Tumor-related Alternatively Spliced Rac1b Is Not Regulated by Rho-GDP Dissociation Inhibitors and Exhibits Selective Downstream Signaling*

Rac1 is a member of the Rho family of small GTPases, which control signaling pathways that regulate actin cytoskeletal dynamics and gene transcription. Rac1 is activated by guanine nucleotide exchange factors and inactivated by GTPase-activating proteins. In addition, Rho-GDP dissociation inhibitors (Rho-GDIs) can inhibit Rac1 by sequestering it in the cytoplasm. We have found previously that colorectal tumors express an alternatively spliced variant, Rac1b, containing 19 additional amino acids following the switch II region. Here we characterized the regulation and downstream signaling of Rac1b. Although little Rac1b protein is expressed in cells, the amount of activated Rac1b protein often exceeds that of activated Rac1, suggesting that Rac1b contributes significantly to the downstream signaling of Rac in cells. The regulation of both Rac1 and Rac1b activities is dependent on guanine nucleotide exchange factors and GTPase-activating proteins, but the difference in their activation is mainly determined by the inability of Rac1b to interact with Rho-GDI. As a consequence, most Rac1b remains bound to the plasma membrane and is not sequestered by Rho-GDI in the cytoplasm. Unlike Rac1, activated Rac1b is unable to induce lamellipodia formation and is unable to bind and activate p21-activated protein kinase nor activate the downstream protein kinase JNK. However, both Rac1 and Rac1b are able to activate NFκB to the same extent. These data suggest that alternative splicing of Rac1 leads to a highly active Rac variant that differs in regulation and downstream signaling.

Rac1 is a member of the Rho family of small GTPases, key molecules in signaling pathways that control cell migration, cell adhesion, cell cycle progression, and tumor formation (1)(2)(3)(4). Rac1 cycles between an active GTP-bound state and an inactive GDP-bound state. In vivo, this process is tightly controlled and spatially regulated by activating guanine nucleotide exchange factors (GEFs), 1 which promote exchange of GDP for GTP, and by inactivating GTPase-activating proteins (GAPs), which accelerate the intrinsic GTPase activity. In addition, Rho-GDP dissociation inhibitors (Rho-GDIs) can regulate Rac1 activity by sequestering the protein in the cytoplasm.
Following its activation Rac1 can interact with several downstream effector molecules and trigger various cellular responses. One such effector, for example, is the protein kinase PAK that becomes activated upon direct interaction with GTP-Rac1 (5). Furthermore, active Rac1 binds to IRSp53, which results in activation of WAVE2 and subsequent stimulation of actin polymerization and lamellipodia formation (6). Active Rac1 can also stimulate the activity of the Jun N-terminal kinase (JNK) cascade (7,8) and the transcription factor NFB (9,10).
Depending on cell type and growth conditions studied, cellular Rac1 signaling regulates epithelial cell-cell adhesion (11)(12)(13), cell spreading, cell motility (14 -17), and cell cycle progression (18). The latter is probably mediated via activation of the transcription factor NFB (10,19). De-regulated Rac1 activity and a disturbed balance between Rac and Rho activities may determine epithelial-mesenchymal transition in tumor cells (20 -22). In contrast to the mutations found in oncogenic Ras, activating mutations in Rac1 have never been found in tumors. However, de-regulated Rac activity, either by overexpression of Rac or by aberrations in Rac regulators, has been demonstrated to influence various aspects of tumorigenicity (3,4,(23)(24)(25)(26).
We have identified previously (27) a splice variant of Rac1, designated Rac1b, in colon tumors. Rac1b was also found in human and mice breast tumors (24) 2 and contains an additional 19 amino acids following the switch II region. When we determined the structure of the human RAC1 gene (28), these 19 amino acids were found encoded in an additional exon of Rac1. Rac1b is thus derived from the RAC1 gene by alternative splicing.
