Identification of AF-6 and canoe as putative targets for Ras.

Ras (Ha-Ras, Ki-Ras, N-Ras) is implicated in the regulation of various cell functions such as gene expression and cell proliferation downstream from specific extracellular signals. Here, we partially purified a Ras-interacting protein with molecular mass of about 180 kDa (p180) from bovine brain membrane extract by glutathione S-transferase (GST)-Ha-Ras affinity column chromatography. This protein bound to the GTPS (guanosine 5′-(3-O-thio)triphosphate, a nonhydrolyzable GTP analog)•GST-Ha-Ras affinity column but not to those containing GDP•GST-Ha-Ras or GTPS•GST-Ha-Ras with a mutation in the effector domain (Ha-Ras). The amino acid sequences of the peptides derived from p180 were almost identical to those of human AF-6 that is identified as the fusion partner of the ALL-1 protein. The ALL-1/AF-6 chimeric protein is the critical product of the t (6:11) abnormality associated with some human leukemia. AF-6 has a GLGF/Dlg homology repeat (DHR) motif and shows a high degree of sequence similarity with Drosophila Canoe, which is assumed to function downstream from Notch in a common developmental pathway. The recombinant N-terminal domain of AF-6 and Canoe specifically interacted with GTPS•GST-Ha-Ras. The known Ras target c-Raf-1 inhibited the interaction of AF-6 with GTPS•GST-Ha-Ras. These results indicate that AF-6 and Canoe are putative targets for Ras.

Ras (Ha-Ras, Ki-Ras, N-Ras) is a signal-transducing guanine nucleotide-binding protein for tyrosine kinase-type receptors such as epidermal growth factor receptors and the Src family, leading to a mitogenic response and differentiation (for reviews, see Refs. 1 and 2). Ras has GDP-bound inactive and GTP-bound active forms, the latter of which makes physical contact with targets. Intensive investigations revealed that the Raf kinase family, consisting of c-Raf-1 (for reviews, see Refs. 3 and 4), A-Raf (5), and B-Raf (6 -9), is one of the direct targets for Ras. The activated Raf phosphorylates MAP 1 kinase kinase and activates it. Consequently the activated MAP kinase kinase activates MAP kinase, leading to the expression of certain genes such as c-fos (for reviews, see Refs. 10 and 11). Several molecules interacting with activated Ras in addition to Raf have been identified in mammals. These include phosphatidylinositol-3-OH kinase (12), Ral GDS (13,14), and Rin1 (15). On the basis of these observations, a variety of Ras targets may account for the pleiotropic functions of Ras. To understand the molecular mechanism of pleiotropic functions of Ras, it is essential to identify novel targets for Ras.
In the present study, we discovered and partially purified another putative target for Ras with a molecular mass of about 180 kDa (p180) by use of GST-Ha-Ras affinity column chromatography and identified it as AF-6 (16), whose structure resembles that of Drosophila Canoe, which is involved in the Notch signaling pathway (17).

EXPERIMENTAL PROCEDURES
Materials and Chemicals-All materials used in the nucleic acid study were purchased from Takara Shuzo Co. Ltd. (Kyoto, Japan). Expression plasmids, pGEX, pMal-c2, and pRSET were obtained from Pharmacia Biotech (Tokyo, Japan), New England Biolabs Inc. (Beverly, MA), and Invitrogen Corp. (San Diego, CA), respectively. Other materials and chemicals were obtained from commercial sources.
[ 35 S]GTP␥S and [ 35 S]methionine were purchased from DuPont NEN. A rabbit polyclonal antibody against a 16-mer peptide corresponding to 561-576 aa of human AF-6 (RVEQQPDYRRQESRTQ) was generated and purified.
