G Protein βγ Subunits Augment UVB-induced Apoptosis by Stimulating the Release of Soluble Heparin-binding Epidermal Growth Factor from Human Keratinocytes*

UV radiation induces various cellular responses by regulating the activity of many UV-responsive enzymes, including MAPKs. The βγ subunit of the heterotrimeric GTP-binding protein (Gβγ) was found to mediate UV-induced p38 activation via epidermal growth factor receptor (EGFR). However, it is not known how Gβγ mediates the UVB-induced activation of EGFR, and thus we undertook this study to elucidate the mechanism. Treatment of HaCaT-immortalized human keratinocytes with conditioned medium obtained from UVB-irradiated cells induced the phosphorylations of EGFR, p38, and ERK but not that of JNK. Blockade of heparin-binding EGF-like growth factor (HB-EGF) by neutralizing antibody or CRM197 toxin inhibited the UVB-induced activations of EGFR, p38, and ERK in normal human epidermal keratinocytes and in HaCaT cells. Treatment with HB-EGF also activated EGFR, p38, and ERK. UVB radiation stimulated the processing of pro-HB-EGF and increased the secretion of soluble HB-EGF in medium, which was quantified by immunoblotting and protein staining. In addition, treatment with CRM179 toxin blocked UV-induced apoptosis, but HB-EGF augmented this apoptosis. Moreover, UVB-induced apoptosis was reduced by inhibiting EGFR or p38. The overexpression of Gβ1γ2 increased EGFR-activating activity and soluble HB-EGF content in conditioned medium, but the sequestration of Gβγ by the carboxyl terminus of G protein-coupled receptor kinase 2 (GRK2ct) produced the opposite effect. The activation of Src increased UVB-induced, Gβγ-mediated HB-EGF secretion, but the inhibition of Src blocked that. Overexpression of Gβγ increased UVB-induced apoptosis, and the overexpression of GRK2ct decreased this apoptosis. We conclude that Gβγ mediates UVB-induced human keratinocyte apoptosis by augmenting the ectodomain shedding of HB-EGF, which sequentially activates EGFR and p38.

Heterotrimeric GTP-binding proteins (G proteins) are composed of ␣, ␤, and ␥ subunits and are signal transducers that convert receptor-bound signals into intracellular signals (20). GPCR activated by agonist binding stimulates the exchange of GDP with GTP in the ␣ subunit of G proteins (G␣), and this leads to the activation and dissociation of G␣ from the ␤␥ subunit of G proteins (G␤␥) (21). These activated G␣ and G␤␥ subunits regulate the activities of various effector molecules, including adenylyl cyclases, phospholipases, phosphodiesterases, ion channels, and MAPKs (22,23). G␤␥ subunits were found to mediate UV-induced p38 activation in an EGFR-dependent manner (24), thus demonstrating that G proteins might mediate MAPK activation not only because of GPCR agonists but also because of stressors like UV radiation (25). However, it is unclear how G␤␥ mediates the UVB-induced activations of EGFR and MAPKs. Thus, in this study we investigated the mechanism whereby G␤␥ mediates the UV-induced activations of EGFR and MAPKs and its physiological significance. From this study, G␤␥ was found to mediate UVB-induced ectodomain shedding of HB-EGF to activate EGFR and MAPKs, which resulted in UVB-induced apoptosis in human keratinocytes.

EXPERIMENTAL PROCEDURES
Reagents-Recombinant human HB-EGF was purchased from R&D Systems, and manumycin A, AG1478, PP2, and PP3 were from Calbiochem.
Expression Plasmids-The expression plasmids for the ␤ and ␥2 subunits of G protein and for the carboxyl terminus of GRK2 were subcloned into pcDNA3 expression vector (25). A dominant negative p38 mutant (p38AF) was kindly provided by Dr. Dongeun Park (Seoul National University). The cDNA coding for the Ras binding domain of cRaf-1 (amino acids 1-149) fused to glutathione S-transferase and the Cdc42 and Rac binding domain of p21-activated kinase (amino acids 67-150) fused to glutathione S-transferase were generously supplied by Dr. Walter Kolch (University of Glasgow, UK) and Dr. Jung-Won Lee (Seoul National University), respectively. An active mutant of Src (Src527F) and kinase-negative mutant of Src (Src295M) were presented by Dr. Eung-Gook Kim (Chungbuk National University).
Cell Cultures and Transfection-HaCaT human keratinocytes were grown in Dulbecco's modified minimal essential media containing 10% fetal bovine serum, 100 IU/ml penicillin, and 50 g/ml streptomycin and incubated in a 5% CO 2 incubator at 37°C. Normal human epidermal keratinocytes (NHEKs) were obtained from human foreskin and cultured as described previously (26). Keratinocytes were grown in keratinocyte growth medium (Clontech), and cells in second-passage were used in the experiments. HaCaT cells were transfected by elec-troporation (24) and irradiated with ultraviolet radiation B (UVB, 312 nm, Vilber Lourmat) 72 h after transfection.