Here we have characterized the Rac1b protein and found that alternative splicing generates a highly activated variant of Rac1 with altered GTPase regulation and selective downstream signaling.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-Cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum (HT29, SW480, ZR75.1, HeLa) or 10% bovine serum (NIH 3T3) (Invitrogen) and regularly checked for mycoplasm infection. For transfections, cells at 50 -75% confluence were transfected using LipofectAMINE Plus (Invitrogen), according to the manufacturer's instructions, and cells were analyzed after 16 or 22 h, respectively. Total amounts of transfected DNA for biochemical assays were 4 g with SW480 or HT29 and 2 g with NIH 3T3 cells (60-mm dishes). For immunofluorescence, 2 or 1 g, respectively, was used in 35-mm dishes. When required the amount of DNA was adjusted with empty vector. Transfection efficiencies were 40 -80% depending on the cell type used and as judged by co-transfection with pEGFP. To analyze Rho-GDI inhibition of endogenous Rac1b, 1.5 ϫ 10 7 HT29 cells were divided into six 60-mm dishes and transfected individually (0.5 g of GDI each dish), and the lysates were subsequently pooled.
DNA Plasmids and Constructs-Rac1 and Rac1b cDNAs were amplified using a ⌬atg 5Ј primer containing a BamHI restriction site and a 3Ј primer containing an EcoRI restriction site (28). The PCR products were cloned in pCR2.1-TOPO (Invitrogen), sequenced (ABI Prism 3100 automated DNA sequencer), and then subcloned into the pcDNA3-Myc plasmid. The c-Myc and HA epitope sequences were inserted as annealed oligonucleotides between the KpnI and BamHI restriction sites of pcDNA3 (Invitrogen). pcDNA3.1-GFP-Rac1b was generated by exchange of a BstXI-EcoR1 fragment between pcDNA3-Myc-Rac1b and pcDNA3.1-GFP-Rac1 (29). pcDNA3-HA-Rho-GDI␣ was generated exchanging the complete GDI cDNA sequence from pGEX-2T-Rho-GDI␣ into pcDNA3-HA plasmid, using the flanking BamHI and EcoRI restriction sites. The Bcr GAP variant 1 cDNA from pCMV5-Bcr-GAP and PAK1 cDNA from pCMV6-Myc-PAK1 were subcloned into pCMV2-FLAG (Sigma). IB was cloned into pcDNA3-Myc following amplification from pcDNA3-IB with primers containing a BamHI and EcoRI restrictions site. All constructs were verified by automatic DNA sequencing.
SDS-PAGE and Western Blotting-Samples were boiled for 10 min, centrifuged at 2,500 ϫ g for 30 s, and resolved in 12-15% SDS-PAGE mini-gels. Proteins were transferred onto polyvinylidene difluoride (Bio-Rad) membranes using a Mini Trans-Blot cell (Bio-Rad). Membranes were probed using the indicated antibodies, and specific binding was detected via a secondary peroxidase-conjugated antibody (Bio-Rad) followed by chemiluminescence. Monoclonal anti-Myc 9E10 and anti-HA 12CA5 were from Roche Applied Science. Polyclonal anti-Myc was from Santa Cruz Biotechnology. Polyclonal anti-HA and monoclonal anti-FLAG were from Sigma. Anti-Rac1 monoclonal antibody was from Upstate Biotechnologies, Inc. Anti-␤-catenin was from Transduction Laboratories. Anti-active JNK was from Promega. Anti-phospho IB was from Hypromatrix, and anti-phospho-PAK1 was from Cell Signaling. Anti-Rac1b serum was produced by Sigma-Genosys according to their standard immunization protocol by coupling the peptide VGETYGKDITSRGKDKPIAC to keyhole limpet hemocyanin. The serum was affinity-purified on a peptide-Sepharose column.
