindIII cDNA fragment of the plasmid pselALDHX, the aldh3-specific clone (1). As a control, ␤-actin mRNA levels were measured using the cDNA plasmid encoding the mouse cytoskeletal ␤-actin (26). These nuclei acid probes were labeled with [ 32 P]dCTP using the random primer labeling procedure (27).
Preparation of CAT Constructs-A CAT construct (pALDH3.5CAT) containing approximately 3.5 kb of 5Ј-flanking region of the aldh3 gene containing functional XREs (5) was used to test for the effects of PAH and PKA inhibitors/activators. A series of deletion mutants of the proximal 5Ј-flanking region of aldh3 were prepared by polymerase chain reaction as described previously (6), namely pI-3, pI-4, pI-5, pI-6, and pI-8. Construct p⌬(Ϫ1054/Ϫ392)ALDH was produced by treating pALDH3.5 with the restriction enzyme PstI to remove a portion of the flanking region from Ϫ1057 bp to Ϫ392 bp relative to the transcription start site; the resulting plasmid was gel purified and religated to yield p⌬(Ϫ1057/Ϫ392)ALDH. Construct p⌬(Ϫ1057/Ϫ930) was prepared using polymerase chain reaction to generate a fragment from Ϫ374 bp to Ϫ930 bp (upstream primer 5Ј-GGAGGACAAAGTGTTGCTATG-3Ј; downstream primer, 5Ј-AGCTGCTGTTCTCTGAGTCC-3Ј). The polymerase chain reaction fragment was incorporated into a pCRII vector and amplified in bacteria. The PstI fragment was liberated from the pCRII vector prior to religation into the PstI-cleaved vector, p⌬(Ϫ1057/ Ϫ392)CAT, to yield construct p⌬(Ϫ1057/Ϫ930) which was identified by EcoRI restriction analysis.
Transient Transfection Assays-HepG2 cells (ATCC, Bethesda) were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, fungizone (2.5 g/ml), penicillin (10 units/ml), and streptomycin (10 units/ml). The cells were grown at 37°C in a 5% carbon dioxide, 95% air atmosphere. The cells were grown to approximately 30 -40% confluence, transfected with 2 g of pCMV-␤gal as a transfection control, and 4 g of pALDH3.5CAT or other reporter plasmid in T25 flasks using the calcium phosphate precipitation/glycerol shock method and harvested as described by Rushmore et al. (10). The chemicals used were prepared as 500 ϫ stock solutions in dimethyl sulfoxide or water, and they were added 24 h after the shock treatment. The chloramphenicol acetyltransferase assay used was a modification of the method of Gorman et al. (28) which includes a xylene extraction of the products from [ 3 H]chloramphenicol and subsequent quantitation of products and reaction mixture by liquid scintillation methods. The reactions were performed in 100-l reaction mixtures containing the cell extract (approximately 120 g of protein) in 0.25 M Tris-HCl buffer, pH 7.5, containing 3.7 mM chloramphenicol (25 nCi), and 5 g nbutyryl-CoA. The samples were extracted with 300 l of xylene and back-extracted with 100 l of Tris-HCl buffer. Aliquots of the xylene phases and the original reaction mixture were assayed for tritium content. ␤-Galactosidase activity was measured spectrophotometrically using chlorophenol red ␤-galactopyranoside as substrate by measuring the absorbance formed after 1 h at 595 nm on a Titretek Uniskan II plate reader (Flow Laboratories, McLean, VA). Protein concentrations were determined using the method of Smith et al. (20). All transient transfection experiments were performed three times in triplicate. Although a single experiment is shown in the data presented, the results seen in three separate transfection experiments were similar.
Statistical Analysis-All data were analyzed using a Student's t test.

