Signaling Mediated by the Closely Related Mammalian ... - CiteSeerX

Vol. 12, 157–167, March 2001

Cell Growth & Differentiation

Signaling Mediated by the Closely Related Mammalian Rho Family GTPases TC10 and Cdc42 Suggests Distinct Functional Pathways1 Gretchen A. Murphy, Stephanie A. Jillian, David Michaelson, Mark R. Philips, Peter D’Eustachio, and Mark G. Rush2

salicylate. These findings point to distinct pathways in which TC10 and Cdc42 may act and distinct modes of regulation of these proteins.

Departments of Biochemistry [G. A. M., S. A. J., P. D., M. G. R.], Cell Biology [D. M., M. R. P.], and Medicine [M. R. P.] and the Kaplan Cancer Center [P. D.], New York University School of Medicine, New York, New York 10016

Introduction

Abstract The mammalian Rho family GTPases TC10 and Cdc42 share many properties. Activated forms of both proteins stimulate transcription mediated by nuclear factor ␬B, serum response factor, and the cyclin D1 promoter; activate c-Jun NH2-terminal kinase; cooperate with activated Raf to transform NIH-3T3 cells; and, by a mechanism independent of all of these effects, induce filopodia formation. In contrast, previously reported differences between TC10 and Cdc42 are not striking. We now present studies of TC10 and Cdc42 in cell culture that reveal clear functional differences: (a) wild-type TC10 localizes predominantly to the plasma membrane and less extensively to a perinuclear membranous compartment, whereas wild-type Cdc42 localizes predominantly to this compartment and less extensively to the plasma membrane; (b) expression of Rho guanine nucleotide dissociation inhibitor ␣ results in a redistribution of wild-type Cdc42 to the cytosol but has no effect on the plasma membrane localization of wild-type TC10; (c) TC10 fails to rescue a Saccharomyces cerevisiae cdc42 mutation, unlike mammalian Cdc42; (d) dominant negative Cdc42, but not dominant negative TC10, inhibits neurite outgrowth in PC12 cells stimulated by nerve growth factor; and (e) activation of nuclear factor ␬B-dependent transcription by Cdc42, but not by TC10, is inhibited by sodium

Received 7/12/00; revised 11/30/00; accepted 1/25/01. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 Supported by Grants CB-100 from the American Cancer Society (to M. G. R.), MCB-9630675 from the National Science Foundation (to P. D.), and AI 36224 and GM 55279 from the NIH (to M. R. P.). G. A. M. was supported by USPHS Training Grant GM 07827, and D. M. was supported by USPHS Training Grant GM 07308. DNA and protein sequence analyses were done at the Research Computer Resource of New York University Medical Center, which is supported by National Science Foundation Grant DIR-8908095. 2 To whom requests for reprints should be addressed, at Department of Biochemistry, New York University School of Medicine, 550 First Avenue, New York, NY 10016. Phone: (212) 263-5128; Fax: (212) 263-8166; Email: [email protected].

The small GTPase TC10 was identified nearly 10 years ago and, except for preliminary RNA expression experiments, had remained essentially uncharacterized until recently (1–3). On the basis of amino acid sequence, TC10 is most closely related to the Cdc42 and Rac members of the mammalian Rho family (67% identity to Cdc42 and 63% identity to Rac). TC10 localizes predominantly to the plasma membrane and exhibits several of the cellular functions characterized for Cdc42, Rac1, and RhoA. Similar to the action of Cdc42 on the actin cytoskeleton, activated (GTP-bound, GTPasedefective) TC10 induces the formation of filopodia, and, like Cdc42 and Rac1, activated TC10 stimulates JNK3 (2, 3). In common with Cdc42, Rac1, and RhoA, activated TC10 stimulates SRF- and NF-␬B-dependent transcription and cooperates with an activated form of the serine/threonine PK Raf in a transformation assay measuring synergistic focus formation. Moreover, dominant negative TC10 blocks oncogenic Ras focus forming activity, suggesting that TC10 function is required for full Ras-mediated transformation (2). Also in common with Cdc42, Rac1, and RhoA, activated TC10 stimulates transcription from a cyclin D1 promoter, providing a more direct link to regulation of the cell cycle (2). From these results, it has been concluded that TC10 is functionally most closely related to Cdc42, consistent with observed sequence identities, and it has been suggested that TC10 may play a pivotal role in actin cytoskeleton rearrangement, activation of gene expression, and oncogenic transformation (2). Subcellular localization studies with TC10 proteins indicate that wild-type and activated (75L) TC10 are distributed in both the plasma membrane and intracellular membranes to varying extents. Plasma membrane localization of TC10 coincides with the distribution of cortical filamentous actin (2), and cortical filamentous actin is required for TC10 plasma membrane localization, as evidenced by disruption of fila-

3 The abbreviations used are: JNK, c-Jun NH2-terminal kinase; GFP, green fluorescent protein; GEF, guanine nucleotide exchange factor; I␬B, inhibitor of nuclear factor ␬B; IKK, I␬B kinase; IL, interleukin; MEKK1, mitogen/extracellular signal-regulated protein kinase kinase 1; NaSal, sodium salicylate; NF-␬B, nuclear factor ␬B; NGF, nerve growth factor; NIK, NF-␬B-inducing kinase; PAK, p21-activated protein kinase; PK, protein kinase; RhoGDI, Rho guanine nucleotide dissociation inhibitor; SRF, serum response factor; TNF, tumor necrosis factor; MLK, mixed lineage kinase; WASP, Wiskott Aldrich syndrome protein; RSK, ribosomal subunit kinase; HA, hemagluttinin; HS, horse serum; FBS, fetal bovine serum.

157

158

TC10 and Cdc42 Signaling

mentous actin with the marine sponge toxin Latrunculin B.4 The presence of a COOH-terminal CAAX (cysteine-aliphaticaliphatic-any) amino acid motif in TC10 indicates that, like other members of the Rho family, the protein is processed by isoprenylation of the cysteine, followed by proteolytic removal of the three COOH-terminal amino acids and carboxymethylation. This processing suggests that TC10 associates with membranes for some period during its normal, regulated function. Indeed, examination of exogenously expressed COOH-terminal-deleted proteins showed that the CAAX motif is essential for the membrane localization observed for both the wild-type and activated proteins (2). As might have been expected from a drastic alteration of normal cellular localization, the activated/COOH-terminal-deleted mutant is defective in TC10-mediated activation of filopodia, gene expression, and transformation, indicating that COOHterminal processing and membrane localization are essential for TC10 signaling functions (2). Other Rho family proteins, such as Cdc42, localize to cellular membranes, sites of cell adhesion, and regions of filamentous actin (4, 5). The subcellular localization of Rho family proteins is an important aspect of their regulation, and Ccd42, Rac1, and RhoA can be regulated by RhoGDIs in a mechanism postulated to involve both redistribution of the GTPases and the direct inhibition of guanine nucleotide exchange (Refs. 5– 8; reviewed in Ref. 9). Guanine nucleotide exchange is the process by which the GTPase releases bound GDP and, in a GTP-rich cellular environment, binds GTP. The process is stimulated by a GEF. Upstream regulators of TC10 and extracellular stimuli that activate TC10 signaling are currently unknown. However, several GEFs and extracellular stimuli are known to activate Cdc42, but none of the many Rho family GEFs examined, such as the Cdc42- and RhoA-stimulating GEF, Dbs, enhance nucleotide exchange on TC10.5 Rho family GTPases and some of their target signaling molecules have been implicated in the NGF-stimulated differentiation process in PC12 cells (for review, see Ref. 10). Undifferentiated PC12 cells derived from a rat pheochromocytoma grow exponentially in culture and maintain a small, round morphology. Treatment with NGF results in a halt in proliferation, followed by characteristic morphological changes (cell body growth, flattening, and spreading) and eventually by extension of neurites (long, actin-containing extensions; for review, see Ref. 11). Activated RhoA induces neurite retraction (12, 13), and dominant negative forms of Rac1 and Cdc42 inhibit NGF-induced neurite outgrowth (13, 14), and although activated Cdc42 does not induce neurites in the absence of NGF, it does stimulate the other characteristic morphological changes associated with the differentiation program (14). Thus, it appears that RhoA controls neurite retraction, whereas Rac1 and Cdc42 control neurite outgrowth. Whereas no role for TC10 in the NGF pathway has been reported, expression of high levels of activated

