ArgBP2γ Interacts with Akt and p21-activated Kinase-1 and Promotes Cell Survival*

Akt/protein kinase B is a major cell survival pathway through phosphorylation of proapoptotic proteins Bad and Bax and of additional apoptotic pathways linked to Forkhead proteins glycogen synthase kinase-3β and ASK1. To further explore the mechanism by which Akt regulates cell survival, we identified an Akt interaction protein by yeast two-hybrid screening. It is highly homologous to ARG-binding protein 2 (ArgBP2) with splicing exon 8 of the coding region of the ArgBP2. As two splicing isoforms (ArgBP2α and -β) of ArgBP2 have been identified (Wang, B., Golemis, E. A., and Kruh, G. D. (1997) J. Biol. Chem. 272, 17542–17550), it was named ArgBP2γ. ArgBP2γ contains four Akt phosphorylation consensus sites, a SoHo motif, and three Src homology (SH) 3 domains and binds to C-terminal proline-rich motifs of Akt through its first and second SH3 domains. It also interacts with p21-activated protein kinase (PAK1) via its first and third SH3 domains, indicating the SH3 domains of ArgBP2γ as docking sites for Akt and PAK1. Akt phosphorylates ArgBP2γ in vitro and in vivo. Expression of ArgBP2γ induces PAK1 activity and overrides apoptosis induced by ectopic expression of Bad or DNA damage. Nonphosphorylatable ArgBP2γ-4A and SH3 domain-truncated mutant ArgBP2γ inhibit Akt-induced PAK1 activation and reduce Akt and PAK1 phosphorylation of Bad and antiapoptotic function. These data indicate that ArgBP2γ is a physiological substrate of Akt, functions as an adaptor for Akt and PAK1, and plays a role in Akt/PAK1 cell survival pathway.

dependent manner (8 -13). Full activation of the Akt requires their phosphorylation at Thr 308 (Akt1), Thr 309 (Akt2), or Thr 305 (Akt3) in the activation loop and Ser 473 (Akt1), Ser 474 (Akt2), or Ser 472 (Akt3) in the C-terminal activation domain (14). Akt1, one of the most studied isoforms, which was originally designated as Akt, has been shown to be an important regulator of several cellular processes, including proliferation, programmed cell death, angiogenesis, and metabolism. There is considerable interest in the identities of the protein substrates of the kinase activity of Akt, which are presumably responsible for mediating its effects on cell behavior. More than 20 substrates have been identified in the literature, most of which have the Akt phosphorylation consensus site RXRXX(S/T). The most widely studied substrates are glycogen synthase kinase 3␤, the proapoptotic protein Bad, and the members of the Forkhead family of proteins. Glycogen synthase kinase 3␤ phosphorylation and inactivation by Akt/PKB can account for some of the metabolic effects of Akt (10), whereas Bad can account for some of the survival-promoting effects (15,16). The most convincingly proven substrates for Akt are the Forkhead transcription factors; biochemical, cell biological, and genetic data combine to make a very strong case for these proteins being direct targets in mammals, worms, and flies. Phosphorylation of Forkhead proteins results in their exclusion from the nucleus, causing loss of transcriptional activity and decreased expression of proteins promoting cell death and cell cycle arrest (17). Other major Akt substrates include MDM2, IB kinase, 21 Cip1/WAF1 , p27 Kip1 , ASK1, estrogen receptor, androgen receptor, TSC2, and XIAP (18 -30).
In addition to acting as a kinase toward many substrates involved in its function, Akt forms complexes with other proteins that are not substrates but rather act as modulators of Akt activity and function. Several such Akt interaction proteins have been identified, including TCL1, CTMP, APPL, and TRB3. TCL1 is an oncoprotein that interacts with the pleckstrin homology domain of Akt. It activates Akt by increasing Akt oligomerization and then promotes Akt nuclear translocation (31,32). By contrast, CTMP has been shown to interact with the C-terminal regulatory region of Akt and inhibit phosphorylation of Akt on Ser 473 , leading to a decrease of Akt kinase activity (33). TRB3 is a mammalian homolog of Drosophila tribbles and inhibits insulin-stimulated Akt activation by physical interaction with Akt (34). APPL is an adaptor protein that contains a pleckstrin homology domain, a phosphotyrosine-binding domain, and a leucine zipper motif and binds to both Akt and p110 catalytic subunit of phosphatidylinositol 3-kinase. However, the function of APPL is currently unknown (35). Furthermore, Hsp90 and Hsp27 also form a complex with Akt and induce Akt activation (36,37). It has also been shown that Akt interacts with JIP1 scaffold protein to inhibit excitotoxin-induced JNK activation in an Akt kinaseindependent manner, thus providing a novel mechanism for Akt antiapoptotic function (38).
