RasGRP1 Represents a Novel Non-protein Kinase C Phorbol Ester Signaling Pathway in Mouse Epidermal Keratinocytes*

The mouse skin model of carcinogenesis has been instrumental in our appreciation of the multistage nature of carcinogenesis. In this system, tumor promotion is a critical step in the generation of tumors and is usually achieved by treatment with the phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA). Although it is generally assumed that protein kinase C (PKC) is the sole receptor for TPA in this system, we sought to evaluate whether non-PKC pathways could also contribute to the effects of phorbol esters in skin. We documented expression of the high affinity non-PKC phorbol ester receptor and Ras activator RasGRP1 in mouse primary keratinocytes. Overexpression of RasGRP1 in keratinocytes increased the level of active GTP-loaded Ras. TPA treatment further elevated this Ras activation in a PKC-independent manner and induced the translocation and down-regulation of RasGRP1. Overexpression of RasGRP1 in keratinocytes also caused apoptosis. Finally, induction of keratinocyte differentiation by elevation of extracellular calcium suppressed expression of endogenous RasGRP1, whereas overexpression of RasGRP1 inhibited expression of the differentiation markers keratins 1 and 10 induced by high calcium in the medium. Taken together, our results demonstrate that RasGRP1 is an additional diacylglycerol/phorbol ester receptor in epidermal keratinocytes and suggest that activation of this novel receptor may contribute to some of the phorbol ester- and Ras-mediated effects in mouse epidermis.

Carcinogenesis is a multistage process that reflects the cooperation between oncogenic mutations and epigenetic mechanisms that alter cellular homeostasis. Understanding the mechanism governing this cooperation is therefore of great importance, as it may identify molecular targets that can be used for interrupting tumor development before malignant progression occurs. One of the experimental models that has been instrumental in investigating the stages of cancer formation is the induction of tumors in mouse skin. In this model, a carcinogen initiates cells by introducing a mutation in a ras gene, and the initiated population expands under treatment with phorbol esters, non-genotoxic compounds with tumor-promoting activities. Although phorbol esters are the prototypes of tumor promoters in skin, the nature of their interaction with Ras signaling in tumor promotion remains to be determined.
The phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA) 1 is the classic mouse skin tumor promoter (1,2). When cultures of basal proliferating epidermal keratinocytes are exposed to this agent, their growth is arrested, and they undergo a process of differentiation (3,4). However, initiated cells that harbor a mutant ras gene are resistant to the differentiation effects of TPA (5,6). In vivo, the consequence of heterogeneity of response is tumor promotion: the clones of the carcinogenaltered cells are selected and expanded because they resist the induction of terminal differentiation and are stimulated to proliferate (7). Historically, protein kinase C (PKC) has been considered responsible for the phorbol ester effects (8). The role of some of the PKC isoforms in the tumor-promoting actions of phorbol esters has been recently assessed using transgenic animals. PKC␦ overexpression in the epidermis induces resistance to tumor promotion by TPA (9). PKC␣ increases cutaneous inflammation, but has no effect on tumor promotion (10). Overexpression of PKC⑀, which appears to be oncogenic in a variety of cell types (11,12), causes a reduction in the papilloma burden, but enhances carcinoma formation (13). One emerging question is whether other receptors for phorbol esters contribute to tumor promotion in skin. There are at least three different families of non-PKC phorbol ester receptors: the chimerins (14), Munc13 (15), and RasGRP (16). All of these receptors can respond to phorbol esters with sensitivities similar to PKC (15,17,18). The chimerins are GTPase-activating proteins for Rac, a small GTPase involved in cytoskeletal rearrangements (14). Although chimerin is present in a variety of tissues (19), expression in the epidermis has not been demonstrated. The Munc13 family comprises mammalian proteins that participate in synaptic transmission and that are expressed in brain (15). Finally, RasGRP1 is a phorbol ester receptor that functions as an activator of Ras (16,18), is transforming in rat fibroblasts (16), and is a candidate oncogene for mouse retrovirus-induced T-cell lymphomas and leukemias (20,21). The two functional characteristics of RasGRP1, the ability to bind to phorbol esters and to modulate Ras activity, make RasGRP1 an attractive candidate for an additional target of the phorbol esters in skin that could directly link phorbol ester signaling with Ras signaling and the tumorigenic process.
