Protein Kinase C δ Increases Kruppel-like Factor 4 Protein, which Drives Involucrin Gene Transcription in Differentiating Keratinocytes*

Background: KLF4 is an important regulator of terminal differentiation and barrier formation in epidermis, but its mechanism of action is not well understood. Results: PKCδ increases KLF4, which increases hINV expression by interaction at specific elements in the hINV promoter. Conclusion: KLF4 drives expression of involucrin in response to differentiation stimuli. Significance: This study provides detailed information regarding the mechanism of KLF4 regulation of keratinocyte differentiation. KLF4 is a member of the Kruppel-like factor family of transcriptional regulators. KLF4 has been shown to be required for normal terminal differentiation of keratinocytes, but the molecular mechanism whereby KLF4 regulates genes associated with the differentiation process has not been studied. In the present study, we explore the impact of KLF4 on expression of involucrin, a gene that is specifically expressed in differentiated keratinocytes. KLF4 overexpression and knockdown studies show that involucrin mRNA and protein level correlates directly with KLF4 level. Moreover, studies of mutant KLF4 proteins indicate that transcriptionally inactive forms do not increase involucrin expression. PKCδ is a regulator of keratinocyte differentiation that increases expression of differentiation-associated target genes, including involucrin. Overexpression of KLF4 augments the PKCδ-dependent increase in involucrin expression, whereas KLF4 knockdown attenuates this response. The KLF4 induction of human involucrin (hINV) promoter activity is mediated via KLF4 binding to a GC-rich element located in the hINV promoter distal regulatory region, a region of the promoter required for in vivo involucrin expression. Mutation of the GC-rich element, an adjacent AP1 factor binding site, or both sites severely attenuates the response. Moreover, loss of KLF4 in an epidermal equivalent model of differentiation results in loss of hINV expression. These studies suggest that KLF4 is part of a multiprotein complex that interacts that the hINV promoter distal regulatory region to drive differentiation-dependent hINV gene expression in epidermis.

KLF4 is a member of the Kruppel-like factor family of transcriptional regulators. KLF4 has been shown to be required for normal terminal differentiation of keratinocytes, but the molecular mechanism whereby KLF4 regulates genes associated with the differentiation process has not been studied. In the present study, we explore the impact of KLF4 on expression of involucrin, a gene that is specifically expressed in differentiated keratinocytes. KLF4 overexpression and knockdown studies show that involucrin mRNA and protein level correlates directly with KLF4 level. Moreover, studies of mutant KLF4 proteins indicate that transcriptionally inactive forms do not increase involucrin expression. PKC␦ is a regulator of keratinocyte differentiation that increases expression of differentiation-associated target genes, including involucrin. Overexpression of KLF4 augments the PKC␦-dependent increase in involucrin expression, whereas KLF4 knockdown attenuates this response. The KLF4 induction of human involucrin (hINV) promoter activity is mediated via KLF4 binding to a GC-rich element located in the hINV promoter distal regulatory region, a region of the promoter required for in vivo involucrin expression. Mutation of the GC-rich element, an adjacent AP1 factor binding site, or both sites severely attenuates the response. Moreover, loss of KLF4 in an epidermal equivalent model of differentiation results in loss of hINV expression. These studies suggest that KLF4 is part of a multiprotein complex that interacts that the hINV promoter distal regulatory region to drive differentiation-dependent hINV gene expression in epidermis. KLF4 2 is a member of the Kruppel-like factor transcription factor family that recognizes GC-rich DNA enhancer elements (1). KLF4, like other members of the Kruppel-like transcription factor family, encodes a highly conserved C-terminal zinc finger DNA binding domain that shares 65% sequence identity with other Kruppel-like proteins (1). In contrast, the N-terminal region is variable and contains a nuclear localization signal and transcriptional activator and repressor domains (1). KLF4 is expressed in differentiated, nonproliferating cells, where it serves to suppress proliferation. Gene expression profiling experiments show that KLF4 suppresses expression of proliferation-associated genes and activates expression of differentiation-associated genes (2)(3)(4)(5). KLF4 has an important regulatory role in epithelia. It is expressed in suprabasal layers in the epidermis, tongue, palate, esophagus, stomach, cornea, and colon (6 -8), and onset of expression is associated with differentiation (6,7). The impact of KLF4 overexpression and knockdown has been studied in mouse epidermis. Keratin 5 promoter-targeted overexpression of KLF4 in mouse epidermis stimulates premature barrier formation (9 -11), whereas KLF4 knock-out mice display impaired barrier function (12). This is consistent with a role for this protein in suppressing proliferation and driving differentiation. Indeed, these findings are consistent with our recent study indicating that KLF4 increases p21 Cip1 expression in normal keratinocytes (13).
