Identification and Characterization of a Phorbol Ester-responsive Element in the Murine 8 S -Lipoxygenase Gene*

Murine 8 S -lipoxygenase (8 S -LOX) is a 12- O -tetradeca-noylphorbol-13-acetate (TPA)-inducible lipoxygenase. That is, it is not detected in normal mouse skin, how-ever, a significant increase in expression is detected in the skin of TPA promotion-sensitive strains of mice after TPA treatment. In this study, we found TPA-induced 8 S -LOX mRNA expression is a result of increased transcription in SSIN primary keratinocytes and further in-vestigated transcriptional regulation of 8 S -LOX expression by cloning its promoter. The cloned 8 S -LOX promoter ( (cid:1) 2 kb) in which a transcription initiation site was mapped at (cid:1) 27 from the ATG has neither a TATA box nor a CCAAT box. However, the promoter was highly responsive to TPA in TPA promotion-sensitive SSIN but not in TPA promotion-resistant C57BL/6J primary keratinocytes. We then identified a Sp1 binding site located (cid:1) 77 to (cid:1) 68 from the ATG that is a TPA-responsive element (TRE) of the promoter and that Sp1, Sp2, and Sp3 proteins bind to the TRE. We also found that the binding of these proteins to the TRE was significantly increased by TPA treatment and inhibition of the binding by mithramycin A decreased TPA-induced promoter activity as well as 8

Tumor promotion is a critical step in two-stage skin carcinogenesis (1). Cells initiated with a subthreshold dose of carcinogen usually do not further develop into visible and detectable tumors without promotion. 12-O-tetradecanoylphorbol-13-acetate (TPA), 1 which is the most commonly used tumor promoter in skin carcinogenesis, has strong effects on mouse skin, causing such events as hyperplasia, inflammation, and ornithine decarboxylase induction (1)(2)(3). Substantial evidences suggest that some lipoxygenase (LOX) metabolites of arachidonic acid are involved in TPA-induced epidermal hyperplasia and tumor promotion and that LOX inhibitors can effectively inhibit skin carcinogenesis (2)(3)(4)(5)(6).
8S-Lipoxygenase is a recently cloned murine lipoxygenase that has 78% homology with human 15S-LOX-2 (13). It metabolizes arachidonic acid as its preferred substrate and produces 8S-HETE. In addition, it was shown that 8S-LOX can convert linoleic acid to 9S-HODE, although with a much lower efficiency (14). So far constitutive enzyme expression has been detected only in mouse brain, footsole, tail, forestomach, and hair follicles (12,13,15). In mouse skin, however, the level of 8S-LOX message, protein, as well as enzyme activity are highly induced by a single topical treatment of TPA (11)(12)(13). Notably, the protein expression was shown to be restricted to the postmitotic, terminally differentiated epidermal compartment, stratum granulosum (13). This characteristic of 8S-LOX protein expression implies a causal relationship between enzyme expression and terminal differentiation in keratinocytes. In fact, we have recently reported that 8S-LOX transgenic mice have a highly differentiated and keratinized epidermis along with abnormally high expression of a differentiation marker, keratin-1 (16). In addition, the ability of 8S-HETE to induce keratinocyte differentiation through peroxisome proliferatoractivated receptor (PPAR)␣ was shown in vitro (16). On the other hand, increased 8S-LOX expression was also detected in the culture of calcium-induced differentiated keratinocytes. 2 Interestingly, this prominent TPA-induced 8S-LOX expression is observed only in TPA promotion-sensitive mice (SENCAR, NMRI, and CD-1) but not in the promotion-resistant C57BL/6J mice (Refs. 14 and 17 and data not shown). The responses of SSIN and C57BL/6J mice to TPA treatment have been reported to be similar with respect to the induction of ornithine decarboxylase and to the synthesis of other arachidonic acid metabolites from cyclooxygenases or lipoxygenases pathway. However, TPA does not induce hyperplasia, edema, or oxidant generation in C57BL/6J mice (17). Considering the amount of evidence showing an association of arachidonic acid metabolites with these events in TPA treated mouse skin (3)(4)(5)18), 8S-HETE, the only arachidonic acid metabolite that is different between SSIN and C57BL/6J mice, may be critical to these events and to tumor promotion.
Despite these potentially important roles of 8S-LOX in mouse skin, the mechanism by which it is regulated by TPA has not been previously reported. In an early study, Fü rstenberger et al. (12) proposed that TPA-induced 8S-LOX enzyme activity depends on protein biosynthesis, based on the observation that treatment of mouse skin with cycloheximide before or during TPA treatment prevented an increase in 8S-LOX activity. Thereafter, Jisaka et al. (13) found that 8S-LOX protein expression was restricted to the stratum granulosum compartment and that increased expression of 8S-LOX by TPA was associated with an expansion of this compartment in TPAinduced hyperplastic skin. Given this observation, they suggested an increase in the number of cells which produce 8S-LOX is one of the mechanisms of the TPA-induced enzyme activity. However, the fact that 8S-LOX message induction occurs as quickly as 3 h after TPA treatment 3 suggests that TPA regulates 8S-LOX gene expression at the message level.
