Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Characterization of RasGRP2, a Plasma Membrane-targeted, Dual Specificity Ras/Rap Exchange Factor

Characterization of RasGRP2, a Plasma Membrane-targeted, Dual Specificity Ras/Rap Exchange Factor THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 275, No. 41, Issue of October 13, pp. 32260 –32267, 2000 © 2000 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Characterization of RasGRP2, a Plasma Membrane-targeted, Dual Specificity Ras/Rap Exchange Factor* Received for publication, July 11, 2000, and in revised form, July 27, 2000 Published, JBC Papers in Press, July 28, 2000, DOI 10.1074/jbc.M006087200 Jodi Clyde-Smith‡, Gint Silins§, Michael Gartside§, Sean Grimmond§, Maria Etheridge‡, Ann Apolloni‡, Nick Hayward§, and John F. Hancock‡¶ From the ‡Queensland Cancer Fund Laboratory of Experimental Oncology, Department of Pathology, University of Queensland Medical School, Herston Road, Brisbane 4006, Queensland and the §Queensland Cancer Fund Research Unit, Queensland Institute of Medical Research, Bancroft Centre, Herston Road, Brisbane 4006, Queensland, Australia Ras proteins operate as molecular switches in signal GTPase-activating proteins (GAPs) which activate and deacti- vate Ras, respectively. A number of RasGEFs have been cloned, transduction pathways downstream of tyrosine kinases and G-protein-coupled receptors. Ras is switched from which share a core catalytic exchange domain and probably the inactive GDP-bound state to the active GTP-bound release GDP from Ras via a common mechanism (1, 2). These state by guanine nucleotide exchange factors (GEFs). various RasGEFs possess different regulatory domains and We report here the cloning and characterization of Ras- have different tissue expression patterns (reviewed in Ref. 3); GRP2, a longer alternatively spliced form of the recently these observations suggest that they operate in signaling path- cloned RapGEF, CalDAG-GEFI. A unique feature of Ras- ways that activate Ras effectors for different biological GRP2 is that it is targeted to the plasma membrane by a responses. combination of N-terminal myristoylation and palmitoy- Sos1 and Sos2, are RasGEFs that are widely expressed in lation. In vivo, RasGRP2 selectively catalyzes nucleotide mammalian cells (4 – 6). Sos is normally localized to the cell exchange on N- and Ki-Ras, but not Ha-Ras. RasGRP2 cytosol, but is recruited to the plasma membrane in response to also catalyzes nucleotide exchange on Rap1, but this growth factor receptor activation. Since Ras must be localized RapGEF activity is less potent than that associated with at the plasma membrane in order to signal, Ras activation by CalDAG-GEFI. The nucleotide exchange activity of Ras- Sos can be efficiently regulated by simply controlling the sub- GRP2 toward N-Ras is stimulated by diacylglycerol and cellular localization of Sos (7, 8). Sos is recruited to the plasma inhibited by calcium. The effects of diacylglycerol and membrane in complexes with the adapter proteins Grb2 and calcium are additive but are not accompanied by any Shc (5, 6), which bind to autophosphorylated tyrosine kinase detectable change in the subcellular localization of Ras- receptors (9 –11) or to transphosphorylated docking proteins GRP2. In contrast, CalDAG-GEFI is localized predomi- (12). Similarly, G-protein-coupled receptors that activate Src nantly to the cytosol and lacks Ras exchange activity in and Shc can also recruit Grb2-Sos to the cell membrane and vivo. However, prolonged exposure to phorbol esters, or activate Ras (13, 14). Although plasma membrane recruitment growth in serum, results in localization of CalDAG-GEFI to the cell membrane and restoration of Ras exchange of Sos is sufficient to activate Ras, it is also clear that more activity. Expression of RasGRP2 or CalDAG-GEFI in complex intermolecular interactions also influence the activity NIH3T3 cells transfected with wild type N-Ras results in of this exchange factor (15–18). an accelerated growth rate but not morphologic trans- A second family of RasGEFs comprises the related proteins formation. Thus, under appropriate growth conditions, RasGRF1 and RasGRF2 (19, 20). RasGRF1 is expressed only in CalDAG-GEFI and RasGRP2 are dual specificity Ras the central nervous system, specifically in the synaptosomal and Rap exchange factors. fraction of neuronal cells (21) where it plays a role in the establishment of long term memory (22). RasGRF2 is expressed both within and without the central nervous system (20). In Ras proteins are small GTPases that operate as binary contrast to Sos, RasGRF1 is constitutively localized to the switches in key signal transduction pathways regulating cell plasma membrane, probably anchored through its N-terminal proliferation and differentiation. The amount of active GTP- pleckstrin homology domain (23). RasGRF1 exchange activity bound Ras is in turn controlled by the actions of two classes of is quiescent until activated by a rise in intracellular calcium regulator, guanine nucleotide exchange factors (GEFs) and (23, 24). RasGRF2 is also regulated by calcium, but unlike RasGRF1, RasGRF2 is localized to the cytosol in resting cells * This work was supported in part by grants from the Queensland until a rise in intracellular calcium levels causes relocalization Cancer Fund (to J. F. H. and N. H.) and the National Health and of RasGRF2 to the cell periphery and activation of its exchange Medical Research Council, Australia (to N. H.). The costs of publication activity (20). A feature common to Sos and RasGRF is the of this article were defrayed in part by the payment of page charges. presence of Dbl homology domains comprising the catalytic This article must therefore be hereby marked “advertisement”inac- cordance with 18 U.S.C. Section 1734 solely to indicate this fact. core sequences of exchange factors for the Rho family of small The nucleotide sequence(s) reported in this paper has been submitted GTPases (25). The binding of phosphatidylinositol 3,4-bisphos- TM to the GenBank /EBI Data Bank with accession number(s) AF043723 phate to a pleckstrin homology domain flanking the Dbl homol- and AF043722. ¶ Supported by the Royal Children’s Hospital Foundation, Queens- land. To whom correspondence should be addressed: Dept. of Pathology, University of Queensland Medical School, Herston Rd., Brisbane 4006, PCR, polymerase chain reaction; ORF, open reading frame; GST, glu- Queensland, Australia. Tel.: 61-7-3365-5340; Fax: 61-7-3365-5511; E- tathione S-transferase; CC, coiled-coil; DMEM, Dulbecco’s modified mail: [email protected]. Eagle’s medium; PAGE, polyacrylamide gel electrophoresis; RBD, Ras The abbreviations used are: GEF, guanine nucleotide exchange binding domain; DGB, diacylglycerol binding; BHK, baby hamster kid- ney; TPA, 12-O-tetradecanoylphorbol-13-acetate. factor; GAP, GTPase-activating protein; EST, expressed sequence tag; 32260 This paper is available on line at http://www.jbc.org This is an Open Access article under the CC BY license. A Novel Ras Exchange Factor 32261 TrisCl, pH 7.5, 150 mM NaCl, 1% aprotinin, 1 mM phenylmethylsulfonyl ogy domain of Sos1 activates exchange activity toward Rac (26), fluoride). Insoluble cellular debris was removed by centrifugation at indicating that such RasGEFs coordinate activation of Ras and 14,000 rpm for 5 min at 4 °C. 10 ml of the supernatant was reserved for Rho family proteins. later immunoblotting; the remainder was diluted in 1 ml of binding Most recently, a new family of Ras subfamily GEFs have buffer (50 mM TrisCl pH7.5, 10 mM MgCl , 0.5 mg/ml bovine serum been identified that are regulated by diacylglycerol and cal- M dithiothreitol, 100 mM NaCl) and incubated for 90 min, albumin, 0.5 m cium. RasGRP (also referred to as CalDAG-GEFII) activates rotating at 4 °C with 20 mg of GST-RafRBD-K85A protein fused to Sepharose (50% suspension). Samples were washed three times with N-, Ha-, and Ki-Ras with varying efficiencies (27, 28), whereas binding buffer before the beads were resuspended in 20 mlof23 SDS- the closely related CalDAG-GEFI (also referred to as PAGE sample buffer. Proteins from the bead capture and 10 mlof HCDC25L (Ref. 29)) activates Rap1 but not Ha-Ras (30). In cleared supernatant were resolved on 15% acrylamide gels, blotted with generating a transcript map of chromosome 11q13 (31), we anti-Ras or anti-HA (Rap) antisera and visualized by chemilumines- independently cloned CalDAG-GEFI and identified an addi- cence. Blots were quantified by phosphorimaging. The amount of GTP- tional 59-coding exon. We report here the characterization of Rap or GTP-Ras captured by the GST-RafRBD-K85A beads was ex- pressed as a percentage of the Ras or Rap present in the supernatant RasGRP2: the widely expressed, longer isoform of CalDAG- (estimated from the 10-ml input). Expression of RasGRP2-Flag or Myc- GEFI. We show that RasGRP2 is constitutively localized to the CalDAG-GEFI was confirmed by immunoblotting using anti-Flag or plasma membrane by N-terminal acylation, a feature not pre- anti-Myc antisera, respectively. viously described for any RasGEF. A consequence of constitu- Cell Culture and Metabolic Labeling Assays—COS cells were trans- tive plasma membrane localization is that RasGRP2 operates fected by electroporation and BHK cells by lipofection (36). Metabolic as an exchange factor for N-Ras and Ki-Ras in addition to being labeling with [ S]methionine was carried out for4hin methionine-free DMEM containing [ S]methionine (50 mCi/ml). Labeling with an activator of Rap1. RasGRP2, like CalDAG-GEFI, cannot 3 3 [ H]palmitic acid (0.5 mCi/ml) or [ H]myristic acid (0.5 mCi/ml) was activate Ha-Ras. CalDAG-GEFI and RasGRP2 also show quite M sodium pyruvate. carried out for4hin DMEM supplemented with 5 m distinct responses to calcium, diacylglycerol, and serum stim- Cleared lysates, normalized for protein content, were immunoprecipi- ulation revealing a series of interesting possibilities for co- tated with M2 anti-Flag Sepharose or 9E10-coated protein G-Sepharose ordinate regulation of Ras and Rap function. as required. Immunoprecipitated proteins were resolved by SDS-PAGE and visualized by fluorography (for details, see Ref. 37). Where indi- EXPERIMENTAL PROCEDURES cated, gels were agitated in 1 M neutral hydroxylamine or 1 M TrisCl, pH 7.5, for1hat room temperature prior to fluorography. Indirect immu- Cloning of RASGRP2—Cosmid cSRL-20 h12, derived from a chromo- nofluorescence for Myc and Flag epitopes was performed as described some 11 cosmid library that contained the human MCG7 (RASGRP2) (36). Prior to harvesting for fractionation studies, COS cells were se- gene (31), was sequenced and the sequence data aligned to the Gen- rum-starved for 4 –16 h as indicated. Hypotonic lysis, homogenization, Banky Expressed Sequence Tag (EST) data base using the program and centrifugation were carried out as described previously (38). BLASTN (32). The cDNA clones for matching EST entries T78563 NIH3T3 Growth Assays—NIH3T3 cells were transfected using cal- (clone 113434) and AA035643 (clone 471819) were obtained from Ge- cium phosphate precipitation as described (39). To generate polyclonal nome Systems Inc. and sequenced to completion with gene-specific and populations, cells seeded on 10-cm plates were co-transfected with 1.5 vector-specific primers. The data base entry of a third matching EST mg of pSV-hyg, together with 1.5 mg of EXV expression plasmids for clone, HIBBP12, contained an additional stretch of nucleotides up- N-Ras, RasGRP2-Flag, Myc-CalDAG-GEFI, or empty vector. After 48 h, stream of the sequences of the other ESTs, and was therefore suspected cultures were selected with 300 mg/ml hygromycin. Surviving cells were to contain an additional upstream exon of RASGRP2, but since this pooled and grown continually in hygromycin. For growth assays, 10 clone was not commercially available the new exon was sought by PCR cells (per well) were plated in multiwell dishes in DMEM containing from a genomic template, bacterial artificial chromosome clone 137C7 5%, 1%, or 0.1% serum. Cells were trypsinized and counted every 24 h (Genome Systems Inc). The composite cDNA sequence of RASGRP2 for 4 days with each time point counted in triplicate. Three independent was suspected to contain a single nucleotide error originating from transfections were analyzed. Expression of RasGRP2-Flag or Myc- within clone 471819 and the correct nucleotide sequence was therefore CalDAG-GEFI was confirmed by immunoblotting or indirect immuno- sought by reverse transcription-polymerase chain reaction of the cDNA fluorescence using anti-Flag or anti-Myc antisera, respectively. Expres- fragment from a human cDNA pool. The resulting