Interaction with PI3-kinase contributes to the cytotoxic activity of Apoptin

(1)

SHORT COMMUNICATION

Interaction with PI3-kinase contributes to the cytotoxic activity ofApoptin

S Maddika

1,2,10,11

_{, E Wiechec}

1,3,11

_{, SR Ande}

1,2,11

_{, IK Poon}

4

_{, U Fischer}

5

_{, S Wesselborg}

6

_{, DA Jans}

4

_,

K Schulze-Osthoff

5

_{and M Los}

1,2,7,8,9

1_{Manitoba Institute of Cell Biology, CancerCare Manitoba, University of Manitoba, Winnipeg, Manitoba, Canada;}2_{Department of}

Biochemistry and Medical Genetics, University of Manitoba, Winnipeg, Manitoba, Canada;3_{Institute of Human Genetics, University}

of Aarhus, Aarhus, Denmark;4_{Nuclear Signaling Laboratory, Department of Biochemistry and Molecular Biology, Monash}

University, Clayton, Victoria, Australia;5_{Institute of Molecular Medicine, University of Du¨sseldorf, Du¨sseldorf, Germany;} 6_{Department of Internal Medicine I, University of Tu¨bingen, Tu¨bingen, Germany;}7_{Manitoba Institute of Child Health, University of}

Manitoba, Winnipeg, Manitoba, Canada;8_{Department of Human Anatomy and Cell Science, University of Manitoba, Winnipeg,}

Manitoba, Canada and9_{BioApplications Enterprises, Winnipeg, Manitoba, Canada}

Apoptin, a small protein from the chicken anemia virus, has attracted attention because ofits specificity in killing tumor cells. Localization ofapoptin in the nucleus of tumor cells has been shown to be vital for proapoptotic activity, however, targeted expression ofapoptin in the nucleus ofnormal cells does not harm the cells, indicating that nuclear localization of apoptin is insufficient for its cytotoxicity. Here, we demonstrate for the first time that apoptin interacts with the SH3 domain ofp85, the regulatory subunit ofphosphoinositide 3-kinase (PI3-K), through its proline-rich region. Apoptin derivatives devoid ofthis proline-rich region do not interact with p85, are unable to activate PI3-K, and show impaired apoptosis induction. Moreover, apoptin mutants containing the proline-rich domain are sufficient to elevate PI3-K activity and to induce apoptosis in cancer cells. Downregulation of p85 leads to nuclear exclusion ofapoptin and impairs cell death induction, indicating that interaction with the p85 PI3-K subunit essentially contributes to the cytotoxic activity ofapoptin.

Oncogene(2008) 27, 3060–3065; doi:10.1038/sj.onc.1210958; published online 3 December 2007

Keywords: anticancer drugs; apoptin; apoptosis; PI3-kinase

Apoptin, a 14 kDa protein encoded by VP3 gene of chicken anemia virus, represents a novel potential anticancer therapeutic, because it induces apoptotic death speciﬁcally in tumor but not normal cells (Fischer and Schulze-Osthoff, 2005; Alvisi et al., 2006). The mechanism by which apoptin is able to distinguish between tumor and normal cells remains to be

elucidated. The subcellular localization of apoptin appears to be important for the selective toxicity towards transformed and cancer cells. In cells resistant to apoptin-mediated cell death, the protein has a cytoplasmic localization, whereas in sensitive cells, apoptin is found in the nucleus (Danen-Van Oorschot et al., 2003). Apoptin was described to be speciﬁcally phosphorylated in transformed cells by an unknown kinase, which was proposed to be important for the nuclear localization and apoptotic activity of apoptin (Rohn et al., 2002). However, it was also shown that a C-terminally truncated apoptin mutant, in which the phosphorylation site was deleted, was still able to translocate to the nucleus and induce apoptosis (Guelen et al., 2004). Thus, the interaction with other molecules or additional modiﬁcations of apoptin that are not present in nontransformed cells may be required for the selective cytotoxicity of apoptin.

Employinga mass spectrometric approach, we previously found that apoptin interacts with the p85 regulatory subunit of phosphoinositide 3-kinase (PI3-K) (Maddika et al., 2007). Class I PI3-Ks play a central role in various regulatory processes, such as cell growth, survival and differentiation, by converting phosphatidylinositol-4,5-phosphate to phosphatidylino-sitol-3,4,5-phosphate, leadingto the activation of AKT kinase and other downstream effectors (Fruman et al., 1998; Vanhaesebroeck and Waterfield, 1999). To in-vestigate the relevance of an interaction of apoptin with the p85 subunit of PI3-K, we performed glutathione S-transferase (GST)–apoptin pull-down assays usingtotal cell extracts of different tumor cell lines, includingPC-3 prostate cancer cells, L929 mouse fibrosarcoma cells, 293 transformed human embryonic kidney cells and MCF-7 human breast cancer cells. We detected that apoptin interacts with the p85 subunit in all the cell lines tested (Figure 1a), indicating that interaction of apoptin with p85 is neither cell type- nor species-specific.

