Increased Synthesis of Phosphocholine Is Required for UV-Induced AP-1 Activation

Home , Phosphocholine

Oncogene (1998) 17, 1845 ± 1853  1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc Increased synthesis of phosphocholine is required for UV-induced AP-1 activation

Zigang Dong, Chuanshu Huang, Wei-Ya Ma, Barbara Malewicz, Wolfgang J Baumann and Zoltan Kiss

The Hormel Institute, University of Minnesota, Austin, Minnesota 55912, USA

Exposure of mammalian cells to UV irradiation critical in the process of tumor promotion (Alani et al., stimulates phosphatidylcholine hydrolysis and activates 1991; Dong et al., 1994, 1995; Domann et al., 1994), the transcription factor AP-1. Since phosphocholine such epigenetic modi®cations may play an important (PCho), a phospholipid metabolite, is a potential role in UV-irradiation-mediated tumor promotion regulator of mitogenesis and carcinogenesis, we exam- (Shea and Parrish, 1991; Radler-Pohl et al., 1993; ined the eect of UV exposure on the formation of PCho Angel, 1995; Huang et al., 1996, 1997). The UV and the possible mediatory role of PCho in UVB-and enhanced gene expression mediated by transcriptional UVC-induced activation of AP-1 in mouse JB6 induction is known as `UV response' (Angel, 1995). epidermal cells. We found that both UVB and UVC The two most thoroughly studied transcription irradiation resulted in increased PCho levels. Hemi- factors that have been implicated in mediating this cholinium-3 (HC-3), an inhibitor of choline kinase, `UV response' are AP-1 and NF-kB (Angel, 1995; strongly inhibited UV-induced AP-1 activity. By con- Huang et al., 1996, 1997). Exposure of cells to UV light trast, relatively low levels of PCho (80 mM) or choline rapidly activates an Src-family tyrosine kinase, (20 mM) nearly doubled UV-induced AP-1 activity, while followed by sequential activation of the Ha-Ras, the higher (2 ± 20 mM) concentrations of PCho alone cytoplasmic serine-threonine kinase Raf-1, p42/p44 stimulated AP-1 activity 6 ± 8-fold. Importantly, HC-3 mitogen-activated protein kinases (MAP kinases; also inhibited only the stimulatory eect of choline, but not of known as extracellular signal-regulated kinases, or PCho, on AP-1 activity. Of the mitogen-activated ERK-1 and ERK-2, respectively) and AP-1 (Devary protein (MAP) kinases involved in the regulation of et al., 1992; Radler-Pohl et al., 1993). Since both c-Src AP-1 activity, UVC stimulated the MAP kinase family and Ha-Ras are involved in the UV-induced signal ERK-1/ERK-2, JNK as well as p38 kinase activity. transduction pathway, the primary signal generated in These UVC eects were all inhibited by HC-3. With the plasma membrane by UV must be initiated UVB, by contrast, only the activation of ERK-1/ERK-2 upstream from Ras/Raf. In this context, it is was inhibited by HC-3. The data suggest that increased important to note that UV irradiation has also been formation of PCho is required for UV-induced activation shown to enhance phospholipid hydrolysis by both of AP-1 by an ERK-1/ERK-2-dependent mechanism. phospholipase C and D (Carsberg et al., 1995). Based on this, it was suggested that the signaling cascade Keywords: AP-1; UV irradiation; phosphocholine leading to the activation of AP-1 by UV may be phosphatidylcholine; MAP kinase generated at the plasma membrane (Devary et al., 1993; Angel, 1995; Karin and Hunter, 1995; Rosette and Karin, 1996). We have recently demonstrated that atypical protein Introduction kinase C (aPKC) isoforms are required for UV-induced AP-1 activation by using cells transfected with an Exposure to ultraviolet (UV) irradiation accounts for a antisense sequence of mouse PKCz or by a dominant high proportion of human skin cancers (IARC, 1992). negative mutant construct of Xenopus PKC l/i and Radiation in the UVC (180 ± 290 nm), UVB (290 ± cells in which the conventional (a, bI, bII and g) and 320 nm) and UVA (320 ± 400 nm) regions induces novel (d, e, Z and y) PKC isoforms were down- tumor formation in mouse skin (Staberg et al., 1983; regulated by chronic treatment with 12-0-tetradecanoyl Strickland, 1986; IARC, 1992). UV irradiation has been phorbol 13-acetate (TPA) (Huang et al., 1996, 1997). shown to act both as a tumor initiator and a tumor Although the upstream eector of aPKC in the UV promoter (Staberg et al., 1983; Strickland, 1986). It has signal transduction cascade is not clear, some lipids or been proposed that damage of important regulatory their metabolites, such as ceramide, phosphatidic acid genes, such as p53 or ras, plays a role in UV-induced and phosphatidylinositol-3,4,5-P3, are believed to be tumor initiation. UV-induced modulation of signaling responsible for the activation of these enzymes molecules may enhance or decrease gene expression (Nakanishi and Exton, 1992; Nakanishi et al., 1993; leading to increased cell proliferation or apoptosis Dominguez et al., 1993; Berra et al., 1995; MuÈ ller et (Ronai and Weinstein, 1988; Stein et al., 1989; Devary al., 1995). et al., 1991). Since the expression of these genes is Phosphocholine (PCho), an intermediate and product of phospholipid metabolism, has recently been shown to stimulate DNA synthesis in NIH3T3 ®broblasts (Cuadrado et al., 1993; JimeÂ nez et al., 1995; Tomono Correspondence: Z Dong Received 1 December 1997; revised 27 April 1998; accepted 28 April et al., 1995; Kiss, 1996; Chung et al., 1997; Kiss and 1998 Mukherjee, 1997). It has also been reported that PCho Phosphocholine in UV-induced signal transduction ZDonget al 1846 levels are regulated by growth factors (Warden and UVB (3 KJ/m2) was about 30 ± 40% more eective Friedkin, 1985), oncogenes (Macara, 1989; Ratnam and than was UVC (60 J/m2) in stimulating the formation Kent, 1995; Kiss and Crilly, 1995) and chemical of [14C]PCho (Figure 2). Interestingly, HC-3 had a carcinogens (Ishidate et al., 1980; Paulson et al., 1989; somewhat greater inhibitory eect on UVB-induced Kiss and Tomono, 1995). It has further been than on UVC-induced formation of [14C]PCho (Figure demonstrated that most human tumors are character- 2). It should be mentioned here again that the cells ized by elevated levels of PCho (Navon et al., 1977; were irradiated with UVC for 45 s, but with UVB for Daly et al., 1987; Evans and Kaplan, 1997) and that 16 min which may in part account for observed high expression of choline kinase activity is required for dierences. When the incubations were terminated, cancer cell growth (HernaÁ ndez-Alcoceba et al., 1997). only a small portion (*1%) of the newly formed Taken together this may suggest that PCho is involved cellular [14C]PCho was detected in the medium of both in the regulation of cell growth and perhaps in the UV-irradiated and the non-irradiated cultures. The carcinogenesis in vivo. Based on the indicated impor- amount of [14C]PCho released was proportional to the tance of PCho as a growth regulator in cells, and on the