UV-induced Expression of Key Component of the Tanning Process, the POMC and MC1R Genes, Is Dependent on the p-38-activated Upstream Stimulating Factor-1 (USF-1)*

Protection against UV-mediated DNA damage and the onset of oncogenesis is afforded by the tanning response in which UV irradiation triggers melanocytes to increase production of melanin that is then transferred to keratinocytes. A key component of the tanning process is the UV-mediated induction of the pro-opiomelanocortin (POMC) and MC1R genes encoding the α-melanocyte-stimulating hormone and its receptor, respectively, which play a crucial role in pigmentation by regulating the intracellular levels of cAMP. How these genes are regulated in response to UV irradiation is not known. Here we have shown that UV-induced activation of the POMC and MC1R promoters is mediated by p38 stress-activated kinase signaling to the transcription factor, upstream stimulating factor-1 (USF-1). Importantly, melanocytes derived from USF-1 -/- mice exhibit a defective UV response and fail to activate POMC and MC1R expression in response to UV irradiation. The results define USF-1 as a critical UV-responsive activator of genes implicated in protection from solar radiation.

Solar ultraviolet (UV) 1 radiation (1) is a major environmental hazard that can generate reactive oxygen species, inducing DNA damage and protein oxidation (2) that subsequently lead to skin inflammation, photo-aging, and skin cancer. The incidence of melanoma, the most dangerous form of skin cancer, is increasing at an alarming rate, with metastatic melanoma being notoriously refractive to treatment.
Protection against UV-mediated DNA damage is afforded by the tanning response in which UV irradiation triggers melanocyte production of melanin. Melanin synthesis takes place in specific organelles, the melanosomes, which are transferred to keratinocytes (3), the neighboring cells generating a protective skin screen. Two key upstream components of the melanin cascade process are the POMC and MC1R genes (4) encoding the ␣-melanocyte-specific hormone (␣-MSH) after cleavage of the POMC (5-8) and its heterotrimeric G-protein receptor, respectively. The MC1R gene is conserved among species and displays a large number of different alleles (9,10). Among them, human genetic variants contribute to the existence of the six different phototypes (I-VI), ranging from white skin and red hair to dark skin and black hair (11)(12)(13), that are associated with a different incidence of skin cancer (14,15). Binding of ␣-MSH to MC1R regulates the intracellular level of cAMP (16,17), which is involved in microphthalmia gene expression (18). The microphthalmia-associated basic helix-loop-helix-leucine zipper (b-HLH-LZ) transcription factor (19,20) is required for the expression of the Tyrosinase (21), TRP-1 (tyrosine-related protein), and Dct (dopachrome tautomerase) genes (17) that encode enzymes implicated in the manufacture of the pigment melanin. Tyrosinase encodes the rate-limiting enzyme for the production of melanin and is absolutely necessary for pigmentation and solar protection.
POMC and MC1R gene expression are UV-inducible (5,22). However, no molecular mechanism has yet been proposed to explain how UV irradiation might activate their expression. The cellular response to stress-inducing agents such as UV irradiation is known to be mediated by the activation of specific kinase cascades, the stress-activated protein kinase families, which include the c-Jun NH 2 -terminal kinase and p38 signaling pathways. The microphthalmia-associated transcription factor (Mitf) and the upstream stimulating factor-1 (USF-1) (23,24) are members of the evolutionary conserved family of b-HLH-LZ; significantly, both are downstream targets of the p38 stress-activated kinase (25)(26)(27). Although Mitf is restricted to the melanocyte lineage, USF-1 is expressed ubiquitously (28), yet both factors can bind the same specific E-box motifs (CANNTG) (20).
Specific interaction between E-box regulatory elements and b-HLH-LZ transcription factors is conferred by the composition of the residues within and outside the core E-box motif. However, a defined E-box motif can be the target site for different transcription factors depending on their relative amount and affinities. In silico analysis of the POMC and MC1R promoters revealed that both proximal promoters displayed several E-box motifs (29 -31) with only a small number conserved in humans, mouse, and rat species that could potentially interact with the Mitf and USF-1 transcription factors to mediate the response of POMC and MC1R to UV irradiation.