In the present report, we have identified an Akt interacting protein, ArgBP2␥, that functions as an adaptor for Akt and PAK1 and is phosphorylated by Akt. In addition, ArgBP2␥ activates PAK1 by direct interaction with PAK1 and mediates Akt-induced PAK1 activation and Akt/PAK1 phosphorylation of Bad.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Screening and Expression Constructs-Yeast twohybrid system 3 was employed to identify Akt interaction protein(s) using the C-terminal regulatory region of Akt as bait following the manufacturer's procedure (Clontech). A human fetal brain library (Clontech) was screened. Full-length cDNA of ArgBP2␥, isolated from a human skeletal muscle library, was subcloned into FLAG-tagged pcDNA3 vector. ArgBP2␥ and Akt mutants were created with the QuikChange multiple site-directed mutagenesis kit (Stratagene). The cytomegalovirus-based expression constructs encoding HA-tagged Akt and Myc-tagged PAK1 have previously been described (8,39).
Glutathione S-transferase (GST) Fusion Protein and Generation of Anti-ArgBP2␥ Antibody-Different portions of ArgBP2␥, including each SH3 domain and the regions containing each Akt phosphorylation site, were cloned into the prokaryotic expression vector pGEX-4T1. Expression and purification of the GST fusion protein were carried out as previously described (40). Polyclonal anti-ArgBP2␥ antibodies were raised in New Zealand White rabbits. Approximately 300 g of GST fusion protein (GST⅐ArgBP2␥/300 -382 and GST⅐ArgBP2␥/500 -573) was used to immunize rabbits every 2 weeks; rabbits were bled 10 days after each booster injection. The anti-ArgBP2␥ antibodies were affinity purified with Affi-Gel protein A (Bio-Rad).
Cell Lines and Transfection-HeLa, COS, and human embryonic kidney (HEK) 293 cells were cultured at 37°C and 5% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. The cells were seeded in 60-mm Petri dishes at a density of 0.5 ϫ 10 6 cells/dish. Following incubation overnight, the cells were transfected with 2 g of DNA/dish using calcium phosphate or Lipofectamine Plus (Invitrogen).
Immunoprecipitation and Immunoblotting-Cells were lysed in a lysis buffer containing 20 mM Tris-HCl (pH 7.5), 137 mM NaCl, 15% (v/v) glycerol, 1% Nonidet P-40, 2 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin and leupeptin, 2 mM benzamidine, 20 mM NaF, 10 mM NaPPi, 1 mM sodium vanadate, and 25 mM ␤-glycerolphosphate. Lysates were subjected to immunoprecipitation and immunoblotting analysis as previously described (30). Briefly, lysates were precleared with protein A-protein G (2:1) -agarose beads. Following the removal of the beads by centrifugation, lysates were incubated with appropriate antibodies in the presence of protein A-protein G (2:1) -agarose beads for 2 h at 4°C. After being washed, the immunoprecipitates were subjected to in vitro kinase assay. Protein expression was determined by probing Western blot of the immunoprecipitates or total cell lysates with the appropriate antibodies as noted in the figure legends.
GST Pulldown Assay-Glutathione-agarose beads coupled to GST alone, GST-SH3 domains of ArgBP2␥, or GST-Akt-pleckstrin homology, -kinase domain, and -C-tail motif were incubated with whole cell lysate (ϳ800 g of protein) for 2 h at 4°C. After being washed four times with lysis buffer, the beads were subjected to Western blot analysis with appropriate antibodies.
In Vitro Kinase Assay-Akt and PAK1 kinase assays were performed as previously described (13,26). Briefly, reactions were carried out in the presence of 10 Ci of [␥- 32  tein were used as exogenous substrate, respectively. After incubation at room temperature for 30 min, the reactions were stopped by adding protein loading buffer and separated in SDS-PAGE gels. Each experiment was repeated three times. The relative amounts of incorporated radioactivity were determined by autoradiography and quantified with a Phosphorimager (Amersham Biosciences).