Rac Pull-down Experiments and Immunoprecipitation-1-1.5 ϫ 10 7 cells were grown in 100-mm dishes and used for endogenous Rac1/ Rac1b pull-down assays. For transfection assays, 1-2.5 ϫ 10 6 cells were seeded in 60-mm dishes, transfected as indicated, and assayed 24 h later. In both cases, cells were washed in cold phosphate-buffered saline and lysed on ice in 750 and 400 l, respectively, of lysis buffer (50 mM Tris-HCl, pH 7.5, 1% Nonidet P-40, 100 mM NaCl, 10% glycerol, 10 mM MgCl 2 , and protease inhibitor mixture (Sigma)). Total lysates were cleared by centrifugation at 2,500 ϫ g for 5 min, and 0.1 volume was added to 2ϫ Laemmli sample buffer. The remaining lysate was incubated for 1 h at 4°C with glutathione S-transferase (GST)-PAK-CRIB domain (PAK-CD) pre-coupled to glutathione-Sepharose beads (Amersham Biosciences), as described (15). Precipitated complexes were washed three times in an excess of lysis buffer. After the final wash, the supernatant was discarded, and 20 l of 2ϫ Laemmli sample buffer was added to the beads. Total lysates and precipitates were then analyzed by Western blot. The immunoprecipitation procedure was identical except that the lysates were incubated with the monoclonal anti-c-Myc or anti-HA antibodies (2 g ml Ϫ1 ) pre-coupled to protein G-agarose beads (Roche Applied Science). Precipitates were analyzed on Western blots with rabbit anti-Myc or anti-HA antibodies, respectively. All results were confirmed in at least three independent experiments.
Immunofluorescence-Cells were grown on 10 ϫ 10-mm glass coverslips, washed twice in phosphate-buffered saline (PBS), immediately fixed with 3.7% formaldehyde in PBS for 20 min at room temperature, and subsequently permeabilized with 0.2% Triton X-100 in PBS for 10 min at room temperature. Cells were then washed 3 times for 5 min in PBS containing 0.05% Tween 20 (PBS-T), incubated for 1 h in PBS-T containing the primary antibody (1:400 anti-␤-catenin, Transduction Laboratories; 1:100 anti-HA, Roche Applied Science) or phalloidin-TRITC, and then washed 3ϫ and incubated for 30 min with anti-mouse antibody conjugated to Texas Red (Jackson ImmunoResearch). Coverslips were washed 3 times, mounted in VectaShield (Vector Laborato-ries), and sealed with nail polish. Images were recorded with a cooled CCD camera or a Zeiss LSM510 confocal microscope and processed with Adobe Photoshop software. All results were confirmed by at least three independent experiments.

Endogenous Rac1b Is a Highly Activated Variant-Rac1b
is an endogenous splice variant of Rac1 (27). In order to study the physiological role of Rac1b, we raised antibodies against the 19-amino acid insert that distinguishes human Rac1b from Rac1. Western blot analysis showed that these antibodies specifically recognize Rac1b but not Rac1 in mouse 3T3 fibroblasts overexpressing both cDNAs (Fig. 1A). With these antibodies several tumor cell lines as well as mouse fibroblasts were subsequently screened for the presence of endogenous Rac1b protein by Western blotting. As shown in Fig. 1B, expression of endogenous Rac1b by both reverse transcriptase-PCR and Western blotting was specifically found in HT29 (colon adenocarcinoma), ZR75.1 (breast carcinoma), and HeLa (cervix carcinoma) cells, whereas SW480 (colon adenocarcinoma) and 3T3 fibroblasts expressed Rac1 only. Commercially available anti-Rac1 antibodies recognized both Rac1 and Rac1b equally well (see Fig. 1A). This allowed us to optimize electrophoretic separation of both Rac1 and Ra1b and to compare the endogenous levels of both proteins. As shown in Fig. 1B, in all three Rac1b positive cell lines the level of Rac1b protein is low compared with that of Rac1. By densitometric analysis of Western blots, we estimated a 20-fold excess of Rac1 over Rac1b protein in HT29 cells.
