- Email: [email protected]
Oxy5, a novel protein from Arabidopsis thaliana, protects mammalian cells from oxidative stress
The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602 www.elsevier.com/locate/ijbcb
Oxy5, a novel protein from Arabidopsis thaliana, protects mammalian cells from oxidative stress Anil Kush a, Kanaga Sabapathy b,* b
a Indo-American Hybrid Seeds, P.O. Box 7099, BSKII Bangalore 560070, India National Cancer Center, Di6ision of Cellular and Molecular Research, 11 Hospital Dri6e, Singapore 169610, Singapore
Received 19 July 2000; accepted 9 February 2001
Abstract The use of molecular oxygen in various cellular processes results in the generation of toxic intracellular by-products termed reactive oxygen species (ROS). In Escherichia coli the oxyR gene product is a transcriptional regulator of the oxyR regulon that is induced in response to hydrogen peroxide-induced oxidative stress (OS). We have previously shown that an annexin-like protein from Arabidopsis thaliana, termed Oxy5, can replace the obligatory role of OxyR in E. coli. Here, we have investigated as to whether oxy5 can function across evolutionary boundaries to protect mammalian cells from OS. Overexpression of the oxyR gene in mammalian tumor cell lines protects them from hydrogen peroxide-induced cell death, and these cells are also highly resistant to the superoxide ion producing compound paraquat. Oxy5 appears to be involved in the detection of calcium flux, as it binds to Ca2 + ions during hydrogen peroxide stress. Moreover, overexpression of Oxy5 leads to lowered protein kinase C activity. Thus, Oxy5 probably functions to sense and initiate protective responses to OS. In addition, Oxy5-overexpressing cells exhibit a reduction in endogenous superoxide ion levels, which concomitantly results in a dramatic decrease in their tumorigenic potential. Taken together, the results demonstrate an antioxidant role for plant oxy5 gene in mammalian cells, which can be potentially utilized in gene therapy programs aimed at reducing the deleterious effects of ROS. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: Annexin; Cell death; Oxidative stress; Oxy5; Reactive oxygen species; Superoxide ions
1. Introduction Oxidative stress (OS), resulting from the detrimental effects of reactive oxygen species (ROS), is an important phenomenon in many biological * Corresponding author. Tel.: + 65-436-8349; fax: +65226-5694. E-mail address: [email protected] (K. Sabapathy).
systems. ROS include superoxide ions (O− 2 ), hydroxyl radicals (OH) and hydrogen peroxide (H2O2). They are involved in lipid peroxidation, protein denaturation and DNA damage [1,2]. These cellular damages can lead to the development of various pathological symptoms observed both through apoptosis and necrosis, in diseases like arthritis, muscular dystrophy and various genetic and nervous disorders like Down’s syn-
1357-2725/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 1 3 5 7 - 2 7 2 5 ( 0 1 ) 0 0 0 4 0 - 1
592
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
drome and amylotrophic lateral sclerosis [1– 3]. Moreover, accumulating evidence suggests that increased cellular concentrations of active or reactive oxygen, organic peroxides and free radicals are important in tumor promotion [4– 6]. The mechanism by which mammalian cells counteract OS is at present not fully understood. In general, ROS are controlled through a protective network referred to as the antioxidant defense, which includes enzymes such as manganese superoxide dismutase (MnSOD), copper zinc superoxide dismutase (CuZnSOD), catalase and glutathione peroxidases. Overexpression of MnSOD has been shown to partially rescue cells from TNF-induced, ROS-mediated cell death [7]. In addition, Bcl-2 was shown to prevent apoptosis by scavenging ROS [8,9]. However, Bcl-2 could prevent cell death by a mechanism other than by its antioxidant function, as it also promotes cell survival in anaerobic conditions and cells that lack mitochondrial DNA [10,11]. In most of the above studies only a partial protective effect was observed, suggesting that other yet unidentified proteins might play an important role in this process. In Escherichia coli, the oxyR gene product is induced upon exposure to OS, which in turn activates the expression of defense genes [12]. The OxyR protein plays a dual role both in sensing and transducing peroxide stress signals [13,14]. By a functional complementation screen, we have previously identified an annexin-like protein from Arabidopsis thaliana, termed Oxy5, which could restore the ability of an E. coli oxyR deletion mutant to grow in the presence of H2O2 [15]. Moreover, we have demonstrated that Oxy5 was upregulated during OS in plants [15]. Oxy5 has high homology with the annexin supergene family and displays high catalase activity [15]. In the present study, we have investigated if this plant gene could serve an antioxidant function across evolutionary boundaries. We found that overexpression of Oxy5 protects mammalian cells against OS-induced cell death. Its mode of action is discussed.
2. Materials and methods
2.1. Cell lines, DNA and antibodies K36 is a leukemia cell line derived from the AKR/J mice [16] and CHO cells were obtained from ATCC. Cells were cultured in 10% serum containing RPMI media. An EcoR1 –Xho1 (1.2 kb) full-length oxy5 cDNA was cloned into pcDNA3 vector in both the sense and antisense orientations (Invitrogen) as described [15]. Both K36 and CHO cells were transfected with oxy5 or antisense-oxy5 expression plasmids and geneticin (G418)-resistant cells were selected. At least two independent pools of cells were used in all experiments. Polyclonal antibodies to Oxy5 were raised in rabbits as described [15].
2.2. RNA and protein analysis Total RNA was extracted from about 106 cells (control or oxy5 -transfected). The oligomer (5%TGG CGA CTC TTA AGG TTT CTG ATT CTG TTC 3%) corresponding to the N-terminus of A. thaliana oxy5 was used as a probe for mRNA analysis by standard Northern blotting techniques. For Ca2 + binding analysis, cells were incubated with 1 mCi of 45Ca2 + ml − 1 for 8 h before H2O2 treatment, lysed and immunoprecipitated using polyclonal antibodies against Oxy5, followed by 10% polyacrylamide SDSPAGE analysis and autoradiography. Levels of Oxy5 were determined by immunoprecipitation of cells labeled with 200 mCi of 35S-methionine for 2 h.
2.3. Cell 6iability assays Viability of cells treated with or without the indicated concentrations of H2O2 was analyzed by their ability to exclude propidium iodide. Percentage of viable cells (cells able to exclude propidium iodide, and therefore not stained for propidium iodide) was determined using FACScan with LYSYS II software. For studies with EGTA or nifedipine, cells were treated with either 1 mM EGTA or 10 mM nifedipine for 6 and 2 h, respectively, before H2O2 treatment
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
and cell viability was analyzed. For PMA studies, cells were treated with 100 nM of staurosporine or 250 nM of PMA for 3 and 2 h, respectively, prior to the H2O2 treatment for 1 h or 15 min, respectively. For O− 2 -induced stress, cells were treated with 200 mM of paraquat for three days and viability determined as described above.
2.4. Protein kinase C acti6ity assay Protein kinase C (PKC) activity was determined using an Amersham kit (RPN-77). Measurement of enzyme activity was based on PKC-mediated transfer of the [gamma-32P] ATP to a synthetic peptide homologous to the PKC-o pseudosubstrate domain in the presence of cofactors, as per the manufacturer’s instructions. PKC activity of control K36 cells treated H2O2 was taken as 100% PKC activity, and the other values are represented as a percentage of this value.
2.5. Measurement of O− 2 production by 4 -MFU-glu A Beckman LS2 fluorescence spectrophotometer was employed to determine the amount of O− 2 released by a fluorometric assay. The cell suspensions were adjusted to a volume of 3 ml (5× 106 cells ml − 1) in PBS containing 8 mM 4-methylbeta-D-umbelliferyl glucopyranoside (4-MUFglu). The assay estimates the amount of superoxide anion radical released, which then cleaves the 4-MUF-glu to 4-MUF, based on the liberation of 4-methylumbelliferone (4-MUF). Control or oxy5 -transfected cells were incubated at room temperature for an hour and the fluorescence intensity of the supernatant was measured, as described [17].