cellular [14C]PCho levels formed, indicating that observed eects of UV on phospholipid hydrolysis, we [14C]PCho release from the cells was not stimulated asked the question whether PCho might be a mediator either by UVC or UVB (data not shown). in UV-induced signal transduction leading to the UVC, and particularly UVB, also enhanced the activation of AP-1. incorporation of [14C]choline into the cellular PtdCho In the present study, we used mouse JB6 epidermal pool (Figure 3). As expected, HC-3 strongly inhibited cells to assess the potential role of PCho and choline PtdCho synthesis in both UV-irradiated and non- kinase in UV-induced AP-1 activation. Hemicholinium- irradiated cultures (Figure 3). Since UV induced similar 3 (HC-3) (Cuadrado et al., 1993) was utilized to inhibit increases in [14C]PCho after 30 min, it is presently not choline kinase activity. Changes in water-soluble clear whether increased labeling of PtdCho was due to phospholipid metabolites were followed by radiolabel- increased de novo synthesis or merely re¯ected the ing and by phosphorus-31 NMR. Our data strongly increase in the cellular [14C]PCho pool available for suggest that activation of AP-1 by UV requires PtdCho synthesis. increased PCho formation. While the [14C]choline experiments strongly suggested that choline kinase activity was stimulated by UV irradiation, these experiments did not provide unambiguous information on the total PCho content of Results the irradiated and non-irradiated cells. We therefore used 31P NMR to directly measure the short-term PCho formation in UV-irradiated JB6 cells eects of UVB and UVC on cell PCho levels. Figure 4 To evaluate the possible contribution of PCho to UV- induced signal transduction, we ®rst compared the relative cellular levels of PCho as well as the amounts of PCho released from the cells into the medium in UV-irradiated and non-irradiated JB6 cells incubated in the presence of [14C]choline. Preliminary experiments had shown that radiolabeling can detect rapid eects of UV-irradiation on the formation of [14C]PCho provided the cells are preincubated for 60 min with [14C]choline. Preincubation with 0.4 mM HC-3 for 30 min did not have an eect on cellular [14C]choline levels (data not shown). Since HC-3 is known to inhibit high-anity choline uptake as well as choline kinase (referenced in Cuadrado et al., 1993), we conclude that at the [14C]choline concentration (25 mM) used, choline entered the cells by a predominantly HC-3-insensitive mechanism. Due to the inhibitory eect of HC-3 on choline kinase activity, the control cells contained about 2.7- times higher [14C]PCho levels than did cells treated with HC-3 for 30 min prior to brief (45 s) UVC exposure (60 J/m2; UVC treatment was terminated at 0 time; Figure 1). In the absence of HC-3, but not in its presence, irradiation with UVC resulted in a rapid Figure 1 Eects of UVC and HC-3 on the formation of increase in the formation of [14C]PCho up to 30 min [14C]PCho in JB6 cells. Serum-starved JB6 cells were ®rst (Figure 1). In separate experiments we also found that incubated in the presence of [14C]choline for 1 h, without HC-3 the dierence in the levels of PCho in non-irradiated (open symbols), or with 400 mM HC-3 present during the last 30 min of the labeling period (closed symbols). The plates were versus UVC-irradiated cells did not increase further then removed from the incubator for 2 min during which time when the incubation time was extended up to 1 h and they either were not UV irradiated (*-*) or were UVC (60 J/m2) that UVC-irradiated cells after 12 h actually contained irradiated (~-~) followed by further incubation in the same 14 less [14C]PCho than did control cells. medium for up to 30 min. [ C]PCho levels were measured as When the eects of UVB and UVC were compared described in Materials and methods. Each data point represents the mean+s.d. of three incubations in a single experiment after 30 min of incubation following UV irradiation, representative of two Phosphocholine in UV-induced signal transduction ZDonget al 1847 shows the 31P NMR spectra (70.5 to 1.5 p.p.m. mediating the UV eects on AP-1 activation. We used regions) of the water-soluble phospholipid metabolites the choline kinase inhibitor HC-3, which we had shown from JB6 cells that had been incubated for 5 min after to strongly inhibit UV-induced formation of both PCho irradiation with UVB (3 KJ/m2; Figure 4A) or UVC and PtdCho. We found that HC-3 caused a dose- (60 J/m2; Figure 4B) compared to non-irradiated cells dependent inhibition of AP-1 activity induced by either (Figure 4C). The major resonances were readily UVB (Figure 6A) or UVC (Figure 6B). At the assigned to PCho (0.010 p.p.m.) and inorganic concentrations tested, HC-3 had no eect on cell phosphorus (Pi; 0.495 p.p.m.) based on comparison proliferation. On the other hand, TPA-induced AP-1 with selected phosphate standards (Figure 4D). No activity was not inhibited by HC-3 (Figure 6C). water-soluble phospholipid metabolites other than Similarly, epidermal growth factor-induced AP-1 PCho became major constituents of JB6 cells upon activity remained unchanged at the HC-3 concentra- short-term UV irradiation. tions tested (data not shown). The results indicated that A quantitative comparison of the NMR spectral data the inhibitory eect of HC-3 on UV-induced AP-1 based on integration of the respective 31P resonances activity was not due to a nonspeci®c inhibition of AP-1. revealed that UVB and particularly UVC irradiation increased cell PCho levels within 5 min after UV Choline and PCho also can stimulate AP-1 activity treatment (Figure 5). While UVB irradiation resulted in a more modest increase (13%) in PCho mass, the If PCho or PtdCho, or both, are required for UV eect of UVC irradiation on cell PCho levels was quite induction of AP-1 activity, then supplementation with substantial (27% increase after 5 min; Figure 5). choline, the precursor of PCho and PtdCho, should result in increased AP-1 activity. Moreover, direct treatment of the cells with PCho also should increase HC-3 treatment inhibits UV-induced AP-1 activity AP-1 activity, assuming the cells can incorporate PCho While UV treatment notably enhanced overall cell as was reported (Cuadrado et al., 1993; Kiss, 1996). To PCho levels, even more signi®cant increases may occur further examine these possibilities, we ®rst demon- within speci®c cellular pools that may aect cell growth. strated in separate experiments that PCho can, indeed, Moreover, UV irradiation clearly stimulated PtdCho enter the JB6 cells. For this purpose, starved JB6 cells synthesis which in turn may also contribute to the were incubated for 16 h in fresh serum-free medium in observed cellular UV eects. We therefore further the absence or presence of 2 mM PCho, and the levels examined the possible role of choline metabolites in of water-soluble phosphometabolites in the cells were