Preparation of Bovine Brain Membrane Extract-The homogenate of bovine brain gray matter, 190 g, was prepared and centrifuged at 20,000 ϫ g for 30 min at 4°C as described (19). The precipitate was suspended into 360 ml of homogenizing buffer (25 mM Tris/HCl at pH 7.5, 5 mM EGTA, 1 mM dithiothreitol, 10 mM MgCl 2 , 10 M (p-amidinophenyl)-methanesulfonyl fluoride, 1 mg/liter leupeptin, 10% sucrose) to prepare the crude membrane fraction (19). The proteins in this fraction were extracted by addition of an equal volume of homogenizing buffer * This study was supported by grants-in-aid for Scientific Research and for Cancer Research from the Ministry of Education, Science, and Culture, Japan (1995) and by a grant from the Yamanouchi Foundation for Research on Metabolic Disease (1995). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
GST-Ha-Ras Affinity Column Chromatography-The membrane extract (16 ml) was eluted through 2.5 ml of a glutathione-Sepharose 4B column (Pharmacia Biotech Inc.) to remove endogenous GST. One-tenth of the pass-through fraction was loaded on 0.25 ml of glutathione-Sepharose 4B columns containing respective GST-small G proteins loaded with guanine nucleotides as described (19). After washing the columns with 0.825 ml of buffer A three times, the bound proteins were coeluted with respective GST-small G proteins by addition of 0.825 ml of buffer A containing 10 mM glutathione and 0.2 M NaCl three times. To prepare affinity-purified p180 for peptide sequencing, the pass-through fraction (16 ml) was loaded onto a 1-ml glutathione-Sepharose column containing 24 nmol of GTP␥S⅐GST-Ha-Ras. The proteins were eluted by addition of 10 ml of buffer A containing 10 mM glutathione and 0.2 M NaCl, and fractions of 1 ml each were collected. The p180 protein appeared in fractions 2-10. The same procedures were repeated 12 times.
Peptide Sequence Analysis of p180 -The affinity-purified p180 was dialyzed three times against distilled water and concentrated by freezedrying. The concentrated samples were separated by SDS-PAGE and transferred onto polyvinylidene difluoride membranes (20). The immobilized p180 was digested, fractionated, and subjected to amino acid sequencing as described (20).
Interaction of in Vitro Translated AF-6, Canoe, and Bacterial AF-6 with GST-Small G Proteins-In vitro translation of pRSET-AF-6 (36 -848 aa), pRSET-AF-6 (36 -206 aa), and pRSET-Canoe (1-217 aa) were performed as described (19). In vitro translated products labeled with [ 35 S]methionine were mixed with glutathione-Sepharose 4B beads containing the respective GST-small G proteins loaded with guanine nucleotides (19). The bound proteins were then coeluted with GST-small G proteins by addition of glutathione. The eluates were resolved by SDS-PAGE and vacuum-dried followed by autoradiography. The shorter N-terminal domain of AF-6 (36 -206 aa) was expressed as an MBP fusion protein (MBP-AF-6) and purified using amylose resin (New England Biolabs). MBP-AF-6 (0.15 nmol) was subjected to the GSTsmall G protein affinity column chromatography as described above. For competition assay with c-Raf-1, 1.5 nmol of MBP-c-Raf-1 was simultaneously added to the incubation mixture.
Other Procedures-SDS-PAGE proceeded as described previously (21). The BLAST program was used for protein homology search (22). Immunoblot analysis of p180 was carried out as described (23).

RESULTS AND DISCUSSION
To detect molecules interacting with Ha-Ras, the bovine brain membrane extract was loaded onto GST-Ha-Ras affinity columns. The proteins bound to the affinity columns were coeluted with GST-Ha-Ras by addition of glutathione. Proteins with a molecular mass of about 180 kDa (p180) and 195 kDa (p195) were detected in the glutathione eluate from the GTP␥S⅐GST-Ha-Ras affinity column but not from those containing GST or GDP⅐GST-Ha-Ras (Fig. 1). Neither p180 nor p195 was detected in the eluate of the affinity column for GTP␥S⅐GST-Ha-Ras A38 , which has a mutation in the effectorinteracting domain (1,2). We further confirmed the specificity of the interaction by affinity column chromatography using GST-R-Ras, GST-RalA, and GST-RhoA. Less p180 and p195 were eluted from the GTP␥S⅐GST-R-Ras affinity column but not from the GDP⅐GST-R-Ras affinity column (data not shown). Neither p180 nor p195 was eluted from the GST-RalA or GST-RhoA affinity column (data not shown).
