Activation of phosphoinositide 3-kinase activity by Cdc42Hs binding to ...

THE JOURNAL OF BIOLWICAL CHEIEMISTRY Vol. 269, No. 29, Issue of July 22, pp. 18727-18730, 1994 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

Communication

(Ziman et al., 1993; Miller and Johnson,1994), it is likely to be involved in processes that influence mammalian cell growth and cytokinesis. Given that the GTP-binding/GTPase cycle of Cdc42 is an essential regulatory feature of its biological actions, we and identify mammalian proteins thatregu(Received for publication, May 9, 1994, and in revised form, others have set out to May 25, 1994) late this cycle. Three such regulators havebeen identified for the human Cdc42 protein (Cdc42Hs). The first of these, the Yi ZhengS, Shubha BagrodiaS, and Cdc42Hs-GTPase-activatingprotein (Cdc42Hs-GAP),’ is a 50Richard A. Cerione kDa protein (Barfod et al., 19931, the functionaldomain of From the Department of Pharmacology, Schurman Hall, which shares homology with various proteins including the Cornell Uniuersity, Zthaca, New York 14853 breakpoint cluster region (Bcr) protein (Heisterkamp et al., The Ras-likeGTPase Cdc42is essential for cell polarity 1985; Diekmann et al., 1991), the Ras-GAP-binding protein and bud site assembly in Saccharomyces cerevisiae by p190 (Settleman et al., 19921, and the 85-kDa regulatory subregulating cell cycle-dependent reorganization of corti- unit of the phosphatidylinositol (PI)3-kinase(Otsu et al., cal cytoskeletal elements. However, its role in mamma- 1991). A second regulator, the GDP dissociation inhibitor, was lian cells is unknown. To identify potential effectors of purified from bovine brain cytosol (Ueda et al., 1990) and inCdc42Hs, weincubated lysates from NIH 3T3 fibroblasts hibits bothGDP dissociation(Leonard et al., 1992) and GTPase or PC12 cells with immobilized glutathione S-transfer- activity (Hart et al., 1992). A third regulator is the Dbl oncoase (GST)-Cdc42Hsfusion proteins bound to different protein, which catalyzes the GDP-GTP exchange reaction and guanine nucleotides and observed aspecific association is a guanine nucleotide exchange factorfor Cdc42Hs and RhoA between the 85-kDa subunit (p85) of phosphatidylinosi- (Hart et al., 1994). to1 3-kinase (PI 3-kinase) and GTPyS (guanosine 5’-3-0While the identification of these different proteins has pro(thi0)triphosphate)-boundGST-Cdc42Hs.Recombinant vided insight into the molecular mechanisms underlying the p85 formed a complexwith GTP+-bound GST-Cdc42Hs regulation of the GTP-binding/GTPase cycle of Cdc42Hs, it and with a GTPase-defective GTP-bound GST-Cdc42Hs- seems unlikelythat any of these regulatory proteins represent Q6lLmutant,but not with aGTP+-bound, effector true effectors or targets for the biological action of Cdc42. Redomain GST-Cdc42HsT35A mutant. Both the Rho-GAP cent studies (Manseret al., 1993, 19941, using a nitrocellulose homology domain of p85 and the Cdc42Hs-GAPcompetifilter overlay assay, have identified proteins that bind to the tively inhibited the binding ofrecombinantp85 to Cdc42Hs. In addition, PI 3-kinase activity immunopre- GTP-bound form of Cdc42Hs and have reported two potential cipitated from cell lysates with anti-p85 antibody was targets, a novel 120-kDa hippocampal tyrosine kinase and a stimulated --fold by GST-Cdc42-GTPyS.Similar inter- 65-kDa brain serinehhreonine kinase. In this work, we have actions were observed between p85 and GST-Racl- used a glutathioneS-transferase Cdc42Hs fusion protein, GTPyS but not between p85 and GST-RhoA-GTPyS. bound to GTPrS, as an affinity resin for binding potential These findings suggest that PI 3-kinase, through the Cdc42Hs targets that are present in different cell lysates. An Rho-GAP homology domain of p85, can couple to the 85-kDa protein wasidentified as one such target andshown to of PI 3-kinase. We have chareffector domain of Cdc42Hs and that p85 may be a targetrepresent the regulatory subunit acterized the interactionbetween Cdc42Hs and p85 and shown for the GTP-bound forms of Cdc42Hs and Racl. that it requires the Rho-GAP homology domain of p85 and the putative effector domain of Cdc42Hs. A GTPase-defective Cdc42Hs mutant binds p85 and concomitantly PI 3-kinase acThe members of the Rho subclass of the Ras superfamily, those of tivity is stimulated. These results, taken together with which includes the Rho, Rac, and Cdc42 proteins, play organiet al. (1993, 19941, raise the possibility that the pathManser zational roles in different cytoskeletal-associatedassembly procways of Rho-subtype proteins (particularlyRac and Cdc42) are esses (Boguski and McCormick, 1993). RhoA and Racl have beenimplicated in cytoskeletal reorganization, where RhoA linked to protein and lipid kinases, perhaps ina manner analappears to mediate growth factor-stimulated stress fiber and ogous to the actions of Ras on the Raf kinase.