2.6. Analysis of tumor growth 1× 106 control or oxy5 -transfected K36 or CHO cells were injected subcutaneously in AKR/J mice or intraperitonially in Balb/c nude mice, respectively, and tumor growth was monitored over the indicated periods of time. Twenty independent mice of each type were used in each group in the study.
593
3. Results
3.1. Expression of Oxy5 protects mammalian cells from hydrogen peroxide-induced cell death To ascertain if Oxy5 affects cellular survival in response to OS, we transfected stably a murine leukemia cell line (K36) and an hamster ovary cell line (CHO) with the full length (1.2 kb) oxy5 cDNA or a pcDNA3 vector expressing the oxy5 gene in the antisense orientation (hereinafter referred to as control cells). Cells were selected for G418 resistance and three independently derived pools were obtained. We have used pools of G418-resistant cells to circumvent the problem of clonal variability. In all experiments, at least two independent pools were tested. Both K36 and CHO cells transfected with oxy5 expressed high levels of oxy5 mRNA (Fig. 1a), and protein (Fig. 2c), in contrast to control cells. Although Oxy5 protein shares a high degree of homology to annexin protein superfamily and especially to human annexin 1 [15], neither the cDNA for oxy5 nor the antibody to Oxy5 cross-reacted with any endogenous mRNA or protein from the control K36 and CHO cells (Fig. 1a, Fig. 2c). We next determined the optimal dose of H2O2, and the duration of treatment that results in substantial cell death of the control cells. Treatment of K36 and CHO cells for 6 h with various concentrations of H2O2 showed that most of the cells were dead at a high dose of 10 mM H2O2, whereas there was little cell death at both 100 mM and 1 mM of H2O2 (Fig. 1b(i) for K36 cells; data not shown for CHO cells). This indicates that these tumor cells are rather resistant to H2O2 stress and requires high levels of H2O2 to succumb to cell death. Therefore, we tested for the optimal duration of treatment with 10 mM of H2O2 at which about 60–75% of cell was killed. At 2 h after H2O2 treatment, most of the cells were dead (Fig. 1b(ii)), whereas about 77% of cells were killed at 1 h after exposure to H2O2. At shorter periods of treatment, between 65 (0.5 h) (Fig. 1b(ii)) and 80% (0.25 h) (Fig. 3b) cells were alive. Hence, we chose a treatment regimen of 10 mM H2O2 for 1 h for all our studies. When exposed to 10 mM H2O2 for 1 h, the Oxy5-overexpressing cells (both
594
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
Fig. 1. Oxy5-overexpressing cells are resistant to H2O2-induced cell death. (a) oxy5 mRNA expression in control and oxy5 -transfected K36 (K) and CHO (C) cells. gapdh expression was analyzed for loading control. (b) Determination of optimal doses (i) and duration (ii) of treatment with H2O2 in inducing cell death in K36 cells. Cell viability was determined as described in Section 2. (c) Control or oxy5 -overexpressing K36 (i) and CHO (ii) cells were treated with 10 mM H2O2 for 1 h and viability was determined. ( +) and ( −) indicates with or without H2O2 treatment, respectively. Vertical bars indicate standard deviation.
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
595
Fig. 2. Ca2 + and H2O2 stress. Cell survival rates were measured with prior treatment with 1 mM EGTA (for 6 h) (a) or 10 mM nifedipine for 2 h (b), followed by H2O2 stress, as described above. (c) Binding of 45Ca2 + to Oxy5 upon H2O2stress. Cells were labeled with 45Ca2 + (upper panel) or 35S-methionine (lower panel) as described in Section 2, lysed and immunoprecipitated with anti-Oxy5 antibody. Upper panel indicates amount of 45Ca2 + bound to Oxy5 and lower panel indicates levels of Oxy5 immunoprecipitated. ( + ) and ( −) indicate with or without H2O2 treatment, respectively.
596
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
Fig. 3. (Continued)
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
K36 and CHO) exhibited a survival rate of 60 and 70%, as opposed to 20 and 18% for the control K36 and CHO cells, respectively (Fig. 1c). Moreover, almost none of the Oxy5-overexpressing cells were dead when treated for 15 min as compared to about 20% death of control DNA transfected cells (Fig. 3b). These results were reproducible with independent pools of both the cell types tested over at least three separate experiments. Thus, these data indicate that overexpression of Oxy5 confers resistance to an otherwise lethal dose of H2O2, and suggests that Oxy5 can participate in the cellular responses to OS in mammalian cells.
3.2. Oxy5 acts as a calcium chelator during H2O2 -induced OS Since H2O2 stress has been shown to induce changes in calcium homeostasis [18,19] and members of the annexin superfamily are known to be calcium-binding proteins [20], we investigated if oxy5 which is highly homologous to the annexin superfamily of genes is causally involved in regulating changes in calcium influx during OS. Pretreatment of control K36 cells with the Ca2 + -chelator EGTA (1 mM for 6 h) or the Ca2 + -channel blocker nifedipine (10 mM for 2 h) rendered these cells more than twice as resistant to H2O2 than untreated cells (Fig. 2a and b). This resistance could be reversed by the addition of exogenous Ca2 + (data not shown). Similar results were obtained with control CHO cells (data not shown). In addition, the small percentage of Oxy5-overexpressing cells that were susceptible to H2O2 treatment was almost completely rescued in the presence of EGTA and nifedipine (Fig. 2a and b). These observations indicate that H2O2-induced cell death requires influx of exogenous Ca2 + and that inhibition of the Ca2 + influx can protect cells from OS-induced cell death. To determine if Oxy5 activity is associated with the Ca2 + influx, cells
597
were cultured in 45Ca2 + supplemented media for 8 h and the H2O2 stress experiments were performed as described earlier. Cells were lysed and immunoprecipitated with polyclonal antibodies against Oxy5 to determine the amount of 45Ca2 + bound to Oxy5. We found significantly higher binding of 45Ca2 + to Oxy5 in Oxy5-overexpressing K36 cells exposed to H2O2, compared to untreated cells (Fig. 2c, upper panel). Similar results were obtained with CHO cells (data not shown). There was no 45Ca2 + binding observed in control cells (Fig. 2c, upper panel), as the antibody did not recognize any endogenous protein (Fig. 2c, lower panel). It thus appears that Oxy5 is involved in the quenching of Ca2 + flux resulting from H2O2 stress, and thereby probably functions as a Ca2 + chelator to sense and initiate a protective response to OS.
3.3. PKC acti6ity is reduced in Oxy5 -o6erexpressing cells It is well established that both H2O2 stress and Ca2 + influx can induce the activity of various serine threonine kinases, including PKC [21– 23]. Therefore, we examined if the resistance conferred against cell death by Oxy5 is a consequence of the modulation of PKC activity. To start with, we pre-treated K36 and CHO cells with the PKC inhibitors staurosporine and H7 to evaluate their effect on cell death. Cell death (both control and oxy5 -transfected cells) due to H2O2 stress was significantly reduced in the presence of the PKC inhibitors (Fig. 3a; and data not shown for H7). Conversely, pre-treatment with the PKC-activator phorbol meristyl acetate (PMA) resulted in a significant acceleration of both the rate and extent of H2O2-induced cell (Fig. 3b). In this experiment, cell death was determined 15 min after H2O2 treatment, a time-point at which only about 20% cells were dead, as this would allow for the analysis of the effect of PMA on augmenting H2O2-in-
Fig. 3. PKC inactivation results in resistance to H2O2-induced cell death. K36 cells were pre-treated with either 100 nM staurosporine for 3 h (a) or 250 nM PMA for 2 h (b), followed by H2O2 stress [1 h in (a) and 15 min in (b)], and cell viability was determined. (c) PKC activity in control or oxy5 -transfected K36 cells without ( −) or with ( + ) H2O2 treatment, as described in legends to Fig. 1. Value of control cells treated with H2O2 was taken as 100% and other values are represented as a percentage of this value.