KLF4 has also been studied at the molecular level. KLF4 activates expression of cyclin-dependent kinase inhibitors (2,13,14) and cooperates with p53 to increase p21 Cip1 gene expression (15). Moreover, KLF4 binds to GC-rich elements in the p53 promoter to suppress transcription (16). KLF4 also binds to the cyclin B2 promoter and represses the cyclin B2 expression (17). We have recently shown that a key controller of keratinocyte differentiation, PKC␦, increases p21 Cip1 mRNA and protein level via a KLF4-dependent mechanism (13). PKC␦ stimulates increased KLF4 binding to the six GC-rich elements in the proximal p21 Cip1 promoter, and mutation of these sites results in loss of the response (13). Thus, KLF4 participates as a regulator to inhibit keratinocyte proliferation.
However, less is known about the molecular details governing KLF4 regulation of genes associated with epidermal differ-entiation. The epidermis is a stratifying epithelium in which the basal layers contain stem cells that undergo intermittent proliferation. These cells then give rise to cells that differentiate to form the suprabasal differentiated layers and that express genes associated with terminal differentiation (18 -20). Information available in the literature suggests that genes associated with differentiation should be increased by KLF4 (3,12,21); however, this has not been studied at the molecular level. Involucrin is a model for the study of gene expression during keratinocyte differentiation (22)(23)(24)(25)(26). It is expressed in the suprabasal epidermal layers, but is absent in basal cells (22,27,28). Activation of hINV transcription relies on a variety of mechanisms, but a prominent mechanism is PKC␦ activation (13, 29 -32). We have demonstrated that PKC␦ activity drives hINV expression via activation of a MEKK1, MEK3/MEK6, and p38␦/ERK signaling cascade that elevates Sp1 and AP1 transcription factor level and binding to the hINV promoter to activate transcription (22, 24 -26, 29, 33-38). 12-O-Tetradecanoylphorbol-13-acetate (TPA), a diacylglycerol analog that activates PKC isoforms, is a pharmacologic agent that activates keratinocyte differentiation via this pathway (23,24,39), as do some other naturally occurring agents (40).
Our previous study showed that PKC␦ activation increases KLF4 mRNA and protein level in normal human keratinocytes (13). In the present study, we extend these studies and examine the effect of the elevation of KLF4 on expression of hINV, an important marker of keratinocyte differentiation. Our results show that KLF4 interacts via a GC-rich element in the hINV promoter distal regulatory region (DRR) to drive transcription. Thus, these findings present the first detailed evidence that KLF4 interacts with response elements on a keratinocyte differentiation-expressed gene to drive transcription.

