|Home | About | Journals | Submit | Contact Us | Français|
15-LOX-1 and its metabolites are involved in colorectal cancer. Recently, we reported that 15-LOX-1 overexpression in HCT-116 human colorectal cancer cells inhibited cell growth by induction of p53 phosphorylation (4). To determine whether the 15-LOX-1 protein or its metabolites are responsible for phosphorylation of p53 in HCT-116 cells, we used HCT-116 cells that expressed a mutant 15-LOX-1. The mutant 15-LOX-1 enzyme, with a substitution of Leu at residue His361, was devoid of enzymatic activity. HCT-116 cells transiently transfected with either native or mutant 15-LOX-1 showed an increase in p53 phosphorylation and an increase in the expression of downstream genes. Thus 15-LOX-1 induces p53 phosphorylation independent of enzymatic activity. Treatment of A549 human lung carcinoma cells with IL-4 increased the expression of 15-LOX-1 and also increased the expression of downstream targets of p53. This confirmed that the activation of p53 was also observed in wild type cells expressing physiological 15-LOX-1. Immunoprecipitation experiments revealed that 15-LOX-1 interacts with, and binds to, DNA-dependent protein kinase (DNA-PK). The binding of 15-LOX-1 to DNA-PK caused an approximate 3.0 fold enhancement in kinase activity, resulting in increased p53 phosphorylation at Ser15. Knockdown of DNA-PK by small interfering RNA (siRNA) significantly reduced p53 phosphorylation. Furthermore, confocal microscopy demonstrated a co-localization of 15-LOX and DNA-PK in the cells. We propose that the 15-LOX-1 protein binds to DNA-PK, increasing its kinase activity, and results in downstream activation of the tumor suppressor p53, thus revealing a new mechanism by which lipoxygenases may influence the phenotype of tumor cells.
Cyclooxygenases (COX) and lipoxygenases (LOX) are two important classes of enzymes that metabolize polyunsaturated acids and influence carcinogenesis. Two isoforms of 15-LOX exist, 15-LOX-1 and 15-LOX-2. The primary arachidonic acid metabolite of 15-LOX is 15-HETE [15(S)-hydroxy-eicosatetraenoic acid], whereas 13(S)-HODE [13(S)-hydroxyoctadecadienoic acid] is the major linoleic acid metabolite formed by 15-LOX. The preferred substrate for 15-LOX-1 is linoleic acid and for 15-LOX-2 it is arachidonic acid (1). The functions of 15-LOX-1 and its metabolites in the development of atherosclerosis, inflammation, and carcinogenesis have been extensively investigated. Recent evidence links 15-LOX-1 to the development or progression of colorectal cancer (2). With a mouse xenograft model, we found that tumors derived from 15-LOX-1 expressing HCT-116 cells were smaller than tumors from vector cells (3). These findings suggest that 15-LOX-1 may act as a tumor suppressor in intestinal cancer. Recently, we reported that 15-LOX-1 overexpression in HCT-116 cells induced an increase in p53 phosphorylation at Ser15. This phosphorylation up-regulates downstream p53 target genes such as p21, MDM2, and nonsteroidal anti-inflammatory drug-activated gene (NAG-1), activates tumor suppression, and leads to inhibition of cell proliferation (4). At least 8 kinases have been identified to induce phosphorylation of p53 at Ser15 (5) including DNA-dependent protein kinase (DNA-PK) (6), ataxia telangiectasia mutated (ATM) kinase (7), and the ataxia telangiectasia and rad-3-related (ATR) kinase (8). Evidence suggests that DNA-PK is a possible target for 15-LOX-1. Here we demonstrate that 15-LOX-1 protein binds to DNA-PK, increases its kinase activity, and induces phosphorylation and activation of p53. This is a unique and novel mechanism for mediating the biological activity of a lipid metabolizing enzyme.
