Calcium-dependent Involucrin Expression Is Inversely Regulated by Protein Kinase C (PKC)α and PKCδ*

Calcium is an important physiologic regulator of keratinocyte function that may regulate keratinocyte differentiation via modulation of protein kinase C (PKC) activity. PKCα and PKCδ are two PKC isoforms that are expressed at high levels in keratinocytes. In the present study, we examine the effect of PKCδ and PKCα on calcium-dependent keratinocyte differentiation as measured by effects on involucrin (hINV) gene expression. Our studies indicate that calcium increaseshINV promoter activity and endogenous hINV gene expression. This response requires PKCδ, as evidenced by the observation that treatment with dominant-negative PKCδ inhibits calcium-dependent hINV promoter activity, whereas wild type PKCδ increases activity. PKCα, in contrast, inhibits calcium-dependent hINV promoter activation, a finding that is consistent with the ability of dominant-negative PKCα and the PKCα inhibitor, Go6976, to increasehINV gene expression. The calcium-dependent regulatory response is mediated by an AP1 transcription factor-binding site located within the hINV promoter distal regulatory region that is also required for PKCδ-dependent regulation; moreover, both calcium and PKCδ produce similar, but not identical, changes in AP1 factor expression. A key question is whether calcium directly influences PKC isoform function. Our studies show that calcium does not regulate PKCα or δ levels or cause a marked redistribution to membranes. However, tyrosine phosphorylation of PKCδ is markedly increased following calcium treatment. These findings suggest that PKCα and PKCδ are required for, and modulate, calcium-dependent keratinocyte differentiation in opposing directions.

Calcium is an important regulator of keratinocyte differentiation. Incubation of cultured keratinocytes with calcium increases differentiation and expression of differentiation-associated genes (1)(2)(3). Moreover, the presence in vivo of an epidermal calcium gradient, with increasing calcium levels in the more differentiated layers, suggests a role for calcium in regulating epidermal differentiation (2, 4 -6). However, the mechanism whereby the increase in extracellular free calcium triggers differentiation is not well understood. One possible mechanism involves the calcium-dependent activation of protein kinase C (PKC) 1 isoforms (7)(8)(9). Keratinocytes express the PKC␣, -␦, -⑀, -, and -isoforms (10). These enzymes control a variety of signaling cascades and transcription factors and function as regulators of keratinocyte differentiation-dependent gene expression (11)(12)(13)(14)(15). In keratinocytes, PKC␣ and PKC␦ are abundant PKC isoforms that have been implicated as regulators of differentiation (16 -19). In the present study, we focus on the role of these isozymes and their effects on calciumdependent regulation of differentiation.
Involucrin, a keratinocyte structural protein that functions as a precursor of the cornified envelope (20 -22), is expressed in a tissue-specific and differentiation-appropriate manner in vivo (23). Moreover, agents that promote keratinocyte differentiation, including calcium, increase hINV levels and hINV promoter activity in cultured keratinocytes (24 -27). A novel PKC, Ras, MEKK1, MEK3/MEK6, p38 pathway has been shown to mediate phorbol ester-dependent activation of hINV gene expression (28 -30). This pathway targets AP1 transcription factors that, in turn, bind to sites within the hINV promoter to activate transcription (31)(32)(33). However, the events leading to calcium-dependent induction of hINV gene expression in normal keratinocytes are not well understood. The goal of the present study is to evaluate the role of PKC in mediating the calcium-dependent increase in hINV gene expression. Our findings suggest that PKC␣ inhibits and PKC␦ enhances the calciumdependent activation of hINV promoter activity and endogenous gene expression.

MATERIALS AND METHODS
Chemicals and Reagents-Keratinocyte serum-free medium (KSFM) was obtained from Invitrogen. Go6976, an inhibitor of classical PKC isoforms, was obtained from Calbiochem. The pGL2-basic plasmid and the chemiluminescent luciferase assay system were purchased from Promega. Isoform-selective rabbit polyclonal antibodies for PKC␣ (sc-208) and PKC␦ (sc-937) were obtained from Santa Cruz Biotechnology and diluted 1:500 for immunoblot. Normal mouse IgG (sc-2025) and horseradish peroxidase-conjugated goat anti-mouse IgG (sc-2005) were from Santa Cruz Biotechnology and used diluted 1:7500. Goat polyclonal Sp1-specific antibody (sc-59), obtained from Santa Cruz Biotechnology, was used for immunoblot at a dilution of 1:500. Rabbit antihuman involucrin polyclonal antibody, used for immunoblot at a dilution of 1:8000, has been described (34). The mouse monoclonal anti-phosphotyrosine (clone 4G10) was obtained from Upstate Biotechnology, Inc., and diluted 1:500 for immunoblot. Mouse monoclonal antihuman ␤-actin (Sigma, clone AC-15) was diluted 1:10,000 for immunoblot. Horseradish peroxidase-conjugated donkey anti-rabbit IgG (NA934) was from Amersham Biosciences and used for immunoblot at a dilution of 1:7500.
Adenoviruses and Plasmids-The hINV promoter constructs used in this study have been described previously (31,32). All nucleotide posi-* This work was supported by a grant from the National Institutes of Health (to R. L. E.). 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.
