Aph2, a Protein with a zf-DHHC Motif, Interacts with c-Abl and Has Pro-apoptotic Activity*

c-Abl is a non-receptor tyrosine kinase implicated in DNA damage-induced cell death and in growth factor receptor signaling. To further understand the function and regulation of c-Abl, a yeast two-hybrid screen was performed to identify c-Abl-interacting proteins. Here we report the identification of Abl-philin 2 (Aph2), encoding a novel protein with a unique cysteine-rich motif (zf-DHHC) and a 53-amino acid stretch sharing homology with the creatine kinase family. Thezf-DHHC domain is highly conserved from yeast to human. Two proteins containing this motif, Akr1p and Erf2p, have been characterized in Saccharomyces cerevisiae, both implicated in signaling pathways. Deletion analysis by two-hybrid assays revealed that the N-terminal portion of Aph2 interacts with the C terminus of c-Abl. Aph2 was demonstrated to interact with c-Abl by co-immunoprecipitation assays. Aph2 is expressed in most tissues tested and is localized in the cytoplasm, mainly in the endoplasmic reticulum (ER). The sequences required for ER location reside in the N terminus and the zf-DHHC motif of Aph2. It has been reported that a portion of c-Abl is localized in the ER. We demonstrate here that Aph2 and c-Abl are co-localized in the ER region. Overexpression of Aph2 leads to apoptosis as justified by TUNEL assays, and the induction of apoptosis requires the N terminus. Co-expression of c-Abl and Aph2 had a synergistic effect on apoptosis induction and led to a decreased expression of both proteins, suggesting either that these two proteins are mutually down-regulated or that cells expressing both c-Abl and Aph2 rapidly disappeared from the culture. These results suggest that Aph2 may be involved in ER stress-induced apoptosis in which c-Abl plays an important role.

Aph2, by a two-hybrid screen. The protein has a cysteine-rich domain that is conserved in many proteins from yeast to mammals. Aph2 is ubiquitously expressed in tissues tested and is mainly localized in the ER of cells. When ectopically expressed, Aph2 induces apoptosis. c-Abl and Aph2 have a synergistic effect in apoptosis induction. These results indicate that Aph2 may cooperate with c-Abl in induction of apoptosis.

MATERIALS AND METHODS
Yeast Two-hybrid Screen-To produce a bait protein, a type I c-Abl cDNA fragment (encoding 1091 residues) was subcloned into pSH2-1 to form the LexA DNA binding domain fusion protein. The bait construct was introduced by the standard LiAc-polyethylene glycol method into a yeast strain, CTY10 -5D, which carried in its genome a ␤-galactosidase reporter under the regulation of a LexA promoter. The library of cDNAs was derived from the murine myeloid cell line WEHI-3 and was transferred into pGADNOT (20). To screen for c-Abl-interacting proteins, the library was used to transform CTY10 -5D carrying the bait and was selected on Leu Ϫ His Ϫ plates. Interaction was detected by staining for ␤-galactosidase activity on nitrocellulose replicas. To identify the sequences of Aph2 that interact with c-Abl, different portions of Aph2 cDNA were synthesized by PCR and subcloned into pGADNOT and then were transformed into CTY10 -5D carrying the bait. ␤-galactosidase activity was detected on the nitrocellulose replica.
Northern Analysis-The tissue distribution of Aph2 was determined by Northern analysis. A blot that contains mRNA from different tissues (CLONTECH) was probed with radio-labeled Aph2. The probe was random primer-labeled full-length Aph2 cDNA.
Cell Culture and Transfection-COS-7 and NIH3T3 cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal calf serum supplemented with glutamine and penicillin. For transfection, DNA isolated with the Maxi-DNA purification kit (Qiagen) was used to transform cells by the lipofectAMINE method following the manufacturer's protocol (Invitrogen).