PCR product was sion of N-Ras was verified by immunoblotting using Y13-259. ligated into pGEM-T (Promega) and the clone, designated MCG7- pGEM, sequenced. The corrected nucleotide sequence of RASGRP2 was RESULTS translated in all possible reading frames and compared with the non- redundant protein data base using the program BLASTX (32). One open Identification of a Full-length cDNA for CalDAG-GEF1/Ras- reading frame (ORF), encoding RasGRP2– 609, was assigned on the GRP2—The coding sequence for a human RasGEF was discov- basis of homology with the Caenorhabditis elegans protein F25B3.3 ered during the generation of a transcript map of a region of (GenBanky accession no. Z70752) and is identical to HCDC25L (Gen- human chromosome band 11q13 (31). The gene, originally des- Banky accession no. Y12336) (29) and CalDAG-GEFI (GenBanky ac- ignated MCG7, was subsequently assigned the locus name cession no. U78170) (30). Conceptual translation of the ORF from the alternate translation start site upstream of the one utilized in RASGRP2 by the human genome project nomenclature com- RasGRP2– 609, contained in the sequence from bacterial artificial chro- mittee. A cDNA corresponding to RASGRP2 was independently mosome clone 137C7, was designated RasGRP2– 671. These sequences isolated by two other groups (29, 30) and subsequently charac- have been deposited under GenBanky accession nos. AF043723 and terized as a 609-amino acid RapGEF referred to as CalDAG- AF043722, respectively. GEFI. However, during our independent isolation of a full- Plasmid Construction—A Myc tag was cloned onto the N terminus of length cDNA for RASGRP2, we identified an additional EST RasGRP2– 609 by recombination of MCG7-pGEM, EST clone 113434, and pGEM-11Z-Myc-tag (33) to give clone Myc-RasGRP2– 609 (referred that contained sequence 59 to that previously reported to con- to as Myc-CalDAG-GEFI hereafter). Myc-CalDAG-GEFI was digested stitute the N terminus of CalDAG-GEFI (see “Experimental with BamHI to remove the Myc tag and ligated into pGEX-2T (Amer- Procedures” for details). This additional sequence together sham Pharmacia Biotech). The final construct has an ORF that fuses with the general layout of the full-length RASGRP2 gene prod- GST to the front of RasGRP2– 609. A C-terminal Flag-tagged uct is shown in Fig. 1. The core exchange factor domains com- RasGRP2– 671 cDNA was generated by PCR. The final Flag-tagged prising the Ras exchange motif and the SCR1, SCR2, and SCR3 construct contains an ORF with the amino acids DYKDDDDKstop (Flag) added immediately after the last residue of RasGRP2 (hereafter motifs plus the diacylglycerol binding (DGB) domain and EF referred to as RasGRP2-Flag). The Myc- and Flag-tagged constructs hands have been described previously (27, 30). A novel feature were directionally cloned into the eukaryotic expression vector EXV. of the N terminus of full-length RasGRP2 is the presence of Ras Guanine Nucleotide Exchange Assays—In vitro and in vivo nu- potential consensus sequences for co-translational myristoyla- cleotide exchange assays were performed as described previously tion and posttranslational palmitoylation. Northern blot anal- (34, 35). ysis showed that RASGRP2 was expressed predominantly as Ras/Rap “Pull-down” Assays—Frozen cells were lysed in 200 mlof ice-cold lysis buffer (0.5% Triton X-100, 0.5% deoxycholate, 20 mM 2.4-kilobase pair transcripts in all tissues tested: heart, brain, 32262 A Novel Ras Exchange Factor FIG.1. Structure of RasGRP2. A, the N-terminal 62 amino acids of RasGRP2 that are not present in CalDAG-GEFI. The underlined residues represent the previously reported first 8 amino acids of CalDAG-GEFI. The sequence of the remainder of RasGRP2 is exactly as reported for CalDAG-GEFI. B, domain structure of RasGRP2. The N-terminal 62 amino acids of RasGRP2 including the acylation motifs (M1P) are replaced with a Myc tag in Myc-CalDAG-GEFI. Motifs are indicated by boxes and include the Ras exchange motif (REM), CDC25 homology domain comprising SCR1–SCR3, CC, two EF hands, and a DGB domain. placenta, lung, liver, skeletal muscle, kidney, and pancreas. Expression was highest in heart, brain, and kidney as vari- ously reported previously (29, 30), but in contrast to the pub- lished data we also detected additional transcripts of 0.8 kilo- base pairs in heart and 4.7 kilobase pairs in heart, lung, and placenta (results not shown). For the sake of brevity, we will refer to the short form of the RASGRP2 gene product as CalDAG-GEFI because this is the previously published name, and the full-length gene product as RasGRP2, after the name assigned to the gene by human genome project nomenclature committee. RasGRP2 Is Palmitoylated and Myristoylated—The N-termi- nal sequence of full-length RasGRP2 comprises MGTQRLCGR. Although the common consensus sequences for myristoylation include serine or threonine at amino acid 16, small hydropho- bic residues are also tolerated, as in the myristoylation motifs of Arf, A kinase, Arg, and Abl. Moreover, the presence of cys- teine at amino acid 17 suggested that the N terminus of RasGRP2 may also be palmitoylated (40). We therefore inves- tigated whether RasGRP2 was lipid-modified in vivo. To this FIG.2. RasGRP2 is lipid-modified and localizes to the plasma end, a Flag epitope was engineered onto the C terminus of the membrane. A, lysates of COS cells expressing RasGRP2-Flag or Myc- protein leaving the N-terminal sequences intact (5RasGRP2- CalDAG-GEFI were prepared in Triton X-114 and partitioned into Flag) (Fig. 1). A second construct, which represents the shorter detergent (D) and aqueous (Aq) phases as described (41). Equal propor- tions of each phase were resolved by SDS-PAGE, immunoblotted with alternative spliced form of RasGRP2, was also engineered by anti-Myc or anti-Flag antisera, and detected by enhanced chemilumi- replacing the N-terminal 62 amino acids of RasGRP2 with a nescence. Hydrophobic proteins partition into the detergent phase, Myc epitope (5Myc-CalDAG-GEFI). RasGRP2-Flag and Myc- whereas hydrophilic proteins partition into the aqueous phase of Triton X-114. By this analysis, RasGRP2 is significantly more hydrophobic CalDAG-GEFI were transiently expressed in COS cells and the than CalDAG-GEFI, consistent with the N terminus of RasGRP2 being relative hydrophobicity of the two proteins assessed in a Triton lipid-modified. B, RasGRP2-Flag and Myc-CalDAG-GEFI proteins ex- X-114 partitioning assay (41). Fig. 2A shows that 60% of Ras- pressed in COS cells were metabolically labeled with [ S]methionine 3 3 GRP2-Flag partitions into the detergent phase of Triton X-114, (Met), [ H]myristic acid (Mys), or [ H]palmitic acid (Pl). Proteins were whereas Myc-CalDAG-GEFI partitions exclusively into the immunoprecipitated (IP) from cell lysates with anti-epitope antisera, resolved by SDS-PAGE, and visualized by fluorography. A duplicate of aqueous phase of Triton X-114. Thus, a significant proportion the tritium-labeled gel was soaked in 1 M hydroxylamine for 1 h prior to of RasGRP2-Flag is hydrophobic, whereas all Myc-CalDAG- fluorography (1hydrox). RasGRP2-Flag incorporates label from both GEFI is hydrophilic. This result is consistent with the N ter- tritiated fatty acids, showing that it is co-translationally myristoylated minus of RasGRP2 being lipid-modified. and posttranslationally palmitoylated. Note that the palmitate but not the myristate label is released by hydroxylamine. CalDAG-GEFI, which COS cells expressing RasGRP2-Flag or Myc-CalDAG-GEFI lacks the consensus sites for these modifications, is labeled by neither were then metabolically labeled with [ H]myristic or fatty acid. C, RasGRP2-Flag and Myc-CalDAG-GEFI were expressed in [ H]palmitic acid and cell lysates analyzed by immunoprecipi- BHK cells and visualized by indirect immunofluorescence using anti- tation and fluorography. Fig. 2B shows that RasGRP2-Flag epitope antisera. Cells expressing RasGRP2-Flag show plasma mem- 3 3 brane staining, whereas cells expressing CalDAG-GEFI show cytosolic incorporated label from [ H]myristic and [ H]palmitic acid, staining. whereas Myc-CalDAG-GEFI incorporated neither label. The metabolic labeling of RasGRP2-Flag was repeated and the im- munoprecipitated proteins divided between two gels, which and therefore shows that the palmitate labeling is not a con- were soaked in 1 M hydroxylamine or TrisCl prior to fluorogra- sequence of metabolism of the palmitate label to myristate. phy. Fig. 2B shows that hydroxylamine treatment caused loss RasGRP2 Is Localized to the Plasma Membrane—The pre- of the palmitate label but not the myristate label from Ras- ceding results demonstrate that RasGRP2-Flag is N-terminally GRP2-Flag. This result indicates that palmitate, but not my- myristoylated and palmitoylated: a combination of lipid modi- ristate, is attached to RasGRP2-Flag via a thioester linkage fications that constitutes a plasma membrane targeting se- A Novel Ras Exchange Factor 32263 FIG.3. CalDAG-GEFI translocates to the membrane fraction in response to prolonged exposure to TPA. Equivalent aliquots of COS cells expressing RasGRP2-Flag or Myc-CalDAG-GEFI were treated with A23187 (10 mM) and/or TPA (100 nM), or Me SO (control) for 2 or 15 min. The cells were harvested and equal proportions of each S100 (cytosolic, S) and P100 (membrane, P) fraction were immuno- blotted with anti-Myc or anti-Flag antisera. Myc-CalDAG-GEFI is found predominantly in the S100 cytosolic fraction, whereas full-length RasGRP2-Flag is equally distributed between S100 and P100 fractions. Neither protein shows any significant change in subcellular distribu- tion in response to a 2-min exposure to calcium ionophor or TPA or both agents together; however, the amount of CalDAG-GEFI found in the P100 fraction increases substantially after a 15-min exposure to TPA. The distribution of RasGRP2-Flag was unchanged after 15 min of exposure to TPA (data not shown). quence (42). To determine the subcellular localization of Ras- GRP2, BHK cells expressing RasGRP2-Flag or Myc-CalDAG- GEFI were examined using indirect immunofluorescence. Fig. 2C shows that RasGRP2-Flag localizes extensively to the plasma membrane, although there is also some soluble cytoso- lic staining. In contrast, Myc-CalDAG-GEFI is confined to the cytosol with no evidence of plasma membrane staining. Simi- larly, when COS cells expressing the epitope-tagged variants of FIG.4. RasGRP2 functions as a RasGEF in vivo. A, wild type RasGRP2 were fractionated and immunoblotted for the Flag or Ki-Ras (B), N-Ras, or Ha-Ras proteins were transiently expressed in COS cells with and without RasGRP2-Flag or Myc-CalDAG-GEFI. The Myc epitopes, RasGRP2-Flag was found to be distributed ap- cells were metabolically labeled with [ P]orthophosphate and Ras pro- proximately equally between the P100 (membrane) and S100 teins immunoprecipitated; guanine nucleotides were eluted from the (cytosolic) fractions (Fig. 3), consistent with the proportion immunoprecipitates, resolved by TLC, and quantified by phosphorim- shown to be lipidated by Triton X-114 partitioning (Fig. 2A). aging. Duplicate aliquots of cells from each electroporation were ana- lyzed by immunoblotting to confirm that all transfected proteins were Myc-CalDAG-GEFI was found predominantly in the S100 frac- equally expressed (data not shown). RasGRP2-Flag selectively loads tion (Fig. 3). N-Ras and Ki-Ras with GTP, whereas Myc-CalDAG-GEFI does not RasGRP has been shown to translocate from cytosol to mem- nucleotide exchange any Ras protein in vivo. B, COS cells expressing brane when cells are treated with 12-O-tetradecanoylphorbol- N-Ras with and without RasGRP2-Flag were metabolically labeled as in A and treated with A23187 (10 mM) and/or TPA (100 nM) for 2 min 13-acetate (TPA) for 2 min (27, 28). We therefore examined prior to harvesting. The TLC plate from a representative experiment whether TPA would affect the subcellular localization of Ras- shows that the GEF activity of RasGRP2-Flag is up-regulated by TPA GRP2. Fig. 3 shows that there was no detectable change in the and down-regulated by an increase in intracellular calcium. C, the subcellular distribution of either RasGRP2-Flag or Myc- results from the experiment shown in panel B and three to five similar CalDAG-GEFI in response to a 2-min exposure to either TPA or experiments in which cells were treated with A23187 and/or TPA for 2 or 15 min prior to harvesting. After 15 min of exposure to TPA, Myc- a calcium ionophor (A23187), or both agents together. How- CalDAG-GEFI potently stimulates GTP loading of N-Ras. The control ever, a substantial proportion of Myc-CalDAG-GEFI did trans- cells (Con) expressing N-Ras without exchange factors were incubated locate to the P100 fraction when cells were treated with TPA for 15 min with A23187 and/or TPA. DMSO,Me SO. for 15 min (Fig. 3). The presence or absence of calcium ionophor during exposure to TPA did not significantly affect the trans- similar magnitude to that reported recently on Ha-Ras by location of Myc-CalDAG-GEFI (Fig. 3). Prolonged exposure to RasGRP (27). Importantly, in control experiments using co- TPA did not alter the subcellular distribution of RasGRP2 transfected mSos1, Ha-Ras was efficiently loaded with GTP (to (data not shown). 