4 5

G. A. Murphy, unpublished observations. K. Rossman and C. Der, personal communication.

TC10 in neuronal cells does promote elongation of previously existing neurites (15). The transcription factor NF-␬B plays an important regulatory role in expression of genes whose products are involved in the inflammatory response and control of cell death (reviewed in Refs. 16 and 17). RhoA, Rac1, Cdc42, and TC10 have all been implicated in the activation of NF-␬B (2, 18). However, the pathways by which these signaling molecules control NF-␬B activation are poorly understood. Normally, NF-␬B is sequestered in the cytosol in an inactive state by an inhibitory subunit, I␬B. Phosphorylation of I␬B, most often by IKK, leads to its degradation by the proteasome and the release of NF-␬B, which is then free to translocate to the nucleus and activate transcription. IKK, in turn, is a heterotrimer composed of ␣, ␤, and ␥ subunits, and phosphorylation of the ␤ subunit, IKK␤, appears to be required for activation. This phosphorylation can be mediated by MEKK1, an upstream activator of JNK and p38 mitogenactivated protein kinases, by NIK, by particular isoforms of PKC, or by PKB. Both MEKK1 and NIK are activated as a result of receptor-mediated cell stimulation by proinflammatory cytokines such as TNF-␣ and IL-1 or by cellular stresses (for reviews, see Refs. 17 and 19). Recent evidence suggests that the nonsteroidal antiinflammatory drugs acetyl salicylate (aspirin) and NaSal reduce inflammation at least in part by suppressing TNF-␣- or IL-1-stimulated NF-␬B activation (20 –22). This inhibition of NF-␬B is independent of the more thoroughly studied effects of salicylates on cyclooxygenase activity and prostaglandin synthesis. The mechanism of action of salicylates in the NF-␬B signaling cascade is controversial but appears to occur at or above IKK (22, 23). Interestingly, salicylates directly inhibit the kinase activity of IKK␤ in vitro (22). No TC10 homologue has been identified in yeast, but the budding yeast (Saccharomyces cerevisiae) homologue of mammalian Cdc42 plays an essential role in establishing and regulating the polarity of dividing cells, in which cell division occurs by budding of a new daughter from a mother cell (24). The processes of polarity establishment and budding are complex and highly regulated, involve actin remodeling into filaments that extend into the daughter cell, and require the activities of several small GTPases (for review, see Refs. 25 and 26). Briefly, the Bud1p GTPase plays a critical role in bud site selection, Cdc42p is essential for the formation of protein complexes at the mother-bud neck that drive bud initiation and eventual bud growth, the Rho1p GTPase is implicated in bud growth and elongation, and Cdc42p is also necessary for formation of the actomyosin contractile ring and septum at cytokinesis. A budding yeast temperature-sensitive mutation in CDC42 (cdc42-1ts) has been described. At a restrictive temperature of 37°C, mutant cells grow in size, replicate DNA, and complete nuclear division but arrest as large, multinucleate cells (27, 28). Shortly after the identification of the yeast CDC42 gene, two groups independently cloned two human isoforms (⬎95% amino acid identity) homologous to the yeast gene product and confirmed that they complement cdc42-1ts (29, 30). One isoform (Cdc42Hs) was isolated from placenta and is ubiquitously expressed, and the other isoform (G25K) was


isolated from and is expressed only in the brain. The human Cdc42 isoforms are ⬃80% identical to the Cdc42p yeast gene product (29, 30). Based on the relatively highly conserved effector domains of TC10 and Cdc42 and on previous findings using mutations in this domain (2), it is possible that TC10 exerts many of its biological effects via mechanisms similar to those used by Cdc42. This possibility is consistent with the observation that GTP-bound TC10 may bind directly to a variety of putative Cdc42 effectors, including PAK1, PAK2, PAK3, myotonic dystrophy kinase-related Cdc42-binding kinase ␣/␤, MLK2, neural WASP, and the recently identified Borg family of putative effectors (3, 31). Nevertheless, distinct differences have also been observed between TC10 and Cdc42. For example, TC10 stimulates formation of filopodia that are longer and more extensive than those stimulated by Cdc42 (2, 3); TC10 fails to bind the known Cdc42 effectors WASP, MLK3, and activated Cdc42-binding kinase 1 (3); TC10 interacts with the actinbinding protein profilin in a GTP-dependent manner, whereas Cdc42 binding to profilin is nucleotide independent;4 and p50Rho-GTPase-activating protein enhances the intrinsic GTP hydrolysis rate of TC10 but not that of Cdc42 (3). To further characterize the differences between TC10 and Cdc42, especially with regard to behavior in vivo, we have investigated (a) the cellular localizations of GFP-tagged constituents in both the presence and absence of overexpressed RhoGDI␣, (b) the comparative abilities of TC10 and Cdc42 to rescue a temperature-sensitive S. cerevisiae cdc42 mutation, (c) the requirements for these GTPases in NGF-stimulated neurite extension, and (d) the effect of NaSal, a known inhibitor of NF-␬B activation, on TC10- or Cdc42-mediated NF-␬B stimulation. Significant differences between TC10 and Cdc42 were observed in all of these investigations, suggesting that distinct functional pathways can be regulated by each of these GTPases.

Results Wild-Type TC10 Localizes Predominantly to the Plasma Membrane in Living Cells. It was reported previously, using epitope-tagged GTPase constructs in fixed cells, that wild-type TC10 protein localizes predominantly to the plasma membrane, whereas the activated (75L) and dominant negative (31N) mutant proteins distribute predominantly in the cytoplasm in perinuclear patterns that are reproducible for the three cell types tested (COS, HeLa, and BHK; Ref. 2). Activated (75L) TC10 binds GTP normally but is defective in GTP hydrolysis, whereas dominant negative (31N) TC10 does not bind GTP, binds GDP poorly, and is predicted to inhibit endogenous TC10 function by trapping GEFs. To confirm the distribution of TC10 observed in fixed cells, GFP-tagged wild-type or mutant TC10 was expressed in HeLa cells, and the localization of expressed protein was visualized in living cells without fixation. Expression of GFP alone resulted in cytoplasmic and nucleoplasmic fluorescence. In contrast, GFP-tagged wild-type TC10 localized predominantly to the plasma membrane and peripheral membrane extensions and less extensively to a perinuclear

Fig. 1. Localization of TC10 in living mammalian cells. HeLa cells actively growing in medium containing 10% calf serum were transfected with plasmids containing GFP-tagged constructs by LipofectAMINE PLUS reagent in serum-free medium, and at 6 h posttransfection, cells were transferred into medium with 10% serum. At 24 h posttransfection, cells were viewed with a Zeiss Axiophot fluorescence microscope, and digitized images were recorded.

membranous compartment (Fig. 1). The TC10 75L (activated) mutant localized to intracellular membranes in a perinuclear pattern similar to that observed in fixed cells, to the plasma membrane, and to peripheral membrane extensions (Fig. 1). The TC10 31N (dominant negative) mutant distributed in an intracellular particulate pattern (Fig. 1). Taken together, these data confirm and extend previous observations in fixed cells. Specifically, the data indicate that

159

160


Fig. 2. Effect of RhoGDI␣ on Rho family protein distribution in living cells. HeLa cells actively growing in medium containing 10% calf serum were transfected with plasmids containing GFP-tagged GTPase constructs in the absence or presence of a 3-fold excess of a RhoGDI␣ expression construct by LipofectAMINE PLUS reagent in serum-free medium, and 6 h posttransfection, cells were maintained in medium containing 10% serum. At 24 h posttransfection, cells were viewed with a Zeiss Axiophot fluorescence microscope, and digitized images were recorded.

