Identification and Characterization of RA-GEF-2, a Rap Guanine Nucleotide Exchange Factor That Serves as a Downstream Target of M-Ras*

The Ras family small GTPase Rap is regulated by an array of specific guanine nucleotide exchange factors (GEFs) in response to upstream stimuli. RA-GEF-1 was identified as a novel Rap GEF, which possesses a Ras/ Rap1-associating (RA) domain. Here we report a protein closely related to RA-GEF-1, named RA-GEF-2. Like RA-GEF-1, a putative cyclic nucleotide monophosphate-binding domain, a Ras exchanger motif, a PSD-95/DlgA/ ZO-1 domain, and an RA domain in addition to the GEF catalytic domain are found in RA-GEF-2. However, RA-GEF-2 displays a different tissue distribution profile from that of RA-GEF-1. RA-GEF-2 stimulates guanine nucleotide exchange of both Rap1 and Rap2, but not Ha-Ras. The RA domain of RA-GEF-2 binds to M-Ras in a GTP-dependent manner, but not to other Ras family GTPases tested, including Ha-Ras, N-Ras, Rap1A, Rap2A, R-Ras, RalA, Rin, Rit, and Rheb, in contrast to the RA domain of

The Ras family small GTPase Rap1 participates in the regulation of a wide variety of cellular responses, including proliferation, differentiation, lymphocyte aggregation, T-cell anergy, and platelet activation (1,2). In contrast, the physiological function of Rap2, a close relative of Rap1, remains largely unknown. Diverse effectors for Rap1, almost all of which are common to Ras effectors, have been identified, although their roles in individual signaling pathways remain obscure. For activation of the effectors, the interaction of the effector region of Rap1 (amino acids [32][33][34][35][36][37][38][39][40] with the Ras/Rap1-binding or RA 1 domain of the effectors is important (3). In addition, the second Ras/Rap1-binding site identified in several Ras/Rap1 effectors, including Raf-1, B-Raf, and yeast adenylyl cyclase, is required for proper effector activation (4 -7). Suppression of Ki-Rasinduced transformation by overexpressed Rap1 is thought to be ascribed to tight binding to the second Ras/Rap1-binding sites of Ras effectors such as Raf-1 without stimulating their activities (5,6).
Ras family GTPases cycle between GTP-bound active and GDP-bound inactive states, serving as a molecular switch of intracellular signaling (8,9). Conversion between GTP-and GDP-bound states is controlled by GEFs and GTPase-activating proteins (8,9). Particularly, GEFs enhance the formation of the GTP-bound active conformation in response to upstream signals mediated by various cell surface receptors. To date, various GEFs for Rap1 have been identified in mammalian cells. C3G binds to the adaptor protein Crk, being involved in tyrosine kinase-dependent activation of Rap1 (10). Epac/cAMP-GEF is activated through direct association with cAMP, thereby stimulating Rap-dependent signaling (11,12). Another Rap GEF, CalDAGGEF1, which contains calcium-and diacylglycerol-binding motifs, has a role in Rap activation in response to these second messengers (13). Additionally, we and other groups recently identified a novel type of the Rap GEF, RA-GEF-1 (also termed PDZ-GEF1, nRapGEP, or CNrasGEF), which exhibits GEF activity toward Rap1 and Rap2, but not Ha-Ras (14 -18). RA-GEF-1 contains putative cNMP-binding, REM, PDZ, and RA domains as well as the GEF catalytic domain. We and others detected no specific cAMP/cGMP binding to the cNMP-binding domain (14 -16, 18), although Pham et al. (17) reported cAMP binding to this domain and subsequent stimulation of Ras GEF activity. The RA domain of RA-GEF-1 binds to Rap1⅐GTP, suggesting that RA-GEF-1 plays an important role downstream of Rap1 as well (14). Indeed, the RA domain is required for translocation of RA-GEF-1 to the perinuclear compartments including the Golgi complex and for the full activation of Rap1, as evidenced by our recent obser-vation that an RA domain mutation that abolishes Rap1 binding compromised RA-GEF-1-dependent activation of Rap1 in vivo (18).