Linoleic acid, arachidonic acid, 13(S)-HODE, and 15(S)-HETE were purchased from Cayman Chemical (Ann Arbor, MI), and Wortmannin and caffeine from Sigma (St. Louis, MO). Antibodies against p53 and actin were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), phospho-p53 (Ser15) from Cell Signaling Technology, Inc. (Danvers, MA), and DNA-PK (Ab-2 antibody) from Calbiochem (EMD Bioscience, Germany). Polyclonal CheY-IgG1 antibody specific for 15-LOX-1 was a generous gift from Dr. Elliot Sigal (9). The BCA protein Assay Kit was from Pierce Biotechnology, Inc. (Rockford, IL). IL-4 was from R&D (Minneapolis, MN)
Human colorectal carcinoma cells, HCT-116, were purchased from ATCC (Manassas, VA) and maintained in McCoy's 5a medium supplemented with 10% fetal bovine serum and penicillin-streptomycin. HCT-116 cells, stably transfected with pcDNA 3.1(+) vector carrying human 15-LOX-1 cDNA, were established as described previously (4). These cells were maintained in McCoy's 5a medium supplemented with 10% fetal bovine serum and 125 µg/ml zeocin (Invitrogen, CA). Human lung carcinoma cell line A549 was purchased from ATCC and grown in Fl2K medium with 10% fetal bovine serum albumin. Cells were treated with 10, 20, or 50ng/ml of IL-4 for 48hrs, then harvested and lysed as described below.
Cell lysates were isolated using RIPA buffer [50 mM Tris-Cl (pH 8.0), 150 mM NaCl, 1% NP40, 0.1% SDS, and 0.5% sodium deoxycholate] containing protease inhibitors (Sigma, St. Louis, MO). Subcellular fractionation was carried out as previously described (10). Briefly, nuclear and cytosolic fractions were prepared using cellular homogenization and differential centrifugation. Nuclei were prepared using a nuclear extract kit (Active Motif, Carlsbad, CA). Protein concentrations were determined with the MicroBCA kit (Pierce, Rockford, IL). Samples containing the same amount of protein were separated by electrophoresis in a 4–12% polyacrylamide gel and transferred onto a nitrocellulose membrane. Membranes were blocked in TBS (Tris buffered saline) and 5% skim milk. Protein bands were probed with the following primary antibodies: CheY-human 15-LOX-1 (dilution 1:5000); phospho-p53 (Ser15) (dilution 1:1000); p53 (dilution 1:3000); DNA-PK (dilution 1:2000) or actin (dilution 1:5000). This was followed by incubation with horseradish peroxidase-conjugated anti-mouse, anti-rabbit, or anti-goat secondary antibody.
The ECL chemilluminescence reagent (Amersham, Piscataway, NJ) was used according to the manufacturer’s instruction to visualize the bands. Samples were compared with 15-LOX-1 and actin standards. The density of protein bands was quantified by Scion Image software.
Mutants of 15-LOX-1 were generated by site-directed mutagenesis using a QuikChange kit (Stratagene). The mutated plasmid was sequenced to confirm the introduction of the mutation and to ensure the absence of non-specific nucleotide changes. Mutant 15-LOX-1 constructs in pcDNA3.1 vector were transfected into HCT-116 cells using LipofectAMINE according to the manufacturer’s protocol (Invitrogen). To obtain stably transfected cells, clones were selected in the presence of zeocin and cultivated in the same medium as wild-type 15-LOX-1 clones. 15-LOX-1 enzymatic activity analyses were performed as previously described (11). Reverse-phase HPLC analysis was performed using an Ultrasphere ODS column (5 mm; 4.6 × 250 mm; Beckman, Philadelphia, PA). All solvents were of HPLC grade and were from Baker (Phillipsburg, NJ). The solvent system consisted of a methanol/water/acetic acid gradient (80/20/0.01, v/v/v) at a flow rate of 1.1 ml/min. Radioactivity was monitored using a Flow Scintillation Analyzer (Packard, Palo Alto, CA) with EcoLume (ICN Biochemicals, Costa Mesa, CA) as the liquid scintillation cocktail. UV analysis was performed by monitoring absorbance at 234 nm with a Waters 486 detector. Authentic standards of 13(S)-HODE (Cayman Chemical, Ann Arbor, MI) were simultaneously injected to monitor retention times. Percent (%) conversion of [14C] LA to its respective products was used to quantify the enzyme activity.