Because the expression level is not always indicative for the active, functional pool of a protein, we analyzed the activation state of Rac1 and Rac1b in these cells by GST-PAK-Crib domain (CD) pull-down assays as described (15). Surprisingly, a significant amount of active Rac1b was pulled down from all Rac1b-expressing cell lines (see Fig. 1B). In order to confirm that the pull-down assay specifically isolated activated Rac1b, we transfected Myc-tagged wild type as well as active Q61L and inactive T17N mutants of Rac1b into cells and tested their activation state by pull-down assays. As shown in Fig. 1C, wild type Rac1b and the active Q61L mutant of Rac1b were clearly pulled down in this assay, whereas the inactive mutant T17N had no affinity for the GST-PAK-CD probe. This indicates that Rac1b binds PAK-CRIB in the active GTP-bound state. Interestingly, as shown in Fig. 1B, active Rac1b can exceed the amount of active Rac1 in cells. In HT29 colon cells, for example, we found three times more active Rac1b than Rac1, as estimated by Western blotting. Thus, although Rac1b protein is present in small amounts in cells, its contribution to the pool of active Rac in vivo can be very significant. Apparently, alternative splicing of Rac1 yields a highly activated variant of Rac1 that can lead to a significant increase in the total amount of activated Rac in cells.
Activation of Rac1b in Vivo Requires GEFs-We next examined the mechanism underlying the high activation level of Rac1b. Previous in vitro studies have indicated that bacterially expressed Rac1b is predominantly in the active, GTP-bound form (24). We therefore wondered whether the activation of Rac1b in vivo was independent of endogenous GEFs. To study this, HT29 cells that express endogenous Rac1 and Rac1b (see Fig. 1B) were incubated for 15 min with different concentrations of either wortmannin or LY294002 to inhibit phosphatidylinositol 3-kinase, or genistein to inhibit tyrosine phosphorylation. Both pathways have been reported previously to mediate activation of endogenous Rac-GEFs such as Tiam1 (15,30) or Vav (31-32), respectively. As shown in Fig. 2A, both inhibitors reduced the activation of Rac1 as well as of Rac1b in a dose-dependent manner. These data suggest that the activa-tions of Rac1 and Rac1b are both dependent on GEFs. Indeed, we found that the Rac activator Tiam1 (33) is able to activate both Rac1 and Rac1b (not shown) and is able to bind Rac1 and Rac1b to the same extent (Fig. 2B). These data indicate that Rac1b is able to associate with GEFs leading to its activation, similarly as found for Rac1.
Bcr-GAP Down-regulates Rac1 as Well as Rac1b-Because GAPs contribute to the activation state of Rho-like GTPases, by the capacity to stimulate their intrinsic GTPase activity, we determined whether Rac1b differs from Rac1 with respect to the down-regulation by GAPs. Recombinant Rac1b has been found to possess a reduced intrinsic GTPase activity when compared with Rac1; however, the addition of GAP protein strongly accelerated the GTPase reaction of both variants. 3 In order to confirm in vivo that GAP-stimulated GTP hydrolysis affects Rac1 as well as Rac1b, we determined the effect of the Bcr protein, an efficient GAP for Rac in vivo (34). Full-length Bcr GAP was co-transfected into SW480 cells together with tagged Myc-Rac1 and Myc-Rac1b. With increasing amounts of Bcr, we found a corresponding decrease in the activities of endogenous Rac1 as well as of exogenous Myc-Rac1 and Myc-Rac1b proteins (Fig. 3). Thus, both Rac1 and Rac1b are able to associate with Bcr and as a consequence are down-regulated because of the acceleration of the intrinsic GTPase activity of both proteins. 3 R. Ahmadian, personal communication. FIG. 1. Rac1b is a highly activated Rac variant. A, specificity of the anti-Rac1b antiserum. NIH 3T3 mouse fibroblasts were transfected as indicated with either empty vector or pcDNA3-Myc-Rac1 or pcDNA3-Myc-Rac1b. Total cell lysates were separated by SDS-PAGE. Western blots were stained as indicated with commercial anti-Rac1 (detects endogenous as well as Myc-tagged exogenous Rac), anti-Myc (detects only exogenous Myc-tagged Rac), or the Rac1b-specific anti-peptide serum (detects only exogenous Myc-Rac1b). B, detection of endogenous Rac1b. Cell lines from breast (ZR75.1), colon (HT29, SW480), cervix (HeLa), as well as 3T3 mouse fibroblasts were analyzed as indicated by reverse transcriptase-PCR and by Western blot using a Rac1b-specific antipeptide serum or a commercial monoclonal anti-Rac1 antibody that recognizes both Rac1 and Rac1b in cells. In addition, pull-down assays are shown to demonstrate the amount of active GTP-bound Rac1 and Rac1b in these cell lines. C, specificity of the GST-PAK-CD pull-down assay for the activation of Rac1b. SW480 colon cells were transfected with wild type Myc-Rac1b or with the corresponding active Q61L or inactive T17N Rac1b mutant cDNAs. Cells were lysed, and active (GTP-bound) Rac1b was pulled down and analyzed by Western blotting as described under "Experimental Procedures." Note that the inactive Rac1b mutant has no affinity for the GST-PAK CRIB domain, whereas wild type-and Rac1b-Q61L bind this domain, similarly to what has been found for Rac1.