Immunoblot and Co-immunoprecipitation Analysis-UVBirradiated cells were harvested and analyzed by Western blotting as described previously (24). The primary antibodies used were as follows: antibody against G␤ (SW) was prepared as described previously (24); antibodies against GRK2, p38, EGFR, and pro-HB-EGF were purchased from Santa Cruz Biotechnology; antibodies against phosphorylated EGFR (Tyr-1068), phosphorylated p38 (Thr-180/Tyr-182), phosphorylated ERK (Thr-202/Tyr-204), phosphorylated JNK (Thr-183/Tyr-185), phosphorylated Src (Tyr-416), and cleaved caspase-3 were from Cell Signaling Technology; antibodies against Cdc42 and Rac1 were from BD Transduction Laboratories; FLAG antibody was from Sigma, and Ras antibody was from Upstate Biotechnology, Inc. Proteins on blots were visualized by incubation with an enhanced chemiluminescence substrate mixture (Pierce) and then exposed to x-ray film (AGFA Curix RPI). Densities of visualized bands were quantified using an image analyzer (Fujifilm, model MultiGauge version 2.3), and relative band densities were expressed as percentages of the corresponding band densities in UVB-irradiated cells. For co-immunoprecipitation, cell lysates were precleared by incubation with protein A-Sepharose beads (Santa Cruz Biotechnology) for 2 h at 4°C. The precleared lysates were incubated with 2 g of antibody against GRK2 overnight at 4°C. The beads were centrifuged, washed three times, and subjected to immunoblot analysis.
Activity Assays of Ras, Cdc42, and Rac1-Ras activity was assayed by analyzing the binding of activated Ras to the Ras binding domain of Raf-1 and the activities of Cdc42 and Rac1 by analyzing the bindings of the activated proteins to the Cdc42/ Rac binding domain of p21-activated kinase (amino acid residues 67-150) fused to glutathione-Sepharose (24).
Analysis of sHB-EGF in Conditioned Medium-The conditioned medium of UVB-irradiated HaCaT cells was collected and lyophilized using a Freeze Dry System (Samwon, Korea) after freezing at Ϫ70°C. The lyophilized medium was solubilized in distilled water and then desalted by ultrafiltration (Centricon 3, Amicon). Desalted samples were separated by SDS-PAGE and analyzed by silver staining (Invitrogen) or by Western blotting using sHB-EGF antibody. Enzyme-linked immunosorbent assay (ELISA) for HB-EGF was developed to measure sHB-EGF. In brief, the conditioned medium (1 ml) from UVB-irradiated HaCaT cells was incubated with heparin beads (Pierce) overnight at 4°C. The reaction mixture was incubated with anti-sHB-EGF antibody (R&D Systems) at room temperature for 2 h and then washed three times with phosphate-buffered saline containing 0.05% Triton X-100. Horseradish peroxidase-conjugated anti-goat IgG was then added and incubated for 1 h. The peroxidase reaction was allowed for 15 min by addition of 3,3Ј,5,5Ј-tetramethylbenzidine substrate solution in H 2 O 2 /citric acid buffer (Pierce). The reaction was stopped by adding 2 M H 2 SO 4 , and the absorbance of supernatant was measured at 450 nm. The concentration of sHB-EGF was calculated from a standard curve and normalized by cell numbers.
Data Analysis-At least four independent experiments were conducted for all analyses. Values are expressed as means Ϯ S.E. The nonparametric Mann-Whitney U test was used to analyze mean values, and p values of Ͻ0.05 were considered statistically significant.

RESULTS
G␤␥ Augmented the Activity That Phosphorylates EGFR, p38, and ERK in the Conditioned Medium from UVB-irradiated Cells-To identify the molecule that mediates the UVB-induced activations of EGFR and MAPKs, conditioned medium from UVB-irradiated HaCaT cells was examined to determine whether it could activate EGFR and MAPKs of nonirradiated cells, because UV irradiation is known to release various growth factors and cytokines that can activate EGFR (16,27). Treatment of nonirradiated HaCaT cells with the conditioned medium obtained from UVB-irradiated HaCaT cells induced the phosphorylations of EGFR, p38, and ERK but not that of JNK, in contrast to UVB, which induced the phosphorylations of all three MAPKs, including JNK (Fig. 1A). This result shows that that the conditioned medium contains molecules that can activate EGFR and p38, indicating that UVB radiation induces the secretion of EGFR-and p38-activating factors into conditioned medium.
