Direct Interaction between the Cytoplasmic Tail of ADAM 12 and the Src Homology 3 Domain of p85α Activates Phosphatidylinositol 3-Kinase in C2C12 Cells*

ADAM 12, a member of the ADAM family of transmembrane metalloprotease-disintegrins, has been implicated previously in the differentiation of skeletal myoblasts. In the present study, we show that the cytoplasmic tail of mouse ADAM 12 interactsin vitro and in vivo with the Src homology 3 domain of the p85α regulatory subunit of phosphatidylinositol (PI) 3-kinase. By site-directed mutagenesis, we have identified three p85α-binding sites in ADAM 12 involving PXXP motifs located at amino acids 825–828, 833–836, and 884–887. Using green fluorescent protein (GFP)-pleckstrin homology (PH) domain fusion protein as a probe for PI 3-kinase lipid products, we have further demonstrated that expression of ADAM 12 in C2C12 cells resulted in translocation of GFP-PH to the plasma membrane. This suggests that transmembrane ADAM 12, by providing docking sites for the Src homology 3 domain of p85α, activates PI 3-kinase by mediating its recruitment to the membrane. Because PI 3-kinase is critical for terminal differentiation of myoblasts, and because expression of ADAM 12 is up-regulated at the onset of the differentiation process, ADAM 12-mediated activation may constitute one of the regulatory mechanisms for PI 3-kinase during myoblast differentiation.

ADAMs 1 (proteins containing a disintegrin and metalloprotease) are a family of transmembrane or secreted glycoproteins that have been implicated in cell surface proteolysis, adhesion, and cell-cell communication (1)(2)(3). A typical ADAM protein contains an N-terminal pro-domain, a metalloprotease domain, a disintegrin-like domain, a cysteine-rich region, and usually an EGF-like repeat, a single transmembrane domain, and a cytoplasmic tail. ADAM 12 has been implicated in differentiation of skeletal muscle precursor cells (myoblasts) in vitro (4) and in vivo (5)(6)(7)(8). ADAM 12 expression has been shown to be dramatically up-regulated during embryonic muscle develop-ment (6) and in regeneration of adult muscle following injury (7,8). The extracellular portion of ADAM 12 contains a zincdependent metalloprotease (9) that is negatively regulated by the presence of the pro-domain (10). The cysteine-rich domain (11,12), disintegrin-like domain (13), or the two domains together (14) have been demonstrated to mediate cell-cell adhesion and communication. The intracellular domain of ADAM 12 has recently been shown to interact with actin cytoskeleton via ␣-actinin-2 (7) and ␣-actinin-1. 2 Moreover, the cytoplasmic tail of ADAM 12 contains several Src homology 3 (SH3) binding motifs, and therefore it has been anticipated to interact with SH3 domain-containing proteins. Indeed, it has been recently demonstrated that ADAM 12 binds to the SH3 domain of nonreceptor protein kinase Src (15,16) and to an adaptor protein, Grb2 (15). Moreover, interaction with ADAM 12 led to stimulation of the enzymatic activity of Src (16).
Phosphatidylinositol 3-kinase (PI 3-kinase) is essential for terminal differentiation of skeletal muscle cells. Two specific and structurally unrelated inhibitors of PI 3-kinase, LY294002 and wortmannin, block myoblast exit from the cell cycle, inhibit expression of muscle specific genes, and abolish the capacity of myoblasts to form myotubes (17,18). Expression of dominantnegative mutants of PI 3-kinase inhibits myoblast fusion and biochemical differentiation (18,19). Moreover, expression of a constitutively active form of PI 3-kinase encoded by a viral oncogene, v-p3k, strongly enhances differentiation and fusion of cultured myoblasts, suggesting that the cellular PI 3-kinase constitutes a rate-limiting step of myogenesis in vitro (18). Insulin growth factors, the well known inducers of myogenic differentiation (20 -22) and potent stimulators of myoblast survival (23), have recently been shown to exert their function through activation of the PI 3-kinase-dependent signaling pathways (24,25). Finally, NF-B and nitric-oxide synthase, identified as downstream effectors of PI 3-kinase in myoblasts, were shown to be critical for myogenic differentiation (26).