In this study, we report that primary mouse keratinocytes express RasGRP1 and that this receptor mediates Ras activation and can translocate and be down-regulated in response to phorbol ester treatment. Moreover, calcium-induced differentiation suppresses the levels of RasGRP1 in a concentrationdependent manner, whereas overexpression of RasGRP1 in epidermal cells inhibits the expression of markers of differentiation and causes cell death, effects that may be mediated by activation of Ras. In summary, our findings indicate that Ras-GRP1 is an additional diacylglycerol/phorbol ester receptor in epidermal keratinocytes and suggest that activation of this novel receptor may contribute to some of the phorbol ester-and Ras-mediated effects in mouse skin.
Primary Keratinocyte Cultures and Infection-Primary cultures of normal mouse keratinocytes were prepared from newborn mice (2-3 days old) by the trypsin flotation procedure (22) with modifications. Briefly, skins were floated, dermis down, in 0.25% trypsin at 4°C overnight. Afterward, the epidermis was separated mechanically from the dermis in S-MEM-high calcium medium (Eagle's minimal essential medium containing 8% Chelex-treated fetal bovine serum and supplemented with 1.2 mM calcium), chopped, and centrifuged. The pellet was washed twice, and the keratinocytes were resuspended in S-MEM-high calcium medium, followed by filtration through a cell strainer to remove cell debris. Cells were plated at 1.5 ϫ 10 6 cells/60-mm Petri dish and incubated overnight at 36°C, after which the medium was changed to S-MEM-low calcium medium (0.04 mM) to maintain the cells under basal proliferating conditions. Keratinocytes were used within 5-6 days after plating. Adenoviral infection was done on keratinocytes plated on 60-mm dishes. The volume of the medium in the dishes was reduced to 1 ml, and adenoviruses were added to the dishes at the appropriate multiplicity of infection, followed by a 4-h incubation at 36°C. After incubation, 3 ml of S-MEM-low calcium medium was added, and the dishes were incubated for an additional 24 -72-h period depending on the experiment.
Reverse Transcription (RT)-PCR-Cultures of primary keratinocytes were harvested, and total RNA was isolated using the high pure isolation kit from Roche Applied Science, treated with DNase, and used for a two-step RT-PCR using Ready-To-Go RT-PCR beads (Amersham Biosciences) according to the manufacturer's protocol. The first reaction step was performed using 1 g of total RNA and 0.5 g of oligo(dT) [12][13][14][15][16][17][18] primer. For the amplification reaction (PCR step), specific primers for mouse RasGRP1 (5Ј-primer, aaa gat ctg ctc cat gac ccc gaa gaa g; and 3Ј-primer, gac atc cca gta aac ccg tct gca gat tc) or the mouse ornithine decarboxylase control (5Ј-primer, cac gct tgc agt caa cat cat tgc caa aa; and 3Ј-primer, cat tga tcc tag cag aag cac agg ctg) were used. To test for DNA contamination, the RT-PCR beads were heated at 95°C for 10 min to inactivate the reverse transcriptase prior to initiating the RT reaction. RT-PCR with the OCD primers also served as a control for DNA contamination, as the PCR product resulting from genomic amplification should be 1237 nucleotides versus the 530-nucleotide product that originated from mRNA.
Preparation and Staining of Keratinocytes-Primary keratinocytes seeded on Permanox slide chambers were fixed with methanol at Ϫ20°C for 10 min. For staining with the m199 monoclonal antibody, blocking was performed utilizing the Vector M.O.M. kit (Vector Labs, Inc.). An 8% solution of bovine serum albumin was utilized as the blocking agent when using the J32 primary polyclonal antibody. The secondary antibodies used were Alexa Fluor 594-conjugated anti-mouse (Molecular Probes, Inc.) and Cy2-conjugated anti-rabbit (Jackson Im-munoResearch Laboratories, Inc.). Staining of actin fibers was done using Alexa Fluor 594-conjugated phalloidin (Molecular Probes, Inc.) according to the manufacturer's protocol.