598
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
duced cell death (Fig. 1b(ii), Fig. 3b). More than 70% of control K36 cells pre-treated with PMA were dead only after 15 min of exposure to 10 mM H2O2, in contrast to about 20% cell death in the absence of the PMA pre-treatment (Fig. 3b). There was no significant amount of cell death induced by PMA treatment alone (Fig. 3b). Similar results were obtained with CHO cells (data not shown). These results suggest that Ca2 + -dependent PKC-like kinase might participate in effecting cell death from OS and the inhibition of its activity can avert the lethality. Strikingly, although the potentiation of H2O2-induced cell death by PMA was markedly reduced in Oxy5overexpressing cells (Fig. 3b), these cells were not completely resistant to H2O2-induced cell death. They underwent cell death to the same extent as the control cells not pre-treated with PMA (Fig. 3b). These results raised the possibility that Oxy5 expression might influence PKC activity and that Oxy5 acts upstream of PKC in the signaling cascade. We therefore investigated whether PKC activity is affected in cells overexpressing Oxy5. When extracts of H2O2-treated cells were analyzed, Oxy5-overexpressing cells were found to have about 25% lower levels of PKC activity compared with control cells (Fig. 3c). Taken together, the data suggests that Oxy5 acts by binding Ca2 + and probably quenches the calcium signaling cascade resulting in lowered PKC activity, which is required to transduce the cell death signal.
3.4. Reduction in superoxide production by Oxy5 -o6erexpressing cells Elevated PKC-like activity has been shown to increase the production of O− 2 [24]. We therefore examined if Oxy5-overexpressing cells, with lower − PKC activity, produce less O− 2 . O2 released in the supernatant was analyzed by a fluorometric assay using 4-MUF-glu that specifically reacts with O− 2 , as described in Section 2 [17]. In brief, this technique exploits the ability of superoxide to cleave the fluorescent substrate 4-MUF-glu to the highly fluorescent 4-MUF, which is detected flurometrically. K36 and CHO cells overexpressing Oxy5 contained between 40 and 50% less O− levels 2
Fig. 4. Reduced superoxide production by Oxy5-overexpressing cells: (a) levels of O− 2 ions were determined in both control and Oxy5-overexpressing K36 and CHO cells, using 4-MUFglu as a substrate, as described in Section 2; (b) survival rate of control or oxy5 -transfected K36 cells was determined in the presence of 200 mM paraquat (O− 2 ions producing herbicide) for three days.
compared to control cells (Fig. 4a). These results indicate that Oxy5 regulates the levels of intracellular superoxide anions, even in the absence of external OS stimuli. We thus reasoned that this would also result in resistance to external O− 2 insults. To test this possibility, cells were cultured in the absence or presence of the O− 2 producing herbicide paraquat to determine their viability. Oxy5-overexpressing cells showed an 80% survival rate compared with 20% survival of the control cells (Fig. 4b). These data indicate that Oxy5
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
expression confers resistance against O− 2 as well as H2O2 on cells, and can function in response to OS.
3.5. Oxy5 and tumorigenicity High levels of O− 2 production have been correlated with increased tumorigenicity [4] and we have previously demonstrated that altering the levels of O− 2 production by K36 tumor cells can reduce their tumorigenic potential [5]. We therefore examined if reduction of O− 2 production by overexpression of Oxy5 would indirectly affect their tumorigenic potential. 1×106 control or oxy5 -transfected K36 or CHO cells were injected subcutaneously in AKR/J mice or intraperitonially in Balb/c nude mice, respectively, and tumor growth was monitored in 20 independent mice of each type. Control K36 cells rapidly formed non-regressing tumors with almost 100% efficiency at the sites of inoculation and all mice developed maximum sized tumors by five weeks (Fig. 5a). The Oxy5-overexpressing K36 cells did form tumors initially, although at a much slower rate. However, these tumors did not develop further up to five weeks, and some of them disappeared by eight weeks (Fig. 5a and c(i)). Similarly, control CHO cells rapidly formed non-regressing tumors in nude mice with 100% efficiency. Tumors were noticed after five to seven days of inoculation and they kept on growing and reached a maximum size by 25 days (Fig. 5b and c(ii)). In contrast, Oxy5-overexpressing CHO cells did not form tumors in nude mice even after 25 days of inoculation (Fig. 5b and c(ii)).
4. Discussion We have demonstrated that the A. thaliana protein Oxy5, when overexpressed, could function as a sensor of the Ca2 + flux induced by an OS signal and subsequently serve as an effector of the protective cellular response leading to resistance against OS. Overexpression of Oxy5 results in lowered PKC levels and decreased production of O− 2 , which correlates with increased resistance to OS. These data demonstrate a direct correlation
599
between the intracellular oxidative state and the resistance to extracellular OS insults. Together, these studies suggest a potential role for plant Oxy5 in regulating intracellular oxidative processes in mammalian cells. Overexpression of Oxy5 also resulted in the suppression of tumor formation in vivo. At present, it is unclear how Oxy5-overexpressing cells, being more resistant to H2O2 in vitro, have a lower tumorigenic potential in vivo. Overexpression of Oxy5 results in two distinct effects: resistance to H2O2 results and decreased production of O− 2 . It could be envisaged that Oxy5 might effect both these processed independently. Decreased O− 2 production has been shown to result in decreased tumorigenicity, as we have previously demonstrated that treatment of mice injected with the K36 leukemia cells with superoxide dismutase (SOD), a superoxide scavenger, resulted in reduced tumorigenicity [5]. This correlates with the reduced tumorigenicity of Oxy5-overexpressing cells. Thus, the tumorigenic potential of cells could be altered by modulating the levels of O− 2 ions. It has been previously shown that cell lines rendered resistant to OS by overexpression of the anti-apoptotic gene, bcl-2, can acquire tumorigenic potential in vivo [8]. However, the results presented here demonstrate that Oxy5 overexpression does not lead to a similar situation, but rather the opposite. Future experiments will address in detail the mechanistic role of Oxy5 in regulating tumorigenesis. Oxy5 shares a high degree of homology to the annexin supergene family and especially to human annexin 1 [15]. Annexins are a ubiquitous family of more than 15 structurally related, calcium-dependent, phospholipid-binding proteins present in eukaryotic cells [25]. The are believed to be involved in diverse functions as membrane trafficking and exocytosis, blood-coagulation, membrane–cytoskeletal interactions, mitogenic signal transduction, and growth control and differentiation [26–31]. Hitherto, an antioxidant role has not been ascribed to any member of this family. Thus, it would be a worthwhile effort to test if one or a few members of the mammalian annexin family have an antioxidant role.
600
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
It has been demonstrated the viral proteins, including adenoviral E1B proteins and cowpox viral CrmA proteins can inhibit cell death in mammalian cells [32,33]. E1B19K is homologous to Bcl-2 and modifies p53 function, thus inhibiting apoptosis [32]. CrmA inhibits the caspases, which
are essential for the execution of the cell death program [33]. Thus, although there is precedence for non-mammalian proteins to serve an essential function in mammalian cells, our data demonstrates an important role for a plant protein in mammalian cells. We are currently attempting to
Fig. 5. Oxy5 and tumorigenicity. (a) Tumor formation rates in AKR/J mice injected subcutaneously with either control or oxy5 -transfected K36 cells, determined over indicated periods of time. Representative results from means of 20 mice are shown. Each point represents mean value of five mice in each group and vertical bars indicate standard deviation. (b) Balb/c nude mice were injected intraperitonially with either control or oxy5 -transfected CHO cells and tumor formation rates was monitored for the indicated time periods. Vertical bars indicate standard deviation. (c) Mice injected with K36: (i) or CHO (ii) cells at the end of the observation period. Arrow indicates site of injection. (a) represents injection site with oxy5 -transfected cells and (v) represents injection site with control cells.