Activation of Phosphoinositide 3-Kinase Activity by Cdc42Hs Binding to p85*

focal adhesion formation(Ridley and Hall,1992) and Raclpromotes growth factor-stimulated membrane ruffling (Ridley et al., 1992). The Saccharomyces cereuisiae Cdc42 protein (Cdc42Sc) is essential for bud site assembly through its ability to control polarized growth and to influence the cortical cytoskeleton (JohnsonandPringle, 1990; Drubin, 1991). At present, the role for Cdc42 in mammalian cells is unknown, although based on its function in budding and fission yeast

* This work was supported by National Institutes of Health Grant GM47458 (to R.A. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ These two authors should be considered equal.

EXPERIMENTAL PROCEDURES

Construction and Expression of GST Fusion Proteins-The cDNAs coding for GST-Cdc42Hs, GST-Racl, GST-RhoA, and GST-Cdc42HsGAP homology domain have been described previously (Hart et al., 1994; Barfod et al., 1994). The human p85a cDNA in the pGEX-2T vector was a gift from Dr. Lewis Cantley (Harvard Medical School).Sequences corresponding to the Rho-GAP homology domain in human p85a (amino acids 71-339) and in the Cdc42HsGAP (C-terminal 235 amino acids) were subclonedinto the pGEX-2Tvector(Barfod et al., 1994). Sitedirected mutagenesis was performed t o generate GST-Cdc42HsQGlL

The abbreviations used are: GAP, GTPase-activating protein; Bcr, break point cluster region protein; PI 3-kinase, phosphatidylinositol 3-kinase; GST, glutathione S-transferase; GTPyS, guanosine 5’-3-0(thioltriphosphate; PAK, pal-activated kinase; ACK, activated Cdc42Hs-associated kinase.