21 6 1.65%, n 5 3) showing that Ha-Ras is competent for RasGRP2 Selectively Activates N- and Ki-Ras but Not Ha- nucleotide exchange in this assay. In contrast, Myc-CalDAG- Ras—We next investigated the ability of RasGRP2 to catalyze GEFI was unable to stimulate GTP loading of any Ras isoform, Ras nucleotide exchange in intact cells. COS cells were trans- consistent with the findings of a previous study (30). These fected with expression plasmids for wild type N-, Ki-, and results indicate that plasma membrane localization may be Ha-Ras, alone and in combination with RasGRP2-Flag or Myc- critical for the RasGEF activity of RasGRP2 in vivo. CalDAG-GEFI. The cells were metabolically labeled with The N-Ras exchange assays were then repeated, but cells [ P]orthophosphate and the relative amounts of GTP and GDP were treated with TPA, A23187, or TPA1A23187 for 2 or 15 captured in anti-Ras immunoprecipitates determined by thin min immediately prior to harvesting. Fig. 4 (B and C) shows layer chromatography (TLC) and phosphorimaging. Fig. 4A that TPA enhanced and A23187 partially inhibited the GTP shows that in the conditions of this assay RasGRP2-Flag stim- loading of N-Ras by RasGRP2-Flag. These opposing effects ulated GTP loading of N-Ras and Ki-Ras but not Ha-Ras. The were additive on RasGRP2 exchange activity. The duration of -fold increase in GTP loading of N-Ras by RasGRP2 is of a exposure to TPA did not affect the degree of stimulation of 32264 A Novel Ras Exchange Factor FIG.5. CalDAG-GEFI functions as a RasGEF in vitro. A, GST- N-Ras was preloaded with [ H]GDP as described (34) and 20 pmol of [ H]GDP-loaded Ras incubated with (M) or without (E) 0.5 mg of re- combinant GST-CalDAG-GEFI for 0, 6, or 20 min at 30 °C in buffer containing 20 mM TrisCl, pH 7.5, 50 mM NaCl, 1 mM MgCl ,1mM dithiothreitol, 1 mM GTP. Reactions were filtered through nitrocellulose FIG.6. Comparison of the RapGEF activity of RasGRP2 and discs and the radioactivity remaining bound to Ras determined by CalDAG-GEFI. A, COS cells expressing HA-Rap1 with and without scintillation counting and expressed as a percentage of that counted in RasGRP2-Flag and Myc-CalDAG-GEFI were treated with TPA (100 nM) the t 5 0 sample (34). All assay points were carried out in duplicate, and and/or A23187 (10 mM)orMe SO for 2 min prior to harvesting. Cells the data shown are the mean of the values. No exchange activity was were lysed and HA-Rap1-GTP was isolated with GST-RafRBD-K85A- evident when GST-Rac or Rho replaced Ras in the assay (data not Sepharose beads as described under “Experimental Procedures.” A rep- shown). B, GST N-Ras preloaded with [ H]GDP was incubated with resentative immunoblot from one of these experiments is shown in recombinant CalDAG-GEFI as in A but the assay buffer was adjusted to panel A. In each lane, the HA-Rap1GTP captured is shown above an 2mM CaCl (l) or 0.01 mM TPA (l) prior to the addition of recombinant aliquot of the total HA-Rap1 present in the binding reaction (see “Ex- CalDAG-GEFI. These results were replicated in two to three independ- perimental Procedures” for details). Greater quantities of HA-Rap1GTP ent experiments. are present when Myc-CalDAG-GEFI is co-expressed than when HA- Rap1 is expressed by itself or with RasGRP2-Flag. 2-min incubations RasGRP2 activity, but with more prolonged exposure to TPA, with A23187 and/or TPA stimulate the GTP loading of HA-Rap1 by either Myc-CalDAG-GEFI or RasGRP2-Flag. RasGRP2-Flag and Myc- the inhibitory effect of A23187 was less marked. In contrast, CalDAG-GEFI expression was confirmed by immunoblotting using an- Myc-CalDAG-GEFI did not stimulate GTP loading of N-Ras ti-epitope antisera. B, the results from the assay shown in panel A and during a 2-min exposure to TPA, but a 15-min treatment with from three to six similar experiments in which cells were stimulated for TPA did result in potent activation of Myc-CalDAG-GEFI N- 2 or 15 min with TPA or A23187 were quantified. Myc-CalDAG-GEFI is a significantly more potent activator of HA-Rap1 than RasGRP2-Flag. Ras exchange activity (Fig. 4C). As with RasGRP2, this TPA- After 2 min of incubation with the chemical modulators, the GTP mediated stimulation of Myc-CalDAG-GEFI exchange activity loading of HA-Rap1 by Myc-CalDAG-GEFI is up-regulated by TPA and was partially inhibited by A23187. A23187, and this effect is additive when both chemicals are adminis- To further characterize CalDAG-GEFI, we evaluated ex- tered together. A similar profile is seen after 15-min incubations. Ras- GRP2-Flag is a weaker RapGEF than Myc-CalDAG-GEFI, but its ac- change activity toward Ras in vitro. A GST fusion of CalDAG- tivity is still excited by a 2-min stimulation with TPA and to a lesser GEFI was generated and incubated with recombinant N-Ras extent by A23187. 15-min incubations with these modulators have less protein preloaded with radiolabeled GDP. Fig. 5 shows that, in potent effects on RasGRP2 than 2-min incubations. Con, control; the presence of excess cold GTP, CalDAG-GEFI catalyzed re- DMSO,Me SO. lease of GDP from N-Ras whereas minimal loss of label oc- curred under control conditions. Thus, CalDAG-GEFI has gua- RasGEF activity of RasGRP2 and CalDAG-GEFI and yielded nine nucleotide exchange activity toward N-Ras in vitro.To identical results to those obtained in the immunoprecipitation investigate whether this GEF activity was sensitive to regula- assay (data not shown). Fig. 6 shows that CalDAG-GEFI effi- tion by calcium or diacylglycerol, the exchange reactions were ciently catalyzed GTP loading of Rap1 in COS cells and that repeated in the presence of 2 mM calcium or 10 mM TPA. this RapGEF activity was further stimulated by both TPA and Interestingly, calcium completely inhibited CalDAG-GEFI-cat- calcium ionophor treatment of cells for 2 or 15 min prior to alyzed release of GDP from N-Ras, whereas TPA had no effect harvesting. In contrast, RasGRP2 was less potent than (Fig. 5). CalDAG-GEFI in catalyzing GTP loading of Rap1. The While this work was in progress, Kawasaki et al. (30) re- RapGEF activity of RasGRP2 was stimulated by 2 min of ported that CalDAG-GEFI has potent RapGEF activity. We treatment with TPA or calcium ionophor, although these re- therefore compared the RapGEF activity of RasGRP2 with that sponses were less marked after 15 min of stimulation (Fig. 6B). of CalDAG-GEFI. To this end we used a “pull-down” assay (43), RasGRP2 Stimulates the Growth of NIH3T3 Cells—We next in which RapGTP is isolated from cell lysates using a GST examined whether overexpression of RasGRP2 in NIH3T3 cells fusion of the Raf-1 Ras binding domain (RBD) and quantified would result in morphological transformation or accelerated by immunoblotting. The RafRBD used in these experiments growth, as has been shown for membrane-targeted Sos1 (8). We contains a K85A substitution, which increases .20-fold the found that neither RasGRP2-Flag nor Myc-CalDAG-GEFI gave amount of RasGTP or RapGTP that can be recovered from a cell rise to foci in NIH3T3 focus assays. Moreover, in co-transfec- lysate. The pull-down assay was validated by evaluating the tion assays with RasGRP2-Flag and a hygromycin resistance Fridman, M., Maruta, H., Gonez, J., Walker, F., Treutlein, H., Zeng, J., and Burgess, A. (2000) J. Biol. Chem. 275, 30363–30371. A Novel Ras Exchange Factor 32265 FIG.7. RasGRP2 and CalDAG-GEFI facilitate the growth of NIH3T3 cells co-expressing N-Ras. A, growth curves of polyclonal cell populations expressing N-Ras (E), N-Ras 1Myc-CalDAG-GEFI (), N-Ras1RasGRP2-Flag (l), or mock-transfected with empty vector (M). Each point is the mean (6 S.E., n 5 9) from three independent experiments. Cell numbers are expressed relative to the number of cells counted on day 1(51), typically 10 . These curves show growth in 5% serum; essentially identical results were obtained in 1% serum (data not shown). B, growth rates, relative to the vector-transfected control populations (51), derived from the growth curves shown in panel A. C, subcellular fractionation of cells expressing Myc-CalDAG-GEFI following growth in serum (1), 16 h of serum-free growth (2), or 16 h of serum-free growth followed by a 2-min add-back (AB) of 20% serum prior to harvesting. Equal proportions of S100 (S) and P100 (P) fraction were immunoblotted (IB) with anti-Myc antisera. D, COS cells expressing N-Ras with and without RasGRP2-Flag and Myc-CalDAG-GEFI were serum-starved for 16 h and then harvested directly (2BCS) or treated for 2 min with 20% bovine calf serum (1BCS) prior to harvesting. Con, control. RasGTP levels were measured in a GST-RafRBD-K85A pull-down assay. The graph shows a representative experiment that was repeated two further times. The presence of serum potently stimulates CalDAG-GEFI RasGEF activity toward N-Ras. Expression of CalDAG-GEFI and RasGRP2-Flag was confirmed by immuno- blotting with anti-Myc or anti-Flag antisera, respectively. vector, hygromycin-resistant colonies appeared no different in conditions of the growth rate assays in Fig. 7 (A and B), carried morphology than vector transfected controls, despite clear ex- out in 5% serum, a substantial amount of Myc-CalDAG-GEFI pression of RasGRP2-Flag by Western blotting (data not becomes membrane-associated (Fig. 7C). In the light of this shown). Similarly, assays of growth in reduced serum (5%, 1%, observation, the serum response of Myc-CalDAG-GEFI was and 0.1%) revealed no differences in growth rates between evaluated in COS cells. Fig. 7D shows that serum is a potent polyclonal cell populations expressing RasGRP2-Flag or Myc- activator of the RasGEF activity of CalDAG-GEFI toward N- CalDAG-GEFI over vector-transfected controls (data not Ras but has minimal effect on the RasGEF activity of shown). Given the Ras isoform selectivity of RasGRP2, we RasGRP2-Flag. repeated these experiments in NIH3T3 cells co-transfected DISCUSSION with a low level of wild type N-Ras plasmid sufficient to give approximately 3-fold overexpression of N-Ras over endogenous In this paper we describe the cloning and characterization of levels. Fig. 7 (A and B) shows that polyclonal cell populations RasGRP2, a longer isoform of CalDAG-GEFI. The basic domain expressing wild type N-Ras had a growth rate approximately structure of RasGRP2/CalDAG-GEFI is similar to that of the twice that of empty vector transfected control cell populations, related protein RasGRP/CalDAG-GEFII, which has been de- but that co-expression of RasGRP2-Flag increased this growth scribed previously. One structural feature not commented on in rate another 3-fold. Somewhat unexpectedly, co-expression of the original description of CalDAG-GEFI is the presence of a Myc-CalDAG-GEFI in the N-Ras-expressing cells also resulted short coiled-coil (CC) domain located between the Ras exchange in an increment in growth rate (Fig. 7, A and B). A possible motif and SCR1 box. This is not present in RasGRP, which explanation of this result is shown in Fig. 7 (C and D). The instead has a more extensive CC domain in a C-terminal ex- biochemical assays of RasGRP2 function all required a period tension that is absent from CalDAG-GEFI. The role of the CC of serum starvation; therefore, the cell fractionation and im- domains in RasGRP and RasGRP2 is unknown. In RasGRF1 a munofluorescence studies in Figs. 2 and 3 were performed CC domain co-operates with IQ and pleckstrin homology do- under these conditions. However, Fig. 7C shows that continu- mains for maximal stimulation of exchange activity by calmod- ous growth in serum, or a 2-min treatment with 20% serum ulin-complexed calcium (23), but since the RasGRPs lack these after a period of serum-free growth, significantly increased the other domains and RasGRP2 activity is inhibited by calcium, amount of Myc-CalDAG-GEFI associated with the P100 frac- the CC domain must have some other role. tion. In contrast, there was no effect of serum on the subcellular The novel N-terminal sequences present in RasGRP2 con- distribution of RasGRP2-Flag (data not shown). Thus, in the tain acylation sites that we show here are both N-myristoylated 32266 A Novel Ras Exchange Factor and palmitoylated. The most likely site for palmitoylation is RasGRP; RasGRP is translocated