To express c-Abl, the open reading frame of type IV c-Abl was cloned into the expression vector pMT21 (21) without any epitope tag (a stop codon was retained in c-Abl). To express Aph2, the open reading frame of Aph2 was synthesized by PCR and cloned into pSR␣ (22) such that the expression protein was tagged by c-myc peptide at the C terminus. To make various deletions of Aph2, PCR-synthesized mutant versions were subcloned into pSR␣ similarly.
Immunoprecipitation and Western Blot-Transfected cells were washed with cold PBS and lysed in TNEN buffer (50 mM Tris-HCl, pH 7.5, 50 mM NaCl, 2 mM EDTA, 0.5% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml each of pepstatin, leupeptin, and aprotinin). The lysate was clarified by centrifugation at 10,000 ϫ g for 10 min at 4°C. Aph2 was immunoprecipitated with monoclonal anti-myc antibody 9E10 (Santa Cruz Biotechnology) and recovered by protein Gagarose. c-Abl was precipitated with polyclonal antibodies (K-12) (Santa Cruz Biotechnology), raised against the kinase domain of c-Abl, and recovered with protein A-agarose. The precipitate was washed in TNEN buffer three times, boiled in 1ϫ loading buffer, and separated on a SDS-PAGE gels. Proteins were transferred to immobilon-P nitrocellulose membranes (Millipore), and Western blot was carried out by standard protocol and visualized with ECL (Amersham Biosciences). To detect protein phosphorylation after c-Abl transfection, cells were lysed in TNEN buffer containing 0.5 mM sodium orthovanadate and 1 mM NaF. To detect phosphorylated proteins on nitrocellulose blots, an antiphosphorylated tyrosine antibody (RC20:HRPO, Transduction Laboratories) was used.
Immunohistochemistry-COS or NIH3T3 cells growing on coverslips were transfected with lipofectAMINE. After 24 h, the cells were permeablized with methanol at Ϫ20°C for 3 min, washed three times with PBS containing 10% bovine calf serum, and incubated with the primary antibodies for 1 h at room temperature. 9E10 was used to detect Aph2, and K-12 was used to detect c-Abl. After interaction with the primary antibody, the slides were washed three times with PBS containing 10% bovine calf serum and then incubated with secondary antibody conjugated with FITC or Texas Red (Sigma) for 1 h at room temperature. DAPI (4Ј,6-diamidino-2-phenylindole) was used to stain the nucleus, and DioC6 was used to stain the ER. The slides were then washed with PBS and examined under a confocal microscope.
TUNEL Assay-COS-7 or NIH3T3 cells were transfected with vector DNA expressing Aph2 by the lipofectAMINE method. Cells cultured for 24 h after transfection were fixed in 10% formalin and then assayed for apoptosis using a kit from Oncor, which uses horseradish peroxidase-conjugated antibodies. Counter-staining was performed with Methyl Green.
Quantitation of Apoptotic Cells-COS-7 or NIH3T3 cells were transfected with pCR3.1cmv-galactosidase (23) and plasmids expressing c-Abl, Aph2, or both and were cultured for 24 h. Plates were washed with PBS and fixed in a phosphate buffer containing formaldehyde and paraldehyde for 5 min at room temperature. In each transfection, empty vectors (PSRѨ or pMT21) were used to make up the same amount of DNA. After being washed in phosphate buffer containing Nonidet P-40, cells were stained with X-gal at 37°C. When the blue color developed, plates were washed and kept in PBS. ␤-galactosidase-positive cells were counted at 40 ϫ10 magnification for five fields of view that represented the surviving cells. For NIH3T3, surviving cells were counted from seven fields of view at the magnification of 10 ϫ10 because of lower transfection efficiency.