Herein, we describe a novel member of the RA-GEF family, designated RA-GEF-2, whose structural features are intimately related to RA-GEF-1. Like RA-GEF-1, RA-GEF-2 exhibits GEF activity toward Rap1 and Rap2, but not Ha-Ras. However, unlike RA-GEF-1, the RA domain of RA-GEF-2 specifically binds to M-Ras⅐GTP, which causes the translocation of RA-GEF-2 to the plasma membrane, where M-Ras exists. Correspondingly, activation of Rap1 that localized in the plasma membrane was observed following coexpression of RA-GEF-2 and activated M-Ras.

Cloning of RA-GEF-2 and Construction of Mammalian Expression
Plasmids-A genomic clone (AC004227) that encodes an open reading frame whose predicted amino acid sequence is highly homologous to RA-GEF-1 was identified through a search of the GenBank data base with the BLAST program. The predicted partial coding sequence in AC004227 was amplified from a cDNA library synthesized from human fetal brain mRNA (Invitrogen) using two gene-specific primers. Upstream and downstream sequences were obtained by 5Ј-and 3Ј-RACE using the same library by a Marathon cDNA amplification procedure (CLONTECH). The complete nucleotide sequence was confirmed by isolating and sequencing multiple clones, and the encoded protein was designated RA-GEF-2. The full-length RA-GEF-2 cDNA was subcloned into the mammalian expression vectors pFLAG-CMV2 (Sigma) and pcDNA3.1HisB (Invitrogen), generating pFLAG-CMV2-RA-GEF-2 and pcDNA3.1HisB-RA-GEF-2, respectively. The cDNA for RA-GEF-2⌬RA, RA-GEF-2 lacking the N-terminal portion of the RA domain (amino acids 749 -779), was constructed by the polymerase chain reaction and subcloned into pFLAG-CMV2, generating pFLAG-CMV2-GEF-2⌬RA. cDNAs for EGFP-tagged RA-GEF-2 and RA-GEF-2⌬RA were constructed by the polymerase chain reaction and subcloned into the mammalian expression vector pCMV2, generating pCMV2-EGFP-GEF-2 and pCMV2-EGFP-RA-GEF-2⌬RA, respectively. cDNAs for wild-type and activated M-Ras were subcloned into pFLAG-CMV2, generating pFLAG-CMV2-M-Ras and pFLAG-CMV2-M-Ras 71L , respectively. The cDNA for HA-tagged M-Ras 71L was subcloned into the mammalian expression vector pEF-BOS (19), generating pEF-BOS-HA-M-Ras 71L . The cDNA for R-Ras was kindly provided by Shintaro Iwashita (Mitsubishi Kasei Institute of Life Sciences, Tokyo, Japan).
Cell Culture and Transfection-COS-7 cells were cultured in DMEM supplemented with 10% fetal calf serum. Expression plasmids were introduced into COS-7 cells by using GenePORTER (Gene Therapy System) or Superfect (Qiagen) according to the manufacturer's protocol.
Effect of M-Ras on GEF Activity in Vitro-Post-translationally modified FLAG-tagged M-Ras was purified from COS-7 cells as follows. COS-7 cells transfected with pFLAG-CMV2-M-Ras were sonicated in sonication buffer (20 mM Tris-HCl, pH 7.4, 100 mM NaCl, 2.5 mM MgCl 2 , 1 mM EDTA, 10 g/ml aprotinin, 1 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride) and centrifuged at 100,000 ϫ g for 1 h. The pellet was resuspended and incubated for 1 h on ice in sonication buffer containing 0.2% Triton X-100. The 15,000 ϫ g supernatant was subjected to affinity purification of FLAG-M-Ras using anti-FLAG M2 resin. For examining the effect of M-Ras on GEF activity of RA-GEF-2 toward Rap1A in vitro, 6ϫHis-M-Ras (purified from E. coli) or FLAG-M-Ras (purified from COS-7 cells) was preloaded with GTP␥S and incubated at 4°C for 2 h with full-length FLAG-RA-GEF-2 purified from Sf9 cells. Subsequently, the mixture was added to 4 pmol of [ 3 H]GDP-preloaded Rap1A at a total volume of 200 l, and GEF assays were performed as described above.