Immunoprecipitation and Immunoblotting-Primary keratinocytes were harvested in lysis buffer (50 mM Tris-HCl (pH 7.4) containing 150 mM NaCl, 1% IGEPAL, 1 mM MgCl 2 , 1 mM dithiothreitol, 1 mM EDTA, 1 mM NaF, 1 mM sodium orthovanadate, and one tablet of Mini-Complete protease inhibitors (Roche Applied Science)). Samples for immunoblotting were prepared in 2ϫ Laemmli sample buffer and boiled before being subjected to SDS-PAGE. For immunoprecipitations, 2 mg of total lysate protein was incubated with 2 g of primary antibody overnight at 4°C, followed by a 2-h incubation with protein A/G-Agarose beads in the cold. The immunocomplexes were pelleted by centrifugation, washed three times with lysis buffer, and resuspended in 2ϫ Laemmli buffer. After boiling, 50 -100 g of lysate protein was loaded onto acrylamide gels for SDS-PAGE. Immunoblotting was done on nitrocellulose membranes. Primary antibodies were incubated overnight at 4°C, followed by a 2-h incubation with a horseradish peroxidase-conjugated secondary antibody at room temperature and chemiluminescence detection using SuperSignal West Dura substrate (Pierce). When fractionation was performed, cells were harvested in 50 mM Tris-HCl supplemented with protease inhibitors and then sonicated for 8 s to cause lysis. Soluble or cytosolic fractions were prepared from aliquots of 60 l of total lysates by ultracentrifugation at 100,000 ϫ g for 1 h at 4°C. The supernatants were collected as the "soluble" fractions.
GTP-loaded Ras Pull-down Assay-Levels of GTP-loaded Ras were measured using glutathione S-transferase/Raf-1 Ras-binding domain as a probe in an affinity precipitation or pull-down assay as described previously (23). Briefly, primary keratinocytes were serum-starved (0.1% fetal bovine serum) overnight, treated with TPA or vehicle, and harvested on ice in magnesium lysis buffer (25 mM HEPES (pH 7.5) containing 150 mM NaCl, 1% IGEPAL, 10% glycerol, 10 mM MgCl 2 , 25 mM NaF, 1 mM EDTA, 1 mM sodium orthovanadate, and one tablet of Mini-Complete protease inhibitors). Lysates were rotated at 4°C and clarified by centrifugation. Lysate protein (300 g) was incubated with glutathione S-transferase/Raf-1 Ras-binding domain conjugated to glutathione-Sepharose 4B beads (Amersham Biosciences) for 1.5 h with rotation in the cold. The affinity complexes were washed three times with lysis buffer, resuspended in 2ϫ Laemmli buffer, boiled, and run on 15% acrylamide gels. Aliquots of the total lysates were run in parallel as a control of the "input" of total Ras in the pull-down assay. Proteins were blotted onto nitrocellulose membranes; immunostaining was performed using anti-pan Ras antibody (clone Ras10).
Flow Cytometry-Primary keratinocytes were trypsinized for 5 min, centrifuged, washed twice with cold phosphate-buffered saline (PBS) containing 0.1% fetal bovine serum, and then fixed in 5 ml of ice-cold 70% ethanol. After incubation on ice for 30 min, cells were centrifuged, washed twice with PBS, and rehydrated in the last wash with PBS by incubation at room temperature for 15 min. After centrifugation, the keratinocytes were incubated for 30 min at 37°C in 200 l of PBS containing 1 g/ml RNase A, followed by an additional 15-min incubation at room temperature with 1 ml of a 3 M propidium iodide solution containing 100 mM Tris-Cl (pH 7.4), 150 mM NaCl, 1 mM CaCl 2 , 0.5 mM MgCl 2 , and 0.1% IGEPAL. Cells were then analyzed by flow cytometry using a FACScan apparatus.
Annexin V/Propidium Iodide Staining-Detection of externalization of phosphatidylserine in apoptotic cells was done using the Vibrant apoptosis assay kit (Molecular Probes, Inc.) according to the manufacture's instructions. Briefly, keratinocytes were plated on Permanox chamber slides; infected with AdLacZ or AdRasGRP1 for 48 h, and then washed with cold PBS, followed by one wash with 1ϫ annexin binding buffer. Slides were incubated with 20 l of fluorescein isothiocyanatelabeled annexin V and 2 l of 100 g/ml propidium iodide in 200 l of 1ϫ annexin binding buffer for 15 min at room temperature. Cells were washed twice with 200 l of 1ϫ annexin binding buffer and then observed under a fluorescence microscope using the appropriate filters.

Expression of RasGRP1 in Primary
Keratinocytes-To study the expression of RasGRP1 in skin, we used primary cultures of mouse keratinocytes. Mouse keratinocytes are the target cells in chemically induced skin carcinogenesis. Cells were cultured for 3-6 days, and total cellular RNA was isolated, treated with DNase to eliminate genomic DNA contamination, and then utilized for RT-PCR using specific primers for mouse RasGRP1. As shown in Fig. 1A, the results revealed a transcript of ϳ600 bp (expected size of 578 bp). The product was excised from the gel, purified, and sequenced to confirm identity to the published sequence for mouse RasGRP1.