In this study, we demonstrate that 8S-LOX is transcriptionally regulated by TPA in SSIN primary keratinocytes and further studied the mechanistic basis of the regulation by cloning and characterizing its promoter. A TPA-responsive element (TRE) of the 8S-LOX promoter was mapped to a Sp1 transcription factor binding site located Ϫ77 to Ϫ68 of the promoter and Sp1, Sp2, and Sp3 were identified as the transcription factors binding to this site. Finally we showed an increased binding of these factors to the TRE by TPA treatment and propose this as a mechanism of TPA-induced 8S-LOX expression in SSIN primary keratinocytes.
Assembly of Reporter Constructs-The parent reporter vector for all 8S-LOX promoter constructs in this report was pGL2 basic-m (20), a generous gift from Dr. Andrew P. Butler (University of Texas M. D. Anderson Cancer Center, Smithville, TX). This plasmid was created by disrupting a cryptic AP1 site in the Simian Virus 40 segment downstream of the luciferase gene in pGL2 basic (Promega, Madison, WI). The mutation was made to eliminate any spurious response to TPA. Clones A and B were combined into a contiguous, non-overlapping fragment for use in a reporter construct by first subcloning them into pGL2 basic-m separately. A NotI/HindIII fragment from the pCR2.1 construct of Clone A was inserted into pGL2 basic-m digested with SmaI and HindIII to create the plasmid pGL2m-CloneA, and a NotI/ KpnI fragment from the pCR2.1 construct of Clone B was inserted into pGL2 basic-m digested with HindIII and KpnI to create pGL2m-CloneB. An EcoRV digestion fragment of pGL2m-CloneA containing all of Clone A that did not overlap with Clone B was then ligated into the EcoRV digestion fragment of pGL2m-CloneB containing both the parent vector and all of Clone B that was unique from Clone A. The segment created from Clones A and B in this new construct was then amplified by PCR with primers 8S-LOXA(Ϫ)1 (5Ј-CCTCCTGAC-CAGTTTAGCTCTCTAC-3Ј) and 8S-LOXS(Ϫ)2248 (5Ј-AACATGAGCT-GAACCAGAAC-3Ј). The 2248-bp fragment of the PCR product (Clone C) was inserted into pCR 2.1 (pCR2.1CloneC), from which it was subcloned into pGL2 basic-m digested with KpnI and HindIII. Truncated promoters for the deletion constructs were generated by PCR amplification or restriction enzyme digestion of pCR 2.1CloneC and by ligation of the desired product into pGL2 basic-m. The insert in each reporter construct was verified by automated sequencing.
For the assembly of pLuc-8S-LOX(Ϫ81/Ϫ65) construct, two complementary oligonucleotides spanning Ϫ81 to Ϫ65 of the cloned 8S-LOX promoter region to which a non-complementary protruding XhoI site is linked at each 5Ј end (5Ј-tcgaCTGATGGGCGGGGCATC-3Ј and 5Ј-tcgaGATGCCCCGCCCATCAG-3Ј) were synthesized (Integrated DNA Technologies, Coralville, IA). After a process of annealing, the resulting double-stranded oligonucleotides were ligated into the pLuc-MCS reporter vector (Stratagene, La Jolla, CA) digested with XhoI.
Rapid Amplification of cDNA Ends (5Ј-RACE)-5Ј-RACE was performed with the SMART TM RACE cDNA amplification kit (Clontech) as described by the manufacturer. First-strand cDNA was synthesized from TPA-treated mouse epidermis total RNA using 8S-LOXA483 (5Ј-GCGAGGCCAACCTTCAATGTAAGTCTTC-3Ј) as the GSP. The same GSP was used to amplify the resulting 5Ј-RACE ready-cDNA, and the desired fragment was further amplified by PCR using 8S-LOXA297 (5Ј-AGGTAGCCACTCCAGCTCGAACCAG-3Ј) as the nested GSP. The nested products were gel-purified and cloned into pCR 2.1 for automated sequencing.
Site-directed Mutagenesis-Seven nucleotides, GGGCGGG, in the Sp1 binding motif in the Ϫ121 deletion construct was mutated to TTTATTT by PCR-based site-directed mutagenesis (21). The initial overlapping fragments were generated by PCR amplification of the wild type Ϫ121 deletion construct, using the upstream outer primer 5Ј-CAACACTCAACCCTATCTCG-3Ј with the mutant primer 5Ј-CCT-CAGCGATGCAAATAAAATCAGACCAGGTTAAG-3Ј and using the downstream outer primer 5Ј-ATAGCCTTATGCAGTTGCTC-3Ј with the mutant primer 5Ј-CTTAACCTGGTCTGATTTTATTTGCATCGCTGA-GG-3Ј (IDT). The product of annealing between the initial overlapping fragments was amplified by PCR using the upstream and downstream outer primers. The final product was digested with PvuI and XbaI and ligated into corresponding sites in pGL2 basic-m. The resulting deletion construct (Ϫ121m) was sequenced to confirm the desired mutation.