We have shown here that the p38-activated USF-1 transcription factor is responsible for UV-induced POMC and MC1R gene expression and that only a restricted set of E-box motifs within the POMC and MC1R proximal promoters mediates the UV response. Using a genetic approach, we defined USF-1 as a key UV-responsive activator of genes implicated in protection from solar radiation.
Immortalized USF-1 Ϫ/Ϫ mouse melanocytes (melan-usf1) were derived from USF-1 knockout mice and were grown as previously described (32), except in RPMI 1640 medium. Mitomycin C-treated XB2 feeder cells were used in establishment and thawing out of frozen stocks of this cell line.
UVB Irradiation-Cells were irradiated using a UV Stratalinker apparatus (Stratagene); physiological UVB irradiation was 312 nm, 50 -80 mJ/cm 2 , i.e. about 1 minimal erythemal dose. Cells were plated at 50 -70% confluence, depending on their doubling time, in 10-and 3.5-cm diameter Petri dishes for RNA and protein preparation, respectively, such that they reached 80% confluence 48 h later (i.e. on UVB induction day). The medium was completely removed following overnight serum starvation and replaced immediately after UVB stimulation. When required, cells were pretreated with p38-specific family kinase inhibitor (10 M SB203580) or with mitogen-activated protein kinase inhibitors (10 M U0126 or PD98059) for 30 min prior to UVB irradiation. Monolayer cells were washed twice with phosphate-buffered saline, harvested by scraping, and then centrifuged at a low speed 3, 5, 8, 24 h following UVB treatment. Cell pellets were resuspended in either 300 l of RNA-plus (Q-BIO-gene) solution for immediate mRNA extraction and storage at Ϫ80°C until use or resuspended in Laemmli buffer for Western blotting analysis.
RNA Extraction and cDNA Synthesis-RNA was extracted by using the RNeasy mini kit (Qiagen) according to the manufacturer's instructions. Cells were treated with RQ DNase (Promega) within the column to ensure the absence of genomic DNA. Reverse transcription reactions were done with the Superscript II reverse transcriptase kit (Invitrogen) according to the manufacturer's instructions, using 1 g of total RNA.
Gene Expression Analysis-Real-time PCR was done in sealed 96well microtiter plates using the SYBR Green TM PCR Master Mix (Applied Biosystems). Gene expression was analyzed by using the ABI Prism 7000 sequence detection system (Applied Biosystems), and results were evaluated with the associated software (version 1.0; Applied Biosystems). The 18 S ribosomal RNA subunit, the values of which remain constant over 24 h, was used as an internal positive control. The relative amounts of POMC, MC1R, EDN1, TYR, and USF-1 transcripts were determined using the cycle threshold method as described by the manufacturer. The mRNA levels at each time point following UVB irradiation are expressed as -fold increase compared with non-irradiated control cells. Each experiment was performed at least twice, and each time point was done in duplicate.
Western Blotting Analysis-Whole cell lysates were resolved by 10% SDS-PAGE using a 200:1 acrylamide: bis-acrylamide ratio. Following blotting, membranes were probed with appropriate primary antibodies and positive signals detected using peroxidase-conjugated anti-rabbit or anti-mouse antibodies. Bound antibodies were visualized by using the ECL Super Signal detection kit (Pierce). The primary antibodies used were a polyclonal anti-USF-1 Ab (USF-C20; Santa Cruz Biotechnology), a polyclonal anti-phospho-p38 Ab (Cell Signaling Technology; Biolabs Inc.), and a monoclonal anti-tubulin antibody (Sigma).
Melanin Content-The melanin content of each cell line was determined by spectrophotometry. Briefly, melanocyte cells were harvested, washed twice with phosphate-buffered saline, and counted. Melanin was solubilized in 0.2 M NaOH (10 6 cells/ml). Melanin concentration was determined by measuring absorbance at 475 nm and comparing with a standard curve of known synthetic melanin concentrations (Sigma). Melanin content is expressed in g/10 6 cells.