In Vitro Association Assay-Polypeptides corresponding to the RA domains and their flanking regions of RA-GEF-2 (amino acids 683-853) and RA-GEF-2⌬RA (amino acids 683-822) were expressed as MBP fusion proteins (MBP-RA WT and MBP-RA MUT , respectively) in E. coli by using pMal-c2 (New England Biolabs). The interaction of the RA domain of RA-GEF-2 with Ras family small GTPases was assessed essentially as described previously (14). Small GTPases (6 pmol) preloaded with GTP␥S or GDP were incubated at 4°C for 2 h with MBP-RA WT or MBP-RA MUT (50 pmol) immobilized on amylose resin in 100 l of binding buffer (14). After extensive washing with binding buffer, bound proteins were eluted from the resin with 10 mM maltose and subjected to SDS-polyacrylamide gel electrophoresis followed by immunoblotting. Ha-Ras and Rap1A were detected with anti-Ha-Ras antibody (F235, Oncogene Science Inc.) and anti-Rap1A antibody (sc-65, Santa Cruz Biotechnology), respectively. Other GTPases were detected with anti-6ϫHis antibody (CLONTECH). For quantitative in vitro association assays, 6ϫHis-M-Ras was loaded with [ 35 S]GTP␥S (3, 500 cpm/pmol) or [ 3 H]GDP (1, 100 cpm/pmol) and incubated with MBP-RA WT (25 pmol) as described above except that unlabeled GTP␥S or GDP (0.1 mM), respectively, was included in the binding reaction. Radioactivities eluted from amylose resin with 10 mM maltose were counted.
Subcellular Fractionation-COS-7 cells cultivated in 20 culture plates (100-mm diameter) were harvested by centrifugation (600 ϫ g for 5 min) and washed with phosphate-buffered saline. Preparation of the plasma membrane fraction was performed essentially as described (21). Cell pellets were resuspended in 0.25 M STM (0.25 M sucrose, 5 mM Tris-HCl, pH 7.4, 1 mM MgCl 2 , 10 g/ml aprotinin, 1 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride), homogenized using a Potter-Elvehjem homogenizer, and centrifuged at 280 ϫ g for 5 min. The supernatant was further centrifuged at 1,500 ϫ g for 10 min, and the pellet was resuspended in 0.25 M STM. 2 M STM (2 M sucrose, 5 mM Tris-HCl, pH 7.4, 1 mM MgCl 2 ) was added to adjust the sucrose concentration to 1.42 M. The suspension was transferred to a centrifugation tube and overlaid with 0.25 M STM. After centrifugation (82,000 ϫ g for 1 h), the pellicle at the 0.25 M/1.42 M interface was collected, resuspended in 0.25 M STM, and centrifuged at 1,500 ϫ g for 10 min. The final pellet (the plasma membrane fraction) was resuspended in 150 l of lysis buffer and subjected to the pull-down assay. Alkaline phosphodiesterase (a marker enzyme for the plasma membrane) and mannosidase II (a marker enzyme for the Golgi apparatus) activities were determined as described by Storrie and Madden (22) to estimate the purity of the fractions.

RESULTS
Cloning of RA-GEF-2-Through a BLAST search of the Gen-Bank data base, we identified a genomic clone (AC004227) encoding an open reading frame closely related to RA-GEF-1, designated RA-GEF-2. Two gene-specific primers designed on the basis of the AC004227 sequence were employed to isolate a partial cDNA clone encoding RA-GEF-2. Subsequently, the full-length cDNA was reconstructed from 5Ј-and 3Ј-RACE products obtained from a human fetal brain cDNA library. The putative start codon matched the Kozak consensus sequence and was preceded by in-frame stop codons. The full-length open reading frame was composed of 4,530 nucleotides, encoding a protein of 1,509 amino acids, and was divided into four contiguous genomic clones (AC005218, AC005576, AC004227, and AC004622 from the N terminus to the C terminus) derived from chromosome 5. RA-GEF-1 and RA-GEF-2 were highly homologous in five domains identified by ISREC ProfileScan (76.2, 80.0, 71.4, 81.6, and 85.9% identity in cNMP-binding, REM, PDZ, RA, and GEF domains, respectively), whereas both N-terminal and C-terminal regions were rather divergent (Fig.  1). Three regions (named structurally conserved regions 1-3), which are highly conserved among diverse Ras/Rap GEFs, were also identified in the GEF domain of RA-GEF-2 (Fig. 1C). Compared with RA-GEF-1, RA-GEF-2 has 145-amino acid extension at its N terminus, but lacks the C-terminal 102 amino acid residues. In particular, the C-terminal region was least homologous (47.5% identity).