Cells were cultured to 40–50% confluency and then transfected with 100 nM of a pooled mixture of four SMART-selected siRNA duplexes (SMART pool; Dharmacon Research) for DNA-PK cs or a nonspecific control siRNA (Dharmacon Research) using LipofectAMINE 2000 (Invitrogen). The cells were incubated overnight with siRNA and then the media changed to complete media for 48 hours.
For immunoprecipitation (12, 13), whole cell extracts (1000 µg) were precleared with 40 µl protein A-Sephoarose at 4°C for 1 h, then incubated with specific antibodies against 15-LOX-1 or DNA-PK cs at 4°C overnight to capture immunocomplexes. After washing, bead-containing pellets were resuspended in lysis buffer for western blotting analysis or for kinase assays. Kinase assays were performed using the Sigma TECT DNA-PK assay kit according to the manufacturer’s protocol. Specificity of the kinase reaction was confirmed by the DNA-PK inhibitor, Wortmannin.
Laser scanning immunofluorescent confocal microscopy was performed with 15-LOX-1 expressing HCT-116 cells (104cells/ml) on glass cover slips as previously described (14). Cells were fixed in 3% paraformaldehyde for 20 min at room temperature, permeabilized for 10 min in 0.1% Triton X-100, and washed three times with phosphate-buffered saline at pH 7.4 (PBS). Nonspecific binding sites were blocked for 60 min with 0.5% gelatin and 1% rabbit serum in PBS. The fixed cells were co-incubated with primary antibodies for 15-LOX-1 and DNA-PK overnight at 4°C. Cells were washed three times and incubated with the appropriately labeled secondary antibody for 60 min. All washes and antibody dilutions were carried out using PBS containing 0.1% gelatin and 1% rabbit serum. After washing, the cover slips were mounted on glass slides with ProLong Antifade (Molecular Probes) according to the manufacturer's directions to inhibit photobleaching. DNA-PK was identified with the mouse Ab2 antibody, followed by an Alexa 488-conjugated donkey anti-mouse antibody (Invitrogen Corp.). 15-LOX-1 was identified with the CheY-IgG1 antibody, followed by an Alexa 594-conjugated rabbit anti-goat antibody (Invitrogen Corp.). The fluorescent signals were analyzed using an LSM 410 inverted confocal laser-scanning microscope (Carl Zeiss, Oberkochen, Germany).
Previous results from our laboratory demonstrated that the overexpression of 15-lipoxygenase-1 induced growth arrest through phosphorylation of p53 in human colorectal cancer cells (4). However, treatment of 15-LOX-1 expressing HCT-116 cells with linoleic acid or arachidonic acid as a substrate, or with 13(S)-HODE or 15(S)-HETE, did not increase p53 phosphorylation (data not shown). This finding suggested that the activation of p53 may be independent of the enzymatic activity of 15-LOX-1. To test this hypothesis, we created a mutant 15-LOX-1 that was devoid of enzymatic activity. The crystal structure of 15-LOX reveals that the mammalian 15-lipoxygenase is composed of two domains, the small N-terminal β- barrel domain and the large C-terminal catalytic domain that contains the substrate binding pocket and the non-heme iron complex. The mammalian 15-LOX enzyme coordinates the metal atom by an octahedral ligand sphere that is completed by five amino residues (His361, His366, His541, His545, and Ile663) and a water molecule (15, 16). To create the mutant 15-LOX-1, we used site-directed mutagenesis in which His361was replaced with Leu and used this mutant to generate cells expressing an inactive 15-lipoxygenase. Both the wild-type and mutant 15-LOX-1 were expressed in HCT-116 cells, with expression levels of 15-LOX-1 higher in the wild type cells (Fig. 1A). We then measured the enzyme activity in wild-type and mutant 15-LOX-1 expressing cells as described in the Materials and Methods. We found that the H361L mutant was totally inactive (data not shown). To investigate whether the inactive mutant 15-LOX-1 can induce phosphorylation of p53, we measured p53 phosphorylation at Ser15 in cells expressing either the mutant or wild type 15-LOX-1 by western blot analysis. Both the wild type and mutant 15-LOX-1 expressing cells displayed higher phosphorylation of p53 at Ser15 as compared to HCT-116 vector-transfected controls (Fig. 1B). Densitometric analysis (Fig. 1C) indicated that the total 15-LOX-1 protein in mutant expression clones was only 37% that of the wild type expressing clones and, correspondingly, p53 phosphorylation was also reduced when