EXPERIMENTAL PROCEDURES
Chemicals and Reagents-Keratinocyte serum-free medium (KSFM) and trypsin were purchased from Invitrogen. Sodium butyrate was from Calbiochem, and actinomycin D was purchased from EMD Chemicals (Gibbstown, NJ). Dimethyl sulfoxide and phorbol ester (TPA) were purchased from Sigma-Aldrich. Rabbit polyclonal antibody for human involucrin (hINV) was produced in our laboratory (41), and rabbit polyclonal antibodies for KLF4 (sc-20691) and PKC␦ (sc-937) were from Santa Cruz Biotechnology (Santa Cruz, CA). Mouse monoclonal antibody for ␤-actin (A5441) was purchased from Sigma-Aldrich. Normal rabbit IgG (sc-2027) was purchased from Santa Cruz Biotechnology. Peroxidase-conjugated antimouse IgG (NXA931) and peroxidase-conjugated anti-rabbit IgG (NA934V) were obtained from GE Healthcare (Buckinghamshire, UK). We present results obtained using the Santa Cruz siRNA (control "scrambled," sc-37007; KLF4, sc-35480), but identical results were observed with KLF4 siRNA directed to other sites in the KLF4 mRNA (Dharmacon, Lafayette, CO, J-005089-10 and J-005089-09). Thus, the observed responses are not due off-site actions of the siRNA.
Promoter Activity-For hINV promoter activity analysis, 2 g of hINV promoter reporter plasmid was mixed with 4 l of FuGENE 6 reagent diluted with 96 l of KSFM. The mixture was incubated at 25°C for 15 min and then added to a 50% confluent culture of primary human epidermal keratinocyte maintained in 2 ml of KSFM in a 9.6-cm 2 dish. For co-transfection experiments, 2 g of p21 Cip1 promoter reporter plasmid and 1 g of hKLF4 expression plasmid were mixed, treated with FuGENE 6, and added to cells as indicated above. After 24 h, the cells were harvested, and extracts were prepared for assay of luciferase activity. Electroporation and siRNA-mediated Knockdown-Keratinocytes were electroporated with siRNA or plasmids using the Amaxa electroporator and the VPD-1002 nucleofection kit. For electroporation, keratinocytes were harvested with trypsin and replated 1 day prior to the day of electroporation. On the day of electroporation, 1 ϫ 10 6 of the replated cells were harvested with trypsin and resuspended in KSFM. The cells are collected at 2000 rpm, washed with 1 ml of sterile phosphate-buffered saline (pH 7.5), and suspended in 100 l of keratinocyte nucleofection solution. The cell suspension, which included 3 g of gene-specific siRNA, was mixed by gentle pipetting and electroporated using the T-018 settings. Warm KSFM (500 l) was added, and the suspension was transferred to a 21.3-cm 2 cell culture dish containing 3.5 ml of KSFM. When required, cells were electroporated a second time with luciferase reporter or expression plasmid. This was accomplished by harvesting the cells with trypsin and resuspending in KSFM. The cells were collected, washed with PBS, and resuspended in nucleofection solution as above. The nucleofection suspension, which included 2 g of plasmid, was electroporated using the T-018 settings. The cells were plated and maintained for various times before extracts were prepared for assay. Our electroporation method delivers nucleic acid reagents with greater than 90% efficiency.
Immunological Analysis-Equivalent amounts of protein were electrophoresed on a 4 -15% denaturing polyacrylamide gradient gel and transferred to nitrocellulose. The membranes were blocked, incubated with a specific primary antibody, washed, and exposed to an appropriate horseradish peroxidaseconjugated secondary antibody. Chemiluminescent detection was used to visualize secondary antibody binding.
KLF4 mRNA Half-life-To analyze KLF4 mRNA decay kinetics, keratinocytes were infected with 15 MOI of Ad5-EV or Ad5-PKC␦ for 24 h prior to the addition of 5 g/ml actinomycin D. At 0, 0.5, 1, 2, 3, and 4 h after actinomycin D addition, RNA was isolated (illustra RNAspin mini kit, GE Healthcare) and analyzed for KLF4 and cyclophilin A mRNA content by quantitative real-time PCR using primers described in the previous section. Values for each mRNA level at each time point are presented as the mean Ϯ S.D. derived from triplicate quantitative RT-PCR reactions of independent samples. First-order decay constants (k) were determined by nonlinear regression analysis (Prism version 3.03, GraphPad) of plots measuring the percentage of KLF4 mRNA remaining versus time of actinomycin D treatment. KLF4 mRNA decay constants are presented as the mean Ϯ S.D. of three independent time-course experiments, permitting pairwise statistical assessment using the Student's t test. Differences were considered significant if p Ͻ 0.005.