Transient Transfections and Luciferase Assays-The assembled 8S-LOX promoter reporter constructs or corresponding parent vector were co-transfected with the pCMV-␤-galactosidase expression vector (Clontech) into primary keratinocytes, 24 h after plating at 1 ϫ 10 6 cells/ 35-mm dish, using FuGENE TM 6 transfection reagent (Roche Applied Science, Indianapolis, IN) as described by the manufacturer. After 16 h, transfected cells were treated for 24 h with either acetone vehicle or TPA. For one set of experiments, cells were treated with 1 M mithramycin A (MMA) (Sigma, St. Louis, MO) 1 h before acetone or TPA treatment. Luciferase activity was measured using the Luciferase Assay System (Promega), and ␤-galactosidase activity was measured using the Galacto-Light TM assay kit (Tropix, Bedford, MA). Light from either assay was detected by a luminometer (Tropix). The protein concentration of each cell lysate was quantified by the BCA protein assay (Pierce, Rockford, IL). Luciferase activity was normalized to ␤-galactosidase activity and protein concentration and then expressed as relative luciferase activity.
Electrophoretic Mobility Shift Assay-Nuclear extracts were prepared as previously described (24)  GGGCGGGGCATC-3Ј (XhoI(Ϫ81/Ϫ65)) was end-labeled with [␥-32 P]ATP by T4 polynucleotide kinase (Amersham Biosciences, Piscataway, NJ), and 15,000 cpm of the labeled probe was incubated with 2 g of nuclear extracts and 1 g of poly(dI-dC) in binding buffer (20 mM Tris-HCl, 60 mM HEPES-KOH (pH 7.9), 300 mM KCl, 60% glycerol, 2.5 mM EDTA, and 5 mM dithiothreitol) for 25 min at room temperature. The probe was then electrophoresed on a 5% non-denaturing polyacrylamide gel and viewed by autoradiography. Binding specificity was tested in parallel assays through the addition of unlabeled probe or consensus oligonucleotides for Sp1, AP1, CREB, or NF-I (Santa Cruz Biotechnology, Santa Cruz, CA) at 100-fold molar excess over the labeled probe. Supershifting was assayed by incubating 2-g nuclear extracts on ice for 30 min with 2 l of Sp1, Sp2, Sp3, or Sp4 antibody (Santa Cruz Biotechnology) before inclusion of the extracts in the binding mixture.
Northern Analysis-Total RNA was isolated from mouse whole skin or primary keratinocytes treated with acetone, TPA, or a combination of TPA and actinomycin D (Sigma) for various time points using Trireagent (Molecular Research Center, Cincinnati, OH) according to the manufacturer's protocol. 10 g of RNA was separated on a formaldehyde-containing 1% agarose gel, transferred onto nylon membrane (Micron Separation, Westboro, MA), and UV cross-linked onto the membrane with a Stratalinker (Stratagene). cDNAs for 8S-LOX, Sp1, and glyceraldehyde-3-phosphate dehydrogenase were labeled with [␣-32 P]dCTP by using a Random Primed DNA Labeling kit (Roche Applied Science) and hybridized to the blot by using the QuickHyb (Stratagene) solution. Specific bands were detected by autoradiography.
Western Analysis-Nuclear extracts prepared from SSIN primary keratinocytes were separated on a 8% SDS-PAGE and transferred onto a polyvinylidene difluoride membrane (Pierce). The blot was blocked with 5% nonfat dried milk in 0.1% Tween 20-Tris buffered saline and probed with antibodies against Sp1, Sp2, Sp3, and ␤-actin (Santa Cruz Biotechnology). After washing, the blot was then probed with horseradish peroxidase-conjugated secondary antibody. The specific bands were detected by enhanced chemiluminescence kit (Amersham Pharmacia Biotech).

Transcriptional Regulation of 8S-Lipoxygenase by TPA-
TPA-induced 8S-LOX expression in SSIN primary keratinocytes occurred quickly and with a pattern that is similar to many immediate early genes, i.e. increased message was observed by 3 h, peaked at 9 h and started to decline thereafter (Fig. 1A). We then treated cultures of the cells with acetone, TPA (30 ng/ml), or the combination of TPA (30 ng/ml) plus actinomycin D (5 g or 10 g/ml, a transcription inhibitor) for 9 h to see whether TPA induces 8S-LOX transcriptionally. 8S-LOX mRNA was readily detected after treatment with TPA but not detected after treatment with acetone or after a combined treatment of actinomycin D and TPA (Fig. 1B). This data strongly suggest that TPA-induced elevation of the 8S-LOX message in SSIN primary keratinocytes occurs through transcriptional activation.
Cloning and Characterization of the Murine 8S-LOX Promoter-To better understand the transcriptional regulation of 8S-LOX by TPA treatment, we cloned the 2248 bp of genomic sequence immediately upstream of the translation start site, using a PCR-based method. Fig. 2 shows the cloned sequence. Unlike regions that have been shown to serve as the promoter for other lipoxygenases (25)(26)(27)(28), this sequence was not GϩCrich. The mononucleotide frequencies were T (28.1%), C (25.8%), A (27.5%), and G (18.5%), and these frequencies were observed even near the downstream end of the fragment. The sequence presented in Fig. 2 has neither a TATA box nor a CCAAT box, however, this sequence contains several putative binding sites for transcription factors such as AP1 (five sites), C/EBP (two sites), GATA (three sites), NF-B (one site), Oct-1 (three sites), and Sp1 (one site).