To identify and map the sites on apoptin responsible for interaction with PI3-K, PC-3 cells were transfected to express either full-length green ﬂuorescent protein (GFP)-apoptin or various deletion mutants tagged with an N-terminal GFP (Figure 1b). Apoptin was

Received 17 September 2007; revised 23 October 2007; accepted 29 October 2007; published online 3 December 2007

Correspondence: Dr M Los, BioApplications Enterprises, Winnipeg, Manitoba, Canada R2V 2N6.

E-mail: mjelos@gmail.com

10_{Current address: Department of Therapeutic Radiology, Yale School} of Medicine, New Haven, CT, USA.

11_{These authors contributed equally to this work}

(2)

immunoprecipitated with anti-GFP antibodies at 24 h post-transfection, and the immune complexes were analysed for interaction with PI3-K by immunoblotting usinganti-p85 antibodies. PI3-K was found in the immunoprecipitates of full-length apoptin and apoptin derivatives that harbored amino acids 74–100 (a proline-rich sequence), implyingthat this region of apoptin is important for interaction with PI3-K (Figure 1b). We then tested whether interaction of PI3-K with apoptin is essential for apoptin-induced cell death by assessingthe percentage of apoptosis induced by the different mutants. In agreement with the interaction data, only those mutants that had the intact interaction site for PI3-K were able to induce apoptosis both in PC-3 and MCF-7 cells (Figure 1c). Interestingly, the mutant Ala-108, which was reported to be noncytotoxic due the substitution of a Thr-phosphorylation site (Rohn et al., 2002), was still able to induce signiﬁcant apoptosis, as longas its interaction site with the PI3-K remained intact. Next, we investigated the apoptin interaction site on p85, which is attached to the p110 catalytic subunit and, dependingon its conformation, can act as a repressor or

inducer of kinase activity. The p85 subunit contains an N-terminal SH3 domain, two proline-rich domains and a breakpoint cluster region homology domain (Figure 2a). p85 also has two SH2 domains (nSH2, cSH2) that are separated by a coiled-coil region, called the inter (i) SH2 domain, which mediates the stable dimerization with the p110 catalytic subunit. Under restingconditions, p85 inhibits PI3-K activity. The bindingof tyrosine-phosphorylated proteins to the nSH2 domain of p85 inhibits PI3-K and is most frequently responsible for kinase activation, however, alternate mechanisms have been reported. For example, the conformational switch between p85 and p110 holoenzyme can also occur upon the interaction of signaling mediators with the SH3 or breakpoint cluster region homology domain (Liu et al., 1993; Prasad et al., 1993; Pleiman et al., 1994; Zheng et al., 1994). To investigate the apoptin interaction site of the p85 subunit, we cotransfected PC-3 cells with a vector encoding full-length apoptin, together with full-length p85 or various deletion mutants. Both full-length p85 and the mutant lackingthe iSH2 domain were

Figure 1 In vivoassociation of apoptin with the p85 subunit of PI3-K. (a) GST pull-down assay performed with PC-3, MCF-7, 293 and L929 cell lysates with either GST or GST–apoptin. The proteins cloned in PGEX-2T (Amersham Biosciences, Mississauga, ON, Canada) were along with the different cell lysates immobilized on glutathione sepharose beads overnight at 4 1C. The beads were washed thrice with ice-cold lysis buffer and bound proteins were separated on sodium dodecyl sulphate–PAGE. The presence of the p85 subunit of PI3-K in apoptin complexes was determined by immunoblottingusinganti-p85 and horseradish peroxidase (HRP) coupled anti-mouse IgG (Upstate Cell Signaling, Lake Placid, NY, USA). (b) Schematic representation of apoptin deletion mutants tagged with N-terminal GFP. The deletion mutants (Poon et al., 2005) and wild type apoptin (1–121) were transfected into PC-3 cells, immunoprecipitated with anti-GFP antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) 18 h post-transfection, and the association of p85 was detected by immunoblotting. (c) Induction of apoptosis by the different apoptin mutants. The percentage of apoptosis was assessed in PC-3 cells by ﬂow cytometric detection of hypodiploid DNA (left panel) and in MCF-7 cells by the MTT assay (right panel) 24 h post-transfection essentially as described previously (Maddika et al., 2005; Burek et al., 2006).