Tissue Distribution of RA-GEF-1 and RA-GEF-2-We next examined mRNA levels for RA-GEF-1 and RA-GEF-2 in various human tissues using a multiple tissue blot membrane (Fig.  2). Probes derived from divergent regions were used under stringent hybridization conditions to compare the distribution of the transcripts. A single band of ϳ8 kilobases was detected for both probes. The RA-GEF-2 transcript was abundant in heart, brain, placenta, lung, and liver, but barely detectable in skeletal muscle, kidney, and pancreas, whereas RA-GEF-1 was expressed in all tissues tested.
GEF Activity of RA-GEF-2-A FLAG-tagged RA-GEF-2 fragment (amino acids 403-1276) containing the GEF domain was Asterisks and plus signs above the aligned residues represent identities and similarities, respectively, between RA-GEF-2 and RA-GEF-1, whereas those below the aligned residues represent identities and similarities, respectively, among all the three proteins. C, sequence alignment of GEF domains of RA-GEF-2, RA-GEF-1 and Epac. SCR, the structurally conserved region. Asterisks and plus signs represent identities and similarities, respectively, as described in B.
expressed in Sf9 cells and purified to near homogeneity. By using this protein, in vitro GEF activities toward various Ras family proteins were examined. RA-GEF-2 showed GEF activity toward Rap1A and Rap2A, but not toward other Ras family members including Ha-Ras, N-Ras, M-Ras, R-Ras, RalA, Rin, Rit, and Rheb, as determined by [ 3 H]GDP-releasing assays (Fig. 3A). [ 35 S]GTP␥S-binding assays were also performed for Rap1A and Rap2A, revealing that GEF activity toward Rap2A was higher than that toward Rap1A (Fig. 3B). Next, in vivo GEF activities of RA-GEF-2 toward Rap1A and Ha-Ras were examined by pull-down assays (Fig. 3C). Rap1A⅐GTP and Ha-Ras⅐GTP were affinity-precipitated by GST-RalGDS-RID and GST-Raf-1-RBD, respectively, from COS-7 cells that expressed increasing amounts of RA-GEF-2 and either Rap1A or Ha-Ras, and then were detected by immunoblotting. In parallel with the in vitro activity, RA-GEF-2 caused the accumulation of Rap1A⅐GTP, but not Ha-Ras⅐GTP. RA-GEF-2-dependent increase in the level of Rap1A⅐GTP, but not Ha-Ras⅐GTP, was also shown by measuring the ratio of Rap1A-or Ha-Ras-bound GDP and GTP (Fig. 3D). Collectively, RA-GEF-2, like RA-GEF-1, acts as a specific GEF for Rap1 and Rap2.
Association of M-Ras with RA-GEF-2 through the RA Domain-The RA domain of RA-GEF-1 binds to Rap1A in a GTPdependent manner (14), suggesting that the RA domain of RA-GEF-2 may also bind to Rap1A⅐GTP. To test this possibility, we employed in vitro association assays using an MBP fusion construct, named MBP-RA WT , consisting of the RA domain and its flanking region of RA-GEF-2. Unexpectedly, MBP-RA WT bound to Rap1A only weakly compared with the binding of the RA-GEF-1 RA domain, and this binding was GTP-independent (Fig. 4A). Thus, we next examined the binding activity of MBP-RA WT to a variety of Ras family proteins. Among Ras family members tested, only M-Ras associated with MBP-RA WT in a GTP-dependent manner, and virtually no binding was detected for other Ras family proteins, such as Ha-Ras, N-Ras, Rap2A, R-Ras, RalA, Rin, Rit, and Rheb (Fig.  4A). Quantitative analyses further confirmed that the association between MBP-RA WT and M-Ras was dose-and GTP-dependent (Fig. 4B). Additionally, the binding of M-Ras⅐GTP was abolished when a 31-amino acid deletion was introduced within the RA domain, suggesting that these highly conserved residues are crucial (Fig. 4C).