To confirm expression of RasGRP1 at the protein level, we performed immunoprecipitations followed by immunoblotting using a monoclonal antibody that has reactivity toward mouse RasGRP1 (m199). This antibody recognized two bands with an apparent molecular mass of ϳ90 kDa, which most likely reflect post-translational modifications. Similar doublets were also observed in lysates from T-cells of mouse origin (24). These bands were not detected when immunoprecipitation was performed with an irrelevant mouse isotype IgG1 (Fig. 1B, second  lane). Immunofluorescence studies using both monoclonal and polyclonal antibodies against RasGRP1 yielded a consistent image for the subcellular distribution of this protein in keratinocytes, which corresponded to cytoplasmic distribution with strong perinuclear localization (Fig. 2). As expected, cells ectopically expressing the GFP fusion protein of RasGRP1 (Ras-GRP1-GFP) showed a localization pattern for the GRP protein similar to that for the native protein (Fig. 2e).
RasGRP1 Activates Ras in Response to Phorbol Ester Treatment in Keratinocytes-Previous studies including our own have demonstrated that RasGRP1 is a high affinity receptor for phorbol esters both in vitro and in vivo (16,18,25). If phorbol esters like TPA activate RasGRP1 in keratinocytes, RasGRP1 is expected to activate its substrate, Ras. We first examined the ability of the phorbol ester TPA to induce Ras activation in mouse keratinocytes under control conditions. To do this, we measured the level of active GTP-loaded Ras using the Rasbinding domain of Raf-1 as a probe in an affinity precipitation (pull-down) assay. In primary keratinocytes, a 15-min treatment with 1 M TPA elevated the GTP-loaded Ras levels 3-4 times over the control levels observed in vehicle-treated cells (Fig. 3, A and B). To test whether the activation was dependent on PKC, we pretreated the cells with the specific PKC inhibitor GF109203X before the addition of TPA. We have previously used similar concentrations of this inhibitor to successfully abrogate PKC activation in keratinocytes (26). As shown in Fig.  3A, activation of Ras by TPA was independent of PKC. When RasGRP1 was overexpressed in the keratinocytes by infection with a recombinant adenovirus (AdRasGRP1), GTP-loaded Ras levels increased 3-fold over control levels observed in cells infected with AdLacZ. This activation was further increased by a 15-min exposure to TPA (1 M) (Fig. 3, C and D), consistent with the hypothesis that RasGRP1 responds as a phorbol ester receptor in mouse epidermal cells.
TPA Induces Translocation and Down-regulation of Ras-GRP1 in Keratinocytes-The preceding results indicate that RasGRP1 activated Ras in keratinocytes in response to phorbol ester treatment. We then investigated whether phorbol ester treatment could cause redistribution of RasGRP1. As shown in   (Fig. 4A), suggesting translocation of RasGRP1 to particulate compartments of the cell. To determine the localization of the translocated Ras-GRP1, we examined its cellular distribution in response to TPA by confocal microscopy in live mouse keratinocytes expressing the RasGRP1-GFP fusion protein. As shown in Fig. 4B, 1 M TPA translocated RasGRP1-GFP to the plasma membrane of the keratinocytes during a 15-min treatment. This result confirms that RasGRP1 is able to redistribute in response to phorbol esters in mouse keratinocytes and indicates translocation to the membrane upon short-term exposure to TPA. Additionally, when mouse epidermal cells were exposed to TPA for 24 h, the level of endogenous RasGRP1 was reduced in a concentrationdependent manner (Fig. 4C), suggesting that RasGRP1, like some members of the PKC family such as PKC␦ and PKC␣, is down-regulated in this system in response to phorbol ester treatment.
RasGRP1 Overexpression Induces Apoptosis-To gain insight into the role of RasGRP1 in keratinocyte biology, we used recombinant AdRasGRP1-HA to overexpress C-terminally HA epitope-tagged rat RasGRP1. With this approach, Ͼ90% of the cells were infected. Fig. 5A shows the levels of expression of the recombinant protein after a 48-h infection at different multiplicities of infection. The tagged rat RasGRP1 protein migrated slightly slower than the endogenous mouse RasGRP1 protein, probably because of the presence of the HA tag. We noticed that the levels of endogenous RasGRP1 were lower in infected cells compared with control keratinocytes. This was evident not only for the RasGRP1-HA overexpressor, but also in keratinocytes infected with the AdLacZ control vector, although to a much lesser extent.