Interaction of apoptin with PI3-kinase S Maddika et al

(3)

immunoprecipitated by the anti-p85 antibody, while other mutants were immunoprecipitated usinganti-hemagglutinin antibodies. Immunodetection of apoptin in the immune complexes revealed that apoptin interaction required an intact SH3 domain of p85 (Figure 2a).

To determine the functional signiﬁcance of the interaction of apoptin with the p85-regulatory subunit, we measured the PI3-K activity by usinga nonradio-active ELISA-based method. We tested whether the activation of PI3-K is a direct consequence of apoptin’s interaction by observingthe PI3-K activity in cells transfected with different apoptin deletion mutants

(Figure 2b). Full-length apoptin induced six-fold activa-tion, while the 1–89 mutant, which may contain only a partially preserved interaction site, weakly activated PI3-K. The mutants 1–111, 74–121, Ala-108 and Glu-108 were able to activate PI3-K similar to the levels of activation by the full-length apoptin. In contrast, the 1–73 and 104–121 mutants were unable to activate beyond basal levels, indicatingthat the intact p85 interaction site on apoptin is crucial for PI3-K activation (Figure 2b).

On the other hand, the observed PI3-K activity in cells cotransfected with full-length apoptin and p85

Figure 2 Apoptin interacts with SH3 domain of p85 and activates PI3-K activity. (a) Schematic representation of the PI3-K p85 deletion mutants highlighting the structural domains. The p85 mutants (Jascur et al., 1997) were cotransfected together with apoptin into PC-3 cells. A 28 h post-transfection, cells were lysed, and p85 was immunoprecipitated usingeither anti-p85 (for wild type and DiSH2 mutants) or anti-HA antibodies (Upstate) for the HA-tagged mutants. Apoptin interaction with the mutants was detected using antiapoptin antibodies. Expression of the deletion mutants was veriﬁed by anti-p85 or anti-HA antibodies. (b) PI3-K activity was measured 24 h after transfectingPC-3 cells with the indicated apoptin mutants usingan ELISA (Echleon Biosciences). (c) The indicated p85 PI3-K mutants were expressed after transfection in PC-3 cells. A 24 h post-transfection, p85 was immunoprecipitated from cell lysates and assessed for PI3-K activity as described above. (d) PC-3 cells were either transfected with apoptin alone, pretreated with wortmannin (5 nM) or LY294002 (1.5 mM) followed by apoptin transfection, or transfected to coexpress apoptin and a dominant-negative (DN) PI3-K mutant (Kauffmann-Zeh et al., 1997).

(4)

deletion mutants conﬁrmed that the interaction of apoptin at the SH3 domain of p85 is essential for PI3-K activity (Figure 2c). The p85 mutants with both the SH3 and iSH2 domains had constitutive PI3-K activity in the presence of apoptin, whereas others lackingthese domains were severely impaired in their activity (Figure 2c). Interestingly, the mutant Nþ I þ C þ iSH2, containingonly the SH2 domains but lackingthe SH3 domain, conferred relatively higher PI3-K activity compared to other deletion mutants and the control, indicatingthat the intact SH3 domain in the wild-type p85 protein might have an inhibitory action on its activity (Figure 2c). In control experiments, different PI3-K inhibitors, includingwortmannin and LY294002, and expression of a dominant-negative PI3-K mutant reduced the apoptin-induced PI3-K activation to the basal levels (Figure 2d). Thus, the interaction of apoptin with the SH3 domain might lead to a conformational change of p85, converting it from an inhibitory to an active state.

To investigate the functional signiﬁcance and the requirement of PI3-K activity towards apoptin-induced cell death, we used RNA interference and tested whether knockingdown endogenous p85 would affect resistance to apoptin-induced cell death. In PC-3 and MCF-7 cells, the expression of p85 was completely abrogated by p85-speciﬁc small interfering(si) RNA but not by a control siRNA (Figure 3a), as assayed 30 h post-transfection. The assay revealed that cells lackingp85 expression were strongly resistant towards apoptin-induced cell death compared to control siRNA-expressingcells (Figure 3b). Similar observations were made in 293 cells (data not shown).