18727

18728

Activation of PI 3-Kinase Activity

by Cdc42Hs Binding to

p85

(glutamine at position 61 is mutated to leucine) and GST-Cdc42HsT35A A Cdc42Hs (threonine a t position 35 is mutated to alanine) by in vitro mutagenesis WCLGST kit (Amersham Corp). GST-fusion proteins were expressed in 1 2 3 4 5 Escherichia coli and purified by glutathione affinity chromatography according to Zheng etal. (1993), except that 10 p~ GDP was added in the lysis buffer. Cell Culture a n d Cell Lysis-NIH 3T3 cells were grown in Dulbecco's modified Eagle's medium supplemented with 5% calf serum (Sigma) under 10% CO,. PC12 cells were grown in Dulbecco's modified Eagle's medium with 5% horse serum (Sigma) and 5% fetal calf serum (Upstate Biotechnology, Inc.) under 5% CO,. Confluent PC12 cells were lysed in buffer A (20 mM Tris (pH 8.0), 100 mM NaCI, 1 mM EDTA, 0.3 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 1% Triton X-100, -D85 10 mg/ml each aprotinin and leupeptin) a t 4 "C for 1h, and lysates were clarified by spinning in the Eppendorf centrifuge (model 5415 C) for 30 min. a t 14,000 rpm. Subconfluent NIH 3T3 cells were washed with phosphate-buffered saline and lysed in buffer B (40 mM Hepes (pH7.41, 100 mM NaCl, 1mM EDTA, 1% Nonidet P-40,lOmM NaF, 1mM sodium orthovanadate, 10 pg/ml each aprotinin and leupeptin) at 4 "C for 20 B Rac I Rho A GTPYS" G D p ~ ~ ~ ~ ~ ' min. The lysates were clarified by spinning ina JA20.1rotor (Beckman) W C L GST - GDP 1 2 3 4 5 6 7 8 at 15,000 x g for 25 min at 4 "C. MW Association of p85 with GST-GTP-bindingProteins-GST-GTP-binding proteins (-50 pg) were incubatedfirst with 10 mM EDTA and then with 0.2-1 mM GDP, GTP, or GTPyS in the presence of excess MgCl, as 190 KDdescribed previously(Hart etal., 1994). Immobilized GST or GST-GTPbinding proteins bound to different guanine nucleotides were incubated with PC12or NIH 3T3cell lysates in the presence of 10 mM MgCl, (final 125KDconcentration, to ensure the bindingof the guanine nucleotides to the GTP-binding proteins) for 1-2 h a t 4 "C. Protein precipitates were washed three times with ice-cold buffer C (20mM Tris (pH 8.0),100 mM -P% NaCI, and 10 mM MgCl,) in thecase of PC12 cell lysates or with buffer B in the case of NIH 3T3cell lysates. Bound proteins were eluted in SDS FIG.1. A, association of p85 from NIH 3T3 cells with immobilized sample buffer, subjected to 7.5% or 10% SDS-polyacrylamide gel electrophoresis, Western-blotted, and probed with anti-p85 monoclonal or Cdc42Hs. Lysates from NIH 3T3cells were incubated withimmobilized polyclonal antibody (UpstateBiotechnology, Inc.). The primary antibod- GST (lane 2 ) , and GST-Cdc42Hs prebound with (lanes 4 and 5 ) or without (-) (lane 3 ) the indicated nucleotides andbound proteins were iesweredetectedwithanti-mouse or anti-rabbitimmunoglobulins coupled to horseradish peroxidase, using the Renaissance chemilumi- Western-blotted and probed with anti-p85 monoclonal antibody as del represents a fraction nescence reagent (DuPont NEN). To measure complex formation be- scribed under "Experimental Procedures." Lane binding reactions.Results are tween recombinant p85a and Cdc42Hs, immobilized GST-p85a was first of the whole cell lysate (WCL) used in the cleaved withthrombin (1 pg/lOO pg GST-fusion protein,Sigma) in representative of five independent experiments. B , association of p85 buffer C for 30 min at 4 "C. Thrombin was removed by incubating the from NIH 3T3 cells with immobilized Racl and RhoA. All procedures were similar to those described for panel A. Lysates from NIH 3T3cells supernatants withagarose-boundp-aminobenzamidinebeads for 1h a t were incubated with immobilized GST (lane 2 ) ; GST-Racl and GST4 "C. Cleaved p85a (2-4 pg) was incubated with immobilized wild type RhoA prebound with or without (-) the indicated nucleotides. Lane 1 and mutant GST-Cdc42Hs in buffer C supplemented with 0.3 mM di- represents 5% of the whole cell lysate (WCL) used in the binding reacthiothreitol for 2 h at 4 "C. For competition assays,2-5 pg of p85a and tions. Results are representativeof two independent experiments. 10 pgof recombinant Rho-GAP homology domain of p85u or10 pg of the functional domainof the Cdc42HsGAP was added simultaneously to theGTPyS complex (data not shown). Similarly, in NIH 3T3 cells immobilized wild type and mutantGST-Cdc42Hs. Bound proteins were overexpressing an activated mutant of the c-Src tyrosine kiwashed three times with buffer C, and p85a was detected as described nase, c-Src(Y527F),an 85-kDa tyrosine-phosphorylated protein above. Immunoprecipitation a n d PI 3-Kinase Assay-PC12 cell lysates were was specifically bound by the GST-Cdc42Hs-GTPyS species prepared as described above, and PI3-kinase was immunoprecipitated (data not shown). Since the 85-kDa regulatory subunit (p85) of with antLp85a polyclonal antibody (Upstate Biotechnology, inc.). Im- the PI 3-kinase is tyrosine-phosphorylated in Src-transformed munocomplexes were collected on protein A-Sepharose beads, washed cells (Fukui and Hanafusa, 1989), this raised the possibility twice with buffer C, and washed once with kinase assay buffer (20mM Tris (pH 7.4), 10 mM MgCI,). Wild type or mutant GST-Cdc42Hs (2-5 pg) that the85-kDa protein was p85. To test this possibility, NIH 3T3 lysates were incubated with immobilized GST-Cdc42Hs in bound to different guanine nucleotides was added to the washed immunoprecipitates and PI 3-kinase activity was assayed by the addition of various nucleotide-bound states andprecipitated proteins were 5 pg of sonicated PI and phosphatidylserine (1:l) (Sigma) and 20 PM Western-blotted and probed with anti-p85a monoclonal anti[y-"'PIATP (200 pCi/ml) in 45 p1 of kinase assay buffer. The reactions body (Fig. L4).p85 bound specifically to theGTPyS-bound form were terminated after 10 min at 25 "C by the addition of 100 pl of 1 M of Cdc42Hs. (Fig. L4, lane 5 ) .No detectable p85 was present in HCI. The lipids were extracted with CHCI,:MeOH (1:l).analyzed by thin layer chromatographyon potassium oxalate-coated silica plates Sci(EM precipitates containing the guanine nucleotide-depleted GSTence), and developed in CHCl~eOH&I,OINH,OH (45:35:8.5:1.5). PI Cdc42Hs or the GDP-bound GST-Cdc42Hs(Fig. lA, lanes 3 and 3-kinase activity associated with wild type and mutant immobilized 4 ) or in precipitates containing GST alone (lane 2). Similar GST-Cdc42Hs precipitates from PC12 cell lysates was measured simi- experiments were performed with Racl, which is 70% identical larly. 32Pcounts on the TLC plates were extracted with MeOH and quan- to Cdc42Hs. Although the recombinant Racl preparationsfretitated by scintillation counting. quently contained other proteins capable of a nonspecific assoRESULTS AND DISCUSSION