Transient Transfection Assays-For in vivo gene expression analysis, cells were plated in supplemented medium in 10-cm diameter Petri dishes and allowed to grow overnight to 80% confluence. Following medium removal, cells were transfected for 1 h using the specific transfection medium OptiMEM (catalogue number 51985-026; Invitrogen) containing up to 5 g of total plasmid DNA, including 4 g of USF-1 constructs (WT pCMV-USF-1 or T153E and T153A mutants), 500 ng each of p38 and MKK6(b)E expression vector (pCDNA p38; pCDNA MKK6(b)E), or empty vector mixed with the Transfast TM reagent (Promega) according to the manufacturer's instructions. After transfection, OptiMEM was removed and replaced with fresh medium. 24 h posttransfection, cells were washed twice with phosphate-buffered saline, harvested, and subjected to RNA extraction as previously described. The RNA was subjected to POMC and MC1R gene expression analysis using real-time PCR.
For luciferase reporter analysis, cells were plated in 12-well plates and transfected as described above using up to 1 g of plasmid DNA. 48 h post-transfection, harvested cells were passively lysed using Promega's buffer. Cell lysates (100 l/well) were screened for luciferase activity using the dual-luciferase reporter system (Promega) according to the manufacturer's instructions. Luciferase activities were quantified using a luminometer apparatus (Turner Design).
POMC and MC1R promoters were isolated by PCR, cut with BamH1 and EcoRI, and cloned into the luciferase reporter pGL3-Basic plasmid (Promega). The POMC and MC1R E-boxes were mutated with the GeneTailor TM mutagenesis system (Invitrogen) using five specific oligonucleotides (sequences available on request).
The pCMV-USF-1 WT and point-mutated constructs have been described previously (27). The p38 expression vector and MKK6(b)E were provided by Dr. Jiahan Han and have been described previously (34).

DNA Binding Assays
Gel Shift Assay-Gel electrophoresis DNA binding assays with crude 501 mel nuclear extracts (4 g) were performed as previously described (35) along with Klenow-labeled probes (0.5 ng). For supershift experiments, the DNA binding reactions were incubated for a further 45 min on ice with a polyclonal anti-USF-1 antibody (200 ng; Santa Cruz Biotechnology) or with nonspecific IgG (anti-rabbit Ab; Amersham Biosciences). The products were loaded onto an 8% polyacrylamide gel (29:1 acrylamide:bis-acrylamide ratio) and subjected to electrophoresis at 200 V for 1 h. The sequences of the double-stranded oligonucleotides used as probes and competitors were as follows: Chromatin Immunoprecipitation Assays-Chromatin immunoprecipitation assays were performed in human and mouse cell lines (melan-a, 501 mel, XB2) essentially as described previously (36) using 10 l of specific antibody (USF-1 Ab; Santa Cruz Biotechnology) or nonspecific IgG (anti-rabbit Ab; Amersham Biosciences). The recovered DNA was subjected to PCR, ensuring that the reaction stayed in the log phase.

RESULTS
POMC and MC1R UV Response-␣-MSH, derived from POMC, and the MC1R receptor are key upstream regulators of the pigmentation cascade (4). Although both ␣-MSH and the MC1R receptor genes are UV-inducible (5,22), the molecular mechanism that mediates their UV responsiveness has yet to be defined. As a first step, we examined constitutive and UVinduced POMC and MC1R expression levels in different skin cell types implicated in the pigmentation process: mouse keratinocytes (XB2), mouse melanocytes (melan-a), and human melanoma (501 mel). Using real-time PCR, we were able to show that POMC and MC1R respond to a physiological dose of UVB irradiation (50 mJ/cm 2 UVB), although time courses and amplitudes differed (Table I).
p38 Kinase Pathway Mediates POMC and MC1R UV-induced Expression-UV irradiation activates stress signaling via both the c-Jun NH 2 -terminal kinase and p38 pathways (25,27,37). To identify the signaling pathway required for regulation of POMC and MC1R UVB expression, we used real-time PCR to examine the response over time of the POMC and MC1R genes to UV irradiation in 501 mel cells in the presence or absence of the specific p38 inhibitor SB203580 (34) or, as controls, inhibitors of the mitogen-activated protein mitogenactivated protein kinase kinase (U0126, PD96059). 501 mel cells were used because they exhibit a robust response of these genes to UVB irradiation. The results (Fig. 1, A and B) indicate that the p38 kinase inhibitor SB203580 inhibited UV-mediated induction of both POMC and MC1R gene expression, strongly suggesting that the p38 stress-activated protein kinase is implicated in the UV responsiveness of POMC and MC1R gene expression. By contrast, neither U0126 nor PD98059 (mitogenactivated protein kinase inhibitors) had any effect (data not shown).