We also tested whether PI3-K inhibition affects subcellular localization of apoptin. Apoptin was mainly localized in the nucleus of PC-3 and MCF-7 cells, but in the presence of wortmannin, apoptin was found mainly distributed in the cytoplasm, as shown either by immunostainingand confocal laser microscopic imaging (Figure 4a) or by subcellular fractionation and western blotting(Figure 4c). The effect of PI3-K inhibition on the localization of apoptin was further studied usingp85 siRNA. Upon inhibition of p85 expression, apoptin was almost completely excluded from the nucleus (Figure 4b), indicating that PI3-K activity is necessary for not only apoptin-induced cell death, but also for the nuclear localization of the protein.

In summary, we demonstrate that apoptin interacts with the p85 subunit of PI3-K in various cell types. Apoptin speciﬁcally associates with the SH3 domain of p85, leadingto the activation of PI3-K presumably by a conformational change. Importantly, downregula-tion of the p85 subunit by RNA interference or PI3-K inhibition severely impairs apoptin-induced cell death.

p85 speciﬁcally interacts with the proline-rich domain of apoptin localized between amino acids 80 and 90. The PKPPSK sequence (amino acids 81–86) of apoptin is a perfect consensus motif (PxxPxR/K) recognized by SH3 domains. Mutants lackingthis particular proline-rich region were unable to interact with the p85, were not able to activate PI3-K pathway and were nontoxic. This clearly implies that apoptin interacts with SH3 domain of p85 and activates PI3-K. Previously, it has been shown that phosphorylation of Thr-108 is speciﬁc for tumor cells and is required for apoptin-induced cell

Figure 3 Downregulation of p85 impairs apoptin-induced apoptosis. (a) PC-3 cells (left panels) and MCF-7 cells (right panels) were transfected with a p85 siRNA plasmid (Upstate Cell Signaling) or control siRNA for 48 h using Lipofectamine (Invitrogen, Burlington, ON, Canada). Downregulation of p85 was veriﬁed by immunoblotting; tubulin was used as loading control. (b) PC-3 cells (left panel) and MCF-7 cells (right panel) were transfected with p85 siRNA or control siRNA vectors, followed by transfection after 48 h to express apoptin. After further 24 and 48 h, cell death was measured as described in Figure 1c.

(5)

death (Rohn et al., 2002). However, phosphorylation of apoptin was severely impaired in a mutant lacking amino acids 80–90, even though the residues within the phosphorylation site remained intact (Rohn et al., 2002). Our results explain this interestingobservation by showingthat amino acids 80–90 are vital for the interaction with PI3-K, and this interaction may be required for further activation of downstream molecules. Mutants of Thr-108, that is Ala-108 and Glu-108, were still cytotoxic, indicatingthat phosphorylation of apoptin is not of crucial importance.

Our data further suggest that the PI3-K pathway is involved in localization of apoptin. Inhibition of the PI3-K pathway usingwortmannin or RNA interference retained apoptin in the cytoplasm and led to its nuclear exclusion, hence impairingcell death. PI3-K has been mainly implicated in lipid signaling at the cell membrane, but increasingevidence indicates that all components of the PI3-K pathway also reside in the nucleus (Neri et al., 2002). The functions regulated by nuclear K are still rather unclear. Although PI3-K is important for apoptin localization, it may not be a

direct regulator, as we did not observe any direct phosphorylation of apoptin by PI3-K in in vitro kinase assays (data not shown). Apoptin phosphorylation may, however, be regulated by downstream effectors of the PI3-K pathway, a possibility that we are currently investigating. In conclusion, the data pre-sented here not only advance our knowledge on the mechanisms that are activated by apoptin to kill cancer cells, but they also demonstrate that the prosurvival PI3-K/Akt pathway, frequently aberrantly active in cancer, may have additional functions in cell death propagation.

Acknowledgements

We thank J Downward and T Mustelin for valuable reagents. SM was funded by a studentship from the Manitoba Health Research Council (MHRC) and the CancerCare Manitoba Foundation (CCMF). EW was supported through ITMHRT. ML thankfully acknowledges the support by the CFI-Canada Research Chair program, MICH-, CIHR, CIHR-IPM-, CCMF- and MHRC-ﬁnanced programs.

Figure 4 Downregulation of p85 affects apoptin localization. (a) The localization of GFP-apoptin in MCF-7 and PC-3 cells was detected by confocal microscopy after transfection in the absence or presence of wortmannin (5 nM). (b) The effect of p85 knockdown on apoptin localization was determined by confocal microscopy, 24 h post-transfection with GFP-apoptin with or without a cotransfected p85 siRNA plasmid. The expression of p85 was detected by immunohistochemistry as described (Janssen et al., 2007) usinganti-p85 and a Cy-3-conjugated secondary antibody. (c) The nuclear/cytoplasmic distribution of apoptin in the presence of wortmannin or p85-inhibitory siRNA in both MCF-7 and PC-3 cells was investigated by subcellular fractionation, followed by western blottingas described previously (Maddika et al., 2005). Scale bars: 10 mm.