Epidermal Equivalents-Normal human keratinocytes (1.5 ϫ 10 6 ), growing in KSFM (25), were harvested with trypsin and electroporated with 3 g of siRNA. The cells were replated and expanded in KSFM for 72 h, and 2 ϫ 10 6 cells were reelectroporated with 3 g of siRNA and then allowed to settle overnight onto Millicell-PCF chambers (diameter ϭ 12 mm, 0.4-m pore size) in KSFM (Millipore, Billerica, MA). The cells were then shifted to EpiLife medium containing 1.4 mM calcium chloride and 5 g/ml of vitamin C and cultured at the air-liquid interface. Fresh medium was added every 2 days, and after 4 days, the epidermal skin equivalents were harvested for preparation of histological sections and isolation of RNA for quantitative RT-PCR analysis.

RESULTS
PKC␦ Increases KLF4 Expression-PKC␦ is an important regulator of keratinocyte differentiation (22)(23)(24)(25)(26)47), and we recently reported that PKC␦ expression increases KLF4 level in keratinocytes (13). We initiated the present studies by confirming the PKC␦-dependent increase in KLF4. Keratinocytes were infected with empty or PKC␦-expressing adenovirus, and after 24 h, cells were harvested for preparation of mRNA. Fig. 1A confirms that increasing PKC␦ level increases KLF4 mRNA and protein. The human KLF4 gene promoter is presently not available, and so to assess the mechanism responsible for this increase, we monitored the impact of increased PKC␦ on KLF4 mRNA turnover. As shown in Fig. 1B, KLF4 mRNA half-life is not significantly altered by increased PKC␦ level (t1 ⁄ 2 ϭ 0.75 h in control versus 0.87 h in PKC␦ overexpressing cells). The lack of change in turnover rate, coupled with the increase in mRNA level, suggests that PKC␦ increases KLF4 mRNA level by a transcriptional mechanism.
KLF4 Increases hINV Expression-The Kruppel-like transcription factor, KLF4, is an Sp1-related transcription factor that interacts with the GC-rich DNA elements. KLF4 is an important transcription factor that is required for epidermal barrier formation, suggesting that it has a role in differentiation (12,21). However, the role of KLF4 in regulating specific differ- entiation-responsive epidermal genes has not been examined. Involucrin is an extensively studied model of differentiationassociated gene expression in epidermis and in epidermal keratinocytes (22, 49 -53). Fig. 2 (A and B) shows that KLF4 expression increases hINV protein and mRNA level, suggesting that KLF4 may regulate hINV gene transcription. To explore this further, we monitored the impact of KLF4 expression and knockdown on hINV promoter activity. Cells were transfected with the hINV full-length promoter construct, pINV-2473 and hKLF4(1-470) (wild type); an hKLF4 mutant lacking the zinc finger domain, hKLF(1-388); and a mutant encoding only the zinc finger domain, hKLF4(335-470) (13). Treatment with wild-type KLF4 increases transcription, but treatment with the inactive mutants, hKLF4(1-388) and hKLF4(335-470), does not (Fig. 2C). hKLF4(1-470) and hKLF4(1-388) were confirmed to be expressed at similar levels by immunoblot, thereby confirming that the difference in hINV promoter activity is not due to a difference in expression of these proteins (Fig. 2C). Because anti-KLF4 binds to KLF4 within amino acids 1-180, expression of hKLF4(335-470) could not be monitored. We also assessed the impact of knockdown of endogenous KLF4 on promoter activity. As shown in Fig. 2D, reduction in KLF4 level is associated with reduced hINV promoter activity.