Effects of Protein Kinase Inhibitors on the Induction of P4501A1 and ALDH-3 by PAH-Previous
studies have demonstrated that protein kinase C-mediated protein phosphorylation may be a determinant in the regulation of PAH-dependent gene expression of aldh3, cyp1A1, gst1, qor, and ugt1.6 (10, 11, 17). As we have previously shown (17), the PKC inhibitor H7 (50 M) potently inhibited BA induction of both cyp1A1 and aldh3 mRNA by over 80% in cultured rat hepatocytes (Fig. 1). In contrast, the PKA inhibitors H8 (50 M) and HA1004 (50 M) had little or no effect on PAH induction of mRNA levels for cyp1A1; in fact, they served as negative controls for inhibitor studies on PKC activity since they only inhibit this activity at very high concentrations (Ͻ200 M). Surprisingly, the PKA inhibitors H8 and HA1004 strikingly potentiated PAH induction of mRNA for aldh3 by over 2-fold (Fig. 1). These results were also seen in the levels of ALDH-3 and P4501A1 proteins (data not shown), demonstrating that the changes seen are caused by the regulation of pretranslational processes.
The time courses for induction of ALDH-3 and EROD (7ethoxyresorufin O-deethylase activity for P4501A1) activity in cultured rat hepatocytes in the presence of 50 M BA or BA plus 50 M H8 were compared (Fig. 2). Maximal induction of EROD activity by BA occurred at day 3, and H8 had no observable effects on either BA-inducible or basal EROD activity. The time courses of induction of ALDH-3 activity by BA or BA plus H8 were very similar. Maximal induction occurred at 4 days, and H8 potentiated PAH induction by 2-3-fold as early as 1 day and as late as 6 days in culture. Basal ALDH-3 activity (6.2 nmol/ min/mg protein) was observed on day 0 but declined to 0.5 nmol/min/mg protein by day 3 in culture. H8 treatment alone at any concentration did not appreciably affect the basal level of ALDH-3 enzyme activity. H8 potentiated PAH induction of ALDH-3 activity over 3-fold in a concentration-dependent manner ( Fig. 3) but had no effect on basal activity at any concentration tested.
PKA activity was measured using a specific peptide substrate labeled with a fluorescent dye, and an appropriate PKA standard with a known activity was used to confirm the specificity of PKA activity (29). Specific PKA enzyme activity in rat hepatocytes was inhibited (Ͼ80%) in a concordant but opposite manner than was observed with the H8 and HA1004 potentiation of PAH induction of ALDH-3 enzyme activity (Fig. 4). As a control, the same concentrations of a structurally similar PKC inhibitor, H7, were observed to have little or no effect on PKA activity. Since higher concentrations of H8 and HA1004 (Ͼ100 M) also inhibit PKC activity (17), we routinely used 50 M concentrations of PKA inhibitors at which hepatocyte PKC activity was unaffected.
Effect of Elevated Intracellular cAMP Levels on PAH Induction and H8 Potentiation of PAH-mediated Induction of aldh3 Expression-The level of intracellular cAMP, and therefore the activation of cAMP-dependent protein kinase, was modulated in primary rat hepatocytes by addition of the cAMP analog, dibutyryl cAMP, or the PKA activator, forskolin. After a period of 24 h in culture and immediately prior to initiating PAH induction, hepatocytes were exposed for 60 min to either dibutyryl cAMP or forskolin. PAH-mediated induction of aldh3 mRNA and H8 potentiation of PAH-mediated induction of aldh3 mRNA levels were inhibited more than 60 and 70%, respectively, by either dibutyryl cAMP or forskolin (Fig. 5). This treatment had no effect on the PAH-mediated induction of levels of cyp1A1 mRNA. Furthermore, the changes in levels of aldh3 mRNAs were concordant with the changes observed for ALDH-3 enzyme activity (Fig. 6) and protein measured by Western immunoblot analysis (data not shown). These results demonstrate that cAMP analogs and activators of adenylyl cyclase led to repression of both PAH induction and H8 potentiation of PAH induction of aldh3 gene expression. This effect is specific for the aldh3 gene, since these agents had little or no effect on expression of cyp1A1.