compared to wild type clones. We calculated the ratio of phosphorylated p53 to 15-LOX-1 protein expression and found that the ratios for the mutant 15-LOX-1 expressing clones were approximately the same as in wild type 15-LOX-1 expressing clones (Fig. 1D). These results suggest that the increase in p53 phosphorylation was not dependent on 15-LOX-1 enzymatic activity. To determine if there is a correlation between the expression of 15-LOX-1 and p53 phosphorylation, we transfected various amounts of wild type and mutant 15-LOX-1 expression plasmids and measured p53 phosphorylation. HCT-116 cells were transiently transfected with 0.5 µg to 2.0 µg of wild type and mutant 15-LOX-1 plasmids. As Fig. 2 shows, an increase in both 15-LOX-1 expression and p53 phosphorylation was observed in cells transfected with wild-type and mutant 15-LOX-1 plasmid as compared to cells transfected with empty vector. Transfection with greater amounts of wild type and mutant 15-LOX-1 plasmid did not increase the expression of 15-LOX-1 and p53 phosphorylation (data not shown). There appears to be a correlation between the expression of either the wild type 15-LOX-1 or the mutant 15-LOX-1 with p53 phosphorylation. Because p53 is activated by phosphorylation, we next analyzed the expression of several downstream target genes of p53. NAG-1, p21, and MDM2 are all established genes regulated by p53. As shown in Fig. 2C, the expression of the mutant and native 15-LOX-1 increased the expression of both the pro and mature forms of NAG-1 and p21. These observations strongly support our hypothesis that 15-LOX-1 increases p53 phosphorylation and activity, independent of enzymatic activity of this lipoxygenase. The results suggest that induction of p53 phosphorylation in the 15-LOX-1 expressing HCT-116 cells is the result of the activation of a kinase by the 15-LOX-1 protein.
DNA-PK, ATM, and ATR are members of the phosphatidylinositol-3 kinase-related kinases (PIKK) family. Previously we reported that the inhibitor of ATM and ATR, caffeine, did not alter the 15-LOX-1 induced phosphorylation of p53 (4), suggesting that DNA-PK mediated the 15-LOX-1 effect of p53 phosphorylation. The DNA-dependent protein kinase (DNA-PK) has been shown to phosphorylate human p53 on Ser15 and Ser37 in vitro (17, 18). DNA-PK is a serine/threonine kinase composed of a large catalytic subunit (DNA-PK cs) of ~460 kDa, and a regulatory component, the Ku70 and Ku80 heterodimer (19). We measured DNA-PK expression by Western blot in 15-LOX-1 expressing cells and found that DNA-PK cs had a somewhat higher expression in 15-LOX-1 expressing cells as compared to HCT-116 vector-containing cells. However, Ku70 and Ku80 were detected at similar levels in both 15-LOX –1 expressing cells and vector-containing cells (data not shown). This result was consistent with a previous observation that DNA-PK cs is an induced gene in 15-LOX-1 expressing HCT-116 cells as determined by RNA microarray analysis (4). To determine which kinases are involved in phosphorylation of p53 at Ser15, we incubated different kinase inhibitors with 15-LOX-1 expressing cells. Wortmannin, which is a specific PIKK inhibitor (20), blocked phosphorylation of p53 at Ser15. As shown in Fig. 3A, complete inhibition of phosphorylation was observed with treatment at 10.0 µM. In contrast, caffeine, which is an inhibitor of ATM/ATR kinase, did not inhibit the phosphorylation of p53 at Ser15 (data not shown). The results indicate that DNA-PK is involved in p53 phosphorylation in 15-LOX-1 expressing cells. To further investigate the relationship of DNA-PK and p53 phosphorylation, we used DNA-PK cs siRNA to block the expression of DNA-PK in 15-LOX-1 expressing HCT-116 cells and to examine the effect on phosphorylation of p53 at Ser15. As shown by western blot analysis, 48 h after transfection of 15-LOX-1 expressing HCT-116 cells with DNA-PK siRNA, there was a decrease in the expression of DNA-PK cs and significantly lower phosphorylation of p53 (Fig. 3C, lane3) as compared to corresponding cells transfected with nonspecific siRNA (Fig. 3C, lane2) or cells with transfection media alone (Fig. 3C, lane 1). DNA-PK cs siRNA did not decrease the expression of 15-LOX-1 or total p53 expression in 15-LOX-1 expressing HCT-116 cells (Fig. 3C) as compared to HCT-116 vector cells (Fig. 3B, lane 1, 2 and 3). Hence, it is clear that DNA-PK is involved in phosphorylation of p53 in 15-LOX-1 expressing HCT-116 cells.
