Finkel-Biskis-Reilly Mouse Osteosarcoma Virus v-fosInhibits the Cellular Response to Ionizing Radiation in a Myristoylation-dependent Manner*

DNA damage is recognized as a central component of carcinogenesis. DNA-damaging agents activate a number of signal transduction pathways that lead to repair of the DNA, apoptosis, or cell cycle arrest. It is reasoned that a cell deficient in DNA repair is more likely to acquire other cancer-promoting mutations. Despite the recent interest in the link between DNA damage and carcinogenesis, retroviral oncogenes have not yet been shown to affect the DNA damage-signaling pathway. In this report, we show that Finkel-Biskis-Reilly mouse osteosarcoma virus (FBR) v-fos, the retroviral homologue of the c-fos proto-oncogene, inhibits the cellular response to ionizing radiation. Cells that express FBR v-Fos show a decreased ability to repair DNA damage caused by ionizing radiation, and these cells show decreased survival in response to ionizing radiation. In addition, FBR v-Fos inhibits DNA-dependent protein kinase, a kinase specifically activated upon exposure to ionizing radiation. These effects were specific to ionizing radiation, as no effect of FBR v-Fos on the UV light signaling pathway was seen. Last, these effects were dependent on a lipid modification required for FBR v-Fos tumorigenesis, that of myristoylation of FBR v-Fos. A non-myristoylated mutant FBR v-Fos caused none of these effects. This study suggests that a retroviral oncogene can lead to an increased genomic instability, which can ultimately increase the carcinogenic potential of a cell.

A hallmark of neoplasms is the loss of genomic integrity (1). Since cancer arises from a stepwise progression of mutations in DNA, a loss in genomic integrity increases the likelihood of acquiring cancer promoting mutations. This is highlighted in hereditary non-polyposis colon cancer in which cancer cells have a mutator phenotype caused by a defect in DNA mismatch repair (2,3). In addition, the human papilloma virus oncoproteins, E6 and E7, have been shown to cause the cell to lose its cell cycle checkpoint controls, ultimately leading to genomic instability (1). Despite the heightened interest in the link between genomic instability, DNA repair, and cancer, the only retroviral oncogene that has been shown to affect DNA repair is the HTLV-I tax protein (4). In this paper, we show that FBR 1 v-Fos, the retroviral homologue of the c-fos proto-oncogene, can inhibit cellular signaling and DNA repair in response to ionizing radiation.
Previously our laboratory has been interested in the mechanism by which FBR v-Fos causes transformation. FBR v-fos differs from c-fos both by an N-terminal viral gag sequence and a C-terminal mouse c-Fox sequence (5). FBR v-Fos heterodimerizes with c-Jun, but displays a loss of function at AP-1 sites (6,7). This loss of function can be mapped to the myristoylation of FBR v-Fos. When the myristoylation site is mutated (G2A v-Fos), the non-myristoylated G2A v-Fos regains the ability to transactivate AP-1 sites (6). Cellular transformation is not affected by FBR v-Fos' myristoylation, but myristoylation of FBR v-Fos increases its carcinogenic potential in vivo. Transgenic mice that express G2A v-Fos only develop lipomas, while mice that express FBR v-Fos develop a wide range of mesenchymal tumors, including rhabdomyosarcomas, liposarcomas, chondrosarcomas, and osteosarcomas (8). The mechanism by which FBR v-Fos causes this increased carcinogenesis is unclear.
In addition to its function in AP-1 transactivation, c-Fos has also been shown to be an integral component of the cell's stress response (9). Since DNA damage has been shown to be an essential component of carcinogenesis, we hypothesized that FBR v-Fos might cause a generalized defect in the stress response and DNA repair. In this paper, we show that FBR v-Fos causes a decreased capacity to repair double-strand DNA breaks caused by ionizing radiation, but it does not cause a decreased capacity to repair DNA damage caused by UV radiation. We show that cells that express FBR v-Fos show decreased survival in response to ionizing radiation, but normal survival when exposed to UV radiation. In addition, FBR v-Fos expression causes decreased activity of DNA-dependent protein kinase (DNA-PK), a kinase specifically activated by ionizing radiation. Last, all of these effects are dependent on the myristoylation of FBR v-Fos as cells that express G2A v-Fos do not show any of these effects. These data imply that while both FBR v-Fos and G2A v-Fos cause efficient transformation, only FBR v-Fos leads to a genomic instability, which increases a cell's carcinogenic potential.