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
elucidate if Oxy5 has a specific role in oxidant-induced cell death or if it would have a broad effect on cell death induced by various means. Moreover, the role of Oxy5 in regulating other cellular proteins involved in cell death is being investigated. The results presented here demonstrate that when Oxy5 is overexpressed in mammalian tumor cell lines, it could confer protections against OS inducing agents. It is not yet clear if there are any mammalian homologues of Oxy5, and if they can be induced upon OS and subsequently act to protect cells. However, endogenous Oxy5 has been shown to be induced by OS in plants and exhibit catalase activity, indicating that this plant protein has a physiological role in plants during OS. Nonetheless, future experiments will identify if Oxy5 has any functional homologues in untransformed mammalian cells. OS, due to the formation of ROS, plays a crucial role in controlling cell survival and cell death. The deregulation of the OS response leads to deleterious effects resulting in pathological conditions. Defining molecular components of the signaling pathway emancipating from OS, or identifying novel proteins that could interfere with the signaling pathway can serve as an attractive strategy for therapeutic intervention. In this respect, the above demonstration that a plant protein could function across evolutionary barriers and serve an antioxidant role in mammalian cells should be of potential clinical importance.
Acknowledgements This work was initiated at the Institute for Molecular and Cell Biology, Singapore. We are grateful to Dr K.M. Hui for his constructive suggestions. Financial support from National Science and Technology Board and National Medical Research Council is greatly appreciated.
References [1] I. Fridovich, in: W.A. Pryor (Ed.), Free radicals in Biology, Academic Press, New York, 1976, pp. 239 –277.
601
[2] B. Demple, C.F. Amabile-Cuvas, Redox redux:the control of oxidative stress responses, Cell 67 (1991) 837 – 839. [3] C. Behl, J.B. Davis, R. Lesley, D. Schubert, Hydrogen peroxide mediates amyloid beta protein toxicity, Cell 77 (1994) 817 – 827. [4] P.A. Cerutti, Prooxidant status and tumor promotion, Science 227 (1985) 375 – 381. [5] K.Y. Chia, S.P. Lim, A.A. Oei, T.K. Sabapathy, K.M. Hui, Acquisition of immunogenicity by AKR leukemic cells following DNA-mediated gene transfer is associated with the reduction of constitutive reactive oxidative species, Int. J. Cancer 57 (1994) 216 – 223. [6] T.J O’Donnell, C.J DeBoer, C.F Nathan, Persistence of human tumor cells in vitro to oxidative cytolysis, J. Clin. Invest. 76 (1985) 80 – 86. [7] G.H.W. Wong, J.H. Elwell, L.W. Oberley, D.V. Goeddel, Manganous superoxide dismutase is essential for cellular resistance to cytotoxicity of tumor necrosis factor, Cell 75 (1989) 923 – 931. [8] D.M. Hockenbery, Z.N. Oltvai, X.M. Yin, C.L Milliman, S.J. Krosmeyer, Bcl-2 functions in an antioxidant pathway to prevent apoptosis, Cell 75 (1993) 241 – 251. [9] D.J. Kane, T.A. Sarafian, R. Anton, H. Hahn, E.B. Gralla, J.S. Valentine, T. Ord, D.E. Bredsen, Bcl-2 inhibition of neuronal death: decreased generation of reactive oxygen species, Science 262 (1993) 1274 – 1277. [10] M.D. Jacobson, M.C. Raff, Programmed cell death and Bcl-2 protection in very low oxygen, Nature 374 (1995) 814 – 816. [11] M.D. Jacobson, Reactive oxygen species and programmed cell death, Trends Biochem. Sci. 21 (1996) 83 – 86. [12] M.F. Christman, R.W. Morgan, F.S. Jacobson, B.N. Ames, Positive control of a regulon for defenses against oxidative stress and some heat-shock protein in a Salmonella typhimurium, Cell 41 (1985) 753 – 762. [13] G. Storz, L.A. Tartaglia, B.N. Ames, Transcriptional regulation of oxidative stress inducible genes: direct activation by oxidation, Science 248 (1990) 189 – 194. [14] M.B. Toledano, I. Kullik, F. Trinh, P.T. Baird, T.D. Schneider, G. Storz, Redox-dependent shift of OxyRDNA contacts along an extended DNA-binding site: a mechanism for differential promoter selection, Cell 78 (1994) 897 – 909. [15] X. Gidrol, P.A. Sabelli, S.R. Yip, A. Kush, Annexin-like protein from Arabidopsis thaliana rescues oxyR mutant of E. coli from H2O2 stress, Proc. Natl. Acad. Sci. 93 (1996) 11268 – 11273. [16] T.K. Sabapathy, Q. Cheng, K.M. Hui, Characterization of tumor-specific cytotoxic effector cells with a novel CD3−/Thy1 + phenotype, Cell. Immunol. 166 (1995) 141 – 153. [17] W.S. Lin, M. Kapoor, Increase in superoxide production by heat-shocked cells of Neurospora crassa, demonstrated by a fluorometric assay, Int. J. Biochem. 24 (1992) 1081 – 1086.
602
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602 [26] R.B. Pepinsky, L.K. Sinclair, Epidermal growth factordependent phosphorylation of lipocortin, Nature 321 (1986) 81 – 84. [27] P. Thiagarajan, J.F. Tait, Binding of annexin V/placental anticoagulant protein to platelets. Evidence for phosphatidylserine exposure in the procoagulant response to activated platelets, J. Biol. Chem. 265 (1990) 17,420 – 17,423. [28] S.M. Violette, I. King, J.L. Browning, R.B. Pepinsky, B.P. Wallner, A.C. Sartorelli, Role of lipocortin 1 in the glucocorticoid induction of the terminal differentiation of a human squamous carcinoma, J. Cell. Physiol. 142 (1990) 70 – 77. [29] J. Theobald, P.D. Smith, S.M. Jacob, S.E. Moss, Expression of annexin VI in A431 carcinoma cells suppresses proliferation: a possible role foe annexin VI in cell growth regulation, Biochem. Biophys. Acta 1223 (1994) 383 – 390. [30] C.E. Creutz, The annexins and exocytosis, Science 258 (1992) 924 – 931. [31] J.D. Croxtall, R.J. Flower, Lipocortin1 mediated dexamethasone-induced growth arrest of the A459 lung adenocarcinoma cell line, Proc. Natl. Acad. Sci. 89 (1992) 3571 – 3575. [32] E. White, Regulation of p53-dependent apoptosis by E1A and E1B, Curr. Top. Microbiol. Immunol. 199 (1995) 34 – 58. [33] T. Komiyama, L.T. Quan, G.S. Salvesan, Inhibition of cysteine and serine proteinases by the cowpox virus serpin CrmA, Adv. Exp. Med. Biol. 389 (1996) 173 – 176.
[18] I. Sakaida, A.P. Thomas, J.L. Farber, Increases in cytosolic calcium ion concentration can be dissociated from the killing of cultured hepatocytes by tert-butyl-hydroperoxide, J. Biol. Chem. 266 (1991) 717 – 722. [19] J.W. Snyder, A. Serroni, J. Savory, J.L. Farber, The absence of extracellular calcium potentiates the killing of cultured hepatocytes by aluminum maltolate, Arch. Biochem. Biophys. 316 (1995) 434 –442. [20] S.E. Moss, D. Smith, Structural evolution of the annexin supergene family, Trends Genet. 10 (1994) 241 –246. [21] R. Gopalakrishna, W.B. Anderson, Reversible oxidative activation of protein kinase c by the mitogen/tumor promoter periodate, Arch. Biochem. Biophys. 285 (1991) 382– 387. [22] M.K Brawn, W.J. Chiou, K.L. Leach, Oxidation-induced activation of protein kinase C in UC1 1MG cells, Free Radic. Res. 22 (1995) 23 –27. [23] M.A. Pronzato, C. Domenicotti, E. Rosso, A. Bellocchio, M. Patrone, U.M. Marinari, E. Melloni, G. Poli, Modulation of rat liver protein kinase C during ‘‘in vivo’’ CC14-induced oxidative stress, Biochem. Biophys. Res. Commun. 194 (1993) 635 –641. [24] J. Pongracz, J.M. Lord, Superoxide production in human neutrophils: evidence for signal redundancy and the involvement of more than one PKC isoenzyme class, Biochem. Biophys. Res. Commun. 247 (1998) 624 –629. [25] F. Russo-Marie, The Annexins, Portland Press, London, 1992, pp. 35 –46.