FIG. 2. Inhibition of GEFs decreases Rac1b activation. A, Rac1b
activation is dependent on phosphatidylinositol 3-kinase and tyrosine kinase activity. HT29 cells expressing endogenous Rac1 and Rac1b were incubated for 15 min with various concentrations of either the phosphatidylinositol 3-kinase inhibitor LY294002 or the tyrosine kinase inhibitor genistein, as indicated. Cells were lysed, and active, endogenous Rac1 and Rac1b were isolated in GST-PAK pull-down assays. B, Rac1b interacts with Tiam1. NIH 3T3 fibroblasts were cotransfected with the active C1199Tiam1 mutant and either wild type Myc-Rac1 or Myc-Rac1b, as indicated, and cells were lysed after 16 h. Rac variants were immunoprecipitated (IP) with anti-Myc antibodies, and the precipitate was stained for the presence of Tiam1 with anti-HA. Note that Tiam1 co-precipitates with both Rac1 and Rac1b. Single transfections of either HA-C1199Tiam1 or Myc-Rac1b served as controls. EV, empty vector.

Rac1b Is Resistant to Down-regulation by Rho-GDI-Subse-
quently we analyzed the effect of Rho-GDI on the activation state of Rac1 and Rac1b. Rho-GDI binds and masks the hydrophobic C-terminal region of Rac, the same region that is responsible for targeting Rac to the plasma membrane (35). Thus, Rho-GDI maintains Rac in the cytoplasm and must dissociate to allow Rac to translocate to the membrane and interact with membrane-associated activators (36). When SW480 cells were co-transfected with increasing amounts of Rho-GDI, activation of both Myc-Rac1 and endogenous Rac1 was completely prevented. In contrast, the activation state of Myc-Rac1b remained unaffected (Fig. 4A). Apparently, Rho-GDI is unable to FIG. 4. Rho-GDI down-regulates Rac1 but not Rac1b. A, exogenous Rac1b remains active in the presence of Rho-GDI. SW480 cells were co-transfected with Myc-Rac1, Myc-Rac1b, and increasing amounts of Rho-GDI. Subsequently, the fraction of activated GTPases was isolated by pull-down assays. Upon co-expression of Rho-GDI, Rac1b was still detected in the active GTP-bound form, whereas Rac1 was not. B, endogenous Rac1b is not affected by Rho-GDI. Rho-GDI was transfected into HT29 cells expressing endogenous Rac1 and Rac1b. Only endogenous GTP-bound Rac1b but not endogenous Rac1 could be isolated by pull-down assays in the presence of Rho-GDI. Apparently Rho-GDI down-regulates the activity of Rac1 but not of Rac1b. EV, empty vector; NT, not transfected.

FIG. 5. Rho-GDI binds Rac1 but not Rac1b.