MATERIALS AND METHODS
Cell Culture and Transfections-HeLa cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum (Sigma), 2 mM L-glutamine (Sigma) and 1% antimycotic/antibiotic (Sigma). Stably transfected cell lines were created by calcium phosphate precipitation with 18 g of FBR v-Fos or G2A v-Fos and 2 g of RSV-neo (for RSV-neo only stable cell lines, 2 g of RSV-neo and 18 g of pGEM-4Z was used). Cells were then selected for 14 days in medium containing 300 g/ml G418 (Life Technologies, Inc.). Approximately 500 -1000 clones were pooled and shown to express FBR v-Fos or G2A v-Fos by radioimmunoprecipitation and Western blotting.
CAT Assays-DNA repair CAT assays were performed as described previously (9,10,11). Briefly, SV40-CAT was damaged by UV radiation (0, 100, 250, and 400 J/m 2 (emission peak: 254 nm, Stratagene)) or by ionizing radiation (0, 3, 7, 15 Gy ( 137 Cs, 2 Gy/0.7 min)). Activity of these constructs was assayed 12 h later. At 12 h, DNA repair has not taken place, so the activity of the construct is directly related to the amount of DNA damage that has been sustained (9 -11). 12 h after transfection, cell lysis was performed by four alternating cycles of freezing and thawing. Transfection efficiency was then standardized by equal expression of ␤-galactosidase via methods described previously (12). Following standardization, CAT activity was measured as described previously (13). The TLC plate was developed by autoradiography, and counts were measured by a PhosphorImager (Molecular Dynamics). 400 J/m 2 and 15 Gy were judged to be the optimal doses of DNA-damaging agent. These plasmids were then assayed at 48 h after transfection (a time at which DNA repair has taken place) to assay for these constructs' ability to be repaired. CAT activity was indistinguishable between irradiated DNA and unirradiated DNA at 48 h. Cells were transfected with these constructs and 5 g of FBR v-Fos or G2A v-Fos, and CAT activity was assayed 48 h after transfection as described above. Activities were standardized to unirradiated construct alone ϩ expression vector. For instance, FBR v-Fos' effect on repair was determined by standardizing FBR v-Fos ϩ irradiated reporter to FBR v-Fos ϩ unirradiated reporter.
DNA-dependent Protein Kinase Assays-Cells were transfected with 5 or 10 g of c-Fos, FBR v-Fos, or G2A v-Fos expression vector. 48 h later, the cells were washed twice in phosphate-buffered saline and suspended in Buffer P (10 mM HEPES (pH 7.2), 25 mM KCl, 10 mM NaCl, 1.1 mM MgCl 2 , 0.1 mM EDTA). Cells were set on ice for 5 min and then centrifuged at 14 K for 10 min to obtain soluble lysate. Soluble lysate was allowed to sit on ice for 30 min to allow all proteins to mix. Transfection efficiency was standardized by ␤-galactosidase assay, and protein concentration was standardized by Bio-Rad protein assay. Cell extract was added to a peptide corresponding to the consensus DNA-PK phosphorylation site in p53 (PESQEAFADLWKK) as described previously (14). [␥-32 P]ATP was added in 25 mM HEPES (pH 7.4), 70 mM KCl, 10 mM MgCl 2 , 1 mM dithiothreitol, 0.2 mM EGTA, 0.1 mM EDTA, and 0.25 mM ATP to a total volume of 40 l. Double-stranded calf thymus (10 g/ml) DNA was added to one tube, while a second reaction was run in parallel without DNA added. After 30 min at 30°C, the DNA-PK mixture was applied to Whatman No. 3MM paper, washed extensively, allowed to air dry, and scintillation counts were then measured. Specific activity was determined by the following method.
Cell Irradiation-Cell irradiation was performed as described previously (15,16). Briefly, cells were plated at 200 cells/dish and irradiated at the given dose. For UV irradiation, medium was removed before irradiation. The same medium was then added to the dish immediately following irradiation. Cells were then cultured for 10 days. They were stained with Giemsa and colonies were counted. Relative survival was determined by comparing cell survival of cells irradiated at a given does to unirradiated cells plated at the same density and cultured for the same amount of time.

