Novel Protein Kinase C Isoforms Regulate Human Keratinocyte Differentiation by Activating a p38δ Mitogen-activated Protein Kinase Cascade That Targets CCAAT/Enhancer-binding Protein α*

The novel protein kinase C (nPKC) isoforms are important regulators of human involucrin (hINV) gene expression during keratinocyte differentiation (Efimova, T., and Eckert, R. L. (2000) J. Biol. Chem. 275, 1601–1607). Although the regulatory mechanism involves mitogen-activated protein kinase (MAPK) activation, the role of individual MAPK isoforms has not been elucidated. We therefore examined the effects of individual nPKCs on MAPK activation. We observe unique changes whereby nPKC expression simultaneously increases p38 activity and decreases ERK1 and ERK2 activity. Although p38α, p38β, and p38δ are expressed in keratinocytes, only a single isoform, p38δ, accounts for the increased p38 activity. Parallel studies indicate that this isoform is also activated by treatment with the keratinocyte regulatory agents, 12-O-tetradecanoylphorbol-13-acetate, calcium, and okadaic acid. These changes in MAPK activity are associated with increased C/EBPα transcription factor expression and DNA binding to the hINV promoter and increased hINV gene expression. Expression of PKCδ, PKCε, or PKCη causes a 10-fold increase in hINV promoter activity, whereas C/EBPα expression produces a 25-fold increase. However, simultaneous expression of both proteins causes a synergistic 100-fold increase in promoter activity. These responses are eliminated by the dominant-negative C/EBP isoform, GADD153, and are also inhibited by dominant-negative forms of Ras, MEKK1, MEK3, and p38. These results suggest that the nPKC isoforms produce a unique shift in MAPK activity via a Ras, MEKK1, MEK3 pathway, to increase p38δ and inhibit ERK1/2 and ultimately increase C/EBPα binding to the hINV promoter and hINV gene expression.

In previous studies, we demonstrated that the novel PKC isoforms PKC␦, PKC⑀, and PKC, but not the conventional and atypical PKC forms, activate keratinocyte differentiation as measured by effects on human involucrin (hINV) gene expression (10,11). Involucrin is a precursor of the keratinocytecornified envelope and a marker of early keratinocyte differentiation (12). Dominant-negative PKC␦ inhibits this response (10). This pathway appears to operate by triggering a cascade that includes Ras, MEKK1, and MEK3. Although this pathway is known to require MAPK activity, whether individual nPKCs activate different MAPKs has not been explored. In the present study we activate the signaling cascades at the PKC level by expression of individual nPKC isoforms in keratinocytes and monitor the effects on MAPK function and downstream responses. Our results show that nPKC activation results in increased p38␦ MAPK activity and reduced ERK1/2 activity. No change in activity of other p38 isoforms (p38␣ and p38␤) is observed. This combination of changes leads to increased C/EBP␣ transcription factor expression and activity and increased hINV gene expression. These results suggest that nPKC inhibits ERK1/2 expression and activates p38␦ via a Ras, MEKK1, MEK3 pathway that targets C/EBP␣ to increase hINV gene expression.

MATERIALS AND METHODS
Reagents-Keratinocyte serum-free medium, gentamicin, trypsin, and Hanks' balanced salt solution were obtained from Invitrogen. Dispase was obtained from Roche Molecular Biochemicals. Phorbol ester (12-O-tetradecanoylphorbol-13-acetate (TPA)) and Me 2 SO were purchased from Sigma. Bis-indolylmaleimide and okadaic acid were from Calbiochem. The pGL2-Basic plasmid and the chemiluminescent luciferase assay system were obtained from Promega. Chemiluminescence was measured using a Berthold luminometer. Oligonucleotides for construction of mutant promoter sequences were synthesized using an Applied Biosystems DNA synthesizer. PKC isoform-selective (PKC␦, sc-937; PKC, sc-215) and C/EBP␣-selective (C/EBP␣, sc-61x) antibodies were obtained from Santa Cruz Biotechnology. The rabbit antiinvolucrin antibody was described previously (13).
Tissue Culture, Cell Transfection, and Luciferase Assay-Normal human foreskin keratinocytes were cultured as described previously (14). Third passage keratinocytes were transfected in 35-mm-diameter dishes when ϳ50% confluent using Fugene-6 reagent as described previously (10,25). The final DNA concentrations in all groups were adjusted to identical levels using empty expression vector. The luciferase activity was assayed immediately using a Promega luciferase assay kit and a Berthold luminometer. All of the assays were performed in triplicate, and each experiment was repeated a minimum of three times. Luciferase activity was normalized per g of protein as described previously (10,14). Transfection efficiency was monitored using a green fluorescent protein expression plasmid.