.
Oxy5, a novel protein from Arabidopsis thaliana, protects mammalian cells from oxidative stress Anil Kush a, Kanaga Sabapathy b,* b
a Indo-American Hybrid Seeds, P.O. Box 7099, BSKII Bangalore 560070, India National Cancer Center, Di6ision of Cellular and Molecular Research, 11 Hospital Dri6e, Singapore 169610, Singapore
Received 19 July 2000; accepted 9 February 2001
Abstract The use of molecular oxygen in various cellular processes results in the generation of toxic intracellular by-products termed reactive oxygen species (ROS). In Escherichia coli the oxyR gene product is a transcriptional regulator of the oxyR regulon that is induced in response to hydrogen peroxide-induced oxidative stress (OS). We have previously shown that an annexin-like protein from Arabidopsis thaliana, termed Oxy5, can replace the obligatory role of OxyR in E. coli. Here, we have investigated as to whether oxy5 can function across evolutionary boundaries to protect mammalian cells from OS. Overexpression of the oxyR gene in mammalian tumor cell lines protects them from hydrogen peroxide-induced cell death, and these cells are also highly resistant to the superoxide ion producing compound paraquat. Oxy5 appears to be involved in the detection of calcium flux, as it binds to Ca2 + ions during hydrogen peroxide stress. Moreover, overexpression of Oxy5 leads to lowered protein kinase C activity. Thus, Oxy5 probably functions to sense and initiate protective responses to OS. In addition, Oxy5-overexpressing cells exhibit a reduction in endogenous superoxide ion levels, which concomitantly results in a dramatic decrease in their tumorigenic potential. Taken together, the results demonstrate an antioxidant role for plant oxy5 gene in mammalian cells, which can be potentially utilized in gene therapy programs aimed at reducing the deleterious effects of ROS. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: Annexin; Cell death; Oxidative stress; Oxy5; Reactive oxygen species; Superoxide ions
1. Introduction Oxidative stress (OS), resulting from the detrimental effects of reactive oxygen species (ROS), is an important phenomenon in many biological * Corresponding author. Tel.: + 65-436-8349; fax: +65226-5694. E-mail address: [email protected] (K. Sabapathy).
systems. ROS include superoxide ions (O− 2 ), hydroxyl radicals (OH) and hydrogen peroxide (H2O2). They are involved in lipid peroxidation, protein denaturation and DNA damage [1,2]. These cellular damages can lead to the development of various pathological symptoms observed both through apoptosis and necrosis, in diseases like arthritis, muscular dystrophy and various genetic and nervous disorders like Down’s syn-
1357-2725/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 1 3 5 7 - 2 7 2 5 ( 0 1 ) 0 0 0 4 0 - 1
592
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
drome and amylotrophic lateral sclerosis [1– 3]. Moreover, accumulating evidence suggests that increased cellular concentrations of active or reactive oxygen, organic peroxides and free radicals are important in tumor promotion [4– 6]. The mechanism by which mammalian cells counteract OS is at present not fully understood. In general, ROS are controlled through a protective network referred to as the antioxidant defense, which includes enzymes such as manganese superoxide dismutase (MnSOD), copper zinc superoxide dismutase (CuZnSOD), catalase and glutathione peroxidases. Overexpression of MnSOD has been shown to partially rescue cells from TNF-induced, ROS-mediated cell death [7]. In addition, Bcl-2 was shown to prevent apoptosis by scavenging ROS [8,9]. However, Bcl-2 could prevent cell death by a mechanism other than by its antioxidant function, as it also promotes cell survival in anaerobic conditions and cells that lack mitochondrial DNA [10,11]. In most of the above studies only a partial protective effect was observed, suggesting that other yet unidentified proteins might play an important role in this process. In Escherichia coli, the oxyR gene product is induced upon exposure to OS, which in turn activates the expression of defense genes [12]. The OxyR protein plays a dual role both in sensing and transducing peroxide stress signals [13,14]. By a functional complementation screen, we have previously identified an annexin-like protein from Arabidopsis thaliana, termed Oxy5, which could restore the ability of an E. coli oxyR deletion mutant to grow in the presence of H2O2 [15]. Moreover, we have demonstrated that Oxy5 was upregulated during OS in plants [15]. Oxy5 has high homology with the annexin supergene family and displays high catalase activity [15]. In the present study, we have investigated if this plant gene could serve an antioxidant function across evolutionary boundaries. We found that overexpression of Oxy5 protects mammalian cells against OS-induced cell death. Its mode of action is discussed.
2. Materials and methods
2.1. Cell lines, DNA and antibodies K36 is a leukemia cell line derived from the AKR/J mice [16] and CHO cells were obtained from ATCC. Cells were cultured in 10% serum containing RPMI media. An EcoR1 –Xho1 (1.2 kb) full-length oxy5 cDNA was cloned into pcDNA3 vector in both the sense and antisense orientations (Invitrogen) as described [15]. Both K36 and CHO cells were transfected with oxy5 or antisense-oxy5 expression plasmids and geneticin (G418)-resistant cells were selected. At least two independent pools of cells were used in all experiments. Polyclonal antibodies to Oxy5 were raised in rabbits as described [15].
2.2. RNA and protein analysis Total RNA was extracted from about 106 cells (control or oxy5 -transfected). The oligomer (5%TGG CGA CTC TTA AGG TTT CTG ATT CTG TTC 3%) corresponding to the N-terminus of A. thaliana oxy5 was used as a probe for mRNA analysis by standard Northern blotting techniques. For Ca2 + binding analysis, cells were incubated with 1 mCi of 45Ca2 + ml − 1 for 8 h before H2O2 treatment, lysed and immunoprecipitated using polyclonal antibodies against Oxy5, followed by 10% polyacrylamide SDSPAGE analysis and autoradiography. Levels of Oxy5 were determined by immunoprecipitation of cells labeled with 200 mCi of 35S-methionine for 2 h.
2.3. Cell 6iability assays Viability of cells treated with or without the indicated concentrations of H2O2 was analyzed by their ability to exclude propidium iodide. Percentage of viable cells (cells able to exclude propidium iodide, and therefore not stained for propidium iodide) was determined using FACScan with LYSYS II software. For studies with EGTA or nifedipine, cells were treated with either 1 mM EGTA or 10 mM nifedipine for 6 and 2 h, respectively, before H2O2 treatment
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
and cell viability was analyzed. For PMA studies, cells were treated with 100 nM of staurosporine or 250 nM of PMA for 3 and 2 h, respectively, prior to the H2O2 treatment for 1 h or 15 min, respectively. For O− 2 -induced stress, cells were treated with 200 mM of paraquat for three days and viability determined as described above.
2.4. Protein kinase C acti6ity assay Protein kinase C (PKC) activity was determined using an Amersham kit (RPN-77). Measurement of enzyme activity was based on PKC-mediated transfer of the [gamma-32P] ATP to a synthetic peptide homologous to the PKC-o pseudosubstrate domain in the presence of cofactors, as per the manufacturer’s instructions. PKC activity of control K36 cells treated H2O2 was taken as 100% PKC activity, and the other values are represented as a percentage of this value.
2.5. Measurement of O− 2 production by 4 -MFU-glu A Beckman LS2 fluorescence spectrophotometer was employed to determine the amount of O− 2 released by a fluorometric assay. The cell suspensions were adjusted to a volume of 3 ml (5× 106 cells ml − 1) in PBS containing 8 mM 4-methylbeta-D-umbelliferyl glucopyranoside (4-MUFglu). The assay estimates the amount of superoxide anion radical released, which then cleaves the 4-MUF-glu to 4-MUF, based on the liberation of 4-methylumbelliferone (4-MUF). Control or oxy5 -transfected cells were incubated at room temperature for an hour and the fluorescence intensity of the supernatant was measured, as described [17].
2.6. Analysis of tumor growth 1× 106 control or oxy5 -transfected K36 or CHO cells were injected subcutaneously in AKR/J mice or intraperitonially in Balb/c nude mice, respectively, and tumor growth was monitored over the indicated periods of time. Twenty independent mice of each type were used in each group in the study.
