Neurotrophin receptor-interacting mage homologue is an inducible inhibitor of apoptosis protein-interacting protein that augments cell death.

The inhibitor of apoptosis proteins (IAPs) have been shown to interact with a growing number of intracellular proteins and pathways to fulfil their anti-apoptotic role. In the search for novel IAP-interacting proteins we identified the neurotrophin receptor-interacting MAGE homologue (NRAGE) as being able to bind to the avian IAP homologue ITA. This interaction requires the RING domain of ITA. NRAGE additionally coimmunoprecipitates with XIAP. When overexpressed in 32D cells NRAGE augments interleukin-3 withdrawal induced apoptosis, possibly through binding endogenous XIAP. Moreover, NRAGE is able to overcome the anti-apoptotic effect of Bcl-2.

The inhibitor of apoptosis proteins (IAPs) have been shown to interact with a growing number of intracellular proteins and pathways to fulfil their anti-apoptotic role. In the search for novel IAP-interacting proteins we identified the neurotrophin receptor-interacting MAGE homologue (NRAGE) as being able to bind to the avian IAP homologue ITA. This interaction requires the RING domain of ITA. NRAGE additionally coimmunoprecipitates with XIAP. When overexpressed in 32D cells NRAGE augments interleukin-3 withdrawal induced apoptosis, possibly through binding endogenous XIAP. Moreover, NRAGE is able to overcome the anti-apoptotic effect of Bcl-2.
The inhibitor of apoptosis proteins (IAPs) 1 are encoded by a recently identified family of highly conserved genes, playing a crucial role in the regulation of apoptosis by interfering with the process of caspase activation (1). The mammalian IAPs consist of the X-linked inhibitor of apoptosis (XIAP, MIHA, ILP), cellular IAP1 (cIAP1, HIAP2, MIHC), cIAP2 (HIAP1, MIHB), NAIP, and survivin (2)(3)(4)(5)(6). The inhibitor of T cell apoptosis (ITA) is an avian IAP homologue (7). Members of this family must contain at least one NH 2 -terminal cysteine-rich baculovirus IAP repeat (BIR) domain, which is required for pro-survival activity, and may also possess a carboxyl-terminal RING zinc finger. cIAP1 and cIAP2 were originally identified as part of the TNFR2-TRAF signaling complex, with the BIR domains being responsible for the interaction with TRAFs 1 and 2 (3). XIAP on the other hand was shown to act as a mediator of bone morphogenic protein signaling linking the receptors to the signaling modulators TAB1 (TAK1-binding protein) and TAK1. The RING domain of XIAP binds to the receptor, while the BIR domains are important for recruiting TAB1 (8). XIAP was also shown to interact with nuclear proteins such as XAF1 (9) and can be diverted away from caspase inhibition by Smac/DIABLO (10,11), as well as being involved in other signaling processes such as JNK1 (c-Jun NH 2 -terminal kinase) (12) and NF-B (8) activation.
Another group of proteins important in the regulation of apoptosis is the Bcl-2 family (13). Bcl-2 itself has been shown to play a crucial role in the prevention of apoptosis at the mitochondrial surface impeding the function of the pro-death family members or by obstructing the release of apoptogenic molecules (13).
NRAGE is a very recently cloned neurotrophin receptorinteracting MAGE homologue, uncovered in a two-hybrid search for proteins interacting with the intracellular domain of the p75 neurotrophin receptor (p75NTR). This protein was shown to be involved in the apoptotic response after nerve growth factor (NGF) binding in neuronal cells (14). NGF can mediate a variety of cellular responses, such as survival, differentiation, growth, and apoptosis in neurons, by binding to two types of cell surface receptors, the Trk receptor tyrosine kinases and the p75NTR (15). Interestingly, NRAGE shows a high homology in its carboxyl-terminal with the melanoma associated antigen (MAGE) family, which codes for antigens that present on many human tumors, although their intracellular function is mostly unknown. One exception is necdin, a postmitotic neuron-specific growth suppressor that can interact with both intracellular proteins such as E2F1 and p53 and viral transforming proteins such as adenovirus E1A (16).
In this study we identified the pro-apoptotic protein NRAGE as a novel IAP-interacting protein in the yeast two-hybrid system. Furthermore, we show that NRAGE coprecipitates with ITA and XIAP, via the RING zinc finger domain of the IAPs. NRAGE can augment cell death upon growth factor withdrawal in the IL-3-dependent promyeloid leukemic 32D cell line, even in the presence of Bcl-2.