wild-type TC10 distributes predominantly to the plasma membrane, 31N (dominant negative) mutant protein distributes predominantly to a perinuclear compartment, and 75L (activated) mutant protein partitions between the two compartments. Under similar conditions, wild-type, dominant negative, and activated Cdc42 all distribute predominantly to endomembranes (especially Golgi) and less extensively to the plasma membrane and cytosol (5). The plasma membrane localization of wild-type TC10 is strikingly different from the extensive Golgi localization observed for wild-type Cdc42 in HeLa (see Fig. 2) and other cell types (5). Moreover, this predominant plasma membrane TC10 localization remains clear even when GFP-TC10 is expressed at very low levels (5). In contrast, low-level expression of GFP-Cdc42 results in decreased plasma membrane and endomembrane fluorescence and increased localization to the cytosol (5). As shown in the following section, the concentration dependence of Cdc42, but not TC10, distribution, probably reflects the fact that Cdc42, but not TC10, can interact with and be solubilized by RhoGDI␣. The Subcellular Distribution of Cdc42, but not TC10, Is Altered by RhoGDI␣. Rho GTPases have been shown to be regulated by RhoGDI by a mechanism thought to involve prenylation-dependent binding and inhibition of guanine nucleotide exchange (32). There are three known isoforms of RhoGDI: (a) ␣/1; (b) ␤/2; and (c) ␥/3. RhoGDI␣ is ubiquitously expressed and interacts strongly with Cdc42, Rac, and

RhoA; whereas the ␤/2 and ␥/3 isoforms are restricted both in their expression and the range of Rho GTPases with which they interact (reviewed in Ref. 9; see Refs. 33 and 34). Because of its ubiquitous expression, we chose to test the effect of the ␣ isoform on the cellular localization of Cdc42, Rac, and TC10. In living mammalian cells, the subcellular distributions of GFP-tagged Rac1 and GFP-tagged Cdc42 proteins are dramatically altered by coexpression with RhoGDI␣ (Fig. 2; Ref. 5). As shown in the first column of photomicrographs of Fig. 2, in the absence of exogenous RhoGDI␣ expression, GFPRac1 localizes to the plasma membrane, whereas GFPCdc42 localizes to intracellular membranes in a perinuclear pattern and to the plasma membrane and peripheral membrane extensions. Coexpression of a 3-fold excess of RhoGDI␣ with GFP-Rac1 or GFP-Cdc42 results in a dramatic and nearly complete redistribution of these GTPases to the cytoplasm and nucleoplasm (Fig. 2, second column). In contrast, coexpression of GFP-TC10 with a 3-fold excess of RhoGDI␣ did not alter the plasma membrane localization of TC10 (Fig. 2). Identical results were obtained for Rac1, Cdc42, and TC10 in Madin-Darby canine kidney cells (5). Consistent with this finding, GFP-Cdc42 and GFP-Rac1 coimmunoprecipitated with RhoGDI␣ from lysates of MadinDarby canine kidney cells coexpressing GFP-Cdc42 or GFPRac1 and a 3-fold excess of RhoGDI␣. In contrast, TC10 failed to interact with RhoGDI␣ in this in vivo binding assay (5). In addition, stimulation of SRF-dependent transcription by activated Cdc42 (12V) is inhibited strongly by expression of RhoGDI␣, whereas stimulation by activated TC10 (75L) is unaffected (5). The results presented in this section demonstrate that, in contrast to Cdc42 and Rac1, TC10 does not interact with RhoGDI␣ in vivo. TC10 Does Not Rescue the Growth Defect of a Temperature-sensitive cdc42 Mutation in S. cerevisiae. No TC10 homologue exists in S. cerevisiae, but the apparent functional similarity of TC10 and Cdc42 in mammalian cells prompted us to examine the ability of TC10 to rescue the growth defect of cdc42ts mutant S. cerevisiae. Expression of wild-type TC10 or Cdc42 (human placental isoform) was driven by the constitutive alcohol dehydrogenase (ADH1) promoter in the cdc42ts or parental (CDC42) yeast strain. At the permissive temperature (23°C), the cdc42ts and parental strains grow and divide normally, but at the restrictive temperature (37°C), only the parental strain divides (Fig. 3A). As shown in Fig. 3A, expression of mammalian Cdc42 rescues the growth defect of the mutant at the restrictive temperature (37°C), whereas expression of TC10 does not. Immunoblotting with highly specific antibodies to TC10 (our antipeptide antibody) and human Cdc42 (Santa Cruz Biotechnology) confirms expression of both exogenous TC10 and Cdc42 at the restrictive temperature (Fig. 3B). Immunofluorescence studies indicate that S. cerevisiae Cdc42p localizes to the plasma membrane, especially near the site of bud emergence and at the tips and sides of the emerging bud (35). To determine whether the failure of TC10 to rescue cdc42ts yeast could be due to inappropriate distribution of the protein, subcellular localization of TC10 in S.


Fig. 3. Lack of rescue by TC10 of the growth defect of a temperature-sensitive cdc42 mutation in S. cerevisiae. A, a cdc42 temperature-sensitive yeast strain or the CDC42 parental strain was transformed with a plasmid that expresses the LEU2 gene and either TC10 or human Cdc42 driven by the constitutive yeast alcohol dehydrogenase (ADH1) promoter. Cells were selected for leucine prototrophy at 23°C (permissive temperature) for 2– 4 days and then examined for rescue of the cdc42 temperature-sensitive growth defect at 37°C (restrictive temperature). B, TC10 or Cdc42 transformants were grown under selection for leucine prototrophy in liquid culture to A ⬇ 1. Cells were then pelleted, washed with PBS plus protease inhibitors, and resuspended in 10 mM sodium azide, and extracts were prepared by rapid boiling and glass bead rupture of the cell wall. Equal amounts of extract were resolved by SDS-PAGE, transferred to nitrocellulose, and immunoblotted for TC10 or Cdc42 using our anti-TC10 antibody or a commercially available anti-Cdc42 antibody (Santa Cruz Biotechnology), and protein expression was detected by chemiluminescence.

Fig. 4. Subcellular localization of TC10 in live S. cerevisiae cells. Yeast cells were transformed with a plasmid containing a LEU2 gene and a GFP-tagged TC10 construct whose expression is driven by the constitutive ADH1 promoter. Transformants were selected by leucine prototrophy at 30°C and grown to A ⬇ 1. Cells were then pelleted, washed, and resuspended in PBS plus protease inhibitors; mounted with Vectashield; and viewed with a Zeiss Axiophot fluorescence microscope.

cerevisiae was examined in live cells by visualizing expression of a GFP-TC10 fusion protein. As shown in Fig. 4, GFP-TC10 localizes predominantly to the plasma membrane and a perinuclear region in parental and cdc42ts cells. The data presented in this section indicate that TC10 does not rescue the cdc42–1ts mutation in S. cerevisiae, although it is expressed and apparently localized to the necessary site of action. These results suggest that TC10 and Cdc42 are

not functionally redundant between mammals and yeast and that TC10 cannot substitute for Cdc42 function. Cdc42, but not TC10, Is Required for NGF-stimulated Neurite Extension in PC12 Cells. Because Cdc42 and Rac1 are implicated in controlling NGF-stimulated neurite outgrowth, we sought to determine whether TC10 also functions in this signaling cascade. Undifferentiated PC12 cells actively growing in culture were transiently transfected with plasmids expressing wild-type, dominant negative, or activated forms of epitope-tagged TC10 or Cdc42 and subsequently treated with NGF for 72 h. Neurite formation in transfected cells was visualized by fixing, permeabilizing, and staining cells with anti-epitope antibody and rhodamine phalloidin (to observe F-actin in neurites). Transfected cells were classified visually as expressing high, intermediate, or low levels of Cdc42 or TC10, and only high expressors were used for the quantitation described here. As shown in Fig. 5A, 53% of nontransfected cells treated with NGF form neurites. NGF-stimulated cells transfected with dominant negative TC10 (31N) exhibit relatively normal neuritogenesis, whereas dominant negative Cdc42 (17N) interferes dramatically with neurite formation (10% of transfected cells have neurites; Fig. 5A). However, NGF-treated cells expressing either dominant negative Cdc42 or dominant negative TC10 retain the ability to halt proliferation and exhibit morphological changes, such as flattening, similar to nontransfected cells treated with NGF (Fig. 5B). Wild-type or activated TC10 and Cdc42 do not dramatically alter NGF-stimulated neurite outgrowth, nor do they stimulate neurite outgrowth in PC12