Recombinant Adenovirus Vectors-nPKC-expressing adenoviruses were previously constructed (26). Adenoviruses encoding constitutively active MEK6 and wild type FLAG-tagged p38 MAPK isoforms ␣, ␤, ␥, and ␦ were obtained from Dr. Y. Wang (27,28). A control (empty) adenovirus, Ad5-EV, was generated by recombining pCA3 plasmid with the pJM17 adenovirus backbone. Recombinant adenoviruses were propagated in 293 cells and purified by cesium chloride centrifugation. The optimal multiplicity of adenoviral infection was determined using the green fluorescent protein-encoding adenovirus. The adenoviruses were administered at the indicated MOI in the presence of 2.5 g/ml polybrene.
Immunoblot Analysis-Total cell or nuclear extracts were prepared from cultured human epidermal keratinocytes as described previously (15). Equal quantities of protein were electrophoresed on a 10% denaturing SDS-polyacrylamide gels and transferred to nitrocellulose. The membranes were blocked, incubated with an indicated primary antibody, washed, and exposed to an appropriate horseradish peroxidaseconjugated secondary antibody. Secondary antibody binding was visualized using chemiluminescent detection methods (29).
MAPK Assays-The activities of MAPKs were measured using nonisotopic p44/42 (ERK1/2), p38 MAPK, and JNK/stress-activated protein kinase assay methods (New England Biolabs) (25). Briefly, keratinocyte total cell lysates were prepared under nondenaturing conditions. Equal amounts of total protein (200 g) were used per each kinase assay. Immobilized, dual phospho-ERK1/2 and phospho-p38 MAPK monoclonal antibodies were used to selectively immunoprecipitate active (phosphorylated) ERK1/2 and p38 kinases, respectively. JNK/ stress-activated protein kinase was selectively precipitated from cell lysates using c-Jun fusion protein glutathione-Sepharose beads. Precipitated kinases were then allowed to phosphorylate substrate proteins (Elk-1 for ERK1/2, ATF-2 for p38, and c-Jun for JNK/stress-activated protein kinase) in a kinase reaction performed in the presence of ATP. Phosphorylation of the substrate proteins was analyzed by immunoblot using phosphorylated substrate-specific antibodies. The activity of adenovirus-delivered FLAG-tagged p38 MAPK isoforms was measured by precipitating the tagged kinases using anti-FLAG M2 mouse monoclonal antibody (F3165; Sigma). Expression of individual FLAG-p38 isoforms was confirmed by immunoblot using anti-FLAG antibody. To measure the activity of endogenous p38␦ isoform, p38␦-specific antibody (SC-7585) was used to selectively immunoprecipitate this enzyme followed by a kinase assay performed as outlined above.

Regulation of MAPK Activity by PKC␦ and PKC-
To identify the mechanism whereby the nPKC isoforms regulate gene expression during keratinocyte differentiation, we transfected keratinocytes with empty adenovirus or adenovirus encoding PKC␦ or PKC. At 48 h, the cells were harvested, and an in vitro kinase assay was performed by immunoprecipitating the activated forms of all p38 isoforms using anti-phospho-p38. Kinase activity was then measured based on the ability of the precipitated kinase to phosphorylate ATF-2. Parallel assays were performed to measure ERK1/2 and JNK1/2 activity. As shown in Fig. 1A, PKC and PKC␦ markedly increase p38 activity. In contrast, ERK1/2 activity is suppressed (P-ELK1). JNK activity was not detected (P-c-Jun). To assure that the observed changes in kinase activity were not due to regulation of kinase level, we measured endogenous kinase levels by immunoblot. Fig. 1B shows that nPKC expression does not alter the level of p38, JNK1/2, or ERK1/2. To evaluate whether the decline in overall ERK activity is selectively associated with reduced ERK1 or ERK2 activity, we compared total ERK1/2 and P-ERK1/2 levels by immunoblot. Fig. 1C shows that although total ERK1/2 levels are not altered by PKC␦ or treatment, both isoforms are proportionately reduced in activity as measured by diminished levels of phosphorylated ERK1/2 (P-ERK1/2). Fig. 1A shows that JNK1/2 is not active. Thus, as