To confirm that EGFR-and p38-activating factors in the conditioned medium from UVB-irradiated cells may mediate UVinduced p38 activation, we examined whether p38 activation induced by conditioned medium also requires the activations of EGFR and Ras, because EGFR and Ras were found to mediate UV-induced p38 activation in a previous study (24). Pretreating HaCaT cells with AG-1478, an EGFR inhibitor, or manumycin A, a Ras inhibitor, significantly reduced p38 activation by conditioned medium (53.5 Ϯ 3.1% and 63.8 Ϯ 2.1%, respectively, p ϭ 0.037, Mann-Whitney) (Fig. 1B). This result indicates that EGFR and Ras are also involved in p38 activation by conditioned medium from UVB-irradiated cells, as in p38 activation by UVB irradiation, suggesting the EGFR-and p38-activating factors in the conditioned medium may mediate UV-induced p38 activation.
G protein ␤␥ subunits were found to mediate the UVinduced EGFR and p38 activations in a previous study (24), and thus we examined the effects of G␤␥ on EGFR-and p38-activating activity in conditioned medium obtained from UVB-irradiated cells. The overexpression of G␤ 1 ␥ 2 augmented the activations of EGFR (133.0 Ϯ 14.3%, p ϭ 0.029) and p38 (173.5 Ϯ 13.8%, p ϭ 0.008) in UVB-irradiated HaCaT cells (Fig. 1C), and the conditioned medium from G␤ 1 ␥ 2 -overexpressing cells increased the activations of EGFR (126.6 Ϯ 4.8%, p ϭ 0.029) and p38 (147.7 Ϯ 5.6%, p ϭ 0.029) when added to nonirradiated cells versus the conditioned medium from vector-transfected cells (Fig. 1D). Conversely, the expression of the carboxyl terminus of G protein-coupled receptor kinase 2 (GRK2ct), which sequesters free G␤␥, attenuated UV-induced EGFR (74.2 Ϯ 9.0%, p ϭ 0.029) and p38 activations (65.0 Ϯ 7.2%, p ϭ 0.008), and the conditioned medium from GRK2ct-transfected cells showed reduced activations of EGFR (88.4 Ϯ 3.1%, p ϭ 0.029) and p38 (70.1 Ϯ 3.6%, p ϭ 0.029) compared with that of the vector-trans-fected control (Fig. 1, C and D). The result indicates that G␤␥ augments the activity that phosphorylates EGFR and p38 in the conditioned medium from UVB-irradiated cells. Furthermore, pretreatment of HaCaT cells with pertussis toxin that blocks activation of inhibitory G proteins, including G␣ is , also partially inhibited UVB-induced activations of EGFR (55.1 Ϯ 6.3%, p ϭ 0.029) and p38 (44.1 Ϯ 2.4%, p ϭ 0.029) (Fig. 1E), which confirms that G proteins, including G␤␥ subunits, are involved in UVB-induced EGFR and p38 activation. However, the incomplete inhibition by pertussis toxin suggests that there are other HaCaT cells were irradiated with UVB light (60 mJ/cm 2 ) and incubated for a further 60 min. The cells were then harvested and lysed, and the conditioned medium was harvested and incubated with nonirradiated HaCaT cells for 60 min (A). Cell lysates were separated by 10% SDS-PAGE and analyzed by Western blotting using antibodies against phosphorylated EGFR, p38, ERK, and JNK. Proteins were visualized by incubating blots with ECL substrate mixture and then exposed to x-ray film. Other HaCaT cells were pretreated with 20 M AG1478 for 30 min or with 5 M manumycin A for 2 h before being treated with the conditioned medium (B). HaCaT cells were transfected with G␤ 1 and G␥ 2 or GRK2ct. After 72 h, the cells were irradiated with UVB light, and cells were harvested for Western blot analysis against phosphorylated EGFR and p38 (C). HaCaT cells were transfected with G␤ 1 and G␥ 2 or GRK2ct. After 72 h, the cells were irradiated with UVB light, and the conditioned medium was harvested and incubated with nonirradiated HaCaT cells for 60 min (D). HaCaT cells were pretreated with 200 ng/ml of pertussis toxin for 18 h before being irradiated with UVB light (E). Cells were harvested and lysed for Western blot analysis against phosphorylated EGFR and p38. The blots shown are representative of at least four independent experiments. factors to mediate UVB-induced EGFR and p38 activations besides G proteins.
HB-EGF Induced the Activations of EGFR and p38 via the Same Pathway as Did UV Light-Because UV light was reported to induce ectodomain shedding of pro-HB-EGF and thus the release of sHB-EGF, which is known to phosphorylate EGFR to elicit various cellular responses (16), sHB-EGF was examined to determine its involvement in UVB-induced p38 activation. First, the effects of pretreating with a mutant of diphtheria toxin, CRM197, a HB-EGF specific inhibitor, or with HB-EGF-neutralizing antibody on UVB-induced MAPK activation were examined. Pretreatment of HaCaT cells with CRM197 or HB-EGF-neutralizing antibody inhibited UVB-induced p38 and ERK activations in HaCaT cells ( Fig. 2A), and similar effects of CRM197 were observed in primary cultured NHEKs (Fig.  2B). Furthermore, pretreatment of the conditioned medium with CRM197 toxin attenuated the conditioned mediuminduced activation of EGFR and p38 (Fig. 2C).
