CCAAT/Enhancer-binding Proteins

The phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA) is a potent inducer of keratinocyte differentiation and of involucrin gene expression. In the present study we show that a CCAAT/enhancer-binding protein (C/EBP) site in the proximal regulatory region is required for the phorbol ester response. Mutation of the C/EBP site results in the loss of basal and TPA-responsive activity. Gel mobility supershift analysis shows that C/EBPα binding to this site is increased by TPA treatment. Moreover, cotransfection of the human involucrin reporter plasmid with C/EBPα increases promoter activity to an extent comparable with TPA treatment. Mutation of the C/EBP-binding site eliminates these responses. Transfection experiments using GADD153 to create C/EBP-null conditions confirm that C/EBP factors are absolutely required for promoter activity and TPA responsiveness. C/EBPβ and C/EBPδ inhibit both TPA- and C/EBPα-dependent promoter activation, indicating functional differences among C/EBP family members. These results suggest that C/EBP transcription factor activity is necessary for basal promoter activity and TPA response of the involucrin gene.

The cornified envelope is a covalently cross-linked layer of protein that is formed by epidermal keratinocytes during the final stages in differentiation (1,2). Involucrin (hINV) 1 is a precursor of the keratinocyte cornified envelope that functions as a glutamyl donor and amine acceptor in the transglutaminasedependent cross-linking reaction (1,(3)(4)(5)(6)(7). For proper envelope formation the transglutaminase enzyme and its substrates (the envelope precursors) must be expressed at the appropriate time and level during the differentiation process. Abnormal expression or lack of expression can result in disease (8 -10).
Involucrin is exclusively expressed in the suprabasal epidermal layers (1,7,(11)(12)(13). The mechanisms that regulate hINV expression during keratinocyte differentiation are an area of active investigation (14 -22). The proximal regulatory region (PRR) of the hINV promoter is located between positions Ϫ241 and Ϫ7 relative to the start of transcription (20,21). The PRR drives nearly one-half of the activity of the promoter. Site specific mutation experiments indicate that an activator protein 1 (AP1)-binding site, AP1-1, located within the PRR, is absolutely required for promoter activity (20). This site interacts with JunB, JunD, and Fra-1 (20). In addition to the AP1-1 site, this region contains a C/EBP site that is necessary for promoter activity (14).
C/EBP factors comprise a family of related bZIP (basic region leucine zipper) DNA-binding proteins that regulate transcription. This family includes C/EBP␣, C/EBP␤, C/EBP␦, GADD153, CHOP, and LAP (23)(24)(25)(26)(27)(28). C/EBP factors have been shown to differentially modulate transcription and differentiation in adipocytes, myelomonocytic cells, and ovarian follicles (24,26,29,30). Based on previous studies showing that the C/EBP-binding site of the hINV promoter is important for activity (14), we hypothesized that C/EBP factors may have a role in regulating hINV expression during keratinocyte differentiation. In the present study we show that each C/EBP protein differentially regulates hINV promoter activity via the hINV promoter C/EBP site and that C/EBP factor binding to this site is dramatically increased following treatment of keratinocytes with 12-O-tetradecanoylphorbol-13-acetate (TPA). We show that C/EBP␣ is a component of this complex. We also use an inhibitory member of the C/EBP family, GADD153, to show that C/EBP factor activity is required for the TPA-dependent increase in activity.

MATERIALS AND METHODS
Reagents-[␥-32 P]ATP (3000 Ci/mmol) was purchased from NEN Life Science Products. Keratinocyte serum-free medium (KSFM), trypsin, Hanks' balanced salt solution, gentamicin, and Lipofectin were obtained from Life Technologies, Inc. The pGL2 plasmid and the chemiluminescent luciferase assay systems were obtained from Promega. Phorbol ester (TPA) and dimethyl sulfoxide (Me 2 SO) were obtained from Sigma. Chemiluminescence was measured using a Berthold luminometer, and synthetic oligonucleotides were synthesized using an Applied Biosystems DNA synthesizer.
Tissue Culture-Human foreskin keratinocytes were isolated and cultured as described (20,21). The cells were passaged at a split ratio of 1:5 when 70% confluent and used for transfection at the third passage.