a positive control to assure that the JNK assay is functional, we treated keratinocytes for 24 h with okadaic acid, prepared extracts, and assayed JNK activity. As shown in Fig. 1D, okadaic acid-dependent JNK activity is readily detected, confirming the validity of the assay. Finally, it is important to confirm that the vector-delivered proteins are expressed. Thus, extracts were prepared from adenovirus expression vector-infected cells, and the level of PKC isoform expression was monitored by immunoblot. The blot shown in Fig. 1E confirms that PKC␦ and are expressed. It should be noted that, as expected (10), endogenous PKC expression is detected when the film exposure time is extended, and overexpression of selected PKC isoforms does not affect the expression of the endogenously expressed PKCs (not shown). Taken together, these results provide evidence that nPKC isoforms ␦ and activate p38, inhibit ERK1/2, and have no effect on JNK1/2 activity. Activation of p38 Isoforms by PKC and PKC␦-p38 MAPK exists as a family of four distinct isoforms (␣, ␤, ␦, and ␥) that have different biological functions (30). To identify which p38 isoform is activated by PKC and ␦, keratinocytes were infected with FLAG-tagged p38␣, ␤, ␥, or ␦ in the presence of empty virus or virus encoding PKC or PKC␦. After 48 h, the cells were harvested, and individual FLAG-p38 isoforms were precipitated using anti-FLAG antibody and assayed for the ability to phosphorylate ATF-2 (29, 31). Fig. 2A shows that adenovirus-delivered FLAG-p38␣, ␥, and ␦ are active in control (empty virus infected) cultures and that p38␦ and ␥ activities are increased in the presence of nPKC expression. In contrast, no p38␤ activity is detected for any of the treatment conditions. Our previous study using quantitative reverse transcription-PCR indicates that p38␥ is not expressed in keratinocytes (29). Thus, the only p38 isoform that is increased in activity in response to nPKC is p38␦. To confirm that the vector-delivered p38 isoforms are expressed, we prepared cell extracts and performed immunoblots using a FLAG-specific antibody. Fig. 2B shows that, with the exception of p38␥, which is expressed at higher levels, each FLAG-p38 isoform is expressed at a relatively similar level.
The above results suggest that p38␦ is the major p38 isoform activated by PKC␦ and . To determine whether endogenous p38␦ behaves in a similar manner, we infected keratinocytes with empty vector or vectors encoding PKC␦ or PKC. After 48 h, endogenous p38␦ was selectively immunoprecipitated, and p38␦ activity was assayed based on the ability to phosphorylate ATF-2. As shown in Fig. 2C, PKC␦ and stimulate the activation of endogenous p38␦. In contrast, the dominant-negative forms of PKC␦ and (dnPKC␦ and dnPKC) do not regulate p38␦ activity, suggesting that this regulation requires nPKC activity. To assure that these changes in p38␦ activity are not the results of altered p38␦ levels, we prepared extracts 1:5000), rabbit anti-JNK1/2 (Sigma J4500 diluted 1:2000), and rabbit anti-ERK1/2 (Sigma M5670 diluted 1:5000) antibodies, followed by detection with horseradish peroxidase-conjugated donkey anti-rabbit IgG (Amersham Biosciences NA934 diluted 1:10,000). These experiments were repeated a minimum of three times with similar results. ␤-Actin levels were monitored as a control for loading. C, cells were treated as in A, and total ERK1/2 level and phosphorylated ERK1/2 level were monitored by immunoblot using rabbit anti-ERK1/2 (Sigma M5670) and mouse monoclonal anti-phospho-ERK1/2 (Santa Cruz Biotechnology, sc-7383) antibodies, respectively. D, keratinocytes were treated for 24 h in the absence (Ϫ) or presence (ϩ) of 100 nM okadaic acid, and JNK1/2 activity was monitored as outlined above. E, normal keratinocytes were infected with empty adenovirus (EV) or PKC␦ or PKC-encoding adenovirus. After 48 h, the cell extracts were prepared, and PKC␦ and PKC levels were measured by immunoblot using rabbit anti-PKC␦ (Santa Cruz Biotechnology, sc-937), and rabbit anti-PKC (Santa Cruz Biotechnology, sc-215).