§ tions are defined relative to the hINV gene transcription start site (32). Expression vectors encoding wild type PKC isoforms, cloned into pcDNA3, were a generous gift of Dr. S. Ohno (35)(36)(37). Dominant-negative PKC␣, cloned in pcDNA3 (K368R mutation in the ATP-binding site), was a gift from Dr. B. Weinstein (38). Adenoviruses encoding wild type and dominant-negative (dn) kinases were kindly provided by Dr. Kuroki (39). Wild type PKC␦ and PKC␣ and dnPKC␦, in which a Lys to Arg mutation was introduced in the ATP-binding site (39), are transcribed, respectively, from the cytomegalovirus and chicken ␤-actin promoter.
Keratinocyte Transfection and Infection-Normal human foreskin keratinocytes were cultured as described previously (32). Third passage keratinocytes, in 9.5-cm 2 dishes, were transfected when ϳ25% confluent. FuGENE 6 transfection reagent was mixed with KSFM at a final concentration of 3% for 5 min at 25°C. This mixture (100 l) was then added to 1 g of plasmid DNA, incubated for an additional 15 min, and then added dropwise to the cells in dishes containing 2 ml of KSFM. After 24 h, the medium was changed to KSFM containing 0.09 or 0.3 mM calcium chloride. After 48 h, the cells were harvested and assayed for luciferase activity. All assays were performed in triplicate, and each experiment was repeated a minimum of three times. Luciferase activity is normalized per g of protein (28). As required, transfection efficiency was determined using a green fluorescent protein-expressing plasmid (29).
For adenovirus infection, keratinocyte cultures in 9.5-cm 2 dishes were transfected with 1 g of pINV-2473 when 30% confluent and incubated for 24 h. The media were then removed, and the cells were incubated with the appropriate adenovirus for 24 h in 1 ml of KSFM containing 2.5 g/ml Polybrene. The cells were then transferred to fresh medium containing 0.09 or 0.3 mM calcium chloride and incubated for 48 h prior to harvest and measurement of luciferase activity (30).
Cell Fractionation-Cells were washed in cold phosphate-buffered saline and scraped into a minimal volume of extraction buffer (20 mM Tris-HCl, pH 7.5, containing 5 mM EDTA, 10 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 5 g/ml leupeptin, 5 g/ml aprotinin, and 1 mM sodium orthovanadate) (41). The suspension was sonicated, and centrifuged at 100,000 ϫ g for 1 h. The supernatant (cytosol) was removed, and the pellet was resuspended in extraction buffer containing 1% Triton X-100, sonicated, incubated on ice for 1 h, and centrifuged at 100,000 ϫ g for 1 h to yield the Triton-soluble fraction. The high speed pellet was resuspended in sample buffer to yield the particulate fraction (41).
Immunofluorescence Microscopy-Keratinocytes were plated on glass coverslips and grown in KSFM containing 0.09 mM calcium. Cells were then incubated for various times in KSFM containing 0.3 mM calcium or 500 nM 12-O-tetradecanoylphorbol-13-acetate (TPA). The cells were then fixed at 4°C for 12 h in 2% paraformaldehyde, permeabilized with 100% methanol for 30 min, blocked in 10% goat serum for 30 min, and incubated for 30 min in primary PKC antibody at a 1:500 dilution in the presence or absence of isoform-specific blocking peptide (PKC␦ peptide, Santa Cruz Biotechnology, sc-937P). The sections were then incubated for 30 min with Oregon Green 514-linked goat antirabbit IgG (Molecular Probes) at a dilution of 1:500. The coverslips were mounted using Gel Mount Media (Biomedia), and fluorescent images were obtained at 100ϫ using a digital Nikon Optiphot microscope.

Calcium Regulates Keratinocyte Differentiation-Involucrin
is a well characterized marker of keratinocyte differentiation (27,43) that has been extensively used as a model to identify mechanisms that regulate differentiation (23,28,29,31,44). We began our studies by confirming that the hINV gene expression is regulated by calcium in our culture system. Keratinocytes were cultured in medium containing 0.09 or 0.3 mM calcium for 48 h, and hINV levels were then monitored by immunoblot. The inset in Fig. 1A shows that calcium treatment causes a 5-fold increase in endogenous hINV expression. We confirmed this response by examining the effects of calcium on hINV promoter activity. Keratinocytes were transfected with the hINV promoter reporter plasmids, pINV-41 or pINV-2473 (32), and then grown for 48 h in 0.09 or 0.3 mM calciumcontaining medium. The activity of the full-length hINV promoter construct, pINV-2473, is increased 5-fold by calcium treatment. In contrast, activity of the minimal promoter construct, pINV-41, which encodes only the hINV gene TATA box (32), is not regulated. We also confirmed that the appropriate PKC isoforms are expressed in the cultured human keratinocytes. Cell extracts were prepared from keratinocytes growing in medium containing 0.09 mM calcium, and samples were electrophoresed for immunodetection using PKC-specific antibodies. Fig. 1B confirms that PKC␣, -␦, -⑀, -, and -are expressed in our model system, as has been reported elsewhere (10,14,29,(45)(46)(47)(48). Moreover, although the results cannot be regarded as quantitative, the film exposure times and protein loading densities required for visualization suggest that PKC␣ and -␦ are the most abundant isoforms.