Plasmid Construction-The structure of the hINV promoter reporter plasmid, pINV-241, has been described (20). To create the C/EBP site mutant, the fragment containing the wild type C/EBP site (5Ј-GCT-GCTTAAG-3Ј) was released as part of a larger fragment by digestion of pINV-241 with ApaI/PstI and replaced with the identical segment containing a mutated C/EBP site (5Ј-GCTGAGATCT-3Ј). The modified nucleotides are underlined. The structure of the mutated AP1-1 site in pINV-241(AP1-1m) has been previously described (20). The junction between the hINV gene sequences and the luciferase reporter gene sequence is identical in all constructs.
Cell Transfection and Luciferase Assay-Keratinocytes (60% confluent) were transfected in 60-mm-diameter dishes. Lipofectin reagent (16 g) and 4.0 g of test plasmid were added to cells in 3 ml of KSFM and incubated for 5 h at 37°C. After 5 h, additional KSFM (3 ml) was added, and the incubation was continued for another 19 h. After 24 h in fresh * This work was supported by Grants AR41456 and GM43751 from the National Institutes of Health (to R. L. E.) and utilized the facilities of the Skin Diseases Research Center of Northeast Ohio (National Institutes of Health Grant AR39750). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ ‡ To whom correspondence should be addressed: Dept. of Physiology and Biophysics, Rm. 532, Case Western Reserve University School of Medicine, 2109 Adelbert Rd, Cleveland, OH 44106-4970. Tel.: 216-368-5530; Fax: 216-368-5586; E-mail: rle2@po.cwru.edu. 1 The abbreviations used are: hINV, human involucrin; AP1, activator protein 1; TPA, 12-O-tetradecanoylphorbol-13-acetate; C/EBP, KSFM, cells were washed and treated for 24 h with KSFM or KSFM containing 50 ng/ml TPA (delivered from a 5 mg/ml stock in Me 2 SO) (20,21). Control groups received the Me 2 SO vehicle. For cotransfection experiments, involucrin reporter plasmid (2.5 g) was transfected with various concentrations of transcription factor expression plasmid. The final expression vector concentration was maintained constant by addition of an empty expression vector. The C/EBP factors C/EBP␤ (CRP2, rat) and C/EBP␦ (CRP3, mouse) were expressed using pMEX (Dr. Peter Johnson of the Frederick Cancer Research Center) (25). C/EBP␣ was obtained from Dr. David Samols (Case Western Reserve University) (31). GADD153 and pCMV-neo were obtained from by Dr. Nikki Holbrook (27,28). The cells were harvested and assayed for luciferase activity as outlined above. For luciferase assay, cells were washed twice with phosphate-buffered saline, dissolved in 250 l of cell culture lysis reagent (Promega), and harvested by scraping. Luciferase assays were performed immediately using a Promega luciferase assay kit (20). The results are expressed as luciferase activity per g of protein. All assays were performed in triplicate, and each experiment was repeated a minimum of three times. The triplicates routinely varied by less than 15%. As a control to assure comparable transfection efficiency, we utilized a green fluorescent protein plasmid and determined the percent of cells transfected by visual inspection (21,32).
Gel Mobility Shift Assay-For mobility shift assays, the reaction (20 l) contained 15% glycerol, 75 mM KCl, 0.375 mM dithiothreitol, 0.375 mM phenylmethylsulfonyl fluoride, 12.5 mM NaCl, 0.1 mg/ml poly(dI-dC), 2.5 g of nuclear extract, and 0.3 ng of radiolabeled DNA. The mixture was incubated for 5 min at room temperature, and samples were immediately electrophoresed at 250 V for 1.5 h on a 5% nondenaturing acrylamide gel using 0.25ϫ TBE running buffer. The gels were then dried for autoradiography. For competition studies, radioinert DNA competitor was added as a 20-or 200-fold molar excess. For gel supershift assays, the complete gel mobility shift assay mixture, without the 32 P-labeled oligonucleotide, was incubated at 4°C for 2 h in the presence of an antibody specific for the C/EBP isoform using 1 g of rabbit IgG per reaction. Dr. Steven McKnight kindly provided the C/EBP␣-specific rabbit polyclonal antibodies generated against amino acids 247-358 of rat C/EBP␣ (C/EBP␣-(247-358)). An additional C/EBP␣-specific antibody was obtained from Santa Cruz Biologicals (catalog no. sc-61X). The 32 P-labeled DNA was then added to the mixture and incubated at room temperature for 5 min. The resulting complexes were electrophoresed on a nondenaturing gel for characterization. C/EBP␤ (sc-150)-and C/EBP␦ (sc-636)-specific antibodies were obtained from Santa Cruz Biologicals.