Mapping the Transcription Start Site-The previously described 8S-LOX cDNA contains 27 bp of 5Ј-untranslated region (13), and this was confirmed in our hands by 5Ј-RACE. Performing 5Ј-RACE with a 30-bp adaptor oligonucleotide and an antisense primer that binds from 483 to 456 bp downstream of the translation start codon abundantly produced a 540-bp fragment containing only the previously reported sequence. This result was verified by 5Ј-RACE with a nested primer that binds 186 bp closer to the predicted cDNA end (data not shown).
TPA Responsiveness of the 8S-LOX Promoter-Because the transcription start site for the 8S-LOX gene appears to fall within the region we cloned, we wanted to determine whether the cloned region contains elements responsive to TPA. A reporter construct containing the entire cloned fragment (Ϫ2248) was transfected into SSIN primary keratinocytes, and the level of luciferase activity after treatment with various concentrations of TPA was measured. Basal luciferase activity was detected from the Ϫ2248 construct and the activity was clearly increased by TPA, whereas luciferase activity from a promoterless control vector was not (Fig. 3A). Induction of the reporter increased with increasing doses of TPA up to 30 ng/ml, however, it decreased thereafter, likely due to the observed toxicity of TPA to SSIN primary keratinocytes.
TPA-induced 8S-LOX mRNA expression was very strong in TPA promotion-sensitive SSIN mouse skin, whereas the expression was quite weak in TPA promotion-resistant C57BL/6J mouse skin even when higher doses of TPA were used (Fig. 3B).
To determine whether this relationship was retained in cultured keratinocytes, we transfected the Ϫ2248 construct into both SSIN and C57BL/6J primary keratinocytes and compared the level of luciferase activity after TPA treatment. Consistent with the expression profile, TPA-induced luciferase activity was also much lower in C57BL/6J primary keratinocytes compared with that in SSIN primary keratinocytes (Fig. 3B). Taken together, these data suggest that the cloned region of 8S-LOX promoter contains at least one TPA-responsive element (TRE), and it mimics, at least in part, the normal regulation of the endogenous gene.
Isolation of a TPA-responsive Region in the 8S-LOX Promoter-To locate TPA-responsive elements in the isolated re-  SSIN primary keratinocytes. A, a Northern blot of total RNA (10 g) isolated from keratinocytes treated for various time periods with acetone or TPA (30 ng/ml). B, a Northern blot of total RNA (10 g) isolated from keratinocytes treated for 9 h with acetone, TPA (30 ng/ml), or a combination of TPA (30 ng/ml) plus actinomycin D (5 or 10 g/ml). Each blot was hybridized with a radiolabeled 8S-LOX cDNA probe and thereafter re-hybridized with a radiolabeled glyceraldehyde-3-phosphate dehydrogenase cDNA probe to control for loading. gion of the 8S-LOX promoter, we inserted progressively shortened segments of the cloned promoter in front of a luciferase gene and transfected the resulting constructs into SSIN primary keratinocytes. Fig. 4 shows the structure of the deletion constructs and the corresponding luciferase activities with or without TPA treatment. Although fold induction of the luciferase activity by TPA was variable to some extent, deletions of the promoter up to Ϫ93 bp from the translation start site did not affect the TPA response of the promoter. However, deletion from Ϫ92 to Ϫ70 led to the complete loss of both basal and TPA-induced activity similar to that seen with a promoterless control vector (pGL2m). It therefore appears that the response of the 8S-LOX promoter to TPA is mediated within a Ϫ92 to Ϫ70 segment of the promoter that is also critical to basic promoter activity.
Interaction between Nuclear Proteins and the TPA-responsive Region of the 8S-LOX Promoter-Because the segment between Ϫ92 and Ϫ70 seems to be critically important to the 8S-LOX promoter function, we looked for the binding of factors within that segment (Fig. 5). The nuclear extracts prepared from TPA-treated SSIN primary keratinocyte were incubated with radiolabeled oligonucleotides spanning from Ϫ92 to Ϫ58 of the 8S-LOX promoter. This binding reaction generated three retarded protein-DNA complexes (complexes I, II, and III, lane 2).
