Cloning and Characterization of the Human Protein Kinase C-η Promoter*

Protein kinase C-η (PKC-η) is predominantly expressed in epithelial tissue, including lung, intestine, and skin. In skin, PKC-η expression is limited to keratinocytes in the upper layers of the epidermis. To investigate regulation of cell type-specific expression of PKC-η, we cloned the 5′-segment of the PKC-η gene from a P1 genomic library. A 9.4-kilobase pair fragment encompassing the 5′-flanking region, first exon, and first intron, was localized on human chromosome 14 (14q22–23). Two major transcription initiation sites identified by reverse transcriptase polymerase chain reaction, primer extension, and S1 nuclease mapping, were located approximately 650 base pairs upstream from the translation start site. The human PKC-η proximal promoter region lacks canonical TATA and CAAT boxes and GC-rich regions. A 1.6-kilobase pair 5′-flanking region displayed maximal promoter activity. This promoter was active in human keratinocytes but not human skin fibroblasts, in accord with endogenous PKC-η gene expression. Stepwise 5′ deletion analysis revealed the presence of adjacent regulatory regions containing silencer and enhancer elements located 1821–1702 base pairs and 1259–1189 base pairs upstream of the transcription initiation site. Deletion of the proximal PKC-η promoter rendered the enhancer element inactive. Both the silencer and enhancer elements regulated heterologous promoters in keratinocytes but not fibroblasts. Electrophoretic mobility shift analysis demonstrated specific protein binding to Ets/heat shock factor and Ets/activator protein-1 consensus sequences in the enhancer and silencer regions, respectively. Mutations of the Ets/heat shock factor binding sites caused loss of functional enhancer activity. These data elucidate transcriptional regulation and tissue-specific expression of the PKC-η gene.

with the 125-bp probe, yielding 15 positive genomic clones of which two were unique, pTHQ5 and pTHQ15. To obtain additional 5Ј-upstream sequence, an EcoRI, XbaI sublibrary was generated from the P1 genomic clone, as described above, and screened with a 110-bp probe corresponding to the 5Ј-end of pTHQ15. Twenty-one additional positive clones were identified, of which two were unique, pTHQ16 and pTHQ35. The alignment of these four unique clones was determined by restriction mapping (BamHI, HindIII, EcoRI, and XbaI), PCR analysis, nucleotide sequencing, and Lasergene computer program (DNAStar Inc., Madison, WI).
DNA Sequence Analysis-All oligonucleotides were synthesized by the Biomedical Research DNA Core Facilities (University of Michigan, Ann Arbor, MI). Nucleic acid sequences of both strands were determined by automated sequencing using an applied Biosystems DNA Sequencer (model 377) or manually by the dideoxy chain termination method (15). Nucleotide sequences were confirmed by manual sequencing using a SequiTherm cycle sequencing kit (Epicentre Technologies, Madison, WI). Sequences spanning the transcription initiation site and intron-exon borders were subcloned into TA cloning vectors (pCR2.1, Invitrogen, San Diego, CA) and subjected to manual sequencing in both directions, as described above. To identify putative cis-regulatory elements, 3.6 kb of 5Ј-flanking DNA of the human PKCgene was analyzed with transcription element search software (Computational Biology and Informatics Laboratory, School of Medicine, University of Pennsylvania).
Cell Culture-Human keratinocyte HaCaT cells and primary human fibroblasts were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (Life Technologies, Inc.). Primary human keratinocytes were grown in complete MCDB 153, in a humidified incubator with a 5% CO 2 atmosphere at 37°C. Primary keratinocyte and fibroblast cultures were established from normal human skin, as described previously (5), and used at passage three or four. All culture media were supplemented with penicillin G (100 units/ml) and streptomycin (100 units/ml).