(6)

References

Alvisi G, Poon IK, Jans DA. (2006). Tumor-speciﬁc nuclear targeting: promises for anti-cancer therapy? Drug Resist Updat 9: 40–50. Burek M, Maddika S, Burek CJ, Daniel PT, Schulze-Osthoff K,

Los M. (2006). Apoptin-induced cell death is modulated by Bcl-2 family members and is Apaf-1 dependent. Oncogene 25: 2213–2222.

Danen-Van Oorschot AA, ZhangYH, Leliveld SR, Rohn JL, Seelen MC, Bolk MW et al. (2003). Importance of nuclear localization of apoptin for tumor-speciﬁc induction of apoptosis. J Biol Chem 278: 27729–27736.

Fischer U, Schulze-Osthoff K. (2005). New approaches and therapeu-tics targeting apoptosis in disease. Pharmacol Rev 57: 187–215. Fruman DA, Meyers RE, Cantley LC. (1998). Phosphoinositide

kinases. Annu Rev Biochem 67: 481–507.

Guelen L, Paterson H, Ga¨ken J, Meyers M, Farzaneh F, Tavassoli M. (2004). TAT-apoptin is efﬁciently delivered and induces apoptosis in cancer cells. Oncogene 23: 1153–1165.

Janssen K, Hofmann TG, Jans DA, Hay RT, Schulze-Osthoff K, Fischer U. (2007). Apoptin is modiﬁed by SUMO conjugation and targeted to promyelocytic leukemia protein nuclear bodies. Onco-gene 26: 1557–1566.

Jascur T, Gilman J, Mustelin T. (1997). Involvement of phosphatidy-linositol 3-kinase in NFAT activation in T cells. J Biol Chem 272: 14483–14488.

Kauffmann-Zeh A, Rodriguez-Viciana P, Ulrich E, Gilbert C, Coffer P, Downward J et al. (1997). Suppression of c-Myc-induced apoptosis by Ras signalling through PI(3)K and PKB. Nature 385: 544–548.

Liu X, Marengere LE, Koch CA, Pawson T. (1993). The v-Src SH3 domain binds phosphatidylinositol 30_{-kinase. Mol Cell Biol 13:} 5225–5232.

Maddika S, Bay GH, Kroczak TJ, Ande SR, Maddika S, Wiechec E et al.(2007). Akt is transferred to the nucleus of cells treated with apoptin, and it participates in apoptin-induced cell death. Cell Prolif 40: 835–848.

Maddika S, Booy EP, Johar D, Gibson SB, Ghavami S, Los M. (2005). Cancer-speciﬁc toxicity of apoptin is independent of death receptors but involves the loss of mitochondrial membrane potential and the release of mitochondrial cell-death mediators by a Nur77-dependent pathway. J Cell Sci 118: 4485–4493.

Neri LM, Borgatti P, Capitani S, Martelli AM. (2002). The nuclear phosphoinositide 3-kinase/AKT pathway: a new second messenger system. Biochim Biophys Acta 1584: 73–80.

Pleiman CM, Hertz WM, Cambier JC. (1994). Activation of phosphatidylinositol-3’ kinase by Src-family kinase SH3 binding to the p85 subunit. Science 263: 1609–1612.

Poon IK, Oro C, Dias MM, ZhangJ, Jans DA. (2005). Apoptin nuclear accumulation is modulated by a CRM1-recognized nuclear export signal that is active in normal but not in tumor cells. Cancer Res 65: 7059–7064.

Prasad KV, Janssen O, Kapeller R, Raab M, Cantley LC, Rudd CE. (1993). Src-homology 3 domain of protein kinase p59fyn mediates bindingto phosphatidylinositol 3-kinase in T cells. Proc Natl Acad Sci USA 90: 7366–7370.

Rohn JL, ZhangYH, Aalbers RI, Otto N, Den HertogJ, Henriquez NV et al. (2002). A tumor-speciﬁc kinase activity regulates the viral death protein Apoptin. J Biol Chem 277: 50820–50827.

Vanhaesebroeck B, Waterﬁeld MD. (1999). Signaling by distinct classes of phosphoinositide 3-kinases. Exp Cell Res 253: 239–254. ZhengY, Bagrodia S, Cerione RA. (1994). Activation of

phospho-inositide 3-kinase activity by Cdc42Hs bindingto p85. J Biol Chem 269: 18727–18730.