To determine the specificity of these binding complexes, we added 100-fold molar excess of unlabeled 8S-LOX oligonucleotide (Ϫ92/Ϫ58) to the binding reaction. Complexes I and II were entirely competed away, however, complex III was not significantly affected (lane 3). Based on the putative transcription factor binding site information presented in Fig. 2, an Sp1 binding motif located between Ϫ77 and Ϫ68 of the promoter was the only predicted site in the fragment between Ϫ92 and Ϫ70. We therefore tried to determine if this putative Sp1 binding motif was responsible for producing these protein-DNA complexes. Interestingly, when nuclear extracts were incubated with radiolabeled Ϫ92/Ϫ58m probe in which the putative Sp1 binding motif was point mutated, complexes I and II were no longer detected (lane 10). Moreover, when 100-fold molar excess of the unlabeled Ϫ92/Ϫ58m probe was added to the binding reaction, it could not compete away complexes I and II (lane 4). The critical role of the putative Sp1 binding site to generate complexes I and II was further confirmed when we chased labeled Ϫ92/Ϫ58 probe with unlabeled Sp1, AP1, CREB, and NF-I consensus oligonucleotides. Only Sp1 consensus oligonucleotide competed away complexes I and II (lane 5), whereas AP1, CREB, and NF-I consensus oligonucleotides did not reduce complex formation (lanes 6 -8). These data collectively demonstrate that a specific interaction between at least FIG. 2. Murine 8S-LOX upstream genomic sequence and putative transcription factor binding sites. Presented above is the sequence of the 2248-bp genomic region immediately upstream of the murine 8S-LOX translation start site (ϩ1). A transcription start site at Ϫ27 is indicated by the bent arrow. The beginning of the 8S-LOX open reading frame is also included for reference. Transcription factor binding motifs that were identified by computer analysis are shown in boldface, underlined, and labeled. one nuclear protein and the Sp1 binding motif generated complexes I and II, whereas nonspecific binding gave rise to complex III.
Identification of a Sp1 Binding Site as a TRE in the 8S-LOX Promoter-To verify that this Sp1 binding site plays a critical role in promoter activity, we mutated the Sp1 binding site in the Ϫ121 construct and transfected the resulting mutant construct (Ϫ121m) into SSIN primary keratinocytes. In these cells, the Ϫ121 construct retained basal activity and responded to TPA as previously shown in Fig. 4. However, the Ϫ121m construct could not generate either basal or TPA-induced luciferase activity (Fig. 6A). These data distinctly prove that the Sp1 binding site encompasses a functionally essential segment of the 8S-LOX promoter.
The inability of TPA to induce luciferase activity from the Ϫ121m construct presents a strong possibility that the Sp1 binding site may also mediate the TPA responsiveness of the promoter, however, we could not further test it in the absence of basal transcription activity. We therefore generated a new reporter construct based on a TATA box containing reporter vector (pLuc-MCS). In the construct, a single copy of the 8S-LOX promoter segment (Ϫ81/Ϫ65), which includes just the Sp1 binding motif (Ϫ77/Ϫ68), was inserted in front of the TATA box in the reporter vector. We then transfected the resulting construct, pLuc-8S-LOX(Ϫ81/Ϫ65), into the SSIN primary keratinocytes and treated the cells with vehicle or TPA to see if the insertion of the Sp1 binding motif could exert TPA responsiveness on the reporter vector. As shown in Fig. 6B, basal luciferase activity from the pLuc-8S-LOX(Ϫ81/Ϫ65) construct was significantly increased compared with that of the control vector and the activity was further increased by TPA treatment (about 2-to 3-fold induction), whereas the activity from control vector was not increased by TPA treatment. This observation clearly demonstrates that the single Sp1 binding site alone can mediate TPA response of the promoter and further suggests that this Sp1 binding site is a TRE of the 8S-LOX promoter.
A Mechanism for TPA-induced 8S-LOX Gene Transcription-To understand how a single Sp1 binding site can mediate TPA responsiveness of the promoter, we first compared nuclear protein binding to the Sp1 binding site between acetone-and TPA-treated SSIN primary keratinocytes. We therefore prepared nuclear extracts from 6 h acetone-or TPA-treated cells and incubated the extracts with radiolabeled oligonucleotides spanning Ϫ81 to Ϫ65 of the 8S-LOX promoter. From this reaction, three retarded protein-DNA complexes were generated, as shown in Fig. 5, and the migration pattern of these complexes was not different between acetone-and TPA-treated cells (Fig. 7A). However, it was obvious that the formation of these binding complexes, especially complexes I and II, were clearly increased by TPA treatment (Fig. 7A). This experiment was repeated with more than five independent nuclear extract preparations, all yielding the same results.