RNA Isolation and Nuclear Protein Extracts-Total RNA was purified from human skin by the guanidinium/cesium trifluoroacetic acid method (16). Total RNA from HaCaT cells, keratinocytes, and fibroblasts was extracted with a commercial kit (RNeasy kit, Qiagen, Chatsworth, CA). Poly(A) ϩ RNA was selected using an OligoTex mRNA Mini Kit (Qiagen). Nuclear extracts were prepared according to the method described by Schreiber et al. (17), with slight modifications. Briefly, HaCaT keratinocytes, human skin fibroblasts, and human keratinocytes were harvested on ice by scraping in phosphate-buffered saline. Cells were then pelleted by centrifugation (1000 ϫ g for 5 min at 4°C) and washed in 10 ml of washing buffer (1 M Tris-HCl, pH 7.5, 5 M NaCl, 1 M KCl, 1 M MgCl 2 ). Cells were washed and suspended in fresh hypotonic buffer (1 M HEPES, pH 7.9, 1 M KCl, 1 M MgCl 2 , 1 M dithiothreitol, 0.3 phenylmethylsulfonyl fluoride), incubated on ice for 10 min, and then homogenized by passing through a 25-gauge needle 10 times. Nuclei were pelleted by centrifugation for 10 min at 4°C and suspended in extraction buffer (1 M HEPES, pH 7.9, 25% glycerol, 5 M NaCl, 1 M MgCl 2 , 0.25 M EDTA, pH 8.0). The suspension was stirred for 1 h at 4°C and centrifuged at 48,000 ϫ g for 30 min at 4°C. Supernatants were stored at Ϫ80°C.
Fluorescence in Situ Hybridization Mapping-P1 clone 7239 was labeled with digoxigenin dUTP by nick translation. Labeled probe was combined with sheared human DNA and hybridized to normal metaphase chromosomes, derived from phytohaemaglutinin-stimulated peripheral blood lymphocytes in a hybridizing solution (50% formamide, 10% dextran sulfate, and 2ϫ SSC). Specific hybridization signals were detected by incubating the hybridized slide with fluoresceinated antidigoxigenin antibodies followed by counterstaining with DAPI. In separate experiments, chromosomes were cohybridized with digoxigeninlabeled P1 clone 7239 and a biotin-labeled probe specific for the centromere of chromosomes 14 and 22 (Genome Systems, St. Louis, MO). Probe detection was accomplished by incubating the hybridized slides with fluoresceinated antidigoxigenin antibodies and biotin-labeled avidin followed by counterstaining with DAPI.
Reverse Transcription Polymerase Chain Reaction-First strand cDNA was synthesized from HaCaT cell total RNA using an antisense 18-bp primer (H, 5Ј-ATCGGTGAGGCAGTGGGG) complementary to the PKC-cDNA sequence spanning positions ϩ704 to ϩ721, 45 bp downstream of the translation initiation codon. HaCaT cell total RNA (1 g) was denatured by heating at 85°C for 10 min, chilled on ice for 5 min, and then reverse-transcribed at 55°C for 30 min. PCR primers used were as follows: A, 5Ј-GAAAAAGAACCTCCCCGCCG; B, 5Ј-AG-GTCAGTCTACCACGCCTCAGGT; C, 5Ј-CACGGCCAGACCCAGCGC-TACAAG; D, 5Ј-ACGCCCCTGGGGTCCGGAAG; E, 5Ј-TCCTGGTTTG-AAGCTCGC; F, 5Ј-CCTGGAGAAGGGGCGAGT; and G, 5Ј-GAACTTC-ATGGTGCCAGACGACA. PCR amplification was performed by denaturing at 97°C for 1 min, annealing at 52°C for 1 min, and extension at 72°C for 35 cycles for 3 min, followed by incubation for 10 min at 72°C. PCR products were analyzed by 1.2% agarose gel electrophoresis and visualized with ethidium bromide. Each RT-PCR experiment contained negative controls in which reverse transcriptase was omitted. In all cases, negative control reactions did not yield detectable PCR product, indicating that PCR products were amplified from bona fide PKC-cDNA and not from contaminating genomic DNA.
Primer Extension-Primer extension assays were performed using an 18-bp antisense primer (I, 5Ј-CTCTTGCTTCTCCTCCTG) complementary to the PKC-cDNA sequence spanning positions ϩ249 to ϩ266. The primer was end-labeled with [␥-32 P]ATP (10 Ci/l, NEN Life Science Products) by T4 polynucleotide kinase (Life Technologies, Inc.). HaCaT cell total RNA (100 g) was denatured at 85°C for 10 min, chilled on ice for 5 min, and then reverse-transcribed. Reverse transcription was performed at 55°C for 30 min with 200 units of reverse transcriptase (Supertranscript II, Life Technologies, Inc.). Radioactive primer extension products were precipitated with ethanol, resuspended in gel loading buffer (90% formamide, 5 mM EDTA, 0.05% bromphenol blue, 0.05% xylencynol), boiled for 5 min, chilled on ice, and loaded on a 6% acrylamide/8 M urea sequencing gel. The gel was transferred to 3 mm Whatman paper, dried, and scanned with a PhosphorImager (Storm 860, Molecular Dynamics, Sunnyvale, CA). The 5Ј ends of primer extension products were determined by sequencing. S1 Nuclease Mapping-A sense primer (C, 5Ј-CACGGCCAGAC-CCAGCGCTACAA G-3Ј) was 5Ј phosphorylated by T4 polynucleotide kinase (Life Technologies, Inc.), and then combined with antisense primer (J, 5Ј-CCTCAGCGGGCCGGGGAA-3Ј) to generate a phosphorylated double-stranded PCR fragment. The PCR fragment was gelpurified and digested with lambda exonuclease (Amersham Pharmacia Biotech) to generate antisense single-stranded DNA. This antisense single-stranded DNA was then end-labeled, as described above. Total RNA from human skin (100 g) was hybridized with the end-labeled antisense probe (10 5 cpm) in 80% formamide, 0.4 M sodium chloride, 0.2 M PIPES, pH 6.5, at 55°C overnight. S1 nuclease (100 units, Roche Molecular Biochemicals) was then added to the digestion buffer (280 mM sodium chloride, 30 mM sodium acetate, 4.5 mM zinc acetate, and 20 g/ml of salmon sperm DNA) for 30 min at 37°C. Remaining DNA-RNA hybridized fragments were analyzed as described above for primer extension analysis.

