The Exchange Factor and Diacylglycerol Receptor RasGRP3 Interacts with Dynein Light Chain 1 through Its C-terminal Domain*

RasGRP3 is an exchange factor for Ras-like small GTPases that is activated in response to the second messenger diacylglycerol. As with other diacylglycerol receptors, RasGRP3 is redistributed upon diacylglycerol or phorbol ester binding. Several factors are important in determining the pattern of translocation, including the potency of the diacylglycerol analog, the affinity of the receptor for phospholipids, and in some cases, protein-protein interactions. However, little is known about the mechanisms that play a role in RasGRP3 redistribution aside from the nature of the ligand. To discover potential protein binding partners for RasGRP3, we screened a human brain cDNA library using a yeast two-hybrid approach. We identified dynein light chain 1 as a novel RasGRP3-interacting protein. The interaction was confirmed both in vitro and in vivo and required the C-terminal domain encompassing the last 127 amino acids of RasGRP3. A truncated mutant form of RasGRP3 that lacked this C-terminal domain was unable to interact with dynein light chain 1 and displayed a dramatically altered subcellular localization, with a strong reticular distribution and perinuclear and nuclear localization. These findings suggest that dynein light chain 1 represents a novel anchoring protein for RasGRP3 that may regulate subcellular localization of the exchange factor and, as such, may participate in the signaling mediated by diacylglycerol through RasGRP3.

RasGRP3 is one of the members of the RasGRP family, a novel group of exchange factors that catalyzes the formation of the active, GTP-bound form of Ras-like small GTPases (1)(2)(3)(4). RasGRP3 has the broadest substrate selectivity of all the Ras-GRP family members and is able to activate H-Ras, R-Ras, and Rap1 (3,5). An important aspect of RasGRP3 regulation is its binding to the second messenger diacylglycerol. This binding is usually accompanied by subcellular redistribution of RasGRP3, which is believed to contribute to co-localization with its substrates (6). In fact, subcellular redistribution in response to diacylglycerol and its ultrapotent analogs, the phorbol esters, is one of the hallmarks of activation of diacylglycerol receptors, such as protein kinase C, chimaerin, and RasGRP (7). The process of translocation for a given receptor is orchestrated by a combination of factors, among them the nature and lipophilicity of the ligand (8) and the affinity for phospholipids like phosphatidylserine (9,10). The interaction of the receptor with specific intracellular proteins can also participate in the translocation events. The role of an adaptor protein as a raft or shuttling molecule has been shown in the case of protein kinase C and its adaptor protein RACK (11). It is tempting to speculate that distinct adaptor proteins participate in translocation of the different diacylglycerol receptors, among them RasGRP3. The identification of such adaptors for Ras-GRP3 should provide insights into the mechanisms involved in RasGRP3 regulation and the subsequent activation of downstream signaling cascades. To this end, we performed a yeast two-hybrid screen to identify proteins that interact with RasGRP3 using a cDNA library from human brain, a tissue where RasGRP3 is expressed (3).
We have identified the dynein light chain 1 (DLC1, also known as LC8) as a novel RasGRP3-interacting protein. DLC1 is a component of cytoplasmic dynein, a molecule involved in cytoskeleton-mediated motility and intracellular transport events (12). Recent studies have shown that DLC1 interacts with a variety of proteins, many of them involved in cell survival and apoptosis, such as BimL (13), neuronal nitric-oxide synthase (14), IkB␣ (15), and Pak1 (16).
In this study, we demonstrated that RasGRP3 specifically interacted with DLC1 both in vitro and in vivo and that the interaction occurred through the C-terminal region of Ras-GRP3 located after the C1 domain and EF-hands, which are responsible for binding to diacylglycerol and calcium, respectively. Moreover, a truncated form of RasGRP3 that lacked the ability to interact with DLC1 displayed a strikingly different subcellular localization from that of full-length RasGRP3. Taken together, these findings provide evidence of a novel interaction between RasGRP3 and DLC1 and suggest that DLC1 could represent an anchoring protein that participates in the subcellular localization of RasGRP3.