To confirm the mediation of UVB-induced MAPK activation by sHB-EGF, we examined the effect of sHB-EGF on MAPK activation. The treatment of HaCaT cells with sHB-EGF resulted in the activations of EGFR, p38, and ERK but not of JNK (Fig. 2D), which corresponded well to the observed effects of conditioned medium from UVB-irradiated cells on the MAPKs (Fig. 1A). Because UV-induced p38 activation was previously found to involve the sequential activations of Ras and Cdc42 (24), sHB-EGF was examined to determine whether it could activate Ras, Cdc42, and Rac1 during the course of p38 activation. In fact, treatment with sHB-EGF induced the activations of Ras and Cdc42 in HaCaT cells (Fig. 2D), which corresponded well to the effect of UVB radiation. However, HB-EGF did not increase Rac1 activity compare with vehicletreated control cells or nonirradiated control cells. This result shows that sHB-EGF activates p38 by activating EGFR, Ras, and Cdc42 as does UVB irradiation and that sHB-EGF mediates the UVB-induced activations of p38 and ERK in NHEK and HaCaT cells, thus suggesting that UVB radiation induces the ectodomain shedding of pro-HB-EGF to mediate the sequential activations of EGFR, p38, and ERK.
UVB Irradiation Stimulated Release of Free G␤␥, Which Augmented Ectodomain Shedding of pro-HB-EGF Induced by UVB Irradiation-Next, to validate that UVB radiation induces ectodomain shedding of pro-HB-EGF, cleavage of pro-HB-EGF after UVB irradiation was analyzed by Western blotting. UVB radiation caused the cleavage of pro-HB-EGF in a dose-dependent manner (30 -180 mJ/cm 2 ) with a concomitant increase in the phosphorylations of EGFR and p38 (Fig. 3A). To prove the secretion of sHB-EGF after UVB irradiation, conditioned medium from UVB-irradiated cells was pooled, concentrated by lyophilization, and then analyzed by SDS-PAGE followed by Western blotting or protein staining. Silver-stained gels showed that the concentrated conditioned medium from UVB-irradiated cells contained more HB-EGF protein than that from mock-irradiated cells (Fig. 3B). The HB-EGF band was identified by applying purified sHB-EGF to gels as a positive control. Western blot analysis also showed that UVB radiation increased the immunoreactivity of HB-EGF in medium (Fig.  3B). A small amount of sHB-EGF was observed in nonirradiated cells, which is speculated to be due to secretion induced by the stress of incubation in serum-free conditions after mock treatment. This result proved that UVB radiation induces the ectodomain shedding of pro-HB-EGF in HaCaT cells.
G␤␥ was found to augment the activity that phosphorylates EGFR and p38 in the conditioned medium, and sHB-EGF shed by UVB was found to mediate UV-induced phosphorylation of EGFR and p38. Therefore, the effect of G␤␥ on UVinduced ectodomain shedding of HB-EGF was examined. The concentrated conditioned medium from UVB-irradiated G␤ 1 ␥ 2 -overexpressing HaCaT cells showed an increase in sHB-EGF protein content to 154.4 Ϯ 14.3% (p ϭ 0.029) versus that from UVB-irradiated vector-transfected cells, and that of GRK2ct-overexpressing cells showed a decrease in sHB-EGF protein to 73.9 Ϯ 6.3% (p ϭ 0.028) (Fig. 3C). This result shows that G␤ 1 ␥ 2 augments ectodomain shedding of pro-HB-EGF induced by UVB irradiation and suggests that G␤ 1 ␥ 2 mediates the activations of EGFR, p38, and ERK by augmenting sHB-EGF secretion following UVB irradiation in HaCaT cells. Moreover, treatment with pertussis toxin inhibited UVBinduced HB-EGF secretion (61.8 Ϯ 4.1%, p ϭ 0.002) (Fig. 3D), which was quantified by ELISA for sHB-EGF. This result supports that G proteins, including G␣ and G␤␥ subunits, are involved in UV-induced ectodomain shedding of HB-EGF, although G protein-independent pathways are also suggested in mediation of sHB-EGF secretion.
Next, the findings that G␤␥ subunits are involved in UV-induced HB-EGF secretion and EGFR activation suggest that UVB irradiation might activate G proteins. Thus, we measured the amount of G␤␥⅐GRK2ct complex as an indicator for G␤␥ activation because only free G␤␥, which is released from G␣ subunit when GPCR is activated, can bind to GRK2 (28). The amount of G␤␥ co-immunoprecipitated with GRK2 was increased by 27.6% (p ϭ 0.029) in UVB-irradiated HaCaT cells compared with nonirradiated cells (Fig. 3E), which means that UVB irradiation activates G␤␥ and thus increase free G␤␥.