We then tried to identify the nuclear protein(s) composing these complexes in acetone-or TPA-treated SSIN primary keratinocytes. Because the segment Ϫ81/Ϫ65 encodes the consensus binding site for Sp1, the effect of antibodies against Sp1, Sp2, Sp3, and Sp4 upon the complexes was first tested (Fig.  7B). Incubating the Sp1 antibody with keratinocyte nuclear extracts led to the disappearance of most of complex I (lanes 3 and 7), the Sp2 antibody supershifted a portion of complex I (lanes 4 and 8), and the Sp3 antibody supershifted some of complex II (lanes 5 and 9). However, incubating the Sp4 antibody with the same extracts did not supershift any of the complexes but rather generated even stronger binding complexes than the original complexes in the absence of the antibody (compare lanes 1 and 6 with lanes 2 and 10). This phenomenon can occur with some antibodies that stabilize protein-DNA interactions (29). Because complex II was not significantly affected by Sp1, Sp2, or Sp3 antibody in our supershift assay conditions, we explored the possible existence of other transcription factor(s) in this complex. Antibodies against c-Jun, c-fos, JunB, Fra1, Fra2, Ap2, TFIIF, ets-1/ets-2, CBP, CREB, c-myc, C/EBP␣, C/EBP␤, C/EBP␦, PPAR␣, PPAR␤, PPAR␥, Smad2/3, USF1, or USF2 were incubated with the nuclear extracts, and none of the antibodies could supershift either complex I or complex II (data not shown). However, as we increased the amount of antibody against Sp1, Sp2, or Sp3 in the binding mixture, a supershifted band grew more intense and most of complex II disappeared (data not shown). These data demonstrate that complexes I and II include Sp1, Sp2, and Sp3, and that complex II, in particular, includes these factors in a stable, tightly bound state. On the other hand, we did not detect any differences in the level of Sp1, Sp2, and Sp3 protein expression between acetone-and TPA-treated cells (Fig. 7C). FIG. 3. TPA induction of the murine 8S-LOX promoter in primary keratinocytes. Shown in A is the relative luciferase activity of whole cell extracts from SSIN primary keratinocytes transiently transfected with 2 g of pGL2 basic-m vector or pGL2m-8S-LOX(Ϫ)2248 plasmid for 16 h and then treated for another 24 h with acetone or with 3, 30, 300, or 1000 ng/ml TPA. The data were obtained from a single experiment repeated two more times with similar results. Each bar displays the mean Ϯ S.D. of relative luciferase activity from triplicate wells. Shown above in B is a Northern blot of total RNA (10 g) isolated from SSIN and C57BL/6J mouse whole skin after a single, topical treatment with acetone or TPA (1 g and 4 g for SSIN and C57BL/6J mice, respectively) for 9 h. The blot was hybridized with a radiolabeled 8S-LOX cDNA probe and thereafter re-hybridized with a radiolabeled Sp1 cDNA probe to control for loading. Shown below in B is the relative luciferase activity of whole cell extracts from SSIN or C57BL/6J primary keratinocytes transiently transfected with 2 g of pGL2m-8S-LOX(Ϫ)2248 plasmid for 16 h and then treated for another 24 h with acetone or TPA (30 ng/ml). Each value is the mean Ϯ S.D. of at least three independent experiments. Taken together, these observations suggest that enhanced binding of Sp1, Sp2, and Sp3 to the Sp1 binding site between Ϫ77 and Ϫ68 within the 8S-LOX promoter mediates TPAinduced expression of the 8S-LOX message in keratinocytes.
Functionality of the Sp1 Binding Site as a TRE-Our evidence for Sp1, Sp2, and Sp3 binding to their cognate site of the 8S-LOX promoter prompted us to explore whether such binding has a functional consequence. Treating SSIN primary keratinocytes with mithramycin A (MMA), an antibiotic that has a GC base-specific binding property (30), before TPA treatment inhibited TPA-induced formation of complexes I and II (Fig. 8A,  compare lanes 2-4). Basal complex formation was also reduced by MMA treatment, although to a much lower extent (Fig. 8A,  compare lanes 2 and 5). Consistent with the result shown in Fig. 8A, MMA treatment before TPA treatment significantly decreased TPA-induced luciferase activity from the Ϫ121 construct in primary keratinocytes (Fig. 8B). Further confirming this, MMA pretreatment reduced TPA-induced 8S-LOX mRNA expression as well (Fig. 8C). These findings suggest that modulation of Sp1, Sp2, and Sp3 binding to the TRE of the 8S-LOX promoter is a mechanism for regulating 8S-LOX expression. DISCUSSION The most notable difference between murine 8S-LOX and other lipoxygenases is that the levels of 8S-LOX message, protein, and activity are very weak in normal mouse skin but are strongly increased after a single topical treatment of TPA (11)(12)(13). In NMRI mouse skin, for example, the activity of 8S-LOX is dramatically increased by TPA treatment, yet the activity of 12S-LOX, which is constitutively expressed, is not, and the activity of 5S-LOX is increased only slightly (11). Recent studies have proposed that the induction of 8S-LOX activity by TPA is a result of protein biosynthesis (12,13), however, the specific mechanism through which TPA induces 8S-LOX expression has not been previously demonstrated.
Since we found that TPA-induced expression of 8S-LOX mRNA in SSIN primary keratinocytes was completely blocked by actinomycin D, we pursued the idea that TPA induction of 8S-LOX expression occurs at the transcription level. We began by cloning the 2248-bp region immediately upstream of the 8S-LOX translation start codon. This region contains a major transcription start site 27 bp upstream from the translation start site, and when inserted into a luciferase reporter construct, it promotes luciferase activity. It thus appears to include the 8S-LOX proximal promoter region. To date, the promoter of three human lipoxygenases (5S-LOX (31), 12S-LOX (26), and 15S-LOX-1 (27)) and the promoter of one mouse lipoxygenase (5S-LOX (28)) have been cloned. Each of these promoters also has a predominant transcription start site FIG. 4. Isolation of a TPA-responsive region within the 8S-LOX promoter. Illustrated above on the left are various deletion constructs tested for TPA induction. Each construct is named according to the distance in nucleotides of its upstream end from the translation start site (ϩ1). A transcription start site (Ϫ27) is indicated by a vertical line. Two g of each construct or pGL2 basic-m, along with 0.125 g of an expression vector for ␤-galactosidase (pCMV-␤-gal; an internal control), were transfected into SSIN primary keratinocytes. After 16 h of transfection, the cells were subsequently treated with acetone or TPA (30 ng/ml) for 24 h. Shown above on the right is the relative luciferase activity from each construct in response to acetone or TPA. Luciferase activity was normalized to both ␤-galactosidase activity and protein concentration and then standardized to the normalized activity from pGL2m-8S-LOX(Ϫ)2248 after acetone treatment. Each value is the mean Ϯ S.D. of at least three independent experiments.