Based on their structures, lipid specificity, and modes of regulation, PI 3-kinases can be divided into three classes (27)(28)(29). Class I enzymes are heterodimers composed of an ϳ110-kDa catalytic subunit and a regulatory subunit. Class I can be further divided into subclasses IA and IB, which are regulated by tyrosine kinases and G protein-coupled receptors, respectively. Each of the class IA regulatory subunits contains two SH2 domains that bind to phosphotyrosine residues in activated tyrosine kinase receptors or adaptor proteins and play critical roles in translocation of the cytosolic PI 3-kinases to the plasma membrane. In addition, two of the class IA regulatory subunits, p85␣ and p85␤, contain a single SH3 domain. Importantly, a p85␣-associated PI 3-kinase appears to be indispensable for myogenesis (18,19).
In the present study, we show that the cytoplasmic tail of ADAM 12 interacts with the SH3 domain of the p85␣ regulatory subunit of PI 3-kinase in vitro and in vivo. By site-directed mutagenesis, we mapped the p85␣-binding sites to three different regions in ADAM 12 cytoplasmic tail. Using green fluorescent protein-pleckstrin homology (GFP-PH) domain fusion protein as a probe for PI 3-kinase lipid products in the plasma membrane of intact cells, we further demonstrate that the expression of ADAM 12 in C2C12 cells leads to activation of PI 3-kinase. Therefore, interaction of the SH3 domain of p85␣ with the cytoplasmic domain of ADAM 12 could be a novel mechanism of activation of PI 3-kinase in myoblasts.

EXPERIMENTAL PROCEDURES
Antibodies-Anti-ADAM 12 antibody has been described previously (16). Mouse monoclonal and rabbit polyclonal anti-p85␣ antibodies were purchased from Transduction Laboratories (Lexington, KY) and Upstate Biotechnology (Lake Placid, NY), respectively.
Expression Constructs-Bacterial expression constructs encoding GST-P1-5, -P1-4, -P3-5, -P12, -P5, calmodulin-binding peptide (CBP)tagged P1-5, and the cytoplasmic tail of integrin ␤ 1A were described previously (16). DNA fragments encoding the P34 region of ADAM 12 (aa 846 -870) and the SH3 domain of mouse p85␣ (aa 1-86) were amplified by polymerase chain reaction (PCR) using mouse skeletal muscle cDNA (CLONTECH, Palo alto, CA) as template, Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA), and appropriate sets of primers. PCR products were cloned into the pGEX-2T vector (Amersham Pharmacia Biotech) between the BamHI and EcoRI sites for expression of glutathione S-transferase (GST) fusion proteins in Eschrichia coli. The construction of the full-length ADAM 12 plasmid and ADAM 12-(⌬1-424), which encodes a truncated form of mouse ADAM 12 (aa 425-903) lacking the pro-and metalloprotease domains and containing an exogenous Ig secretion signal, has been described previously (16). To engineer a membrane-targeted ADAM 12 cytoplasmic tail, a myristoylation motif corresponding to the first 7 amino acids of mouse c-Src (ATGGGCAGCAACAAGAGCAAG) was added in-frame to the 5Ј-end of the DNA fragment encoding the ADAM 12 cytoplasmic tail (aa 732-903). The amplification product was cloned into the pIRESpuro vector (CLONTECH) between the NotI and ClaI sites. The PH domain of mouse ARNO, comprising amino acids 262-400, was amplified using primers 5Ј-CTCGCTGGATCCCGAGAGGGCTGGCTCCTAAA-3Ј and 5Ј-CTCGCTGAATTCTCAGGGTTGTTCTTGCTTCT-3Ј and mouse skeletal muscle cDNA as template. The PCR product was cloned into the pEGFP-C1 vector (CLONTECH) between the BglII and EcoRI sites.