EXPERIMENTAL PROCEDURES
Plasmid Construction-Molecular cloning of ITA was described previously (7). The full-length cDNA was inserted into the ApaI and EcoRI sites of pcDNA3 (Invitrogen). A 5Ј PCR product was amplified using the primers 5Ј-tccagcaacagttgtcgactatgaacataatg-3Ј and 3Ј-tgggtctcgaggaaaactg-5Ј, with pCDNA3-ITA as a template and Pfu polymerase (Stratagene). The PCR product was digested with SalI and SmaI, ligated with pCDNA3-ITA SmaI/EcoRI, and inserted first into the twohybrid vector pPC86 at SalI/EcoRI then into pPC97 at SalI/AaT2 (17). pCDNA3HA-ITA construction was described previously (18). The expression vectors for truncated ITA isoforms were generated by PCR using HF DNA polymerase (CLONTECH) on the pCDNA3HA-ITA tem-* This work was supported by Mildred Scheel Stiftung Grant Az 10-1446-Ra4. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
plate. For ITA⌬BIR, the KpnI site containing forward primer 5Ј-gcgcggtacccctgcttcgtgtgaaaggaggag-3Ј together with the reverse primer 5Јtagctgaaagtaaactcccaagg-3Ј were used, the PCR product digested at the KpnI and MroI sites, and cloned into a MroI-and KpnI-digested pcDNA3HA-ITA. For ITA⌬RING, the forward primer 5Ј-ggtttccgaggtgtgagtatctg-3Ј together with the EcoRI containing reverse primer 5Јgcgcgaattcttatgttctctcttcttgcaatctcc-3Ј were used and the PCR product digested at the MroI and EcoRI sites and cloned into MroI-and EcoRIdigested pcDNA3HA-ITA.
Cloning of NRAGE-The largest two-hybrid insert was labeled with [ 32 P]dCTP using the rediprime TM random labeling system (Amersham Pharmacia Biotech) according to manufacturer's instructions and used to screen a bacterially expressed S194 murine plasmacytoma cDNA library as described previously (20). The full-length NRAGE cDNA was amplified with Pfu polymerase from a resultant 2.8-kilobase pair putative full-length clone with the forward primer 5Ј-gcgaattctgccaccatggactacaagaaggacgatgagatgatgacaaactggaaaccatccag-3Ј, incorporating an EcoRI site and a flag tag, and the reverse primer 5Ј-gcgcctcgagttactcaacccagaagaagccaatg-3Ј, and inserted into pCFG5-IEGZ (provided by D. Lindemann) at the Klenow-treated EcoRI site. This allowed expression of both the gene of interest from the retroviral long terminal repeat and the green fluorescent protein and Zeocin resistance genes from an internal ribosomal entry site.
Yeast Two-hybrid System-The yeast two-hybrid screen was performed using a PC12 cDNA library cloned into pPC86HA3 (17). The yeast strain Y190 was manipulated according to the MATCHMAKER Library User Manual (CLONTECH). A sequential transformation protocol was used with pPC97-ITA as bait. Positive clones were identified by growth on SD/-TRP/-LEU/-HIS/ ϩ 25 mM 3-amino-1,2,4-triazole (Sigma) plates and activity of the lacZ reporter gene in filter assays. Direct two-hybrid tests were performed according to standard protocols.
Cell Culture and Transfection-The human embryonic kidney cell line HEK293 and Phoenix TM packaging cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) heat-inactivated fetal calf serum (FCS), 100 units of penicillin and streptomycin/ ml, and 2 mM L-glutamine, at 37°C in humidified air with 6% CO 2 . The 32D cells were obtained from Dr. J. S. Greenberger and grown in RPMI 1640 (Life Technologies, Inc.) supplemented with 10% FCS, 100 units of penicillin and streptomycin/ml, 2 mM L-glutamine, and 15% of WEHIconditioned medium at 37°C, 5% CO 2 , and 90% humidity. 32D cells expressing a human Bcl-2 protein were obtained from J. Reed (21).
Cells were transfected by a modified calcium phosphate coprecipitation method (22). HEK293 cells were starved in Dulbecco's modified Eagle's medium with 0.3% (v/v) FCS for 48 h after transfection.