PKC-Promoter/CAT Constructs and Site-directed Mutant
Clones-A series of 5Ј deletion fragments were cloned upstream of the promoterless pCAT3-Basic CAT reporter gene (Promega, Madison, WI). Briefly, 5Ј deletion DNA fragments were obtained by PCR using pTHQ35 as a template and a series of primers extending to the 5Ј end from position ϩ708 of the PKC-cDNA. These PCR fragments were gel purified and subcloned into a TA cloning vector (pCR2.1, Invitrogen, San Diego, CA). Purified pCR2.1 DNA was digested with KpnI and XhoI, and the excised inserts were gel purified and ligated into pCAT3-Basic. Silencer and enhancer-like fragments spanning positions Ϫ2207 to Ϫ1645 and positions Ϫ1641 to Ϫ1084, respectively, were amplified by PCR and then inserted into the pBLCAT2 and pCAT3 promoter vectors (Promega). All constructs were subjected to restriction enzyme digestion analysis and nucleotide sequencing to verify correct sequence and orientation. The PKC-enhancer region was mutated using a QuikChange site-directed mutagenesis kit (Stratagene). All mutated clones were sequenced to confirm correct sequences and orientation.
Transient Transfection and CAT Assays-Plasmid DNA was introduced into primary human keratinocytes and skin fibroblasts using LipofectAMINE-PLUS (Life Technologies, Inc.). HaCaT cells were transfected using Superfect (Qiagen). Plasmid DNA (5 g) containing the ␤-galactosidase gene (pCMV␤, CLONTECH Laboratories, Inc., Palo Alto, CA) was co-transfected with PKC-/CAT constructs (10 g) to provide an internal standard for transfection efficiency. Cells were harvested 48 h after transfection in extraction buffer (250 mM Tris, pH 7.5) and assayed for ␤-galactosidase activity (18). Aliquots containing identical ␤-galactosidase activity were used for each CAT assay (19). CAT activity was expressed as the percentage of the total chloramphenicol that was acetylated.

Cloning of the 5Ј-Upstream Region of the Human PKC-
Gene-Screening of a P1 genomic clone yielded a total of 36 positive clones, four of which were unique (pTHQ5, pTHQ15, pTHQ16, and pTHQ35) (Fig. 1A). These four unique clones were characterized by restriction enzyme mapping (BamHI, HindIII, EcoRI, and XbaI), PCR screening, and nucleotide sequencing. These four positive clones varied from 3.9 to 5.2 kb in size as determined by 1.2% agarose gel electrophoresis. Alignment of the four unique clones yielded 9.4 kb of continuous genomic sequencing, displaying the 5Ј-flanking region, exon I, and intron I (Fig. 1B).
Genomic Organization and Nucleotide Sequencing of the 5Ј-Upstream Region of the PKC-Gene-Sequencing of the four unique clones (pTHQ5, pTHQ15, PTHQ16, and pTHQ35) in both directions (Fig. 1B) yielded 4.2 kb of sequence upstream of the 5Ј end of the published PKC-cDNA (Fig. 2). This upstream region lacked canonical TATA and CAAT boxes commonly found within 100 bp upstream of transcription initiation sites. The 5Ј-upstream sequences did contain several putative cis-acting regulatory elements. These included consensus binding sites for transcription factors Sp1 (20), AP-1 (21), the Ets family (22,23), and others (Fig. 2). An exon 1/intron 1 junction was identified 363 bp downstream from the ATG translation start site, where the genomic sequence diverged from the PKC-cDNA sequence (Fig. 2). The first exon encodes 121 amino acids (Fig. 2). Further sequencing of exon 1 in the 3Ј direction yielded no cDNA sequences. It follows that the second exon was separated by at least 4.7-kb intronic sequences.