within 2 min, whereas 15 Cys-7, since this is the only cysteine residue in the N-terminal min of TPA treatment is needed to bring CalDAG-GEFI to the 70 amino acids of RasGRP2. We cannot exclude a more remote cell membrane. Taking these data together, we propose that, palmitoylation site, but palmitoylation of other N-myristoy- although RasGRP2 is localized at the plasma membrane, it may in part be sequestered away from Ras until interactions lated proteins is usually close to the N terminus (40). The results of the Triton X-114 assay indicate that only 60% of between TPA and the DGB domain relocate RasGRP2 to a different plasma membrane microdomain. The mechanism for RasGRP2-Flag expressed in COS cells undergoes lipid modifi- the effects of TPA could then be the same for RasGRP, cation. Since co-translational myristoylation precedes post- CalDAG-GEFI, and RasGRP2, namely to co-localize the Ras- translational palmitoylation, it is likely that this result reflects GEF and Ras substrate in the same plasma membrane mi- incomplete myristoylation of RasGRP2. We have observed pre- crodomain. In support of this model, it is interesting to note viously that myristoylation of proteins overexpressed in COS that the GEF activity of RasGRP was highest when it was cells can be incomplete (37); this, coupled with the slightly membrane-targeted using a Ras localization signal (28), a atypical myristoylation motif, probably accounts for the incom- strategy that should exactly co-localize Ras and RasGRP in the plete acylation of RasGRP2. The N-terminal sequence of Ras- same microdomain. An alternative explanation that the TPA GRP also contains a potential N-myristoylation site, but it has stimulates RasGRP2 via an allosteric effect seems unlikely, not yet been determined whether RasGRP is acylated (28). since stimulation was not evident in vitro even with high con- RasGRP2 is localized extensively to the plasma membrane in centrations of TPA. Serum stimulation caused a much more contrast to CalDAG-GEFI, which is almost exclusively cytosolic rapid membrane translocation and activation of CalDAG-GEFI in quiescent cells. These data clearly implicate the N-terminal than TPA, perhaps reflecting selective generation of diacylg- 60 amino acids of full-length RasGRP2 as containing a plasma lycerol in physiologically relevant plasma membrane microdo- membrane localization motif. Since the combination of myris- mains, rather than general loading of cell membranes with toylation and palmitoylation is well established as a targeting DGB binding sites. signal (44), we propose that dual acylation targets RasGRP2 to RasGRP2 selectively activates N-Ras and Ki-Ras, but not the plasma membrane. Ha-Ras. In contrast, RasGRP activates Ha-Ras and mRasGRP CalDAG-GEFI has been reported to be a RapGEF with no when co-transfected with N-, K-, or Ha-Ras results in mitogen- RasGEF activity (30). We show here, however, that CalDAG- activated protein kinase activation, consistent with activation GEFI is indeed a RasGEF, albeit with an isoform specificity of all three isoforms (27, 28). One explanation of these data is that is restricted to N- and Ki-Ras. Ha-Ras was the only Ras simply that sequence differences between RasGRP and Ras- isoform tested by Kawasaki et al. (30), and we are in full GRP2 result in different substrate specificities. An alternative agreement with the conclusions of their study that CalDAG- hypothesis invokes recent work showing that Ras isoforms GEFI cannot activate this Ras isoform. More intriguingly, operate in functionally distinct microdomains of the plasma CalDAG-GEFI and RasGRP2 have different basal exchange membrane (45); signaling by Ha-Ras, anchored to the plasma activities and show distinct in vivo responses to stimulation by membrane by palmitoylation and farnesylation, is dependent TPA and calcium. CalDAG-GEFI has potent RapGEF activity, on the integrity of cholesterol-rich microdomains, whereas sig- which is stimulated both by acute elevations of cytosolic cal- naling by Ki-Ras, anchored by a charged polybasic domain and cium and exposure to TPA, but it has almost no N-Ras GEF farnesylation, does not require lipid rafts. In this context, the activity until subject to prolonged stimulation by TPA or acute different Ras isoform specificity of RasGRP2 and RasGRP may stimulation by serum. The TPA-stimulated RasGEF activity of be explained by postulating that RasGRP2 gains access only to CalDAG-GEFI is inhibited by elevations of cytosolic calcium. In the N- and Ki-Ras microdomains because N-terminal palmitoy- contrast, RasGRP2 is a less potent RapGEF but a more potent lation imposes a restraint on RasGRP2 mobility within the RasGEF than CalDAG-GEFI. The RapGEF activity of Ras- plasma membrane that is not experienced by RasGRP. GRP2 is stimulated by TPA and calcium, whereas the RasGEF Somewhat surprisingly, moderate overexpression of plasma activity is stimulated by TPA, but inhibited by calcium. membrane-localized RasGRP2 did not morphologically trans- The activation of RasGEF activity in CalDAG-GEFI by TPA form NIH3T3 cells, although we were able to show an N-Ras- or serum stimulation correlates temporally with translocation dependent stimulation of cell growth by both RasGRP2 and of CalDAG-GEFI to the plasma membrane. We propose there- CalDAG-GEFI. In fibroblasts, activated Rap can partially an- fore that this major change in subcellular localization is fun- tagonize Ras transforming activity (46). It is therefore tempt- damental to Ras activation by CalDAG-GEFI, analogous to the ing to speculate that the lack of transforming ability of Ras- mechanisms described for Sos1 and RasGRF2. Similarly, there GRP2 and CalDAG-GEFI may reflect their co-activation of Ras is a significant translocation of RasGRP to the plasma mem- and Rap. In this context it is worth noting that, since calcium brane in response to TPA treatment, concomitant with the inhibits the RasGEF activity of RasGRP2/CalDAG-GEFI while stimulation of RasGEF activity (27). RasGRP2 has constitutive stimulating the RapGEF activity, a calcium signal effectively RasGEF activity, presumably reflecting constitutive plasma shifts the GEF activity of RasGRP2 and CalDAG-GEFI from membrane localization. However, in contrast to CalDAG-GEFI Ras to Rap. Together, these observations therefore reveal an and RasGRP, there is no detectable change in the subcellular interesting and novel molecular mechanism for utilizing a sin- distribution of RasGRP2 to correspond with TPA stimulation of gle exchange factor to co-ordinate Ras and Rap signaling RasGEF activity. events. Is there a common mechanism underlying these TPA effects? Acknowledgments—We thank Masha Fridman and Tony Burgess for The DGB domain of RasGRP is clearly responsible for TPA- the GST-RafRBD-K85A construct and for sharing results prior to pub- mediated membrane translocation because deletion of the do- lication, Marilyn Resh for helpful advice, and David Bowtell for the main blocks RasGRP membrane binding (27) and the isolated mSos1 cDNA. DGB domain can target heterologous proteins to the plasma REFERENCES membrane in a TPA-dependent manner (28). The DGB domain 1. Boriack-Sjordan, P. A., Margarit, S. M., Bar-Sagi, D., and Kuriyan, J. (1998) of CalDAG-GEFI can effect a similar change in subcellular Nature 394, 337–343 localization, since TPA treatment results in membrane trans- 2. Wittinghofer, F. (1998) Nature 394, 317–320 location of CalDAG-GEFI, albeit with different kinetics from 3. Bos, J. L. (1997) Biochim. Biophys. Acta 1333, M19-M31 A Novel Ras Exchange Factor 32267 4. Simon, M. A., Bowtell, D. D. L., Dodson, G. S., Laverty, T. R., and Rubin, G. M. 25. Cerione, R. A., and Zheng, Y. (1996) Curr. Opin. Cell Biol. 8, 216 –222 (1991) Cell 67, 701–716 26. Nimnual, A. S., Yatsula, B. A., and Bar Sagi, D. (1998) Science 279, 560 –563 5. Li, N., Batzer, A., Daly, R., Yajnik, V., Skolnik, E., Chardin, P., Bar Sagi, D., 27. Ebinu, J. O., Bottorff, D. A., Chan, E. Y. W., Stang, S. L., R. J., D., and Stone, Margolis, B., and Schlessinger, J. (1993) Nature 363, 85– 88 J. C. (1998) Science 280, 1082–1086 6. Chardin, P., J. H., C., Gale, N. W., Van Aelst, L., Schlessinger, J., Wigler, 28. Tognon, C. E., Kirk, H. E., Passmore, L. A., Whitehead, I. P., Der, C. J., and M. H., and Bar-Sagi, D. (1993) Science 260, 1338 –1343 Kay, R. J. (1998) Mol. Cell. Biol. 18, 6995–7008 7. Buday, L., and Downward, J. (1993) Cell 73, 611– 620 29. Kedra, D., Serossi, E., Fransson, I., Trifunovic, J., Clark, M., Lagercrantz, J., 8. Aronheim, A., Engelberg, D., Li, N., al Alawi, N., Schlessinger, J., and Karin, Blennow, E., Mehlin, H., and Dumanski, J. (1997) Hum. Genet. 100, M. (1994) Cell 78, 949 –961 611– 619 9. Ravichandran, K. S., Lorenz, U., Shoelson, S. E., and Burakoff, S. J. (1995) 30. Kawasaki, H., Springett, G. M., Toki, S., Canales, J. J., Harlan, P., Blumen- Mol. Cell. Biol. 15, 593– 600 stiel, J. P., Chen, E. J., Bany, I. A., Mochizuki, N., Ashbacher, A., Matsuda, 10. Lioubin, M. N., Myles, G. M., Carlberg, K., Bowtell, D., and Rohrschneider, M., Housman, D. E., and Graybiel, A. M. (1998) Proc. Natl. Acad. Sci. L. R. (1994) Mol. Cell. Biol. 14, 5682–5691 U. S. A. 95, 13278 –13283 11. Harmer, S. L., and DeFranco, A. L. (1997) Mol. Cell. Biol. 17, 4087– 4095 31. Bergman, L., Silins, G., Grimmond, S., Hummerich, H., Stewart, C., Little, P., 12. Kouhara, H., Hadari, Y. R., Spivak-Kroizman, Schilling, J., Bar-Sagi, D., Lax, and Hayward, N. K. (1999) Genomics 55, 49 –56 I., and Schlessinger, J. (1997) Cell 89, 693–702 32. Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990) J. 13. Lutterell, L., Ferguson, S., Daaka, Y., Miller, W., Maudsley, S., Della Roca, G., Mol. Biol. 215, 403– 410 Lin, F.-T., Kawakatsu, H., Owada, K., Luttrell, D., Caron, M., and Lefkow- 33. Evan, G., Lewis, G., Ramsey, G., and Bishop, J. (1985) Mol. Cell. Biol. 5, itz, R. (1999) Science 283, 655– 661 3610 –3615 14. Chen, Y., Grall, D., Salcini, A., Pelicci, P., Poussegur, J., and Van Obberghen- 34. Porfiri, E., Evans, T., Chardin, P., and Hancock, J. F. (1994) J. Biol. Chem. Schilling, E. (1996) EMBO J. 15, 1037–1044 269, 22672–22677 15. Chen, R., Corbalan-Garcia, S., and Bar-Sagi, D. (1997) EMBO J. 16, 35. Downward, J., Graves, J. D., Warne, P. H., Rayter, S., and Cantrell, D. A. 1351–1359 (1990) Nature 346, 719 –723 16. Qian, X., Vass, W., Papageorge, A., Anborgh, P., and Lowy, D. (1998) Mol. Cell. 36. Roy, S., McPherson, R. A., Apolloni, A., Yan, J., Clyde-Smith, J., and Hancock, Biol. 18, 771–778 J. F. (1998) Mol. Cell. Biol. 18, 3947–3955 17. Wang, W., Fisher, E. M., Jia, Q., Dunn, J. M., Porfiri, E., Downward, J., and 37. Cadwallader, K., Paterson, H., Macdonald, S. G., and Hancock, J. F. (1994) Egan, S. E. (1995) Nat. Genet. 10, 294 –300 Mol. Cell. Biol. 14, 4722– 4730 18. Byrne, J. L., Paterson, H. F., and Marshall, C. J. (1996) Oncogene 13, 38. Roy, S., Lane, A., Yan, J., McPherson, R., and Hancock, J. F. (1997) J. Biol. 2055–2065 Chem. 272, 20139 –20145 19. Shou, C., Farnsworth, C. L., Neel, B. G., and Feig, L. A. (1992) Nature 358, 39. Huang, D. C. S., Marshall, C. J., and Hancock, J. F. (1993) Mol. Cell. Biol. 13, 351–354 2420 –2431 20. Fam, N. P., Fan, W. T., Wang, Z., Zhang, L. J., Chen, H., and Moran, M. F. 40. Wedegaertner, P. B., Wilson, P. T., and Bourne, H. R. (1995) J. Biol. Chem. (1997) Mol. Cell. Biol. 17, 1396 –1406 270, 503–506 21. Sturani, E., Abbondio, A., Branduardi, P., Ferrari, C., Zippel, R., Martegani, 41. Gutierrez, L., Magee, A. I., Marshall, C. J., and Hancock, J. F. (1989) EMBO E., Vanoni, M., and Denis Donini, S. (1997) Exp. Cell Res. 235, 117–123 J. 8, 1093–1098 22. Brambilla, R., Gnesutta, N., Minichiello, L., White, G., Roylance, A. J., Herron, 42. Resh, M. (1994) Cell 76, 411– 413 C. E., Ramsey, M., Wolfer, D. P., Cestari, V., Rossi Arnaud, C., Grant, S. G., 43. de Rooij, J., and Bos, J. L. (1997) Oncogene 14, 623– 625 Chapman, P. F., Lipp, H. P., Sturani, E., and Klein, R. (1997) Nature 390, 44. Resh, M. D. (1996) Cell Signal 8, 403– 412 281–286 23. Buchsbaum, R., Telliez, J. B., Goonesekera, S., and Feig, L. A. (1996) Mol. Cell. 45. Roy, S., Luetterforst, R., Harding, A., Apolloni, A., Etheridge, M., Stang, E., Rolls, B., Hancock, J. F., and Parton, R. G. (1999) Nat. Cell Biol. 1, 98 –105 Biol. 16, 4888 – 4896 24. Farnsworth, C. L., Freshney, N. W., Rosen, L. B., Ghosh, A., Greenberg, M. E., 46. Kitayama, H., Sugimoto, Y., Matsuzaki, T., Ikawa, Y., and Noda, M. (1989) and Feig, L. A. (1995) Nature 376, 524 –527 Cell 56, 77– 84 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Biological Chemistry Unpaywall

Characterization of RasGRP2, a Plasma Membrane-targeted, Dual Specificity Ras/Rap Exchange Factor

Download PDF
8 pages

Loading next page...