161

162


Fig. 5. Quantitation of the effects of dominant negative TC10 and Cdc42 proteins on neuritogenesis of PC12 cells. PC12 cells grown on collagencoated dishes were mock transfected or transfected with T7-epitopetagged TC10 or HA-epitope-tagged Cdc42 constructs using LipofectAMINE 2000. Six h posttransfection, cells were transferred to medium containing 20% serum, cultured for 18 h, and then transferred to medium containing 1% serum plus 50 ng/ml NGF. Every 24 h thereafter, half the culture medium was removed and replaced with fresh medium plus 50 ng/ml NGF. At 96 h posttransfection, cells were fixed, permeabilized, and stained with anti-T7 antibody or anti-HA antibody and rhodamine-phalloidin. Cells were mounted with Vectashield and viewed with a Zeiss Axiophot fluorescence microscope. For each construct, a total of ⬃150 cells from multiple transfections were analyzed for neurite outgrowth and morphological changes. Neurites are classified as at least two cell bodies in length, and morphological changes are classified as cellular flattening and the formation of small extensions. In the presence of NGF, nontransfected cells exhibit a high level of neurite outgrowth (53% of the cells) and morphological change (⬃36% of the cells). The columns represent the percentage of transfected (TC10 31N, Cdc42 17N) or nontransfected cells (None) with neurites (A) or morphological changes (B).

cells that are not treated with NGF. In addition, activated Cdc42 (61L), but not activated TC10 (75L), stimulates morphological changes (cell body growth, flattening, and spreading) in a small percentage of non-NGF-treated cells (data not shown). Taken together, the results presented in this section indicate that TC10 is not a component of NGF-stimulated neurite outgrowth in PC12 cells. Dominant negative TC10, unlike

Fig. 6. Effects of NaSal on activation of NF-␬B transcription by TC10, Cdc42, or MEKK1. COS cells were cotransfected with plasmids expressing wild-type or activated Cdc42, wild-type or activated TC10, or a constitutively active MEKK1 and a reporter plasmid containing NF-␬B binding elements fused to a firefly luciferase gene. Six h posttransfection, cells were maintained in serum-free medium for 20 h and then transferred to fresh serum-free medium in the absence (⫺) or presence (⫹) of 20 mM NaSal for 6 h. Samples to be tested for recovery of NF-␬B activation were transferred to fresh medium in the absence (⫺) of NaSal for an additional 20 h. Cells were then harvested, lysed, and examined for luciferase expression. ⫺, mock NaSal treatment; ⫹, 6-h NaSal treatment; ⫺/⫺, mock NaSal treatment followed by a change to fresh serum-free medium and incubation for an additional 20 h; ⫹/⫺, NaSal treatment for 6 h followed by a change to fresh serum-free medium and incubation for an additional 20 h. The columns represent the average percentage of activation relative to TC10 75L (for TC10 and TC10 75L samples), Cdc42 61L (for Cdc42 and Cdc42 61L samples), or MEKK1 (for MEKK1 samples) in the absence of NaSal for three experiments, and the error bars show the range. The level of GTPase or MEKK1 expression was determined by immunoblotting to be similar for all samples.

dominant negative Cdc42, does not interfere with NGF-stimulated neurite extension, and activated TC10, unlike activated Cdc42, does not induce morphological changes in cells not treated with NGF. Stimulation of NF-␬B-dependent Transcription by Activated Cdc42, but not by Activated TC10, Is Inhibited by NaSal. Previous studies indicated that TNF-␣ and/or IL-1 signaling to NF-␬B activation involves activation of IKK and phosphorylation of I␬B and is sensitive to the nonsteroidal anti-inflammatory compound NaSal. To explore the possibility that TC10- or Cdc42-mediated activation of NF-␬B occurs by a pathway sensitive to salicylate inhibition, we expressed epitope-tagged TC10 75L, Cdc42 61L, or NH2-terminal truncated (constitutively active, ⌬) MEKK1 with an NF-␬B luciferase reporter construct in COS cells. Cells were cultured in 20 mM NaSal for 6 h before harvesting. As shown in Fig. 6, activation of NF-␬B by Cdc42 61L and ⌬MEKK1 is decreased ⬃50% by treatment with NaSal. However, activation by TC10 75L is not greatly affected by such treatment. As also shown in Fig. 6, NF-␬B activation was fully restored by


by TC10 is not, the data presented in this section demonstrate that this inhibition is reversible and that NaSal treatment is not nonspecifically toxic. The results also confirm previous reports for MEKK1 that show that NaSal-mediated inhibition of NF-␬B activation is independent of inhibition of cyclooxygenase activity and prostaglandin synthesis (22).

Discussion

Fig. 7. Specificity of the inhibition of NF-␬B-dependent activation by NaSal. COS cells were cotransfected with plasmids expressing wild-type or activated (75L) TC10 (A and B), wild-type or activated (61L) Cdc42 (A), or a constitutively active MEKK1 (B) and a reporter plasmid containing NF-␬B binding elements fused to a firefly luciferase gene. Six h posttransfection, cells were maintained in serum-free medium for 20 h and then transferred to fresh serum-free medium in the absence (⫺) or presence (⫹) of 1 mM acetaminophen (A) or 5 ␮M indomethacin (B) for 6 h. Cells were then harvested, lysed, and examined for luciferase expression. The columns represent the average percentage of activation relative to TC10 75L (for TC10 and TC10 75L samples), Cdc42 61L (for Cdc42 and Cdc42 61L samples), or MEKK1 (for MEKK1 samples) in the absence of drug treatment for two experiments, and the error bars show the range. The level of GTPase or MEKK1 expression was determined by immunoblotting to be similar for all samples.

removal of NaSal followed by incubation of cells in serumfree medium for an additional 24 h. Furthermore, treatment of cells with levels of acetaminophen (Fig. 7A) or indomethacin (Fig. 7B) sufficient to inhibit cyclooxygenase activity had no effect on TC10-, Cdc42-, or MEKK1-mediated NF-␬B activation, confirming that the NaSal effects on Cdc42 61L and ⌬MEKK1 signaling are independent of cyclooxygenase activity. In addition to showing that the activation of NF-␬B by Cdc42 or MEKK1 is inhibited by NaSal, whereas activation

TC10 and Cdc42 share many properties, especially when assayed by studying mammalian cells transfected with activated (GTPase-defective) mutants. Both proteins induce the formation of filopodia, act synergistically with Raf in transformation assays, and stimulate transcription mediated by JNK, SRF, NF-␬B, or the cyclin D1 promoter. They also interact in vitro and/or in vivo with some of the same known or postulated effectors (3, 31). Nevertheless, a few clear differences between these GTPases are known, including the length and extent of induced filopodia (greater for TC10 than for Cdc42) and differential interactions with some effectors and regulators such as WASP, MLK3, activated Cdc42-binding kinase 1, profilin, and p50Rho-GTPase-activating protein. These data support the hypothesis that TC10 and Cdc42 are functionally distinct. Although they stimulate the same process, such as NF-␬B-dependent transcription, they need not do so through the same overall pathway. In addition, for some processes such as NGF-stimulated neurite extension, only one of the GTPases may be required. The data presented in this study represent the results of a series of experiments performed to further examine the similarities and differences between TC10 and Cdc42, with an emphasis on function and regulation in vivo. To begin, we examined cellular localization in HeLa cells using GFPtagged fusion proteins. The results shown in Fig. 1 confirm our previous findings using immunostaining of fixed cells and extend them to include plasma membrane localization of activated TC10. Wild-type TC10 localizes predominantly to the plasma membrane, activated TC10 (75L) localizes predominantly to the plasma membrane and a perinuclear membrane compartment, and dominant negative TC10 (31N) localizes predominantly to a perinuclear region. The results shown in Fig. 2, which compare data obtained from cells expressing either wild-type TC10, Cdc42, or Rac1, demonstrate two highly reproducible findings. First, whereas TC10 and Rac1 localize predominantly to the plasma membrane, Cdc42 localizes to both the plasma membrane and a perinuclear membrane compartment. Second, whereas overexpression of RhoGDI␣ clearly causes the solubilization and relocalization of Cdc42 and Rac1 to the cytoplasm and nucleoplasm, RhoGDI␣ expression has no major effect on the distribution of TC10. These results, along with the observations that GFP-Cdc42 and GFP-Rac1, but not GFP-TC10, coimmunoprecipitate with RhoGDI␣ in cell extracts and that SRF stimulation by activated GFP-Cdc42 (12V) but not by activated GFP-TC10 (75L) is inhibited by RhoGDI␣ (5),4 indicate that this ubiquitously expressed RhoGDI isoform is unlikely to be a regulator of TC10 function. Whether TC10 is regulated by tissue-specific RhoGDIs, such as GDI␤ or GDI␥, or whether its activity is regulated at another level remains to be examined.