FIG. 1. PKC and ␦ regulate MAPK activity. A, normal human keratinocytes were infected with empty adenovirus (EV) or infected with adenovirus encoding PKC or ␦ at a MOI of 15, and 48 h later the cells were lysed for extract preparation. Activated p38, JNK1/2, and ERK1/2 were precipitated, respectively, using mouse monoclonal antiphospho-p38␣, ␤, ␦, and ␥ (New England Biolabs 9219), c-Jun fusion beads (New England Biolabs 9811), and mouse monoclonal anti-phospho-ERK1/2 (New England Biolabs 9109). Activity of precipitated p38, ERK1/2, and JNK1/2 kinases was monitored, respectively, by measuring phosphorylation of ATF-2 using rabbit anti-phospho-ATF-2 (New England Biolabs 9221S), c-Jun using rabbit anti-phospho-c-Jun (New England Biolabs 9810), and ELK1 using rabbit anti-phospho-ELK1 (New England Biolabs 9181). B, p38, JNK1/2, and ERK1/2 levels were assayed by immunoblot using rabbit anti-p38 (Sigma M0800 diluted FIG. 2. Regulation of p38 MAPK isoforms by PKC and ␦. A, to measure the enzymatic activity of individual p38 isoforms in response to nPKC expression, keratinocytes were co-infected with empty adenoviral vector (EV) or adenovirus encoding PKC␦ or PKC and with FLAG-p38␣, ␤, ␥, or ␦. After 48 h, individual FLAG-tagged p38 isoforms were immunoprecipitated (200 g of protein/sample) using mouse monoclonal anti-FLAG antibody M2 (Sigma F3165, 5 g/precipitation) and 30 l of protein A/G-agarose (Santa Cruz Biotechnology, sc-2003), and p38 activity was monitored based on the ability of the precipitated kinase to phosphorylate ATF-2. B, immunoblot showing that adenovirus-delivered p38 ␣, ␤, ␥, and ␦ are expressed at comparable levels in keratinocytes. The cells were lysed 48 h after infection, and each p38 isoform was detected by immunoblot using peroxidase-conjugated mouse monoclonal anti-FLAG antibody (Sigma A8592, diluted 1:10000). C, endogenous p38␦ activity is regulated by nPKCs. The keratinocytes were infected with empty adenovirus (EV) or adenovirus encoding PKC, PKC␦, dnPKC, or dnPKC␦. After 48 h, the cells were harvested and endogenous p38␦ was precipitated using goat anti-p38␦ (Santa Cruz sc-7585, 5 g/precipitation). p38 kinase activity was monitored based on ability to phosphorylate ATF-2 (P-ATF-2) as described for Fig.  1. In parallel, the p38␦ levels were monitored by immunoblot using goat anti-p38␦ at a dilution of 1:1000. ␤-Actin levels were monitored to control for gel loading. from cells at 48 h after nPKC-encoding virus infection and measured p38␦ levels by immunoblot. The middle panel of Fig.  2C shows that PKC expression does not alter p38␦ levels. The bottom panel is a ␤-actin immunoblot that is included to assure even loading.
Regulation of ERK1/2 and p38␦ Activity by Differentiating Agents-If nPKC activation is important for response to agents that regulate keratinocyte differentiation, we would anticipate that treatment with these agents would produce a change in MAPK activity similar to that of expression of nPKC isoforms. To address this issue, cells were infected with a FLAG-p38␦ encoding adenovirus. After 24 h, the cells were treated with 0.3 mM calcium, 100 nM okadaic acid, or 50 ng/ml TPA, agents that regulate nPKC activity or the activity of enzymes immediately downstream of the nPKCs. At the indicated times, the extracts were prepared, and FLAG-38␦ was precipitated and assayed for activity. As measured by the ability to phosphorylate ATF-2, each of these agents induced strong and sustained activation of p38␦; however, the time course of the calcium-dependent activation was slower (Fig. 3A). It should be noted that okadaic acid-and TPA-dependent p38␦ activity was maximally increased by 6 h and remained maximal until 24 h (not shown). ERK1/2 activity was measured directly by monitoring the level of phosphorylated ERK1/2 following treatment for the time indicated. Fig. 3B shows that phosphorylated ERK1 and ERK2 levels are reduced by TPA and okadaic acid treatment. However, calcium did not reduce the level of phosphorylated ERK1 or ERK2. These changes in p38␦ and ERK1/2 activity were not associated with changes in p38␦ or ERK1/2 levels as measured by immunoblot (not shown).
PKC␦ and PKC Regulate C/EBP␣ Expression-The above results suggest that both differentiating agents and nPKC isoforms regulate MAPKs in similar manners. An important aspect of how the nPKC-dependent MAPK cascades regulate keratinocyte differentiation is identification of the downstream targets that mediates activation of differentiation-dependent gene expression. C/EBP␣ has been proposed as a key transcriptional regulator of differentiation (15,32,33). Our previous study suggests that TPA-dependent regulation of differentiation-associated gene expression requires C/EBP␣ activity (15). We next determined whether nPKC isoforms regulate C/EBP function. We infected keratinocytes with native or dominantnegative forms of PKC␦ or , and after 48 h monitored C/EBP␣ level. Fig. 4A shows that C/EBP␣, consistent with previous reports (15), is expressed in control (empty vector-infected) cells. The present experiment shows that this expression is markedly increased by PKC␦ and PKC and markedly reduced by expression of dnPKC␦ or dnPKC. Thus, C/EBP␣ levels increase in a PKC␦ and PKC-dependent manner. Based on the above results, we suggest that PKC␦ and increase C/EBP␣ expression via a p38␦-dependent mechanism. To test this hypothesis, we infected keratinocytes with empty adenovirus or adenoviruses encoding p38␦, constitutively active MEK6 (caMEK6), or both. After 48 h, the cells were harvested, and the C/EBP␣ levels were monitored by immunoblot. As shown in Fig. 4B, adenovirus-dependent expression of p38␦ causes a marked increase in C/EBP␣ level. This experiment also demonstrates that caMEK6 enhances C/EBP␣ expression. MEK6 is a MAPK kinase that is an immediate upstream activator of p38 (30,31). Thus, the fact that p38␦ increases C/EBP␣ expression that is further increased when both MEK6 and p38␦ are present further suggests that the C/EBP␣ increase is p38␦-dependent.