TUNEL Assay-HeLa cells were transfected with appropriate plasmids as noted in the figure legends, seeded into 60-mm diameter dishes, and grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum for 24 h. Following treatment with or without VP16, doxorubicin, or staurosporine, apoptotic cells were determined by terminal deoxynucleotidyltransferase-mediated dUTP nick-end labeling (TUNEL) using an in situ cell death detection kit (Roche Applied Science). These experiments were performed in duplicate.
In Vivo [ 32 P]P i Labeling-COS7 cells were co-transfected with FLAG-ArgBP2␥, wild type, and constitutively active Akt or pcDNA3 and labeled with [ 32 P]P i (0.5 mCi/ml) in phosphate-and serum-free minimum essential medium for 4 h. Cell lysates were subjected to immunoprecipitation with anti-FLAG antibody (Sigma). The immunoprecipitates were separated by SDS-PAGE and transferred to membrane. The phosphorylated ArgBP2␥ band was examined by autoradiography.

Identification of Akt/PKB␣-binding Protein
ArgBP2␥-In an attempt to identify protein(s) specifically interacting with Akt1, the C-terminal regulatory domain of Akt1 (410 -480 amino acids), the most diverse region between three isoforms of Akt, was used as bait in a yeast two-hybrid screening. A human fetal brain cDNA library was used in this screen because Akt1 is highly expressed in brain (1,9). A total of 32 clones that specifically interacted with the bait were identified. Sequence analysis revealed that two of the clones contained overlapping sequences of a cDNA (Fig. 1A). The largest clone contained an 182-amino acid open reading frame with a conserved SH3 domain, named ArgBP2␥ as it matches the sequence of exons 5-11 of the ArgBP2␣ and the ArgBP2␤ with splicing exon 8 ( Fig. 1A) (40). Additional cDNA clones were isolated from a human brain cDNA library by plaque hybridization using the largest clone as radiolabeled probe. Sequence analysis revealed that the full length of the open reading frame of ArgBP2␥ encoded a 620-amino acid protein. ArgBP2␥ contains a SoHo (sorbin homology) domain, a serine/threonine-rich region, three SH3 motifs, and four Akt phosphorylation consensus sites (Fig.  1A). To confirm the association of Akt with ArgBP2␥ that was identified by yeast two-hybrid system, HEK293 cells were cotransfected with FLAG-ArgBP2␥ and HA-Akt. Immunoprecipitation was performed with anti-FLAG and detected with anti-HA antibody or vice versa. As shown in Fig. 1B, HA-Akt was detected in the FLAG-ArgBP2␥ immunoprecipitates and ArgBP2␥ was coimmunoprecipitated by anti-HA antibody. In addition, immunoprecipitation revealed that Akt also interacts with ArgBP2␣ and ArgBP2␤ ( Fig. 1C and data not shown). Further, anti-ArgBP2␥ antibody was generated and its specificity was determined by immunoprecipitation and/or immunoblotting. Fig. 1, D and E, shows that this antibody recognized both endogenous and exogenous ArgBP2␥ and could be used for immunoprecipitation. The interaction between endogenous Akt and ArgBP2␥ was detected in HeLa cells (Fig. 1F). However, it failed to detect Akt association with ArgBP2␣ and ArgBP2␤ in HeLa cells (Fig. 1F), suggesting that the HeLa cells predominantly express the ArgBP2␥ isoform.
Definition of Domains Involved in Akt-ArgBP2␥ Interaction-Sequence alignment analysis showed that ArgBP2␥ belongs to a newly defined vinexin adaptor protein family (41). Although the function of the SoHo domain is currently unclear, the SH3 domain is known to bind to proline-rich sequences containing the PXXP core sequence. As the C-terminal region of Akt1, which was used as the bait for yeast two-hybrid screen, contains two proline-rich motifs, we examined whether SH3 domains of ArgBP2␥ are required for interaction with Akt. Fig.  2A shows that Akt was detected in the ArgBP2␥, but not ArgBP2␥-⌬SH3, immunocomplex. We next performed GST pulldown assay to determine which SH3 domain(s) of ArgBP2␥ interacts with Akt and which motif of the C terminus of Akt binds to ArgBP2␥. Cell lysates were incubated with GST fusion proteins derived from different regions of Akt and ArgBP2␥ that were immobilized on GSH beads. GST alone was used as a control. After stringent washes, the GST⅐Akt pulldown products and GST⅐ArgBP2␥ bound proteins were subjected to immunoblotting analysis with anti-ArgBP2␥ or anti-Akt antibody, respectively. As illustrated in Fig. 2, B and C, the C-terminal regulatory region of Akt interacts with the first and second SH3 domains of ArgBP2␥.