FBR v-Fos Causes Decreased Repair of Double-strand DNA
Breaks-Since the cellular proto-oncogene c-fos has been shown to be an integral component of a cell's stress response (9), and since the stress response is tightly linked to DNA repair and cancer (2, 3), we hypothesized that FBR v-Fos could cause a defect in DNA repair. To test this hypothesis, DNA repair assays were performed (9 -11) using either UV light or ionizing radiation as the insulting agent. UV light causes nicks in DNA, while ionizing radiation causes double-strand DNA breaks. SV40-CAT was damaged with the two sources at various doses, and damage was titrated to obtain the optimal amount. Expression was measured 12 h after transfection. At 12 h, DNA repair has not occurred, so a lack of signal represents unrepaired CAT DNA (9 -11). 400 J/m 2 for UV-treated CAT and 15 Gy for x-ray-treated CAT were judged to be the optimal amounts of damage (Fig. 1A). Assays performed with these plasmids 12 h after transfection showed significant DNA damage (CAT activity ϭ 85% reduction (UV) and 69% reduction (x-ray) from unirradiated SV40-CAT, respectively). Both of these damaged plasmids were able to be repaired effectively as activity of the damaged plasmids matched the activity of the undamaged plasmids 48 h after transfection, a time at which DNA repair has occurred (control, Figs. 1, C and D).
To assess the effects of FBR v-Fos on nicked DNA and double-strand breaks, FBR v-Fos or G2A v-Fos were co-transfected into HeLa cells with the damaged reporter plasmids. CAT assays were performed 48 h after transfection, and activity was standardized to control for the effect of FBR v-Fos or G2A v-Fos on the reporter plasmid. A representative CAT assay is shown in Fig. 1B. When FBR v-Fos was co-transfected with 15 Gytreated reporter, repair of the reporter plasmid was significantly impaired (Fig. 1C). This result is even more dramatic when it is considered that unrepaired reporter (12-h time point, Fig. 1A) had ϳ30% residual activity (residual activity is shown as a solid line in Fig. 1C). In contrast, G2A v-Fos did not impair reporter function (Fig. 1, B and C), indicating that FBR v-Fos' myristoylation plays a significant role in its inhibition of DNA repair. These results were specific to DNA treated with ionizing radiation. When the same assay was performed with reporter plasmid treated with 400 J/m 2 UV, there was no significant difference between reporter plasmid co-transfected with FBR v-Fos or G2A v-Fos (Fig. 1D) despite higher amounts of damage to the 400 J/m 2 -treated reporter (Fig. 1, A and solid  line in E). Thus, FBR v-Fos' effect on DNA repair is specific to double-strand-break repair.
FBR v-Fos Inhibits DNA-dependent Protein Kinase-Since FBR v-Fos causes a defect in a cell's ability to repair doublestrand DNA breaks, we were interested in determining whether FBR v-Fos affected the repair process itself or if FBR v-Fos affected the signaling pathway that leads to repair of double strand DNA breaks. To delineate between these possibilities, we made use of a kinase that becomes activated in response to ionizing radiation. DNA-PK activity has been shown to increase when the cell is exposed to ionizing radiation. Upon activation, DNA-PK has been shown to phosphorylate p53, SP-1, c-Jun, Oct-1, and RNA polymerase II (reviewed in Ref. 21). Cells that lack DNA-PK are hypersensitive to ionizing radiation, and these cells cannot undergo VDJ recombination (a process requiring the resolution of double-strand DNA breaks) (15)(16)(17)(18)(19)(20). We reasoned that if FBR v-Fos inhibited DNA-PK, this would be evidence that FBR v-Fos affects the signaling pathway. On the other hand, if DNA-PK was unaffected by FBR v-Fos, this would indicate that FBR v-Fos affected the repair machinery directly. To test this, FBR v-Fos, c-Fos, or G2A v-Fos were transfected into HeLa cells. 48 h after transfection, whole cell lysates were prepared. Protein concentration and transfection efficiency were standardized. Our transfection efficiency was approximately 10%, and given this transfection efficiency, we would expect to see a result in only 10% of the cells. To correct for this, after cell lysis we allowed the cells to sit on ice for 30 min. This allowed the recombinantly overexpressed FBR v-Fos equal access to all cellular proteins in the lysate. DNA-PK activity was then measured (14). Reactions were run in the presence and absence of DNA to isolate DNAdependent kinase activity (described under "Materials and Methods"). Neither c-Fos nor G2A v-Fos affected endogenous DNA-PK activity ( Fig. 2A). In contrast, FBR v-Fos caused a (cpm disc with DNA added Ϫ cpm disc without DNA) ϫ (1000 pmol/min) incubation time ϫ cpm 5 l of 1.2 mM ATP ϫ mg of total protein added 6 ϫ 10 Ϫ9 mol of ATP/5 l (Eq. 1) 5-fold decrease in DNA-PK activity ( Fig. 2A). Next, HeLa cell lines, which stably expressed RSV-neo, G2A v-Fos, or FBR v-Fos, were constructed. Briefly, cells were transfected with either FBR v-Fos or G2A v-Fos and the neo R gene in a 9:1 ratio. 500 -1000 clones from each respective cell line were pooled. A control cell line containing only neo R was also constructed (RSV-neo cells) The FBR v-Fos and G2A v-Fos stable cell lines expressed the 75-kDa FBR v-Fos or G2A v-Fos protein, and neither of these proteins were present in the RSV-neo only stable cell lines (data not shown). DNA-PK activity was measured in these cell lines. Again, FBR v-Fos causes a significant decrease in DNA-PK activity, and this decrease is again dependent on FBR v-Fos' myristoylation (Fig. 2B).