Synergistic Activation of Target Gene Expression by nPKC Isoforms and C/EBP␣-The above studies suggest that nPKCs, via activation of p38␦ and C/EBP␣, regulate differentiationassociated gene expression in keratinocytes. To evaluate the impact of this regulation on differentiation, we studied involucrin gene expression. Involucrin is a marker of keratinocyte differentiation that is regulated via a C/EBP␣-dependent mechanism (15). We transfected keratinocytes with pINV-241, a luciferase-linked hINV promoter reporter plasmid (14). The cells were then treated with expression plasmids encoding PKC␦, PKC⑀, PKC, or C/EBP␣. As shown in Fig. 5A, C/EBP␣ expression causes a 25-fold increase in promoter activity, whereas each nPKC isoform (nPKC␦, ⑀, and ) produces a 10-fold activity increase. Remarkably, simultaneous expression of nPKC with C/EBP␣ results in a Ͼ100-fold activation. The

FIG. 3. Regulation of MAPK activity by differentiating agents.
A, keratinocytes were infected with 15 MOI of empty adenovirus (EV) or adenovirus encoding FLAG-p38␦. After 24 h, the cells were treated with 0.3 mM calcium, 50 ng/ml TPA, or 100 nM okadaic acid. None of these treatments altered the level of FLAG-p38␦ (not shown). At the indicated times, the extracts were prepared, and FLAG-p38␦ was precipitated using anti-FLAG. The precipitated kinase was assayed for activity based on the ability to phosphorylate ATF-2, and phospho-ATF2 was detected using a specific antibody. B, keratinocytes were treated with 0.3 mM calcium, 50 ng/ml TPA, or 100 nM okadaic acid. At the indicated times, the total cell extracts were prepared and assayed by immunoblot for phosphorylated ERK1/2 or ␤-actin. Binding of the primary antibody was detected using secondary chemiluminescence detection reagents.
inset of Fig. 5A confirms by immunoblot that each of the three PKC isoforms tested (␦, ⑀, ) and C/EBP␣ are expressed at appropriate levels from the delivery vector.
To confirm the importance of C/EBP␣ for PKC-dependent activation of gene expression, we treated pINV-241-transfected cells with PKC␦, ⑀, or in the presence or absence of a dominant-inactivating C/EBP form, GADD153 (34 -36) and monitored the effects on promoter activity. We first confirmed (Fig.  5B, inset) that vector-produced GADD153 does accumulate within the cell. Fig. 5B shows that GADD153 completely suppresses the PKC-dependent increase in transcriptional activity. Thus, functional C/EBP␣ protein is required for PKC-dependent gene activation. We next tested the ability of nPKC isoforms and C/EBP␣ to activate a hINV promoter construct that contains a mutated C/EBP transcription factor-binding site. As shown in Fig. 5C, mutation of this site eliminates the ability of the nPKCs and C/EBP␣ to enhance gene expression.
The results presented in Fig. 4 demonstrate that PKC␦ and increase C/EBP␣ expression. To further assess the importance of this increase for regulation of hINV transcription, we measured C/EBP␣ binding to the C/EBP-binding site in PKC␦or PKC-expressing cells. A 32 P-labeled oligonucleotide encoding the hINV promoter C/EBP-binding site (15) was incubated with nuclear extracts prepared from keratinocytes that had been treated for 48 h with empty vector or vectors encoding PKC␦ or PKC. The bands were then visualized by gel mobility shift (15). Fig. 5D shows that binding of C/EBP␣ to the C/EBPbinding site is minimal in empty vector-infected cells but is markedly increased in cells expressing PKC␦ or . This binding is specific, because supplementing the gel shift mixture with a 100-fold molar excess of radioinert homologous oligonucleotide successfully competes the signal (Fig. 5D, PKC␦ ϩ 100ϫ). Additional specificity studies were not performed, because supershift analysis has identified C/EBP␣ as the major binding entity at this site (15). A similar increase in C/EBP␣ binding has been observed at this site following TPA treatment of keratinocytes (15).