USF-1 Transcription Factor and POMC, MC1R UV-induced Expression-The b-HLH-LZ transcription activators Mitf and USF-1 are both expressed in melanocytes, bind the same E-box motifs in vitro and in vivo (20), and-significantly-both are phosphorylated and activated by the p38 stress-activated kinase (26,27). Because the POMC and MC1R promoters contain E-box elements that are potentially responsive to both Mitf or USF-1 (see below), either factor could potentially mediate the response of these genes to UV irradiation. We therefore sought a genetic approach to resolve this key issue. Because Mitf-null mice lack all pigment cells (38), it was not possible to examine the UV responsiveness of Mitf Ϫ/Ϫ melanocytes. By contrast, USF-1-null mice exhibit normal levels of pigmentation (39), indicating that USF-1 is unlikely to play a major role in melanocyte development. We therefore derived a melanocyte cell line (melan-usf1) from newborn USF-1 knockout mice. In agreement with the genetic modification of the melan-usf1 cell line and unlike the WT melanocyte cell line, no USF-1 transcripts were detected by quantitative real-time PCR ( Fig. 2A), although we observed similar steady state expression of the POMC, MC1R, Tyrosinase, and EDN1 genes ( Fig. 2A) as well as TYRP1 and DCT mRNA levels (not shown). In agreement with the lack of USF-1 mRNA, no USF-1 protein was detected by Western blotting analysis (Fig. 2B). As expected, inactivation of the USF-1 gene did not affect the p38 kinase signaling pathway, which was still activated by UV irradiation as determined by Western blotting for phosphorylated p38 (Fig. 2B) 3 h post-UV irradiation in both WT and USF-1-null melanocytes and inhibited by SB 203580 (not shown). Western blotting also confirmed that in WT cells the presence of phosphorylated p38 correlated with the appearance of a low mobility form of USF-1 that corresponds to the p38-phosphorylated USF-1. Further characterization of the USF-1-negative melanocytes (melan-usf1) suggested they shared most characteristics of their WT counterparts. Both cell lines are dendritic and express the melanosomal proteins (Tyrosinase, Tyrp1, and Dct) with a similar cytoplasmic distribution (Fig. 2C). Moreover, the lack of USF-1 does not affect the accumulation of a basal level of melanin (19.62 and 18.1 g/10 6 cells, respectively) (Fig. 2C). The doubling time of the USF-1-null melanocytes was evaluated by time lapse analysis and appeared to be slightly longer than that of WT cells (about 24 and 20 h, respectively), although M phase was completed in about 70 min and no mitotic abnormalities could be observed in the daughter cells (not shown).
In contrast to WT melanocytes, those lacking USF-1 exhibited a dramatically altered ability to respond to UV irradiation.  Although the levels of both POMC and MC1R mRNA peaked after 3 h in the melan-a cell line (Ͼ7-fold increase), no substantial effect of UV irradiation was detected in expression of these genes in the USF-1-null melanocytes in the 8 h following UV exposure (Fig. 2, D and E) or at 48 h (not shown), ruling out the possibility that the kinetics of activation were severely delayed. As a positive control for the ability of UV irradiation to activate gene USF-1-independent expression, we chose the Endothelin-1 gene (EDN1) that is implicated in the pigmentation process independently of the POMC and MC1R genes and that has been previously shown to be up-regulated in response to UV irradiation (5,22). Although no clear regulatory pathway has been identified for the EDN1 UV response, in silico analysis failed to find any E-box element within the EDN1 promoter with the potential to bind either Mitf or USF-1. Following UV irradiation, the EDN1 gene was induced to a similar degree in both WT and USF-1-null melanocytes. Taken together these data strongly suggest that the USF-1 transcription factor is implicated in the up-regulation of POMC and MC1R gene expression following exposure to UVB irradiation.