Immunoprecipitations, Immunoblots, and Antibodies-Transfected HEK293 cells and 32D cells were lysed at 4°C in Nonidet P-40 buffer (10 mM Hepes, pH 7.4, 145 mM KCl, 5 mM MgCl 2 , 1 mM EGTA, 0.2% IGEPAL, and a protease/phosphatase inhibitor mix containing 1 mM Pefabloc, 1 mM sodium orthovanadate, 5 mM benzamidine, 5 g/ml aprotinin, 5 g/ml leupeptin). Cell debris was removed by centrifugation, and 500 g of protein supernatants were precleared with 10 l of protein-A/G-agarose (Amersham Pharmacia Biotech). The lysates were then incubated with the appropriate antiserum coupled to 30 l of protein-A/G-agarose overnight at 4°C. The immune complexes were washed twice in Nonidet P-40 buffer containing 500 mM NaCl. After resuspension in Laemmli buffer and SDS-PAGE, proteins were detected by immunoblot. Anti-FLAG M2 monoclonal antibody (Eastman Kodak Co.) was a generous gift from S. Feller (Institut fü r Medizinische Strahlenkunde und Zellforschung). Goat polyclonal anti-FLAG raised against the OctA probe and rabbit polyclonal ERK2 antibody were purchased from Santa Cruz. HA-tagged proteins were detected with a monoclonal anti-HA antibody (12CA5). Anti-XIAP rabbit polyclonal antibody was purchased from R & D systems anti-mouse or anti-rabbit horseradish peroxidase-linked F(abЈ)2 fragments (Amersham Pharmacia Biotech), and anti-goat IgG horseradish peroxidase (Santa Cruz) was also employed.
Establishment of Stable Cell Lines-The protocol for transfection of Phoenix TM virus packaging cells and establishment of stably transduced 32D cells has been described previously (23). Retroviral expressed proteins were confirmed after selection with Zeocin by Western blotting.
Cell Survival Assays-For the analysis of cell survival, cells were washed three times in tissue culture medium containing 10% FCS but without WEHI cell supernatant. 0.5 ϫ 10 6 cells in a total volume of 1 ml were dispensed into a single well of a flat bottom 24-well tissue culture plate. Cell viability was routinely assessed by staining the cells in trypan blue (Sigma). For protein analysis, the cell number and volume were scaled up by ϫ20.

RESULTS
ITA Binds to Overlapping NH 2 -terminal NRAGE Fragments-In the search for IAP-interacting proteins, a yeast twohybrid screen was performed with the full-length ITA cDNA as bait, using a PC12 cDNA library. The ITA-GAL4DB fusion protein expressed in yeast cells did not autonomously activate the lacZ reporter genes and was not toxic (data not shown). 32 His-positive clones were obtained from this screen, of which 8 were ␤-galactosidase positive, and after elimination of false positives, 3 clones remained. Sequence analysis of these clones revealed that they represented three overlapping clones from the same cDNA (Fig. 1A, upper panel). Although they bore no homology to any previously described gene, an electronically assembled gene THC179960 from The Institute of Genomic Research did show a high degree of identity; THC179960 was later identified as NRAGE (14).

XIAP Interacts Directly with NRAGE NH 2 -terminal Clones
in Vivo-To determine whether NRAGE could interact with other IAP family members, direct yeast two-hybrid tests were carried out between the strongest interacting clone (⌬NRAGEp 60 ) and c-IAP1, c-IAP2, XIAP, and NAIP. XIAP interacted with ⌬NRAGEp 60 with an equal intensity of ␤-galactosidase activity compared with ITA (Fig. 1A, lower panel). XIAP also bound to ⌬NRAGEp 70 and ⌬NRAGEp 80 . However, c-IAP1, c-IAP2, and NAIP did not interact with any of the clones in direct yeast-two-hybrid tests.