PKC␦ and PKC Regulate Expression of the Endogenous hINV Gene-
The above studies demonstrate that the novel PKC isoforms activate hINV promoter expression. To confirm the physiological relevance of this observation, we examined the effects of PKC and p38␦ on expression of endogenous hINV. Keratinocytes were infected with empty expression vector or vectors encoding PKC, p38␦, or both. After 48 h, the cells were harvested, and the total cell extracts were prepared for immunoblot. Fig. 5E shows that involucrin levels are minimal in cells infected with empty vector. However, hINV levels are markedly increased in cells expressing PKC or p38␦ and further increased in cells expressing both PKC and p38␦.
p38 MAPK and ERK1/2 Regulation of nPKC-dependent hINV Promoter Activity-The results presented in Figs. 1 and 2 are consistent with the idea that p38␦ functions as an activator and ERK1/2 as an inhibitor of the nPKC-dependent regulation. To test this, we monitored the ability of dominant-negative forms of p38, ERK1 and ERK2, to alter the PKC␦-dependent increase in hINV promoter. The keratinocytes were transfected with PKC␦ in the absence and presence of dnp38, dnERK1, or dnERK2. As shown in Fig. 6A, PKC␦ produces an 8-fold increase in hINV promoter activity. Co-expression of dnp38 with PKC␦ completely inhibits this response. In contrast, dnERK1 markedly enhances the PKC␦-dependent activation. dnERK2 does not alter the regulation. These results suggest that p38 is a mediator of the PKC␦-dependent regulation, and ERK1 is an inhibitor. Identical results were observed for PKC⑀-and PKCdependent gene activation (not shown).
Ras, MEKK1, and MEK3 Activity Are Required for nPKC-dependent Regulation-To identify kinases that mediate transfer of the nPKC-dependent signal to the MAPKs, we transfected keratinocytes with PKC␦ in the absence and presence of dnRas, dnRaf1, dnMEKK1, dnMEK3, and dnJNK. Fig. 6B shows that dnRas, dnMEKK1, and dnMEK3 inhibit the PKC␦-dependent increase in hINV promoter activity. In contrast, dnRaf1 and dnJNK do not influence this increase. Identical results were observed for PKC⑀ and (not shown).

Novel PKC Isoforms Drive Keratinocyte Differentiation via
Activation of p38␦ MAPK-Epidermal keratinocytes and other stratified squamous surface epithelia undergo a regulated process of differentiation (12). They begin as proliferative cells and are ultimately converted to nonproliferating cells called corneocytes that assemble the epidermal surface (12). This change involves a remarkable number of gene activation events that result in the expression of a group of proteins, including involucrin, that are designed to construct this barrier (12,37,38). Specific PKC isoforms are implicated in this process. For example, recent studies show that the nPKC isoforms activate keratinocyte differentiation (10,26) and that this response is associated with growth inhibition (26). This response is also associated with increased expression of differentiation-associated marker genes, including type I transglutaminase (26) and hINV (10). A major goal is identification of the cascades that link nPKC activation to increased expression of target genes. In this regard, we have shown that an nPKC, Ras, MEKK1, MEK3, and p38 cascade regulates hINV gene expression (10,25). Although these studies showed that p38 is an important intermediary in this cascade, they did not identify which of the four known p38 isoforms are required (25). Identifying the active isoform is important, because individual p38 isoforms are known to produce varying biological responses in a cell type-specific manner (39 -42), and three p38 isoforms, ␣, ␤, and ␦, are expressed in keratinocytes (29). Our present study indicates that, in keratinocytes, the nPKCs selectively activate p38␦ and not p38␣ or p38␤. These results are in agreement FIG. 4. Novel PKC-and p38␦-dependent regulation of C/EBP␣ level. A, keratinocytes, growing in 50-cm 2 dishes were infected with empty adenovirus vector (EV) or adenovirus encoding wild type or dominant-negative forms of PKC␦ or PKC at a MOI of 15. After 48 h, the cells were harvested, and an equivalent