163

164


In S. cerevisiae, Cdc42p is an essential protein required for actin polymerization events involved in bud initiation and cytokinesis, and human Cdc42 can rescue the growth and budding defects of a yeast cdc42ts mutation (Fig. 3). Because both the yeast and human proteins stimulate initiation of actin polymerization in a variety of in vitro assays (25, 36, 37), this rescue is consistent with a conserved function for Cdc42 in actin dynamics. A TC10 homologue does not exist in S. cerevisiae, but the similar filopodial induction phenotypes of human Cdc42 and TC10 in animal cells (due presumably to the actin polymerizing activity of these proteins) prompted us to examine whether TC10 could mimic Cdc42 function well enough to rescue the yeast cdc42ts mutant. As shown in Figs. 3 and 4, it does not. This result indicates that the role played by Cdc42 in yeast requires interactions that are not compatible with TC10. For example, TC10 might not respond to yeast GEFs or be able to stimulate specific effectors. To distinguish between these two possibilities, we attempted to rescue cdc42ts yeast with activated Cdc42 (61L) or TC10 (75L) mutants, although activated Cdc42 expression in wild-type yeast was reported to be lethal (38). As expected, Cdc42 61L expression was lethal in both the parental and cdc42ts strains at the permissive temperature. Furthermore, at the nonpermissive temperature, Cdc42 61L was lethal in the parental strain and failed to rescue cdc42ts. In contrast, TC10 75L expression had no effect on the growth of either strain at either temperature (data not shown). The simplest explanation for these results is that overexpressed activated TC10, unlike Cdc42, is unable to overstimulate potentially lethal targets in yeast and that wild-type TC10 is unable to rescue cdc42ts due to its inability to stimulate an effector pathway. However, the fact that activated Cdc42 itself does not rescue cdc42ts indicates that this simple explanation should be treated with caution. The inability of TC10 to rescue cdc42ts yeast also suggests, albeit indirectly, that some essential regulators and/or effectors of these proteins, such as those noted earlier in this “Discussion,” are likely to differ in animal cells as well. From a broader perspective, the rescue failure also suggests that TC10 and Cdc42 are not functionally redundant proteins, a hypothesis that can be tested eventually by studying selective knockouts in mice. Rho family GTPases have been implicated in the differentiation, including neuritogenesis, of NGF-treated PC12 cells, and potential effectors and signaling pathways are being elucidated. Basically, Cdc42 and Rac1 appear to function in neurite extension, whereas RhoA induces retraction. The effect of RhoA on neurite retraction appears to be mediated by the effector Rho kinases that activate myosin-based contractility (39). The Cdc42 and Rac requirement for neurite extension appears to be mediated by the effector kinases myotonic dystrophy kinase-related Cdc42-binding kinase ␣ and PAK (14, 39) and perhaps by the mitogen-activated protein kinase family members p38 and JNK (40, 41), both of which are required for NGF-stimulated neuritogenesis. Also, as noted in “Results,” although activated Cdc42 and Rac do not induce PC12 cells to form neurites in the absence of NGF, dominant negative forms of these proteins inhibit neu-

rite formation induced by NGF. It should be noted, however, that the latter observation is consistent with both a direct role for Rac1 and Cdc42 in the NGF pathway and with the ability of the dominant negative forms of these proteins to inhibit an exchange factor required by another GTPase that is itself a component of the pathway. The above-mentioned findings suggested that Cdc42 and Rac1 are downstream components of a pathway initiated by the NGF receptor. The fact that actin polymerization is a major component of both neurite and filopodial extension, along with the observation that activated TC10 is a more effective filopodial inducer than Cdc42, prompted us to examine the role of TC10 in neuritogenesis. As reported in “Results” and in agreement with findings obtained for activated Cdc42, activated TC10 exhibited no ability to induce neurite extension in non-NGF-treated PC12 cells. However, in contrast with findings obtained for activated Cdc42, activated TC10 exhibited no ability to induce morphological changes such as spreading and flattening in such cells. Also in contrast with Cdc42, where expression of the dominant negative protein inhibited NGF-stimulated neurite extension by 80%, expression of dominant negative TC10 had little effect (Fig. 5). In addition, a role for TC10 in neurite retraction is highly unlikely because wild-type or activated TC10 had no effect on NGF-stimulated neurite outgrowth. Taken together, the results presented here suggest that TC10, unlike Cdc42, is not a component of the NGF-stimulated neuritogenesis pathway. It should also be noted that these results differ significantly from those obtained when examining the roles of Cdc42 and TC10 in activated Ras-induced transformation, where dominant negative Cdc42 and TC10 each reduce focus forming activity significantly (Ref. 2; reviewed in Ref. 42). Because of its role in mediating inflammation, NF-␬B and the mechanisms by which it is activated are areas of intense investigation. Several extracellular stimuli and cellular stresses activate NF-␬B through complex and interconnecting pathways. For example, TNF-␣ binds and activates its receptor to recruit additional signaling molecules that result in the activation of the serine/threonine kinase NIK. Alternatively, IL-1 binds and activates its receptor to recruit signaling molecules that in turn result in the activation of the serine/threonine kinase MEKK1. These two inflammatory signaling pathways then converge because both NIK and MEKK1 activate IKK, apparently through phosphorylation of its ␤-subunit, IKK␤. IKK, in turn, catalyzes the phosphorylation of I␬B, targeting it for degradation, and freeing NF-␬B to activate transcription of genes, many of which mediate the inflammatory response (reviewed in Refs. 17 and 19). Inflammation can also be mediated by eicosanoids, particularly the prostaglandins, whose synthesis depends on the activity of cyclooxygenases. Indeed, it has been known for nearly 30 years that the anti-inflammatory properties of aspirin and other salicylates can be attributed, at least in part, to their inhibition of cyclooxygenase activity. The inhibition of cyclooxygenases by salicylates is now well established, but it appears that the anti-inflammatory properties of these drugs are due to other mechanisms as well. Specifically, recent reports indicate that both the TNF-␣


and IL-1 signaling pathways to NF-␬B activation are blocked by salicylates (20 –22) and that the kinase activity of IKK␤ is inhibited by salicylates in vitro (22). As might be expected from the pathways described above, the stimulation of NF␬B-dependent transcription by an activated form of MEKK1 is also inhibited by salicylates (Figs. 6 and 7; Ref. 22). Because the effect of such drugs on TC10- or Cdc42-mediated activation of NF-␬B had not been reported previously, we chose to study the effect of NaSal at a concentration used by other investigators and known to inhibit NF-␬B activation, as well as cell adhesion stimulated by extracellular factors and/or upstream kinase activators (23, 43). As shown in Figs. 6 and 7, the results indicate that Cdc42mediated, but not TC10-mediated, activation of NF-␬B is sensitive to salicylate treatment, suggesting that Cdc42 is an integral component of the signaling cascades described above, but that TC10 activates NF-␬B by a mechanism that is distinct, at least in part, from that used by Cdc42. Of interest in connection with these findings are the observation that TC10 and Cdc42 do not show a synergistic activation of NF-␬B when coexpressed in COS cells4 and the fact that Cdc42 has been shown to bind and activate MEKK1 (44). An interaction between TC10 and MEKK1 has not yet been reported. In addition to IKK, I␬B can be phosphorylated and presumably targeted for degradation by casein kinase II, RSK2, and an unknown tyrosine kinase (for reviews, see Refs. 17 and 19; Ref. 45). Because RSK2, like IKK␤, is inhibited directly by salicylates, it is unlikely that TC10 activated NF-␬B through RSK2, but other phosphorylation pathways are possible. Alternatively, TC10 may mediate NF-␬B activation by a mechanism independent of I␬B phosphorylation. In summary, the data reported in this study emphasize key differences between TC10- and Cdc42-mediated signaling, including cellular localization and regulation by RhoGDI␣, involvement in NGF-induced neuritogenesis, and the pathway mediating activation of NF-␬B. At this time, the quantitative significance of such differences is difficult to assess because processes contributing to the upstream activation of TC10 have not yet been identified. These differences, along with the inability of TC10 to restore Cdc42p function in yeast, clearly show that TC10 and Cdc42 have not evolved as functionally redundant signaling molecules.