593
3. Results
3.1. Expression of Oxy5 protects mammalian cells from hydrogen peroxide-induced cell death To ascertain if Oxy5 affects cellular survival in response to OS, we transfected stably a murine leukemia cell line (K36) and an hamster ovary cell line (CHO) with the full length (1.2 kb) oxy5 cDNA or a pcDNA3 vector expressing the oxy5 gene in the antisense orientation (hereinafter referred to as control cells). Cells were selected for G418 resistance and three independently derived pools were obtained. We have used pools of G418-resistant cells to circumvent the problem of clonal variability. In all experiments, at least two independent pools were tested. Both K36 and CHO cells transfected with oxy5 expressed high levels of oxy5 mRNA (Fig. 1a), and protein (Fig. 2c), in contrast to control cells. Although Oxy5 protein shares a high degree of homology to annexin protein superfamily and especially to human annexin 1 [15], neither the cDNA for oxy5 nor the antibody to Oxy5 cross-reacted with any endogenous mRNA or protein from the control K36 and CHO cells (Fig. 1a, Fig. 2c). We next determined the optimal dose of H2O2, and the duration of treatment that results in substantial cell death of the control cells. Treatment of K36 and CHO cells for 6 h with various concentrations of H2O2 showed that most of the cells were dead at a high dose of 10 mM H2O2, whereas there was little cell death at both 100 mM and 1 mM of H2O2 (Fig. 1b(i) for K36 cells; data not shown for CHO cells). This indicates that these tumor cells are rather resistant to H2O2 stress and requires high levels of H2O2 to succumb to cell death. Therefore, we tested for the optimal duration of treatment with 10 mM of H2O2 at which about 60–75% of cell was killed. At 2 h after H2O2 treatment, most of the cells were dead (Fig. 1b(ii)), whereas about 77% of cells were killed at 1 h after exposure to H2O2. At shorter periods of treatment, between 65 (0.5 h) (Fig. 1b(ii)) and 80% (0.25 h) (Fig. 3b) cells were alive. Hence, we chose a treatment regimen of 10 mM H2O2 for 1 h for all our studies. When exposed to 10 mM H2O2 for 1 h, the Oxy5-overexpressing cells (both
594
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
Fig. 1. Oxy5-overexpressing cells are resistant to H2O2-induced cell death. (a) oxy5 mRNA expression in control and oxy5 -transfected K36 (K) and CHO (C) cells. gapdh expression was analyzed for loading control. (b) Determination of optimal doses (i) and duration (ii) of treatment with H2O2 in inducing cell death in K36 cells. Cell viability was determined as described in Section 2. (c) Control or oxy5 -overexpressing K36 (i) and CHO (ii) cells were treated with 10 mM H2O2 for 1 h and viability was determined. ( +) and ( −) indicates with or without H2O2 treatment, respectively. Vertical bars indicate standard deviation.
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
595
Fig. 2. Ca2 + and H2O2 stress. Cell survival rates were measured with prior treatment with 1 mM EGTA (for 6 h) (a) or 10 mM nifedipine for 2 h (b), followed by H2O2 stress, as described above. (c) Binding of 45Ca2 + to Oxy5 upon H2O2stress. Cells were labeled with 45Ca2 + (upper panel) or 35S-methionine (lower panel) as described in Section 2, lysed and immunoprecipitated with anti-Oxy5 antibody. Upper panel indicates amount of 45Ca2 + bound to Oxy5 and lower panel indicates levels of Oxy5 immunoprecipitated. ( + ) and ( −) indicate with or without H2O2 treatment, respectively.
596
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
Fig. 3. (Continued)
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
K36 and CHO) exhibited a survival rate of 60 and 70%, as opposed to 20 and 18% for the control K36 and CHO cells, respectively (Fig. 1c). Moreover, almost none of the Oxy5-overexpressing cells were dead when treated for 15 min as compared to about 20% death of control DNA transfected cells (Fig. 3b). These results were reproducible with independent pools of both the cell types tested over at least three separate experiments. Thus, these data indicate that overexpression of Oxy5 confers resistance to an otherwise lethal dose of H2O2, and suggests that Oxy5 can participate in the cellular responses to OS in mammalian cells.
3.2. Oxy5 acts as a calcium chelator during H2O2 -induced OS Since H2O2 stress has been shown to induce changes in calcium homeostasis [18,19] and members of the annexin superfamily are known to be calcium-binding proteins [20], we investigated if oxy5 which is highly homologous to the annexin superfamily of genes is causally involved in regulating changes in calcium influx during OS. Pretreatment of control K36 cells with the Ca2 + -chelator EGTA (1 mM for 6 h) or the Ca2 + -channel blocker nifedipine (10 mM for 2 h) rendered these cells more than twice as resistant to H2O2 than untreated cells (Fig. 2a and b). This resistance could be reversed by the addition of exogenous Ca2 + (data not shown). Similar results were obtained with control CHO cells (data not shown). In addition, the small percentage of Oxy5-overexpressing cells that were susceptible to H2O2 treatment was almost completely rescued in the presence of EGTA and nifedipine (Fig. 2a and b). These observations indicate that H2O2-induced cell death requires influx of exogenous Ca2 + and that inhibition of the Ca2 + influx can protect cells from OS-induced cell death. To determine if Oxy5 activity is associated with the Ca2 + influx, cells
597
were cultured in 45Ca2 + supplemented media for 8 h and the H2O2 stress experiments were performed as described earlier. Cells were lysed and immunoprecipitated with polyclonal antibodies against Oxy5 to determine the amount of 45Ca2 + bound to Oxy5. We found significantly higher binding of 45Ca2 + to Oxy5 in Oxy5-overexpressing K36 cells exposed to H2O2, compared to untreated cells (Fig. 2c, upper panel). Similar results were obtained with CHO cells (data not shown). There was no 45Ca2 + binding observed in control cells (Fig. 2c, upper panel), as the antibody did not recognize any endogenous protein (Fig. 2c, lower panel). It thus appears that Oxy5 is involved in the quenching of Ca2 + flux resulting from H2O2 stress, and thereby probably functions as a Ca2 + chelator to sense and initiate a protective response to OS.
3.3. PKC acti6ity is reduced in Oxy5 -o6erexpressing cells It is well established that both H2O2 stress and Ca2 + influx can induce the activity of various serine threonine kinases, including PKC [21– 23]. Therefore, we examined if the resistance conferred against cell death by Oxy5 is a consequence of the modulation of PKC activity. To start with, we pre-treated K36 and CHO cells with the PKC inhibitors staurosporine and H7 to evaluate their effect on cell death. Cell death (both control and oxy5 -transfected cells) due to H2O2 stress was significantly reduced in the presence of the PKC inhibitors (Fig. 3a; and data not shown for H7). Conversely, pre-treatment with the PKC-activator phorbol meristyl acetate (PMA) resulted in a significant acceleration of both the rate and extent of H2O2-induced cell (Fig. 3b). In this experiment, cell death was determined 15 min after H2O2 treatment, a time-point at which only about 20% cells were dead, as this would allow for the analysis of the effect of PMA on augmenting H2O2-in-
Fig. 3. PKC inactivation results in resistance to H2O2-induced cell death. K36 cells were pre-treated with either 100 nM staurosporine for 3 h (a) or 250 nM PMA for 2 h (b), followed by H2O2 stress [1 h in (a) and 15 min in (b)], and cell viability was determined. (c) PKC activity in control or oxy5 -transfected K36 cells without ( −) or with ( + ) H2O2 treatment, as described in legends to Fig. 1. Value of control cells treated with H2O2 was taken as 100% and other values are represented as a percentage of this value.