RESULTS
Cloning and Expression of Aph2-To search for c-Abl-interacting proteins, a yeast two-hybrid screen was carried out. The nearly full-length type I c-Abl (aa 1-1091) was used as bait and fused to the DNA binding domain of LexA. The library consists of fusions between the GAL4 activation domain and cDNAs isolated from WEHI-3 myeloid lymphoma. A total of one million yeast colonies were screened, and four were found to carry one previously unidentified gene named Aph2 for abl-philin 2. The gene was also isolated in a separate screen using a C-terminal portion of Abl as bait (20). To obtain the full-length cDNA, the 5Ј-terminal DNA of an initial two-hybrid clone (a 400-bp fragment) was used to screen a mouse liver cDNA library. Clones were isolated, subcloned into pBluescript, and sequenced. The longest cDNA sequence has been deposited to GenBank TM (accession number AF176814), and the predicted protein sequence is shown in Fig. 1A. The Aph2 gene encodes a 361-amino acid protein.
Aph2 protein shows no strong homology to any known proteins in GenBank TM . A search of the expression data base revealed that Aph2 has a homologue in the human (AAH04535). A search for domains using the domain architecture retrieval tool (DART) revealed two distinct domains, zf-DHHC (aa 157-210) (24) and a 53-amino acid sequence (aa 205-257) homologous to the consensus sequence of ATP:guanido phosphotransferase, represented by creatine kinase. The zf-DHHC domain is distinct from the Zinc-finger, LIM, and Ring finger domains and is found in many proteins including Akr1p and Erf2p, both of which are involved in signaling pathways in Saccharomyces cerevisiae (25)(26)(27)(28). Fig. 1B shows the alignment of five such proteins. Creatine kinase is the major enzyme in vertebrates that converts creatine to phosphocreatine using ATP as phosphate donor. Phosphocreatine is responsible for transporting energy from the site of ATP synthesis and is also used as an energy buffer (29).
Northern analysis was carried out to determine the expression pattern of Aph2 in mouse tissues. Fig. 2A shows that Aph2 has two transcripts, a major one that is ϳ2 kb in size and a minor one of ϳ3.5 kb. The 2-kb mRNA probably corresponds to the cDNA of Aph2, whereas the 3.5-kb mRNA may represent an alternatively spliced form. Aph2 is expressed in all tissues examined, with a much weaker expression in the spleen. The commercial blot has an even loading of RNA from different tissue as judged by probing with actin (20). The ubiquitous expression suggests that the Aph2 protein may play a fundamental role in cell physiology. To test the expression of Aph2 at the protein level, Aph2 was tagged with glutathione S-transferase or maltose-binding protein (MBP) at the N terminus and was expressed in bacteria. A band migrating at a portion of markers Ͼ200 kDa was detected on a SDS-PAGE gel (data not shown), suggesting that Aph2 protein exists as aggregates. To express Aph2 in mammalian cells, Aph2 was Myc-tagged at the C terminus. When expressed in COS-7 cells, Aph2 again was found to migrate as a Ͼ200-kDa protein on SDS-PAGE gels (Fig. 2B). It may be a multimer or a conjugate with other proteins through disulfide bonds because the zf-DHHC motif has many cysteine residues. The behavior was unusual for proteins boiled in SDS in the presence of reducing agents. Addition to the lysis buffer of 100 mM dithiothreitol or treatment with 1 mM iodoacetamide did not improve the solubility of Aph2. Lysis in radioimmune precipitation assay (RIPA) buffer also gave similar results. However, proper expression of Aph2 was confirmed by the observation that a fusion protein between MBP and Aph2 ran at the correct size (Fig. 2B). This result also suggests that addition of MBP to the N-terminal end may have disrupted the tertiary structure of Aph2 and therefore made the protein incapable of conjugating with other proteins. Deletion of the cysteine-rich domain also resulted in a protein that runs at correct molecular weight on SDS-PAGE gel (Fig. 2C). An N-terminal fragment of Aph2 (aa 1-226) and a C-terminal fragment of Aph2 (aa 150 -361) also were expressed correctly. These data indicate that Aph2 can be expressed correctly in mammalian cells, although it may exist in a form of oligomer or in a form of conjugate with other proteins.