Immunological Detection of C/EBP␣-Cultured keratinocytes, grown in KSFM, were treated for 24 h with KSFM or with KSFM containing 50 ng/ml TPA before preparation of nuclear extracts as described previously (20,33). Equal quantities of nuclear protein were electrophoresed on a 10% denaturing polyacrylamide gel and transferred to nitrocellulose. C/EBP␣ was detected using rabbit anti-human C/EBP␣ at a dilution of 1:500 (catalog no. sc-61X, Santa Cruz Biologicals) followed by a goat anti-rabbit IgG secondary antibody used at a 1:10,000 dilution. Secondary antibody binding was visualized using chemiluminescent detection reagents.

RESULTS
hINV Promoter C/EBP Site Is Necessary for TPA-dependent Regulation-A C/EBP transcription factor binding site located in the hINV promoter proximal regulatory region is required for hINV promoter activity (14). Fig. 1 shows that basal promoter activity (open bars) is reduced to 5-10% of control in the absence of a functional C/EBP site. In addition, TPA treatment increases promoter activity 8-fold; this TPA-dependent activation is absent when the C/EBP site is mutated (solid bars). These results demonstrate a requirement for the C/EBP site for both basal and TPA-activated promoter activity.
C/EBP Proteins Regulate Promoter Activity-We next tested the effects of C/EBP transcription factors on promoter activity. Fig. 2A shows that transfection of increasing concentrations of C/EBP␣ expression plasmid with a constant amount of hINV reporter plasmid increases hINV promoter activity in a concentration-dependent manner in the absence of TPA treatment. Expression is maximally increased by 0.4 g of C/EBP␣ plasmid, and the level is not significantly increased at higher concentrations. As different C/EBP heterodimers are known to differentially regulate gene expression, we determined whether C/EBP␤ and C/EBP␦ can regulate the C/EBP␣-dependent activation. As shown in Fig. 2B, C/EBP␤ and -␦ are equally efficient inhibitors of the C/EBP␣-dependent activity. In addition, as shown in Fig. 3, the C/EBP proteins do not regulate promoter activity when the C/EBP-binding site is mutated. These results (i) show that C/EBP proteins can regulate hINV promoter activity in the absence of TPA treatment, (ii) provide evidence for a dynamic regulatory interaction among C/EBP proteins, and (iii) demonstrate that the proteins act through the C/EBP-binding site.
GADD153 Inhibits C/EBP␣ and TPA-dependent Promoter Activity-To obtain additional evidence of a role for C/EBP factors in TPA-dependent activation, we treated keratinocytes with TPA in the presence of GADD153, a C/EBP family member that inhibits the activity of other C/EBP proteins by inhibiting interaction of the C/EBP-GADD153 complex with DNA (23,27). This treatment creates an environment in which C/EBP activity is selectively eliminated. As shown in Fig. 4A, transfection of TPA-treated cells with increasing concentrations of GADD153 produces a concentration-dependent inhibition of TPA-dependent promoter activity. As shown in Fig. 2, C/EBP␣ activates the promoter in the absence of TPA treatment. We would predict that GADD153 would also inhibit the C/EBP␣-dependent activation. Fig. 4B shows that GADD153 produces a concentration-dependent inhibition of C/EBP␣-dependent activation of promoter activity. In addition, GADD153 produced a concentration-dependent inhibition of basal promoter activity; 0.4 g of GADD153 expression plasmid/transfection reduced basal activity to 5% of normal (not shown).