Materials and Methods DNA Constructs. Wild-type TC10 cDNA in pMT2 (1) was used as a PCR template to generate TC10 open reading frame cassettes for subcloning into pCGT (2). The cassette contained 5⬘ XbaI and 3⬘ BamHI sites and was directionally cloned downstream of and in frame with the T7 epitope tag of the vector. Wild-type TC10 subcloned into pCGT was then used to generate point mutants coding for Q75L and T31N amino acid changes, using the Stratagene PCR-based QuikChange System according to the manufacturer’s protocol (2). For TC10 localization studies in live cells, wild-type and mutant (75L or 31N) TC10 in pCGT were used as PCR templates to generate open reading frame cassettes for subcloning into pEGFP-C3 (a generous gift of Dr. Marcela Nadal;

New York University, New York, NY). The cassettes contained 5⬘ EcoRI and 3⬘ BamHI sites and were directionally cloned downstream of and in frame with the GFP tag of the vector. For S. cerevisiae rescue studies, wild-type and Q75L human TC10 and wild-type and Q61L human Cdc42 (placental isoform) in pEGFP-C3 were used as PCR templates to generate five open reading frame cassettes, TC10, TC10 75L, GFP-TC10, Cdc42, and Cdc42 61L. The cassettes contained 5⬘ BamHI and 3⬘ EcoRI sites and were cloned directionally into pGAD.MS (a generous gift of Dr. Adam Hittelman; New York University) for constitutive expression driven by the alcohol dehydrogenase (ADH1) promoter. pGAD.MS also carries a LEU2 gene for plasmid selection and maintenance. The mammalian expression vectors for GFP-Rac, GFPCdc42 (placental isoform), and RhoGDI␣ were constructed as described elsewhere (5). Additional expression constructs were obtained as generous gifts. Wild-type, activated (61L), and dominant negative (17N) Cdc42 mammalian (placental isoform in pKH3, HA tag) expression constructs were gifts of Dr. Danny Manor (Cornell University, Ithaca, NY; Ref. 46). The mammalian expression vector for the NF-␬B-luciferase reporter was a gift of Dr. Channing Der (University of North Carolina, Chapel Hill, NC; Ref. 47). Activated (amino acids 1–320 deleted) EE (EEEEYMPME)-epitope-tagged MEKK1 (⌬MEKK1) expression construct was a gift of Dr. Ed Skolnik (New York University; Ref. 44). Mammalian Cell Culture and Transfection Conditions. COS and HeLa cells were grown at 37°C, 10% CO2 in DMEM supplemented with 10% calf serum (Life Technologies, Inc.). Transient transfections (1 ␮g of DNA per plate unless indicated otherwise) were carried out in serum-free medium using LipofectAMINE PLUS Reagent and a standard protocol supplied by the manufacturer (Life Technologies, Inc.). For cell localization studies, cells were transferred to medium supplemented with 10% serum 6 h posttransfection and cultured for an additional 18 h. For NF-␬B reporter studies, cells were maintained in serum-free medium for at least 32 h after transfection. PC12 cells (a generous gift of Dr. Liang-Tung Yang; New York University) were grown at 37°C, 10% CO2 in DMEM supplemented with 10% HS and 10% FBS (Life Technologies, Inc.) on collagen (Cohesion Technologies)-coated dishes. Neurite outgrowth was induced by treatment of cells with 50 ng/ml NGF (Harlan Bioproducts) in medium supplemented with 0.5% HS and 0.5% FBS. For cell survival and neurite maintenance, cells were treated with NGF every 24 h up to 72 h by replacing half of the culture medium with fresh serum-free medium containing 50 ng/ml NGF. Transient transfections were carried out in serum-free medium using LipofectAMINE 2000 and a standard protocol (Life Technologies, Inc.). Transfected cells were cultured for 18 h in medium supplemented with 10% HS and 10% FBS before NGF treatment. Design and Production of Anti-TC10 Antipeptide Antibody. The 14-amino acid peptide TLARLNDMKEKPIC (TC10 residues 139 –152) was synthesized, purified by highperformance liquid chromatography, and conjugated via the

165

166


COOH-terminal cysteine to keyhole limpet hemocyanin (Pierce) through activated maleimide according to the protocol supplied by the manufacturer. The extent of conjugation was determined by free sulfhydryl reactivity using Ellman’s Reagent (Pierce). The conjugate was introduced into two different rabbits following a typical protocol, and sera were collected at 35 and 56 days after the initial inoculation (Cocalico Biologicals). Serum from test bleeds of both rabbits was titrated against a bacterially expressed, purified TC10 fusion protein and determined to contain antibody that recognizes and is specific for TC10. Affinity-purified antibody does not recognize Ran, Rab11, RalA, RhoA, Rac1, Cdc42, or H-Ras.6 Live Cell Intracellular Localization Assay. To examine TC10 localization in live cells, HeLa cells grown to 70% confluence in 35-mm dishes (Matek Corp.) designed for use in an inverted microscope were transfected with plasmids containing GFP-tagged GTPases (pEGFP) in the absence or presence of a 3-fold excess of RhoGDI␣ expression construct. At 24 h posttransfection, cells were washed with PBS and viewed with a Zeiss Axiophot fluorescence microscope, and digitized images were collected. Yeast Rescue Assay. cdc42ts (MAT␣ ura3 his4 leu2 trp1 gal2 cdc42–1) and parental (MAT␣ ura3 his4 leu2 trp1) S. cerevisiae strains (generous gifts of Dr. Erfie Bi; University of Pennsylvania, Philadelphia, PA) were maintained in YPAD (rich medium ⫹ 2% glucose) and transformed with plasmids to express TC10, Cdc42, or GFP-TC10 (pGAD.MS) by a LiOAc/polyethylene glycol method as described previously (48), except that DMSO was added to 10% during a 10-min 42°C heat shock. Cells were selected for leucine prototrophy after 2– 4 days at 23°C (permissive temperature), and transformants were recovered. Strains to be examined for growth at the restrictive temperature were grown at 23°C in 2–3 ml of synthetic complete medium minus leucine (for strains transformed with pGAD.MS) or synthetic complete (for nontransformed strains) to A600 nm ⬇ 1, diluted to equivalent A600 nm, and streaked on solid medium in duplicate. One plate was incubated at 23°C (permissive temperature) for 2– 4 days, and the other was incubated at 37°C (restrictive temperature) for 2– 4 days. For cultures expressing GFP-TC10, 10 ␮l of culture at A600 nm ⬇ 1 were spotted on a glass slide, covered with a glass coverslip, and examined in a Zeiss Axiophot fluorescence microscope. For immunoblot analysis, TC10 and Cdc42 transformants were grown under selection for leucine prototrophy to A600 nm ⬇ 1. Lysates were prepared by rapid boiling and glass bead rupture (49). Briefly, cells were transferred to a tube containing 50 mM Tris (pH 7.5) and 10 mM NaN3 on ice, pelleted in a clinical centrifuge, resuspended in 30 ml of ESB [2% SDS, 80 mM Tris (pH 6.8), 10% glycerol, 1.5 mM DTT, and 0.1 mg/ml bromphenol blue], and heated quickly for 3 min at 100°C. Immediately after heating, 0.2 mm acidwashed glass beads (Sigma) were added, tubes were vor-