In the present work, we have used the GTPyS-bound form of GST-Cdc42Hs as an affinity reagent for identifying potential targetsleffectors of Cdc42. When screening 35S-methionine-labeled NIH 3T3 cells for proteins that bind specifically to the GST-Cdc42Hs-GTPySspecies, we found that an85-kDa protein co-precipitated with glutathione-agarose-bound GST-Cdc42Hs-

ciation with the anti-p85 antibody, we consistently found a specific association between the GTPyS-bound GST-Racl and p85 (Fig. l B , lane 5 ) .In contrast, GST-RhoA showed no detectable binding to p85 (Fig. l B , lanes 6-8). However, GST-Racl, GST-Cdc42Hs, and GST-RhoA were able to bind to the RasGAP-associated p190 (from NIH 3T3 cell lysates): which has S. Bagrodia and J. Settleman, unpublished results.

Activation of PI 3-Kinase Activity by Cdc42Hs Binding to

p85

18729

p85-

1

1

2

3

4

2

3

4

5

6

7

8

9

1

0

5

FIG.3. Direct association of recombinant p85 with immobiFIG.2. Association of p86 from PC12 cells with immobilized lized wild type and mutant Cdc42Hs. Recombinant p85 was incuwild type and mutant Cdc42Hs. Procedures were similar to those batedwith immobilized GST (lane 11, GST-Cdc42Hs-GDP (lane 2). described for Fig. L4 except that anti-p85polyclonal antibody was used GST-Cdc42Hs-GTPyS (lane 3), GST-Cdc42Hs-GTPyS and (-10 pg) for immunoprobing. Lysates from PC12 cells were incubated with im- Rho-GAP homology domain of p85a (lane 4 ), GST-Cdc42Hs-GTPyS and mobilized GST (lane 1 j, GST-Cdc42Hs-GDP (lane 2 ) or GTPyS(lane 3 ) , (-10 pg) Cdc42Hs-GAP homology domain (lane 51, GST-Cdc42HsQ61L GST-Cdc42HsT35A-GTPyS (lane 4j, andGST-Cdc42HsQGlL-GTP (lane GTP (lane 6),and GST-Cdc42HsQGlLGTP and Rho-GAP homology 5 j. Results are representativeof three independent experiments. domain of p85a (lane 7), GST-Cdc42HsQGlL-GTP and Cdc42Hs-GAP homology domain (lane 81, and GST-Cdc42HsT35A-GTPyS (lane 9). Lane 10 represents 0.5 pg of recombinant p85. Bound proteins were been shown to function as a GAP for all three GTP-binding detected by probing Western blots with anti-p85 polyclonal antibody. proteins (Settlemanet al., 1992; Ridley et al., 1993). This indi- Results are representativeof two independent experiments.