USF-1-expressing Vector Restores POMC and MC1R UV Response in USF-1-null Cells-To confirm that the defective response of USF-1-null melanocytes to UV irradiation was caused by the absence of USF-1, we transiently transfected the USF-1-negative melanocyte cell line with vectors expressing either WT USF-1 or mutant forms of USF-1 in which the p38 phosphorylation site (Thr-153) (27) is mutated to either alanine or glutamic acid. The glutamic acid substitution acts to mimic p38 phosphorylation and renders USF-1 largely independent of p38. Cells were then exposed to physiological doses of UVB irradiation (80 mJ/cm 2 ). 3 h post-irradiation, endogenous POMC and MC1R mRNA levels were quantified by quantitative real-time PCR. The results (Fig. 3) reveal that transfection with the WT USF-1 expression vector did not significantly induce endogenous POMC and MC1R. However, following UVB irradiation, a significant induction was observed (Ͼ2-fold) compared with the controls (empty vector) in which no modification occurred following UVB irradiation. Note that in these experiments the transfection efficiency was around 60%; as such, the induction of expression observed must be seen against the background originating from the untransfected cells. Transfection with the T153E mutant, mimicking the transcriptionally activated p38-phosphorylated form of USF-1, induced POMC and MC1R gene expression significantly (2.63-and 4-fold, respectively). Similar results were obtained when the WT USF-1 expression vector was transfected in combination with the p38␣ and the constitutively activated MKK6 kinase (MKK6(b)E). As expected, the T153A mutant, which cannot be phosphorylated by p38, failed to activate either gene. Because ectopic expression of USF-1 is able to restore the activation of POMC and MC1R gene expression following UV irradiation of USF-1-null melanocytes, the data strongly suggest that USF-1 is implicated in the UV-associated transcriptional control of both genes.
USF-1 Interacts with POMC and MCR1 Promoters-In silico analysis revealed that within 750 bp of the transcription start site both the MC1R and POMC promoters displayed several Tyrp1, and Dct proteins were determined by fluorescent microscopy following immunostaining. Tyrosinase, Tyrp1, and Dct primary antibodies were detected using a TRITC-coupled secondary antibody. Melanin content was determined and is expressed in g/10 6 cells. UVinduced POMC (D) and MC1R (E) gene expression was abolished in the melan-usf1 cell line, whereas UV-induced Endothelin-1 (F) gene expression was USF-1-independent. Real-time PCR was performed using mRNA from WT and null USF-1 melanocytes at the indicated times (3 and 8 h) after UVB irradiation (50 mJ/cm 2 ).

FIG. 2. Loss of UV-induced POMC and MC1R gene expression in the USF-1 knockout melanocyte cell line.
A, constitutive expression levels of the POMC, MC1R, USF-1, Tyrosinase (Tyr), and Endothelin-1 (EDN1) genes in melan-usf1 (USF-1 Ϫ/Ϫ) and melan-a cell lines were determined by using real-time PCR. mRNA level is expressed in arbitrary units compared with 18 S values, using the POMCmelan-a value as an internal reference (100 units). B, Western blotting analysis using the anti-USF-1 antibody (Santa Cruz), anti-P-p38 (Cell Signaling) of WT (melan-a), and null USF-1 melanocytes (melan-USF1) following UV irradiation. Anti-␤-tubulin antibody was used as a protein loading control. C, characterization and comparison of the melan-USF1 to WT melanocytes (melan-a). The localizations of the tyrosinase, E-box motifs (CANNTG) that would potentially bind USF-1. Of these only a small number are conserved in humans, mouse, and rat species, namely the Ϫ741 CAGCTG and Ϫ461 CAT-GTG E-box motifs in the MC1R promoter and the CAGGTG element at Ϫ200 in the POMC promoter (Fig. 4A). Initial in vitro binding assays (Fig. 4B) using radiolabeled oligonucleotide probes corresponding to the Ϫ200 and Ϫ461 E-box motifs from the POMC and MC1R promoters, respectively, together with extract from 501 mel melanoma cells revealed a specific DNA-protein complex that was supershifted with anti-USF-1 antibody, but not with a control antibody (not shown). Similar results were observed with the Ϫ183 Tyrosinase E-box element, which is known to bind the USF-1 transcription factor (27). No protein-DNA interaction was observed with the Ϫ741 MC1R E-box element or with a non-conserved E-box located at Ϫ596 (not shown).