As SH3 domain binds to the PXXP motif and Akt contains two proline-rich regions ( 424 PFKP 427 and 467 PHFP 470 ) within its C terminus (1, 2), we next defined the ArgBP2␥ binding sites of Akt. Akt-P424A, Akt-P467A, and Akt-P424/467A were created by converting proline(s) to alanine(s). GST pulldown assay showed that mutation of either Pro 424 or Pro 467 abrogated Akt interaction with ArgBP2␥, indicating that both proline-rich motifs are required for binding to the SH3 domain of ArgBP2␥ (Fig. 2D). Further, in vitro Akt kinase assay and Western blotting analysis with anti-phospho-Ser 473 Akt antibody showed that ArgBP2␥ has no significant effect on activation of Akt ( Fig. 2E and data not shown).
Akt Phosphorylates ArgBP2␥ in Vitro and in Vivo-Because ArgBP2␥ contains four putative Akt phosphorylation consensus sites (Ser 232 , Thr 234 , Ser 305 , and Thr 379 ), we next determined whether ArgBP2␥ is phosphorylated by Akt. GST fusion proteins containing each putative Akt phosphorylation site of ArgBP2␥ were bacterially expressed and were used as substrate for in vitro Akt kinase assay. Triple experiments showed that all four sites of ArgBP2␥ were phosphorylated by Akt with lower levels of Ser 305 (Fig. 3A).
To determine whether Akt phosphorylates ArgBP2␥ in vivo, COS7 cells were transfected with FLAG-ArgBP2␥ together with wild type, dominant negative, or constitutively active Akt and labeled with [ 32 P]orthophosphate. The labeled ArgBP2␥ was immunoprecipitated with anti-FLAG antibody. Western blot revealed that both constitutively active Akt and insulinlike growth factor 1-induced Akt phosphorylated ArgBP2␥ (Fig.  3B). Moreover, an increased density of a slower migration band, i.e. the phosphorylation isoform of ArgBP2␥, was observed in constitutively active Akt-transfected and insulin-like growth factor 1-stimulated cells (Fig. 3B). These results indicate that ArgBP2␥ is a physiological substrate of Akt.
ArgBP2␥ Interacts with and Activates PAK1-A previous study has shown that Akt activates PAK1 by an uncharacterized mechanism (42). As PAK1 contains four proline-rich motifs in its N-terminal regulatory region and is activated by NCK through interaction between its proline-rich motif and the SH3 domain of NCK, we assumed that ArgBP2␥ could mediate Akt-induced PAK1 activation by interaction with PAK1. Coimmunoprecipitation showed that ArgBP2␥ and PAK1 interact at physiological protein concentration in HeLa cells (Fig. 4A). To define the binding domains, various truncated PAK1 and ArgBP2␥ plasmids were created. Immunoprecipitation and GST pulldown experiments revealed that the first and third SH3 domains of ArgBP2␥ bind to the first proline region of PAK1 (Fig. 4, B-D). Fig. 2B shows the first and second SH3 domains of ArgBP2␥ associate with Akt. Taken collectively, these data indicate that the SH3 domains of ArgBP2␥ function as docking sites for Akt and PAK1.
To determine the effect of the interaction of PAK1 and ArgBP2␥ on PAK1 kinase activity, HEK293 cells were transfected with Myc-PAK1 and immunoprecipitated with anti-Myc antibody. The immunoprecipitates were incubated with individual GST-fused SH3 domains of ArgBP2␥, as well as the second SH3 domain of NCK as a positive control, and subjected to in vitro kinase assay using myelin basic protein as a substrate. As shown in Fig. 5A, the first and third SH3 domains of ArgBP2␥, as well as NCK, stimulated PAK1 activity, indicating that ArgBP2␥ is able to directly activate PAK1 in vitro through their interaction.