PKC␦ Activity Is Required for Calcium-dependent Regulation of hINV Gene Expression-Because of their relative abundance, and the fact that they have been implicated as mediating differentiation-dependent regulation in keratinocytes (14,18,28,29,39,40,49), we focused on the PKC␦ and PKC␣ isoforms. We began by studying the role of PKC␦. Keratinocytes were co-transfected with pINV-2473 and PKC␦-encoding vector and then treated with 0.09 or 0.3 mM calcium for 48 h. Cell extracts were then prepared and assayed for hINV promoter activity. As shown in Fig. 2A, both basal and calcium-stimulated hINV promoter activity is increased by PKC␦. This suggests that cotreatment with calcium and PKC␦ can enhance promoter activity but does not indicate whether PKC␦ activity is required for the calcium response. To determine whether PKC␦ activity is required for calcium regulation, we used a dominantnegative form of PKC␦. In this experiment cells were treated with pINV-2473 and 24 h later with dnPKC␦-encoding virus and then incubated with 0.09 or 0.3 mM calcium for 48 h. Fig.  2B shows that dnPKC␦ nearly completely inhibits the calciumdependent increase in hINV promoter activity. In contrast, dnPKC␦ expression does not alter base-line promoter activity. To confirm that PKC␦ and dnPKC␦ isoforms are expressed, we treated cells with empty vector (EV) or expression vectors encoding PKC␦ or dnPKC␦. Extracts were then prepared for immunoblot. As shown in Fig. 2C, this analysis confirms that the PKC␦ and dnPKC␦ expression vectors produce each respective protein in keratinocytes and that these products co-migrate with the endogenous PKC␦. These results suggest that calcium-associated regulation of hINV promoter activity requires PKC␦ activity. We next determined whether the endogenous gene displays a similar sensitivity. In Fig. 2D, cells were incubated with 0.09 (Ϫ) or 0.3 mM (ϩ) calcium in the presence of empty vector (EV) or PKC␦-encoding adenovirus. After 48 h, hINV protein levels were measured by immunoblot. Treatment with 0.3 mM calcium or PKC␦ causes a 2.5-fold increase in hINV protein level. Stimulation with both 0.3 mM calcium and PKC␦ results in a 5.5-fold increase. To determine whether PKC␦ activity is required for calcium regulation of endogenous hINV level, we infected keratinocytes with empty vector (EV) or vector encoding dominant-negative PKC␦ (dnPKC␦), and then treated with 0.3 mM calcium for 48 h. Fig. 2E shows that dnPKC␦ efficiently inhibited the calcium-dependent increase in endogenous hINV levels.
PKC␣ Suppresses Basal and Calcium-dependent Promoter Activity-PKC␣ is a classical PKC isoform that has an important regulatory role in mouse keratinocytes (18,19). To determine whether PKC␣ influences calcium-dependent regulation of differentiation, we transfected normal keratinocytes with pINV-2473 and increasing concentrations of PKC␣ expression plasmid and incubated for 48 h in the presence of 0.09 or 0.3 mM calcium. As shown in Fig. 3A, PKC␣ causes a concentrationdependent reduction in calcium-dependent promoter activity.
To provide additional evidence for this regulation, we used a plasmid encoding the dominant-negative form of PKC␣ (dnPKC␣). Cells were transfected with pINV-2473 in the presence or absence of dnPKC␣ and then treated for 48 h with high or low calcium. Expression of dnPKC␣ results in an enhanced calcium-dependent increase in promoter activity (Fig. 3B). The immunoblot in Fig. 3B (inset) confirms that the expression vectors produce the appropriate proteins. To determine whether the endogenous hINV gene is regulated in a similar manner, keratinocytes were treated for 48 h in 0.09 or 0.3 mM calcium-containing medium in the absence or presence of 1 M Go6976. Go6976 is an inhibitor of classical PKC isoforms, including PKC␣ (50,51). Calcium causes a 3-fold increase in hINV protein level (Fig. 3C). In cells treated with 0.3 mM calcium and Go6976, hINV levels increase 4-fold. Interestingly, Go6976 also increases hINV levels in cells treated with 0.09 mM calcium, suggesting that PKC␣ may also function to inhibit basal transcription. Based on these results, we predict that PKC␣ expression should inhibit calcium-dependent activation of endogenous hINV gene expression. To assess this possibility cells were treated with 0.3 mM calcium in the presence or absence of a PKC␣-encoding adenovirus. As shown in Fig. 3D, the presence of PKC␣ inhibits the calcium-dependent increase in endogenous hINV gene expression.