A, Rho-GDI co-immunoprecipitates with Rac1 but not Rac1b. SW480 cells were co-transfected with Myc-Rac1 and Myc-Rac1b and lysates immunoprecipitated (IP) with anti-Myc. When co-expressed, Rho-GDI was co-precipitated by Myc-Rac1 but not by Myc-Rac1b. B, Rho-GDI cannot displace Rac1b from the plasma membrane. Confocal immunofluorescence images showing exogenous wild type GFP-Rac1 and GFP-Rac1b co-localizing at the plasma membrane with ␤-catenin (␤-Cat) in HT29 cells. Upon co-expression of Rho-GDI, GFP-Rac1 is released in the cytoplasm, whereas GFP-Rac1b remains localized at the plasma membrane. Bars, 10 m. down-regulate the activation of Rac1b. To support this observation, Rho-GDI was overexpressed in HT29 cells or HeLa cells, which contain endogenous levels of both Rac1 and Rac1b (see Fig. 1B). Upon introduction of Rho-GDI, endogenous active Rac1b could still be isolated from these cells, whereas endogenous active Rac1 was no longer detectable (Fig. 4B, shown for  HT29). These data suggest that Rho-GDI somehow is unable to down-regulate the activation state of Rac1b.
We therefore determined whether the extra 19 amino acids in Rac1b could physically interfere with binding of Rho-GDI. From the available crystal structures of Rho-GDI complexed with Rac or Cdc42 (36 -37), it appears that 19 additional amino acids following the switch II region might disturb the binding to the C-terminal sandwich domain of Rho-GDI. To address this question, we transfected HA-Rho-GDI together with either Myc-Rac1 or Myc-Rac1b into SW480 cells and immunoprecipitated with either anti-Myc or anti-HA antibodies. Indeed, Rho-GDI could only be co-immunoprecipitated with Rac1 but not with Rac1b (Fig. 5A), and vice versa the immunoprecipitation of HA-Rho-GDI contained Myc-Rac1 but not Myc-Rac1b (data not shown). From these studies we concluded that Rho-GDI is unable to interact with Rac1b, which explains its inability to down-regulate Rac1b activity. The impaired binding of Rac1b to Rho-GDI might thus contribute to the high activation status of Rac1b found in cells.
Besides inhibition of nucleotide exchange, Rho-GDI can also retain Rho GTPases in an inactive cytosolic complex. We therefore analyzed the intracellular localization of Rac1 and Rac1b in HT29 cells following overexpression of Rho-GDI. As shown in Fig. 5, Rac1 localized to the plasma membrane at sites of cell-cell contact but became cytoplasmic following co-expression of Rho-GDI, consistent with studies reported previously (38 -39). In contrast, Rac1b localized at the plasma membrane but remained at the plasma membrane, even in the presence of high amounts of Rho-GDI (Fig. 5B). These data are consistent with the inability of Rho-GDI to associate with Rac1b and illustrate in vivo that Rho-GDI is unable to bind Rac1b. The impaired Rho-GDI binding will favor a more permanent localization of Rac1b at the plasma membrane, where it can be continuously activated by membrane-localized GEFs.
Rac1b Differs in Downstream Signaling-We next asked whether the high activation status that was found for endogenous Rac1b in vivo could lead to increased Rac-mediated downstream signaling. Hallmark features of activated Rac1 are the FIG. 6. Rac1b downstream signaling is selective. A, active Rac1b does not induce lamellipodia formation. SW480 cells were transfected with either constitutively activated Myc-Rac1b-Q61L or Myc-Rac1-Q61L. Cells were fixed 12 h after transfection and double-labeled with anti-Myc to visualize transfected cells and with phalloidin-TRITC to reveal the actin cytoskeleton. Lamellipodia formation is only seen in colon cells expressing constitutively active Rac1 but not Rac1b. Note that both proteins localize at the plasma membrane. B, activated Rac1b does not activate PAK. SW480 cells were co-transfected with FLAG-tagged PAK and either empty vector (EV) or Myc-tagged Rac1 or Rac1b mutants, as indicated. Cells were lysed 16 h after transfection and analyzed by Western blot with anti-Myc antibodies to visualize the transfected Rac mutants, with anti-FLAG antibodies to reveal equal quantities of PAK in the cells, and with anti-phospho-PAK1 to detect activation of PAK. C, activated Rac1b does not activate JNK. SW480 cells were co-transfected with FLAG-JNK and either empty vector (EV) or MKK4 as positive control, or Myc-Rac1 or Myc-Rac1b mutants, as indicated. Cells were lysed 16 h after transfection and analyzed by Western blotting with anti-Myc antibodies to visualize transfected Rac mutants, with anti-FLAG antibodies to reveal equal quantities of JNK in the cells, and with anti-active JNK to detect activation of JNK. D, activated Rac1b does not interact with PAK. SW480 cells were co-transfected with Myc-tagged PAK and GFP-tagged Rac1 or Rac1b mutants, as indicated. Cells were lysed 16 h after transfection and lysates immunoprecipitated (IP) with anti-GFP. Whereas active GFP-Rac1 co-precipitated well with PAK, the interaction between Rac1b and PAK was not significant.