Figure 2 Inhibition of UVC- and UVB-induced formation of [14C]PCho by HC-3. Serum-starved JB6 cells were incubated for 1 h in the presence of [14C]choline and in the absence (&)or presence (&) of 400 mM HC-3 which was added for the last 30 min of labeling. The samples were then removed from the incubator for 18 min during which time samples were taken for analysis. Sample I was taken immediately, while the other samples Figure 3 Inhibition of UVC- and UVB-induced formation of (at 228C) remained untreated (Sample II) or were treated with [14C]PtdCho by HC-3. Incorporation of [14C]choline into PtdCho UVC (60 J/m2; Sample III) or with UVB (3 KJ/m2; Sample IV). in the absence (&) or presence of 400 mM HC-3 (&) was Then, the plates (Samples II ± IV) were returned into the determined in the experiment presented in Figure 2. The samples incubator and the cells were further incubated at 378C for were either taken after 1 h of incubation with radiolabeled choline 12 min. This protocol assured that the samples (II ± IV) were kept (Sample I), or after an additional 30 min of incubation without for the same time period (18 min) at 228C and then for 12 min at irradiation (Sample II) or upon irradiation with UVC (Sample 378 in the presence of [14C]choline following the 1 h prelabeling III) or UVB (Sample IV) as described in the legend to Figure 2 period. The data are means+s.d. of six incubations in a single and in Materials and methods. The data are the means+s.d. of experiment representative of two six incubations in a single experiment representative of two Phosphocholine in UV-induced signal transduction ZDonget al 1848 monitored by 31P NMR (Figure 7). A comparison of the spectra does show that the PCho levels of cells grown in PCho supplemented medium were about 1.6- fold higher (Figure 7b) than those in controls (Figure 7a), while the levels of other water-soluble phosphates, such as glycerophosphocholine, glycerophosphoethano-