Calcium Regulation of PKC␣ and PKC␦ Function-We next asked whether calcium treatment modifies PKC␦ or PKC␣ function. To assay calcium effects on PKC level, keratinocytes were treated with calcium for 0 -48 h, and PKC␣ and PKC␦ levels were measured by immunoblot. Fig. 4A shows that total PKC␣ and PKC␦ levels are not influenced by calcium. To study PKC isoform localization, we fractionated cell extracts into 100,000 ϫ g pellet, cytosol, and the Triton-soluble portion of the 100,000 ϫ g pellet (41). We initially assessed the relative distribution among the pellet, cytosol, and Triton-soluble pellet fractions. To determine whether calcium influences this distribution, we treated keratinocytes for various times in the presence of 0.3 mM calcium, and we assayed the cytosol, Tritonsoluble particulate, and pellet fractions for changes in level of PKC␦ and -␣ by immunoblot. As shown in Fig. 4B, treatment with calcium does not visibly alter the distribution of either FIG. 1. Calcium regulation of human involucrin gene expression. A, normal human keratinocytes growing in 9.5-cm 2 dishes were transfected with 1 g of pINV-41 or pINV-2473. After 24 h, the cells were treated with 0.09 or 0.3 mM calcium for 48 h. Cell extracts were prepared and assayed for luciferase activity. This experiment was repeated four times with similar results. The error bars represent the mean Ϯ S.D. In the inset, normal human foreskin keratinocytes were cultured in medium containing either 0.09 or 0.3 mM calcium chloride for 48 h. Cells were harvested into sample buffer and boiled, and 20 g of protein per lane was electrophoresed in a denaturing 8% polyacrylamide gel. Involucrin was detected using rabbit anti-hINV generated using recombinant human involucrin (34). ␤-Actin levels were monitored as a control to normalize gel loading. B, total cell extracts were prepared from keratinocytes growing in 0.09 mM calcium-containing medium and assayed for immunoblot of PKC␣, -␦, -⑀, -, and -expression. The amount of protein loaded per lane and the film exposure times are indicated. Binding to the primary antibody was detected using an appropriate secondary antibody and visualized using chemiluminescence. PKC isoform. Particulate fraction ␤-actin levels were monitored as a control. We next performed a calcium concentrationresponse curve to determine whether higher levels of calcium may cause PKC translocation. As shown in Fig. 4C, calcium concentrations ranging from 0.09 to 1.8 mM do not cause PKC translocation to membranes. As a positive control for PKC mobilization, we treated keratinocytes for 30 min with 500 nM TPA. Our results confirm, as reported previously (41), that TPA mobilizes PKC␦ and PKC␣ from the cytosol to the Tritonsoluble fraction (Fig. 4D). To confirm the above results visually, we treated keratinocytes for various times with 0.3 mM calcium, and we monitored PKC␣ and PKC␦ subcellular localization by fluorescence microscopy. As shown in Fig. 4E, elevated calcium did not promote detectable translocation of PKC␦. However, a 30-min treatment with 500 nM TPA caused mobilization of PKC␦. We could not monitor PKC␣ movement by immunohistology due to technical difficulties with the antibody; however, the biochemical analysis clearly showed redistribution from cytosol to membrane (see Fig. 4D).
Several agents are known to stimulate phosphorylation of PKC␦ on tyrosine (52,53), and tyrosine phosphorylation can regulate PKC␦ activity and substrate specificity (54,55). We were interested to determine whether calcium treatment produces a covalent modification of PKC␦. To detect tyrosine phos-phorylation, endogenous PKC␦ was immunoprecipitated with anti-PKC␦, and phosphotyrosine was assayed by immunoblot. As shown in Fig. 5, anti-PKC␦ precipitates endogenous PKC␦ from cells treated with low or high calcium (PKC␦ blot). Nonspecific anti-IgG, in contrast, does not precipitate PKC␦. The phosphotyrosine blot of the precipitated material demonstrates that calcium treatment increases PKC␦ tyrosine phosphorylation. We confirmed this finding using extracts prepared from keratinocytes transfected with PKC␦-encoding plasmid (expressed PKC␦). Keratinocytes were transfected with empty plasmid or PKC␦-encoding plasmid, and extracts were prepared at 72 h post-transfection. Fig. 5 shows that expressed PKC␦ can be precipitated and, as with the endogenous enzyme, is tyrosine-phosphorylated following calcium treatment.
Location of hINV Promoter Calcium-and PKC-response Elements-The above results indicate that PKC␦ and -␣ influence calcium-dependent regulation of hINV gene expression. To identify the region of the hINV promoter responsible for this regulation, keratinocytes were transfected with the constructs shown in Fig. 6A and then treated with 0.09 or 0.3 mM calcium for 48 h. As shown in Fig. 6B, calcium increased pINV-2473 and pINV-2216 promoter activity by 4.3-and 3-fold, respectively. In contrast, shorter constructs displayed a reduced calcium-dependent response. This suggests that the Ϫ2473/Ϫ2100 seg- ment contains the calcium-responsive element(s). We next determined whether the AP1 and Sp1 sites, previously shown to be present in this region (Fig. 6A) (31), are required for regulation. As shown in Fig. 6C, mutation of either the AP1 or Sp1, or both sites, reduces basal transcription and eliminates or reduces the calcium-dependent increase.