Although normal mouse keratinocytes in low Ca 2ϩ medium (0.04 mM) showed characteristics of basal cell morphology, in-FIG. 3. TPA-induced activation of Ras through RasGRP1 in mouse keratinocytes. A, primary mouse keratinocytes were serum-starved overnight and then treated for 15 min with TPA or vehicle (Me 2 SO (DMSO)). When indicated, GF109203X (GF) was added to the cells for 30 min before and during TPA treatment. Active GTP-loaded Ras (Ras GTP ) was affinity-precipitated as described under "Experimental Procedures." Levels of Ras (total Ras) were determined by Western blotting on an aliquot of the lysate sample used as input for the affinity precipitation. Similar results were obtained in three independent experiments. B, the intensities of the bands were determined by densitometry. The graph indicates the mean Ϯ S.E. of three independent experiments. C, keratinocytes were infected with control AdLacZ or AdRasGRP1 for 24 h, subjected to serum starvation overnight, and then treated with TPA or vehicle as described for A. Similar results were obtained in four independent experiments. D, the intensities of the bands were determined by densitometry. The graph indicates the mean Ϯ S.E. of four independent experiments. fection with AdRasGRP1 caused significant rounding and detachment of the cells (Fig. 5B). When mouse keratinocytes were analyzed by staining with phalloidin to visualize the actin cytoskeleton, we observed that ectopic expression of RasGPR1 caused the disappearance of actin stress fibers (Fig. 5C). To assess whether these changes were causing apoptosis, we monitored three parameters that are associated with the apoptotic process: cleavage of the caspase substrate PARP, externalization of phosphatidylserine (annexin V binding), and DNA fragmentation (hypodiploid DNA or sub-G 1 peak). As shown in Fig.  5A, RasGRP1 overexpression resulted in increased cleavage of PARP, indicating caspase activation. Furthermore, by labeling with annexin V and propidium iodide, we observed an increase in both early apoptosis (binding of annexin V to the plasma membrane with no propidium iodide staining of the nucleus) and late apoptosis (annexin V and propidium iodide staining) after 48 h of infection with AdRasGRP1 (Fig. 5D). In contrast, cell cultures infected with the control adenovirus (AdLacZ) displayed a considerably lower number of cells staining positive for annexin V and/or propidium iodide (Fig. 5D). DNA content analysis of the cell population attached to the dishes at 48 h post-infection showed that the hypodiploid peak (sub-G 1 ), which corresponds to cells with fragmented DNA due to apoptotic degradation, was significantly higher in cells overexpressing RasGRP1 (15.29 Ϯ 2.12%, n ϭ 3) compared with control (LacZ) cells (0.35 Ϯ 0.19%, n ϭ 3) (Fig. 5E). Taken together, these results confirm that RasGRP1 overexpression induces apoptosis in primary keratinocytes.
Activation of the Raf/MEK/ERK Cascade in Response to Ras-GRP1 and TPA-Activation of Ras in the epidermis selectively activates the Raf/MEK/ERK cascade (27,28). We have shown that RasGRP1 overexpression caused a significant increase in the levels of active Ras in primary keratinocytes (Fig. 3, C and D). As expected, the level of active phospho-Thr 202 /Tyr 204 ERK was significantly elevated in primary keratinocytes infected with AdRasGPR1 (Fig. 6A). No changes were observed in the total amount of ERK protein (Fig. 6A), indicating that the increment in phospho-ERK was the result of phosphorylation and not of up-regulation of ERK levels in the cells. When cells overexpressing RasGRP1 were treated with 1 M TPA for a short period of time (15 min), a further and robust increase in ERK phosphorylation was observed (Fig. 6B). However, this increase was sensitive to PKC inhibition by GF109203X treatment. To investigate this, we examined the effect of TPA on Raf-1, the first step in the cascade downstream of Ras activation. Raf-1 was analyzed by measuring the level of phospho-Ser 388 , a phosphorylation that is absolutely required for activation of Raf-1 and that closely coincides with Raf-1 kinase activity (29,30). In serum-starved cells, the addition of TPA caused an increase in the amount of phospho-Ser 338 Raf-1 (Fig.  6B). However, as was observed with ERK phosphorylation, inhibition of PKC by GF109203X treatment completely abrogated Raf-1 phosphorylation (Fig. 6B). These results indicate that, although TPA increases GTP-loaded Ras levels in a PKCindependent manner, PKC is required for TPA-mediated ERK phosphorylation through activation of Raf-1.