Cell Culture and Transfections-C2C12 and COS-7 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 2 mM glutamine at 37°C in the presence of 5% CO 2 under humidified atmosphere. Transfection of C2C12 cells (5 ϫ 10 5 cells/100-mm plate) or COS-7 cells (2 ϫ 10 6 cells/100-mm plate) was performed using LipofectAMINE Plus (Life Technologies, Inc.) according to the manufacturer's instructions. Expression of the recombinant proteins was analyzed 38 h after transfection.
Mutagenesis-Site-directed mutagenesis was performed to introduce alanine substitutions for Pro 825 , Pro 828 , Pro 833 , Pro 836 , Pro 884 , or Pro 887 of ADAM 12. Mutants were generated by annealing mutagenic primers to a double-stranded plasmid containing ADAM 12-(⌬1-424) insert. Pfx Platinum DNA polymerase (Life Technologies, Inc.) was used during PCR to extend the appropriate mutagenic primers. PCR products were digested twice with DpnI (Promega, Madison, WI) and then transformed into E. coli XL1 Epicurian-Blue Supercompetent cells (Stratagene). Plasmids were purified using the EndoFree Plasmid Maxi Kit (Qiagen, Valencia, CA). The identity of all of the mutants was verified by DNA sequencing (Iowa State University, Ames, Iowa).
Protein Expression and Purification-All of the GST fusion proteins and CBP-tagged proteins were expressed as soluble forms and purified on glutathione-Sepharose columns (Amersham Pharmacia Biotech) or calmodulin affinity resin (Stratagene), respectively, according to the manufacturers' instructions.
Immunoprecipitation-Transfected COS-7 cells were solubilized with buffer A (2 ml of buffer/100-mm plate). Cell extracts were subjected to centrifugation (15,000 ϫ g, 20 min), and the supernatant (1 ml) was mixed with protein G-Sepharose (20 l; Amersham Pharmacia Biotech) and incubated for 1 h at 4°C (pre-clearing). After removal of protein G-Sepharose, the cell lysate was incubated with anti-ADAM 12 antibody (1.5 g/ml lysate) or anti-p85␣ monoclonal antibody (2.5 g/ml lysate) for 4 h at 4°C and then with protein G-Sepharose (20 l) for 30 min at 4°C. The immunoprecipitates were washed four times with buffer B and eluted with SDS-gel loading buffer. Eluates were analyzed in 8% SDS-PAGE followed by immunoblotting first with anti-p85␣ polyclonal antibody or anti-ADAM 12 antibody and then with HRPcoupled secondary antibody.
Immunoblotting-Protein samples were separated by 8% SDS-PAGE and transferred onto nitrocellulose membranes. The membranes were first washed in blocking buffer (DPBS, 3% (w/v) dry milk, and 0.3% (v/v) Tween 20) for 1 h and then incubated with blocking buffer supplemented with a primary antibody followed by a HRP-conjugated secondary antibody. The antigen-antibody complexes were visualized by chemiluminescent detection (SuperSignal West Pico, Pierce). The following concentrations of primary antibodies were used: polyclonal anti-ADAM 12 antibody, 0.3 g/ml; monoclonal anti-p85␣ antibody, 0.25 g/ml; polyclonal anti-p85␣ antibody, 1 g/ml.
Immunostaining-C2C12 cells were seeded on glass coverslips and transfected with expression vectors. Two days after transfection, cells were fixed with 3.7% paraformaldehyde in DPBS. After permeabilizing cells with 0.1% Triton X-100 in DPBS for 5 min, cells were incubated with anti-ADAM 12 antibody (1:500 dilution) for 1 h and then with rhodamine-conjugated anti-rabbit IgG antibody (1:500 dilution) for 30 min. The coverslips were rinsed with DPBS, mounted onto slides with 20 l of 10% (w/v) Mowiol 4-88 (Calbiochem) in 25% glycerol, and viewed on a Zeiss LSM410 laser scanning confocal microscope. In the experiment in which PI 3-kinase inhibitors were used, 6 h before fixation, cells were treated with growth medium containing 1 M wortmannin or 50 M LY294002 (Calbiochem) dissolved in dimethyl sulfoxide (Me 2 SO). Control cells were treated with Me 2 SO only.