lamine or Pi, were practically not aected. To further characterize the incorporation of PCho into JB6 cells, we incubated the cells with 2 mM [14C]PCho (450 000 d.p.m./well) for 12 h and found that about 3% and 2% of the total radiolabel was associated with the cellular PCho and PtdCho pools, respectively. In a similar experiment with 2 mM [14C]choline as precursor, 22%, 16% and 2% of the total radiolabel was associated with cell PtdCho, PCho and choline, respectively (data not shown). These data suggest that JB6 cells can incorporate signi®cant amounts of PCho, despite the presence of the phosphate group in the molecule. However, incorporation of PCho into JB6 cells is clearly much less ecient than is incorporation of choline. We then compared the eects of choline and PCho at dierent concentrations on AP-1 activity. For this purpose, JB6 cells were cultured for 24 h in 0.1% serum MEM which contains 7 mM choline. As shown in Figure 8, addition of 20mM choline to the medium indeed stimulated AP-1 activity about twofold, while a further increase in choline concentration up to 20 mM was not accompanied by a further increase in AP-1 Figure 4 Short-term eect of UV irradiation on the levels of activity. By comparison, low concentrations of PCho water-soluble phosphate metabolites in JB6 cells as measured by (up to 200 mM) enhanced AP-1 activity also about phosphorus-31 NMR. The water-soluble phosphate metabolites were extracted from JB6 cells after incubation for 5 min following twofold. However, PCho at 2 and 20 mM concentra- irradiation with UVB (3 KJ/m2; A) or with UVC (60 J/m2; B), or tions stimulated AP-1 activity six- and eightfold, without irradiation (C). The 31P NMR spectrum of a respectively (Figure 8). representative standard mixture containing glycerophosphoetha- nolamine (GroPEtn), phosphoethanolamine (PEtn), glycerophosphocholine (GroPCho) and PCho is shown in D. The water- soluble metabolites were extracted and the spectra were acquired as described in Materials and methods

Figure 6 Eect of HC-3 on UV-stimulated AP-1 activity. Eight6103 JB6 AP-1 reporter stably transfected Cl 41 cells, P+ 1-1, suspended in 5% FBS/MEM were added per well using 96- well plates. After culturing at 378C overnight, the cells were starved by replacing the medium with 0.1% FBS/MEM medium for 12 ± 20 h. The cells were then treated with HC-3 at dierent Figure 5 Quantitative comparison of the short-term eect of concentrations for 30 min and then were irradiated with UVB 2 2 UVB and UVC irradiation on cell PCho levels. JB6 cells were not (3 KJ/m ; A) or UVC (60 J/m ; B), or were treated with 40 nM irradiated ( ) or were irradiated with UVB (3 KJ/m2; )or TPA (C). After 24 h, AP-1 activity was measured using the with UVC (60 J/m2; ), and then incubated for 5 min. Cellular luciferase assay. The results are presented as relative luciferase PCho levels were measured by 31P NMR as described in Materials activity and expressed as means+s.d. of three incubations in a and methods. The data are representative of two determinations single experiment representative of three Phosphocholine in UV-induced signal transduction ZDonget al 1849 In UVB-irradiated cell cultures, extracellular PCho potentiating eect of PCho was not increased at (80 mM) was able to nearly double the stimulatory higher concentrations (data not shown). Interestingly, eect of UV on AP-1 activity (Figure 9). The a similar potentiation of the UVB eect by choline required only 20 mM choline (Figure 9); this is not signi®cantly dierent from the choline concentration which induced AP-1 activity without UV irradiation. Because much lower PCho concentrations were required to potentiate the eect of UV than to stimulate AP-1 activity without UV, we next examined the possibility that UV light may stimulate cellular PCho uptake. However, we found that UV irradiation had no eect on the uptake of [14C]PCho (data not shown). Similarly, in the experiments illustrated in Figures 1 and 2, UV did not signi®cantly alter cellular [14C]choline levels. The observation that, in the micromolar range, choline was more eective in stimulating AP-1 activity than was PCho suggested that either the choline mode of action did not require the conversion of choline to PCho, or that PCho was less eective because of its restricted transport across the plasma membrane. To examine the ®rst possibility, we followed the eects of choline and PCho on AP-1 activity in cells that had been pretreated with HC-3 for 30 min. The data presented in Figure 1 illustrate that pretreatment of Figure 7 Incorporation of PCho into JB6 cells as monitored by the cells with HC-3 for 30 min was sucient to phosphorus-31 NMR. Starved JB6 cells were incubated for 16 h strongly inhibit choline kinase activity. Hence, in fresh serum-free medium in the absence (A) or presence (B)of possible inhibition of HC-3 uptake in the presence of 2mM PCho. The water-soluble cell metabolites were extracted and 31P spectra were acquired as described in Materials and choline or PCho should not aect choline kinase methods inhibition. As is shown in Figure 10, HC-3 at 200 ± 800 mM concentrations inhibited the approximately twofold stimulatory eect of 20 mM choline on AP-1

Figure 8 Induction of AP-1 activity in JB6 cells by choline and Figure 9 Synergistic eects of choline and PCho with UVB PCho. 86103 JB6 AP-1-luciferase stably transfected P+ 1-1 cells irradiation on the induction of AP-1 activity. 86103 JB6 P+ 1-1 suspended in 5% FBS/MEM medium were seeded per well using cells were seeded per well using 96-well plates. After culturing 96-well plates. After incubation at 378C overnight, the cells were overnight at 378C, the cells were starved for 24 h by incubating starved by replacing the medium with 0.1% FBS/MEM for 24 h. them in 0.1% FBS/MEM. Untreated and UVB (3 KJ/m2)- Then the cells were or were not exposed to choline or PCho. After irradiated cells were then further incubated in medium with no incubation for 24 h, AP-1 activity was measured using the other addition (control), or in medium supplemented with 20 mM luciferase assay. Each point represents the mean+s.d. of three choline or 80 mM PCho as indicated. The data are the means+s.d. incubations in a single experiment representative of three of three incubations in a single experiment representative of two Phosphocholine in UV-induced signal transduction ZDonget al 1850 activity by about 50 ± 80%. By contrast, HC-3 even at the highest concentration used inhibited the eect of 20 mM PCho (about an eightfold increase) only by about 12%.