To provide direct evidence that the USF-1 transcription factor interacts in vivo with the POMC and MC1R promoters, we performed a chromatin immunoprecipitation assay using a specific anti-USF-1 antibody and a nonspecific IgG as a negative control. Three different cell types (mouse keratinocytes and melanocytes and human melanoma) that express the POMC, MC1R, and Tyrosinase genes to different levels were used to assess the binding of USF-1. After cross-linking and immunoprecipitation, the recovered DNA was subject to PCR using primers spanning the USF-1 binding E-box motifs in the POMC and MC1R promoters (Fig. 4A). As a positive control, the chromatin immunoprecipitation data for the POMC and MC1R genes were compared with those for Tyrosinase, a known USF-1 target gene that is expressed specifically in the melanocyte lineage (27). The HSP70 promoter was used as a negative control. The results are shown in Fig. 4C. In the human 501 mel melanoma cell line, a specific PCR product corresponding to USF-1-bound POMC, MC1R, or Tyrosinase promoter DNA was readily detected using the anti-USF-1 antibody, but not with the IgG control. No PCR product was observed for the HSP70 promoter or for the other promoters if PCR primers were used that did not span a conserved E-box motif (data not shown). Although the efficiency of the immunoprecipitation was significantly reduced in the mouse XB2 keratinocyte and melan-a melanocyte cell lines compared with the human 501 mel cells, the results were similar with the exception that no USF-1 chromatin immunoprecipitation was observed at the Tyrosinase promoter in the XB2 keratinocytes, most likely because Tyrosinase expression is limited to the melanocyte lineage.
In Vitro Transcriptional Regulation of POMC and MC1R Genes-To determine whether USF-1 binding to E-box motifs in the POMC and MC1R promoters was relevant for regulation of these genes in response to UV irradiation, transient transfection experiments were performed in 501 mel cells. POMC and MC1R proximal promoters linked to the pGL3-luciferase reporter plasmid were used to assess whether both promoters are UV-inducible in a USF-1-and E-box-dependent fashion. Cells transfected with WT or E-box-mutated POMC or MC1R promoter reporters were exposed to UVB irradiation and luciferase activity determined. The results (Fig. 5A) showed that, although expression from the WT POMC and MC1R promoters was induced 3-and 5-fold, respectively, following UVB irradiation, mutation of the Ϫ200 POMC E-box element and the Ϫ461 MC1R one abrogated UV-induced gene expression, indicating that the USF-1-bound E-box motifs are indeed UV target sites. By contrast, deletion of the Ϫ741 MC1R E-box motif, which is conserved in human and mouse species but which failed to bind the USF-1 transcription factor in vitro, did not affect the UV response. The requirement of the stress-activated form of USF-1 in POMC and MC1R gene regulation was further confirmed using this reconstituted system. POMC and MC1R promoter reporters were highly induced in the presence of the constitutive activated USF-1 form (T153E), whereas WT USF-1 had no effect. Moreover, mutation of the USF-1-specific binding site abrogated gene expression (Fig. 5B). Taken together, these results suggest that stress-activated USF-1 and specific E-box are critically required for activation of the MC1R and POMC genes in response to UV irradiation. DISCUSSION The tanning response is the major molecular response toward UV irradiation. It protects the skin and avoids oncogenic transformation to a certain extent. It is thus important to understand how target proteins mediate the transcriptional response to UV irradiation and to general stress signals. We have previously shown that USF-1 is an in vivo target for the stress-responsive p38 kinase (27). In this study, we demonstrated that the USF-1 transcription factor is a general key regulator of UV-induced gene expression, implicated at several steps of the tanning response. Our results elucidate the general melanocyte pigmentation process in response to UV induction.
The ␣-melanocyte-stimulating hormone, encoded by the Proopiomelanocortin gene, and its receptor, encoded by the melanocortin receptor 1 (MC1R), are two upstream partners of the pigmentation cascade (4, 7). Using real-time PCR, we were able to quantify and characterize POMC and MC1R transcripts. Levels of both transcripts increased following exposure to physiological doses of UVB irradiation in all cell lines tested (XB2 keratinocytes, melan-a melanocytes, and 501 mel melanomas), with differing kinetics. In accordance with these data, POMC and MC1R promoter-luciferase reporters were both activated following UV induction. Moreover, UV-induced gene expression is dependent on the stress-responsive p38 family kinases, as SB 203580, a specific inhibitor of the closely related p38␣ and p38␤ kinases (34), abrogated UV-induced transcription of POMC and MC1R.