amount (7 g) of nuclear extract was electrophoresed on a 10% denaturing polyacrylamide gel. The separated proteins were transferred to nitrocellulose, and the C/EBP␣ level was measured by immunoblot with rabbit anti-C/EBP␣ antibody (Santa Cruz, sc-61x) used at a dilution of 1:500. Binding of the primary antibody was detected using peroxidase-conjugated donkey anti-rabbit IgG (Amersham Biosciences). The bands were visualized using chemiluminescent detection methods. B, keratinocytes were infected with empty virus (EV) or adenoviruses encoding p38␦ or caMEK6. MEK6 is a known activator of p38␦ in keratinocytes (31). Total viral load was adjusted to 30 MOI/each treatment group using empty virus. At 48 h post-infection, the nuclear extracts were prepared and assayed for C/EBP␣ level by immunoblot. FIG. 5. A, novel PKC and C/EBP␣ synergistically increase hINV promoter activity. Cultured human epidermal keratinocytes were transfected with 2 g of pINV-241 in the presence of 2 g of control plasmid or plasmid encoding C/EBP␣, PKC␦, PKC⑀, or PKC. The final concentration of plasmid was adjusted to 4 g/group for all treatments using empty plasmid. After 48 h, the cells were harvested, and the extracts were prepared and assayed for luciferase activity. The inset is an immunoblot demonstrating that each expression vector produces the appropriate encoded protein. B, keratinocytes were transfected with 2 g of pINV-241 in the presence of 2 g of control plasmid or plasmid encoding GADD153, PKC␦, PKC⑀, or PKC. The final concentration of plasmid was adjusted to 4 g/group for all treatments using empty plasmid. After 48 h, the cells were harvested, and luciferase activity was monitored as outlined in A. C, keratinocytes were transfected with pINV241(C/EBPm) in the presence of C/EBP␣, PKC␦, PKC⑀, or PKC. After 48 h, luciferase activity was monitored as in A. D, to measure the effects of PKC␦ and PKC expression on C/EBP␣ binding to DNA, the cells were infected with empty adenovirus or adenovirus encoding PKC␦ or PKC. After 48 h, the cells were harvested, and the nuclear extracts were prepared. To detect C/EBP␣ binding to DNA, nuclear extract (2 g) was incubated with double-stranded, 32 P end-labeled oligonucleotide encoding the hINV promoter C/EBP site (5Ј-GGTTTGCTGCTTAAGATGCCTG, C/EBP-binding site in bold type) (15). The lane labeled Ϫ shows the migration of free probe in the absence of nuclear extract. The other lanes contain nuclear extract prepared from cells treated with the indicated adenovirus. A 100-fold molar excess of radioinert homologous oligonucleotide was included in the reaction mixture, as indicated, to demonstrate specificity of the binding. Additional controls demonstrating that the binding is due to C/EBP␣ are included in our with the findings of a previous study showing that PKC-activating agents increase p38␦ activity in HeLa cells (43). We further confirm that stimuli that enhance keratinocyte differentiation via PKC-associated mechanisms, including calcium and TPA, increase p38␦ activity. Thus, these results confirm that p38␦ is an important pro-differentiation isoform in several contexts.
In addition to the selective activation of the p38␦ isoform, nPKC isoforms also suppress ERK1/2 activity. This suggests that a reduction in ERK1/2 activity may also be essential for keratinocyte differentiation and for inhibition of proliferation. In contrast, JNK1/2 activity is not altered. Agents that regulate keratinocyte differentiation, including calcium, TPA, and okadaic acid, shared the common ability to increase p38␦ activity. However, these agents differ with respect to suppression of ERK1/2 activity, because TPA and okadaic acid inhibited ERK1/2, but calcium treatment did not. Thus, it appears that nPKCs activate a subset of the differentiation-related signaling changes. However, it is a response (i.e. p38␦ activation) that is shared by all of the differentiating agents and appears to be necessary for activation of hINV gene expression. It is possible that separate cascades regulate different aspects of the differentiation process, and additional studies will be necessary to sort this out.