Chromosomal Localization of the PKC-Gene-Chromosomal fluorescent in situ hybridization using PKC-P1 clone 7239 as probe resulted in specific labeling of the long arm of a group D chromosome (Fig. 3). This chromosome was tentatively identified as number 14 based on DAPI staining. This identification was confirmed by double staining with the P1 clone probe and a biotin-labeled probe that was specific for the centromeres of chromosomes 14 and 22. This double staining resulted in specific labeling of the centromere of chromosome 14 in red and the long arm in green (Fig. 3). Based on ten independent measurements, the PKC-hybridization signal was located 47% of the distance from the centromere to the telomere of chromosome arm 14q, an area that corresponds to the interface between bands 14q22-23.
Identification of Transcription Start Site-As stated above, the proximal promoter region of the PKC-gene lacked sequences indicative of transcription initiation sites. For this reason, several complementary approaches were taken to determine the transcription start site. The first approach utilized RT-PCR to delineate the region of transcription initiation. First-strand cDNA was synthesized from HaCaT cell total RNA using antisense primer H (Fig. 4), which hybridized 45 bp downstream of the translation initiation site. This first strand cDNA was then used as a template for PCR amplification with six primer pairs consisting of a common antisense primer G (Fig. 4) and a series of sense primers located 334 bp to 974 bp upstream from G (Fig. 4). PCR amplification using the same six primer pairs with cloned PKC-genomic DNA as template served as positive control. PCR products of the expected sizes were observed with all six primer pairs using PKC-DNA as template (Fig. 4). RT-PCR products of the expected sizes were also observed with primer pairs G/F (334 bp), G/E (484 bp), and G/D (588 bp) (Fig. 4). In contrast, no RT-PCR products were found with primer pairs G/C (705 bp), G/B (799 bp), or G/A (974 bp) (Fig. 4), indicating that genomic sequences contained within primers A-C were not present in the PKC-cDNA. These data indicate that the 5Ј end of the PKC-transcript lies within the 93-bp region between primers D and C.
We next employed primer extension analysis to narrow down the location of the transcription initiation site. Antisense primer I (Fig. 5A), which annealed 279 bp downstream of primer C (Fig. 4), was end-labeled and annealed to HaCaT cell total RNA and extended by Superscript II reverse transcriptase. Two major primer extension products were observed, approximately 260 bp in length (Fig. 5B, lane 3). No extended products were observed with yeast tRNA as template, a control for the specificity of primer hybridization (Fig. 5B, lane 2). Sequence analysis revealed that the two alternative transcription start sites were separated by 12 nucleotides (Fig. 5C).
To confirm the primer extension results, we performed S1 nuclease mapping using a probe, generated with antisense primer J (Fig. 3A) and sense primer C (Fig. 4), that extended 26 nucleotides upstream of the distal transcription start site identified by primer extension analysis. Two major S1 nucleaseprotected fragments of 431 and 418 nucleotides were observed with HaCaT cell total RNA as template (Fig. 5D, lane 4). The sizes of these two protected fragments exactly matched the expected transcription start sites identified by primer extension. No protected fragments were observed using yeast tRNA as template (Fig. 5D, lane 3). Taken together, the above RT- PCR, primer extension, and S1 nuclease mapping data demonstrate that the transcription initiation site of the human PKCgene is located approximately 650 bp upstream from the ATG translation initiation site.