An important issue is whether the PKC␦-associated hINV promoter activation is mediated via these same elements. To evaluate this, keratinocytes were transfected with each reporter construct in the presence of empty expression vector (ϪPKC␦) or PKC␦-encoding expression vector (ϩPKC␦). As shown in Fig. 7A, PKC␦ markedly increases the activity of constructs pINV-2473 and pINV-2216, suggesting that this region contains a response element. To determine whether the AP1 site is required for activity, we tested a construct in which this site is mutated, pINV-2473(AP1-5m) (31). The results presented in Fig. 7B indicate that the AP1-5 site is required for PKC␦-dependent regulation. A parallel experiment using pINV-2473(Sp1m) shows that mutation of the hINV promoter Sp1 site results in a smaller reduction in calcium-dependent activation (Fig. 7B).
Calcium and PKC Regulation of AP1 Factor Expression-The common requirement for an intact AP1-5 site for both PKC␦and calcium-dependent regulation of hINV gene expression suggests that each stimulus may regulate AP1 factor expression. To evaluate this possibility, we infected keratinocytes with empty vector or PKC␦-encoding adenovirus, and after 48 h we prepared nuclear extracts to assay for AP1 factor levels by immunoblot. Fig. 8A shows that PKC␦ expression increases JunB, c-Fos, and Fra-2 expression and decreases Fra-1 and c-Jun expression. In contrast, JunD levels are not altered. In parallel experiments, we treated keratinocytes for 48 h in medium containing 0.09 or 0.3 mM calcium. As shown in Fig. 8B, although calcium produces similar changes as compared with those observed with PKC␦, Fra-2 levels are increased by PKC␦ but not by calcium. The distal regulatory region of the hINV promoter also includes a functionally important Sp1-binding site (23). Sp1 binds at this site and cooperates with AP1 factors to regulate gene expression (31). We therefore evaluated whether calcium alters Sp1 expression. Fig. 8C shows that nuclear Sp1 levels are substantially elevated in response to a 48-h treatment with 0.3 mM calcium. To further confirm a role for PKC␦ in the regulation of transcription factor levels, we treated cells with dnPKC␦-encoding virus and then treated for 48 h with 0.09 or 0.3 mM calcium prior to preparation of nuclear extracts. As shown in Fig. 8D, the calcium-dependent changes in AP1 factor and Sp1 factor levels are completely inhibited in the presence of dnPKC␦. DISCUSSION Calcium is an important regulator of human and mouse keratinocyte differentiation (15,56). Calcium regulation is manifest in vivo by the presence of an epidermal calcium gradient in which free calcium levels increase in the superficial epidermal layers (6,57). In cultured keratinocytes, intracellular diacylglycerol and intracellular free calcium levels increase with keratinocyte differentiation (58,59), suggesting that these agents may drive differentiation via activation of downstream signaling (28). Because these agents are known activators of PKC, it is likely that some of the calcium-dependent regulation is transmitted via a protein kinase C-dependent mechanism (17). However, detailed information regarding the signal transduction mechanisms mediating this response is limited. A major goal of the present study is to assess the role of PKC␣ and -␦ as mediators of calcium-dependent regulation.
PKC and Calcium Regulate hINV Gene Expression-Previous studies (29,60,61) suggest that calcium regulates hINV gene expression at the mRNA and protein level and suggest that novel PKC isoforms mediate the phorbol ester-dependent increase in hINV gene expression. The PKC regulation is transmitted via a pathway that includes novel PKC, Ras, MEKK1, MEK3, and p38 MAPK (28,29). Because addition of exogenous calcium results in an increase in intracellular keratinocyte diacylglycerol levels (62), it is possible that calcium activates the novel PKC isoforms via a diacylglycerol-dependent mechanism that targets this pathway. Thus, we have investigated whether PKC␦ activity is required for calcium-dependent regulation of hINV gene expression. Our studies, using a dominant-negative mutant of PKC␦, show that inactivation of PKC␦ results in a loss of calcium-dependent hINV promoter activity. In contrast to the PKC␦-associated regulation, PKC␣ suppresses the calcium-associated increase in hINV promoter activity. Consistent with this, an inhibitor of classical PKC isoform function, Go6976, promotes an increase in endogenous hINV gene expression, and inhibition of PKC␦ by dominantnegative PKC␦ inhibits this increase. The opposing effects of FIG. 3. PKC␣ inhibits calcium-dependent hINV gene expression. A, keratinocytes, growing in 0.09 mM calcium medium, were transfected with 1 g of pINV-2473 and 0 -2.5 g of PKC␣ expression vector at a total plasmid concentration of 3.5 g (maintained by addition of empty expression vector). After 24 h, the cells were shifted to medium containing 0.09 or 0.3 mM calcium. At 48 h after calcium addition, cell extracts were prepared for assay of luciferase activity. The values present the mean Ϯ S.D. Similar results were observed in three separate experiments. B, keratinocytes were grown and treated exactly as in A, except that they were transfected with 0 -2 g of dnPKC␣. The inset shows an immunoblot, using anti-PKC␣, demonstrating that the PKC␣ and dnPKC␣ expression vectors produce the corresponding proteins. Endogenous PKC␣ is detected in cells transfected with empty vector (EV). Extracts were isolated 48 h after transfection with 2 g of empty plasmid or plasmid encoding PKC␣ or dnPKC␣. Whole cell lysates were prepared, and equivalent amounts of protein were electrophoresed on an 8% gel, transferred to membrane, and blotted with anti-PKC␣. C, keratinocytes were grown for 48 h in medium containing 0.09 or 0.3 mM calcium in the presence or absence of 1 M Go6976. Go6976 treatment was initiated 45 min prior to calcium treatment. At 48 h, the cells were harvested in sample buffer, and 20 g of whole cell lysate was electrophoresed on an 8% polyacrylamide gel. hINV protein level was assessed by immunoblot, and ␤-actin was used as a loading control. D, keratinocytes were infected with 8 m.o.i. of EV or PKC␣encoding adenovirus and then treated in the presence of 0.09 or 0.3 mM calcium for 48 h. The cells were then harvested in sample buffer, and 20 g of whole cell lysate was electrophoresed on an 8% polyacrylamide gel. hINV and ␤-actin protein levels were assessed by immunoblot.
