If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
* This work was supported by National Institutes of Health Grant AR45787. This is contribution 01-316-J from the Kansas Agricultural Experimental Station.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
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.
protein containing adisintegrin andmetalloprotease
Arf nucleotide binding site opener
Src homology 3 domain
Src homology 2 domain
green fluorescent protein
pleckstrin homology domain
Dulbecco's phosphate-buffered saline
polyacrylamide gel electrophoresis
polymerase chain reaction
ADAMs1 (proteins containing adisintegrin andmetalloprotease) are a family of transmembrane or secreted glycoproteins that have been implicated in cell surface proteolysis, adhesion, and cell-cell communication (
). 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 (
Y. Cao and A. Zolkiewska, unpublished observation.
2Y. Cao and A. Zolkiewska, unpublished observation.
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 non-receptor protein kinase Src (
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 (
). 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 myogenesisin vitro (
). Class I enzymes are heterodimers composed of an ∼110-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, p85α and p85β, contain a single SH3 domain. Importantly, a p85α-associated PI 3-kinase appears to be indispensable for myogenesis (
In the present study, we show that the cytoplasmic tail of ADAM 12 interacts with the SH3 domain of the p85α regulatory subunit of PI 3-kinase in vitro and in vivo. By site-directed mutagenesis, we mapped the p85α-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 p85α with the cytoplasmic domain of ADAM 12 could be a novel mechanism of activation of PI 3-kinase in myoblasts.
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 → C orientation) or PXXPX(R/K) (class II ligands, binding in the C → N orientation) (
). 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 p85α 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 p85α 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 β1A.
Next, we investigated whether the full-length p85α 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 p85α was passed through the column. As shown in Fig.3A, p85α was retained on the GST-P1–5 but not on the GST column, demonstrating that the full-size p85α 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 p85α (Fig. 3B). Moreover, a truncated form ADAM 12, ADAM 12-(Δ1–424), 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 (
), 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 ∼115 to ∼120 kDa; predicted molecular mass, 95 kDa) and ADAM 12-(Δ1–424) (∼60–70 kDa; predicted molecular mass, 52 kDa) were the result of the variable extent of protein glycosylation (
To gain insight into the localization of p85α-binding sites in ADAM 12, we expressed GST fusion proteins containing the following fragments of the ADAM 12 cytoplasmic domain: P1–4 (aa 794–870, spanning the first four SH3 binding motifs), P3–5 (aa 846–903, spanning SH3 binding motifs 3–5), P12 (aa 794–845, containing sites 1 and 2), P34 (aa 846–870, containing sites 3 and 4), and P5 (aa 871–903, containing site 5 only) (see 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 p85α, 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-(Δ1–424) 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-(Δ1–424) to the SH3 domain of p85α. 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-(Δ1–424) and the SH3 domain of p85α, it was necessary to eliminate all three SH3-binding sites.
To examine whether p85α and the cytoplasmic domain of ADAM 12 interact in vivo, C2C12 cells were transfected with a vector encoding ADAM 12-(Δ1–424), ADAM 12-(Δ1–424) triple mutant M1M2M5, or a vector without insert, followed by immunoprecipitation of ADAM 12-(Δ1–424) or p85α and Western blot analysis of the co-imunoprecipitating proteins. The co-immunoprecipitation 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 (
). As shown in Fig.6, p85α was detected in the anti-ADAM 12 immunoprecipitate obtained from ADAM 12 (Δ1–424)-transfected and not from cells transfected with ADAM 12-(Δ1–424) mutant M1M2M5 or from control cells. Reciprocally, ADAM 12-(Δ1–424) was co-immunoprecipitated with anti-p85α 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 p85α 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 p85α with proline-rich ligands has not been reported to increase the specific activity of PI 3-kinase (
), 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 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)P3, one of the major products of PI 3-kinase (
). Co-transfection of C2C12 cells with ADAM 12-(Δ1–424) and GFP-PH resulted in the accumulation of GFP-PH at the plasma membrane in the regions of strong ADAM 12-(Δ1–424) 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-(Δ1–424) triple mutant M1M2M5, which was unable to bind to p85α (Fig. 7, I andJ). Finally, incubation of ADAM 12-(Δ1–424)- and GFP-PH-cotransfected cells with LY294002 (Fig. 7, K andL) 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.
In this work, we have demonstrated that the cytoplasmic domain of ADAM 12 interacts with the SH3 domain of p85α, a regulatory subunit of PI 3-kinase, both in vitro and in vivo. We have identified three p85α-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 p85α in vitro. Moreover, there was very little synergy between the three sites, and the interaction with p85α 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 p85α.
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 SH3-containg proteins, including p85α. 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 p85α, 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 (
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 (
). 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 (
). 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 (
). 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 p85α, 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 (