RasGRP1 and Keratinocyte Differentiation-Phorbol esters like TPA have been shown to induce morphological changes in cultured keratinocytes that resemble some aspects of keratino-cyte differentiation (4,31). Because differentiation is typically controlled by the calcium concentration in the media, we investigated whether RasGRP1 levels could be also modulated by changes in extracellular calcium. As shown in Fig. 7A, cells exposed to calcium showed morphological changes indicative of keratinocyte differentiation, characterized by cell flattening and shedding from the plates. Although the level of RasGRP1 was similar in cells cultured at an extracellular calcium concentration of 0.12 mM compared with 0.04 mM, there was a decrease of 80% in the level of endogenous RasGRP1 in cells cultured for 24 h in medium at 1.2 mM calcium (Fig. 7A). These results suggest that RasGRP1 levels, as has been observed for some of the PKC isoforms, may be regulated in vitro by the status of differentiation of the keratinocytes. To further examine the impact of RasGRP1 activation on differentiation, we FIG. 5. Overexpression of RasGRP1 in mouse keratinocytes causes cell death and morphological changes in the cells. A, primary keratinocytes were infected with either control AdLacZ or AdRasGRP1. After 48 h, cells were harvested, and lysates were subjected to immunoblotting for detecting expression of RasGRP1, cleaved PARP, and actin. Similar results were obtained in three independent experiments. B, shown are representative microphotographs of keratinocytes expressing LacZ or RasGRP1 after 48 h of infection with the corresponding adenoviruses. Many cells overexpressing RasGRP1 had already detached from the dishes. C, keratinocytes expressing either GFP or the RasGRP1-GFP fusion protein were fixed with 4% paraformaldehyde, permeabilized with Triton X-100, and then stained with phalloidin for visualization of the actin cytoskeleton as described under "Experimental Procedures." The left panels depict primary keratinocytes expressing either RasGRP1-GFP or GFP alone and microphotographed using a fluorescein isothiocyanate fluorescence channel. The right panels correspond to cells expressing either RasGRP1-GFP or GFP, stained for actin with Alexa Fluor 594-conjugated phalloidin (red fluorescence), and imaged by fluorescence microscopy using a Texas Red channel. The images shown in B and C are from a representative experiment. Two additional experiments yielded similar results. D, keratinocytes infected for 48 h with AdLacZ or AdRasGRP1 were stained with fluorescein isothiocyanate (FITC)-labeled annexin V (green) and propidium iodide (PI) (red) as described under "Experimental Procedures." Red arrows indicate late apoptotic keratinocytes, which show both annexin V and propidium iodide staining; the green arrow shows an early apoptotic cell with only annexin V binding. Results are representative of three independent experiments. E, keratinocytes were infected with AdRasGRP1 or the control adenovirus for 48 h. Cells were then stained with propidium iodide for flow cytometry analysis. The percentage of hypodiploid cells (sub-G 1 peak) is indicated in each panel (n ϭ 3). MOI, multiplicity of infection; Ab, antibody.
looked at the effect of RasGRP1 overexpression on the differentiation induced by calcium in murine keratinocytes. Previous reports have revealed that ectopic expression of oncogenic Ras can inhibit calcium-induced differentiation of mouse keratinocytes (27,32). Keratins 1 and 10 are markers of keratinocyte differentiation induced by elevated calcium. We found that overexpression of RasGRP1, consistent with its capacity to activate Ras, inhibited the expression of the differentiation markers keratins 1 and 10 caused by increases in extracellular calcium (Fig. 7B).

DISCUSSION
The mouse epidermis is a tissue responsive to phorbol esters like TPA, which is the classic tumor-promoting agent utilized in the two-stage carcinogenesis protocol in skin. The PKC family members are the best known receptors for phorbol esters, and at least five isoforms (33) have been identified in keratinocytes, the target cells in skin carcinogenesis. In this report, we have shown that RasGRP1, a non-PKC phorbol ester receptor, is also expressed in mouse epidermal keratinocytes and can mediate phorbol ester effects independently of PKC.
RasGRP is a family of guanyl nucleotide exchange factors for the Ras-like GTPases, and it is composed of at least four members: RasGRP1 (16), RasGRP2 (34,35), RasGRP3 (23,36), and RasGRP4 (37). Expression of the RasGRP proteins is high in brain and hematopoietic tissues (16,25,34), although RasGRP transcripts have been detected also in lung, heart, and kidney by RT-PCR (36). Using a multiple-tissue panel of human normalized first-strand cDNA, we were able to detect RasGRP1 transcripts not only in thymus, primary leukocytes, spleen, and brain, but also in testis, placenta and kidney, 2 although expression of RasGRP1 at the protein level in those tissues still needs to be confirmed. The RasGRP1 protein extracted from primary keratinocyte cultures from newborn mice migrated as a double band on SDS gels, a finding that is consistent with previous observations of RasGRP1 from T-cells of mouse origin (24). Under our experimental conditions, the RasGRP1 doublet was observed only when using the monoclonal antibody raised against full-length rat RasGRP1 (m199) and not with a polyclonal antibody raised against the N terminus of RasGRP1 (data not shown), suggesting that the RasGRP1 doublets may represent a post-translational modification or proteolysis involving the N-terminal region of RasGRP1. It should be noted, however, that the recombinant rat RasGRP1-HA protein showed only one band on blots stained with the m199 antibody. In any case, the C-terminally tagged RasGRP1 proteins (HAand GFP-tagged) were functional, as shown in this study as well as in a previous report in which the C-terminally epitopetagged RasGRP proteins were successfully utilized for in vitro studies (23).