TPA Treatment Increases Binding to the C/EBP Site-The above described results provide functional evidence suggesting a role for C/EBP proteins as regulators of hINVgene expression. To determine whether TPA treatment alters C/EBP binding, we treated cultures with or without TPA and assayed binding by gel mobility shift assay using the double-stranded hINV C/EBP site oligonucleotide, 5Ј-GGTTTGCTGCTTAA-GATGCCTG-3Ј (C/EBP-binding site in bold). As shown in Fig.  5, gel mobility shift assay indicates that binding to 32 P-labeled C/EBP-binding site is increased by Ͼ10-fold following treatment with TPA (lanes 1 and 2). The binding is specific, because addition of a 20-or 200-fold molar excess of radioinert C/EBP oligonucleotide inhibits the binding (lanes 3 and 4). No inhibition of binding to 32 P-labeled C/EBP oligonucleotide is observed with oligonucleotides encoding consensus AP2 or ets binding sites or a mutant C/EBP-binding site (not shown). In addition, two different C/EBP␣-specific antibodies identify the complex prepared from TPA-treated cells as containing C/EBP␣ (lanes 5 and 6). C/EBP␤-or C/EBP␦-specific antibodies do not detect protein binding (not shown).
TPA Treatment Does Not Increase C/EBP␣ Level-The gel mobility shift experiment indicate that C/EBP␣ binding to the hINV C/EBP site is increased by TPA treatment. To determine whether this is caused by a change in C/EBP␣ level, we measured C/EBP␣ levels in nuclear extracts prepared from untreated and TPA-treated keratinocytes. As shown in Fig. 6, TPA treatment did not increase C/EBP␣ protein levels.
The Role of AP1 and C/EBP Sites-Previous studies indicate that AP1 is an important mediator of TPA-dependent regulation (20 -22). Because the AP1-binding site is located adjacent to the C/EBP-binding site (34), we wanted to assess the role of C/EBP relative to AP1. We therefore evaluated whether both sites are required for TPA-and C/EBP␣-dependent activation of hINV promoter activity. As shown in Fig. 7, mutation of either site results in a loss of TPA-and C/EBP␣-dependent activation.

DISCUSSION
The hINV Promoter Proximal Regulatory Region-The PRR of the hINVpromoter is the region immediately upstream of the TATA box. This region includes nucleotides Ϫ241 to Ϫ7, drives 50% of the activity of the promoter, and contains the consensus binding site for CCAAT/enhancer-binding protein AP1, the ets factors, and activator protein-2 (14). Mutation of the C/EBP site results in the loss of promoter activity. In contrast, mutation of the ets factor binding sites (EBS-1, EBS-2) or the AP2binding site changed basal promoter activity but did not effect TPA regulation. 2 Our present results suggest that the C/EBP site is essential for TPA-dependent promoter activity. These results suggest a dual role for C/EBP in maintaining basal promoter activity and in mediating activation in response to TPA.
C/EBP Proteins Differentially Regulate hINV Promoter Activity-C/EBP proteins comprise a family of bZIP domain proteins that readily form homo-and heterodimers (24,25). These proteins are important regulators of cell differentiation. For example, C/EBP␣ is a transcriptional activator in adipocytes (35). High levels of C/EBP␤ and C/EBP␦ accumulate early in adipocyte differentiation but are replaced by C/EBP␣ during late differentiation. C/EBP␤ and -␦ appear to play an important role in accelerating differentiation (24), and ectopic expression of C/EBP␤ in multipotent NIH-3T3 cells causes them to convert to committed adipoblasts. The accumulation of C/EBP␣ late in differentiation is correlated with expression of markers of adipose differentiation (24,36). C/EBP proteins also differentially regulate gene expression in the liver (37,38).
C/EBP proteins are also expressed in keratinocytes (39,40). C/EBP␤ is a negative regulator of human papillomavirus transcription in keratinocytes (39), and human C/EBP␣ is highly expressed in the epidermis (40). Little is known regarding their role as regulators of expression of keratinocyte genes; however, C/EBP␣ is expressed in suprabasal epidermis and in differentiated keratinocytes (40). A recent report indicates that C/EBP␣, C/EBP␤, and GADD153 are differentially expressed during keratinocyte differentiation (41). C/EBP␦ expression in the epidermis has not been studied. A recent abstract suggests that C/EBP␤ directs expression of cytokeratin 10, a keratin that is expressed in the suprabasal epidermal layers (42). Our present experiments show that C/EBP proteins alter pINV-241 promoter activity in a C/EBP site-dependent manner. C/EBP␣ increases basal transcription as efficiently as TPA treatment.