The establishment of an immortal USF-1 knockout melanocyte line (39) allowed us to study gene expression profiles following UV induction in the absence of one of the stressresponsive p38 kinase targets while keeping the p38 kinase cascade intact. POMC and MC1R gene expression following UV irradiation was abrogated in the USF-1 knockout melanocyte strain, suggesting a major role for the USF-1 transcription factor. The involvement of the activated phosphorylated form of USF-1 was confirmed by the fact that UV-induced POMC and MC1R expression was restored in the USF-1-depleted line following transfection of the T153E-USF-1 expression vector mimicking the p38-activated USF-1 form.
Interestingly, the expression of POMC and MC1R following UV induction was dependent on the presence of the USF-1 transcription factor, whereas their constitutive gene expression does not seem to be. Similar constitutive levels of POMC and MC1R transcription were indeed observed in WT melanocytes (melan-a) and in USF-1-depleted ones. Moreover, the constitutive expression levels of POMC and MC1R were not affected when USF-1 knockout cells were transfected with the WT-USF-1 expression vector alone. Introduction of the WT-USF-1 expression vector into the USF-1-depleted cell line resulted in the up-regulation of POMC and MC1R genes only in the presence of UV irradiation. A similar gene regulation model involving the E-box target motif was observed for the Tyrosinase gene. The tissue-specific microphthalmia transcription factor is required for constitutive gene expression, and the ubiquitously expressed USF-1 transcription factor is responsible for UVinduced Tyrosinase gene expression (27). Consistent with this dual regulation model, the Ϫ461 MC1R conserved E-box motif has been shown to bind the Mitf transcription factor, resulting in transcriptional activation (40). This regulation model is possible because of the nature of the USF-1 and Mitf transcription factors. Both proteins are members of the evolutionarily conserved family of b-HLH-LZ transcription factors (41)(42)(43). Members of this family interact with symmetrical E-boxes with the consensus sequence 5Ј-CANNTG-3Ј (44 -46). The specificity of the E-box binding site toward the b-HLH-LZ family members is conferred by the composition of the residues within and outside the core E-box motif (41). However, a defined E-box motif can be the target site for different transcription factors. The relative amounts of the potential transcription factors and their binding affinities, which can be modified by post-translational modifications, will differ at particular time points, leading to a specific protein-DNA interaction (28,35). We thus hypothesize that following UV irradiation, phosphorylation of the USF-1 transcription factor renders possible its specific interaction with target E-box motifs.
The shortness of the E-box sequence, six nucleotides with potential degenerate sites, favors its broad distribution throughout the genome. However, only a restricted set of E-box supershift experiments, DNA binding reactions were incubated for a further 45 min on ice with a polyclonal anti-USF1 antibody or nonspecific IgG. Retarded complexes are indicated, and the supershifted complexes are indicated by a black arrow. C, USF-1 binds in vivo the POMC and MC1R promoters. Chromatin immunoprecipitation assays were performed on XB2 keratinocytes, melan-a melanocytes, and 501 mel melanoma cells using anti-USF1 antibody or nonspecific IgG. Recovered DNA was subjected to PCR using primers specific for POMC and MC1R promoters and centered on the conserved E-box motifs as shown in panel A. Specific primers for the Tyr promoter were used as positive control, and specific primers for Hsp70 were used as negative control.  (34) used for the chromatin immunoprecipitation assays are shown relative to the transcription start sites of each promoter. The slight numeration discrepancy between the human and mouse POMC promoters is because of the presence of an additional non-conserved sequence in the human POMC promoter upstream from the start site. B, USF-1 binds specific E-box motifs within the human POMC and MC1R promoters. In vitro DNA binding assays were performed using 501 mel nuclear extracts. Oligonucleotides used as probes were centered on the Ϫ200 POMC-E-box, the Ϫ461 MC1R-E-box, and the Ϫ183 Tyr-E-box. For motifs are actual DNA control elements. One way to identify such elements to a certain extent would be to compare promoter sequences, hypothesizing that crucial regulatory elements would be conserved over evolution and present within different species. Accordingly, the DNA control E-box element of the Tyrosinase promoter gene is conserved in human, mouse, quail, and turtle (20). We thus focused on the POMC and MC1R conserved E-box motif in human, mouse, and rat promoters to design DNA interaction assays to identify USF-1 target sites. Chromatin immunoprecipitation assays using anti-USF-1 antibodies and electrophoresis DNA binding assays allowed us to define only one DNA control E-box element for each promoter, located at 461 and 200 bp upstream, respectively, from the POMC and MC1R start sites. The function of these elements was further studied by in vitro transfection assays, implicating for the first time these E-box motifs as UV target cis-elements. Interestingly, one conserved MC1R E-box motif, located 741 bp upstream from the start site, failed to bind in vitro the USF-1 transcription factor. Deletion of this motif did not significantly modify the MC1R promoter-luciferase expression level in the in vitro transfection assays. However, it remains possible that the Ϫ741 conserved E-box motif is a potential target site for another member of the E-box transcription factor family under specific circumstances.