598
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
duced cell death (Fig. 1b(ii), Fig. 3b). More than 70% of control K36 cells pre-treated with PMA were dead only after 15 min of exposure to 10 mM H2O2, in contrast to about 20% cell death in the absence of the PMA pre-treatment (Fig. 3b). There was no significant amount of cell death induced by PMA treatment alone (Fig. 3b). Similar results were obtained with CHO cells (data not shown). These results suggest that Ca2 + -dependent PKC-like kinase might participate in effecting cell death from OS and the inhibition of its activity can avert the lethality. Strikingly, although the potentiation of H2O2-induced cell death by PMA was markedly reduced in Oxy5overexpressing cells (Fig. 3b), these cells were not completely resistant to H2O2-induced cell death. They underwent cell death to the same extent as the control cells not pre-treated with PMA (Fig. 3b). These results raised the possibility that Oxy5 expression might influence PKC activity and that Oxy5 acts upstream of PKC in the signaling cascade. We therefore investigated whether PKC activity is affected in cells overexpressing Oxy5. When extracts of H2O2-treated cells were analyzed, Oxy5-overexpressing cells were found to have about 25% lower levels of PKC activity compared with control cells (Fig. 3c). Taken together, the data suggests that Oxy5 acts by binding Ca2 + and probably quenches the calcium signaling cascade resulting in lowered PKC activity, which is required to transduce the cell death signal.
3.4. Reduction in superoxide production by Oxy5 -o6erexpressing cells Elevated PKC-like activity has been shown to increase the production of O− 2 [24]. We therefore examined if Oxy5-overexpressing cells, with lower − PKC activity, produce less O− 2 . O2 released in the supernatant was analyzed by a fluorometric assay using 4-MUF-glu that specifically reacts with O− 2 , as described in Section 2 [17]. In brief, this technique exploits the ability of superoxide to cleave the fluorescent substrate 4-MUF-glu to the highly fluorescent 4-MUF, which is detected flurometrically. K36 and CHO cells overexpressing Oxy5 contained between 40 and 50% less O− levels 2
Fig. 4. Reduced superoxide production by Oxy5-overexpressing cells: (a) levels of O− 2 ions were determined in both control and Oxy5-overexpressing K36 and CHO cells, using 4-MUFglu as a substrate, as described in Section 2; (b) survival rate of control or oxy5 -transfected K36 cells was determined in the presence of 200 mM paraquat (O− 2 ions producing herbicide) for three days.
compared to control cells (Fig. 4a). These results indicate that Oxy5 regulates the levels of intracellular superoxide anions, even in the absence of external OS stimuli. We thus reasoned that this would also result in resistance to external O− 2 insults. To test this possibility, cells were cultured in the absence or presence of the O− 2 producing herbicide paraquat to determine their viability. Oxy5-overexpressing cells showed an 80% survival rate compared with 20% survival of the control cells (Fig. 4b). These data indicate that Oxy5
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
expression confers resistance against O− 2 as well as H2O2 on cells, and can function in response to OS.
3.5. Oxy5 and tumorigenicity High levels of O− 2 production have been correlated with increased tumorigenicity [4] and we have previously demonstrated that altering the levels of O− 2 production by K36 tumor cells can reduce their tumorigenic potential [5]. We therefore examined if reduction of O− 2 production by overexpression of Oxy5 would indirectly affect their tumorigenic potential. 1×106 control or oxy5 -transfected K36 or CHO cells were injected subcutaneously in AKR/J mice or intraperitonially in Balb/c nude mice, respectively, and tumor growth was monitored in 20 independent mice of each type. Control K36 cells rapidly formed non-regressing tumors with almost 100% efficiency at the sites of inoculation and all mice developed maximum sized tumors by five weeks (Fig. 5a). The Oxy5-overexpressing K36 cells did form tumors initially, although at a much slower rate. However, these tumors did not develop further up to five weeks, and some of them disappeared by eight weeks (Fig. 5a and c(i)). Similarly, control CHO cells rapidly formed non-regressing tumors in nude mice with 100% efficiency. Tumors were noticed after five to seven days of inoculation and they kept on growing and reached a maximum size by 25 days (Fig. 5b and c(ii)). In contrast, Oxy5-overexpressing CHO cells did not form tumors in nude mice even after 25 days of inoculation (Fig. 5b and c(ii)).
4. Discussion We have demonstrated that the A. thaliana protein Oxy5, when overexpressed, could function as a sensor of the Ca2 + flux induced by an OS signal and subsequently serve as an effector of the protective cellular response leading to resistance against OS. Overexpression of Oxy5 results in lowered PKC levels and decreased production of O− 2 , which correlates with increased resistance to OS. These data demonstrate a direct correlation
599
between the intracellular oxidative state and the resistance to extracellular OS insults. Together, these studies suggest a potential role for plant Oxy5 in regulating intracellular oxidative processes in mammalian cells. Overexpression of Oxy5 also resulted in the suppression of tumor formation in vivo. At present, it is unclear how Oxy5-overexpressing cells, being more resistant to H2O2 in vitro, have a lower tumorigenic potential in vivo. Overexpression of Oxy5 results in two distinct effects: resistance to H2O2 results and decreased production of O− 2 . It could be envisaged that Oxy5 might effect both these processed independently. Decreased O− 2 production has been shown to result in decreased tumorigenicity, as we have previously demonstrated that treatment of mice injected with the K36 leukemia cells with superoxide dismutase (SOD), a superoxide scavenger, resulted in reduced tumorigenicity [5]. This correlates with the reduced tumorigenicity of Oxy5-overexpressing cells. Thus, the tumorigenic potential of cells could be altered by modulating the levels of O− 2 ions. It has been previously shown that cell lines rendered resistant to OS by overexpression of the anti-apoptotic gene, bcl-2, can acquire tumorigenic potential in vivo [8]. However, the results presented here demonstrate that Oxy5 overexpression does not lead to a similar situation, but rather the opposite. Future experiments will address in detail the mechanistic role of Oxy5 in regulating tumorigenesis. Oxy5 shares a high degree of homology to the annexin supergene family and especially to human annexin 1 [15]. Annexins are a ubiquitous family of more than 15 structurally related, calcium-dependent, phospholipid-binding proteins present in eukaryotic cells [25]. The are believed to be involved in diverse functions as membrane trafficking and exocytosis, blood-coagulation, membrane–cytoskeletal interactions, mitogenic signal transduction, and growth control and differentiation [26–31]. Hitherto, an antioxidant role has not been ascribed to any member of this family. Thus, it would be a worthwhile effort to test if one or a few members of the mammalian annexin family have an antioxidant role.
600
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
It has been demonstrated the viral proteins, including adenoviral E1B proteins and cowpox viral CrmA proteins can inhibit cell death in mammalian cells [32,33]. E1B19K is homologous to Bcl-2 and modifies p53 function, thus inhibiting apoptosis [32]. CrmA inhibits the caspases, which
are essential for the execution of the cell death program [33]. Thus, although there is precedence for non-mammalian proteins to serve an essential function in mammalian cells, our data demonstrates an important role for a plant protein in mammalian cells. We are currently attempting to
Fig. 5. Oxy5 and tumorigenicity. (a) Tumor formation rates in AKR/J mice injected subcutaneously with either control or oxy5 -transfected K36 cells, determined over indicated periods of time. Representative results from means of 20 mice are shown. Each point represents mean value of five mice in each group and vertical bars indicate standard deviation. (b) Balb/c nude mice were injected intraperitonially with either control or oxy5 -transfected CHO cells and tumor formation rates was monitored for the indicated time periods. Vertical bars indicate standard deviation. (c) Mice injected with K36: (i) or CHO (ii) cells at the end of the observation period. Arrow indicates site of injection. (a) represents injection site with oxy5 -transfected cells and (v) represents injection site with control cells.