Modulation by HC-3 of UV induced activation of ERKs, Jun N-terminal kinases (JNKs) and P38 kinase ERK-1/ERK-2, JNKs and P38 kinase are members of the MAP kinase family and are thought to serve as upstream kinases responsible for AP-1 activation and phosphorylation of c-Jun protein. We have recently also found (Huang et al., 1997) that, like JNKs, ERK- 1/ERK-2 are important mediators of UV-induced AP-1 activation. To examine which of these enzymes may be involved in UV-induced AP-1 activation mediated by PCho, we studied the eects of HC-3 on UV-induced phosphorylation of ERK-1/ERK-2. As is shown in Figure 11A, both UVB and particularly UVC irradiation greatly enhanced phosphorylation (activation) of ERK-1/ERK-2. In both cases HC-3, particularly at the highest concentration used (200 mM), signi®cantly inhibited the eects of UVB and UVC on ERK-1/ERK-2 activation (P50.01, n=3). Interestingly, while both UVB and UVC enhanced the activities of JNKs (Figure 11B) and p38 kinase (Figure 11C), 200 mM HC-3 inhibited only the Figure 10 Eects of HC-3 on choline- and PCho-induced AP-1 eects of UVC on both enzyme activities. activity. 86103 JB P+ 1-1 cells were seeded per well using 96-well plates. After culturing at 378C overnight, the cells were starved for 24 h by replacing the medium with 0.1% FBS/MEM. Starved Discussion cells were treated with HC-3 for 30 min and then with choline (20 mM) or PCho (20 mM), as indicated, for 24 h and luciferase activity was measured. The data are the means+s.d. of three Although UV-induced modulation of signal transduction clearly plays an important role in UV-induced tumor promotion (Shea and Parrish, 1991; Radler-Pohl et al., 1993; Angel, 1995; Huang et al., 1996, 1997), the prevalent mechanisms are presently not well under- stood (IARC, 1992). UV-induced signal transduction is believed to originate at the cell membrane, perhaps involving increased phospholipid hydrolysis (Carsberg et al., 1995) and is thought to be linked to the nucleus via a signaling cascade involving Src-like tyrosine kinase, Ras, Raf kinase and MAP kinases resulting in the activation of transcription factors such as AP-1,