 
/lp/unpaywall/characterization-of-rasgrp2-a-plasma-membrane-targeted-dual-f0tuDx78SW

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Unpaywall
ISSN
0021-9258
DOI
10.1074/jbc.m006087200
Publisher site
See Article on Publisher Site

Abstract

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 275, No. 41, Issue of October 13, pp. 32260 –32267, 2000 © 2000 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Characterization of RasGRP2, a Plasma Membrane-targeted, Dual Specificity Ras/Rap Exchange Factor* Received for publication, July 11, 2000, and in revised form, July 27, 2000 Published, JBC Papers in Press, July 28, 2000, DOI 10.1074/jbc.M006087200 Jodi Clyde-Smith‡, Gint Silins§, Michael Gartside§, Sean Grimmond§, Maria Etheridge‡, Ann Apolloni‡, Nick Hayward§, and John F. Hancock‡¶ From the ‡Queensland Cancer Fund Laboratory of Experimental Oncology, Department of Pathology, University of Queensland Medical School, Herston Road, Brisbane 4006, Queensland and the §Queensland Cancer Fund Research Unit, Queensland Institute of Medical Research, Bancroft Centre, Herston Road, Brisbane 4006, Queensland, Australia Ras proteins operate as molecular switches in signal GTPase-activating proteins (GAPs) which activate and deacti- vate Ras, respectively. A number of RasGEFs have been cloned, transduction pathways downstream of tyrosine kinases and G-protein-coupled receptors. Ras is switched from which share a core catalytic exchange domain and probably the inactive GDP-bound state to the active GTP-bound release GDP from Ras via a common mechanism (1, 2). These state by guanine nucleotide exchange factors (GEFs). various RasGEFs possess different regulatory domains and We report here the cloning and characterization of Ras- have different tissue expression patterns (reviewed in Ref. 3); GRP2, a longer alternatively spliced form of the recently these observations suggest that they operate in signaling path- cloned RapGEF, CalDAG-GEFI. A unique feature of Ras- ways that activate Ras effectors for different biological GRP2 is that it is targeted to the plasma membrane by a responses. combination of N-terminal myristoylation and palmitoy- Sos1 and Sos2, are RasGEFs that are widely expressed in lation. In vivo, RasGRP2 selectively catalyzes nucleotide mammalian cells (4 – 6). Sos is normally localized to the cell exchange on N- and Ki-Ras, but not Ha-Ras. RasGRP2 cytosol, but is recruited to the plasma membrane in response to also catalyzes nucleotide exchange on Rap1, but this growth factor receptor activation. Since Ras must be localized RapGEF activity is less potent than that associated with at the plasma membrane in order to signal, Ras activation by CalDAG-GEFI. The nucleotide exchange activity of Ras- Sos can be efficiently regulated by simply controlling the sub- GRP2 toward N-Ras is stimulated by diacylglycerol and cellular localization of Sos (7, 8). Sos is recruited to the plasma inhibited by calcium. The effects of diacylglycerol and membrane in complexes with the adapter proteins Grb2 and calcium are additive but are not accompanied by any Shc (5, 6), which bind to autophosphorylated tyrosine kinase detectable change in the subcellular localization of Ras- receptors (9 –11) or to transphosphorylated docking proteins GRP2. In contrast, CalDAG-GEFI is localized predomi- (12). Similarly, G-protein-coupled receptors that activate Src nantly to the cytosol and lacks Ras exchange activity in and Shc can also recruit Grb2-Sos to the cell membrane and vivo. However, prolonged exposure to phorbol esters, or activate Ras (13, 14). Although plasma membrane recruitment growth in serum, results in localization of CalDAG-GEFI to the cell membrane and restoration of Ras exchange of Sos is sufficient to activate Ras, it is also clear that more activity. Expression of RasGRP2 or CalDAG-GEFI in complex intermolecular interactions also influence the activity NIH3T3 cells transfected with wild type N-Ras results in of this exchange factor (15–18). an accelerated growth rate but not morphologic trans- A second family of RasGEFs comprises the related proteins formation. Thus, under appropriate growth conditions, RasGRF1 and RasGRF2 (19, 20). RasGRF1 is expressed only in CalDAG-GEFI and RasGRP2 are dual specificity Ras the central nervous system, specifically in the synaptosomal and Rap exchange factors. fraction of neuronal cells (21) where it plays a role in the establishment of long term memory (22). RasGRF2 is expressed both within and without the central nervous system (20). In Ras proteins are small GTPases that operate as binary contrast to Sos, RasGRF1 is constitutively localized to the switches in key signal transduction pathways regulating cell plasma membrane, probably anchored through its N-terminal proliferation and differentiation. The amount of active GTP- pleckstrin homology domain (23). RasGRF1 exchange activity bound Ras is in turn controlled by the actions of two classes of is quiescent until activated by a rise in intracellular calcium regulator, guanine nucleotide exchange factors (GEFs) and (23, 24). RasGRF2 is also regulated by calcium, but unlike RasGRF1, RasGRF2 is localized to the cytosol in resting cells * This work was supported in part by grants from the Queensland until a rise in intracellular calcium levels causes relocalization Cancer Fund (to J. F. H. and N. H.) and the National Health and of RasGRF2 to the cell periphery and activation of its exchange Medical Research Council, Australia (to N. H.). The costs of publication activity (20). A feature common to Sos and RasGRF is the of this article were defrayed in part by the payment of page charges. presence of Dbl homology domains comprising the catalytic This article must therefore be hereby marked “advertisement”inac- cordance with 18 U.S.C. Section 1734 solely to indicate this fact. core sequences of exchange factors for the Rho family of small The nucleotide sequence(s) reported in this paper has been submitted GTPases (25). The binding of phosphatidylinositol 3,4-bisphos- TM to the GenBank /EBI Data Bank with accession number(s) AF043723 phate to a pleckstrin homology domain flanking the Dbl homol- and AF043722. ¶ Supported by the Royal Children’s Hospital Foundation, Queens- land. To whom correspondence should be addressed: Dept. of Pathology, University of Queensland Medical School, Herston Rd., Brisbane 4006, PCR, polymerase chain reaction; ORF, open reading frame; GST, glu- Queensland, Australia. Tel.: 61-7-3365-5340; Fax: 61-7-3365-5511; E- tathione S-transferase; CC, coiled-coil; DMEM, Dulbecco’s modified mail: [email protected]. Eagle’s medium; PAGE, polyacrylamide gel electrophoresis; RBD, Ras The abbreviations used are: GEF, guanine nucleotide exchange binding domain; DGB, diacylglycerol binding; BHK, baby hamster kid- ney; TPA, 12-O-tetradecanoylphorbol-13-acetate. factor; GAP, GTPase-activating protein; EST, expressed sequence tag; 32260 This paper is available on line at http://www.jbc.org This is an Open Access article under the CC BY license. A Novel Ras Exchange Factor 32261 TrisCl, pH 7.5, 150 mM NaCl, 1% aprotinin, 1 mM phenylmethylsulfonyl ogy domain of Sos1 activates exchange activity toward Rac (26), fluoride). Insoluble cellular debris was removed by centrifugation at indicating that such RasGEFs coordinate activation of Ras and 14,000 rpm for 5 min at 4 °C. 10 ml of the supernatant was reserved for Rho family proteins. later immunoblotting; the remainder was diluted in 1 ml of binding Most recently, a new family of Ras subfamily GEFs have buffer (50 mM TrisCl pH7.5, 10 mM MgCl , 0.5 mg/ml bovine serum been identified that are regulated by diacylglycerol and cal- M dithiothreitol, 100 mM NaCl) and incubated for 90 min, albumin, 0.5 m cium. RasGRP (also referred to as CalDAG-GEFII) activates rotating at 4 °C with 20 mg of GST-RafRBD-K85A protein fused to Sepharose (50% suspension). Samples were washed three times with N-, Ha-, and Ki-Ras with varying efficiencies (27, 28), whereas binding buffer before the beads were resuspended in 20 mlof23 SDS- the closely related CalDAG-GEFI (also referred to as PAGE sample buffer. Proteins from the bead capture and 10 mlof HCDC25L (Ref. 29)) activates Rap1 but not Ha-Ras (30). In cleared supernatant were resolved on 15% acrylamide gels, blotted with generating a transcript map of chromosome 11q13 (31), we anti-Ras or anti-HA (Rap) antisera and visualized by chemilumines- independently cloned CalDAG-GEFI and identified an addi- cence. Blots were quantified by phosphorimaging. The amount of GTP- tional 59-coding exon. We report here the characterization of Rap or GTP-Ras captured by the GST-RafRBD-K85A beads was ex- pressed as a percentage of the Ras or Rap present in the supernatant RasGRP2: the widely expressed, longer isoform of CalDAG- (estimated from the 10-ml input). Expression of RasGRP2-Flag or Myc- GEFI. We show that RasGRP2 is constitutively localized to the CalDAG-GEFI was confirmed by immunoblotting using anti-Flag or plasma membrane by N-terminal acylation, a feature not pre- anti-Myc antisera, respectively. viously described for any RasGEF. A consequence of constitu- Cell Culture and Metabolic Labeling Assays—COS cells were trans- tive plasma membrane localization is that RasGRP2 operates fected by electroporation and BHK cells by lipofection (36). Metabolic as an exchange factor for N-Ras and Ki-Ras in addition to being labeling with [ S]methionine was carried out for4hin methionine-free DMEM containing [ S]methionine (50 mCi/ml). Labeling with an activator of Rap1. RasGRP2, like CalDAG-GEFI, cannot 3 3 [ H]palmitic acid (0.5 mCi/ml) or [ H]myristic acid (0.5 mCi/ml) was activate Ha-Ras. CalDAG-GEFI and RasGRP2 also show quite M sodium pyruvate. carried out for4hin DMEM supplemented with 5 m distinct responses to calcium, diacylglycerol, and serum stim- Cleared lysates, normalized for protein content, were immunoprecipi- ulation revealing a series of interesting possibilities for co- tated with M2 anti-Flag Sepharose or 9E10-coated protein G-Sepharose ordinate regulation of Ras and Rap function. as required. Immunoprecipitated proteins were resolved by SDS-PAGE and visualized by fluorography (for details, see Ref. 37). Where indi- EXPERIMENTAL PROCEDURES cated, gels were agitated in 1 M neutral hydroxylamine or 1 M TrisCl, pH 7.5, for1hat room temperature prior to fluorography. Indirect immu- Cloning of RASGRP2—Cosmid cSRL-20 h12, derived from a chromo- nofluorescence for Myc and Flag epitopes was performed as described some 11 cosmid library that contained the human MCG7 (RASGRP2) (36). Prior to harvesting for fractionation studies, COS cells were se- gene (31), was sequenced and the sequence data aligned to the Gen- rum-starved for 4 –16 h as indicated. Hypotonic lysis, homogenization, Banky Expressed Sequence Tag (EST) data base using the program and centrifugation were carried out as described previously (38). BLASTN (32). The cDNA clones for matching EST entries T78563 NIH3T3 Growth Assays—NIH3T3 cells were transfected using cal- (clone 113434) and AA035643 (clone 471819) were obtained from Ge- cium phosphate precipitation as described (39). To generate polyclonal nome Systems Inc. and sequenced to completion with gene-specific and populations, cells seeded on 10-cm plates were co-transfected with 1.5 vector-specific primers. The data base entry of a third matching EST mg of pSV-hyg, together with 1.5 mg of EXV expression plasmids for clone, HIBBP12, contained an additional stretch of nucleotides up- N-Ras, RasGRP2-Flag, Myc-CalDAG-GEFI, or empty vector. After 48 h, stream of the sequences of the other ESTs, and was therefore suspected cultures were selected with 300 mg/ml hygromycin. Surviving cells were to contain an additional upstream exon of RASGRP2, but since this pooled and grown continually in hygromycin. For growth assays, 10 clone was not commercially available the new exon was sought by PCR cells (per well) were plated in multiwell dishes in DMEM containing from a genomic template, bacterial artificial chromosome clone 137C7 5%, 1%, or 0.1% serum. Cells were trypsinized and counted every 24 h (Genome Systems Inc). The composite cDNA sequence of RASGRP2 for 4 days with each time point counted in triplicate. Three independent was suspected to contain a single nucleotide error originating from transfections were analyzed. Expression of RasGRP2-Flag or Myc- within clone 471819 and the correct nucleotide sequence was therefore CalDAG-GEFI was confirmed by immunoblotting or indirect immuno- sought by reverse transcription-polymerase chain reaction of the cDNA fluorescence using anti-Flag or anti-Myc antisera, respectively. Expres- fragment from a human