KLF4 Is Required for Activation of hINV Gene Expression-To examine the interplay between PKC␦ and KLF4, we examined the impact of KLF4 knockdown on PKC␦-and TPA-dependent activation of hINV gene expression. TPA is a strong inducer of keratinocyte differentiation and hINV expression (39). TPA is a diacylglycerol analog that is known to activate calcium-and phospholipid-dependent PKC isoforms (54). Fig.  3 (A and B), shows that KLF4 siRNA reduces the ability of PKC␦ and TPA to increase hINV mRNA level, and Fig. 3C shows that KLF4 siRNA reduces TPA-stimulated hINV promoter activity. Consistent with these findings, KLF4 overexpression enhances the PKC␦- (Fig. 3D) and TPA- (Fig. 3E) dependent increase in hINV promoter activity.
KLF4 Activation of hINV Expression Requires the DRR-Our next goal was to identify hINV promoter regulatory sites responsible for the KLF4-dependent increase in transcription. Previously identified functional domains, including the DRR and the proximal regulatory region (PRR), are indicated in Fig.  4A. We tested the response of a series of truncated hINV promoter-luciferase reporter constructs to challenge with KLF4 expression plasmid (37). Activity of the full-length promoter construct, pINV-2473, was increased by KLF4, but the truncated constructs did not respond (Fig. 4B). These studies suggest that a KLF4-responsive element is present in the segment spanning nucleotides Ϫ2473 to Ϫ2136. This region encodes the hINV promoter DRR (39). pINV-41 encodes the hINV minimal promoter (39) and, as expected, has low activity and does not respond to KLF4 stimulation.
The DRR encodes previously characterized AP1 and GC-rich response elements located at nucleotides Ϫ2122/Ϫ2107 (38,39). Because the KLF4-associated increase in hINV promoter activity is abolished in the absence of this segment (Fig. 4B), we determined whether the AP1-5 or the GC-rich response elements are required for the KLF4-dependent increase. Fig. 5A shows a schematic of the hINV luciferase promoter reporter plasmids used in this experiment. Fig. 5B shows that mutation of either the AP1-5 or the GC-rich response elements, or both elements, reduces KLF4-stimulated activity of the full-length promoter. This reduction is also observed (Fig. 5C) using a con- increases hINV mRNA level. Keratinocytes were infected with the indicated virus, and after 24 h, RNA was harvested for detection of KLF4 mRNA by quantitative RT-PCR. The values are the mean Ϯ S.D., n ϭ 3. The asterisk indicates a significant increase over control as determined by Student's t test, p Ͻ 0.005. Similar results were observed in each of three experiments. C, keratinocytes were transfected with 2 g of pINV-2473 luciferase reporter plasmid in the presence of 1 g of empty vector or the indicated KLF4 expression vector. At 24 h, the cells were harvested, and extracts were assayed for luciferase activity. The asterisk indicates a significant increase (n ϭ 3) as determined by Student's t test, p Ͻ 0.005. Wild-type and mutant KLF4 expression was monitored by immunoblot and normalized to the level of ␤-actin. hKLF4(335-470) could not be monitored because the antibody epitope is deleted from this mutant. D, keratinocytes were electroporated with 3 g of the indicated siRNA, and after 48 h, cells were transfected with 4 g of luciferase reporter plasmid. After an additional 24 h, extracts were prepared for luciferase activity assay and immunoblot. The values are mean Ϯ S.D., n ϭ 3. The asterisk indicates a significant reduction in luciferase activity, p Ͻ 0.005. Similar results were observed in each of three experiments. Identical results were observed using several KLF4 siRNA, indicating that the observed responses are not due to off-target effects. struct, pINV(Ϫ2473/Ϫ2088), that encodes a smaller segment of the promoter containing only the DRR region (nucleotides Ϫ2473/Ϫ2088) linked to the minimal promoter (Fig. 5A). This finding indicates that the DRR is necessary and sufficient for the response and rules out a requirement for the PRR AP1-1 site (39). This eliminates the possibility that the PRR has a role in this regulation. pINV-41 encodes the hINV basal promoter (39) and, as expected, does not respond.