ITA Interacts with NRAGE via Its RING Zinc Finger Domain-To independently confirm the ITA-NRAGE two-hybrid interaction we performed coimmunoprecipitation experiments in a higher eukaryotic cell system. The full coding sequence of NRAGE was cloned into a mammalian expression vector as described under "Experimental Procedures," and HEK293 cells were cotransfected with HA-tagged ITA and FLAG-tagged NRAGE. NRAGE was immunoprecipitated and subsequently immunoblotted using a goat anti-FLAG antibody (Fig. 1B). As shown in Fig. 1B, upper panel, NRAGE was detectable after immunoprecipitation with the 12CA5 antibody using lysates from HA-ITA-expressing lysates (Fig. 1B, lane 2), but not from control cells (Fig. 1B, lane 5). Furthermore, NRAGE coimmunoprecipitated with an ITA deletion construct containing only the RING domain (Fig. 1B, lane 4), but not with the deletion construct containing only the BIR domains of ITA (Fig. 1B, lane  3). These interactions were able to withstand high stringency washes with 500 mM NaCl. In Fig. 1B, lower panel, the presence of the immunoprecipitated ITA proteins used in the coprecipitation experiments (lanes 2-5) was confirmed by stripping and reprobing the same membrane with the 12CA5 monoclonal antibody. In summary, these experiments demonstrate that ITA interacts with NRAGE with high affinity and that this interaction occurs specifically via its carboxyl-terminal RING zinc finger domain.
XIAP Binds to NRAGE-The same strategy was applied to confirm the XIAP-NRAGE binding. HEK293 cells were transiently transfected with vectors expressing FLAG-NRAGE and GST-XIAP or FLAG-NRAGE and GST only. In Fig. 1C, upper left panel, NRAGE is coimmunoprecipitated with GST-XIAP (lane 2), whereas no coprecipitated protein was detected in the control sample (lane 3). NRAGE was immunoprecipitated and immunoblotted as a positive control (lane 1). The blot was stripped and reprobed using the rabbit polyclonal XIAP antibody, shown in the lower left panel.
In Fig. 1C, righthand panels, the reverse experiment was performed. GST-XIAP was coimmunoprecipitated with NRAGE (lane 2 upper right panel). Once again, this interaction was stable enough to withstand high salt washing conditions (500 mM NaCl). Immnunoprecipitated and immunoblotted GST-XIAP was used as the positive control (lane 1), while GST containing lysate was used as a negative control in this experiment (lane 3). The blots were stripped and reprobed using the FLAG antibody to detect tagged NRAGE, shown in the lower right panel. These experiments confirm NRAGE as a novel XIAP-interacting protein, with a strong binding affinity.
NRAGE Expression in the IL-3-dependent 32D Cell Line Augments Factor Withdrawal-induced Apoptosis-IL-3-dependent 32D cells have been well studied with regard to the process of apoptosis induced by growth factor withdrawal (24,25). For this reason we established NRAGE and vector stably transduced cell lines, to examine the contribution of NRAGE to this process. NRAGE expression was determined by Western blotting (data not shown). The proliferation rate of these cells was not affected by the stable integration of the retroviral vector or NRAGE protein expression (data not shown). Cell survival after IL-3 withdrawal was measured by trypan blue exclusion. As shown in Fig. 2A, the vector-transduced cells behave identically to the parental cells. In contrast, expression of NRAGE had a dramatic effect on the kinetics of cell death in 32D cells. The survival curves for NRAGE and the vector/parental lines diverge very early on (12 h), and this trend continues throughout the experiment. At the 24-h time point about three times as many dead cells were observed with the NRAGE-expressing cell pool. This indicates that NRAGE greatly accelerates the cell death process upon growth factor withdrawal in this IL-3dependent promyeloid leukemic cell line.

NRAGE Binds XIAP in Vivo upon IL-3 Withdrawal in 32D
Cells-Over a 24-h period we analyzed the interaction of NRAGE with endogenous XIAP in 32D cells. Fig. 3A, upper panel, shows the binding of NRAGE with XIAP is undetectable in proliferating 32D cells; however, 8 -12 h after IL-3 withdrawal we observed an induction of binding, trailing off thereafter, and becoming undetectable at 16 h, as seen in Fig. 3A, upper panel.
To investigate the consequence of NRAGE-XIAP interaction in 32D cells after IL-3 withdrawal, we examined endogenous XIAP protein levels. Fig. 3B, upper panel, shows that XIAP cleavage occurs very early in NRAGE-transduced cells and leads to rapid loss of XIAP protein. In contrast, Fig. 3C shows XIAP cleavage is barely detectable in vector-transduced cells at 24-h IL-3 withdrawal, and the loss of XIAP protein occurs at a greatly reduced rate compared with NRAGE expressing cells.