cDNA pool. The resulting PCR product was sion of N-Ras was verified by immunoblotting using Y13-259. ligated into pGEM-T (Promega) and the clone, designated MCG7- pGEM, sequenced. The corrected nucleotide sequence of RASGRP2 was RESULTS translated in all possible reading frames and compared with the non- redundant protein data base using the program BLASTX (32). One open Identification of a Full-length cDNA for CalDAG-GEF1/Ras- reading frame (ORF), encoding RasGRP2– 609, was assigned on the GRP2—The coding sequence for a human RasGEF was discov- basis of homology with the Caenorhabditis elegans protein F25B3.3 ered during the generation of a transcript map of a region of (GenBanky accession no. Z70752) and is identical to HCDC25L (Gen- human chromosome band 11q13 (31). The gene, originally des- Banky accession no. Y12336) (29) and CalDAG-GEFI (GenBanky ac- ignated MCG7, was subsequently assigned the locus name cession no. U78170) (30). Conceptual translation of the ORF from the alternate translation start site upstream of the one utilized in RASGRP2 by the human genome project nomenclature com- RasGRP2– 609, contained in the sequence from bacterial artificial chro- mittee. A cDNA corresponding to RASGRP2 was independently mosome clone 137C7, was designated RasGRP2– 671. These sequences isolated by two other groups (29, 30) and subsequently charac- have been deposited under GenBanky accession nos. AF043723 and terized as a 609-amino acid RapGEF referred to as CalDAG- AF043722, respectively. GEFI. However, during our independent isolation of a full- Plasmid Construction—A Myc tag was cloned onto the N terminus of length cDNA for RASGRP2, we identified an additional EST RasGRP2– 609 by recombination of MCG7-pGEM, EST clone 113434, and pGEM-11Z-Myc-tag (33) to give clone Myc-RasGRP2– 609 (referred that contained sequence 59 to that previously reported to con- to as Myc-CalDAG-GEFI hereafter). Myc-CalDAG-GEFI was digested stitute the N terminus of CalDAG-GEFI (see “Experimental with BamHI to remove the Myc tag and ligated into pGEX-2T (Amer- Procedures” for details). This additional sequence together sham Pharmacia Biotech). The final construct has an ORF that fuses with the general layout of the full-length RASGRP2 gene prod- GST to the front of RasGRP2– 609. A C-terminal Flag-tagged uct is shown in Fig. 1. The core exchange factor domains com- RasGRP2– 671 cDNA was generated by PCR. The final Flag-tagged prising the Ras exchange motif and the SCR1, SCR2, and SCR3 construct contains an ORF with the amino acids DYKDDDDKstop (Flag) added immediately after the last residue of RasGRP2 (hereafter motifs plus the diacylglycerol binding (DGB) domain and EF referred to as RasGRP2-Flag). The Myc- and Flag-tagged constructs hands have been described previously (27, 30). A novel feature were directionally cloned into the eukaryotic expression vector EXV. of the N terminus of full-length RasGRP2 is the presence of Ras Guanine Nucleotide Exchange Assays—In vitro and in vivo nu- potential consensus sequences for co-translational myristoyla- cleotide exchange assays were performed as described previously tion and posttranslational palmitoylation. Northern blot anal- (34, 35). ysis showed that RASGRP2 was expressed predominantly as Ras/Rap “Pull-down” Assays—Frozen cells were lysed in 200 mlof ice-cold lysis buffer (0.5% Triton X-100, 0.5% deoxycholate, 20 mM 2.4-kilobase pair transcripts in all tissues tested: heart, brain, 32262 A Novel Ras Exchange Factor FIG.1. Structure of RasGRP2. A, the N-terminal 62 amino acids of RasGRP2 that are not present in CalDAG-GEFI. The underlined residues represent the previously reported first 8 amino acids of CalDAG-GEFI. The sequence of the remainder of RasGRP2 is exactly as reported for CalDAG-GEFI. B, domain structure of RasGRP2. The N-terminal 62 amino acids of RasGRP2 including the acylation motifs (M1P) are replaced with a Myc tag in Myc-CalDAG-GEFI. Motifs are indicated by boxes and include the Ras exchange motif (REM), CDC25 homology domain comprising SCR1–SCR3, CC, two EF hands, and a DGB domain. placenta, lung, liver, skeletal muscle, kidney, and pancreas. Expression was highest in heart, brain, and kidney as vari- ously reported previously (29, 30), but in contrast to the pub- lished data we also detected additional transcripts of 0.8 kilo- base pairs in heart and 4.7 kilobase pairs in heart, lung, and placenta (results not shown). For the sake of brevity, we will refer to the short form of the RASGRP2 gene product as CalDAG-GEFI because this is the previously published name, and the full-length gene product as RasGRP2, after the name assigned to the gene by human genome project nomenclature committee. RasGRP2 Is Palmitoylated and Myristoylated—The N-termi- nal sequence of full-length RasGRP2 comprises MGTQRLCGR. Although the common consensus sequences for myristoylation include serine or threonine at amino acid 16, small hydropho- bic residues are also tolerated, as in the myristoylation motifs of Arf, A kinase, Arg, and Abl. Moreover, the presence of cys- teine at amino acid 17 suggested that the N terminus of RasGRP2 may also be palmitoylated (40). We therefore inves- tigated whether RasGRP2 was lipid-modified in vivo. To this FIG.2. RasGRP2 is lipid-modified and localizes to the plasma end, a Flag epitope was engineered onto the C terminus of the membrane. A, lysates of COS cells expressing RasGRP2-Flag or Myc- protein leaving the N-terminal sequences intact (5RasGRP2- CalDAG-GEFI were prepared in Triton X-114 and partitioned into Flag) (Fig. 1). A second construct, which represents the shorter detergent (D) and aqueous (Aq) phases as described (41). Equal propor- tions of each phase were resolved by SDS-PAGE, immunoblotted with alternative spliced form of RasGRP2, was also engineered by anti-Myc or anti-Flag antisera, and detected by enhanced chemilumi- replacing the N-terminal 62 amino acids of RasGRP2 with a nescence. Hydrophobic proteins partition into the detergent phase, Myc epitope (5Myc-CalDAG-GEFI). RasGRP2-Flag and Myc- whereas hydrophilic proteins partition into the aqueous phase of Triton X-114. By this analysis, RasGRP2 is significantly more hydrophobic CalDAG-GEFI were transiently expressed in COS cells and the than CalDAG-GEFI, consistent with the N terminus of RasGRP2 being relative hydrophobicity of the two proteins assessed in a Triton lipid-modified. B, RasGRP2-Flag and Myc-CalDAG-GEFI proteins ex- X-114 partitioning assay (41). Fig. 2A shows that 60% of Ras- pressed in COS cells were metabolically labeled with [ S]methionine 3 3 GRP2-Flag partitions into the detergent phase of Triton X-114, (Met), [ H]myristic acid (Mys), or [ H]palmitic acid (Pl). Proteins were whereas Myc-CalDAG-GEFI partitions exclusively into the immunoprecipitated (IP) from cell lysates with anti-epitope antisera, resolved by SDS-PAGE, and visualized by fluorography. A duplicate of aqueous phase of Triton X-114. Thus, a significant proportion the tritium-labeled gel was soaked in 1 M hydroxylamine for 1 h prior to of RasGRP2-Flag is hydrophobic, whereas all Myc-CalDAG- fluorography (1hydrox). RasGRP2-Flag incorporates label from both GEFI is hydrophilic. This result is consistent with the N ter- tritiated fatty acids, showing that it is co-translationally myristoylated minus of RasGRP2 being lipid-modified. and posttranslationally palmitoylated. Note that the palmitate but not the myristate label is released by hydroxylamine. CalDAG-GEFI, which COS cells expressing RasGRP2-Flag or Myc-CalDAG-GEFI lacks the consensus sites for these modifications, is labeled by neither were then metabolically labeled with [ H]myristic or fatty acid. C, RasGRP2-Flag and Myc-CalDAG-GEFI were expressed in [ H]palmitic acid and cell lysates analyzed by immunoprecipi- BHK cells and visualized by indirect immunofluorescence using anti- tation and fluorography. Fig. 2B shows that RasGRP2-Flag epitope antisera. Cells expressing RasGRP2-Flag show plasma mem- 3 3 brane staining, whereas cells expressing CalDAG-GEFI show cytosolic incorporated label from [ H]myristic and [ H]palmitic acid, staining. whereas Myc-CalDAG-GEFI incorporated neither label. The metabolic labeling of RasGRP2-Flag was repeated and the im- munoprecipitated proteins divided between two gels, which and therefore shows that the palmitate labeling is not a con- were soaked in 1 M hydroxylamine or TrisCl prior to fluorogra- sequence of metabolism of the palmitate label to myristate. phy. Fig. 2B shows that hydroxylamine treatment caused loss RasGRP2 Is Localized to the Plasma Membrane—The pre- of the palmitate label but not the myristate label from Ras- ceding results demonstrate that RasGRP2-Flag is N-terminally GRP2-Flag. This result indicates that palmitate, but not my- myristoylated and palmitoylated: a combination of lipid modi- ristate, is attached to RasGRP2-Flag via a thioester linkage fications that constitutes a plasma membrane targeting se- A Novel Ras Exchange Factor 32263 FIG.3. CalDAG-GEFI translocates to the membrane fraction in response to prolonged exposure to TPA. Equivalent aliquots of COS cells expressing RasGRP2-Flag or Myc-CalDAG-GEFI were treated with A23187 (10 mM) and/or TPA (100 nM), or Me SO (control) for 2 or 15 min. The cells were harvested and equal proportions of each S100 (cytosolic, S) and P100 (membrane, P) fraction were immuno- blotted with anti-Myc or anti-Flag antisera. Myc-CalDAG-GEFI is found predominantly in the S100 cytosolic fraction, whereas full-length RasGRP2-Flag is equally distributed between S100 and P100 fractions. Neither protein shows any significant change in subcellular distribu- tion in response to a 2-min exposure to calcium ionophor or TPA or both agents together; however, the amount of CalDAG-GEFI found in the P100 fraction increases substantially after a 15-min exposure to TPA. The distribution of RasGRP2-Flag was unchanged after 15 min of exposure to TPA (data not shown). quence (42). To determine the subcellular localization of Ras- GRP2, BHK cells expressing RasGRP2-Flag or Myc-CalDAG- GEFI were examined using indirect immunofluorescence. Fig. 2C shows that RasGRP2-Flag localizes extensively to the plasma membrane, although there is also some soluble cytoso- lic staining. In contrast, Myc-CalDAG-GEFI is confined to the cytosol with no evidence of plasma membrane staining. Simi- larly, when COS cells expressing the epitope-tagged variants of FIG.4. RasGRP2 functions as a RasGEF in vivo. A, wild type RasGRP2 were fractionated and immunoblotted for the Flag or Ki-Ras (B), N-Ras, or Ha-Ras proteins were transiently expressed in COS cells with and without RasGRP2-Flag or Myc-CalDAG-GEFI. The Myc epitopes, RasGRP2-Flag was found to be distributed ap- cells were metabolically labeled with [ P]orthophosphate and Ras pro- proximately equally between the P100 (membrane) and S100 teins immunoprecipitated; guanine nucleotides were eluted from the (cytosolic) fractions (Fig. 3), consistent with the proportion immunoprecipitates, resolved by TLC, and quantified by phosphorim- shown to be lipidated by Triton X-114 partitioning (Fig. 2A). aging. Duplicate aliquots of cells from each electroporation were ana- lyzed by immunoblotting to confirm that all transfected proteins were Myc-CalDAG-GEFI was found predominantly in the S100 frac- equally expressed (data not shown). RasGRP2-Flag selectively loads tion (Fig. 3). N-Ras and Ki-Ras with GTP, whereas Myc-CalDAG-GEFI does not RasGRP has been shown to translocate from cytosol to mem- nucleotide exchange any Ras protein in vivo. B, COS cells expressing brane when cells are treated with 12-O-tetradecanoylphorbol- N-Ras with and without RasGRP2-Flag were metabolically labeled as in A and treated with A23187 (10 mM) and/or TPA (100 nM) for 2 min 13-acetate (TPA) for 2 min (27, 28). We therefore examined prior to harvesting. The TLC plate from a representative experiment whether TPA would affect the subcellular localization of Ras- shows that the GEF activity of RasGRP2-Flag is up-regulated by TPA GRP2. Fig. 3 shows that there was no detectable change in the and down-regulated by an increase in intracellular calcium. C, the subcellular distribution of either RasGRP2-Flag or Myc- results from the experiment shown in panel B and three to five similar CalDAG-GEFI in response to a 2-min exposure to either TPA or experiments in which cells were treated with A23187 and/or TPA for 2 or 15 min prior to harvesting. After 15 min of exposure to TPA, Myc- a calcium ionophor (A23187), or both agents together. How- CalDAG-GEFI potently stimulates GTP loading of N-Ras. The control ever, a substantial proportion of Myc-CalDAG-GEFI did trans- cells (Con) expressing N-Ras without exchange factors were incubated locate to the P100 fraction when cells were treated with TPA for 15 min with A23187 and/or TPA. DMSO,Me SO. for 15 min (Fig. 3). The presence or absence of calcium ionophor during exposure to TPA did not significantly affect the trans- similar magnitude to that reported recently on Ha-Ras by location of Myc-CalDAG-GEFI (Fig. 3). Prolonged exposure to RasGRP (27). Importantly, in control experiments using co- TPA did not alter the subcellular distribution of RasGRP2 transfected mSos1, Ha-Ras was efficiently loaded with GTP (to (data not shown). 