FBR v-Fos Affects Cell Viability in Response to
Ionizing Radiation-To assess the in vivo effect of FBR v-Fos' on doublestrand break repair, the stable cell lines expressing RSV-neo, FBR v-Fos, and G2A v-Fos were utilized. FBR v-Fos, G2A v-Fos, and RSV-neo cell lines were plated at equal cell densities and exposed to various doses of irradiation from a 137 Cs source or a UV source. Colonies were allowed to grow for 10 days before they were stained and counted. Because we were concerned about the effects that expression of a viral oncogene would have on normal cell growth, FBR v-Fos, G2A v-Fos, and RSV-neo cells' growth were measured relative to themselves. For example, at a given radiation dose, FBR v-Fos cells' survival was measured relative to an unirradiated plate of FBR v-Fos cells plated at the same density and grown for the same amount of time. In response to ionizing radiation, RSV-neo cells show a decrease in cell survival as the radiation dose increases (Fig. 3A), but this decrease is small, indicating that HeLa cells can survive exposure to ionizing radiation. Relative to RSV-neo cells, FBR v-Fos cells show a significant decrease in cell survival at all doses of ionizing radiation (Fig. 3A). G2A v-Fos-expressing cells retain wild-type sensitivity to radiation (Fig. 3A), so FBR v-Fos' effect on cell survival is myristoylationdependent. In contrast to the results seen with ionizing radiation, these cells show no differences in response to UV light (Fig. 3B). Even at doses that cause 90% lethality, no difference in survival is seen. These data support the findings seen with the damaged CAT reporter plasmid and the DNA-PK activity inhibition. FBR v-Fos' effect is specific to ionizing radiation, the process in which DNA-PK has been implicated as performing a critical function. B, representative CAT assay of repair 15 Gy-irradiated CAT. Unirradiated SV40-CAT or 15 Gy-damaged CAT was either transfected alone or was co-transfected with either FBR v-Fos or G2A v-Fos as indicated in the figure. CAT assays were performed at 48 h. In each experiment, ␤-galactosidase expression was used to standardize transfection efficiency. Each experiment was performed three times. C, quantitation of results of Fig. 1B. CAT activities were normalized to the undamaged reporter, the undamaged reporter ϩ FBR v-Fos, or the undamaged reporter ϩ G2A v-Fos and were set to 100%. Normalized percent activities and S.E. values are as follows: undamaged reporter, 100.00; 15 Gy-damaged reporter, 118.33 Ϯ 22.67; undamaged reporter ϩ FBR v-Fos, 100.00; 15 Gy-damaged reporter ϩ FBR v-Fos, 49.83 Ϯ 7.29; undamaged reporter ϩ G2A v-Fos, 100.00; 15 Gy-damaged reporter ϩ G2A v-Fos, 93.97 Ϯ 2.73. The solid line represents residual CAT activity that was present 12 h after transfection (Fig. 6A). D, quantitation of repair of 400 J/m 2 treated CAT. CAT activities were normalized to the undamaged reporter, the undamaged reporter ϩ FBR v-Fos, or the undamaged reporter ϩ G2A v-Fos and were set to 100%. Normalized percent activities and S.E. values are as follows: undamaged reporter, 100.00; 400 J/m 2 -damaged reporter, 90.00 Ϯ 2.89; undamaged reporter ϩ FBR v-Fos, 100.00; 400 J/m 2 -damaged reporter ϩ FBR v-Fos, 86.83 Ϯ 8.07; undamaged reporter ϩ G2A v-Fos, 100.00; 400 J/m 2 -damaged reporter ϩ G2A v-Fos, 82.00 Ϯ 3.93. The solid line represents residual CAT activity that was present 12 h after transfection.