Cell-specific Regulation of the 5Ј-Upstream Region of the PKC-Gene-In human skin in vivo, PKC-is expressed in FIG. 2. Nucleotide sequence of the 5 region through first intron of the human PKC-gene. Two major transcription initiation sites are indicated by bent arrows at position ϩ1. The coding sequence of the first exon is shown as codon triplets, and the first intron is in lowercase letters. Potential cis-acting elements are boxed with their names indicated above. The 5Ј end of published cDNA is indicated by an asterisk. epidermal keratinocytes but not in dermal fibroblasts (5). RT-PCR detected PKC-transcripts in human skin, primary cultured human skin keratinocytes, and human keratinocyte Ha-CaT cells but not in human skin fibroblasts (Fig. 6A). We therefore examined whether the 5Ј-flanking region of the PKC-gene also displayed cell type-specific regulation. The full-length 5Ј-flanking sequence of the PKC-gene (Ϫ3458 to ϩ708) was inserted into the promoterless pCAT3-Basic reporter gene and transiently transfected into primary human keratinocytes, human keratinocyte HaCaT cells, and human skin fibroblasts. Transient transfection of the pCAT3-Control reporter plasmid (containing the SV40 promoter and SV40 enhancer) was used as positive control. The promoterless pCAT3-Basic reporter gene was unable to drive expression of CAT activity, whereas the pCAT3-Control reporter gene expressed high levels of CAT activity in each of the three cell types (Fig. 6B). The PKC-gene reporter construct displayed significant activity in keratinocytes and HaCaT cells (at least 10-fold higher than the promoterless CAT expression vector) but was inactive in fibroblasts (Fig. 6B). These data demonstrate that the upstream region of the PKC-gene contains regulatory elements sufficient to drive cell type-specific transcription.
Functional Analysis of the 5Ј-Flanking Region of the PKC-Gene-To identify cis-acting regulatory elements in the 5Јflanking region of the PKC-gene, we constructed a series of 5Ј deletion CAT expression vectors and transiently transfected them into primary keratinocytes and keratinocyte HaCaT cells (Fig. 7). The promoter activities of the various constructs obtained from 6 to 11 independent experiments in HaCaT cells are summarized in Fig. 7. Similar data were obtained for primary keratinocytes (data not shown). All CAT constructs were also expressed in human skin fibroblasts but were found to be inactive (data not shown).
As described above (Fig. 6B), the largest construct, a 4.1-kb fragment extending from Ϫ3458 to ϩ708 (Fig. 7, construct  83CN), displayed significant CAT activity. The activity of this construct was assigned a relative level of 1.0. 5Ј deletion of 1251 bp (Fig. 7, construct 82CN) did not significantly alter CAT activity. However, further 5Ј deletion of 566 bp (Fig. 7, construct 87CN) resulted in a 10-fold increase in CAT activity, suggesting the presence of a silencer-like element(s) between Ϫ2207 bp and Ϫ1641 bp. Additional 5Ј deletion of 442 bp (Fig.  7, construct 89CN) decreased CAT activity approximately 50%. Further 5Ј truncation of 115 bp (Fig. 7, construct 78C) reduced CAT activity to its initial relative value of 1.0, suggesting that the region between Ϫ1641 and Ϫ1084 contains an enhancerlike element(s). Further 5Ј deletion (Fig. 7, construct 78C-75C) essentially abolished CAT activity.
We tested two constructs, Ϫ1641 to Ϫ611 (Fig. 7, construct 124C) and Ϫ1199 to Ϫ611 (Fig. 7, construct 125C), to determine whether the enhancer-like region identified between Ϫ1641 and Ϫ1084 bp could function independently of the PKC-prox-

FIG. 5. Mapping of the transcription initiation site of the human PKC-gene.
A, schematic diagram of probes used for primer extension and S1 nuclease determination of PKC-transcription start site. B, primer extension analysis of PKC-mRNA. Lane 1, 32 P-labeled molecular size markers; lane 2, negative control using yeast tRNA (100 g) as template; lane 3, primer extension fragments of HaCaT total RNA (100 g) using PKC-specific primer I (see A above). Arrowheads point to two major primer extension products. C, sequencing gel of primer extension products identifying 5Ј-end of two major transcription start sites (circles). D, S1 nuclease protection assay. Lane 1, 32 P-labeled molecular size markers; lane 2, 457-bp 32 P-labeled antisense probe generated from end-labeled primer J (see A above and under "Experimental Procedures"). A 32 P-labeled probe was hybridized to yeast tRNA (100 g) (lane 3) or human skin total RNA (100 g) (lane 4) and digested with S1 nuclease, as described under "Experimental Procedures." Arrowheads indicate position of protected fragments. imal promoter. Neither of these constructs expressed CAT activity (Fig. 7), indicating that the proximal promoter was required for PKC-gene transcription.