Plasma Membrane Translocation of RA-GEF-2 through the Binding to M-Ras-RA-GEF-1 is translocated to the perinuclear compartments, including the Golgi apparatus, upon interaction with Rap1A⅐GTP (18). Thereafter, RA-GEF-1 serves as an amplifier of Rap1A signaling by yielding the GTP-bound form of Rap1A through the action of the GEF domain (18). In view of the observations that the RA domain of RA-GEF-2, unlike that of RA-GEF-1, specifically binds to M-Ras⅐GTP, it is feasible that RA-GEF-2 colocalizes with M-Ras in the cell and plays a role downstream of M-Ras. As an initial step to clarify this point, localization of Rap1A, M-Ras, and RA-GEF-2 was examined by immunofluorescence microscopy. Although Rap1A was localized mainly in the perinuclear region as reported previously (23)(24)(25)(26), small amounts of Rap1A were detectable in the plasma membrane (Fig. 5A). In contrast, M-Ras was localized mainly in the plasma membrane (Fig. 5B). Both RA-GEF-2 and RA-GEF-2⌬RA were distributed to the cytoplasm, when expressed alone in serum-starved cells (Fig. 5B). Upon coexpression of M-Ras 71L , significant amounts of RA-GEF-2 were conveyed to the plasma membrane and colocalized with M-Ras 71L (Fig. 5C). In contrast, RA-GEF-2⌬RA did not colocalize with M-Ras in the plasma membrane, indicating that the interaction mediated by the RA domain is critical. Ha-Ras is also localized in the plasma membrane, but does not bind to the RA domain of RA-GEF-2 (Fig. 4A). As expected, RA-GEF-2 was not translocated to the plasma membrane, when coexpressed with Ha-Ras 12V (Fig. 5E).
Effect of M-Ras Binding on GEF Activity in Vivo and in Vitro-Given that RA-GEF-2 is an effector for M-Ras, M-Ras GTP-bound forms of HA-Rap1A and HA-Ha-Ras were precipitated from total cellular extracts by the use of GST-RalGDS-RID (for Rap1A) and GST-Raf-1-RBD (for Ha-Ras) and detected by immunoblotting using anti-HA antibody. The amounts of HA-Rap1A and HA-Ha-Ras in 1/10 aliquots of the extracts were also monitored by immunoblotting (Input). D, analysis of the Rap1A-or Ha-Ras-bound GDP/GTP ratio. FLAG-Rap1A or FLAG-Ha-Ras was coexpressed with RA-GEF-2. After labeling cells with [ 32 P]orthophosphate, FLAG-Rap1A and FLAG-Ha-Ras were immunoprecipitated with anti-FLAG antibody. Protein-bound GDP and GTP were separated by thin layer chromatography, and radioactivities were quantitated. Ori, the origin of the chromatography. may modulate the activity of RA-GEF-2. To this end, the effect of M-Ras on GEF activity of RA-GEF-2 was assessed. Fulllength recombinant RA-GEF-2 was purified from Sf9 cells, and its GEF activity was measured by [ 3 H]GDP releasing assays in the presence and absence of M-Ras⅐GTP. Full-length RA-GEF-2 exhibited GEF activity toward Rap1A in a dose-dependent manner, which remained totally unaffected following the addition of the GTP-bound form of M-Ras purified from E. coli (Fig.  6A). Considering that post-translational modifications of Ras and Rap1 are required for the activation of their effectors, such as Raf kinases (4 -6) and yeast adenylyl cyclase (7,20), posttranslational modifications of M-Ras may be critical for the regulation of RA-GEF-2. Thus, we next examined the effect of post-translationally modified M-Ras purified from COS-7 cells. The addition of post-translationally modified M-Ras also showed no significant effect on GEF activity in vitro (Fig. 6B).