Effect of Protein Phosphatase Inhibitors on the PAH Induc- tion of ALDH-3 and on the H8 Potentiation of PAH Induction of ALDH-3-Okadaic acid is a potent and specific inhibitor of protein phosphatases 1 and 2A and appears to exert its effects by preventing dephosphorylation of specific transcription factor proteins in cells. Okadaic acid decreased both PAH induction and H8 potentiation of PAH induction of ALDH-3 activity by over 50 and 80%, respectively, in a concentration-dependent manner as expected (Fig. 7). Interestingly, H8 potentiation of PAH induction of ALDH-3 enzyme activity was more sensitive to okadaic acid than was PAH induction of ALDH-3 activity. For example, okadaic acid (15-20 nM) resulted in 50% inhibition of ALDH-3 activity induced by BA alone, whereas 5-10 nM okadaic acid resulted in 50% inhibition of ALDH-3 activity induced by treatment with BA plus H8 (Fig. 7).
To assess whether these effects of okadaic acid involve a pretranslational mechanism of action, we performed Northern analysis of mRNA isolated from cells treated with 50 M H8, 50 M BA, 20 nM okadaic acid, or combinations of these agents (Fig. 8). These results document that the decrease in aldh3 mRNA levels and enzyme activity by okadaic acid are apparently caused by changes in pretranslational processes specific for aldh3 expression. Addition of okadaic acid to cultured rat hepatocytes treated with BA or BA plus H8 had no effect on PAH induction of cytochrome P4501A1 and levels of its mRNA (data not shown).  (1,5,6,30) have developed a highly inducible 5Ј-flanking construct containing 3.5 kb of upstream sequence of the rat aldh3 gene in a chloramphenicol acetyltransferase (CAT) reporter gene. Basal level expression of this gene is cell type-specific, and liver displays very low basal expression (1, 2). Unlike the native gene in rat primary hepatocyte cultures, there is measurable basal expression of the pALDH3.5 plasmid construct during transient transfection into HepG2 cells. We observed a 2-4-fold increase in pALDH3.5CAT expression in the presence of the PKA inhibitor H8 in the absence (Fig. 9A) or presence of 1,2benzanthracene (Fig. 9B). In these experiments, we noted that H7 and H8 also stimulated the expression of the transfection control plasmid, pCMV-␤gal. Therefore, we have expressed the data relative to total cellular protein, since our routine transfection efficiency with pCMV-␤gal normally yields ␤-galactosidase activity which varied less than Ϯ20% within a single transfection experiment. The PKC inhibitor, H7, used as a control for the structurally related H8 compound, had no effect on basal expression at the concentrations used, suggesting that the mechanism of regulation by PKA activators and inhibitors isindependentoftheactionoftheAhR(ligand-orphosphorylationdependent) or PKC activation. Inhibition was seen at higher concentrations of H7 and H8 (50 M), but this concentration of either agent was toxic to HepG2 cells in our hands.

Effect of PKA Inhibitors on Basal and PAH-inducible Expression of pALDH3.5 CAT-Lindahl and co-workers
Subsequent transfection experiments demonstrated that forskolin, dibutyryl cAMP, and okadaic acid inhibited the transient expression of this reporter gene by 60% in the presence of BA and by 75% in the presence of BA plus H8, similar to the manner they affect expression of the native gene in cultured rat hepatocytes (Fig. 10). In addition, both basal and PAH-induced reporter gene activity was inhibited in a similar manner to the PKC inhibitor H7 (Fig. 9), suggesting that the regulation of the aldh3 gene by PKA activity functions independently of the AhR action.
Deletion Analysis to Define cis-Acting Elements Associated with Regulation by PKA-To map the region of the 5Ј-flanking region of the aldh3 gene responsible for a putative PKA-responsive element, we tested a series of 5Ј-flanking CAT constructs derived from the rat gene (5,6). Of these constructs (Fig. 11), only two CAT constructs were positively regulated by the PKA inhibitor H8, namely pALDH3.5 and pI-8. The shorter deletion CAT constructs pI-3, pI-4, pI-5, and pI-6 displayed higher basal level expression than that observed with pALDH3.5CAT and pI-8CAT, suggesting that a cis-acting negative regulatory element exists located between nucleotides Ϫ1125 and Ϫ991 of the 5Ј-flanking region of aldh3. Due to the effects of PKA activators and inhibitors seen in experiments utilizing pALDH3.5, this element appears to be regulated by a PKA-dependent phosphorylation event.