G␤␥ Induced Src Activation to Stimulate UVB-induced Ectodomain Shedding of pro-HB-EGF-Src is known to be involved in ectodomain shedding of HB-EGF by various stimuli (29), and thus, the involvement of Src in UVB-induced, G␤␥-mediated HB-EGF secretion was examined. Pretreatment of HaCaT cells with PP2, an Src inhibitor, before UVB irradiation resulted in the blockade of UVB-induced EGFR phosphorylation. However, PP3, known as an analogue of PP2 and used for a negative control of PP2, did not inhibit the UVB-induced EGFR phosphorylation (Fig. 4A). This result suggests that Src is involved in UVB-induced EGFR activation. Then to confirm the involvement of Src in UVB-induced HB-EGF secretion, HaCaT cells were transfected with an activated form (Src527F) or kinase-negative mutant (Src295M) of Src. The conditioned medium from UVB-irradiated Src295M-overexpressing cells almost completely failed to activate p38 of HaCaT cells (Fig.  4B). On the other hand, the conditioned medium from UVBirradiated Src527F-overexpressing cells slightly increased the HaCaT cells were irradiated with various doses of UVB light, incubated for 60 min, and harvested for Western blot analysis using antibodies specific to the cytoplasmic domain of HF-EGF, phosphorylated EGFR, or phosphorylated p38 (A). Conditioned media from UVB-irradiated HaCaT cells were collected and concentrated by lyophilization, and then subjected to 15% SDS-PAGE followed by silver staining or Western blotting analysis using an antibody specific for sHB-EGF (B). HaCaT cells were transfected with G␤ 1 and G␥ 2 or GRK2ct. After 72 h, the cells were irradiated with UVB light. The conditioned media from UVB-irradiated G␤␥ or GRK2ct expressing HaCaT cells were collected and analyzed by Western blotting against sHB-EGF (C). HaCaT cells were pretreated with 200 ng/ml of pertussis toxin for 18 h, then irradiated with UVB light, and incubated for 60 min. Conditioned media from them were collected and subjected to ELISA against sHB-EGF (D). HaCaT cells were irradiated with UVB light, incubated for 60 min, and harvested for co-immunoprecipitation against GRK2 or G␤ as described under "Experimental Procedures" (E). The blots shown are representative of at least 4 -5 independent experiments, and the histograms show average and standard errors. The asterisk means significantly different from the control (p Ͻ 0.05, Mann-Whitney U test).
UVB-induced HB-EGF secretion (Fig. 4B). From these results, it was confirmed that Src is involved in UVB-induced HB-EGF secretion in HaCaT cells.
Overexpression of G␤␥ resulted in the increase of UVB-induced Src activation, and overexpression of GRK2ct resulted in the decrease of UVB-induced Src activation (Fig. 4C). Moreover, PP2 inhibited the UVB-induced, G␤␥-mediated HB-EGF secretion and PP3 did not inhibit the secretion (Fig. 4D). This result shows that Src acts downstream to G␤␥ in mediating UVB-induced HB-EGF secretion.

HB-EGF Mediated UVB-induced Apoptosis in Primary Cultured NHEKs and HaCaT
Cells-To probe the physiological significance UVB-induced MAPK activation mediated by HB-EGF, we examined the effect of HB-EGF on UVB-induced apoptosis. First, when HaCaT cells were irradiated with increasing doses of UVB, caspase-3 cleavage was found to gradually increase in a dose-dependent manner, indicating a gradual increase in apoptosis. The phosphorylations of EGFR and p38 were also increased on increasing UVB dose in a similar manner (Fig. 5A), suggesting that EGFR and p38 might be involved in UVB-induced apoptosis. Thus, to assess the role of HB-EGF on UVB-induced apoptosis, HB-EGF was blocked by pretreating with CRM197. Pretreating HaCaT cells with CRM197 reduced the UVB-induced cleavage of caspase-3 by 32.9% (p ϭ 0.029) versus the UVB-irradiated control, although CRM197 alone without UVB irradiation slightly enhanced caspase-3 cleavage (Fig. 5B). Pretreatment with CRM197 also reduced PARP cleavage by 34.7% (p ϭ 0.029) and annexin V-stained cells by 39.6% (p ϭ 0.029). Similarly, pretreatment with CRM197 also blocked UVB-induced caspase-3 cleavage in pri-mary cultured NHEKs (p ϭ 0.029) (Fig. 5C). In contrast, treatment with sHB-EGF enhanced UVB-induced caspase-3 cleavage by 19.6% (p ϭ 0.008) versus the untreated control, although sHB-EGF alone slightly decreased caspase-3 cleavage (Fig. 5D). Treatment with conditioned medium obtained from UVB-irradiated cells did not increase apoptosis of HaCaT cells either (Fig. 5E). These results show that HB-EGF alone is not enough to induce apoptosis but can augment the UVB-induced apoptosis of NHEKs and HaCaT cells, suggesting that other additional molecules activated by UVB irradiation are involved in the apoptosis.