PKC␦ and -␣ on hINV promoter activity and endogenous gene expression are an interesting finding, as PKC␦ and -␣ have been shown to appose each other in other contexts. For example, PKC␦ is shown to be an activator of apoptosis in keratinocytes and other cell types (63,64), whereas PKC␣ produces anti-apoptotic responses in several cell types (65)(66)(67)(68). Because calcium addition induces both hINV expression and other changes in keratinocytes leading to differentiation-related cell death, our results are consistent with the idea that PKC␦ is a downstream mediator of these effects. This also supports the general hypothesis that PKC␦ and PKC␣ play opposing regulatory roles, i.e. PKC␦ is a pro-apoptosis, pro-differentiation mediator, whereas PKC␣ is a pro-proliferation regulator. This concept is supported by several additional studies (39, 41, 63, 68 -71) but is not supported by others (12, 17), pointing to the complexity of the regulation.
Studies in cultured mouse keratinocytes suggest that PKC inhibits calcium-dependent activation of genes that are normally expressed early (K1, K10) in differentiation (49). In contrast, PKC activation appears to increase expression of the late markers, loricrin and filaggrin (49). In addition, PKC␣ positively regulates calcium-dependent induction of loricrin and filaggrin gene expression in mouse cells but does not influence calcium-dependent K1 expression (19). This suggests that PKC activation produces differential effects on different classes of genes during differentiation. Our present study suggests that PKC␣ inhibits expression of involucrin in human keratinocytes, suggesting a role of PKC␣ in inhibiting spinous layer markers. Because, hINV is first expressed in the late spinous layer, it is possible that PKC␣ functions to keep hINV gene expression off during early spinous differentiation. PKC␦, in contrast, may activate hINV gene expression in the late spi- FIG. 4. Calcium and PKC isoform expression and distribution. A, keratinocytes, growing in 0.09 mM calcium-containing medium, were treated for 0 -48 h in 0.3 mM calcium-containing medium. The cells were then harvested directly into sample buffer, and 20 g of total protein was electrophoresed on an 8% polyacrylamide gel, and the samples were transferred to nitrocellulose for blotting with antibodies specific for each PKC␣, PKC␦, or ␤-actin. B, keratinocytes were grown in high and low calcium for 0 -48 h as outlined above. The cells were then harvested for preparation of Triton-soluble, cytosol, and pellet fractions. An equivalent amount of each fraction, based on total cell number, was electrophoresed on an 8% acrylamide gel, and PKC␦ and -␣ levels were measured by immunoblot. ␤-Actin levels in the pellet fraction were also monitored as a control for loading. C, keratinocytes were grown in the indicated calcium concentration for 48 h. The cells were then harvested for preparation of Triton-soluble, cytosol, and pellet fractions. An equivalent amount of each fraction, based on total cell number, was electrophoresed on an 8% acrylamide gel, and PKC␦ and -␣ level was measured by immunoblot. ␤-Actin levels in the pellet fraction were also monitored as a loading control. D, keratinocytes were treated for 48 h with 0.09 or 0.3 mM calcium or for 30 min with 500 nM TPA. The cells were harvested, fractionated as above, and PKC␣ and -␦ levels were measured by immunoblot. ␤-Actin levels were monitored as an internal control for loading. E, keratinocytes, growing on glass coverslips, were incubated for various times in KSFM containing 0.3 or 1.8 mM calcium or 500 nM TPA. The cells were fixed and permeabilized, and PKC␦ was detected using the PKC␦ isoform-specific primary antibody and Oregon Green 514-linked goat anti-rabbit IgG secondary antibody. Digital images were obtained using a Nikon Optiphot microscope. The panel marked nonspecific (NS) included, during the primary antibody incubation, PKC␦ antibody blocking peptide at a 5-fold weight excess to the antibody. The arrows in PKC␦ 30 min TPA treatment indicate membrane-associated PKC␦. nous and granular layers. One previous study (72) examined the role of PKC as a regulator of hINV gene expression. In contrast to our findings, these investigators showed that TPAdependent hINV promoter activity is increased by PKC␣ and is not influenced by PKC␦. However, this study differs from the present study in several important respects. First, the cells used were SV40 large T antigen-immortalized keratinocytes. Second, the hINV promoter construct used in this study did not contain the sequences identified in the present report. In addition, studies in our laboratory, using an extensive set of immortalized keratinocyte cell lines, suggest that regulation of hINV gene expression is markedly attenuated and abnormal in most transformed cell lines.