We next examined KLF4 interaction at the DRR using chromatin immunoprecipitation. Cells were infected with empty or KLF4-encoding virus, and after 48 h, extracts were prepared for chromatin immunoprecipitation. Fig. 6A shows increased KLF4 interaction at the DRR (nucleotides Ϫ2218/Ϫ2055) in KLF4-expressing cells. In contrast, as a control, we examined KLF4 interaction at a DNA region (nucleotides Ϫ1040/Ϫ919) that does not encode AP1 or GC-rich elements. As shown in Fig. 6B, KLF4 does not bind to this segment.
KLF4 Knockdown Suppresses Keratinocyte Differentiation and Reduces hINV Expression-To assess the impact of KLF4 knockdown on hINV expression during differentiation, we used an epidermal equivalent model that mimics in vivo epidermal differentiation (55). Keratinocytes were electroporated with control or KLF4 siRNA and then plated into Millicell chambers to test for the ability to form a stratified epidermal equivalent. Fig. 7A confirms that KLF4 siRNA reduces KLF4 mRNA and protein level. Fig. 7B shows that that cells treated with control siRNA produce a well differentiated epidermal equivalent that includes a multiple stratified layer and a cornified layer. In contrast, KLF4 siRNA-treated cells form a disordered stratified structure in which suprabasal cells retain their Identical results were observed using several KLF4 siRNA, indicating that the observed responses are not due to off-target effects. D, KLF4 augments the PKC␦-dependent increase in hINV promoter activity. Keratinocytes were transfected with 2 g of hINV luciferase reporter plasmid, 1 g of KLF4 expression vector, and 1 g of PKC␦ expression vector. After 24 h, the cells were harvested, and extracts were assayed for luciferase activity. E, KLF4 augments the TPA-dependent increase in hINV promoter activity. Keratinocytes were transfected with 2 g of involucrin luciferase reporter plasmid in the presence or absence of 1 g of KLF4 expression vector. After 24 h, the cells were treated with or without TPA (50 ng/ml) for 24 h prior to assay for luciferase activity. In all panels, the values are the mean Ϯ S.D. (n ϭ 3), and the asterisks indicate a significant increase or decrease as determined using the Student's t test, p Ͻ 0.005. Similar results were obtained in each of three independent experiments. nuclei (arrows), showing that KLF4 knockdown impedes the differentiation process. Counting of nucleated suprabasal cells reveals 8 Ϯ 2 suprabasal nucleated cells in control siRNAtreated cultures, versus 41 Ϯ 3 in KLF4 siRNA-treated cells (n ϭ 5, p Ͻ 0.05). This is evidence of reduced differentiation. Fig. 7C shows that hINV mRNA level is reduced in these cultures, thus confirming that KLF4 is required for hINV expression under in vivo-like differentiation conditions.

DISCUSSION
Keratinocyte differentiation is a complex process that requires the coordinated activation of a variety of genes (22). Previous studies show that PKC␦ is an important driver of cell differentiation (13) and that PKC␦ stimulates a MEKK1, MEK3/MEK6, p38␦ cascade that triggers differentiation by increasing Sp1 and AP1 transcription factor level (22). In particular, the expression of involucrin, a marker of differentiation, is increased via AP1 and Sp1 factor interaction with DNA elements in the hINV promoter DRR (23)(24)(25)(26). However, it is unlikely that these transcription factors are the only one involved in regulating involucrin gene expression.