To further analyze the role of M-Ras, RA-GEF-2-dependent increase in the Rap1A⅐GTP level was examined by pull-down assays following coexpression of M-Ras 71L . We measured the level of Rap1A⅐GTP in the plasma membrane fraction in addition to that in the total cellular extract because a subset of RA-GEF-2 translocated to the plasma membrane when coexpressed with M-Ras 71L (Fig. 5C). The purity of the plasma membrane fraction was assessed by measuring activities of alkaline phosphodiesterase (a plasma membrane marker) and mannosidase II (a Golgi marker). The specific activity of alkaline phosphodiesterase was increased by 6.2-fold in the plasma membrane fraction compared with that in the total cellular extract, whereas that of mannosidase II was decreased by 2.5-fold. When precipitated from the total cellular extracts, the Rap1A⅐GTP level was slightly decreased upon coexpression of M-Ras 71L , whereas RA-GEF-2⌬RA-dependent Rap1A⅐GTP formation was unaffected by M-Ras 71L (Fig. 6C). In marked con-trast, RA-GEF-2-dependent increase in the level of Rap1A⅐GTP in the plasma membrane fraction was enhanced when M-Ras 71L was coexpressed (Fig. 6D). Considering that the plasma membrane fraction used in our assays may include some contamination of the Golgi fraction, Rap1A⅐GTP formation in the plasma membrane upon coexpression of M-Ras 71L may actually be more remarkable. Taken together, RA-GEF-2 translocates to the plasma membrane through the interaction with M-Ras⅐GTP, thereby serving as a GEF specific to plasma membrane-located Rap1A.

DISCUSSION
The cDNA encoding a novel Rap1 GEF closely related to RA-GEF-1, designated RA-GEF-2, was isolated from a human fetal brain library. Although RA-GEF-1 and RA-GEF-2 show a striking sequence homology in their RA domains, the binding specificity is completely different. The RA domain of RA-GEF-1 interacts with Rap1⅐GTP, leading to the translocation of RA-GEF-1 to the perinuclear compartments including the Golgi apparatus (14). In marked contrast, the RA domain of RA-GEF-2 binds to M-Ras, but not Rap1, in a GTP-dependent manner. Upon binding, RA-GEF-2 translocates to the plasma membrane, where it may act as a GEF specific to plasma membrane-localized Rap1. In fact, M-Ras 71L -dependent increase in the level of the GTP-bound form of plasma membrane-localized Rap1 was observed when coexpressed with RA-GEF-2 (Fig. 6D). However, total amounts of Rap1⅐GTP in cells expressing Rap1 and RA-GEF-2 were decreased in the presence of M-Ras 71L (Fig. 6C). This may be ascribed to the translocation of RA-GEF-2 to the plasma membrane that may cause seques-tration of RA-GEF-2 from a major population of Rap1, which is not localized in the plasma membrane.
It should be noted that we observed no significant increase or decrease in GEF activity in vitro upon the association of the RA domain of RA-GEF-2 with M-Ras⅐GTP (Figs. 6, A and B). Likewise, in vitro GEF activity of RA-GEF-1 is thought to remain unchanged upon the binding to Rap1⅐GTP at its RA domain on the basis of the analysis of an RA domain-deletion mutant (18). Rather, it is feasible that Rap1 binding promotes the translocation of RA-GEF-1 to specific subcellular compartments, such as the Golgi apparatus, where RA-GEF-1 enhances the formation of Rap1⅐GTP (18). By analogy with this mechanism, translocation of RA-GEF-2 to the plasma membrane may be a crucial event in signal transduction through M-Ras and RA-GEF-2.