In the study, we found that RasGRP1 was able to respond as a phorbol ester receptor in mouse keratinocytes, as evidenced by our findings on TPA-induced translocation and down-regulation and the direct activation of Ras. The ability of RasGRP1 to respond to phorbol esters depends on the presence of the C1 domain, which has high homology to the C1 domain present in PKC (18). Our earlier studies demonstrated that the C1 domain of RasGRP1 recognizes different diacylglycerol analogs with affinities in the nanomolar range, similar to those shown by some of the PKC isoforms (18), and that ectopically expressed RasGRP1 is able to translocate in response to phorbol esters (16). In mouse keratinocytes, endogenous RasGRP1 also translocated in response to TPA. Using GFP-tagged RasGRP1, we observed that the translocation occurred mainly to the plasma membrane during the first 15 min of exposure to TPA. A similar pattern of translocation has been shown for mouse RasGRP1 ectopically expressed in NIH3T3 cells (25). Interestingly, Ras proteins can signal not only from the plasma membrane, but also from endomembrane compartments (38); and in recent studies, RasGRP1 has been shown to mediate Ras activation in the Golgi (39,40). In this context, it is interesting to note that our previous studies on another member of the Ras-GRP family, RasGRP3, demonstrate different patterns of translocation to the plasma membrane and the Golgi apparatus at different concentrations of TPA (23), suggesting that the pattern of redistribution may be concentration-and/or time-dependent. We are currently investigating whether this is also the case for RasGRP1 redistribution in response to TPA in keratinocytes.
In the case of the PKC isoforms, activation by phorbol esters is usually accompanied by subsequent down-regulation, typically between 3 and 24 h after phorbol ester treatment. We 2 P. S. Lorenzo, unpublished data.
FIG. 6. Activation of ERK in mouse keratinocytes overexpressing RasGRP1. A, keratinocytes were infected with AdLacZ or AdRasGRP1 (HA epitope-tagged) for 48 h and then harvested, and the lysates were subjected to SDS-PAGE, followed by immunoblotting for detection of levels of RasGRP1, cleaved PARP, phospho-ERK (pERK), total ERK (TERK), and actin. B, GTP-loaded Ras (Ras GTP ) levels in keratinocytes expressing either LacZ or RasGRP1 were affinity-precipitated as described under "Experimental Procedures." Total lysates used in the experiment were also subjected to SDS-PAGE, followed by blotting for staining with anti-phospho-ERK, anti-phospho-Ser 338 Raf-1 (pS338-Raf-1), and anti-Ras antibodies. Results are from a representative experiment. Two additional experiments yielded similar results. MOI, multiplicity of infection.
have observed that this is also the case for endogenous Ras-GRP1 in keratinocytes. It should be noted, however, that a previous study has shown that RasGRP1 is not down-regulated by TPA treatment in a murine T-cell clone (24). Divergent responses in different biological systems to phorbol ester-induced translocation and down-regulation have been extensively documented in the literature for the PKC isoforms. For example, although TPA causes PKC␣ and PKC␤I down-regulation in glioma cells, the same treatment of melanoma cells does not lead to changes in the levels of PKC (41). Therefore, the possibility remains that, as with PKC, the differences in RasGRP1 down-regulation in response to phorbol esters may be the result of the different cell types examined.