Our studies also indicate that calcium treatment is associated with enhanced phosphorylation of PKC␦. This result is in agreement with a recent report (40) in mouse keratinocytes showing a calcium-dependent increase in phosphorylated PKC␦ in cultured keratinocytes. Phosphorylated PKC␦ was also detected in vivo in the mouse epidermis (40). PKC␦ phosphorylation can activate or inhibit the enzyme, depending upon the stimulus (52,53,73,74). Moreover, the direction of change in catalytic activity may be substrate-dependent (54). Thus, although our studies clearly show that calcium treatment produced covalent changes in PKC␦, further studies will be necessary to determine whether the tyrosine phosphorylation of PKC␦ in our system activates or inhibits the enzyme.
In addition, a surprising finding from our study is that calcium addition did not induce significant mobilization of PKC␣ or PKC␦ to membrane fractions. Membrane mobilization is usually thought to be necessary for PKC activity but may not be absolutely required. The apparent lack of calcium-dependent PKC mobilization in our study is not an artifact, because TPA treatment did, as reported previously (41), mobilize PKC␦ and PKC␣. It is possible that the PKC isoforms are cycling to and from the membrane at a steady rate that is not detected in our assays and that active, membrane-associated, forms are thus generated continually. Such cycling has been reported for PKC␦ in ceramide-treated cells (75). Increased membrane-associated PKC activity has also been reported in calcium-treated mouse keratinocytes (76), and this is associated with PKC␣ and -␦ movement to membranes (77). It is also possible that PKC, resident at the membrane before calcium treatment, simply becomes active in the presence of calcium. For example, PKC␦ is activated by H 2 O 2 in Chinese hamster ovary cells in the absence of membrane translocation (78). Moreover, this translocation-independent activation is associated with tyrosine phosphorylation of PKC␦ (78). Thus, the phosphorylation of PKC␦ described in the present study may be important in this context. Although tyrosine phosphorylation has been reported to reduce PKC␦ activity in mouse keratinocytes (79), the effect of this modification is likely to be context-dependent. Additional studies will be required to understand the effect of this modification in our system. To detect tyrosine phosphorylation of endogenous PKC␦, keratinocytes were grown in the presence of 0.09 or 0.3 mM calcium for 48 h, and extracts were prepared in lysis buffer for immunoprecipitation using anti-PKC␦ or anti-IgG. The precipitate was electrophoresed on an 8% acrylamide gel, and the separated proteins were detected using antiphosphotyrosine (P-Tyrosine blot). The blot was then stripped and incubated with anti-PKC␦ to ensure that equivalent amounts of PKC␦ were precipitated (PKC␦ blot). To confirm this using expressed PKC␦, keratinocytes were transfected with 10 g of PKC␦-encoding plasmid per 50-cm 2 dish. At 24 h after transfection, the cells were treated with 0.09 or 0.3 mM calcium for 48 h. Cell lysates were prepared and processed as outlined for endogenous PKC␦.  (32). The open box adjacent to the AP1-5 site indicates the Sp1-binding site (31,32). The nucleotide positions indicated below the line define the left end of each promoter construct. B, keratinocytes were transfected with 1 g of each hINV promoter reporter construct, and after 24 h, the cells were treated for 48 h with KSFM containing 0.09 or 0.3 mM calcium. The cells were then harvested, and cell lysates were assayed for luciferase activity. The numbers in parentheses indicate the fold increase in response to calcium treatment. Similar results were observed in each of five separate experiments. C, cells were transfected with 1 g of intact pINV-2473 or pINV-2473 containing inactivating mutations in the AP1-5 (AP1-5m), Sp1 (Sp1m), or both sites (AP1-5m/Sp1m). At 24 h post-transfection, the cells were shifted for 48 h to medium containing 0.09 or 0.3 mM calcium. Cell extracts were then prepared and assayed for luciferase activity that is normalized per g of protein (28). The values represent the mean Ϯ S.D. Similar results were observed in four separate experiments.
PKC Regulation Targets the hINV Promoter Distal Regulatory Region-In some systems, AP1 transcription factors are the downstream targets of PKC-dependent regulation (16,17). For example, recent studies (28 -30, 80) show that a Ras, MEKK1, MEK3, p38 MAPK cascade mediates the phorbol esterdependent increase in hINV gene expression and that this cascade targets AP1, Sp1, and C/EBP transcription factors. These factors, in turn, interact with selected binding motifs within the hINV promoter to regulate expression (31,32). These motifs are localized in two major regions, the proximal regulatory region and the distal regulatory region (23,32). Our present promoter truncation studies identify the distal regulatory region as containing the calcium-response elements. Targeted mutation of the Sp1 and AP1-5 sites reveals that both sites are required for the calcium-dependent response. Mutation of the AP1-5 site results in the complete elimination of calcium-dependent regulation. Mutation of the Sp1 site results in partial loss of the calcium-dependent response. These findings are particularly interesting, as they suggest that calciumdependent regulation of hINV gene expression shares common features with phorbol ester-dependent regulation. Moreover, this segment encompasses a DNA regulatory region that is required for tissue-specific (epidermis) and differentiation-appropriate (suprabasal layers) expression of hINV in transgenic mice (23,32,81).