RESULTS
The minimal motifs required for binding to the Src homology 3 (SH3) domains have been determined as (R/K)XXPXXP (class I ligands, binding to SH3 domains in the N 3 C orientation) or PXXPX(R/K) (class II ligands, binding in the C 3 N orientation) (31,32). The cytoplasmic domain of ADAM 12 contains three class I and two class II SH3 binding motifs that are conserved between mouse and human (Fig. 1, A and B).
To examine whether the SH3 binding motifs in ADAM 12 can mediate interaction with the p85␣ regulatory subunit of PI 3-kinase, we expressed and affinity-purified a fragment of mouse ADAM 12 that spans all five SH3 binding motifs, P1-5 (residues 794 -903, Fig. 1B), and is fused to CBP. The SH3 domain of mouse p85␣ was expressed as a GST fusion protein and purified on glutathione affinity resin. As shown in Fig. 2, the GST-SH3 fusion protein bound directly to CBP-P1-5 but not to a control protein composed of CBP and the cytoplasmic domain of mouse integrin ␤ 1A .
Next, we investigated whether the full-length p85␣ protein and full-length ADAM 12 could participate in the binding. The GST-P1-5 fusion protein was immobilized on a glutathione column, and the lysate from C2C12 mouse myoblasts containing the endogenous p85␣ was passed through the column. As shown in Fig. 3A, p85␣ was retained on the GST-P1-5 but not on the GST column, demonstrating that the full-size p85␣ was capable of interaction with the cytoplasmic domain of ADAM 12. Similarly, the full-length ADAM 12 expressed in COS-7 cells was retained on the GST-SH3 but not on the GST column, consistent with the notion that the full-length transmembrane form of ADAM 12 bound to the SH3 domain of p85␣ (Fig. 3B). Moreover, a truncated form ADAM 12, ADAM 12-(⌬1-424), lacking the N-terminal pro-and metalloprotease domains, containing an exogenous secretion signal, and corresponding to a biologically active form of ADAM 12 described previously (4,16), bound equally well to GST-SH3 protein (Fig. 3B), suggesting that the pro-and metalloprotease domains are not required for binding. Multiple forms of recombinant ADAM 12 (ranging from ϳ115 to ϳ120 kDa; predicted molecular mass, 95 kDa) and ADAM 12-(⌬1-424) (ϳ60 -70 kDa; predicted molecular mass, 52 kDa) were the result of the variable extent of protein glycosylation (14,16).
To gain insight into the localization of p85␣-binding sites in ADAM 12, we expressed GST fusion proteins containing the following fragments of the ADAM 12 cytoplasmic domain: P1-4  1 and 3) was immobilized on calmodulin affinity columns. Purified GST-SH3 (lanes 1 and 2) or GST protein (lanes 3 and 4) was loaded on the columns, the columns were washed and eluted with gel loading buffer, and the eluates were subjected to SDS-PAGE and Coomassie Blue staining.  Fig. 1B). The recombinant GST fusion proteins were purified, electrophoresed, transferred to a nitrocellulose membrane, and subjected to an overlay binding assay using purified, biotinylated GST-SH3 protein. As shown in Fig. 4, GST-SH3 bound to all ADAM 12 fragments except P34, suggesting the presence of at least two different binding sites located in the P12 and P5 regions, respectively.