ArgBP2␥ Mediates Akt-induced PAK1 Activation, and Both SH3 Domain and Akt Phosphorylation of ArgBP2␥ Are Required for Activation of PAK1 in Vivo-We next examined whether ArgBP2␥ activates PAK1 and mediates Akt-stimulated PAK1 activation in the intact cells. In vitro kinase analysis showed that ectopic expression of wild type ArgBP2␥ or constitutively active Akt is sufficient to activate PAK1. Coexpression of both exhibited a synergistic effect on activation of PAK1, and the level of PAK1 activity is almost similar to that of constitutively active PAK1-T423E (Fig. 5B). These data suggest that ArgBP2␥ is a positive regulator upstream PAK1 and mediates Akt-induced PAK1 activation.
We further determined whether Akt activation of PAK1 mediated by ArgBP2␥ depends on their binding. A deletion mutation of ArgBP2␥ (ArgBP2␥-⌬SH3) that binds neither to PAK1 nor Akt was created by truncation of its C-terminal three SH3 domains ( Figs. 2A and 4B). The kinase assay revealed that ArgBP2␥-⌬SH3 not only failed to activate PAK1 but also inhibited constitutively active Akt-induced PAK1 (Fig. 5B). Because Akt phosphorylates ArgBP2␥, the effect of Akt phosphorylation of ArgBP2␥ on PAK1 activation was evaluated by in vitro PAK1 kinase assay in the cells transfected with Akt and nonphosphorylatable ArgBP2␥-4A, prepared by converting the Akt phosphorylation serine/threonine sites to alanines. Fig. 5B   FIG. 3. Akt phosphorylates ArgBP2␥ in vitro and in vivo. A, in vitro kinase assay. HEK293 cells were transfected with the indicated Akt expression plasmids and immunoprecipitated with anti-HA antibody. The immunoprecipitates were subjected to in vitro kinase assay using the GST-fused putative Akt phosphorylation site of ArgBP2␥ as substrate. B, in vivo [ 32 P]orthophosphate labeling. COS7 cells were transfected with the indicated plasmids and incubated with [ 32 P]orthophosphate for 4 h. Immunoprecipitates were prepared with anti-FLAG antibody and separated by SDS-PAGE. Following transfer, the membrane was exposed to a film (top) and detected with anti-FLAG antibody (bottom).
shows that the nonphosphorylatable ArgBP2␥-4A failed to induce PAK1 activation even though SH3 domains of ArgBP2␥ were intact. Moreover, ArgBP2␥-4A considerably reduced constitutively active Akt-stimulated PAK1 activation. Taken collectively, we concluded that ArgBP2␥ mediates Akt activation of PAK1 and that SH3 domains of ArgBP2␥ are critical for PAK1 activation induced by Akt and/or ArgBP2␥. Furthermore, Akt activation of PAK1 in the intact cell requires phosphorylation of ArgBP2␥ by Akt, which may result in the conformation change of ArgBP2␥ that leads to exposing the SH3 domains to interaction with PAK1.
ArgBP2␥ Induces Bad Phosphorylation and Regulates Akt and PAK1 Phosphorylation of Bad-Previous studies demonstrate that PAK1 promotes cell survival through phosphorylation of Bad at Ser 112 and Ser 136 (42,43) and that the antiapoptotic function of Akt is mediated by phosphorylation of Bad at Ser 136 (15,16). As ArgBP2␥ is an adaptor for Akt and PAK1 and is phosphorylated by Akt, we next examined the effects of ArgBP2␥ on Bad phosphorylation by PAK1 and Akt using in vitro kinase assay. As shown in Fig. 6A, ectopic expression of wild type ArgBP2␥ enhanced PAK1 phosphorylation of Bad. In  ). B, the SH3 domain of ArgBP2␥ binds to PAK1. HEK293 cells were transfected with FLAG-ArgBP2␥ or FLAG-ArgBP2␥-⌬SH3. Following 48 h of culture, cells were lysed, immunoprecipitated with anti-FLAG antibody, and detected with anti-PAK1 antibody. C, the first proline-rich motif of PAK1 is required for interaction with ArgBP2␥. Myc-tagged proline-rich region deletion mutants of PAK1, as shown in the top panel, were introduced into HEK293 cells. Immunoprecipitation was carried out with anti-Myc antibody and immunoblotted with anti-ArgBP2␥ antibody. D, GST pulldown assay. The cell lysates prepared from Myc-PAK1-transfected HEK293 cells were incubated with the indicated GST fusion proteins. The GST-NCK-SH3 domain, which is known to bind to PAK1, was used as positive control. Bound PAK1 was detected by Western blot analysis of GST pulldown products with anti-Myc antibody.  addition, the coexpression of constitutively active Akt and PAK1, especially in combination with wild type ArgBP2␥, further elevated Bad phosphorylation (Fig. 6, A and B). However, Akt/PAK1-induced Bad phosphorylation was reduced by expression of either nonphosphorylatable or SH3 domain deletion mutant ArgBP2␥ (Fig. 6B), suggesting that ArgBP2␥ has a role in regulation of Akt and PAK1 phosphorylation of Bad through its interaction with Akt and PAK1 in a phosphorylation-dependent manner.