To identify the GTP␥S⅐Ha-Ras-interacting molecule, p180 was subjected to amino acid sequencing as described under "Experimental Procedures." Six peptide sequences derived from p180 were determined. These were: 1) STATTQDVLE; 2) DMPETSFTR; 3) LPYLVELSPDG; 4) PGIVQETTFDLG; 5) YAPDDIPNINS; and 6) LLLEWQFQK. All six peptide sequences were almost identical to the deduced amino acid sequence of human AF-6, which is the fusion partner of the ALL-1 protein (16). The ALL-1/AF-6 chimeric protein is the critical product of the t (6:11) abnormality associated with some human leukemia. Furthermore, p180 was recognized by the antibody raised against human AF-6 (Fig. 2). The calculated molecular mass of human AF-6 is 181,777 Da, which is close to the apparent molecular mass of p180 estimated by SDS-PAGE. We therefore concluded that p180 is the bovine counterpart of human AF-6 and hereafter referred to it as AF-6. Since this antibody cross-reacted with p195 weakly (data not shown), p195 may be an isoform or an alternatively spliced form of AF-6.
To address whether or not recombinant AF-6 interacts with GTP␥S⅐Ha-Ras, GST-small G proteins immobilized on beads were mixed with the in vitro translated N-terminal domain of AF-6 (36 -848 aa), and interacting proteins were coeluted with GST-small G proteins by the addition of glutathione. The in vitro translated AF-6 (36 -848 aa) was coeluted with GTP␥S⅐GST-Ha-Ras but weakly with GDP⅐GST-Ha-Ras, GTP␥S⅐GST-Ha-Ras A38 , GST-R-Ras, GST-RalA, and GST-RhoA (Fig. 3a). The band with GTP␥S⅐GST-Ha-Ras A38 was very faint (lane 5), and the bands with GDP⅐GST-RalA and GTP␥S⅐GST-RalA (lanes 8 and 9) were a little bit stronger than those with GDP⅐GST-R-Ras and GTP␥S⅐GST-R-Ras (lanes 6  and 7). The weak bands detected in the eluates other than that from GTP␥S⅐GST-Ha-Ras may result from the weak interaction of AF-6 with the respective small G proteins. Although some AF-6 in the membrane extract was slightly retained on the GTP␥S⅐GST-R-Ras affinity column (data not shown), the in vitro translated AF-6 was not. This may be due to the lower affinity of AF-6 for GTP␥S⅐GST-R-Ras than that for GTP␥S⅐GST-Ha-Ras. To determine the Ras-interacting domain of AF-6 more accurately, a similar experiment was performed using the shorter N-terminal domain of AF-6 (36 -206 aa). A similar retention was observed when the shorter N-terminal domain of AF-6 (36 -206 aa) was employed (Fig. 3b).
A homology search of the GenBank protein data base revealed a high degree of sequence similarity of AF-6 with Drosophila Canoe (Fig. 4), which is assumed to function down- stream from Notch in a common developmental pathway (17). Since Canoe was presumed to interact with Ras in the same manner as AF-6, the interaction of Canoe was investigated. The in vitro translated N-terminal domain of Canoe (1-217 aa) was also coeluted with GTP␥S⅐GST-Ha-Ras and scarcely with GST, GDP⅐GST-Ha-Ras, and GTP␥S⅐GST-Ha-Ras A38 (Fig. 3b).