NRAGE Overcomes Bcl-2-mediated Cell Protection-Forced Bcl-2 expression in 32D cells has been shown to be very effective in rescuing the cells from the induction of apoptosis by a variety of stimuli, including withdrawal (21). We therefore wanted to test whether it could also block the death-inducing activity of NRAGE. 32D cells stably expressing Bcl-2 were infected with either the retroviral control vector or the retrovirus containing FLAG-NRAGE and selected as described above. 24 h after IL-3 withdrawal, the 32D cells expressing Bcl-2 gained a 100% increase in survival compared with the wild type cells, and no significant difference was seen in the 32DBcl-2 vector-transduced cells. However the 32DBcl-2 cells expressing NRAGE showed a dramatic reduction in survival, completely abolishing any protection given by Bcl-2 (Fig. 2B).
This would suggest that NRAGE is a pro-apoptotic protein despite Bcl-2 expression, after growth factor withdrawal.

DISCUSSION
Besides Bcl-2 proteins, the IAP family members constitute the largest group of pro-survival proteins. Members of this heterogeneous family are mainly defined by the presence of a BIR domain, a novel zinc binding fold, and the ability to suppress apoptosis (26). BIR domains mediate caspase binding and inhibition and are both necessary and sufficient for the antiapoptotic function of IAP proteins (27). While the anti-apoptotic function of Bcl-2 is well understood (13), very little is known about the mechanisms that control IAP function. In an attempt do define possible regulators we carried out a yeast two-hybrid screen with the avian IAP and ITA and can demonstrate that the newly discovered NRAGE binds to the IAP proteins ITA and XIAP. For ITA, the interaction with NRAGE occurs independent of the BIR motif but requires the presence of the RING domain, which is present in c-IAP1, c-IAP2, and XIAP but is lacking in NAIP. Consistent with this we also fail to detect interaction of NAIP with NRAGE. Moreover, our results demonstrate that expression of NRAGE in the IL-3-dependent promyeloid cell line 32D accelerates cell death following growth factor withdrawal even in the presence of Bcl-2. XIAP was previously shown to undergo caspase-mediated cleavage during apoptosis (32). Coexpression of NRAGE led to accelerated XIAP cleavage and further degradation of XIAP during apoptosis in 32D cells, while in the control cells this process occurred at a much reduced rate.
The role of the RING domain of IAPs in the process of apoptosis suppression remains controversial. While the RING domain of c-IAP2 has been shown to be important in TNFmediated NF-B survival signaling (28), it was not necessary for caspase inhibition (29). Furthermore, a recent report demonstrated that the RING domains of cIAP1 and XIAP possess ubiquitin ligase activity upon apoptosis induction, resulting in degradation of these proteins (30). It is tempting to speculate that proteins binding to this region might regulate this process.
NRAGE has been shown initially to induce apoptosis signaling by blocking the physical association of p75NTR with TrkA (15). However, in contrast to one of the observations of Salehi et al. (14), we saw no effect on proliferation by stable expression of NRAGE, 2 which could be due to the surviving cell population expressing a level of NRAGE low enough not to induce cell cycle arrest, but still able to interfere with apoptosis. Similarly, IAP proteins cIAP1 and cIAP2 have first been identified as part of the related TNF-receptor signaling complex (3). The ability of NRAGE to interfere with the intrinsic cell death pathway that is triggered by IL-3 withdrawal points to a novel function of NRAGE in apoptosis signaling. It is commonly assumed that Bcl-2 proteins interfere with the process of caspase activation at the mitochondrial level (13), while the IAPs have been shown to act at the receptor level (3) and post-mitochondrially in the assembly of the apoptosome (31). We performed in vitro caspase cleavage assays utilizing recombinant NRAGE and XIAP proteins, but were unable to see an influence of NRAGE on XIAP anticaspase activity (data not shown). We currently have little understanding of exactly which step NRAGE influences XIAP cleavage, and it thus remains to be established whether NRAGE may function on the level of promoting caspase recruitment rather than altering enzyme activity.
Interestingly, we observed a very high level of expression of nrage in mouse brain tissue. 3 It has been previously described that in embryonic chick primary neurons, ITA expression is required for NGF-mediated survival (18). It remains to be seen whether IAP-NRAGE determines the signaling outcome from the NGF receptor. In summary, our data suggest that IAP activity may be controlled by heterodimerization with proteins such as NRAGE. Future work will dissect the mechanisms that govern the interplay of these molecules in survival control.