To investigate the effects of ArgBP2␥ on Akt1/PAK1 phosphorylation of Bad intact cells, HEK293 cells were transfected with HA-Bad and different forms of ArgBP2␥ with or without PAK1 and Akt. HA-Bad immunoprecipitates were subjected to Western blot analysis with anti-phospho-Bad-Ser 112 and Ser 136 antibodies. As shown in Fig. 6C, co-expression of PAK1 with wild type ArgBP2␥ induced Bad phosphorylation at Ser 112 and Ser 136 to a similar level of constitutively active PAK1. Consistent with previous findings (15,16,42,43), ectopic expression of constitutively active Akt induced phosphorylation of Bad only at Ser 136 , whereas constitutively active PAK1 phosphorylated Bad at both Ser 112 and Ser 136 . However, co-expression of wild type ArgBP2␥, Akt, and PAK1 significantly induced Bad phosphorylation at Ser 112 and Ser 136 , whereas expression of either nonphosphorylatable or SH3 domain-truncated ArgBP2␥ failed to enhance and, indeed, inhibited the Bad phosphorylation induced by PAK1 and constitutively active Akt (Fig. 6, C and  D). These data indicate that Akt and PAK1 phosphorylation of Bad is positively regulated by ArgBP2␥.
ArgBP2␥ Reduces Cell Death Induced by Ectopic Expression of Bad or DNA Damage-We next examined the effects of ArgBP2␥ on Bad-induced programmed cell death. HEK293 cells were transiently transfected with Bad, Bad/ArgBP2␥, or pcDNA3 vector alone. TUNEL assay was performed after 48 h of the transfection. Triple experiments revealed that Badtransfected HEK293 cells underwent apoptosis. However, ArgBP2␥ partially inhibited the apoptosis induced by Bad (Fig.  7A). Further, the effect of ArgBP2␥ on PAK1 antiapoptotic function was further investigated. Ectopic expression of PAK1 alone partially inhibited Bad-induced apoptosis. Co-expression of ArgBP2␥ and PAK1, however, exhibited more antiapoptotic FIG. 7. ArgBP2␥ protects cells from cell death induced by ectopic expression of Bad and DNA damage. A and B, TUNEL assay. HEK293 cells were transfected with the indicated expression plasmids. Apoptotic cells were detected by TUNEL assay and quantified 48 h after transfection. C, HeLa cells were stably transfected with wild type and nonphosphorylatable ArgBP2␥-4A. The cells transfected with pcDNA3 alone were used as control. After treatment with VP16, doxorubicin, or staurosporine, cell viability was evaluated with MTT assay. Each experiment was repeated three times. D, schematic illustration of ArgBP2␥ effects on Akt and PAK1 pathway and Bad phosphorylation. effect than that of expression of either ArgBP2␥ or PAK1 alone. Moreover, nonphosphorylatable ArgBP2␥-4A and SH3 domaintruncated ArgBP2␥ had no significant effect on PAK1-inhibited apoptosis (Fig. 7B). Ectopic expression of wild type, but not nonphosphorylatable, ArgBP2␥ also exhibited antiapoptotic effects on DNA damage agent-induced programmed cell death, which includes VP16, doxorubicin, and staurosporine (Fig. 7C). DISCUSSION In the present study, we identified an Akt interaction protein, ArgBP2␥, that functions as an adaptor protein that binds to Akt and PAK1. ArgBP2␥ is a physiological substrate of Akt and mediates Akt activation of PAK1 and phosphorylation of Bad in an interaction-and phosphorylation-dependent manner. In addition, we observed that ArgBP2␥ interacts with other isoforms of Akt, including Akt2 and Akt3 (data not shown), even though the C terminus of Akt1, the most diverse region among three isoforms of Akt, was used as bait for yeast two-hybrid screening.