Since the expression of the pI-8 construct, but not pI-5, was affected by H8, we prepared an internal deletion in the 3.5-kb construct around the PstI site at Ϫ1057 bp. Specifically, the construct, p⌬(Ϫ1057/Ϫ930)ALDH, was prepared from pALDH3.5 which had an internal deletion between nucleotides Ϫ1057 and Ϫ930 of the flanking region (Fig. 11). This construct had higher levels of basal expression like the pI-5 or shorter constructs and was not regulated by the PKA inhibitors, demonstrating loss of the PKA-regulated element from the reporter gene. Given the results seen with the deletion analyses, we propose that the PKA-responsive elements associated with regulation by PKA lies within a 66-bp region from nucleotides Ϫ1057 and Ϫ991.

DISCUSSION
A number of extracellular signals regulate various intracellular processes via a complex cascade of receptors, transducers, and second messengers (31,32). These second messengers, cyclic nucleotides including cAMP and cGMP, are well characterized biochemical systems (33,34). The resulting activation of cyclic nucleotide-dependent protein kinases is the basis for many effects of cyclic nucleotides on cellular function. Recently, a number of in vivo and in vitro studies have been focused on the effects of protein kinase activation on the specific signal transduction pathways in the control of cyp1A1 gene expression. Notably, protein kinase C-mediated phosphorylation (15,16) appears to be a determinant in the regulation of PAHinducible cyp1A1 gene expression. Our past experiments suggested that PKC-dependent phosphorylation of AhR appears to be required for PAH induction of all of the xenobiotic metabolizing enzymes in the Ah gene battery, not just for cyp1A1 (17).
Other studies have reported that P450-mediated biotransformation activities associated with phenobarbital-inducible isozymes, including cyp2B1, cyp2B2, and cyp3A1, were depressed by agents that elevated cAMP levels in rat liver and in rat hepatocytes (35), indicating a negative regulatory role for cAMP-dependent signal transduction pathway on gene induction by this barbiturate. This is not the case for cyp1A1, the control gene used in the studies reported herein.
In the current study, we used rat hepatocytes and transient transfection of reporter genes in HepG2 cells to investigate the regulatory control exerted by intracellular cAMP on basal and PAH-induced aldh3 gene expression. H8 and HA1004 are effective inhibitors of cyclic nucleotide-dependent protein kinases (29) and are especially effective as PKA and PKG inhibitors.
Our results indicate that these inhibitors potentiate PAH induction of ALDH-3 activity in a concentration-dependent manner, but neither PKA inhibitor had any effect on PAH induction of P4501A1. Protein kinase A activity was shown to be effectively inhibited by H8 and HA1004 but not by H7 in cultured hepatocytes, demonstrating an inverse correlation between potentiation of PAH-dependent regulation of aldh3 gene expression and inhibition of specific PKA activity.