induction of lamellipodia formation, as well as the activation of the protein kinases PAK and JNK (1). As described for fibroblasts, overexpression of Rac1-Q61L in SW480 colon carcinoma cells induced lamellipodia symmetrically along the entire cell surface (Fig. 6A). Surprisingly, expression of wild type Rac1b or Rac1b-Q61L did not lead to significant induction of lamellipodia in SW480 cells, although the protein was localized at the plasma membrane, preferentially at sites in contact to neighboring cells (Fig. 6A). Apparently, Rac1b is unable to induce cytoskeletal changes similar to those found with activated Rac1.
In order to further investigate Rac-mediated activation of downstream protein kinases, the active Q61L and inactive T17N mutants of both Rac1 and Rac1b were co-expressed in SW480 cells with either PAK1 or JNK. Subsequently, cell lysates were analyzed by Western blotting using antibodies specific for the activated, phosphorylated kinases. Activated Rac1 stimulated PAK (Fig. 6B) and JNK (Fig. 6C) activation, whereas the dominant negative mutant Rac1-N17 did not, consistent with previous observations with these Rac mutants (7,8). In contrast, neither of the Rac1b mutants was able to induce any meaningful activation of these kinases (Fig. 6, B and C). Thus in contrast to Rac1, Rac1b is unable to activate PAK and to stimulate the JNK pathway. The inability of active Rac1b to activate PAK reflects impaired interaction because only active Rac1 was able to co-immunoprecipitate full-length PAK (Fig. 6D) Interestingly, analysis of other Rac-mediated signaling pathways revealed that both activated Rac1 and Rac1b are able to stimulate the NFB pathway. The transcription factor NFB becomes active following phosphorylation and subsequent degradation of its cytosolic inhibitor IB (40). As shown in Fig. 7A, expression of both Rac1-Q61L as well as Rac1b-Q61L induced IB phosphorylation, whereas the inactive N17 mutants of both Rac variants did not. In addition, expression of both Rac1-Q61L as well as Rac1b-Q61L induced translocation of the DNAbinding subunit NFB-p65 (RelA) from the cytosol into the cell nucleus (Fig. 7B). These data suggest that Rac1b is able to stimulate only a restricted subset of downstream signaling pathways that are mediated by Rac1. DISCUSSION The data described in this paper demonstrate that alternative splicing of the small GTPase Rac1 profoundly alters its regulation and downstream signaling properties. This difference is due an extra domain of 19 amino acids resulting from inclusion of exon 3b into the Rac1 mRNA. It is interesting to note that this alternative splicing event is specific for the human RAC1 gene because we did not find any homologous, additional exons in the human RAC2, RAC3, or CDC42 gene structures.