6

P. Pe´rez de la Ossa and G. Murphy, unpublished observations.

texed for 2 min, and 70 ␮l of ESB were added. Extract (25 ␮l) was resolved by SDS-PAGE, transferred to nitrocellulose, and immunoblotted for TC10 or Cdc42 using our antipeptide TC10 antibody or anti-Cdc42 antibody (Santa Cruz Biotechnology), and protein expression was detected by chemiluminesence. Neurite Outgrowth Assay. PC12 cells (70% confluent in 35-mm dishes) were transfected with wild-type or mutant TC10 (pCGT) or wild-type or mutant Cdc42 (pKH3) in serumfree medium. Six h posttransfection, cells were transferred to medium supplemented with 10% HS and 10% FBS for 18 h and then transferred to medium supplemented with 0.5% HS and 0.5% FBS plus or minus 50 ng/ml NGF. Cells were successively treated every 24 h up to 72 h by replacing half of the culture medium with fresh serum-free medium containing 50 ng/ml NGF. At 96 h posttransfection, cells were fixed by removing half of the medium and adding an equal volume of 4% paraformaldehyde and then permeabilized with 0.2% Triton X-100; stained with anti-T7 (TC10) or anti-HA (Cdc42) antibody, rhodamine-phalloidin, and 4⬘,6diamidino-2-phenylindole; mounted with Vectashield (Vector Laboratories); and viewed in a Zeiss Axiophot fluorescence microscope. Neurites were counted as extensions at least two cell bodies in length. Morphological changes were counted as alterations in cell shape and size in the absence of neurite formation. NF-␬B Activation Assay. COS cells (70% confluent in 35-mm dishes) were transfected with wild-type or mutant T7-tagged TC10 (pCGT), wild-type or mutant HA-tagged Cdc42 (pKH3), or EE-tagged ⌬MEKK1 and a reporter plasmid containing three NF-␬B binding elements fused to a firefly luciferase gene (2, 47). Six h posttransfection, cells were transferred to fresh serum-free medium, cultured for 20 h, and then transferred to fresh serum-free medium in the absence or presence of 20 mM NaSal, 1 mM acetaminophen, or 5 ␮M indomethacin (Sigma) for 6 h. Samples to be tested for recovery of NF-␬B activity were transferred to fresh serum-free medium in the absence of NaSal for an additional 20 h. Cells were harvested, lysed, and examined for luciferase activity as described previously (2). The level of T7-tagged GTPase expression was nearly identical for all T7-tagged samples analyzed, the level of HA-tagged GTPase expression was nearly identical for all HA-tagged samples analyzed, and the level of EE-tagged ⌬MEKK1 expression was nearly identical for all EE-tagged samples examined.

References 1. Drivas, G. T., Shih, A., Coutavas, E., Rush, M. G., and D’Eustachio, P. Characterization of four novel ras-like genes expressed in a human teratocarcinoma cell line. Mol. Cell. Biol., 10: 1793–1798, 1990. 2. Murphy, G. A., Solski, P. A., Jillian, S. A., Pe´rez de la Ossa, P., D’Eustachio, P., Der, C. J., and Rush, M. G. Cellular functions of TC10, a Rho family GTPase: regulation of morphology, signal transduction and cell growth. Oncogene, 18: 3831–3845, 1999. 3. Neudauer, C. L., Joberty, G., Tatsis, N., and Macara, I. G. Distinct cellular effects and interactions of the Rho-family GTPase TC10. Curr. Biol., 8: 1151–1160, 1998. 4. Erickson, J. W., Zhang, C. J., Kahn, R. A., Evans, T., and Cerione, R. A. Mammalian Cdc42 is a brefeldin A-sensitive component of the Golgi apparatus. J. Biol. Chem. 271: 26850 –26854, 1996. 5. Michaelson, D., Silletti, J., Murphy, G., D’Eustachio, P., Rush, M., and Philips, M. R. Differential localization of Rho GTPases in live cells: regu-


lation by hypervariable regions and RhoGDI binding. J. Cell Biol., 152: 111–126, 2001. 6. Fukumoto, Y., Kaibuchi, K., Hori, Y., Fujioka, H., Araki, S., Ueda, T., Kikuchi, A., and Takai, Y. Molecular cloning and characterization of a novel type of regulatory protein (GDI) for the rho proteins, ras p21-like small GTP-binding proteins. Oncogene, 5: 1321–1328, 1990. 7. Gosser, Y. Q., Nomanbhoy, T. K., Aghazadeh, B., Manor, D., Combs, C., Cerione, R. A., and Rosen, M. K. C-terminal binding domain of Rho GDP-dissociation inhibitor directs N-terminal inhibitory peptide to GTPases. Nature (Lond.), 387: 814 – 819, 1997. 8. Nomanbhoy, T. K., Erickson, J. W., and Cerione, R. A. Kinetics of Cdc42 membrane extraction by Rho-GDI monitored by real-time fluorescence resonance energy transfer. Biochemistry, 38: 1744 –1750, 1999. 9. Olofsson, B. Rho guanine dissociation inhibitors: pivotal molecules in cellular signaling. Cell Signalling, 11: 545–554, 1999. 10. Mackay, D. J. G., Nobes, C. D., and Hall, A. The Rho’s progress: a potential role during neuritogenesis for the Rho family of GTPases. Trends Neurosci., 18: 496 –501, 1995. 11. Kaplan, D. R. Studying signal transduction in neuronal cells: the Trk/NGF system. Prog. Brain Res., 117: 35– 46, 1998. 12. Jalink, K., van Corven, E. J., Hengeveld, T., Morii, N., Narumiya, S., and Moolenaar, W. H. Inhibition of lysophosphatidate- and thrombininduced neurite retraction and neuronal cell rounding by ADP ribosylation of the small GTP-binding protein Rho. J. Cell Biol., 126: 801– 810, 1994. 13. Kozma, R., Sarner, S., Ahmed, S., and Lim, L. Rho family GTPases and neuronal growth cone remodeling: relationship between increased complexity induced by Cdc42Hs, Rac1, and acetylcholine and collapse induced by RhoA and lysophosphatidic acid. Mol. Cell. Biol., 17: 1201– 1211, 1997. 14. Daniels, R. H., Hall, P. S., and Bokoch, G. M. Membrane targeting of p21-activated kinase 1 (PAK1) induces neurite outgrowth from PC12 cells. EMBO J., 17: 754 –764, 1998. 15. Tanabe, K., Tachibana, T., Yamashita, T., Chen, Y. H., Yoneda, Y., Ochi, T., Tohyama, M., Yoshikawa, H., and Kiyama, H. The small GTPbinding protein TC10 promotes nerve elongation in neuronal cells, and its expression is induced during nerve regeneration in rats. J. Neurosci., 20: 4138 – 4144, 2000. 16. Ghosh, S., May, M. J., and Kopp, E. B. NF-␬B and Rel proteins: evolutionarily conserved mediators of immune responses. Annu. Rev. Immunol., 16: 225–260, 1998. 17. Mercurio, F., and Manning, A. M. Multiple signals converging on NF-␬B. Curr. Opin. Cell Biol., 11: 226 –232, 1999. 18. Perona, R., Montaner, R., Saniger, L., Sanchez-Perez, I., Bravo, R., and Lacal, J. C. Activation of nuclear factor-␬B by Rho, CDC42, and Rac-1 proteins. Genes Dev., 11: 463– 475, 1997. 19. Karin, M. The beginning of the end: I␬B kinase (IKK) and NF-␬B activation. J. Biol. Chem., 274: 27339 –27342, 1999. 20. Kopp, E., and Ghosh, S. Inhibition of NF-␬B by sodium salicylate and aspirin. Science (Washington DC), 265: 956 –959, 1994. 21. Schwenger, P., Alpert, D., Skolnik, E. Y., and Vilcek, J. Activation of p38 mitogen-activated protein kinase by sodium salicylate leads to inhibition of tumor necrosis factor-induced I␬B␣ phosphorylation and degradation. Mol. Cell. Biol., 18: 78 – 84, 1998. 22. Yin, M. J., Yamamoto, Y., and Gaynor, R. B. The anti-inflammatory agents aspirin and salicylate inhibit the activity of I␬B kinase-␤. Nature (Lond.), 396: 77– 80, 1998. 23. Alpert, D. A., Schwenger, P., Han, J., and Vilcek, J. Cell stress and MKK6b-mediated p38 MAP kinase activation inhibit tumor necrosis factor-induced I␬B phosphorylation and NF-␬B activation. J. Biol. Chem., 274: 22176 –22183, 1999. 24. Miller, P. J., and Johnson, D. I. Cdc42p GTPase is involved in controlling polarized cell growth in Schizosaccharomyces pombe. Mol. Cell. Biol., 14: 1075–1083, 1994. 25. Cabib, E., Drgonova, J., and Drgon, T. Role of small G proteins in yeast cell polarization and wall biosynthesis. Annu. Rev. Biochem., 67: 307–333, 1998. 26. Johnson, D. I. Cdc42: an essential Rho-type GTPase controlling eukaryotic cell polarity. Microbiol. Mol. Biol. Rev., 63: 54 –105, 1999. 27. Adams, A. E., Johnson, D. I., Longnecker, R. M., Sloat, B. F., and Pringle, J. R. CDC42 and CDC43, two additional genes involved in budding and the establishment of cell polarity in the yeast Saccharomyces cerevisiae. J. Cell Biol., 111: 131–142, 1990. 28. Johnson, D. I., and Pringle, J. R. Molecular characterization of CDC42, a Saccharomyces cerevisiae gene involved in the development of cell polarity. J. Cell Biol., 111: 143–152, 1990.