Expression Constructs-pGBKT7 (TRP1, 2 m) Gal4p DNAbinding domain vector was obtained from BD Biosciences Clontech. pTD1-1 (SV40 large T-antigen in pACT2) was recovered from Y187 (pTD1-1) using the YEASTMAKER yeast plasmid isolation kit (BD Biosciences Clontech). pACT2B2 (LEU2, 2 m) Gal4p activation domain vector was reconstituted by deleting the 2.1-kb BglII fragment, including the SV40 large T-antigen and an hemagglutinin epitope, from pTD1-1. pGBKT7-RasGRP yeast two-hybrid bait plasmids, for expression of the different RasGRP proteins fused to the Gal4p DNAbinding domain, were prepared by standard molecular biology procedures using the following cDNAs: rat RasGRP1, human RasGRP2 (short spliced version, isoform 2), and human Ras-GRP3 (KIAA0846). RasGRP3EC (amino acids 408 -564 of Ras-GRP3) and RasGRP3CSt (amino acids 408 -691 of RasGRP3) yeast two-hybrid bait plasmids were generated by introducing the corresponding cDNAs into pGBKT7. pACT2-K6, containing a hybrid cDNA insert that includes the full-length coding sequence of cytoplasmic dynein light chain DLC1 (NCBI accession code NM_003746) downstream of a mitochondrial genomic sequence and poly(A) tail, was isolated through screening of a pretransformed adult human brain cDNA library for clones displaying a positive interaction with pGBKT7-Ras-GRP3 in a yeast two-hybrid assay (see below). The DLC1 cDNA was isolated from pACT2-K6 through PCR amplification of the open reading frame and cloned into a pACT2 (LEU2, 2) Gal4p activation domain vector plasmid to create pACT2-DLC1. DLC1 was subcloned into pGEX-2T (Amersham Biosciences) for expression of a glutathione S-transferase (GST)-DLC1 fusion protein in E. coli. Mammalian expression constructs for N-terminal FLAG-tagged proteins were generated by subcloning the corresponding RasGRP cDNAs into pCMV-Tag2A/B/C expression vectors (Stratagene, La Jolla, CA). Plasmids for expression of N-terminal green fluorescent protein (GFP)-tagged proteins were generated using the pEGFP-C1 vector (BD Biosciences Clontech). Correct sequence, orientation, and reading frame of inserts were verified by DNA sequencing at the University of Hawaii at Manoa by the Greenwood Molecular Biology Facility, Cardiovascular Research Center/Cancer Research Center DNA Sequencing/ Genotyping Facility or Center for Genomics, Proteomics, and Bioinformatics Research Initiative.
Yeast Transformations and Two-hybrid Screening-Plasmids were transformed into yeast strains using a lithium acetate transformation procedure as outlined in the BD Biosciences Clontech Matchmaker GAL4 Two-hybrid System 3 and Libraries User Manual. The pGBKT7-RasGRP3 yeast two-hybrid bait plasmid was transformed into reporter strain AH109 to create AH109 (pGBKT7-RasGRP3). An adult human brain cDNA library pretransformed in yeast strain Y187 (BD Biosciences Clontech pretransformed human brain Matchmaker cDNA library) was screened in a yeast two-hybrid assay through large scale mating to AH109 (pGBKT7-RasGRP3) following the manufacturer's instructions. Positive clones were selected on medium lacking adenine, histidine, leucine, and tryptophan to identify diploids containing a pACT2-cDNA library plasmid and the pGBKT7-RasGRP3 plasmid and expressing the ADE2 and HIS3 two-hybrid reporter genes. Positive clones were further screened for expression of an additional reporter gene, MEL1 (encoding melibiase/␣-galactosidase), using the chromogenic substrate 5-bromo-4-chloro-3-indolyl ␣-D-galactopyranoside (X-␣-gal; BD Biosciences). Plasmids were recovered using the YEASTMAKER yeast plasmid isolation kit (BD Biosciences) or Zymoprep yeast plasmid minipreparation kit (Zymo Research, Orange, CA).