Our data show that overexpression of RasGRP1 by itself caused activation of Ras in mouse keratinocytes. TPA treatment further increased this level of Ras activation in a PKCindependent manner. However, the same PKC inhibitory treatment (GF109203X) that did not abrogate TPA-induced FIG. 7. Effects of RasGRP1 on calcium-induced differentiation of mouse keratinocytes. A, primary mouse keratinocytes were switched from 0.04 mM Ca 2ϩ to either 0.12 or 1.2 mM Ca 2ϩ for 24 h. After harvesting the cells, total lysates were subjected to SDS-PAGE, blotted, and stained with antibody against RasGRP1 or actin. The upper panels depict microphotographs of keratinocytes exposed to the different concentrations of extracellular calcium, and the lower left panels show representative Western blots. The levels of RasGRP1 were measured by densitometry and converted to percentage of the RasGRP1 levels observed at 0.04 mM Ca 2ϩ . The graph in the lower right panel indicates the mean Ϯ S.E. of four independent experiments. B, primary keratinocytes were infected with either AdLacZ or AdRasGRP1 at a multiplicity of infection of 100 for 48 h, after which the calcium concentration in the medium was increased from 0.04 to either 0.12 or 1.2 mM. After another 24 h, cells were harvested, and total lysates were separated by electrophoresis, blotted, and stained using antibodies against RasGRP1, keratin 10 (K10), keratin 1 (K1), and actin. Results shown in each panel are from a representative experiment. Similar results were obtained in four independent experiments.
Ras activation was effective in reducing TPA-mediated Raf-1 and ERK phosphorylation. This result suggests that, although RasGRP1 can directly activate Ras in response to phorbol esters, PKC is still required for the downstream stimulation of the Raf/MEK/ERK pathway. It also indicates that the point of PKC contribution in the cascade is at the level of Raf-1 because PKC inhibition abrogated TPA-induced phosphorylation of Raf-1 at Ser 338 , a phosphorylation that is indispensable for Raf-1 activity (29). An important emerging concept is that the different families of diacylglycerol/phorbol ester receptors often function together, providing opportunities for synergy or antagonism. Thus, RasGRP3 is phosphorylated by PKC coincident with Ras activation after stimulation in Ramos B cells (42). Conversely, diacylglycerol both activates RasGRP1 and causes the activation and translocation of diacylglycerol kinase, which abrogates the activation of the GRP protein (43,44). Our findings with RasGRP1 in keratinocytes provide another striking example in which both the RasGRP and PKC cascades feed into the same signaling pathway.
In mouse keratinocytes, RasGRP1 overexpression led to apoptosis. This effect appears to be a consequence of increased keratinocyte detachment from the substratum. Consistent with alterations in cell attachment, we observed the disappearance of actin stress fibers in keratinocytes overexpressing RasGRP1. Keratinocyte cell death is a mechanism often associated with cell differentiation; however, our present findings show that ectopic expression of RasGRP1 inhibited two markers of keratinocyte differentiation induced by elevated calcium: keratins 1 and 10. Moreover, high calcium caused a significant reduction in the levels of endogenous RasGRP1. These observation suggest that apoptosis induced by RasGRP1 is not associated with keratinocyte differentiation. In epidermal keratinocytes, activation of Ras pathways has been associated with both proliferative effects (27,32) and cell arrest (45,46). Moreover, although some studies indicate that the introduction of a ras oncogene results in resistance to calcium-mediated differentiation (47), other work shows that oncogenic Ras induces features of differentiation in keratinocytes (45). These reported contradictions appear to be the result of different study models such as cell lines versus primary cells, but particularly of the Ras signal strength because strong and sustained activation of the Ras/ Raf/MEK/ERK cascade seems to trigger growth arrest as a way to counteract uncontrolled mitogenic signaling. A recent study demonstrates that regulated Ras and Raf activation inhibits differentiation and promotes proliferation in mouse epidermis in vivo (27). It would be interesting to test whether different RasGRP1 dosages promote proliferation of keratinocytes rather than apoptosis.
One of the implications of this study is that RasGRP1 represents another target for phorbol esters that may participate in tumor promotion in mouse skin. In chemical skin carcinogenesis, phorbol esters like TPA promote tumor growth by contributing to the expansion of carcinogen-initiated mutant Ras populations of keratinocytes. How can Ras-GRP1 cooperate with Ras mutants in the initiated cells during the tumor promotion events, when Ras mutants are generally unresponsive to exchange factors? One possibility is that RasGRP1 acts on the initiated cells not by activating mutant Ras, but through modulation of the wild-type Ras proteins. Alternatively, RasGRP1 could participate in TPA signaling in the normal population of keratinocytes surrounding the initiated cells rather than in the initiated keratinocytes. In fact, some of the effects of phorbol esters on the proliferation of carcinogen-initiated cell populations have been attributed to growth inhibition of the surrounding normal keratinocytes (48). Further studies utilizing animal models such as the RasGRP1 knockout mouse recently developed (49) should contribute to the elucidation of the role of Ras-GRP1 in TPA-mediated tumor promotion.