NF-kB, TCF/ELK, and ATF2 (Devary et al. 1992; Radler-Pohl et al., 1993; Angel, 1995; Karin and Hunter, 1995; Huang et al., 1996, 1997). Evidence in support of this model has mainly been derived from use of pharmacological inhibitors and of dominant negative mutants of Src, Ras and Raf (Devary et al., 1992; Radler-Pohl et al., 1993; Angel, 1995). We recently demonstrated using a mouse PKCx antisense and a dominant negative mutant construct of PKCl/i that atypical PKC (aPKC) is required for UV-induced AP-1 activation (Huang et al., 1996, 1997). In the present study, we provide the ®rst experi- Figure 11 Eects of HC-3 on UV-induced ERK-1/ERK-2, JNK mental evidence that the pathway by which UV and p38 kinase activities. JB6 P+ 1-1 cells were grown in 6-well activates AP-1, which is a prerequisite for tumor plates at 378C for 24 h and the near-con¯uent cells were starved promotion, requires the generation of PCho. Findings for 48 h by replacing the medium with 0.1% FBS/MEM. Four hours prior to UV exposure, the medium was replaced with in support of this mechanism are: (i) HC-3, an serum-free MEM. Then, the cells either treated or not treated inhibitor of choline kinase, inhibited UV-induced AP- with 25 ± 200 mM HC-3 for 30 min, as indicated, were irradiated 1 activation; (ii) PCho and particularly choline, both with UVB (4 KJ/m2) or UVC (60 J/m2). The cells were extracted stimulated AP-1 activity and enhanced the eects of and phosphorylated ERK-1/ERK-2 (A) JNK (B) and p38 (C) UV; (iii) HC-3 inhibited only the eect of choline, but proteins were determined as described in Materials and methods. Protein kinase C-a (PKCa)(D) was included as an internal not of PCho, on AP-1 activation; and (iv) UV control for protein loading irradiation stimulated choline kinase activity which Phosphocholine in UV-induced signal transduction ZDonget al 1851 was associated with higher cellular levels of PCho and PCho content, UV induced a relatively greater increase increased utilization of PCho for PtdCho synthesis. in PtdCho synthesis. Thus, it appears that most of the We had previously shown that normal NIH3T3 PCho formed by UV-stimulated choline kinase was ®broblasts did not incorporate signi®cant amounts of actually utilized for PtdCho synthesis. Hence, increased radiolabeled PCho (Chung et al., 1997).By contrast, PtdCho synthesis may well play a role in mediating the Ha-Ras-transformed NIH3T3 cells and MCF-7 breast UV eects. Alternatively, a relatively large increase in a carcinoma cells were shown to take up PCho from the hypothetical small cellular `signal transduction' PCho incubation medium (Kiss, 1996). Clearly, JB6 cells, pool may contribute to the eects of UV. Clearly, which are initiated cells, resemble transformed cells in further experiments are needed to determine whether that they take up limited amounts of PCho from the activation of choline kinase or increased PtdCho extracellular space. Yet, the uptake of PCho by JB6 synthesis is the most important event in facilitating cells is clearly less ecient than is the uptake of an optimal UV response. choline. This may explain why much higher concentra- It should also be emphasized here that, in addition tions of PCho than choline were required to induce to mitogenesis, PCho may also be involved in the AP-1 activity. On the other hand, both the choline regulation of carcinogenesis. This possibility is strongly uptake mechanism and the choline kinase system to supported by observations that the PCho level in produce PCho have limited capacity. This may explain human cancers is generally high (Evans and Kaplan, why, in the absence of UV irradiation, AP-1 was 1977; Daly et al., 1987; Navon et al., 1997) and that activated to a greater extent by PCho than by choline choline kinase is activated by oncogenes (Macara, at millimolar concentrations. Further evidence in 1989; Ratnam and Kent, 1995; Kiss and Crilly, 1995) support of a mechanism involving choline kinase in as well as by chemical carcinogens (Ishidate et al., UV activation was the observation that much lower 1980; Paulson, 1989; Kiss and Tomono, 1995). These concentrations of PCho were required to enhance AP-1 ®ndings, combined with the observation that AP-1 activity with than without UV irradiation. This activation is an integral part of UV-induced tumor suggests that the amounts of PCho produced by UV promotion (Devary et al., 1992; Radler-Pohl et al., irradiation were not high enough to elicit full AP-1 1993) and that AP-1 activation by UV involves activation. increased PCho formation strongly suggest that PCho It was reported many years ago that growth factors and/or PtdCho indeed play a key role in UV-induced are capable of stimulating the production of PCho carcinogenesis. (Warden and Friedkin, 1985), suggesting that PCho In summary, we have found that in JB6 cells UV may play some role in growth regulation. Indeed, irradiation, choline, and PCho, each alone or in previous studies from our (Tomono et al., 1995; Kiss synergy, can induce the transactivation of AP-1. UV- and Chung, 1996; Chung et al., 1997; Kiss and induced AP-1 activity was correlated with increased Mukherjee, 1997) and from other laboratories PCho and PtdCho synthesis. Consistent with a role of (Cuadrado et al., 1993; Jimenez et al., 1995) have choline metabolites in mediating the UV eects, the shown that PCho has mitogenic activity. Thus, by choline kinase inhibitor HC-3 strongly inhibited UV using various approaches including microinjection of induction of AP-1 activity. Further studies on the PCho into cells, it was possible to demonstrate that in precise mechanism by which PCho contributes to UV- NIH3T3 ®broblasts PCho is required for the mitogenic induced signal transduction should lead to a better activity of platelet-derived growth factor and ®broblast understanding of UV-induced skin degeneration growth factor (Cuadrado et al., 1993; Jimenez et al., associated with aging and cancer. 1995). We also have shown that in NIH3T3 cells and in several other cell lines PCho can greatly potentiate the eects of insulin and ATP (Tomono et al., 1995; Kiss and Chung, 1996; Chung et al., 1997) and of Materials and methods sphingosine-1-phosphate (Chung et al., 1997; Kiss and Mukherjee, 1997) on mitogenesis by both ERK- Materials 1/ERK-2-dependent and -independent mechanisms. In Eagle's minimal essential medium (MEM) and fetal bovine this context, it comes as no surprise that the primary serum (FBS) were purchased from Bio Whittaker Bios- signal transduction pathway mediating the eects of ciences: L-glutamine was from Gibco; gentamicin was UV and PCho on AP-1 activation would involve ERK- bought from Quality Biologicals, Inc.; formamide was 1/ERK-2. from Fluka; the luciferase assay substrate was purchased In NIH3T3 ®broblasts, PCho aects mitogenesis via from Promega; TPA, PCho and choline were bought from an extracellular target (Chung et al., 1997), although in Sigma Co.; and the SAPK/JNK, ERK-1/ERK-2 and p38 special situations intracellular PCho may also be kinase assay kits were from New England Biolabs; HC-3 involved in growth regulation (Cuadrado et al., was from Calbiochem; and [methyl-14C]choline chloride (55 mCi/mmol) was bought from Amersham

1993). In our present system, PCho is likely to act . within the UV-induced cascade via an intracellular target, because the release of PCho from the cells into Cell culture the extracellular space, both with or without UV, was + minimal. On the other hand, JB6 cells incorporated The tumor promotion-sensitive mouse JB6 P epidermal cell line, C1 41, and its AP-1 luciferase reporter some PCho from the medium, which may explain why transfectant, P+1-1, were cultured in an incubator at higher concentrations of added PCho were able to 378C(5%CO2) using MEM supplemented with 5% fetal stimulate AP-1 activity. It is important to emphasize bovine serum (FBS), 2 mML-glutamine and gentamicin that while rapid stimulation of choline kinase by UVB (25 mg/ml) (Dong et al., 1994, 1995; Huang et al., 1996, and UVC led to only small increases in total cellular 1997). Phosphocholine in UV-induced signal transduction ZDonget al 1852 AP-1 activity assay choline were eluted with 18 ml of water and 12.5 ml of 1 M HCl, respectively. Finally, PtdCho was separated from Con¯uent monolayers of JB6 P+ 1-1 cells were trypsinized, other phospholipids by one-directional thin-layer chroma- and 86103 viable cells suspended in 100 mlof5%FBS/ tography on silica gel H plates using chloroform/methanol/ MEM were added into each well of 96-well microplates. 28% ammonia (65 : 25 : 5, by vol) as developing solvent. The plates were incubated at 378C in a humidi®ed