GST Pulldown and Co-immunoprecipitation Assays-CHO-K1 cells were transiently transfected with RasGRP plasmids using FuGENE 6 transfection reagent according to the manufacturer's instructions (Roche Applied Science). Fortyeight to seventy-two hours later, cells were washed twice with ice-cold TBS (25 mM Tris-HCl, pH 7.4, 0.45% NaCl) and resuspended in 500 l of lysis buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM MgCl 2 , 1% IGEPAL, 1 mM dithiothreitol, 5% glycerol, and protease inhibitors). For immunoprecipitations, the lysis buffer also included 25 mM NaF. Complete Mini, EDTA-free protease inhibitor tablets were purchased from Roche Applied Science. Lysates were briefly vortexed, incubated on ice for 5 min, and then centrifuged at 13,000 rpm for 15 min at 4°C. The resulting supernatants were collected for GST pulldown or co-immunoprecipitation assays.
For GST pulldown studies, a GST-DLC1 fusion protein was produced in E. coli. Cells were lysed by sonication on ice in phosphate-buffered saline in the presence of protease inhibitors, followed by addition of Triton X-100 to a 1% final concentration and subsequent incubation for 30 min on ice. The lysate was partially purified by centrifugation at 18,000 ϫ g for 20 min at 4°C. The resulting supernatant was used as a GST-DLC1 probe in a pulldown assay. GST alone was also produced to use as a negative control in the assay. The GST-DLC1 probe was incubated with rotation at 4°C for 1-3 h with glutathione-agarose beads (SwellGel Immobilized Glutathione Discs; Pierce) in lysis buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM MgCl 2 , 1% IGEPAL, 1 mM dithiothreitol, 5% glycerol, and protease inhibitors). GST-DLC1-glutathione-agarose beads were washed three times with lysis buffer to remove unbound GST-DLC1 and then aliquoted for incubation with cell lysates. For each cell lysate, an aliquot containing 500 g of total protein was incubated with the bead-coupled GST-DLC1 probe at 4°C for 1 h with rotation. After incubation, beads were washed three times with lysis buffer, resuspended in 2ϫ Laemmli buffer con-taining 5% ␤-mercaptoethanol, and boiled for 10 min. Samples were analyzed by SDS-PAGE and immunoblotting using the anti-FLAG M2 mouse monoclonal antibody to detect FLAGtagged RasGRP proteins. For co-immunoprecipitation assays, aliquots of cell lysates containing equal amounts (ϳ1 mg) of total protein were incubated with 24.5 g of anti-FLAG M2 mouse monoclonal antibody overnight with rotation at 4°C. Immunocomplexes were then precipitated by rotation for 2 h at 4°C with 50 l of protein A/G PLUS-agarose beads (Santa Cruz Biotechnology, Santa Cruz, CA) that had been prewashed and resuspended in lysis buffer. Beads were washed four times with lysis buffer, resuspended in 2ϫ Laemmli buffer containing 5% ␤-mercaptoethanol, and boiled for 10 min. Anti-FLAG immunoprecipitated samples and total lysates were analyzed by SDS-PAGE and immunoblotting, using anti-DLC1 rat IgM to detect endogenous DLC1 and anti-FLAG rabbit polyclonal antibody to detect FLAG-tagged RasGRP proteins.