c-Abl Interacts with Aph2-To determine which domain of the protein interacts with c-Abl, different portions of Aph2 were fused in-frame to the DNA binding domain of LexA, and two hybrid assays were carried out (Fig. 3A). We found that any deletion of Aph2 dramatically reduced its affinity for c-Abl. Several fragments truncated at the C terminus displayed quite strong interactions as long as the zf-DHHC domain was retained, whereas all deletions of the N terminus, even in the presence of the zf-DHHC domain, completely abolished the interaction. These results suggest that the N-terminal portion of the protein is important for interaction with c-Abl, although no obvious protein-protein interaction motif is apparent. Similarly, different portions of c-Abl were fused to the DNA binding domain of GAL4 to identify the domains important for the interaction (Fig. 3B). We used the full-length Aph2 fused to the LexA activation domain. These assays suggested that the Cterminal portion of c-Abl interacts with Aph2.
To confirm the interaction in mammalian cells, type IV c-Abl and Aph2 were co-expressed in COS-7 cells, and co-immunoprecipitation assays were performed. First, Aph2-Myc and associated proteins were immunoprecipitated with antibody against the c-Myc tag, fractionated on an 8% SDS-PAGE gel, and transferred to nitrocellulose membrane. The blot was probed with antibodies against c-Abl. As shown in Fig. 3C, c-Abl was brought down with Aph2, although pre-immune serum did not precipitate c-Abl. In cells that only express c-Abl anti-Myc antibody did not precipitate c-Abl, confirming the interaction between c-Abl and Aph2 (data not shown). Reciprocally, c-Abl and associated proteins were precipitated with anti-c-Abl antibodies, separated on an 8% SDS-PAGE gel, and transferred to nitrocellulose membrane. The blot was probed with anti-Myc antibody (Fig. 3D). Aph2 was brought down by c-Abl, whereas the pre-immune serum did not bring down Aph2. In cells that only express Aph2, anti-c-Abl antibodies did not precipitate Aph2 (data not shown). These data suggest that there is an interaction between c-Abl and Aph2 in mammalian cells.
Aph2 is not an efficient substrate for c-Abl. In COS-7 cells co-expressing Aph2 and c-Abl, Aph2 was immunoprecipitated using anti-Myc antibody, and tyrosine phosphorylation was determined by Western blot using antibodies that specifically recognize phosphotyrosine. We could not detect any phosphorylated Aph2, although c-Abl autophosphorylation and Abi-1 phosphorylation could be detected easily using the same setting (data not shown).
Localization of Aph2-To study the function of Aph2, the cellular localization of the protein was determined by indirect fluorescent immunocytochemistry using the Myc-tagged Aph2. In COS-7 cells Aph2 was detected exclusively in the cytoplasm, especially in the perinuclear region, which is indicative of endoplasmic reticulum (Fig. 4, A and B). To confirm that Aph2 is localized in the ER, an ER-specific dye (DiOC6) was used. As shown in Fig. 4, C-E, strong co-staining exists between Aph2 and ER. We found no obvious ER retention signal in the Aph2 protein. We next tested for the co-localization of Aph2 and c-Abl. Fig. 4, F-H shows that c-Abl and Aph2 are co-localized, although c-Abl is also localized near the cell surface and in the nucleus of some cells. Our results are consistent with a recent report suggesting that Abl resides on ER and that upon ER stress Abl may translocate to mitochondria and induce apoptosis (30). The localization of Aph2 to the ER suggests that it may play a role related to the ER function of c-Abl.
Using various deletions, we found that the C-terminal portion of Aph2 (aa 226 -361) expressed separately was dispersed throughout the cytoplasm, in contrast to the strong perinuclear staining of other fragments of Aph2 (aa 1-226, 150 -361, and deletion of the zf-DHHC region) (Fig. 5). These data suggest that more than one signal for ER location may exist, with one residing in the N-terminal portion and another in the zf-DHHC region.