We found that alternative splicing affects the regulation of Rac1b activity. The inclusion of the extra domain prevents the binding to Rho-GDI, one of the regulators of Rac activity. As a consequence, Rac1b is unable to cycle between the plasma membrane and the cytoplasm. It has been shown that Rho-GDI binds Rac1 and masks the hydrophobic C-terminal region that is responsible for targeting Rac to the plasma membrane (36). Thus, Rho-GDI maintains Rac in the cytoplasm and must dissociate to allow Rac to translocate to the membrane and to interact with membrane-associated activators (36). By in vitro studies, it has been demonstrated that release of Rho-GDI is required for translocation of Rac1 to plasma membranes and for the subsequent interaction with exchange factors, such as Tiam1 (41). In addition, integrins induce spatially controlled Rac-effector coupling in vivo by directing Rac1 to membranes and dissociating it from Rho-GDI (42). Because Rac1b is unable to interact with Rho-GDI, it is more persistently localized at the plasma membrane. This leaves the Rac1b variant in a preferred spatial position to become activated. The lack of binding to Rho-GDI, and the increased GTP-GDP cycling rate of recombinant Rac1b as demonstrated in vitro (24) most likely explain the high activation state found for endogenous Rac1b in cells. Even at low amounts of endogenous Rac1b protein in FIG. 7. Rac1 as well as Rac1b activate the NFB pathway. A, both activated Rac1 and Rac1b induce phosphorylation of IB. Mouse NIH 3T3 cells were co-transfected with IB and either empty vector (EV) or Myc-Rac1-Q61L, Myc-Rac1-N17, Myc-Rac1b-Q61L, or Myc-Rac1b-N17. Cells were lysed 16 h after transfection, and IB phosphorylation was visualized following Western blotting using a phosphospecific anti-IB antibody. Note that both active Rac1 and Rac1b induced IB phosphorylation, whereas their cognate dominant negative mutants were ineffective. B, activated Rac1b induces nuclear translocation of transcription factor NFB-p65. Mouse NIH 3T3 cells were co-transfected with pcDNA3-HA-RelA and either empty vector or GFP-Rac1-Q61L or GFP-Rac1b-Q61L. Cells were fixed 22 h after transfection and labeled with anti-HA antibodies to visualize RelA expression. Both active Rac1 and Rac1b induced nuclear translocation of RelA. cells, a considerable portion is present in an activated state and may contribute to specific Rac-mediated downstream signaling.
Our in vivo data also demonstrate that Rac1b retained the ability to cycle between the GDP-and GTP-bound form and rather exists in a high steady-state activation level. Recombinant Rac1b was shown previously in vitro to have intrinsic GTPase activity, which is accelerated upon addition of recombinant GAP protein (24). 3 In addition, we found that Bcr downregulates Rac1 and Rac1b in vivo equally well. Moreover, inhibition of the activity of cellular GEFs revealed that the activation of Rac1b depends on continuous action of GEFs. More specifically, the Rac activator Tiam1 associates with both Rac1 and Rac1b and is able to activate both proteins. All these data strongly suggest that the activity of Rac1b can be regulated by GEFs and GAPs and that Rac1b does not represent a constitutive activated splice variant of Rac1.
Intriguingly, the extra 19 amino acids in Rac1b appear to confer selectivity in downstream Rac signaling properties. In cultured cells, we found that activated Rac1b could neither stimulate formation of lamellipodia or activation of the protein kinase PAK nor the JNK pathway. In this respect it should be noted that the PAK-CRIB domain used in the pull-down assays binds equally to Rac1 and Rac1b, whereas full-length PAK does not significantly associate with Rac1b (Fig. 6D). Previous work has indicated that CRIB domain-containing fragments of effectors such as WASP or PAK are necessary for specific binding to active Rac or Cdc42. However, a productive interaction with the corresponding full-length proteins requires additional binding to regions outside this motif (43). Interestingly, activated Rac1b and Rac1 induced the phosphorylation of IB and nuclear translocation of NFB-p65 to the same extent, suggesting that the activation of Rac1b in cells promotes only specific downstream signaling pathways. Many downstream signaling pathways of Rac have been described that are independent of the ability of Rac1 to induce cytoskeletal changes or to activate PAK or the Jun kinase mitogen-activated protein kinase cascade (44 -48). These pathways often involve transcriptional regulation of genes. So far, the expression of Rac1b has been found in breast and colon tumors. This might indicate that this splice variant somehow facilitates the growth or survival of these epithelial tumors. Currently, we are studying which of the many Rac-specific downstream signaling pathways are activated by Rac1b and which of those pathways could play a role in the formation or progression of epithelial tumors.