Confocal Microscopy Studies-For the immunofluorescence studies, CHO-K1 cells were seeded onto Permanox chamber slides and transfected 18 h later with the FLAG-tagged Ras-GRP3 expression constructs using FuGENE 6 (Roche Applied Science). Forty-eight hours later, the cells were fixed with 3.7% neutral formaldehyde at 37°C for 10 min, followed by 5-min permeabilization with 0.2% Triton X-100. Slides were blocked with 10% normal goat serum diluted in phosphate-buffered saline for 1 h at room temperature, followed by a 2-h incubation with 20 g/ml of the anti-FLAG M2 monoclonal antibody diluted in 1% normal goat serum. After extensive washes in phosphate-buffered saline, a 1:200 dilution of Alexa Fluor 594 goat anti-mouse secondary antibody (Invitrogen) was added to the slides in 1% normal goat serum for 1 h in the dark. Cells were washed several times in phosphate-buffered saline, mounted in Vectashield mounting medium (Vector Laboratories, Burlingame, CA), and viewed with a Leica TCS SP5 or a Zeiss LM5 Pascal scanning confocal microscope. Cells transfected with GFP-tagged constructs were either fixed as above or used in live cell experiments. For the latter, CHO-K1 cells were seeded onto 40-mm circular glass coverslips and transfected with the GFP constructs using FuGENE 6. Forty-eight hours later, the coverslips were enclosed in a Bioptechs Focht Chamber (Bioptechs, Butler, PA), which was attached to the microscope stage of a Leica TCS SP5 confocal microscope with a custom stage adaptor, connected to a temperature controller at 37°C, and perfused with phenol red-free medium supplemented with 8% fetal bovine serum using a Lambda microperfusion pump. A 1-M dilution of 12-O-tetradecanoyl phorbol 13-acetate (TPA) in medium was perfused for 30 min, during which time sequential images of the same cell were collected.
Ras-GTP Pulldown Assay and Western Blots-Levels of GTPloaded Ras (Ras GTP ) were measured as previously described with some modifications (6). Briefly, CHO-K1 cells were transiently transfected with RasGRP3 expression constructs for 48 h, and cytoplasmic fractions were prepared to be used as lysates for the pulldown assay. To evaluate TPA-mediated activation of Ras, cells were serum starved overnight and then treated with vehicle (Me 2 SO) or 1 M TPA for 30 min before harvesting them in lysis buffer for the pulldown assay. Lysates were mixed, incubated on ice for 5 min, and then clarified by centrifugation at 13,000 rpm for 15 min at 4°C. Lysate protein (500 g) was incubated with the Ras-binding domain (RBD) of Raf-1 conjugated to glutathione S-transferase (GST-Raf-1-RBD probe) and bound to glutathione beads for 1 h with rotation in the cold. The affinity complexes were washed three times with lysis buffer and then resuspended in Laemmli buffer containing 5% ␤-mercaptoethanol, boiled, and resolved on 15% acrylamide gels. Total lysate protein (25 g) was run in parallel as control of the input of total Ras in the pulldown assay. Proteins were blotted onto nitrocellulose membranes. Immunostaining was done using an anti-Ras antibody. Rap1 pulldown assays were done using the EZ-Detect Rap1 activation kit according to the manufacturer's instructions (Pierce). Western blot bands were analyzed by densitometry using UN-SCAN-IT gel software (Silk Scientific, Orem, UT) or Scion Image software (Scion Corp., Frederick, MD), and statistical significance was calculated in GraphPad Prism (GraphPad Software, San Diego, CA) using nonpaired Student's t test or one-way analysis of variance followed by Tukey's multiple comparison test. Fractionation of cell lysates into cytoplasmic and nuclear fractions was performed using the nuclear extraction kit (Active Motif, Carlsbad, CA).

DLC1 Is a Novel RasGRP3-interacting Protein Isolated in a
Yeast Two-hybrid Screen-To identify binding partners of Ras-GRP3, we performed a yeast two-hybrid screen of a human brain cDNA library using the full-length human RasGRP3 protein as bait. One of the clones isolated from the screen contained the full-length coding sequence of DLC1 (NCBI Accession Number NM_003746) downstream of a mitochondrial genomic sequence and poly(A) tail. The DLC1 cDNA was isolated from the clone, and the two-hybrid interaction with Ras-GRP3 was confirmed by cotransforming the yeast reporter strain AH109 and assaying for expression of the yeast two-hybrid reporter genes ADE2 and HIS3 by growth on nutrientdeficient plates (Fig. 1A). Because RasGRP3 shares a significant degree of sequence similarity with other members of the Ras-GRP family (Fig. 1B), we tested whether RasGRP1 and RasGRP2 (short spliced version) could also interact with DLC1. Neither protein interacted with DLC1 in the yeast two-hybrid assay, suggesting that the association was specific for RasGRP3.