Ectopic Expression of Aph2 Leads to Apoptosis-Immunostaining for Aph2 in both COS-7 (Fig. 4, A and B) and NIH3T3 (data not shown) revealed an interesting aspect of Aph2 function. Both cell types expressing Aph2 were round and smaller, with the nuclei condensed. Some of the cells show nuclei with abnormal shapes, suggesting that theses cells are undergoing apoptosis (Fig. 4B). In these cells, Aph2 is organized in a unique structure that is exclusive of the nucleus. In some cells, nuclei were split into two by Aph2. The cells transfected with vector alone did not show this phenotype. To confirm that these cells were apoptotic, TUNEL assays were performed. As shown in Fig. 6A, COS cells transfected with vector were not apoptotic, whereas many Aph2 transfected cells were stained positive for apoptosis.
Aph2 and c-Abl Work Synergistically in the Induction of Apoptosis-To quantitate apoptosis, an assay was adopted in which the transfected surviving cells were counted (23). In this assay, cells are transfected with genes of interest and a con- FIG. 3. Interaction between aph2 and c-Abl. Deletion analysis of Aph2 using two-hybrid interaction. A and B, plasmids expressing different mutant forms of Aph2 were introduced into CTY10 -5D that was already expressing a LexA-Abl fusion. ␤-galactosidase activity was measured as an indicator for interaction. (ϩϩϩ), dark blue after 1 h at 30°C; (ϩ), light blue after 4 h; (Ϫ), no blue after 24 h. C, c-Abl can be co-immunoprecipitated with Aph2. COS cells co-expressing c-Abl and Aph2 were harvested 36 h after the start of transfection and were lysed in TNEN buffer. Aph2 was precipitated with anti-Myc antibody, and c-Abl was detected with anti-c-Abl antibody K-12. D, Aph2 co-immunoprecipitated with c-Abl. COS cells co-expressing c-Abl and Aph2 were lysed, c-Abl was precipitated with K-12, and Aph2 was detected with anti-Myc antibody 9E10. IP, immunoprecipitate. NLS, nuclear localization signal; PTK, protein tyrosine kinase. struct that expresses ␤-galactosidase. If the expression of the gene of interest causes apoptosis, the number of ␤-galactosidase-positive cells will be reduced. This method has been used successfully to study the effect of c-Abl on apoptosis and in other cases (10). In our experiments, we used both COS-7 and NIH3T3 cells. The data presented below are from COS-7 cells because of its high transfection efficiency, but similar results were obtained from these two cell lines. To confirm the apoptotic effect of c-Abl and Aph2 using the quantitative assay, COS cells were transfected with 1.0 g of DNA expressing c-Abl, 0.4 g of DNA expressing Aph2, or both, along with 0.2 g of pCR3.1cmv-galactosidase DNA per well in six-well plates. After 36 h, the plates were washed with PBS, fixed in paraformaldehyde/formalin, permeabilized, and stained with X-gal at 37°C. After the color had developed, plates were washed with and stored in PBS. ␤-galactosidasepositive cells were counted from five randomly chosen views under the magnification of ϫ400. Each experiment was repeated three times. Although the transfection efficiency varied slightly, the percentage of transfected cells that survived was similar (Fig. 6B). c-Abl at 1.0 g did not have any significant effect on apoptosis in COS-7 cells. When c-Abl was increased to 3.0 g, the proportion of apoptotic cells was dramatically increased to 40% (data not shown), consistent with previous results (10). Aph2 at 0.4 g was a stronger inducer of apoptosis with 25% of the transfected cells dead. Co-expression of c-Abl and Aph2 (1.0 and 0.4 g, respectively) had a much stronger effect on apoptosis (65% transfected cells dead) than the combination of individual expression, suggesting that c-Abl and Aph2 have a synergistic effect. This also was confirmed by our observation that more cells have condensed nuclei when c-Abl and Aph2 are co-expressed in NIH3T3 cells (data not shown). Similarly, both p53 and p73 were found to induce more apoptosis when co-expressed with c-Abl (10,31).