FIG . 2. FBR v-Fos inhibits DNA-dependent protein kinase. A, c-Fos, FBR v-Fos, and G2A v-Fos were transiently transfected into HeLa cells at approximately 30% confluence. After 44 h whole cell lysates were generated and allowed to sit on ice for 30 min to allow the protein product of the transfected gene to mix with all cellular proteins. ␤-Galactosidase was used to standardize transfection efficiency, and protein concentrations were determined. DNA-PK activity was measured by phosphorylation of a peptide corresponding to the DNA-PK target site in p53 in the absence and presence of DNA. Specific activity values in picomoles of 32  to ionizing radiation. Of these kinases, DNA-PK has been shown to be particularly important, as cells that contain mutant, nonfunctional DNA-PK cannot repair double-strand breaks or survive ionizing radiation exposure (15)(16)(17)(18)(19)(20). These activated kinases ultimately affect gene expression as ionizing radiation has been shown to induce a number of transcription factors including c-Jun (29), Egr-1 (27), and NF-B (30). Despite the increasing knowledge of a cell's response to ionizing radiation, no oncogenes have yet been shown to affect this signal transduction network. In this work, we have analyzed the effect of the retroviral oncogene, FBR v-fos, on this pathway. We have found that cells that express FBR v-Fos show a decreased capacity to repair double-strand DNA breaks (Fig. 1). These cells show decreased DNA-PK activity (Fig. 2), and they show a decreased ability to survive exposure to ionizing radiation (Fig. 3). These data imply that FBR v-Fos could be causing a genomic instability in cells in which it is expressed. This could in turn lead to a mutator phenotype, making it more likely that a cell which expresses FBR v-Fos becomes carcinogenic.
FBR v-Fos' role in cancer has been a mystery. It was originally isolated by virtue of its ability to increase the frequency of bone cancers in mice exposed to ionizing radiation (31). Upon discovery of c-Fos and c-Fos' role in AP-1 transactivation, it was assumed that FBR v-Fos mimicked c-Fos' transactivation of AP-1. This was shown not to be true, however, as AP-1 activity has been shown to be neither necessary (6 -8, 32) or sufficient (33) for transformation by Fos. Despite the fact that FBR v-Fos and G2A v-Fos transform tissue culture cells at an equal frequency (8), in vivo, only FBR v-Fos leads to a variety of malignant cancers that ultimately kill the animal (8). Due to the differences in FBR v-Fos' and G2A v-Fos' effects on carcinogenesis and transformation, we favor a two-step model of carcinogenesis by FBR v-Fos. First, transformation must occur. Only the Fos sequences present in both of FBR v-Fos and G2A v-Fos are needed for transformation (8,34), so both G2A v-Fos and FBR v-Fos satisfy this requirement for carcinogenesis. The second step for carcinogenesis is one of genomic instability. An unstable genome allows the cell to obtain a variety of mutations in a short amount of time. This would increase the likelihood of a cell acquiring even more cancer-promoting mutations. Since cells that express FBR v-Fos show decreased ability to repair damaged DNA, and cells that express G2A v-Fos show no defect in DNA repair, only FBR v-Fos satisfies the second requirement. These findings support the observation that both FBR v-Fos and G2A v-Fos can cause transformation in vitro, but only FBR v-Fos causes cancer in vivo (8).
The precise role by which FBR v-Fos causes this effect on the cell's response to ionizing radiation is unclear. FBR v-Fos could be acting as a transcriptional repressor, inhibiting the expression of a gene that is crucial to the x-ray response. FBR v-Fos could also act as a transcriptional activator to increase expression of a gene that antagonizes the x-ray response. Last, FBR v-Fos could have a growth stimulatory effect, pushing the cell through its cell cycle in the face of x-ray-induced DNA damage. Future experiments will help to define FBR v-Fos' role in the ionizing radiation signaling pathway.