To map the region of RasGRP3 responsible for DLC1 binding, we used a truncated form of RasGRP3 consisting of C-terminal amino acids 408 -691 (RasGRP3CSt) or a shorter variant (RasGRP3EC) consisting of RasGRP3 amino acids 408 -564, which include the EF-hands and C1 domain but lacked the last 127 amino acids (Fig. 1B). We found that only RasGRP3CSt was able to interact with DLC1 in the yeast two-hybrid assay, suggesting that the region located within the last 127 amino acids of RasGRP3 was critical for its interaction with DLC1 in the yeast two-hybrid system. Sequence analysis of this C-terminal region of RasGRP3 revealed the presence of a TQT motif at amino acids 615-617 (Fig. 1B). Similar TQT motifs, contained within the (R/K)XTQT or AATQT consensus sequences, have been reported to be binding sites for DLC1 (17,18). The presence of a TQT motif in the RasGRP3 C-terminal domain was consistent with the interaction of this domain with DLC1.
DLC1 Interacts with RasGRP3 in Vitro-To confirm the specific interaction of RasGRP3 with DLC1 in vitro, we performed a pulldown assay using a GST fusion protein of DLC1 (GST-DLC1). GST-DLC1 was immobilized on glutathioneagarose beads and used as a probe to affinity precipitate a FLAG-tagged version of human RasGRP3 (FLAG-RasGRP3) transiently expressed in CHO-K1 cells. As shown in Fig. 2A, FLAG-RasGRP3 was specifically precipitated by GST-DLC1 and not by the GST protein alone. Mock cells (transfected with empty FLAG vector) were included as a negative control.
The yeast two-hybrid results suggested that the C-terminal region of RasGRP3 encompassing the last 127 amino acids was essential for interaction with DLC1. To examine whether this domain was also necessary for binding to DLC1 in vitro, we transfected CHO-K1 cells with plasmids encoding the FLAG-tagged versions of either a truncated mutant variant of RasGRP3 that lacked the C-terminal domain (RasGRP3⌬C) or the isolated RasGRP3 C-terminal region (RasGRP3COnly) (Fig. 2B) with RasGRP2 was also evaluated. Consistent with the results from the yeast two-hybrid assays, RasGRP2 did not interact with DLC1 in vitro (Fig. 2B, lane 1). Taken together, these experiments confirmed the specific interaction of Ras-GRP3 with DLC1 and further substantiated the participation of the RasGRP3 C-terminal region in DLC1 binding.
RasGRP3 Associates with DLC1 in Vivo-The findings in the yeast two-hybrid system and the GST pulldown in vitro assay strongly suggested that RasGRP3 could interact with DLC1 in vivo. To validate this association, we performed co-immunoprecipitation assays. We transiently expressed FLAG-tagged RasGRP3 and its variants in CHO-K1 cells and immunoprecipitated the proteins using an anti-FLAG antibody. For comparison, we used FLAG-RasGRP2, which was negative for interaction with DLC1 in the yeast two-hybrid screen and in the in vitro assay. As shown in Fig. 3  Approximately equal numbers of cotransformed cells were plated on synthetic medium lacking leucine and tryptophan (SD-Leu-Trp) to select for cotransformants and on synthetic medium lacking adenine, histidine, leucine, and tryptophan (SD-Ade-His-Leu-Trp) to select for cotransformants in which the ADE2 and HIS3 twohybrid reporter genes were expressed to indicate a positive interaction. Growth was recorded after 23 days. Similar results were obtained with pACT2-DLC1, which contains only the DLC1 gene from pACT2-K6; however, cells transformed with this plasmid exhibited a dramatic growth defect. B, schematic representation of the RasGRP proteins tested in the yeast two-hybrid assay. The results of the two-hybrid assay are indicated on the right (Ϫ, no interaction; ϩ, positive interaction).