FIG. 5. Specific interaction between nuclear proteins and a Sp1 binding motif in the TPA-responsive region of the 8S-LOX promoter. Nuclear extracts were prepared from SSIN primary keratinocytes after treatment with TPA (30 ng/ml) for 6 h. Two micrograms of extracts were incubated with a 32 P-end-labeled oligonucleotide spanning the 8S-LOX promoter segment from Ϫ92 to Ϫ58 (Ϫ92/Ϫ58; lanes [2][3][4][5][6][7][8] or incubated with a similarly labeled oligonucleotide covering the same segment but containing TTTATTT in place of the Sp1 binding motif, GGGCGGG (Ϫ92/Ϫ58m; lane 10). Binding specificity was confirmed by chasing labeled Ϫ92/Ϫ58 with a 100-fold molar excess of unlabeled Ϫ92/Ϫ58 (lane 3), Ϫ92/Ϫ58m (lane 4), or consensus oligonucleotides for Sp1, AP1, CREB, and NF-I (lanes [5][6][7][8]. Protein-DNA complexes were resolved on a 5% non-denaturing polyacrylamide gel, and the positions of three protein-DNA complexes (Complexes I, II, and III) were noted. Labeled probe in the absence of nuclear extract migrated as shown in lanes 1 and 9. within 100 bp of the translation start site, and, like the cloned region from 8S-LOX, each lacks TATA and CCAAT boxes.
The 8S-LOX promoter not only drives activity of a reporter gene but also displays strong induction by TPA. From deletion and mutation analyses of the 8S-LOX promoter in SSIN primary keratinocytes, we identified the Sp1 binding motif between positions Ϫ77 and Ϫ68 of the promoter as being critical for basal and TPA-induced transcription. We then presented strong evidence that this site is a functional Sp1, Sp2, and Sp3 binding element.
At least one Sp1 binding site has been identified in the promoters of other lipoxygenases, and Sp1 has been shown to play a critical role in the basal or induced expression of such promoters in various cell types. The mouse 5S-LOX promoter has one Sp1 site at Ϫ189 to Ϫ184, and the human 5S-LOX promoter has five Sp1 binding sites located at Ϫ179 to Ϫ145 from ATG (25). Disrupting the single Sp1 binding site in the mouse 5S-LOX promoter dramatically reduces basal activity from a luciferase reporter plasmid transfected into mouse monocyte-macrophage cells (28). Adding or deleting Sp1 sites within the human 5S-LOX promoter greatly affects transcription from a CAT reporter in Schneider cells (32), and the Sp1 binding sites in the human promoter are important regulatory regions for TPA-induced expression (25). Notably, mutations in the Sp1 sites in the human 5S-LOX promoter are related to development of asthma (32) and breast cancer (33). Sp1 binding sites are also important in the human 12S-LOX promoter. That promoter has five Sp1 binding sites, and two of them (located at Ϫ158 to Ϫ150 and Ϫ123 to Ϫ114) are essential for basal and epidermal growth factor (EGF)-induced transcription in human epidermal carcinoma A431 cells (34).
If Sp1, Sp2, and Sp3 participate in the induction of 8S-LOX transcription by TPA, it will obviously be important to understand how this occurs. TPA treatment has been observed to increase the mRNA and protein expression of Sp1, as well as to enhance the binding of Sp1, in Chinese hamster ovary cells (35) and chronic myelogenous leukemia cells (36). So, we examined whether Sp1, Sp2, or Sp3 might participate in TPA induction of the 8S-LOX promoter through one of these mechanisms. In fact, we found a significant increase of these proteins binding to FIG. 6. Identification of a Sp1 binding site as a TRE in the 8S-LOX promoter. Shown in A are relative luciferase activities from deletion constructs and the pGL2 basic-m in response to acetone or TPA. The seven nucleotides of Sp1 binding motif (GGGCGGG) in the Ϫ121 construct were mutated to TTTATTT in the Ϫ121m. Shown in B is a structure of the pLuc-8S-LOX(Ϫ81/Ϫ65) construct and the relative luciferase activities from the construct and from its parent vector, pLuc-MCS, in response to acetone or TPA. A 8S-LOX promoter segment between Ϫ81 and Ϫ65 inserted in front of the TATA box (open box) in the pLuc-MCS vector is illustrated as a hatched box, and the Sp1 binding motif located Ϫ77 to Ϫ68 is depicted as a solid box. The relative luciferase activities in A and B were obtained from SSIN primary keratinocytes transfected with 2 g of each construct along with its parent vector for 16 h and thereafter treated with acetone or TPA (30 ng/ml) for another 24 h. The data are the mean Ϯ S.D. of at least three independent experiments.