KLF4 is an important candidate regulator that has not been extensively studied at the mechanistic level in epidermis. Previous studies in mouse transgenic mouse models indicate that KLF4 regulates keratinocyte differentiation (3, 10, 12, 21). Specifically, KLF4 impacts the terminal stages in differentiation. KLF4 knock-out mice die shortly after birth due to loss of skin barrier function, which is accompanied by loss of integrity of late stage differentiation structures such as the cornified envelope (12). These defects are retained when KLF4 knock-out mouse skin is grafted onto nude mice, suggesting that the absence of KLF4 creates an intrinsic defect in the keratinocytes (12). Consistent with these findings, overexpression of KLF4 in mouse epidermis, using the keratin 5 promoter, accelerates barrier formation, which is associated with increase epidermal stratification and increased expression of cornified envelope markers (10). Although global profiling has identified potential KLF4 tar-  and GC-rich element are indicated. The distances are in nucleotides relative to the transcription start site. B and C, keratinocytes (KERn) were transfected with 2 g of the indicated reporter plasmid and 1 g of empty vector or hKLF4expression vector, and after 24 h, the cells were harvested, and extracts were prepared for luciferase activity assay. The values are mean Ϯ S.D., n ϭ 3. In all cases, the asterisk indicates a significant reduction in luciferase activity as determined using the Student's t test, p Ͻ 0.005. get genes in mouse epidermis (3), detailed molecular studies examining KLF4 target gene expression have not been performed. Involucrin is a structural protein that is specifically expressed in differentiated cells of the suprabasal epidermis and is a key precursor protein involved in keratinocyte cornified envelope formation (41, 56 -58). Given that KLF4 regulates envelope formation and that involucrin is required for appropriate envelope formation, we surmised that KLF4 may increase involucrin gene expression and that the involucrin gene may be a model for KLF4 action in epidermis.
Mechanism of KLF4 Regulation of Involucrin Expression-We previously showed that hINV expression occurs specifically in differentiated cells and involves interaction of AP1 and Sp1 transcription factors at specific sites in the hINV promoter distal regulatory region (22). This is observed in cultured cells (36,40,59), and the DRR is required to drive suprabasal expression in vivo when the human gene is placed in transgenic mice (49,50,52,53,58). Expression is activated by agents and kinases that stimulate keratinocyte differentiation including calcium, phorbol ester, and protein kinase C activity (22). The role of KLF4 in driving expression of differentiation-associated gene in epidermis has not been examined at the molecular level. In the present study, we explore the idea that KLF4 may drive expression of involucrin. We first show that KLF4 expression increases hINV mRNA and protein level, but that this increase is not observed with a transcriptionally inactive KLF4 mutant. The fact that vector-mediated KLF4 increases expression of involucrin, a marker of differentiation, is consistent with the proposed role of KLF4 in driving keratinocyte terminal differentiation (12). Moreover, knockdown of endogenous KLF4 reduces basal hINV mRNA level.
Promoter deletion experiments reveal that the KLF4 regulation requires the promoter DRR, which encodes closely juxtaposed AP1 and GC-rich (Sp1) binding sites. These sites have been implicated in regulation of hINV expression in cultured cells (38,39,60) and in transgenic mice in vivo (22, 49 -53). KLF4 increases activity of the full-length promoter and a promoter segment that encodes only the DRR and flanking sequences. The fact that KLF4 activates the latter construct indicates that the hINV promoter proximal regulatory region is not required for this regulation. This is important because the PRR does have activity in culture models (24,39). Chromatin immunoprecipitation analysis reveals that KLF4 is enriched at the DRR. It is known that KLF4 interacts at GC-rich elements to drive gene expression and that this interaction requires the KLF4 zinc finger DNA binding domain (1,61). The Sp1 site of the hINV promoter DRR is a GC-rich sequence (37,38) that would be predicted to bind KLF4 (1,61). Indeed mutation of this DNA element results in reduced KLF4-dependent hINV promoter activity. In addition to the GC-rich element, the DRR encodes an AP1 site located immediately upstream of the GC-rich element (50). This site binds AP1 transcription factors including junB, junD, and Fra-1 (39). It is interesting that mutation of this site also reduces the ability of KLF4 to drive transcription. However, this is not unexpected. The AP1 and GC-rich sites are separated by a single nucleotide, and we have previously shown that a large transcriptional complex forms over this region (37,38,62) and that mutation of either site disrupts this interaction (38,60). Thus, we propose that KLF1 forms a complex on this site with AP1 transcription factors and additional adaptor proteins to drive gene expression. These additional proteins are likely to include the p300 histone deacetylase (62). It is interesting that p300 has been reported as a co-regulator with KLF4 (1). In this context, it is plausible that mutation of either the AP1 site or the GC-rich element would reduce KLF4-dependent transcription.