RasGRP3 and the RasGRP3⌬C Mutant Protein Display Different Subcellular Localization
Patterns-We hypothesized that if interaction with DLC1 was an important determinant of the subcellular localization of RasGRP3, disruption of the interaction should affect RasGRP3 distribution in the cell. Immunofluorescence analysis of CHO-K1 cells expressing either FLAG-RasGRP3 or the FLAG-RasGRP3⌬C mutant protein, which was defective for interaction with DLC1, revealed significant differences in localization of these proteins. Whereas full-length Ras-GRP3 mainly localized in the cytoplasm with some perinuclear staining, RasGRP3⌬C displayed a strong reticular distribution and staining of the nuclear envelope (Fig. 4A, FLAG panel). We observed the same results using GFP-tagged proteins in fixed as well as in live cells (Fig. 4A, GFP-fixed and GFP-live, respectively). The FLAG-RasGRP3Conly protein, which was able to bind to DLC1 both in vitro and in vivo, displayed a subcellular localization very similar to that of full-length RasGRP3 (supplemental Fig. S1).
Interestingly, the RasGRP3⌬C mutant protein also formed punctate and filamentous structures inside the nucleus (Fig. 4A,  right column). Fractionation of FLAG-RasGRP3 or FLAG-RasGRP3⌬C expressing cells into cytoplasmic and nuclear frac-tions confirmed accumulation of the RasGRP3 mutant protein in the nucleus (Fig. 4B).
The subcellular location of RasGRP3 should determine its accessibility to substrates and changes in response to activators such as TPA and various diacylglycerol analogs (6,19,20). To evaluate whether there was a difference in the response to TPA between RasGRP3 and RasGRP3⌬C, we transfected CHO-K1 cells with the corresponding GFP-tagged constructs and examined TPA-induced translocation of the proteins in live cells by confocal microscopy. As we have seen before using HEK293 cells (6), 1 M TPA caused a rapid redistribution of GFP-RasGRP3, mainly to internal membranes (Fig. 4C, left column). Translocation of GFP-RasGRP3⌬C in response to TPA was less evident because of the strong reticular localization of this protein under non-stimulated conditions. Nevertheless, there was some redistribution of the protein in response to TPA, particularly to the cell periphery (Fig. 4C,  right column). As noted before, there was marked localization of RasGRP3⌬C to the nucleus. Interestingly, we also observed the formation of a few filamentous and punctate structures in the nucleus of RasGRP3-expressing cells upon TPA treatment (Fig. 4C, left column).
The striking difference in subcellular localization between RasGRP3 and the RasGRP3⌬C mutant protein led us to investigate whether there were any consequences on the downstream RasGRP3 signaling. To this end, we measured activation of one of the substrates for RasGRP3, the small GTPase Ras. In CHO-K1 cells grown in complete medium, levels of active Ras (Ras GTP ) were only slightly elevated in the  RasGRP3 expressors compared with the mock (vector-transfected control) cells (Fig. 5A). However, ectopic expression of the RasGRP3⌬C mutant protein caused a Ն2.0-fold increase in Ras GTP levels compared with those of the RasGRP3 expressors and control cells (Fig. 5A). We also observed a tendency for elevated Ras GTP levels in RasGRP3⌬C-expressing cells under serum starvation conditions, albeit not statistically significant when compared with the levels of active Ras in control or RasGRP3-expressing cells (Fig. 5B). Stimulation with 1 M TPA further increased the levels of active, GTP-bound Ras in both RasGRP3 and RasGRP3⌬C overexpressors (Fig. 5B). Activation of the Rasrelated small GTPase Rap1 in response to TPA in CHO-K1 cells was insensitive to ectopic expression of either RasGRP3 or the mutant protein (supplemental Fig. S2).

DISCUSSION
In the present work, we have demonstrated that DLC1 is a novel RasGRP3-interacting protein. Preliminary mapping of the RasGRP3 domain involved in the association with DLC1 pointed at a region located within the last 127 amino acids of human RasGRP3. Interestingly, we identified a TQT motif at amino acids 615-617 within this region. The (R/K)XTQT consensus motif, where X is any amino acid, is one of the motifs reported to be responsible for DLC1 binding (17). The TQT motif present in Ras-GRP3 lacks the (R/K) amino acid but contains an alanine immediately before the threonine, similar to an AATQT site identified in p53-binding protein 1, another DLC1-interacting molecule (18).