We examined whether or not AF-6 directly interacts with GTP␥S⅐Ha-Ras. The shorter N-terminal domain of AF-6 (36 -206 aa) was expressed as an MBP fusion protein (MBP-AF-6) and mixed with immobilized GST-Ha-Ras. Interacting proteins were coeluted with GST-Ha-Ras by the addition of glutathione. MBP-AF-6 was coeluted with GTP␥S⅐GST-Ha-Ras but not with GST, GDP⅐GST-Ha-Ras, or GTP␥S⅐GST-Ha-Ras A38 (Fig. 3c).
The band corresponding to molecular mass of about 55 kDa may be a degraded product of MBP-AF-6.
The apparent K d values for MBP-AF-6 and MBP-c-Raf-1 were estimated to be about 250 and 200 nM, respectively, under the conditions (data not shown). Since c-Raf-1 interacts with activated Ras via the effector domain (1, 2), we examined whether or not c-Raf-1 competes with AF-6 for interaction with activated Ha-Ras. An excess amount of MBP-c-Raf-1 inhibited the interaction of the MBP-AF-6 with GTP␥S⅐GST-Ha-Ras (Fig. 3c).
In this study, we purified a Ras-interacting protein (p180) from a bovine brain membrane extract. We identified it as AF-6, which has a GLGF/DHR motif and shows a high degree of sequence similarity with Drosophila Canoe (16,17). The recombinant AF-6 and Canoe specifically interacted with activated Ha-Ras. Furthermore, c-Raf-1 inhibited the interaction of AF-6 with activated Ha-Ras. These results indicate that AF-6 and Canoe serve as putative targets for Ras.
We showed that activated Ras interacted with the N-terminal domains of AF-6 and Canoe. These domains show a high degree of sequence similarity to each other, indicating that this unique domain confers specificity for the GTP⅐Ras complex. The direct interaction of c-Raf-1, A-Raf, B-Raf, phosphatidylinositol-3-OH kinase, Ral GDS, and Rin1 with activated Ras has been demonstrated (3-9, 12-15). The Ras-interacting interfaces of these proteins have been determined. There is no obvious homology among Ras-interacting interfaces of c-Raf-1, phosphatidylinositol-3-OH kinase, Ral GDS, Rin1, and AF-6/ Canoe, indicating that activated Ras can recognize a variety of target interfaces. This diversity of Ras-interacting interfaces may allow a range of downstream pathways from Ras to induce appropriate cellular responses to extracellular signals.
AF-6 and Canoe are homologous to each other and share a common domain organization (Fig. 4) (16, 17). The most highly conserved region among them is a GLGF/DHR motif, which is found in a number of other proteins including Drosophila discslarge tumor suppressor gene product (Dlg) (24), dishevelled gene product (25,26), an intracellular protein-tyrosine phosphatase (PTP-meg) (27), postsynaptic density protein 95 (PSD-95) (28), and a tight junction-associated protein ZO-1 (29,30). The GLGF/DHR motif is thought to function to localize them at the specialized sites of cell-cell contact by forming a complex with specific proteins such as protein 4.1 homologues (31). The structural feature of AF-6 and Canoe suggests that they locate at the junction of plasma membrane and cytoskeleton, where they may regulate signal transduction and cytoskeleton.
The N terminus of AF-6 flanked by the GLGF/DHR motif also shares a high homology with that of Canoe, to both of which activated Ras binds specifically. Canoe has been postulated to function downstream from Notch and to mediate interactions between the Notch cascade and other signaling pathways (17). Although AF-6 function remains obscure, the similar structural feature and property of AF-6 and Canoe suggest that the AF-6/Canoe family may serve as an intracellular signaling component controlled by two distinct signaling pathways such as Ras and Notch. Our preliminary experiments suggest that Canoe is genetically linked to Ras1 in Drosophila eye development. Further studies are required to understand the roles of AF-6/Canoe family in signal transduction.