21 6 1.65%, n 5 3) showing that Ha-Ras is competent for RasGRP2 Selectively Activates N- and Ki-Ras but Not Ha- nucleotide exchange in this assay. In contrast, Myc-CalDAG- Ras—We next investigated the ability of RasGRP2 to catalyze GEFI was unable to stimulate GTP loading of any Ras isoform, Ras nucleotide exchange in intact cells. COS cells were trans- consistent with the findings of a previous study (30). These fected with expression plasmids for wild type N-, Ki-, and results indicate that plasma membrane localization may be Ha-Ras, alone and in combination with RasGRP2-Flag or Myc- critical for the RasGEF activity of RasGRP2 in vivo. CalDAG-GEFI. The cells were metabolically labeled with The N-Ras exchange assays were then repeated, but cells [ P]orthophosphate and the relative amounts of GTP and GDP were treated with TPA, A23187, or TPA1A23187 for 2 or 15 captured in anti-Ras immunoprecipitates determined by thin min immediately prior to harvesting. Fig. 4 (B and C) shows layer chromatography (TLC) and phosphorimaging. Fig. 4A that TPA enhanced and A23187 partially inhibited the GTP shows that in the conditions of this assay RasGRP2-Flag stim- loading of N-Ras by RasGRP2-Flag. These opposing effects ulated GTP loading of N-Ras and Ki-Ras but not Ha-Ras. The were additive on RasGRP2 exchange activity. The duration of -fold increase in GTP loading of N-Ras by RasGRP2 is of a exposure to TPA did not affect the degree of stimulation of 32264 A Novel Ras Exchange Factor FIG.5. CalDAG-GEFI functions as a RasGEF in vitro. A, GST- N-Ras was preloaded with [ H]GDP as described (34) and 20 pmol of [ H]GDP-loaded Ras incubated with (M) or without (E) 0.5 mg of re- combinant GST-CalDAG-GEFI for 0, 6, or 20 min at 30 °C in buffer containing 20 mM TrisCl, pH 7.5, 50 mM NaCl, 1 mM MgCl ,1mM dithiothreitol, 1 mM GTP. Reactions were filtered through nitrocellulose FIG.6. Comparison of the RapGEF activity of RasGRP2 and discs and the radioactivity remaining bound to Ras determined by CalDAG-GEFI. A, COS cells expressing HA-Rap1 with and without scintillation counting and expressed as a percentage of that counted in RasGRP2-Flag and Myc-CalDAG-GEFI were treated with TPA (100 nM) the t 5 0 sample (34). All assay points were carried out in duplicate, and and/or A23187 (10 mM)orMe SO for 2 min prior to harvesting. Cells the data shown are the mean of the values. No exchange activity was were lysed and HA-Rap1-GTP was isolated with GST-RafRBD-K85A- evident when GST-Rac or Rho replaced Ras in the assay (data not Sepharose beads as described under “Experimental Procedures.” A rep- shown). B, GST N-Ras preloaded with [ H]GDP was incubated with resentative immunoblot from one of these experiments is shown in recombinant CalDAG-GEFI as in A but the assay buffer was adjusted to panel A. In each lane, the HA-Rap1GTP captured is shown above an 2mM CaCl (l) or 0.01 mM TPA (l) prior to the addition of recombinant aliquot of the total HA-Rap1 present in the binding reaction (see “Ex- CalDAG-GEFI. These results were replicated in two to three independ- perimental Procedures” for details). Greater quantities of HA-Rap1GTP ent experiments. are present when Myc-CalDAG-GEFI is co-expressed than when HA- Rap1 is expressed by itself or with RasGRP2-Flag. 2-min incubations RasGRP2 activity, but with more prolonged exposure to TPA, with A23187 and/or TPA stimulate the GTP loading of HA-Rap1 by either Myc-CalDAG-GEFI or RasGRP2-Flag. RasGRP2-Flag and Myc- the inhibitory effect of A23187 was less marked. In contrast, CalDAG-GEFI expression was confirmed by immunoblotting using an- Myc-CalDAG-GEFI did not stimulate GTP loading of N-Ras ti-epitope antisera. B, the results from the assay shown in panel A and during a 2-min exposure to TPA, but a 15-min treatment with from three to six similar experiments in which cells were stimulated for TPA did result in potent activation of Myc-CalDAG-GEFI N- 2 or 15 min with TPA or A23187 were quantified. Myc-CalDAG-GEFI is a significantly more potent activator of HA-Rap1 than RasGRP2-Flag. Ras exchange activity (Fig. 4C). As with RasGRP2, this TPA- After 2 min of incubation with the chemical modulators, the GTP mediated stimulation of Myc-CalDAG-GEFI exchange activity loading of HA-Rap1 by Myc-CalDAG-GEFI is up-regulated by TPA and was partially inhibited by A23187. A23187, and this effect is additive when both chemicals are adminis- To further characterize CalDAG-GEFI, we evaluated ex- tered together. A similar profile is seen after 15-min incubations. Ras- GRP2-Flag is a weaker RapGEF than Myc-CalDAG-GEFI, but its ac- change activity toward Ras in vitro. A GST fusion of CalDAG- tivity is still excited by a 2-min stimulation with TPA and to a lesser GEFI was generated and incubated with recombinant N-Ras extent by A23187. 15-min incubations with these modulators have less protein preloaded with radiolabeled GDP. Fig. 5 shows that, in potent effects on RasGRP2 than 2-min incubations. Con, control; the presence of excess cold GTP, CalDAG-GEFI catalyzed re- DMSO,Me SO. lease of GDP from N-Ras whereas minimal loss of label oc- curred under control conditions. Thus, CalDAG-GEFI has gua- RasGEF activity of RasGRP2 and CalDAG-GEFI and yielded nine nucleotide exchange activity toward N-Ras in vitro.To identical results to those obtained in the immunoprecipitation investigate whether this GEF activity was sensitive to regula- assay (data not shown). Fig. 6 shows that CalDAG-GEFI effi- tion by calcium or diacylglycerol, the exchange reactions were ciently catalyzed GTP loading of Rap1 in COS cells and that repeated in the presence of 2 mM calcium or 10 mM TPA. this RapGEF activity was further stimulated by both TPA and Interestingly, calcium completely inhibited CalDAG-GEFI-cat- calcium ionophor treatment of cells for 2 or 15 min prior to alyzed release of GDP from N-Ras, whereas TPA had no effect harvesting. In contrast, RasGRP2 was less potent than (Fig. 5). CalDAG-GEFI in catalyzing GTP loading of Rap1. The While this work was in progress, Kawasaki et al. (30) re- RapGEF activity of RasGRP2 was stimulated by 2 min of ported that CalDAG-GEFI has potent RapGEF activity. We treatment with TPA or calcium ionophor, although these re- therefore compared the RapGEF activity of RasGRP2 with that sponses were less marked after 15 min of stimulation (Fig. 6B). of CalDAG-GEFI. To this end we used a “pull-down” assay (43), RasGRP2 Stimulates the Growth of NIH3T3 Cells—We next in which RapGTP is isolated from cell lysates using a GST examined whether overexpression of RasGRP2 in NIH3T3 cells fusion of the Raf-1 Ras binding domain (RBD) and quantified would result in morphological transformation or accelerated by immunoblotting. The RafRBD used in these experiments growth, as has been shown for membrane-targeted Sos1 (8). We contains a K85A substitution, which increases .20-fold the found that neither RasGRP2-Flag nor Myc-CalDAG-GEFI gave amount of RasGTP or RapGTP that can be recovered from a cell rise to foci in NIH3T3 focus assays. Moreover, in co-transfec- lysate. The pull-down assay was validated by evaluating the tion assays with RasGRP2-Flag and a hygromycin resistance Fridman, M., Maruta, H., Gonez, J., Walker, F., Treutlein, H., Zeng, J., and Burgess, A. (2000) J. Biol. Chem. 275, 30363–30371. A Novel Ras Exchange Factor 32265 FIG.7. RasGRP2 and CalDAG-GEFI facilitate the growth of NIH3T3 cells co-expressing N-Ras. A, growth curves of polyclonal cell populations expressing N-Ras (E), N-Ras 1Myc-CalDAG-GEFI (), N-Ras1RasGRP2-Flag (l), or mock-transfected with empty vector (M). Each point is the mean (6 S.E., n 5 9) from three independent experiments. Cell numbers are expressed relative to the number of cells counted on day 1(51), typically 10 . These curves show growth in 5% serum; essentially identical results were obtained in 1% serum (data not shown). B, growth rates, relative to the vector-transfected control populations (51), derived from the growth curves shown in panel A. C, subcellular fractionation of cells expressing Myc-CalDAG-GEFI following growth in serum (1), 16 h of serum-free growth (2), or 16 h of serum-free growth followed by a 2-min add-back (AB) of 20% serum prior to harvesting. Equal proportions of S100 (S) and P100 (P) fraction were immunoblotted (IB) with anti-Myc antisera. D, COS cells expressing N-Ras with and without RasGRP2-Flag and Myc-CalDAG-GEFI were serum-starved for 16 h and then harvested directly (2BCS) or treated for 2 min with 20% bovine calf serum (1BCS) prior to harvesting. Con, control. RasGTP levels were measured in a GST-RafRBD-K85A pull-down assay. The graph shows a representative experiment that was repeated two further times. The presence of serum potently stimulates CalDAG-GEFI RasGEF activity toward N-Ras. Expression of CalDAG-GEFI and RasGRP2-Flag was confirmed by immuno- blotting with anti-Myc or anti-Flag antisera, respectively. vector, hygromycin-resistant colonies appeared no different in conditions of the growth rate assays in Fig. 7 (A and B), carried morphology than vector transfected controls, despite clear ex- out in 5% serum, a substantial amount of Myc-CalDAG-GEFI pression of RasGRP2-Flag by Western blotting (data not becomes membrane-associated (Fig. 7C). In the light of this shown). Similarly, assays of growth in reduced serum (5%, 1%, observation, the serum response of Myc-CalDAG-GEFI was and 0.1%) revealed no differences in growth rates between evaluated in COS cells. Fig. 7D shows that serum is a potent polyclonal cell populations expressing RasGRP2-Flag or Myc- activator of the RasGEF activity of CalDAG-GEFI toward N- CalDAG-GEFI over vector-transfected controls (data not Ras but has minimal effect on the RasGEF activity of shown). Given the Ras isoform selectivity of RasGRP2, we RasGRP2-Flag. repeated these experiments in NIH3T3 cells co-transfected DISCUSSION with a low level of wild type N-Ras plasmid sufficient to give approximately 3-fold overexpression of N-Ras over endogenous In this paper we describe the cloning and characterization of levels. Fig. 7 (A and B) shows that polyclonal cell populations RasGRP2, a longer isoform of CalDAG-GEFI. The basic domain expressing wild type N-Ras had a growth rate approximately structure of RasGRP2/CalDAG-GEFI is similar to that of the twice that of empty vector transfected control cell populations, related protein RasGRP/CalDAG-GEFII, which has been de- but that co-expression of RasGRP2-Flag increased this growth scribed previously. One structural feature not commented on in rate another 3-fold. Somewhat unexpectedly, co-expression of the original description of CalDAG-GEFI is the presence of a Myc-CalDAG-GEFI in the N-Ras-expressing cells also resulted short coiled-coil (CC) domain located between the Ras exchange in an increment in growth rate (Fig. 7, A and B). A possible motif and SCR1 box. This is not present in RasGRP, which explanation of this result is shown in Fig. 7 (C and D). The instead has a more extensive CC domain in a C-terminal ex- biochemical assays of RasGRP2 function all required a period tension that is absent from CalDAG-GEFI. The role of the CC of serum starvation; therefore, the cell fractionation and im- domains in RasGRP and RasGRP2 is unknown. In RasGRF1 a munofluorescence studies in Figs. 2 and 3 were performed CC domain co-operates with IQ and pleckstrin homology do- under these conditions. However, Fig. 7C shows that continu- mains for maximal stimulation of exchange activity by calmod- ous growth in serum, or a 2-min treatment with 20% serum ulin-complexed calcium (23), but since the RasGRPs lack these after a period of serum-free growth, significantly increased the other domains and RasGRP2 activity is inhibited by calcium, amount of Myc-CalDAG-GEFI associated with the P100 frac- the CC domain must have some other role. tion. In contrast, there was no effect of serum on the subcellular The novel N-terminal sequences present in RasGRP2 con- distribution of RasGRP2-Flag (data not shown). Thus, in the tain acylation sites that we show here are both N-myristoylated 32266 A Novel Ras Exchange Factor and