A RasGRP3 mutant protein lacking the C-terminal domain, including the TQT consensus motif, failed to interact with DLC1 both in vitro and in vivo in mammalian cells. This C-terminal region does not contain any other known domains, repeats, or motifs and is located after the C1 domain and EF-hands of RasGRP3. In addition to loss of DLC1 binding, this RasGRP3 mutant protein displayed a dramatically different subcellular localization compared with full-length RasGRP3, with strong reticular partitioning and staining of the nuclear envelope. Additionally, we observed a significant nuclear accumulation of the RasGRP3 mutant protein, which formed punctate and filamentous structures. We also noticed formation of punctate nuclear structures in some of the cells transfected with full-length RasGRP3 upon activation with phorbol esters. Localization of RasGRP3 in the nucleus has not been reported before, and its significance is puzzling. However, forms of H-Ras and Rap1, which are both potential substrates of RasGRP3, have been found in the nuclei of N-nitrosodiethylamine-induced liver tumor cells and squamous cell carcinoma lines, respectively (21,22). It is interesting to note that DLC1 has been suggested to act as a chaperone for nuclear translocation of estrogen receptor in breast cancer cells (23) and for nuclear accumulation of p53 in response to DNA damage (18) but was also reported to function together with DLC2 to prevent nuclear localization of the brain-enriched protein PMES-2 (24). Thus, DLC1 appears to regulate the nuclear localization of certain proteins, but whether DLC1 promotes or inhibits nuclear accumulation appears to depend on its interacting partners. The unusual localization of the RasGRP3 C-terminal deletion protein in the nucleus may be related to its loss of interaction with DLC1; however, we cannot preclude the possibility that interactions with other proteins besides DLC1 could occur at the C terminus of RasGRP3 and contribute to RasGRP3 localization.
The ectopic expression of RasGRP3⌬C increased the levels of active Ras in CHO-K1 cells. The effect of the RasGRP3⌬C mutant protein was significant in the presence of serum (complete medium) compared with serum starvation conditions. One could argue that RasGRP3 mislocalization could lead to closer proximity to Ras, consequently inducing higher levels of Ras activation in response to growth factor stimulation. This finding supports the concept that localization of RasGRP3, potentially mediated in part by the interaction with DLC1, could play an important role in controlling downstream signaling pathways.
Treatment with the phorbol ester TPA, which changed the subcellular distribution of RasGRP3 and, to some extent, of the mutant protein, induced a further increase in Ras activation in both RasGRP3-and RasGRP3⌬C-expressing cells. TPA appeared to be equally potent at activating Ras through Ras-GRP3 and its mutant protein, in contrast with the effect of complete medium. This could indicate that activation of RaGRP3 via TPA may proceed independently of its interaction with DLC1. Alternatively, TPA may dissociate the RasGRP3-DLC1 interaction as part of its mechanism of RasGRP3 activa-tion. This dissociation would not be needed in the case of the Ras-GRP3⌬C mutant protein. However, preliminary experiments did not show a significant change in the amount of DLC1 co-precipitating with RasGRP3 upon TPA treatment (data not shown), arguing against the second hypothesis. Thus, the relevance of the RasGRP3-DLC1 interaction for RasGRP3 signaling appears to vary depending on the stimulus. DLC1 is a highly conserved protein that is ubiquitously distributed among cell lines and tissues. Its ability to interact with a variety of unrelated proteins that play roles in apoptosis and proliferation suggests that DLC1 may act to bring diverse signaling molecules together. In the context of our present findings, it is tempting to speculate that DLC1 may act as an anchoring protein for RasGRP3, preventing it from adopting an active distribution in the absence of activating signals. Further experiments measuring association kinetics of the RasGRP3-DLC1 interaction in response to diacylglycerol and other physiological stimuli, as well as analysis of substitutions within the TQT consensus motif in RasGRP3, are needed to more completely define the RasGRP3-DLC1 association and examine the significance of this interaction in RasGRP3 signaling and function.