Both the N portion and the C portion of Aph2 were tested in the quantitative apoptosis assay. The N terminus Aph2 had a similar effect on apoptosis as the full-length protein, whereas the C terminus had little effect on apoptosis (Fig. 6C). Coexpression of c-Abl with the N terminus of Aph2 also had a synergistic effect, while the C-terminal Aph2 had a much weaker effect (Fig. 6C). These results suggest that the apoptotic effect of Aph2 can be executed by its N portion.
Overexpression of c-Abl leads to apoptosis in many cells lines, and this effect does not require kinase activity; the kinase-defective c-Abl mutant also causes apoptosis when overexpressed (10). To test whether the synergistic effect of c-Abl and Aph2 requires the kinase activity a c-Abl mutant, K290R, was used in similar experiments as shown in Fig. 6B. We did not detect a significant difference between wild type and the mutant c-Abl at 1.0 g per well (Fig. 6C). The mutant c-Abl also had a synergistic effect on apoptosis with Aph2, confirming that the kinase activity is not involved in this aspect of c-Abl function.
Expression of Aph2 and c-Abl Is Mutually Down-regulated-In the course of co-expression experiments of c-Abl and Aph2, we noted a profound effect of high level expression of each protein on the other. At a constant level of c-Abl DNA and increasing amounts of Aph2 DNA, we noticed that the c-Abl expression level was reduced dramatically by the expression of Aph2 (Fig. 7). Similarly, when cells were transfected with a fixed amount of Aph2 DNA, increasing the amount of c-Abl DNA dramatically reduced the level of Aph2 expression (data not shown). On the other hand, Abi-1, another c-Abl-interacting protein, did not down-regulate c-Abl expression when coexpressed (data not shown). These results suggest either that cells co-expressing Aph2 and c-Abl have a higher tendency to undergo apoptosis, that Aph2 and c-Abl proteins are mutually down-regulated, or both. The results shown in Fig. 6 favor the conclusion that down-regulation is attributable to a loss from the cultures of those cells expressing the two proteins, resulting from the synergistic effects of c-Abl and Aph2 on apoptosis induction. Aph2 is under the control of a hybrid SV40/HTLV promoter, and c-Abl is under the control of the adenovirus major late promoter. The down-regulation could not be explained by competition for transcriptional machinery. The C portion of Aph2, under the control of the same promoter, did not have a similar effect. The N portion of Aph2 did downregulate the expression of c-Abl (data not shown). These results on apoptosis induction (Fig. 6C), and down-regulation of c-Abl obtained with truncated forms of Aph2 also support our conclusion that the down-regulation is attributable to a loss from the cultures of those cells expressing both c-Abl and Aph2. DISCUSSION We have cloned and characterized a novel gene, aph2, encoding a c-Abl-interacting protein. This protein has a cysteine-rich motif (zf-DHHC) that is found in many proteins from yeast to human. Most of these protein sequences are predicted from genomic DNA sequences and therefore are not well characterized. Other cysteine-rich motifs include the better-known Zinc finger, Ring finger, and LIM domains. Although Zinc finger domains often are involved in protein-DNA interaction, both Ring finger and LIM are involved in protein-protein interactions. In addition, Ring finger domains also play an important role in protein ubiquitination and degradation. The zf-DHHC motif of Aph2 probably does not bind DNA because Aph2 was localized in the ER. Two yeast proteins (Akr1p and Erf2p) containing the motif were not localized in the nucleus either. The motif may be involved in disulfide bond formation inter-or intra-molecularly. Aph2 expressed in either bacteria or mammalian cells is prone to aggregation and runs as a Ͼ200 K D band on SDS-PAGE gels. Deletion of the domain facilitated the solubilization of Aph2.