atmosphere of 5% CO2. Twelve to 24 h later, the cells were starved by culturing in 0.1% FBS/MEM for 12 h, and Phosphorylation of choline as measured by phosphorus-31 NMR then exposed to either UVB (3 KJ/m2;15min)orUVC (60 J/m2; 45 s) irradiation. When appropriate, HC-3 was JB6 cells, grown to near con¯uency in 150 cm2 dishes in added to the cells prior to UV irradiation. The cells were 5% FBS/MEM, were starved in 0.1% FBS/MEM for 12 h, then incubated for 24 h to induce AP-1 activity. The cells then incubated in fresh serum-free medium for 4 h, which were extracted with lysis buer, and luciferase activity was was followed by exposure to either UVB (3 KJ/m2)orUVC measured using a luminometer (monolight 2010). The (60 J/m2). After 5 min of incubation at 378Cpost results are expressed as relative AP-1 activity (Huang et irradiation, the cells were quickly washed with 265mlof al., 1996, 1997). chilled 20 mM Tris buered saline (pH 7.3), scraped and then transferred into tubes using a total of 5.5 ml of methanol/chloroform/water (4 : 1 : 0.5, by vol). The extract Phosphorylation of JNK, P38 kinase, ERK-1 and ERK-2 was adjusted by adding 7 ml of chloroform and 2 ml of water to arrive at a Folch mixture. After phase separation, Western immunoblot analyses of the phosphorylated forms the aqueous upper phase was evaporated under a stream of of ERK-1/ERK-2, JNK and P38 kinases were done as N and redissolved in 0.7 ml of D OandthepHwas described (Evans and Kaplan, 1977). Brie¯y, JB6 Cl 41 2 2 adjusted to 1.9 by adding 1 aqueous HCl prior to 31P cells were cultured in monolayers in six-well plates. The M NMR analysis. cells were starved in 0.1% FBS/MEM for 48 h at 378Cand Phosphorus-31 NMR spectra were recorded at then incubated in fresh 0.1% FBS/MEM for another 3 ± 121.42 MHz on a Varian Unity-300 instrument (Varian 4hat378C. Before the cells were exposed to UV, they were Associates, Palo Alto, CA) using a 5 mm variable incubated with or without HC-3 for 30 min. The cells were temperature probe (25.0+0.18C). Standard single-pulse then lysed and immunoblot analysis were done by using experiments entailed an 808 pulse of 8 ms, an acquisition phospho-speci®c antibodies against the phosphorylated time of 3 s and a pulse delay of 8 s with the decoupler gated tyrosine 204 of ERK-1 and ERK-2, the threonine 183/ on during the acquisition only. At a spectral width of tyrosine 185 of JNK and the threonine 180/tyrosine 182 of 4300 Hz, typically 6400 transients were collected. The data P38 kinases. were then zero-®lled and Fourier transformed after applying 0.2 Hz exponential linebroadening. The spectra were refer-

DNA synthesis assay enced relative to 85% H3PO4 serving as external standard (0.00 p.p.m.). Peak assignments were veri®ed by comparison The rate of DNA synthesis was determined using the with authentic standards. Spectral data for quantitative [3H]thymidine incorporation assay as described (Huang et analyses were acquired in the absolute mode and quantita- al., 1996; Dong et al., 1997). tions were made by digital integration of peak areas relative to external quantitative standards. At the 100 nmol level, the accuracy was better than 99%. Formation of [14C]PCho and [14C]PtdCho from [14C]choline in JB6 cells Con¯uent JB6 cell cultures, grown in 12-well plates and kept in 0.1% FBS/MEM for 12 h, were incubated for 1 h in the presence of 25 mM [methyl-14C]choline (1 mCi/ml) Abbreviations with 0.4 mM HC-3 being added for the last 30 min of the UV, ultraviolet irradiation; AP-1, activator protein; PKC, labeling period as speci®ed. Cells were then irradiated with protein kinase C; TPA, 12-O-tetradecanoyl phorbol 13- UVC (60 J/m2)orUVB(3KJ/m2). At the time points acetate; PCho, phosphocholine; ERK, extracellular signal- indicated, the incubation medium (0.5 ml) was transferred regulated kinase; JNK, c-jun N-terminal kinase; MAP, into a tube containing 4.5 ml of chloroform/methanol mitogen activated protein; HC-3, hemicholinium-3; MEM, (1 : 1, v/v) and then 2 ml of ice-cold methanol was added Eagle's minimal essential medium; FBS, fetal bovine into each well. The cells were scraped into methanol, and serum; PtdCho, phosphatidylcholine; NMR, nuclear mag- the methanol extract was transferred into 2 ml of chloro- netic resonance. form. After phase separation [14C]choline and its metabolites were separated on Dowex 50W-packed columns (Bio- Acknowledgements Rad Econo-columns; 1 ml bed volume) as described This work was supported by The Hormel Foundation and previously (Kiss and Anderson, 1994). Brie¯y, the initial by National Institutes of Health Grants AA09292 (to ZK) ¯ow-through (4.5 ml) together with the following 5 ml and CA74916 (to ZD) and BRS Shared Instrumentation water wash contained glycerophosphocholine. PCho and Grant RR04654 (to WJB).