G protein ␤␥ subunits were found to mediate the UV-induced activations of EGFR and p38 (24), and to augment ectodomain shedding of pro-HB-EGF in this study. Thus, we examined whether G␤␥-augmented activations of EGFR and p38 following UVB irradiation is mediated by sHB-EGF. Overexpression of G␤␥ increased the UV-induced activations of EGFR and p38, but the expression of GRK2ct had the opposite effect. However, pretreatment of HaCaT cells with CRM197, a HB-EGF inhibitor, significantly inhibited G␤ 1 ␥ 2 -augmented UVB-induced EGFR activation from 133.0 Ϯ 14.3% to 60.9 Ϯ 7.5% (p ϭ 0.029) and p38 activation from 173.5 Ϯ 13.8% to 96.2 Ϯ 4.3%, p ϭ 0.008), suggesting that HB-EGF acts downstream of G␤␥ in the mediation of UVB-induced EGFR and p38 activation (Fig. 6B). Similar inhibition of UVB-induced activa- tion of EGFR and p38 by CRM197 was observed in the absence of G␤␥ overexpression, suggesting HB-EGF is involved in the activation of EGFR and p38 in physiological conditions. This result shows that G␤ 1 ␥ 2 augments the UVB-induced activation of EGFR and p38 via sHB-EGF and suggests that G␤ 1 ␥ 2 augments UVB-induced apoptosis by mediating HB-EGF secretion that sequentially activates EGFR and p38 in keratinocytes.
Because UVB is known to activate EGFR, p38, and ERK, their roles in UVB-induced apoptosis were examined in HaCaT cells. Pretreatment with an EGFR kinase-specific inhibitor, AG1478, reduced the UVB-induced cleavage of caspase-3 by 40.9% (p ϭ 0.029) versus the vehicle-treated control (Fig. 6C), and also decreased UVB-induced annexin V staining by 45.4% (p ϭ 0.037, data not shown). Next, to assess the role of p38 activation in regulating UV-induced apoptosis, we examined the effect of dominant negative p38 mutant (p38 AF) on caspase-3 cleavage. The overexpression of p38 AF inhibited UVB-induced caspase-3 cleavage by 27.9% (p ϭ 0.029) (Fig. 6D), but treatment with PD98059 to inhibit ERK activation did not significantly change UVB-induced caspase-3 cleavage (data not shown). This result indicates that activation of EGFR and p38, but not ERK, is involved in the UVB-induced apoptosis of HaCaT cells.

DISCUSSION
This study was performed to elucidate the mechanism whereby G␤␥ mediates the UV-induced activations of EGFR and MAPKs and its physiological significance. The main finding of this study is that G␤␥ mediates the UVB-induced ectodomain shedding of HB-EGF in an Src-dependent way to activate EGFR and MAPKs, which augments UVB-induced apoptosis in keratinocytes.
This main finding is supported first by the result that UVB radiation induced the processing of pro-HB-EGF and the secretion of sHB-EGF into medium, which enabled the conditioned medium to activate EGFR, Ras, p38, and ERK. Second, sHB-EGF activated EGFR, Ras, Cdc42, p38, and ERK in the same way as UVB (24), and blocking HB-EGF reduced the UVB-induced activations of EGFR, p38, and ERK in NHEK and HaCaT cells. Third, UVB irradiation increased free active G␤␥, and G␤␥ augmented UVB-induced ectodomain shedding of HB-EGF. Moreover, blocking HB-EGF inhibited the G␤␥-augmented UVB-induced activations of EGFR and p38. It is widely known that UV irradiation activates MAPKs like ERK, JNK, and p38 and thus regulates the gene expressions required for various cellular responses, including apoptosis and carcinogenesis (9). The activation of MAPKs by HaCaT cells were irradiated with various doses of UVB light as described, incubated for 12 h, and then harvested and subjected to Western blotting analysis using antibodies against cleaved caspase-3, phosphorylated EGFR, and p38 (A). Primary cultured NHEKs or HaCaT cells were pretreated with 10 g/ml of CRM197 toxin or 20 ng/ml of sHB-EGF for 30 min, then irradiated with UVB light (60 mJ/cm 2 ), and incubated for 12 h. Cells were harvested and subjected to Western blotting to follow caspase-3 and PARP cleavage, and to FACS analysis after annexin V staining (B-D). HaCaT cells were incubated with the conditioned medium obtained from UVB-irradiated or nonirradiated cells for 6 h, and then harvested for Western blot analysis against caspase-3 cleavage (E). The blots shown are representative of at least 4 -5 independent experiments, and the histograms show average and standard errors. The asterisk means significantly different from control cells treated with UVB light only (p Ͻ 0.05, Mann-Whitney U test).