palmitoylated. The most likely site for palmitoylation is RasGRP; RasGRP is translocated within 2 min, whereas 15 Cys-7, since this is the only cysteine residue in the N-terminal min of TPA treatment is needed to bring CalDAG-GEFI to the 70 amino acids of RasGRP2. We cannot exclude a more remote cell membrane. Taking these data together, we propose that, palmitoylation site, but palmitoylation of other N-myristoy- although RasGRP2 is localized at the plasma membrane, it may in part be sequestered away from Ras until interactions lated proteins is usually close to the N terminus (40). The results of the Triton X-114 assay indicate that only 60% of between TPA and the DGB domain relocate RasGRP2 to a different plasma membrane microdomain. The mechanism for RasGRP2-Flag expressed in COS cells undergoes lipid modifi- the effects of TPA could then be the same for RasGRP, cation. Since co-translational myristoylation precedes post- CalDAG-GEFI, and RasGRP2, namely to co-localize the Ras- translational palmitoylation, it is likely that this result reflects GEF and Ras substrate in the same plasma membrane mi- incomplete myristoylation of RasGRP2. We have observed pre- crodomain. In support of this model, it is interesting to note viously that myristoylation of proteins overexpressed in COS that the GEF activity of RasGRP was highest when it was cells can be incomplete (37); this, coupled with the slightly membrane-targeted using a Ras localization signal (28), a atypical myristoylation motif, probably accounts for the incom- strategy that should exactly co-localize Ras and RasGRP in the plete acylation of RasGRP2. The N-terminal sequence of Ras- same microdomain. An alternative explanation that the TPA GRP also contains a potential N-myristoylation site, but it has stimulates RasGRP2 via an allosteric effect seems unlikely, not yet been determined whether RasGRP is acylated (28). since stimulation was not evident in vitro even with high con- RasGRP2 is localized extensively to the plasma membrane in centrations of TPA. Serum stimulation caused a much more contrast to CalDAG-GEFI, which is almost exclusively cytosolic rapid membrane translocation and activation of CalDAG-GEFI in quiescent cells. These data clearly implicate the N-terminal than TPA, perhaps reflecting selective generation of diacylg- 60 amino acids of full-length RasGRP2 as containing a plasma lycerol in physiologically relevant plasma membrane microdo- membrane localization motif. Since the combination of myris- mains, rather than general loading of cell membranes with toylation and palmitoylation is well established as a targeting DGB binding sites. signal (44), we propose that dual acylation targets RasGRP2 to RasGRP2 selectively activates N-Ras and Ki-Ras, but not the plasma membrane. Ha-Ras. In contrast, RasGRP activates Ha-Ras and mRasGRP CalDAG-GEFI has been reported to be a RapGEF with no when co-transfected with N-, K-, or Ha-Ras results in mitogen- RasGEF activity (30). We show here, however, that CalDAG- activated protein kinase activation, consistent with activation GEFI is indeed a RasGEF, albeit with an isoform specificity of all three isoforms (27, 28). One explanation of these data is that is restricted to N- and Ki-Ras. Ha-Ras was the only Ras simply that sequence differences between RasGRP and Ras- isoform tested by Kawasaki et al. (30), and we are in full GRP2 result in different substrate specificities. An alternative agreement with the conclusions of their study that CalDAG- hypothesis invokes recent work showing that Ras isoforms GEFI cannot activate this Ras isoform. More intriguingly, operate in functionally distinct microdomains of the plasma CalDAG-GEFI and RasGRP2 have different basal exchange membrane (45); signaling by Ha-Ras, anchored to the plasma activities and show distinct in vivo responses to stimulation by membrane by palmitoylation and farnesylation, is dependent TPA and calcium. CalDAG-GEFI has potent RapGEF activity, on the integrity of cholesterol-rich microdomains, whereas sig- which is stimulated both by acute elevations of cytosolic cal- naling by Ki-Ras, anchored by a charged polybasic domain and cium and exposure to TPA, but it has almost no N-Ras GEF farnesylation, does not require lipid rafts. In this context, the activity until subject to prolonged stimulation by TPA or acute different Ras isoform specificity of RasGRP2 and RasGRP may stimulation by serum. The TPA-stimulated RasGEF activity of be explained by postulating that RasGRP2 gains access only to CalDAG-GEFI is inhibited by elevations of cytosolic calcium. In the N- and Ki-Ras microdomains because N-terminal palmitoy- contrast, RasGRP2 is a less potent RapGEF but a more potent lation imposes a restraint on RasGRP2 mobility within the RasGEF than CalDAG-GEFI. The RapGEF activity of Ras- plasma membrane that is not experienced by RasGRP. GRP2 is stimulated by TPA and calcium, whereas the RasGEF Somewhat surprisingly, moderate overexpression of plasma activity is stimulated by TPA, but inhibited by calcium. membrane-localized RasGRP2 did not morphologically trans- The activation of RasGEF activity in CalDAG-GEFI by TPA form NIH3T3 cells, although we were able to show an N-Ras- or serum stimulation correlates temporally with translocation dependent stimulation of cell growth by both RasGRP2 and of CalDAG-GEFI to the plasma membrane. We propose there- CalDAG-GEFI. In fibroblasts, activated Rap can partially an- fore that this major change in subcellular localization is fun- tagonize Ras transforming activity (46). It is therefore tempt- damental to Ras activation by CalDAG-GEFI, analogous to the ing to speculate that the lack of transforming ability of Ras- mechanisms described for Sos1 and RasGRF2. Similarly, there GRP2 and CalDAG-GEFI may reflect their co-activation of Ras is a significant translocation of RasGRP to the plasma mem- and Rap. In this context it is worth noting that, since calcium brane in response to TPA treatment, concomitant with the inhibits the RasGEF activity of RasGRP2/CalDAG-GEFI while stimulation of RasGEF activity (27). RasGRP2 has constitutive stimulating the RapGEF activity, a calcium signal effectively RasGEF activity, presumably reflecting constitutive plasma shifts the GEF activity of RasGRP2 and CalDAG-GEFI from membrane localization. However, in contrast to CalDAG-GEFI Ras to Rap. Together, these observations therefore reveal an and RasGRP, there is no detectable change in the subcellular interesting and novel molecular mechanism for utilizing a sin- distribution of RasGRP2 to correspond with TPA stimulation of gle exchange factor to co-ordinate Ras and Rap signaling RasGEF activity. events. Is there a common mechanism underlying these TPA effects? Acknowledgments—We thank Masha Fridman and Tony Burgess for The DGB domain of RasGRP is clearly responsible for TPA- the GST-RafRBD-K85A construct and for sharing results prior to pub- mediated membrane translocation because deletion of the do- lication, Marilyn Resh for helpful advice, and David Bowtell for the main blocks RasGRP membrane binding (27) and the isolated mSos1 cDNA. DGB domain can target heterologous proteins to the plasma REFERENCES membrane in a TPA-dependent manner (28). The DGB domain 1. Boriack-Sjordan, P. A., Margarit, S. M., Bar-Sagi, D., and Kuriyan, J. (1998) of CalDAG-GEFI can effect a similar change in subcellular Nature 394, 337–343 localization, since TPA treatment results in membrane trans- 2. Wittinghofer, F. (1998) Nature 394, 317–320 location of CalDAG-GEFI, albeit with different kinetics from 3. Bos, J. L. (1997) Biochim. Biophys. Acta 1333, M19-M31 A Novel Ras Exchange Factor 32267 4. Simon, M. A., Bowtell, D. D. L., Dodson, G. S., Laverty, T. R., and Rubin, G. M. 25. Cerione, R. A., and Zheng, Y. (1996) Curr. Opin. Cell Biol. 8, 216 –222 (1991) Cell 67, 701–716 26. Nimnual, A. S., Yatsula, B. A., and Bar Sagi, D. (1998) Science 279, 560 –563 5. Li, N., Batzer, A., Daly, R., Yajnik, V., Skolnik, E., Chardin, P., Bar Sagi, D., 27. Ebinu, J. O., Bottorff, D. A., Chan, E. Y. W., Stang, S. L., R. J., D., and Stone, Margolis, B., and Schlessinger, J. (1993) Nature 363, 85– 88 J. C. (1998) Science 280, 1082–1086 6. Chardin, P., J. H., C., Gale, N. W., Van Aelst, L., Schlessinger, J., Wigler, 28. Tognon, C. E., Kirk, H. E., Passmore, L. A., Whitehead, I. P., Der, C. J., and M. H., and Bar-Sagi, D. (1993) Science 260, 1338 –1343 Kay, R. J. (1998) Mol. Cell. Biol. 18, 6995–7008 7. Buday, L., and Downward, J. (1993) Cell 73, 611– 620 29. Kedra, D., Serossi, E., Fransson, I., Trifunovic, J., Clark, M., Lagercrantz, J., 8. Aronheim, A., Engelberg, D., Li, N., al Alawi, N., Schlessinger, J., and Karin, Blennow, E., Mehlin, H., and Dumanski, J. (1997) Hum. Genet. 100, M. (1994) Cell 78, 949 –961 611– 619 9. Ravichandran, K. S., Lorenz, U., Shoelson, S. E., and Burakoff, S. J. (1995) 30. Kawasaki, H., Springett, G. M., Toki, S., Canales, J. J., Harlan, P., Blumen- Mol. Cell. Biol. 15, 593– 600 stiel, J. P., Chen, E. J., Bany, I. A., Mochizuki, N., Ashbacher, A., Matsuda, 10. Lioubin, M. N., Myles, G. M., Carlberg, K., Bowtell, D., and Rohrschneider, M., Housman, D. E., and Graybiel, A. M. (1998) Proc. Natl. Acad. Sci. L. R. (1994) Mol. Cell. Biol. 14, 5682–5691 U. S. A. 95, 13278 –13283 11. Harmer, S. L., and DeFranco, A. L. (1997) Mol. Cell. Biol. 17, 4087– 4095 31. Bergman, L., Silins, G., Grimmond, S., Hummerich, H., Stewart, C., Little, P., 12. Kouhara, H., Hadari, Y. R., Spivak-Kroizman, Schilling, J., Bar-Sagi, D., Lax, and Hayward, N. K. (1999) Genomics 55, 49 –56 I., and Schlessinger, J. (1997) Cell 89, 693–702 32. Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990) J. 13. Lutterell, L., Ferguson, S., Daaka, Y., Miller, W., Maudsley, S., Della Roca, G., Mol. Biol. 215, 403– 410 Lin, F.-T., Kawakatsu, H., Owada, K., Luttrell, D., Caron, M., and Lefkow- 33. Evan, G., Lewis, G., Ramsey, G., and Bishop, J. (1985) Mol. Cell. Biol. 5, itz, R. (1999) Science 283, 655– 661 3610 –3615 14. Chen, Y., Grall, D., Salcini, A., Pelicci, P., Poussegur, J., and Van Obberghen- 34. Porfiri, E., Evans, T., Chardin, P., and Hancock, J. F. (1994) J. Biol. Chem. Schilling, E. (1996) EMBO J. 15, 1037–1044 269, 22672–22677 15. Chen, R., Corbalan-Garcia, S., and Bar-Sagi, D. (1997) EMBO J. 16, 35. Downward, J., Graves, J. D., Warne, P. H., Rayter, S., and Cantrell, D. A. 1351–1359 (1990) Nature 346, 719 –723 16. Qian, X., Vass, W., Papageorge, A., Anborgh, P., and Lowy, D. (1998) Mol. Cell. 36. Roy, S., McPherson, R. A., Apolloni, A., Yan, J., Clyde-Smith, J., and Hancock, Biol. 18, 771–778 J. F. (1998) Mol. Cell. Biol. 18, 3947–3955 17. Wang, W., Fisher, E. M., Jia, Q., Dunn, J. M., Porfiri, E., Downward, J., and 37. Cadwallader, K., Paterson, H., Macdonald, S. G., and Hancock, J. F. (1994) Egan, S. E. (1995) Nat. Genet. 10, 294 –300 Mol. Cell. Biol. 14, 4722– 4730 18. Byrne, J. L., Paterson, H. F., and Marshall, C. J. (1996) Oncogene 13, 38. Roy, S., Lane, A., Yan, J., McPherson, R., and Hancock, J. F. (1997) J. Biol. 2055–2065 Chem. 272, 20139 –20145 19. Shou, C., Farnsworth, C. L., Neel, B. G., and Feig, L. A. (1992) Nature 358, 39. Huang, D. C. S., Marshall, C. J., and Hancock, J. F. (1993) Mol. Cell. Biol. 13, 351–354 2420 –2431 20. Fam, N. P., Fan, W. T., Wang, Z., Zhang, L. J., Chen, H., and Moran, M. F. 40. Wedegaertner, P. B., Wilson, P. T., and Bourne, H. R. (1995) J. Biol. Chem. (1997) Mol. Cell. Biol. 17, 1396 –1406 270, 503–506 21. Sturani, E., Abbondio, A., Branduardi, P., Ferrari, C., Zippel, R., Martegani, 41. Gutierrez, L., Magee, A. I., Marshall, C. J., and Hancock, J. F. (1989) EMBO E., Vanoni, M., and Denis Donini, S. (1997) Exp. Cell Res. 235, 117–123 J. 8, 1093–1098 22. Brambilla, R., Gnesutta, N., Minichiello, L., White, G., Roylance, A. J., Herron, 42. Resh, M. (1994) Cell 76, 411– 413 C. E., Ramsey, M., Wolfer, D. P., Cestari, V., Rossi Arnaud, C., Grant, S. G., 43. de Rooij, J., and Bos, J. L. (1997) Oncogene 14, 623– 625 Chapman, P. F., Lipp, H. P., Sturani, E., and Klein, R. (1997) Nature 390, 44. Resh, M. D. (1996) Cell Signal 8, 403– 412 281–286 23. Buchsbaum, R., Telliez, J. B., Goonesekera, S., and Feig, L. A. (1996) Mol. Cell. 45. Roy, S., Luetterforst, R., Harding, A., Apolloni, A., Etheridge, M., Stang, E., Rolls, B., Hancock, J. F., and Parton, R. G. (1999) Nat. Cell Biol. 1, 98 –105 Biol. 16, 4888 – 4896 24. Farnsworth, C. L., Freshney, N. W., Rosen, L. B., Ghosh, A., Greenberg, M. E., 46. Kitayama, H., Sugimoto, Y., Matsuzaki, T., Ikawa, Y., and Noda, M. (1989) and Feig, L. A. (1995) Nature 376, 524 –527 Cell 56, 77– 84

Journal

Journal of Biological Chemistry – Unpaywall

Published: Oct 1, 2000

There are no references for this article.