As H8 is also a potent protein kinase G inhibitor, we were concerned that the inhibition of this protein kinase (PKG) by H8 might be responsible for the H8 potentiation effect. However, elevation of intracellular cAMP levels, achieved through the use of a membrane-permeable cAMP analog, dibutyryl cAMP, also resulted in inhibition of PAH-mediated induction and H8 potentiation of PAH-mediated induction of aldh3 gene expression. Striking inhibition of BA induction and H8 potentiation of BA induction of aldh3 gene expression also was noted when intracellular cAMP levels were modulated via adenylate cyclase activation in hepatocytes with 25 M forskolin. In recent studies by Bjornsson et al. (36) and Sidhu and Omiecinski (35), treatment of primary rat hepatocytes with forskolin results in dramatic stimulation of intracellular cAMP levels and decrease in expression of cyp2B. Finally, okadaic acid is a potent, selective inhibitor of protein phosphatase type 1 and 2A and apparently exerts its effects by inhibiting protein dephosphorylation in cells. Okadaic acid decreased PAH induction and H8 potentiation of PAH induction of ALDH-3 activity and mRNA levels in a concentration-dependent manner as expected. Dibutyryl cAMP, forskolin, and okadaic acid had no FIG. 9. Effects of PKA and PKC inhibitors on PAH induction of pALDH3.5CAT in HepG2 cells. pALDH3.5CAT was transiently transfected into HepG2 cells, and either 50 M of H7 or H8 was added 1 h prior to treatment of the cells with BA. The cellular protein was harvested, and CAT activity and cell protein were measured as described under "Methods." The CAT activity was normalized relative to cell protein. *, statistically different (p Ͻ 0.01) from untreated cells (A) or cells treated with BA (B). effect on the PAH induction of cyp1A1 gene expression under any condition. Since PKC inhibitors had pronounced effect on PAH induction of P4501A1, but okadaic acid did not, PKC-dependent phosphorylation of the AhR must be near maximal in cultured primary rat hepatocytes (17).
Transient transfection experiments utilizing a CAT construct containing 3.5 kb of the 5Ј-flanking region of the aldh3 gene demonstrated that the effect of protein kinase A inhibitors, namely H8 and HA1004, is at the level of transcription. Furthermore, agents that increase cAMP levels or prevent dephosphorylation of some critical transcription factor act to inhibit the expression of this gene. The results strongly suggest a specific and negative regulatory role for the cAMP-dependent signal transduction pathway in basal and PAH-inducible aldh3 gene expression.
Using deletion analysis of the regulatory 5Ј-flanking region of this gene, we have identified a region that likely contains a negative regulatory element (Ϫ1057 and Ϫ991 bp; CTGCAGT-GGC TGCTATGGCG CAAAGCCACC AGACAAGAGA AATTT CCAACTCCAT TGTAATC TTAT) that apparently regulates expression of this gene by a PKA-dependent process. The mechanism for this decreased expression has not been fully established. In Saccharomyces cerevisiae, the aldehyde dehydrogenase 2 (ALD2) gene (37) has been shown to be negatively regulated by protein kinase A, and its 5Ј-flanking region contains two classical cAMP-responsive elements. The 66-bp region (Ϫ1057 to Ϫ991) of the rat aldh3 gene (Fig. 11) and that of its immediate flanking regions contain no unique sequences FIG. 11. Deletion analysis of the 5-flanking region of the aldh3 gene. A series of CAT constructs were prepared with deletions in the 3.5-kb 5Ј-flanking regulatory region of the aldh3 gene and were tested by transient transfection into HepG2 cells. The cells were treated with either dimethyl sulfoxide or H8 (20 M) in dimethyl sulfoxide 24 h after transfection, and the cellular protein was harvested at 48 h after transfection. CAT activity and cell protein were measured as described under "Methods." The CAT activity was normalized to unity relative to cell protein, and the standard deviation (n ϭ 3) was less than 10% of the average normalized CAT activity. *, statistically different from basal activity of the given construct (p Ͻ 0.005); **, statistically different from basal activity of pALDH3.5 or pI-8 constructs (p Ͻ 0.005).
with high consensus to other canonical cAMP-responsive element binding protein binding elements (classical or nonclassical, Ref. 38) or other elements suggesting that the rat gene may be uniquely regulated by a novel PKA-regulated transcription factor.
Our studies demonstrate that there are apparently two protein phosphorylation mechanisms involved in the regulation of aldh3 gene expression in hepatocytes (Fig. 12). One involves the AhR, which requires PKC phosphorylation to positively regulate aldh3 gene in the presence of PAH ligands. The second appears to involve a cAMP-responsive process, which may require PKA phosphorylation to negatively regulate aldh3 gene expression (Fig. 12). Future studies will characterize the cisand trans-acting elements involved in the negative regulation of the class 3 aldehyde dehydrogenase gene of rat.