Novel PKCs Increase C/EBP␣ Level and Binding to the hINV Promoter-Regulation of hINV gene expression in keratinocytes requires input from several transcriptional regulators, including the C/EBP transcription factors (14, 44 -46). The observation that C/EBP␣ binding to the hINV promoter C/EBPbinding site is increased in TPA-treated keratinocytes suggests a role for protein kinase C in this regulation (15). Our present studies directly test this possibility and show that the nPKCs have an important role. nPKC isoform expression results in increased C/EBP␣ level and DNA binding, suggesting that previous report (15). E, PKC and p38␦ enhance endogenous hINV levels. The keratinocytes were infected with the empty adenovirus (EV) or with adenoviruses encoding PKC or p38␦ at the indicated MOI. After 48 h, the total cell extracts were prepared, and involucrin expression was monitored by immunoblot as described previously (10). FIG. 6. Ras, MEKK1, MEK3, and p38 kinases are required for nPKCdependent regulation of differentiation. A, keratinocytes were transfected with 2 g of pINV-241 in the presence or absence of 1 g of plasmid encoding PKC␦ (ϩPKC␦) and or 1 g of control plasmid (ϪPKC␦). In addition, each group was transfected with control plasmid (C) or plasmid encoding dnp38, dnERK1, or dnERK2. After 48 h, the cells were harvested and assayed for luciferase activity. The total quantity of plasmid in each group was maintained constant by the addition of control plasmid. B, keratinocytes were transfected with 2 g of pINV-241 in the presence or absence of 1 g of plasmid encoding PKC␦ (ϩPKC␦) or 1 g of control plasmid (ϪPKC␦). In addition, each group was transfected with additional control plasmid (C) or plasmid encoding dnRas, dnRaf1, dnMEK1, dnMEK3, or dnJNK. After 48 h, the cells were harvested and assayed for luciferase activity. The total quantity of plasmid in each group was maintained constant by addition of control plasmid. C, pathway of C/EBP␣ factor and hINV gene regulation by nPKC, Ras, MEKK1, MEK3, and p38␦ and ERK1/2. The model indicates that stimuli that activate ERK1/2 have a negative effect on hINV gene transcription, whereas a pathway that includes nPKC, Ras, MEKK1, ME3, and p38␦ activates C/EBP␣ level and DNA binding and increases hINV gene expression. nPKCs control C/EBP␣ level/activity. Consistent with this idea, dominant-negative forms of PKC␦ and do not increase C/EBP␣ level, and the nPKC-dependent activation of hINV gene expression is inhibited by GADD153, a dominant-negative C/EBP isoform. In addition, dominant-negative kinase studies indicate that the nPKC-and C/EBP␣-dependent activation of gene expression requires Ras, MEKK1, and MEK3 activity. Based on these results, we propose the pathway of regulation shown in Fig. 6C in which a nPKC, Ras, MEKK1-dependent signal is transferred to C/EBP␣ via MEK3 and p38␦. Further evidence indicates that ERK1/2 activity is simultaneously suppressed and that this suppression is due to inhibition of activity of both ERK1 and ERK2. Inhibition of ERK1/2 activity using dominant-negative kinases results in increased hINV promoter activity, consistent with an inhibitory role for ERK. ERK1 appears to be the stronger suppressor. Thus, these findings tie nPKC activation to increased p38␦ activity and tie reduced ERK1 activity to increased C/EBP␣ binding at the hINV promoter C/EBP-binding site and to increased hINV gene activation.
p38 also regulates differentiation via activation of C/EBP factors in other cell types. In adipocytes p38 MAPK phosphorylates C/EBP-homologous protein via activation of p38␣ and ␤ (47), suggesting that CHOP serves as a link between p38 MAPK and regulation of adipose cell differentiation (48,49). In contrast, our studies clearly identify p38␦ as carrying the signal downstream to the C/EBP␣ protein. This may reflect a cell type-specific difference in the wiring of the signaling cascades and in the selection of the target C/EBP factor.
Role of C/EBP during in Vivo Keratinocyte Differentiation-C/EBP factors are expressed in a regulated manner during keratinocyte differentiation. C/EBP␣ is expressed in the suprabasal, differentiated compartments, whereas C/EBP␤ is present in the undifferentiated basal layers (50). C/EBP␤ and CHOP are in the nuclei of mid-layer cells (51). In keratinocyte cultures differentiated by suspension in semi-solid media, both C/EBP␣ and C/EBP␤ are present, whereas only C/EBP␤ is present in adherent, undifferentiated, cells (52). Thus, in general, C/EBP␣ levels are increased with increased differentiation. The present results are consistent with a model wherein C/EBP␣ enhances hINV gene expression. In contrast, C/EBP␤ and ␦ inhibits hINV expression in cultured cells (15) or does not influence hINV expression in C/EBP␤ knockout mice (38). This may, at least in part, explain why hINV is expressed in the suprabasal epidermal layers, where C/EBP␣ is expressed, and not in the basal layers, where C/EBP␤ predominates. Although our results with hINV are consistent with this model, the regulation is not simple, because other markers of keratinocyte differentiation, K1 and K10, are increased by C/EBP␤ (33,38,51), and C/EBP␤ is a negative regulator of human papillomavirus type 11 promoter transcription (53). It is possible that this difference in regulatory potential of the various C/EBP isoforms may be related to other transcription factors associated with C/EBP in the regulatory complex. Thus, whether C/EBP acts with AP1 factors, as appears to be the case for hINV, or other factors, may determine the regulatory outcome. Additional studies will be necessary to understand the role of these interactions.
In summary, our results indicate that nPKC isoforms simultaneously inhibit ERK1/2 activity and selectively activate the ␦ isoform of p38 MAPK via a cascade that includes Ras, MEKK1, and MEK3. The results are consistent with a mechanism whereby p38␦ increases C/EBP␣ transcription factor level and C/EBP␣ DNA binding to hINV promoter regulator elements to increase hINV gene expression.