The functions of zf-DHHC-bearing proteins may be diverse. Two such proteins from yeast have been characterized to date, Akr1p and Erf2p. Akr1p interacts physically with the pheromone receptor-coupled G protein, and a genetic interaction has been observed with cdc42p and also with an ARF GAP, Gcs1 (25)(26)(27)32). Erf2p is required for the proper localization and palmitoylation of Ras proteins in yeast (28). These proteins do not suggest any simple role for the motif. Another clue regarding zf-DHHC function is that in some proteins it coexists with ankyrin repeats, which are important for proteinprotein interactions, e.g. Akr1p and an uncharacterized protein (NP-014667) of S. cerevisiae, a hypothetical protein (T33870) of Caenorhabditis elegans, a hypothetical protein (AAF48554) of Drosophila melanogaster, and an uncharacterized protein (NP_061901) of human. Whether a functional connection exists between zf-DHHC and ankyrin repeats is unclear. Two more mammalian proteins containing zf-DHHC also have been deposited into GenBank TM . One is named as rec (reduced expression in cancer, accession number NP_057437), and the other is a protein from the human pancreatic library (C16ORF1). The function of either rec or C16ORF1 is not known.
Aph2 has a 53-amino acid stretch sharing homology with FIG. 7. Down-regulation of c-Abl by expression of Aph2. COS cells were transfected with 0.5 g of DNA expressing type IV c-Abl, 1.0 g of DNA expressing c-Abl, and an increasing amount of DNA expressing Myc-tagged Aph2. Cells were collected 36 h from the start of transfection and were lysed in TNEN buffer. 20 g of total proteins were separated on an 8% SDS-PAGE gel and transferred to Immobilon nitrocellulose membrane. The blot was probed with 9E10 to detect Aph2 and with K-12 to detect c-Abl. ATP:guanido phosphotransferase consensus sequence found in creatine and arginine kinases. The homology resides in domain I (N-terminal) of the creatine kinase family, comprising ␣-helices 4 and 5, which may be involved in substrate interaction. Domain II is important for ATP binding and enzymatic activity (29). Because Aph2 has homology to only 53 amino acids of the 380-amino acid consensus sequence, it is unlikely that Aph2 plays a role in the energy buffer or transport system.
How does overexpression of Aph2 induce apoptosis? Based on the localization of Aph2 and the fact that cyclosporin A could not protect cells from apoptosis induced by Aph2 (data not shown), it is likely that Aph2 induces apoptosis in an ER stress-related pathway. Accumulation of misfolded proteins or an alteration in calcium homeostasis causes ER stress, which leads to apoptosis of cells if the damages are not fixed. Caspase 12 is localized in the ER and is responsible for ER stressinduced apoptosis. Bcl2, an anti-apoptotic protein, also was found to be located in the ER in addition to mitochondria. Depletion of ER calcium by the overexpression of Bcl2 plays an important anti-apoptotic role (33). Overexpression of Aph2 somehow may stress the ER and subsequently induce apoptosis. Whether Aph2 affects protein folding or calcium homeostasis needs to be studied. c-Abl has been implicated to play a role in ER stress-induced apoptosis (30). c-Abl was found to be located in the ER, and treatment with tunicamycin or Brefeldin A led to translocation of c-Abl from the ER to mitochondria.
Aph2 has a synergistic effect on apoptosis induction with c-Abl. What are the mechanisms that underlie the synergy? Unfortunately, even the mechanisms under which overexpression of c-Abl induces apoptosis are still unclear. One hypothesis is that apoptosis is caused by the nuclear c-Abl, while the cytoplasmic c-Abl may play an anti-apoptotic effect, as seen with v-Abl and Bcr-Abl, both of which are localized solely in the cytoplasm (1). Recent experiments demonstrated that Bcr-Abl, once brought into the nucleus, could induce apoptosis, thus supporting this hypothesis (16). We propose that Aph2 may block the anti-apoptotic effect of cytoplasmic c-Abl by physical interaction while allowing the nuclear c-Abl to act. There are other c-Abl-interacting proteins that also are involved in apoptosis. p53 and p73, both of which interact with c-Abl and induce apoptosis when ectopically expressed, have a synergistic effect on apoptosis induction (10,31).