We suspected that an interaction of 15-LOX-1 with DNA-PK enhanced the kinase activity resulting in p53 phosphorylation. To look more directly for this interaction, we first immunoprecipitated DNA-PK cs from whole cell lysates with a DNA-PK cs antibody and examined the washed precipitate in the presence of the 15-LOX-1 antibody by western blot. We found that 15-LOX-1 co-immunoprecipitated with DNA-PK cs in 15-LOX-1 expressing cells (Fig. 4A, lane1), but not in HCT-116 vector transfected cells (Fig. 4A, lane 3). In control precipitations with a non-immune antiserum, 15-LOX-1 was not detected from either the 15-LOX-1 expressing cells or HCT-116 vector cells (Fig. 4A, lanes 2 and 4). In a reverse experiment, 15-LOX-1 was immunoprecipitated followed by immunoblotting for DNA-PK. The results from this approach agree with the previous experiment and indicate the association of these two proteins (Fig. 4B).
To determine where in the cells DNA-PK might be associating with 15-LOX-1, we used a double immunostaining technique in transfected HCT-116 cells. Cells were plated on cover slips, permeabilized, incubated with DNA-PK and 15-LOX-1 antibodies, and then fluorescently labeled by secondary antibodies [Alexa Fluor 488-conjugated donkey anti-mouse IgG (green fluorescence) for DNA-PK and Alexa Fluor 594-conjugated goat anti-rabbit antibody (red fluorescence) for 15-LOX-1]. DNA-PK appeared in the nucleus in nearly 100% of the cells as expected. 15-LOX-1 was not detected in HCT-116 vector-transfected cells (data not shown). However, 15-LOX-1 expressing HCT-116 cells showed strong cytoplasmic staining for 15-LOX-1 (Fig. 5A) in the vast majority of cells. Three different coverslips from each of the two experiments were examined and of the 67 cells that displayed 15-LOX-1 expression, 33 showed significant staining in the nucleus in the same areas as DNA-PK (Fig. 5B and 5D), as shown by the yellow color on the merged image (Fig. 5D). To confirm that this location was the nucleus, cells were stained with 4', 6-diamidino-2-phenylindole (DAPI), as shown in Fig. 6A. Treatment of the 15-LOX-1 expressing HCT-116 cells with ionophore A23187 did not induce either the translocation of 15-LOX-1 to the nucleus or an increase of p53 phosphorylation (data not shown). These results indicate that calcium is not a prerequisite for localization of 15-LOX-1 in the nuclei in the 15-LOX-1 expressing HCT-116 cells. Thus, although we do not have unequivocal evidence to conclude that there is a direct interaction between 15-LOX-1 and DNA-PK, both the microscopy and the immunoprecipitation results are consistent, suggesting that 15-LOX-1 and DNA-PK may be part of the same complex within the nuclei of HCT-116 cells.
We hypothesized that when the 15-LOX-1 protein interacts or binds to DNA-PK, it results in higher kinase activity. Thus, quantification of DNA-PK activity was performed. As shown in Fig. 7, an increase in DNA-PK kinase activity (1.6-fold) was observed in 15-LOX-1 expressing whole cell extract lysates. Moreover, a 3-fold increase in kinase activity was detected in immunoprecipitated 15-LOX-1 expressing cell lysates as compared to the kinase activity of HCT-116 vector whole cell lysates. The addition of 10.0 µM of DNA-PK inhibitor Wortmannin abolished the kinase activity in 15-LOX-1 expressing cell lysates. These results are consistent with our hypothesis that the binding of 15-LOX-1 to DNA-PK increases the kinase activity.