PKC-Silencer and Enhancer Regions Regulate Heterologous Promoter Activity-Next, we further characterized the putative silencer-like region (Ϫ2207 to Ϫ1641) and enhancerlike region (Ϫ1641 to Ϫ1084) in the 5Ј-flanking region of the PKC-gene. The enhancer and silencer sequences were separately cloned upstream of the minimal thymidine kinase promoter in the CAT reporter plasmid pBLCAT2, and the SV40 promoter CAT reporter plasmid pCAT3-promoter. The constructs were then transiently transfected into HaCaT cells, human keratinocytes, and human skin fibroblasts. In HaCaT cells, the silencer region reduced pBLCAT2 activity by 80% and pCAT3-promoter activity by 70% (Fig. 8). The enhancer region elevated both pBLCAT2 activity (3.5-fold) and pCAT3-promoter activity (4-fold). Similar results were obtained in cultured human keratinocytes (data not shown). In contrast, neither the silencer nor the enhancer regions had any effect on reporter gene activity in skin fibroblasts (data not shown).
Characterization of Enhancer Sequences-To more precisely identify enhancer sequences, we analyzed a series of 5Ј deletion CAT constructs within the enhancer region (Ϫ1641 to Ϫ1079) for enhancer activity in HaCaT cells. 5Ј deletion of 355 bp increased enhancer activity 30% (Fig. 9, constructs 96CNN through 130P). 5Ј deletion of an additional 27 bp reduced enhancer activity to its initial level (Fig. 9A, construct 131P). Deletion of an additional 70 bp abolished enhancer activity (Fig. 9A, construct 101P). These data indicate that enhancer sequences lie between Ϫ1259 and Ϫ1189 of the PKC-distal promoter.
Computer analysis of enhancer sequences identified a GArich region that contained three GGA repeats that constituted overlapping potential binding sites for Ets (Ϫ1233/Ϫ1226) and HSF (Ϫ1243/Ϫ1220) transcription factors. EMSA, using a probe spanning the GA-rich region (Ϫ1244/Ϫ1221) revealed specific protein-DNA complexes with HaCaT cell nuclear extract (Fig. 9B, lanes 1-4). In addition, competition experiments with excess unlabeled probes containing consensus binding sites for Ets or HSF transcription factors effectively reduced binding of HaCaT cell nuclear proteins to the PKC-GA-rich region (Ϫ1244/Ϫ1221) probe (data not shown). Incubation of nuclear extracts with antibody to Ets 1/2 (Fig. 9B, lanes 5 and  6) or HSF-1 (Fig. 9B, lanes 7 and 8) yielded weak supershifted complexes.
To test whether the three GGA repeats contribute to enhancer activity, each GGA within the enhancer (Ϫ1286 to Ϫ1079) was mutated, and the effect of mutation on enhancer activity was determined in HaCaT cells. Mutation of two or three GGA sequences reduced enhancer activity 75%, whereas enhancer activity was reduced 20% with mutation of any one of the three GGA sequences (Fig. 9C). Mutations of GGA sequences revealed a similar pattern in EMSA. Mutation of one GGA sequence had minimal effect on formation of DNA-protein complexes, whereas mutation of two or three GGA sequences substantially reduced the intensity of specific retarded complexes (data not shown).
Characterization of Silencer Sequences-To more precisely identify sequences within the silencer region (Ϫ2207 to Ϫ1641),  (FB). B, CAT reporter activity of the human PKCpromoter. The PKC-5Ј-flanking region (Ϫ3458 to ϩ708) was subcloned into the promoterless CAT construct (pCAT3-Basic). This promoter/CAT construct and a ␤-galactosidase expression vector (pCMV␤) were co-transfected into primary human keratinocytes (KC), human keratinocyte HaCaT cells (HaCaT), and primary human skin fibroblasts (FB). Aliquots corresponding to identical ␤-galactosidase activity were used to measure CAT activity. Positive (ϩ, pCAT3-CTRL) and negative (Ϫ, pCAT3-Basic) control (Ctrl) CAT reporter constructs were transfected into each cell type. A representative CAT assay autoradiogram is shown.

FIG. 7. Deletion analysis of the 5-flanking region of the human PKC-gene.
Varying lengths of the 5Ј-flanking region of the PKCgene were amplified by PCR using genomic clone pTHQ35 as a template and then inserted into a promoterless CAT construct (pCAT3-Basic). These constructs were transiently co-transfected with a ␤-galactosidase expression plasmid (pCMV␤) into HaCaT cells. Aliquots corresponding to identical ␤-galactosidase activity were used for each CAT assay. CAT activity of each construct was expressed relative to that of the fulllength PKC-promoter/CAT vector (83CN, Ϫ3458 to ϩ708), which was assigned a relative activity of 1. Silencer (Ϫ2207 to Ϫ1641) and enhancer (Ϫ1641 to Ϫ1084) regions that were identified in the PKC-promoter (see Fig. 7) were inserted upstream of the thymidine kinase (TK) or SV40 promoters in CAT reporter plasmids pBLCAT2 and pCAT3-Promoter, respectively. These CAT reporter constructs were then transiently transfected into HaCaT cells, as described under "Experimental Procedures." Results are the means Ϯ S.D. of three experiments. A representative CAT assay autoradiogram is shown.