References

Alani R, Brown P, Binetruy B, Dosaker H, Rosenberg RK, Carsberg CJ, Ohanian J and Friedmann PS. (1995). Biochem. Angel P, Karin M and Birrer MJ. (1991). Mol. Cell Biol., J., 305, 471 ± 477. 11, 6286 ± 6295. Chung T, Crilly KS, Anderson WH, Mukherjee JJ and Kiss Angel P. (1995). In Induced Gene Expression (Baeuerle PA, Z. (1997). J. Biol. Chem., 272, 3064 ± 3072. (ed) Birkhauser Boston, Cambridge, MA. pp. 63 ± 92. Cuadrado A, Carnero A, Dol® F, JimeÂ nez B and Lacal JC. Berra E, Diaz-Mecco MT, LozanoÂ J, Frutos S, Municio (1993). Oncogene, 8, 2959 ± 2968. MM, SaÂ nchez P, Sanz L and Moscat J. (1995). EMBO J., Daly PF, Lyon RC, Faustino PJ and Cohen JS. (1987). J. 14, 6157 ± 6163. Biol. Chem., 262, 14875 ± 14878. Phosphocholine in UV-induced signal transduction ZDonget al 1853 Devary Y, Gottlieb RA, Lau LF and Karin M. (1991). Mol. Kiss Z and Anderson WH. (1994). Biochem. J., 300, 751 ± Cell. Biol., 11, 2804 ± 2811. 756. Devary Y, Gottlieb RA, Smeal T and Karin M. (1992). Cell, Kiss Z and Chung T. (1996). Biochem. Biophys. Res. 71, 1081 ± 1091. Commun., 218, 505 ± 509. Devary Y, Rosette C, DiDonato JA and Karin M. (1993). Kiss Z and Crilly KS. (1995). FEBS Lett., 357, 279 ± 282. Science, 261, 1442 ± 1445. Kiss Z and Mukherjee JJ. (1997). FEBS Lett., 412, 197 ± 200. Domann FE, Levy JP, Birrer MJ and Bowden GT. (1994). Kiss Z and Tomono M. (1995). FEBS Lett., 358, 243 ± 246. Cell Growth Di., 5, 9 ± 16. Macara IG. (1989). Mol. Cell. Biol., 9, 325 ± 328. Dominguez I, Sanz L, Arenzana-Seisdedos F, Diaz-Meco MuÈ ller G, Ayoub M, Storz P, Rennecke J, Fabbro D and MT, Virelizier JL and Moscat J. (1993). Mol. Cell. Biol., P®zenmaier K. (1995). EMBO J., 14, 1961 ± 1969. 13, 1290 ± 1295. Nakanishi H and Exton JH. (1992). J. Biol. Chem., 267, Dong Z, Birrer M, Matrisian L and Colburn NH. (1994). 16347 ± 16354. Proc. Natl. Acad. Sci. USA, 91, 609 ± 613. Nakanishi H, Brewer KA and Exton JH. (1993). J. Biol. Dong Z, Huang C, Brown RE and Ma W-Y. (1997). J. Biol. Chem., 268, 13 ± 16. Chem., 272, 9962 ± 9970. Navon G, Ogawa S, Shulman RG and Yamane T. (1977). Dong Z, Watts RG, Sun Y and Colburn NH. (1995). Int. J. Proc. Natl. Acad. Sci. USA, 74, 87 ± 91. Oncol., 7, 359 ± 364. Paulson BK, Porter TJ and Kent C. (1989). Biochim. Evans FE and Kaplan NO. (1977). Proc. Natl. Acad. Sci. Biophys. Acta, 1004, 274 ± 277. USA, 74, 4909 ± 4913. Radler-PohlA,SachsenmaierC,GebelS,AuerH-P,Bruder HernaÁ ndez-Alcoceba R, Saniger L, Campos J, NuÂnezMC, JT, Rapp U, Angel P, Rahmsdorf HJ and Herrlich P. Khaless F, Gallo MA, Espinosa A and Lacal JC. (1997). (1993). EMBO J., 12, 1005 ± 1012. Oncogene, 15, 2289 ± 2301. Ratnam S and Kent C. (1995). Arch. Biochem. Biophys., 323, Huang C, May W-Y, Bowden GT and Dong Z. N (1996). J. 313 ± 322. Biol. Chem., 271, 31262 ± 31268. Ronai Z and Weinstein IB. (1988). J. Virol., 62, 1057 ± 1060. Huang C, Ma W-Y and Dong Z. (1997). Oncogene, 14, Rosette C and Karin M. (1996). Science, 274, 1194 ± 1197. 1945 ± 1954. Shea CR and Parrish JA. (1991). In Physiology, Biochemistry Huang C, Ma W-Y, Hanenberger D, Cleary M, Bowden GT and Molecular Biology of Skin Goldsmith CA, (ed) Oxford and Dong Z. (1997). J. Biol. Chem., 272, 26325 ± 26331. University Press: New York. pp. 910 ± 927. IARC. (1992). IARC Monographs on the Evaluation of Staberg B, Wolf HC, Klemp P, Poulsen T and Bradhagen H. Carcinogenic Risks to Humans, Vol. 55, Oxford Press: (1983). J. Invest. Dermatol., 81, 517 ± 519. Lyon, France. Stein B, Rahmsdorf HJ, Steer A, Lit®n M and Herrlich P. Ishidate K, Tsuruoka M and Nakazawa Y. (1980). Biochem. (1989). Mol. Cell. Biol., 9, 5169 ± 5181. Biophys. Res. Commun., 96, 946 ± 952. Strickland PT. (1986). J. Invest. Dermatol., 87, 272 ± 275. JimeÂ nez B, del Peso L, Montaner S, Esteve P and Lacal JC. Tomono M, Crilly KS and Kiss Z. (1995). Biochem. Biophys. (1995). J. Cell. Biochem., 57, 141 ± 149. Res. Commun., 213, 980 ± 985. Karin M and Hunter T. (1995). Current Biol., 5, 747 ± 757. Warden CH and Friedkin M. (1985). J. Biol. Chem., 260, Kiss Z. (1996). Arch. Biochem. Biophys., 355, 191 ± 196. 6006 ± 6011.