29. Munemitsu, S., Innis, M. A., Clark, R., McCormick, F., Ullrich, A., and Polakis, P. Molecular cloning and expression of a G25K cDNA, the human homolog of the yeast cell cycle gene CDC42. Mol. Cell. Biol., 10: 5977– 5982, 1990. 30. Shinjo, K., Koland, J. G., Hart, M. J., Narasimhan, V., Johnson, D. I., Evans, T., and Cerione, R. A. Molecular cloning of the gene for the human placental GTP-binding protein Gp (G25K): identification of this GTP-binding protein as the human homolog of the yeast cell-division-cycle protein CDC42. Proc. Natl. Acad. Sci. USA, 87: 9853–9857, 1990. 31. Joberty, G., Perlungher, R. R., and Macara, I. G. The Borgs, a new family of Cdc42 and TC10 GTPase-interacting proteins. Mol. Cell. Biol., 19: 6585– 6597, 1999. 32. Mondal, M. S., Wang, Z., Seeds, A. M., and Rando, R. R. The specific binding of small molecule isoprenoids to rhoGDP dissociation inhibitor (rhoGDI). Biochemistry, 39: 406 – 412, 2000. 33. Adra, C. N., Manor, D., Ko, J. L., Zhu, S., Horiuchi, T., Aelst, L. V., Cerione, R. A., and Lim, B. RhoGDI␥: a GDP-dissociation inhibitor for Rho proteins with preferential expression in brain and pancreas. Proc. Natl. Acad. Sci. USA, 94: 4279 – 4284, 1997. 34. Zalcman, G., Closson, V., Camonis, J., Honore´, N., Rousseau-Merck, M-F., Tavitian, A., and Olofsson, B. RhoGDI-3 is a new GDP dissociation inhibitor (GDI). J. Biol. Chem., 271: 30366 –30374, 1996. 35. Mulholland, J., Preuss, D., Moon, A., Wong, A., Drubin, D., and Botstein, D. Ultrastructure of the yeast actin cytoskeleton and its association with the plasma membrane. J. Cell Biol., 125: 381–391, 1994. 36. Rohatgi, R., Ma, L., Miki, H., Lopez, M., Kirchhausen, T., Takenawa, T., and Kirschner, M. W. The interaction between N-WASP and the Arp2/3 complex links Cdc42-dependent signals to actin assembly. Cell, 97: 221– 231, 1999. 37. Suetsugu, S., Miki, H., and Takenawa, T. The essential role of profilin in the assembly of actin for microspike formation. EMBO J., 17: 6516 – 6526, 1998. 38. Ziman, M., O’Brien, J. M., Ouellette L. A., Church W. R., and Johnson, D. I. Mutational analysis of CDC42Sc, a Saccharomyces cerevisiae gene that encodes a putative GTP-binding protein involved in the control of cell polarity. Mol. Cell. Biol., 11: 3537–3544, 1991. 39. Chen, X. Q., Tan, I., Leung, T., and Lim, L. The myotonic dystrophy kinase-related Cdc42-binding kinase is involved in the regulation of neurite outgrowth in PC12 cells. J. Biol. Chem., 274: 19901–19905, 1999. 40. Morooka, T., and Nishida, E. Requirement of p38 mitogen-activated protein kinase for neuronal differentiation in PC12 cells. J. Biol. Chem., 273: 24285–24288, 1998. 41. Kita, Y., Kimura, K. D., Kobayashi, M., Ihara, S., Kaibuchi, K., Kuroda, S., Ui, M., Iba, H., Konishi, H., Kikkawa, U., Nagata, S., and Fukui, Y. Microinjection of activated phosphatidylinositol-3 kinase induces process outgrowth in rat PC12 cells through the Rac-JNK signal transduction pathway. J. Cell Sci., 111: 907–915, 1998. 42. Zohn, I. M., Campbell, S. L., Khosravi-Far, R., Rossman, K. L., and Der, C. J. Rho family proteins and Ras transformation: the RHOad less traveled gets congested. Oncogene, 17: 1415–1438, 1998. 43. Pillinger, M. H., Capodici, C., Rosenthal, P., Kheterpal, N., Hanft, S., Philips, M. R., and Weissman, G. Modes of action of aspirin-like drugs: salicylates inhibit Erk activation and integrin-dependent neutrophil adhesion. Proc. Natl. Acad. Sci. USA, 95: 14540 –14545, 1998. 44. Fanger, G. R., Johnson, N. L, and Johnson, G. L. MEK kinases are regulated by EGF and selectively interact with Rac/Cdc42. EMBO J., 16: 4961– 4972, 1997. 45. Stevenson, M. A., Zhao, M. J., Asea, A., Coleman, C. N., and Calderwood, S. K. Salicylic acid and aspirin inhibit the activity of RSK2 kinase and repress RSK2-dependent transcription of cyclic AMP response element binding protein- and NF-␬B-responsive genes. J. Immunol., 163: 5608 –5616, 1999. 46. Wu, W. J., Lin, R., Cerione, R. A., and Manor, D. Transformation activity of Cdc42 requires a region unique to Rho-related proteins. J. Biol. Chem., 273: 16655–16658, 1998. 47. Galang, C. K., Garcia-Ramirez, J., Solski, P. A., Westwick, J. K., Der, C. J., Neznanov, N. N., Oshima, R. G., and Hauser, C. A. Oncogenic Neu/ErbB-2 increases ets, AP-1, and NF-␬B-dependent gene expression, and inhibiting ets activation blocks Neu-mediated cellular transformation. J. Biol. Chem., 271: 7992–7998, 1996. 48. Agatep, R., Kirkpatrick, R. D., Parchaliuk, D. L., Woods, R. A., and Gietz, R. D. Transformation of Saccharomyces cerevisiae by the lithium acetate/single-stranded carrier DNA/polyethylene glycol protocol. Tech. Tips Online (http://tto.biomednet.com), Core Protocol #01525, 1998. 49. Adams, A., Gottschling, D. E., Kaiser, C. A., and Stearns, T. Methods in Yeast Genetics: A Cold Spring Harbor Laboratory Course Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, 1997.

167