we analyzed a series of 5Ј deletion CAT constructs for silencer activity in HaCaT cells. The silencer region inhibited promoter activity approximately 50% (Fig. 10A, construct 114B). Deletion of 262 bp did not alter silencer activity (Fig. 10A, constructs 116B and 118B). 5Ј deletion of an additional 124 bp resulted in complete inhibition of thymidine kinase promoter activity (Fig. 10A, construct 120B). Further 5Ј deletion of 119 bp abolished all silencer activity (Fig. 10A, construct 473). These data indicate that silencer region sequences lie between Ϫ1821 and Ϫ1702 of the PKC-distal promoter.
Computer analysis of the silencer sequences (Ϫ1821 to Ϫ1702) identified AP-1 (-1740 to -1734) and Ets (Ϫ1757 to Ϫ1754 and Ϫ1715 to Ϫ1712) consensus binding sites. EMSA, using a probe containing the AP-1 and Ets elements (Ϫ1772 to Ϫ1703), revealed specific protein-DNA complexes with HaCaT cell nuclear extract (Fig. 10B, lanes 1-4). However, incubation of nuclear extracts with antibody to c-Jun, c-Fos, or Ets 1/2 individually or in combination (Fig. 10B, lanes 5-10) did not reveal any supershifted complexes. In addition, excess unlabeled probes containing consensus binding sequences for AP-1 (c-Jun/c-Fos) or Ets did not compete for binding of HaCaT nuclear extract proteins to the silencer probe (data not shown). These data suggest that neither AP-1 nor Ets proteins bind to the PKC-silencer element. DISCUSSION PKC-expression is highly tissue-and cell type-specific: it is expressed predominantly in epithelial tissues such as skin, Open boxes with ϫ represent mutation of the GGA motif to TTC. The indicated constructs were transiently co-transfected with ␤-galactosidase expression plasmids into HaCaT cells. Aliquots were normalized for ␤-galactosidase activity and assayed for CAT activity. The CAT activity of each construct was expressed relative to the wild-type enhancer CAT that was assigned a value of 100. Data are the means Ϯ S.D. of three experiments. lung, and intestine (5,8,11,24). In skin, PKC-gene expression is restricted to keratinocytes undergoing terminal differentiation (8,9,13), suggesting that PKC-participates in regulation of skin barrier formation. Characterization of the PKC-promoter is essential for understanding cell type-specific and differentiation-related expression of PKC-. In this study, we performed molecular cloning, chromosomal localization, and functional characterization of the upstream regulatory region of the human PKC-gene. A 9.4-kb fragment of genomic DNA was isolated, encompassing the 5Ј-flanking region, the first exon, and the first intron. The first exon is encoded by 121 amino acids and is separated from the second exon by at least 4.7-kb intronic sequences. PKC-␤ and PKC-␥, the only other PKC family members whose gene sequences have been partially characterized, also have large gaps between their first and second exons (25,26). The PKCgene contains two major transcription initiation sites, which are located approximately 650 bp upstream from the ATG translation initiation site. The 3.6-kb 5Ј-flanking region of the PKCgene lacked canonical TATA and CAAT boxes adjacent to the transcription start sites. Both PKC-␥ and PKC-␤ also lack TATA and CAAT boxes (25,26). These features are consistent with identification of multiple transcription initiation sites in a TATA-less promoter region. Many genes that lack TATA and CAAT boxes at the usual positions have multiple transcription initiation sites (27)(28)(29)(30). TATA and CAAT elements, in reverse order, were found further upstream at Ϫ3448, Ϫ3428, and Ϫ2677, respectively. However, we could not detect any transcription in the vicinity of these TATA or CAAT boxes employing either primer extension, S1 nuclease protection, or RT-PCR. Moreover, no significant transcriptional activity in CAT constructs was observed in this region, suggesting that the TATA and CAAT elements were not functional. The proximal promoter region for PKC-is not GC-rich. However, the 5Јuntranslated region is extremely rich in GC (80%).