UV irradiation involves both the induction of EGFR ligands, including HB-EGF, and the inactivation of phosphatases that would otherwise inactivate the receptor tyrosine kinase (30). UV was reported to induce the ectodomain shedding of pro-HB-EGF in Vero-H cells (16), pterygium-derived epithelial cells (31), and human keratinocytes (32) and the shed sHB-EGF binds to EGFR and activates MAPKs to induce diverse cellular responses (33). The ectodomain shedding of HB-EGF is also induced by various other stimuli, including phorbol esters, calcium ionophore, and GPCR ligands such as lysophosphatidic acid and angiotensin II (16,34). The GPCR-induced ectodomain shedding of HB-EGF attracts much attention because this post-translational modification plays an important role in GPCR-induced EGF receptor transactivation (15,35). These studies suggest the involvement of G␤␥ released from activated heterotrimeric G proteins in the ectodomain shedding of HB- EGF, and hence indirectly support our finding that G␤␥ mediates the UVB-induced activations of p38 and ERK by augmenting the ectodomain shedding of HB-EGF in human keratinocytes. Although no GPCR has been found to be activated by UV radiation, we previously reported that G␤␥ mediates the UV-induced sequential activations of EGFR, Ras, Cdc42, and p38 (24). This study shows for the first time, to the best of our knowledge, that G␤␥ mediates the UVB-induced shedding of pro-HB-EGF to activate EGFR, p38, and ERK in human keratinocytes. Moreover, this study also showed that UV irradiation activates GPCR leading to dissociation of free active G␤␥ subunits from G␣ subunit, which indicates that G proteins may play some physiological roles in UVB-induced EGFR and MAPKs activation. Although a similar involvement of G␤␥ in the ectodomain shedding of HB-EGF was reported in COS-7 cells expressing ␣ 2A -adrenergic receptors, this differs in that it involves the activation of a specific GPCR by a selective ␣ 2A -adrenergic receptor agonist, UK14304, not by nonspecific stress like UVB (33). However, the partial inhibition of UVBinduced activation of p38 by sequestration of G␤␥ with GIRK2ct and by treatment with pertussis toxin or EGFR inhibitor suggests that G␤␥-HB-EGF-EGFR-independent signaling pathways are also involved in the UV-induced activation of p38. This is supported by the reports that UV light secretes not only EGFR ligands but also many other growth factors, including tumor necrosis factor and interleukin-1, that activate p38 (7,36,37).
In addition, the present study shows that G␤␥ mediates UVB-induced HB-EGF shedding in an Src-dependent pathway in HaCaT cells. This finding was supported by results that G␤␥ mediated the UVB-induced Src activation, that Src mediated UVB-induced activation of EFGR and p38, and that Src kinase inhibitor PP2 blocked UVB-induced and G␤␥-mediated HB-EGF secretion. Our findings coincide with recent papers reporting the involvement of Src in HB-EGF processing and EGR activation from the result that PP2 blocked lysophosphatidic acid-induced activation of membrane-type matrix metalloproteinase, pro-HB-EGF shedding, and EGFR transactivation (29) and that Src kinase mediated wound-induced EGFR transactivation (38). Several other molecules, including p38, Pyk2, protein kinase C, and metalloproteinase, have been reported to be involved in the ectodomain shedding of HB-EGF (34), and thus analyses of the effects of G␤␥ on these signaling molecules together with responsible metalloproteinases would help to elucidate the mechanisms involved. Another question that remains to be answered is how UV radiation activates the GPCR that is required to release G␤␥. It is speculated that UVB may activate GPCRs by inducing the secretions of GPCR agonists for autocrine signaling or that it may do so by some agonist-independent mechanism. It is also theoretically possible that an unidentified GPCR may interact specifically with UV radiation and trigger intracellular signaling as rhodopsin does with visible light in the retina. The present study also shows that G␤␥ mediates the UV-induced shedding of HB-EGF to activate p38 and ERK, but not JNK, in human keratinocytes, which suggests that UVB radiation activates JNK via a different mechanism from that of p38 and ERK activation. Our findings are in agreement with a report that ERK and p38 activation by angiotensin II requires EGFR transactivation, whereas JNK activation is regulated without the involvement of EGFR transactivation in vascular smooth muscle cells (39).