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602
elucidate if Oxy5 has a specific role in oxidant-induced cell death or if it would have a broad effect on cell death induced by various means. Moreover, the role of Oxy5 in regulating other cellular proteins involved in cell death is being investigated. The results presented here demonstrate that when Oxy5 is overexpressed in mammalian tumor cell lines, it could confer protections against OS inducing agents. It is not yet clear if there are any mammalian homologues of Oxy5, and if they can be induced upon OS and subsequently act to protect cells. However, endogenous Oxy5 has been shown to be induced by OS in plants and exhibit catalase activity, indicating that this plant protein has a physiological role in plants during OS. Nonetheless, future experiments will identify if Oxy5 has any functional homologues in untransformed mammalian cells. OS, due to the formation of ROS, plays a crucial role in controlling cell survival and cell death. The deregulation of the OS response leads to deleterious effects resulting in pathological conditions. Defining molecular components of the signaling pathway emancipating from OS, or identifying novel proteins that could interfere with the signaling pathway can serve as an attractive strategy for therapeutic intervention. In this respect, the above demonstration that a plant protein could function across evolutionary barriers and serve an antioxidant role in mammalian cells should be of potential clinical importance.
Acknowledgements This work was initiated at the Institute for Molecular and Cell Biology, Singapore. We are grateful to Dr K.M. Hui for his constructive suggestions. Financial support from National Science and Technology Board and National Medical Research Council is greatly appreciated.
References [1] I. Fridovich, in: W.A. Pryor (Ed.), Free radicals in Biology, Academic Press, New York, 1976, pp. 239 –277.
601
[2] B. Demple, C.F. Amabile-Cuvas, Redox redux:the control of oxidative stress responses, Cell 67 (1991) 837 – 839. [3] C. Behl, J.B. Davis, R. Lesley, D. Schubert, Hydrogen peroxide mediates amyloid beta protein toxicity, Cell 77 (1994) 817 – 827. [4] P.A. Cerutti, Prooxidant status and tumor promotion, Science 227 (1985) 375 – 381. [5] K.Y. Chia, S.P. Lim, A.A. Oei, T.K. Sabapathy, K.M. Hui, Acquisition of immunogenicity by AKR leukemic cells following DNA-mediated gene transfer is associated with the reduction of constitutive reactive oxidative species, Int. J. Cancer 57 (1994) 216 – 223. [6] T.J O’Donnell, C.J DeBoer, C.F Nathan, Persistence of human tumor cells in vitro to oxidative cytolysis, J. Clin. Invest. 76 (1985) 80 – 86. [7] G.H.W. Wong, J.H. Elwell, L.W. Oberley, D.V. Goeddel, Manganous superoxide dismutase is essential for cellular resistance to cytotoxicity of tumor necrosis factor, Cell 75 (1989) 923 – 931. [8] D.M. Hockenbery, Z.N. Oltvai, X.M. Yin, C.L Milliman, S.J. Krosmeyer, Bcl-2 functions in an antioxidant pathway to prevent apoptosis, Cell 75 (1993) 241 – 251. [9] D.J. Kane, T.A. Sarafian, R. Anton, H. Hahn, E.B. Gralla, J.S. Valentine, T. Ord, D.E. Bredsen, Bcl-2 inhibition of neuronal death: decreased generation of reactive oxygen species, Science 262 (1993) 1274 – 1277. [10] M.D. Jacobson, M.C. Raff, Programmed cell death and Bcl-2 protection in very low oxygen, Nature 374 (1995) 814 – 816. [11] M.D. Jacobson, Reactive oxygen species and programmed cell death, Trends Biochem. Sci. 21 (1996) 83 – 86. [12] M.F. Christman, R.W. Morgan, F.S. Jacobson, B.N. Ames, Positive control of a regulon for defenses against oxidative stress and some heat-shock protein in a Salmonella typhimurium, Cell 41 (1985) 753 – 762. [13] G. Storz, L.A. Tartaglia, B.N. Ames, Transcriptional regulation of oxidative stress inducible genes: direct activation by oxidation, Science 248 (1990) 189 – 194. [14] M.B. Toledano, I. Kullik, F. Trinh, P.T. Baird, T.D. Schneider, G. Storz, Redox-dependent shift of OxyRDNA contacts along an extended DNA-binding site: a mechanism for differential promoter selection, Cell 78 (1994) 897 – 909. [15] X. Gidrol, P.A. Sabelli, S.R. Yip, A. Kush, Annexin-like protein from Arabidopsis thaliana rescues oxyR mutant of E. coli from H2O2 stress, Proc. Natl. Acad. Sci. 93 (1996) 11268 – 11273. [16] T.K. Sabapathy, Q. Cheng, K.M. Hui, Characterization of tumor-specific cytotoxic effector cells with a novel CD3−/Thy1 + phenotype, Cell. Immunol. 166 (1995) 141 – 153. [17] W.S. Lin, M. Kapoor, Increase in superoxide production by heat-shocked cells of Neurospora crassa, demonstrated by a fluorometric assay, Int. J. Biochem. 24 (1992) 1081 – 1086.
602
A. Kush, K. Sabapathy / The International Journal of Biochemistry & Cell Biology 33 (2001) 591–602 [26] R.B. Pepinsky, L.K. Sinclair, Epidermal growth factordependent phosphorylation of lipocortin, Nature 321 (1986) 81 – 84. [27] P. Thiagarajan, J.F. Tait, Binding of annexin V/placental anticoagulant protein to platelets. Evidence for phosphatidylserine exposure in the procoagulant response to activated platelets, J. Biol. Chem. 265 (1990) 17,420 – 17,423. [28] S.M. Violette, I. King, J.L. Browning, R.B. Pepinsky, B.P. Wallner, A.C. Sartorelli, Role of lipocortin 1 in the glucocorticoid induction of the terminal differentiation of a human squamous carcinoma, J. Cell. Physiol. 142 (1990) 70 – 77. [29] J. Theobald, P.D. Smith, S.M. Jacob, S.E. Moss, Expression of annexin VI in A431 carcinoma cells suppresses proliferation: a possible role foe annexin VI in cell growth regulation, Biochem. Biophys. Acta 1223 (1994) 383 – 390. [30] C.E. Creutz, The annexins and exocytosis, Science 258 (1992) 924 – 931. [31] J.D. Croxtall, R.J. Flower, Lipocortin1 mediated dexamethasone-induced growth arrest of the A459 lung adenocarcinoma cell line, Proc. Natl. Acad. Sci. 89 (1992) 3571 – 3575. [32] E. White, Regulation of p53-dependent apoptosis by E1A and E1B, Curr. Top. Microbiol. Immunol. 199 (1995) 34 – 58. [33] T. Komiyama, L.T. Quan, G.S. Salvesan, Inhibition of cysteine and serine proteinases by the cowpox virus serpin CrmA, Adv. Exp. Med. Biol. 389 (1996) 173 – 176.
[18] I. Sakaida, A.P. Thomas, J.L. Farber, Increases in cytosolic calcium ion concentration can be dissociated from the killing of cultured hepatocytes by tert-butyl-hydroperoxide, J. Biol. Chem. 266 (1991) 717 – 722. [19] J.W. Snyder, A. Serroni, J. Savory, J.L. Farber, The absence of extracellular calcium potentiates the killing of cultured hepatocytes by aluminum maltolate, Arch. Biochem. Biophys. 316 (1995) 434 –442. [20] S.E. Moss, D. Smith, Structural evolution of the annexin supergene family, Trends Genet. 10 (1994) 241 –246. [21] R. Gopalakrishna, W.B. Anderson, Reversible oxidative activation of protein kinase c by the mitogen/tumor promoter periodate, Arch. Biochem. Biophys. 285 (1991) 382– 387. [22] M.K Brawn, W.J. Chiou, K.L. Leach, Oxidation-induced activation of protein kinase C in UC1 1MG cells, Free Radic. Res. 22 (1995) 23 –27. [23] M.A. Pronzato, C. Domenicotti, E. Rosso, A. Bellocchio, M. Patrone, U.M. Marinari, E. Melloni, G. Poli, Modulation of rat liver protein kinase C during ‘‘in vivo’’ CC14-induced oxidative stress, Biochem. Biophys. Res. Commun. 194 (1993) 635 –641. [24] J. Pongracz, J.M. Lord, Superoxide production in human neutrophils: evidence for signal redundancy and the involvement of more than one PKC isoenzyme class, Biochem. Biophys. Res. Commun. 247 (1998) 624 –629. [25] F. Russo-Marie, The Annexins, Portland Press, London, 1992, pp. 35 –46.
.