To further determine which of the SH3 binding motifs was responsible for the interaction with p85␣, we used site-directed mutagenesis to disable individual SH3-binding sites in ADAM 12. In each mutant, PXXP motifs, which constitute the core of SH3-binding sites, were replaced with sequences AXXA, leading to a full disruption of any potential interactions involving the mutated sites. Mutants M1, M2, and M5 had a single SH3-binding site disabled (1, 2, or 5, respectively). Double mutants M1M2, M1M5, and M2M5, had only one binding site that remained functional (site 5, 2, or 1, respectively). The triple mutant M1M2M5 had all three SH3-binding sites disabled. The sequences of the regions in ADAM 12 that were subjected to mutagenesis are shown in Fig. 5A. COS-7 cells were transfected with a vector encoding ADAM 12-(⌬1-424) or the same construct containing M1, M2, M5, M1M2, M1M5, M2M5, or M1M2M5 mutations, as described above. Expression of the recombinant proteins was analyzed by Western blotting using anti-ADAM 12 antibody. As shown in Fig. 5B, the mutations did not affect the stability of the recombinant proteins or the levels of protein expression. As shown further in Fig. 5C, the elimination of a single SH3-binding site in the M1, M2, or M5 mutants did not inhibit binding of ADAM 12-(⌬1-424) to the SH3 domain of p85␣. Moreover, simultaneous mutations in any two of the three SH3-binding sites did not abolish the binding. To efficiently inhibit the interaction between ADAM 12-(⌬1-424) and the SH3 domain of p85␣, it was necessary to eliminate all three SH3-binding sites.
To examine whether p85␣ and the cytoplasmic domain of ADAM 12 interact in vivo, C2C12 cells were transfected with a vector encoding ADAM 12-(⌬1-424), ADAM 12-(⌬1-424) triple mutant M1M2M5, or a vector without insert, followed by immunoprecipitation of ADAM 12-(⌬1-424) or p85␣ and Western blot analysis of the co-imunoprecipitating proteins. The coimmunoprecipitation experiment employed the truncated rather than the full-length version of ADAM 12, because the truncated form, lacking the pro-and metalloprotease domains, is transported to the cell surface much more efficiently than the full-length protein (16,33). As shown in Fig. 6, p85␣ was detected in the anti-ADAM 12 immunoprecipitate obtained from ADAM 12 (⌬1-424)-transfected and not from cells transfected with ADAM 12-(⌬1-424) mutant M1M2M5 or from control cells. Reciprocally, ADAM 12-(⌬1-424) was co-immunoprecipitated with anti-p85␣ antibody, suggesting that the two proteins formed a complex in intact cells.
Finally, we addressed the question of the effect of the interaction with ADAM 12 on PI 3-kinase activity. We reasoned that the interaction of p85␣ with the cytoplasmic domain of ADAM 12 could recruit PI 3-kinase to the plasma membrane, where the enzyme could get direct access to its lipid substrates. Because the occupancy of the SH3 domain of p85␣ with prolinerich ligands has not been reported to increase the specific activity of PI 3-kinase (27-29), we decided to measure the activation status of PI 3-kinase by monitoring the amount of PI 3-kinase lipid products in intact cells using a GFP-PH domain  1-6), or GST alone (lane 7). B, direct binding of GST-SH3 to ADAM 12 fragments. The GST fusion proteins containing ADAM 12 fragments (lanes 1-6) or GST alone (lane 7) were electrophoresed, transferred to a nitrocellulose membrane, and incubated with biotinylated GST-SH3 protein, followed by incubation with HRP-conjugated streptavidin and visualization by a chemiluminescence detection method.