ArgBP2␥ contains four Akt phosphorylation consensus sites, three of which are highly phosphorylated by Akt and the other (Ser 305 ) with a lower phosphorylation level (2). However, nonphosphorylatable ArgBP2␥-4A, but not ArgBP2␥-3A, abrogates the function of ArgBP2␥ in regulation of Akt and PAK1 signaling. This suggests that Akt phosphorylation of four serine/ threonine sites of ArgBP2␥ is required for its normal cellular function. A recent study has demonstrated that CTMP binds to the C-terminal region of Akt and inhibits Akt kinase activity by decreasing phosphorylation of Ser 473 of Akt (33). Unlike CTMP, however, ArgBP2␥ has no effect on Akt activation even though it interacts with the C-terminal region of Akt.
In addition to the N-terminal serine/threonine-rich region, a major characteristic of ArgBP2␥ is that it contains an N-terminal SoHo domain and three SH3 domains in the C-terminal region. Based on sequence and structure similarity, ArgBP2␥ belongs to a recently identified adaptor protein family that includes vinexin and CAP/ponsin (41). This protein family shares one SoHo domain in the N-terminal region and three SH3 domains in the C termini. The SoHo domain was defined because this region exhibits a high degree of similarity to the biologically active peptide sorbin, which consists of a 153-amino acid polypeptide (41). The function of the SoHo domain has not been documented, but a recent study shows that this domain binds to the membrane protein flotillin that is crucial for the localization of SH3-binding proteins, such as Cbl, to the lipid raft and propagation of the downstream signal (44). Accumulated studies show that members of this adaptor protein family interact with a number of signal molecules through their SH3 domains to regulate cell survival, growth, adhesion, and cytoskeletal organization (41). For instance, human vinexin has been shown to bind to PAK to regulate the anchorage dependence of extracellular signal-regulated kinase 2 activation (41). ArgBP2␣ and -␤ interact with the proline-rich regions of Arg and c-Abl and are phosphorylated in v-Abl-transformed cells (40). We have demonstrated in this study that SH3 domains of ArgBP2␥ interact with C-terminal proline-rich regions of Akt as well as the first proline-rich motif of PAK1.
PAK1 is implicated in the regulation of several cellular processes, including activation of LIM kinase to regulate cytoskeleton organization (45), phosphorylation of Bad (42), and stimulation of NFB and mitogen-activated protein kinase pathways (46,47) to promote cell survival and cell growth. PAK1 is composed of a C-terminal catalytic domain and an N-terminal regulatory region that contains a p21-binding domain (PBD) and proline-rich regions. It has been shown that PAK1 kinase activity is repressed by an intramolecular interaction between the regulatory and catalytic domains and acti-vated by GTPase Rac or cdc42 binding to the PBD domain and by NCK interaction with the first proline-rich region (38,39,48). A previous study shows that constitutively active Akt induces PAK1 activation (42). However, the underlying detailed molecular mechanism is currently unknown. It has recently been shown that Akt modulates the association of PAK1 with Nck by phosphorylation of serine 21 of PAK1 (49). In the present report, we provided evidence that ArgBP2␥, in addition to direct activation of PAK1, mediates Akt-induced PAK1 activation through a phosphorylation-and interaction-dependent mechanism (Fig. 7D).
Previous studies have demonstrated that Bad is a primary target of Akt and PAK1 and is phosphorylated by Akt at Ser 136 and by PAK1 at Ser 112 and Ser 136 , which results in reduced interaction between Bad and Bcl-2 or Bcl-x L and increased association of Bad with 14 -3-3 (50). We demonstrated in this study that ArgBP2␥ overrides Bad-and DNA damage-induced cell death. As an adaptor protein, ArgBP2␥ enhances PAK1 and Akt phosphorylation of Bad at both Ser 112 and Ser 136 , indicating that ArgBP2␥ could play an important role in Akt/ PAK1 cell survival pathway.