The tanning response is characterized by the increased production of melanin that acts as a factor in protection against UV irradiation. The increase in melanin production is achieved by activation of genes such as MC1R and POMC that increase intracellular cAMP levels and thereby elevate the activity and expression of melanogenic enzymes. In addition to the direct effect of UV irradiation on melanocytes, a significant component of the tanning response is mediated by keratinocyte-derived paracrine signals such as Endothelin 1 and FGF2 (47)(48)(49) that are known to signal at least in part via the p38 stressactivated kinase (49), the pathway identified here as being crucial for UV-induced activation of POMC and MC1R expres- FIG. 5. Transcriptional regulation of the human POMC and MC1R luciferase-reporter. A, POMC and MC1R luciferase-reporters respond to UV irradiation. A POMC luciferase-promoter extending from positions Ϫ245 to ϩ1 and two MC1R luciferase-promoters extending from positions Ϫ862 and Ϫ724 to ϩ1 were transfected into 501 mel cells, along with a mutated form affecting Ϫ200 Ebox POMC luciferase-promoter and Ϫ461 E-box MC1R luciferase-promoter. E-box motifs conserved in human and mouse sequences are shown in black (f), whereas non-conserved ones are in white (Ⅺ). Mutated E-box are cross-marked (ϫ). Cells were exposed to UVB irradiation (80 mJ/ cm 2 ) 24 h post-transfection when indicated. Cells were harvested 48 h posttransfection, and luciferase activity was determined and expressed as -fold induction compared with the controls (empty vectors). B, specific E-box motifs within the POMC and MC1R promoter-luciferase target the constitutive-activated USF-1 transcription factor. 501 mel cells were transiently transfected with WT promoter-luciferase constructs and mutated Ebox motif (ϫ) promoter-luciferase constructs. Cells were co-transfected with vectors expressing WT USF-1 (pCMV-USF-1-WT) or with the T153E-USF-1 form, which mimics phospho-activated USF-1 transcription factor. Transcription levels are expressed as -fold induction compared with the controls (empty vectors). sion. How EDN1 and FGF2 signaling might be linked to elevated POMC and MC1R expression was not known. Indeed, the fact that the MC1R promoter is bound and regulated by Mitf (40), a known p38-responsive transcription factor, seemed to point to Mitf as the likely candidate. However, the fact that melanocytes derived from USF-1-null mice retain their capacity to activate the p38 signaling pathway and EDN1 expression, but are defective in their ability to activate POMC and MC1R expression in response to UV irradiation, unequivocally defines USF-1 as a key component of the tanning response. Interestingly, ␣-MSH, encoded by the POMC gene, acts synergistically with Endothelin-1 and FGF2 in promoting melanogenesis (50). Because activation of the p38 kinase by Endothelin-1 and FGF2 would lead to increased activity of USF-1 and thereby activation of POMC expression and increased MSH expression, there exists a potential positive feedback loop within the tanning response that would facilitate the accumulation of protective pigment in response to UV irradiation and keratinocyte signaling to melanocytes. Our results have provided new insights into the molecular pathways and cross-talk implicated in gene regulation upon solar irradiation, leading to a general tanning response model (Fig. 6).
In addition to their role in pigmentation, ␣-MSH and USF-1 mediate many diverse physiological actions, including immunomodulation and anti-inflammatory effects (51), which are stress response events, and cell survival and apoptosis (52,53). This suggests that USF-1 and ␣-MSH may interact further at the molecular level. Their cooperation may thus not be limited to the pigmentation process and may be much more complex in response to a stress signal. Use of the USF-1 knockout melanocyte cell line should help us to understand the molecular mechanisms underlying these processes.