fusion protein as a probe. The PH domain in the fusion protein was derived from ARNO (Arf nucleotide binding site opener) and was previously shown to bind with a high affinity to PI(3,4,5)P 3 , one of the major products of PI 3-kinase (30), and to translocate from the cytoplasm to the plasma membrane in insulin-stimulated adipocytes (34). Co-transfection of C2C12 cells with ADAM 12-(⌬1-424) and GFP-PH resulted in the accumulation of GFP-PH at the plasma membrane in the regions of strong ADAM 12-(⌬1-424) staining (Fig. 7, A-D). Similarly, GFP-PH localized to the plasma membrane in cells that have been co-transfected with myristoylated, membrane-anchored cytoplasmic domain of ADAM 12 (Fig. 7, E and F). On the contrary, GFP-PH was poorly recruited to the plasma membrane in cells that were transfected with GFP-PH only (Fig. 7, G and H) or in cells co-transfected with GFP-PH and ADAM 12-(⌬1-424) triple mutant M1M2M5, which was unable to bind to p85␣ (Fig. 7, I and J). Finally, incubation of ADAM 12-(⌬1-424)-and GFP-PH-cotransfected cells with LY294002 (Fig. 7, K and L) or wortmannin (not shown), two specific inhibitors of PI 3-kinase, greatly diminished the amount of GFP-PH at the membrane. This indicated that the translocation of GFP-PH to the membrane was PI 3-kinase-dependent and was not a result of direct interaction between GFP-PH and ADAM 12.

DISCUSSION
In this work, we have demonstrated that the cytoplasmic domain of ADAM 12 interacts with the SH3 domain of p85␣, a regulatory subunit of PI 3-kinase, both in vitro and in vivo. We have identified three p85␣-binding sites in ADAM 12 involving PXXP motifs located at amino acids 825-828, 833-836, and 884 -887. Site-directed mutagenesis established that any one of these sites is sufficient to mediate interaction with p85␣ in vitro. Moreover, there was very little synergy between the three sites, and the interaction with p85␣ of ADAM 12 containing three, two, or just one binding site intact was essentially the same. No single site seemed to be critical for the binding, as disruption of any of the three sites did not affect the interaction with p85␣.
ADAM 12 is the first member of the ADAM family reported to interact with PI 3-kinase. Importantly, several other members of the family contain proline-rich sequences in their cytoplasmic domains and, potentially, they may interact with SH3containg proteins, including p85␣. It has to be stressed, however, that the presence of SH3 binding motifs does not necessarily warrant productive SH3-mediated protein-protein interactions. Specifically, although ADAM 12 contains five legitimate SH3 binding motifs, only three of them (motifs 1, 2, and 5) mediated interactions with p85␣, and sites 3 and 4 were nonfunctional. Similarly, we have recently demonstrated that the interaction of ADAM 12 with the SH3 domain of protein tyrosine kinase Src required binding sites 1 or 2, whereas sites 3-5 were not active (16).
Activation of PI 3-kinase requires its translocation to the plasma membrane, where the enzyme is positioned in the proximity of its lipid substrates. The most common mechanism of the translocation to the membrane involves the interaction of SH2 domains in the regulatory subunit of PI 3-kinase with activated, tyrosine-phosphorylated growth factor receptors or adaptor molecules (27)(28)(29). In addition to mediating recruitment to the membrane, interaction of SH2 domains with phosphopeptides further increases the specific activity of the catalytic subunit of PI 3-kinase, leading to full activation of the enzyme (35,36). It has to be stressed, however, that translocation to the plasma membrane alone is sufficient to activate PI 3-kinase, as demonstrated by targeting of the p110 catalytic subunit to the membrane by either N-terminal myristoylation or C-terminal farnesylation (37). These membrane-bound forms of p110 produced constitutively active PI 3-kinases and induced PI 3-kinase-dependent responses in the absence of growth factor stimulation. Transmembrane ADAM 12, by providing docking sites for the SH3 domain of p85␣, could therefore play an important role in the activation of PI 3-kinase by directly recruiting it to the membrane. At the present moment, it is not clear whether other mechanisms contribute further to the activation of PI 3-kinase at the membrane. Nevertheless, because PI 3-kinase is critical for terminal differentiation of myoblasts (17)(18)(19), and because expression of ADAM 12 is dramatically up-regulated at the onset of myoblast differentiation (4, 6, 8), ADAM 12-mediated recruitment to the mem- brane may constitute one of the regulatory mechanisms for PI 3-kinase during the differentiation process.