To examine the functionality of the PKC-promoter and to test the tissue specificity of PKC-gene expression, we introduced a series of PKC-promoter/CAT reporter gene constructs into human keratinocyte HaCaT cells, primary human skin keratinocytes, and human skin fibroblasts. Our results clearly demonstrate that the full-length (3.6 kb) 5Ј-flanking sequence of the PKC-gene contains functional promoter and cis-elements that are capable of conferring keratinocyte-specific expression of the PKC-gene. The promoter CAT constructs were not active in skin fibroblasts, consistent with a lack of endogenous PKC-mRNA expression in these cells.
Deletion analysis of PKC-promoter/CAT constructs revealed the presence of silencer and enhancer elements in the distal 5Ј-upstream region. The silencer region repressed (80%), and the enhancer region stimulated (4-fold) basal activity of two heterologous promoters when transfected into human keratinocyte HaCaT cells or primary human keratinocytes. In contrast, neither the silencer nor the enhancer were active in skin fibroblasts. These data suggest that the silencer and enhancer elements may play an important role in keratinocytespecific expression of the PKC-gene.
5Ј deletion analysis narrowed down enhancer sequences to 70 bp lying between Ϫ1259 and Ϫ1189. This region is GA-rich, containing overlapping consensus binding sites for Ets and HSF transcription factors. Mutation of these binding sites resulted in loss of DNA binding and enhancer activity. Mutation of two sites resulted in greater loss of DNA binding and enhancer activity than mutation of one site. These data suggest that two binding sites are required to form stable DNA complexes. Both Ets and HSF members bind DNA as dimeric or trimeric complexes, respectively (31). Unlabeled consensus Ets and HSF probes effectively competed for binding of HaCaT cell nuclear proteins to enhancer sequences. However, antibodies to Ets 1/2 and HSF-1 yielded only weak supershifted bands, suggesting that other proteins are present in the retarded complexes.
Deletion analysis narrowed down silencer activity to 119 bp between Ϫ1821 and Ϫ1702 of the PKC-promoter. These sequences completely inhibited activity of the heterologous thymidine kinase promoter. Consensus AP-1 (Ϫ1740 to Ϫ1734) and Ets (Ϫ1257 to Ϫ1254 and Ϫ1215 to Ϫ1212) transcription factor binding sites were located within the silencer, and a DNA probe containing the AP-1 and Ets elements formed specific complexes with HaCaT cell nuclear proteins. These data are consistent with binding of AP-1 and Ets family members to FIG. 10. Characterization of silencer sequences in the human PKC-promoter. A, 5Ј deletion analysis of the silencer region of the human PKC-gene. The indicated 5Ј deletion constructs were transiently co-transfected with a ␤-galactosidase expression plasmid (pCMV␤) into HaCaT cells. Aliquots were normalized to ␤-galactosidase activity and assayed for CAT activity. A representative CAT assay autoradiogram is shown. Results are the means Ϯ S.D. for three experiments. B, binding of HaCaT cell nuclear proteins to silencer sequences. Labeled probe corresponding to silencer sequences Ϫ1772/Ϫ1703 was incubated with HaCaT nuclear extract (5 g) and analyzed by EMSA. Lane 1, no excess unlabeled probe; lanes 2-4, 10-, 20-, and 50-fold excess unlabeled probe, respectively. The open triangle indicates specific retarded complexes. For antibody supershifts (lanes 5-10), nuclear extract was incubated with antibody to c-Jun, Ets 1/2, or c-Fos as indicated, prior to addition of silencer probe and EMSA. No supershifted complexes were observed. Results are representative of three experiments. silencer sequences.
3Ј deletion of sequences between the transcription initiation site and the enhancer region in the PKC-promoter abolished enhancer activity, indicating that the proximal promoter was essential for enhancer function. The proximal promoter alone (i.e. without the enhancer), however, could not drive transcription. The proximal promoter contains several potential Sp1 and Ets motifs. Both Sp1 and Ets transcription factor families are expressed in a variety of cell types (20,32), including human keratinocytes (33,34). Sp1 and Ets are known to function in conjunction with other transcription factors to positively regulate several differentiation-related genes in keratinocytes (35). For example, Sp1 cooperates with Ets-like recognition motifs to up-regulate transglutaminase type 3 gene expression in keratinocytes (34). Sp1 and Ets are also essential for expression of SPRR2A (36) and involucrin (37) genes, which are involved in keratinocyte terminal differentiation (36). Interestingly, skin expresses a novel Ets family member named Jen (38) or ESE-1 (39), whose expression coincides with that of PKC-, i.e. restricted to keratinocytes undergoing the later stages of differentiation. Sp1 and Ets members, along with other transcription factors such as AP-1 and AP-2, may coordinately regulate expression of PKC-and other genes involved in keratinocyte differentiation (35).