the Sp1 binding site of the 8S-LOX promoter in the SSIN primary keratinocytes after TPA treatment. We did not measure a significant change in the endogenous level of Sp1, Sp2, or Sp3 protein, however, when keratinocytes were treated with TPA. Therefore, it appears likely that TPA transduces its effect on 8S-LOX expression through increased binding of Sp1, Sp2, and Sp3 to the Sp1 binding site of the promoter rather than through increased expression of these proteins in SSIN primary keratinocytes. The data, which show TPA-induced 8S-LOX gene transcription is inhibited by treating cells with Sp1 binding inhibitor, MMA, further support this conclusion. However, this is not in agreement with related observations of others. Binding of Sp1 to the human 5S-LOX promoter in HL-60 cells and binding of Sp1 to the rat ornithine decarboxylase promoter in Reuber H35 rat hepatoma cells are both unaffected by TPA treatment, although TPA activates both promoters (25,37). Sp1 binding to the 12S-LOX promoter in human epidermal carcinoma A431 cells is also unchanged after EGF treatment, although EGF induces 12S-LOX promoter activity through Sp1 sites (34). Thus, it appears that an induction of gene transcription by Sp1 occurs in a gene-specific as well as cell-type-specific manner through different mechanisms.
The functional consequences of elevated 8S-LOX expression by TPA in mouse skin are still not known. Considering previous studies, which suggest 8S-LOX is associated with differentiation (13, 16), however, our finding of transcriptional activation of 8S-LOX by TPA provides a deeper insight into a function of this gene, at least in part, in the process of keratinocyte differentiation. That is, because the protein expression was found prominently in differentiated keratinocytes, it has been an unresolved question as to whether 8S-LOX induces keratinocyte differentiation or differentiated keratinocytes produce 8S-LOX. However, the fact that 8S-LOX is transcriptionally activated by TPA in a proliferating basal keratinocyte population indicates a more active participation of this gene in the process of keratinocyte differentiation. Interestingly, forced overexpression of 8S-LOX in C57BL/6J mice caused a highly differentiated as well as thinner epidermis and, moreover, resulted in fewer tumors than in wild type mice in a two-stage skin carcinogenesis protocol. 3 Considering that the failure to fully differentiate is a defining characteristic of malignant cells, it is possible that the 8S-LOX gene could be a novel target for skin cancer prevention by modulating its expression.
In summary, our results demonstrate that TPA regulates 8S-LOX expression at the transcriptional level through an increased Sp1, Sp2, and Sp3 binding to the Sp1 binding site in the 8S-LOX promoter. However, this finding raises another question of how TPA alters DNA binding ability of those factors. Considering recent reports showing that protein kinase C (PKC, a cellular receptor for TPA)-mediated Sp1 phosphorylation increases Sp1 binding to a Sp1 binding site (38 -40), it is possible that phosphorylation of Sp1, Sp2, and Sp3 following FIG. 7. Increased binding of Sp1, Sp2, and Sp3 to the TRE of the 8S-LOX promoter after TPA treatment. A, an EMSA showing an effect of TPA on protein-DNA complex formation. Two independent nuclear extract preparations were made from SSIN primary keratinocytes after treatment with acetone (lanes 2 and 4) or TPA (30 ng/ml; lanes 3 and 5) for 6 h. Two micrograms of extracts was incubated with a 32 P-end-labeled oligonucleotide spanning the 8S-LOX promoter segment from Ϫ81 to Ϫ65. Labeled probe in the absence of nuclear extract migrated as shown in lane 1. In B, the proteins complexed to the labeled probes were identified by preincubating 2 g of nuclear extracts with 2 l of anti-Sp1 (lanes 3 and 7), -Sp2 (lanes 4 and 8), -Sp3 (lanes 5 and 9), or -Sp4 (lanes 6 and 10) antibody before addition to the binding reaction. Protein-DNA complexes were resolved on a 5% non-denaturing polyacrylamide gel. The positions of three protein-DNA complexes (Complexes I, II, and III) are noted, and the positions of supershifted bands are indicated by one (*) or two asterisks (**). Shown in C is a Western blot of nuclear proteins (20 g) prepared from SSIN primary keratinocytes after treatment with acetone or TPA (30 ng/ml) for various time periods. The blot was hybridized with an antibody against Sp1, Sp2, or Sp3 and thereafter re-hybridized with a ␤-actin antibody to control for loading. the activation of PKC may enhance their DNA binding activity to the 8S-LOX promoter. On the other hand, we can not exclude a possibility of PKC-independent post-translational activation of Sp1. Torgeman et al. (41) reported that TPA-stimulated Sp1 DNA binding activity was not diminished by a PKC-specific inhibitor. Furthermore, they showed the Sp1 binding stimulation was mediated by formation of a Sp1-p53 protein complex following TPA treatment (42). This observation suggests TPA may modulate DNA binding activity of Sp1 by regulating its interaction with other transcription factors or cofactors. Therefore, pursuit of this question will require extensive work in the future and may well provide significant insight into how TPA promotes tumorigenesis on a global level.