In contrast, C/EBP␤ and -␦ can suppress C/EBP␣-dependent activation. This is consistent with the ability of these proteins to form zippered heterodimers and with the fact that different 2 R. L. Eckert, unpublished observation. C/EBP complexes differentially regulate gene expression (23)(24)(25)(26)(27)(28). This result shows that differential expression of C/EBP factors during keratinocyte differentiation could, in principle, differentially regulate involucrin gene expression.
TPA-dependent Increase in C/EBP␣ DNA Binding Is Not Associated with an Increase in C/EBP␣ Concentration-Our studies indicate that the TPA-dependent activation of hINV promoter activity is correlated with increased binding of C/EBP␣ to the hINV promoter C/EBP site but that concentration of C/EBP␣ protein is not altered by TPA treatment. Recent in vitro studies show that C/EBP is a good substrate for protein kinase C but indicate that protein kinase C-dependent phosphorylation reduces C/EBP␣ binding to DNA (43). This reduction in DNA binding is thought to result from phosphorylation of serine residues within the DNA-binding region. However, as discussed by the authors (43), this scenario has not been confirmed in cells. It is possible that additional factors are phosphorylated by protein kinase C that assist C/EBP␣ to bind DNA or that, in vivo, serine phosphorylation actually enhances DNA binding. Moreover, there may be cell type-specific effects. Additional detailed studies will be necessary to determine how protein kinase C alters C/EBP␣ phosphorylation in keratinocytes. However, it is clear that the increased C/EBP␣ DNA binding is not the result of changes in C/EBP␣ concentration.
GADD153 Suppresses C/EBP␣-and TPA-dependent Transcription-GADD153 is a unique member of the C/EBP transcription factor family that forms transcriptionally inactive complexes with other C/EBP proteins. GADD153 forms heterodimeric complexes with other C/EBP proteins that are unable to bind to DNA and thus do not regulate transcription (27,28). Thus, GADD153 can be used selectively to create C/EBPnull conditions (i.e. to selectively knock out C/EBP function). We show that GADD153 inhibits basal transcription. The fact that GADD153 suppresses basal promoter activity suggests that endogenous C/EBP proteins are involved in maintaining basal promoter function. When C/EBP levels are artificially elevated by transfection of C/EBP␣, GADD153 also inhibits this response, verifying the C/EBP protein requirement for promoter activity. We also show that the TPA-dependent activation of hINV promoter activity is suppressed by GADD153, suggesting that TPA-dependent activation involves signal transduction events that include C/EBP factors. This suggestion is supported by the finding that TPA treatment increases complex formation at the hINV C/EBP-binding site and that C/EBP␣ is a part of this complex. C/EBP␣ is the primary C/EBP factor that interacts with the hINV C/EBP site in cells maintained under basal conditions; no binding of C/EBP␤ or -␦ was detected. A possible mechanism to explain the increased promoter activity in response to TPA is that TPA increases C/EBP␣ binding to the promoter. Consistent with this possibility, gel mobility supershift experiments suggest that C/EBP␣ is an abundant component of the C/EBP site binding activity present in TPA-treated cells.
Regulation of hINV Promoter Activity-Involucrin gene expression and promoter activity are regulated by a variety of transcription factors. TPA is a potent inducer of keratinocyte differentiation and an efficient activator of hINVgene expression and promoter activity (20). TPA-dependent activation of promoter activity has been shown to require the presence of functional AP1 sites (20 -22). The present study demonstrates the role of C/EBP factors. These results are interesting for several reasons. First, the C/EBP and AP1 sites in the hINV promoter PRR are separated by only 10 nucleotides (34). This close proximity should permit AP1-C/EBP factor interaction. Second, both the AP1 and C/EBP-binding sites must be intact for either TPA-or C/EBP␣-dependent promoter activation. This requirement suggests that C/EBP and AP1 factors interact to produce the response and that members of both families are required for activation. Understanding the mechanism of this regulation will require further study. Third, promoter ac-tivity requires the presence of both AP1 and C/EBP proteins (i.e. dominant inhibitory proteins of either class inactivate the promoter) (44). Fourth, both C/EBP and AP1 factor binding activity is increased by TPA treatment (20). Taken together these features suggest that a multiprotein complex, including C/EBP and AP1, may be required for hINV gene expression. Our results suggest that this complex includes C/EBP␣.