PKC␦ and Calcium Regulate AP1 and Sp1 Factor Expression-One common mechanism whereby calcium and PKC␦ may regulate hINV gene expression is through alteration of transcription factor levels. Our results show that treatment with either calcium or PKC␦ increases c-Fos and JunB and decreases c-Jun and Fra-1 levels. Fra-2 levels, in contrast, are increased by PKC␦ but not by calcium. In addition, the calciumassociated change in AP1 factor level requires PKC␦ activity. These results suggest that regulation via both upstream modulators converges on AP1 factors. Because the extent of transcriptional activation or repression is a function of the particular AP1 heterodimers that are formed, any relative change in AP1 factor level may alter gene expression (82,83). Interestingly, the PKC␦ and calcium treatment produce similar changes in AP1 factor expression. In mouse keratinocytes, increased AP1 factor expression is also associated with cell confluence and enhanced differentiation (17). In addition to the increase in AP1 levels, Sp1 levels also increase in the presence of calcium. Moreover, as measured using dominant-negative PKC␦, the increase in Sp1 requires PKC␦ activity. This suggests that Sp1 transcription factors may help mediate calciumdependent gene expression via a PKC␦-dependent mechanism. Sp1 has been reported to be a key participant in the phorbol ester-dependent induction of gene expression (80,84). Additional studies will be required to determine the mechanism FIG. 8. Calcium-versus PKC␦-dependent regulation of AP1 and Sp1 transcription factor levels. A, keratinocytes growing in 0.09 mM calcium-containing medium were infected with empty adenovirus (EV) or PKC␦-encoding adenovirus at 8 m.o.i. and maintained for 48 h. At 48 h nuclear extracts were prepared, and equivalent quantities of protein were electrophoresed on an 8% gel and transferred to nitrocellulose for immunoblot with AP1 factor-selective antibodies (c-Fos, 1:500; Fra-1, 1:2500; Fra-2, 1:500; c-Jun, 1:5000; JunB, 1:5000; JunD 1:2500). Primary antibody binding was detected using peroxidaselinked secondary antibodies, and binding was visualized by chemiluminescence. B, keratinocytes were grown in KSFM containing 0.09 or 0.3 mM calcium. After 48 h nuclear extracts were prepared, and equivalent quantities of protein were electrophoresed on an 8% gel and transferred to nitrocellulose for immunoblot with AP1 factor-selective antibodies as above. C, nuclear extracts were prepared from keratinocytes following calcium treatment as described in Fig. 7B. Sp1 levels were monitored by immunoblot using an Sp1-specific antibody (dilution ϭ 1:500). D, keratinocytes were infected with 8 m.o.i. of dnPKC␦-encoding vector (ϩ) and then incubated for 48 h in the presence of 0.09 or 0.3 mM calcium. Nuclear extracts were prepared and assayed for AP1 factor or Sp1 expression by immunoblot. A control group treated with 0.3 mM calcium and empty vector produced the same regulatory responses shown in B (not shown). It should be noted that the experiments in each panel were performed as separate experiments, and so signal intensity cannot be compared among panels. FIG. 7. Localization of PKC␦-responsive region in hINV promoter. A, keratinocytes were transfected with 1 g of hINV promoter plasmid and 2 g of PKC␦ expression plasmid (ϩPKC␦) or empty expression plasmid (ϪPKC␦) in low calcium medium. After 48 h, the cells were harvested, and cell lysates were assayed for luciferase activity. The numbers in parentheses indicate the fold increase in response to calcium treatment. Similar results were observed in each of five separate experiments. B, cells were transfected with 1 g of intact pINV-2473, pINV-2473(AP1-5m), which contains an inactivating mutation at the AP1-5 site (31), or pINV-2473(Sp1m), which contains an inactivating mutation at the Sp1 site (31), and 2 g of PKC␦ expression plasmid or empty control plasmid. After 48 h, cell extracts were prepared and assayed for luciferase activity that is normalized per g of protein (28). The values represent the mean Ϯ S.D. Similar results were observed in five separate experiments.
whereby AP1 and Sp1 factors regulate calcium-dependent hINV gene expression; however, it is possible that Sp1 may facilitate the response by assisting AP1 factor binding to DNA (31).
In summary, our results are consistent with the hypothesis that PKC␦ and calcium activate keratinocyte differentiation via a mechanism that results in increased expression of AP1 and Sp1 transcription factors. Moreover, PKC␦ and PKC␣ appear to produce opposing effects on calcium-dependent keratinocyte differentiation, PKC␦ being an activator and PKC␣ functioning as an inhibitor of involucrin gene activation.