The pro-inflammatory cytokine, IL-4, increases the expression of 15-LOX-1 and induces apoptosis in the lung epithelial carcinoma cell line A-549 (21). Because these cells are wild type for p53, this result is supportive of the notion for the activation of p53 by the lipoxygenase. Thus these cells appeared to be a suitable system to determine if activation of p53 by 15-LOX-1. First we confirmed the induction of 15-LOX-1 by IL-4 in these cells and then measured the total p53, phosphorylated p53, and the expression of several downstream target genes of p53. As shown in Figure 8, we observed a concentration dependent increase in phosphorylated p53 and p21 with IL-4 treatment. No change in total p53 expression was observed. The expression of NAG-1 was not measured because it is suppressed by inflammatory cytokines like IL-4 (data not shown). This finding indicates that the activation of p53 by 15-LOX-1 can be observed in cells naturally expressing 15-LOX-1.
The role of 15-LOX-1 in cancer is complex because the biological activity is highly tissue specific with opposing activity observed in different tissues. In human prostate cancer, the expression of 15-LOX-1 is associated with an increase in prostate cancer (22, 23). In contrast, 15-LOX-1 appears to act as a “tumor suppressor” in human colorectal cancer (3, 24, 25). Restoring 15-LOX-1 expression induces apoptosis (26, 27) and inhibits the growth of human colorectal cells (3, 27, 28). Investigations from several laboratories (29, 30, 31) suggest that an increase in the expression of 15-LOX-1 by a number of NSAIDs or COX inhibitors may contribute to the cancer prevention by these drugs. The mechanism by which 15-LOX-1 alters colorectal cancer is not fully understood and several possible mechanisms have been examined. Recently, we reported that the expression of 15-LOX-1 increases the activity of the tumor suppressor p53 by increasing the phosphorylation at Ser15, an important phosphorylation site in the activation of p53 by kinases. One interesting, but unexplained, observation was the inactivity of 13(S)-HODE or 15(S)-HETE, which are the primary metabolites of 15-LOX-1, to increase p53 phosphorylation in HCT-116 cells (4). Incubation of 15-LOX-1 expressing HCT-116 cells with the substrates arachidonic acid or linoleic acid did increase the formation of metabolites, but did not result in an increase in the phosphorylation of p53, suggesting that metabolism is not required for the p53 activation (data not shown). This finding rules out the metabolites of 15-LOX-1 pathway including secondary metabolites, such as 4-oxo-2-nonenal, as responsible for the increase in p53 phosphorylation. 4-oxo-2-nonenal (4-ONE), one of secondary metabolites from hydroperoxide metabolites of 15-LOX-1 (32, 33), can induce phosphorylation of p53 in human neuroblastoma SH-SY5Y cells (34) and we observed a similar response in HCT-116 cells (data not shown).
These data suggest the possibility for the 15-LOX-1 protein to act directly, independent of its enzymatic activity, in the activation of the p53 pathway via an increase in p53 phosphorylation. To test this hypothesis, we prepared a mutant 15-LOX-1 devoid of enzymatic activity. With stably transfected HCT-116 cells, we observed an increase in p53 phosphorylation that correlated with the expression levels of the native and mutant 15-LOX-1 expression. The ratio of p53 phosphorylation to 15-LOX-1 expression in mutant 15-LOX-1 expression clones was approximately the same in wild type 15-LOX-1 expression clones. Secondly, transient transfection of wild type and mutant 15-LOX-1 plasmid into HCT-116 cells increased the phosphorylation of p53 dependent on the level of 15-LOX-1 protein expressions. In addition the expression of 15-LOX-1 increased the expression of p21, NAG-1 and MDM2 all targets of p53. In addition, we observed similar effects in the human lung cells, A-549 which are wild-type for p53. Treatment of these cells with IL-4 induces 15-LOX-1 expression and induces apoptosis (21). The presence of 15-LOX-1 in A-549 cells treated with IL-4 was confirmed and we observed an increase in the expressions of p21, phosphorylated p53 at Ser15, and MDM2. Thus the activation of p53 by 15-LOX-1 was observed in cells not genetically modified that physiologically express 15-LOX-1 and in cells engineered to overexpress 15-LOX-1. These data support the hypothesis that the expression of 15-LOX-1 protein induces p53 phosphorylation independent of its enzymatic activity.