This study also shows that G␤␥ augments UVB-induced apoptosis by mediating the ectodomain shedding of HB-EGF, which activates EGFR and p38 sequentially in human keratinocytes. This finding is supported by the following results. First, the overexpression of G␤ 1 ␥ 2 subunits enhanced UVB-induced apoptosis, and the overexpression of GRK2ct inhibited UVBinduced apoptosis. Second, G␤␥ augmented the ectodomain shedding of pro-HB-EGF to trigger sHB-EGF signaling. Third, blocking HB-EGF activity with CRM197 diphtheria toxin protected NHEK and HaCaT cells from UVB-induced apoptosis, and HB-EGF treatment enhanced UVB-induced apoptosis. Fourth, blocking EGFR and p38, downstream of HB-EGF, also reduced UVB-induced keratinocyte apoptosis. The induction of apoptosis by UVB might contribute to the prevention of photocarcinogenesis by removing damaged cells that bear the risk of becoming malignant (40), and the regulation of apoptosis by G proteins suggests that G protein signaling system might modify the photocarcinogenic process, i.e. photocarcinogenesis can be prevented by regulating G protein signaling systems. However, the partial inhibition of UVB-induced apoptosis by GRK2ct overexpression or treatment with CRM197 toxin indicates that UVB induces apoptosis in both G␤␥-/HB-EGFdependent and -independent pathways. HB-EGF has been reported generally to increase cell proliferation and to protect cells against apoptosis, as was shown by its protection of cancer cells from the doxorubicin-induced apoptosis (41). However, the inhibition of HB-EGF by treating with CRM197 toxin did not induce apoptosis, but rather protected hematopoietic 32D cells (42) and human luteinized granulosa cells (43) from apoptosis. The pro-apoptotic effect of HB-EGF was attributed to pro-HB-EGF that showed biological activity quite distinct from that of sHB-EGF, which exerts anti-apoptotic effects. However, our study indicates that sHB-EGF, the release of which is mediated by G␤␥, can enhance UVB-induced apoptosis in human keratinocytes. This finding is supported indirectly by the result that sHB-EGF stimulates p38, which has been reported to translocate Bax to mitochondria, an essential requirement of UVB-induced human keratinocyte apoptosis (44). Furthermore, pro-HB-EGF was reported to stimulate cell survival by interacting with HB-EGF-binding protein (45); thus, it has been speculated that some other factors, like binding proteins, are involved in determining whether HB-EGF promotes or inhibits apoptosis. In addition, the finding that HB-EGF alone does not induce apoptosis implicates that other signaling pathways are required to be activated by UVB for HB-EGF to induce apoptosis. HB-EGF binds to and activates the EGF receptor (ErbB1) and the related receptor tyrosine kinase (ErbB4) (46). Moreover, EGFR activation is an important factor in the control of many fundamental cellular processes, including cell survival and proliferation. EGFR activation usually protects cells against cellular stress and death signals (47), but the activation of EGFR by cellular stressors such as UV irradiation and hypoxic or osmotic stress induces cell death (48,49). This study adds an example of the apoptosis-inducing role of EGFR by showing that the activation of EGFR by HB-EGF following UVB radiation induces human keratinocyte apoptosis. This finding agrees well with a previous finding that the inhibition of EGFR family protects HaCaT cells against UVB-induced apoptosis (48). The present study indicates that the activation of p38 by HB-EGF resulting from G␤␥-mediated ectodomain shedding following UVB irradiation induces apoptosis by showing that the overexpression of dominant negative mutant p38 inhibited UVB-induced apoptosis in keratinocytes. Our finding is supported by a recent report that p38 MAPK mediated the UVB-induced apoptosis of human keratinocytes by translocating Bax to mitochondria (44). Moreover, this pro-apoptotic role of p38 is supported by previous reports that found that impaired p38 activity (by inhibitors or dominant negative p38 mutants) enhanced resistance to UVB-induced apoptosis in various cells (50,51). On the other hand, p38 was reported to play a critical role in the survival of UVA-irradiated HaCaT cells, because of the finding that a blockade of p38 activation increased UVA-induced apoptosis (52). They reported that UVA light did not induce significant apoptosis in HaCaT cells under their experimental conditions, but we observed that UVB light induced obvious apoptosis. The opposite effect of p38 on UV-induced apoptosis may have resulted from differences in the wavelength and intensity of the UV light used, because UV light is known to induce different cellular signalings and biological responses at different wavelengths and intensities (9). In addition, because unlike UVA, UVB light causes sustained p38 activation in keratinocytes (53), a difference in the duration of p38 activation induced by UVA and UVB light might contribute to the different effect of p38 on UV-induced keratinocyte apoptosis.
In summary, this study shows that G protein ␤␥ subunits augment UVB-induced human keratinocyte apoptosis by mediating the ectodomain shedding of HB-EGF in an Src-dependent manner, which sequentially activates EGFR and p38. These findings provide new insights into the mechanism of MAPK activations by UVB light, and of the physiological response to the UVB-induced ectodomain shedding of HB-EGF in keratinocytes. In particular, the importance of G protein signaling is emphasized in the regulation of UV responses, including keratinocyte apoptosis, which might be involved in the photocarcinogenic process.