PKC␦ Increases KLF4 Level-As mentioned above, PKC␦ is an important regulator of keratinocyte differentiation. Vectormediated overexpression of PKC␦ or treatment with phorbol FIGURE 6. KLF4 interacts with AP1-5/GC-rich response element in the hINV promoter DRR. A and B, keratinocytes (KERn) were infected with 15 MOI of tAd5-EV or tAd5-hKLF4. After 48 h, cells were prepared for ChIP assay. ChIP was performed as described under "Experimental Procedures" using hINV promoter-derived PCR primers encoding the indicated range of nucleotides. In all panels, the values are the mean Ϯ S.D. (n ϭ 3), and the asterisk indicates a significant increase or decrease as determined using the Student's t test, p Ͻ 0.005. FIGURE 7. KLF4 is required for hINV expression during keratinocyte differentiation. A and C, keratinocytes (KERn) were electroporated with control or KLF4 siRNA and seeded into Millicell wells to form epidermal equivalents. After 4 days, the epidermal equivalents were harvested, and extracts were prepared for assay of KLF4 and hINV level. B, epidermal equivalents were sectioned and stained with hematoxylin/eosin to assess differentiation status. Bar ϭ 100 m, m indicates the membrane, and c indicates the cornified layer. The arrows indicate nuclei that are retained in cells in the suprabasal layers of the KLF4 siRNA-treated epidermal equivalent. Identical results were obtained with three independent sets of control and KLF4 siRNA. In all cases, the values are the mean Ϯ S.D. (n ϭ 3), and the asterisks indicate a significant increase or decrease, p Ͻ 0.005. ester activates this kinase (25,31,47,63,64). Our present studies show that PKC␦ overexpression or treatment with TPA increases KLF4 level in keratinocytes. This is an interesting finding as very little is known about pathways that regulate KLF4 expression. It has been shown that ⌬N-p63␣ suppresses expression in normal epidermal keratinocytes via a mechanism that involves ⌬N-p63␣ direct interaction on the KLF4 promoter (48). ⌬N-p63␣ also regulates KLF4 gene transcription in HaCaT keratinocytes, as determined using a KLF4 promoterluciferase reporter construct (48). Our actinomycin D inhibitor studies indicate that mRNA stability is not altered in control versus PKC␦-expressing keratinocytes, suggesting that the increase in KLF4 expression is mediated via transcriptional mechanisms. Thus, our results are consistent with previously described transcriptional mechanisms for regulation of KLF4 expression (48).
Moreover, these findings are biologically cohesive as ⌬N-p63␣ expression in basal keratinocytes is designed to maintain cell viability in this layer, and so it is expected that ⌬Np63␣ would suppress expression of the pro-differentiation KLF4 protein (⌬Np63␣ is absent in the upper epidermis, and this is associated with KLF4 expression) (48). Our results show that PKC␦, a pro-differentiation regulator, increases KLF4 expression in keratinocytes to drive differentiation. Moreover, it is interesting that KLF4 is now implicated in two processes in keratinocytes: suppression of proliferation and enhancement of differentiation. Our recent study shows that PKC␦ increases p21 Cip1 expression via a mechanism that requires KLF4 (13).
KLF4 Is Required for hINV Expression in Differentiated Keratinocytes-To confirm that KLF4 is required for hINV expression, we used an epidermal equivalent system that mimics in vivo differentiation. In this system, keratinocytes are plated onto a semipermeable membrane and grown at the airliquid interface. Under these conditions, keratinocytes assemble a highly differentiated multilayered epidermal equivalent that includes a basal proliferative layer, intermediate layers, and a cornified terminal layer (55). Our studies show that knockdown of KLF4 in these cultures severely compromises the differentiation process, resulting in formation of a disordered morphology and the absence of a cornified layer. We further show that hINV expression is markedly reduced. These findings strongly suggest that KLF4 has an in vivo role in controlling hINV gene expression.
Based on these studies, we propose that PKC␦ induction of KLF4 expression is a key event in the suprabasal epidermis that drives hINV gene expression and keratinocyte terminal differentiation.