Our previous data suggested that DNA-PK was required for p53 phosphorylation at Ser15 and is responsible for the increase in p53 phosphorylation observed in 15-LOX-1 expressing cells. In this report, DNA-PK siRNA and the DNA-PK inhibitor Wortmannin, confirmed the involvement of DNA-PK in p53 phosphorylation through overexpression of 15-LOX-1. The inhibition of DNA-PK resulted in a reduction in phosphorylation of p53. Incubation of the cells with the siRNA also inhibited the 15-LOX-1-induced increase in p53 phosphorylation. Furthermore, co-immunoprecipitation experiments demonstrated a physical association between DNA-PK and 15-LOX-1, with the binding of 15-LOX-1 to DNA-PK cs resulting in an increase in the kinase activity. Confocal microscopy showed DNA-PK cs and 15-LOX-1 co-localized in 15-LOX-1 expressing HCT-116 cells (Fig. 5). The co-localization of 15-LOX-1 and DNA-PK cs and the enhanced kinase activity observed in vitro suggest that 15-LOX-1 directly activates the kinase. The nuclear localization of 15-LOX-1 (Fig.6) is particularly interesting for several reasons. Nonsteroidal anti-inflammatory drugs (NSAIDs) induced the expression of 15-LOX-1 and stimulated apoptosis. These findings were observed only at high concentrations of metabolites (27). Therefore, nuclear 15-LOX-1 might interact directly with nuclear protein DNA-PK to achieve activation of p53. Although 15-LOX-1 is a cytosolic enzyme, it can associate with the cell membrane (35). For example, the rabbit reticulocyte-type 15-LOX (15-LOX-1) associates with biomembranes in a calcium-dependent manner (36). 15-LOX-1 also has been found both in the cytosol and the plasma membrane in non-stimulated human dendritic cells in the absence of calcium (37). In our laboratory, early studies using human colorectal carcinoma cells demonstrated that 15-LOX-1 was detected in both nuclear and microsomal fractions, with very little in the cytosol in the presence of calcium (10). The mechanism of translocation of 15-LOX-1 is not clear, although the N-terminal C2-like domain in the enzyme is thought to be responsible for membrane binding in a calcium dependent manner (38). In this study, the findings that 15-LOX-1 and DNA-PK co-localized in nuclei in 15-LOX-1 expressing cells raises the question of whether DNA-PK may act as a transporter for the 15-LOX-1. We suspect that the translocation of 15-LOX-1 might be regulated by DNA-PK interaction. However, we could not find any reports in the literature on the binding of 15-LOX-1 with other proteins. Our results clearly demonstrate that 15-LOX-1 directly interacts and binds with the kinase DNA-PK.
The binding of lipoxygenases with other proteins is not without precedent. 5-lipoxygenase is reported to interact with coactosin-like proteins, and this interaction appears to play a modulatory role in actin dynamics (39). Platelet type 12-LOX is distributed in the membrane fraction (40) and may interact with some cytoskeletal proteins such as keractin and lamin (41). However, we could not find any reports in the literature on the binding of 15-LOX-1 with other proteins. Our results clearly demonstrate that 15-LOX-1 interacts and binds with the kinase DNA-PK. The binding of 15-LOX-1 to DNA-PK increases the kinase activity and leads to increased phosphorylation, and presumably, activation of p53. This represents a new and novel mechanism for the activation of a tumor suppressor, p53, by a lipid metabolizing enzyme.
This work was supported, in part, by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences.
We thank Drs. Paul Wade and Robert Langenbach of NIEHS for their careful reading of the manuscript.