GFR/MR-GEF is another member of the RA domain-containing Rap1 GEF family (27,28). The RA domain of GFR/MR-GEF specifically interacts with the GTP-bound form of M-Ras in vivo and in vitro, suggesting that GFR/MR-GEF as well functions downstream of M-Ras (28). Interestingly, primary structures of RA domains are much less homologous between RA-GEF-2 and GFR/MR-GEF (19.7% identity) in comparison to the homology between RA-GEF-1 and RA-GEF-2 (81.6% identity), although both RA-GEF-2 and GFR/MR-GEF, but not RA-GEF-1, exhibit specific binding ability toward M-Ras⅐GTP. GFR/MR-GEF-dependent increase in the total Rap1⅐GTP level within the cell was suppressed when coexpressed with activated M-Ras, which is similar to our results described in Fig. 6C, and thus negative roles for M-Ras and GFR/MR-GEF in the regulation of the Rap1 pathway were suggested (28). However, considering our observation that the level of the GTP-bound form of plasma membrane-localized Rap1 in RA-GEF-2-expressing cells was significantly increased upon coexpression of the activated form of M-Ras, it is likely that in vivo GEF activity of GFR/MR-GEF may also be up-regulated by M-Ras⅐GTP at the plasma membrane.
Another example of a molecule that possesses both RA and GEF domains, thereby regulating Ras family-mediated signaling via subcellular translocation, is PLC⑀. PLC⑀ was originally isolated as a Ras effector, which in fact was activated by Ras in a liposome-based reconstitution system as well as in cotransfected cells (29 -31). Presumably, PLC⑀ is an effector for Rap1 as well because the RA domain of PLC⑀ associates with Rap1 in a GTP-dependent manner, and epidermal growth factor stimulation induces PLC⑀ to translocate to the perinuclear region where Rap1 is colocalized (29). Our recent characterization of the CDC25 homology domain of PLC⑀ revealed that this domain acts as a GEF for Rap1 and is required for prolonged activation of the Rap1 pathway particularly in the perinuclear region (32). These observations suggest a mechanism whereby PLC⑀ acts not only as a Ras/Rap1-regulated PLC, but also as a signal amplifier that continuously generates the GTP-bound form of Rap1 in a specific subcellular compartment.
In addition to RA and GEF domains, RA-GEF-2 contains putative cNMP-binding and PDZ domains (Fig. 1). Neither cAMP nor cGMP binds to the putative cNMP-binding domain of RA-GEF-1 (14 -16, 18). Likewise, RA-GEF-2 interacted with neither cAMP nor cGMP, 2 and thus the function of this domain remains totally unknown. The PDZ domain is responsible for protein-protein interaction, leading to the formation of a functional signaling complex at specific subcellular sites (33). To date, binding partners of PDZ domains of RA-GEF-1 and RA-GEF-2 have not been identified. Considering the totally different subcellular localization of RA-GEF-1 and RA-GEF-2 upon the binding to GTPases through the RA domains, their PDZ domains may interact with a distinct set of proteins.
Although little is known regarding the physiological function of M-Ras, it is feasible that M-Ras may direct signaling pathways downstream of cell surface receptors because M-Ras is regulated by GEFs such as Sos1 and Ras-GRF1, which are known to act downstream of a wide variety of receptors (34). Moreover, an activated form of M-Ras displays transforming potential (34,35), implying that M-Ras, like Ha-, Ki-, and N-Ras, may be involved in the modulation of cell growth and differentiation. In fact, an activated mutant of M-Ras interacts with the Ras effector AF6 (34). However, M-Ras does not bind to multiple other known Ras effectors, such as Raf-1 and phosphoinositide 3-kinase (34), and therefore signaling pathways downstream of M-Ras remain to be clarified. Our findings that the Rap-specific GEF RA-GEF-2 serves as an effector of M-Ras in the plasma membrane may provide insights into the understanding of downstream pathways of M-Ras. Although speculative, RA-GEF-2 may be implicated in the activation of Raf/ extracellular signal-regulated kinase signaling pathways through Rap1 because Rap1 activates B-Raf in various types of cells (6,36,37). Future studies will reveal the role of a novel signaling cascade involving M-Ras and Rap1 in receptor-mediated cell responses.