cates that, althoughGST-RhoA does not bind p85, i t can associate specifically with other proteins in the lysates. cases, In all we verified that similar amountsof GTP-binding protein were incubated with the cell lysates (as assessed by [35SlGTPyS binding measurements). The results shown in Fig. 1were reminiscent of those obtained with a recently identified 65-kDa serine/threonine kinase where both the GTP-bound forms of Cdc42Hs and Racl (but notRhoA) were capable of binding to this protein kinase (Manser et al., 1994). We then set out to determine if the Cdc42Hdp85 interaction could be detected in other cell types and found that this complex formation could also be observed in rat pheochromocytoma (PC121cells. Fig. 2 (lane 3 ) shows that theGTPyS-bound form of GST-Cdc42Hs complexed to p85, whereas the GDP-bound form of the GST-GTP-binding protein showed much less ability to bind to p85 (lane 2 ) and GST alone showed no complex formation (lane I). A putative effector domain mutant of Cdc42Hs (based on the known effector domain for the Ras GTP-binding protein; Marshall (1993)), containing a threonine + alanine substitution a t position 35 (Cdc42HsT35A1, did not bind to p85 (Fig. 2 A , lane 4 ) . However, a GTPase-defective Cdc42Hs mutant containing a glutamine + leucine substitution at position 61 (Cdc42HsQGlL) binds p85 (Fig. 2 A , lane 51, even when Cdc42Hs is in the GTP-bound state. This differs from the case for the GTP-bound wild type Cdc42Hs, which has a relatively high intrinsic GTPase activity (Hart et al., 1991) and consequently showsin vitrobinding characteristics similar to those of the GDP-bound form of the protein. We next examined if the interactionbetween the GST-Cdc42GTPyS and p85 was direct andinvolved the Rho-GAP homology domain of p85. These studieswere performed with recombinant p85a expressed in E. coli. Fig. 3 shows the binding of recombinant p85 to immobilized GST-fusion proteins. No significant binding occurred between p85 and GST alone (Fig. 3, lane 1); a relatively weak interaction was observed between p85 and theGDP-bound form of GST-Cdc42Hs (Fig. 3, lane 2 ) and the strongest interaction occurred between p85 and the GST-Cdc42-GTPyS (Fig. 3, lane 3 ) . As expected, the GSTCdc42HsQGlL-GTP complex (lane 6) butnotthe GSTCdc42T35A-GTPySmutant (lane 9)bound to p85. The addition of the Rho-GAP homology domain of p85 to these binding reactions significantly attenuated the interaction of p85 with the GST-Cdc42-GTPyS (lane 4 ) and GST-Cdc42QGlL-GTPspecies (lane 7). Similar inhibitory effects were observed when using the recombinant Cdc42Hs-GAP, i.e. binding of recombinant p85 to both the Cdc42Hs-GTPyS (lane 5 ) and Cdc42HsQGlL-GTP (lane 8 ) complexes was inhibited. These results suggest that the Rho-GAP homology domain of p85 contains thebinding site for the Cdc42Hs-GTP (or -GTPyS) species. However, so far we have notfound p85 to serve asa GAP for Cdc42Hs (even when adding up to 10 pg of p85/GAP assay; Zheng et al. (199311, nor for Racl or RhoA (data not shown).

An important question concerned whether thebinding of p85 to theGTP-bound (or GTPyS-bound) form of Cdc42Hs had any functional effects. Fig. 4A shows that in the cases where the GST-Cdc42Hs bound to p85 in PC12 cell lysates, there was an accompanying binding of the PI3-kinase activity, as measured by the formation of the kinase product phosphatidylinositol 3-phosphate. No product formation was observed with GST alone (Fig. 4A, lane 1 ) or with the GST-Cdc42HsT35A-GTPyS (lane 41, and little product formation was elicited by the GSTCdc42Hs-GDP (lane 2 ) species. However, product formation was observed both with theGST-Cdc42Hs-GTPyS (lane 3 ) and with the GST-Cdc42HsQGlL-GTP complexes (lane 5). These results indicate that the 110-kDa catalytic domain of the PI 3-kinase also co-precipitates with the p85-Cdc42Hs complex. To see if the binding of PI 3-kinase to Cdc42Hs stimulated the activity of PI 3-kinase, anti-p85 immunoprecipitatesfrom PC12 cell lysates were mixed with different guanine nucleotidebound forms of GST-Cdc42Hs and subjected t o a PI 3-kinase assay (Fig. 4B).The GTPyS-bound form of GST-Cdc42Hs (lane 3 ) or the GTP-bound form of the GST-Cdc42HsQGlL (lane 5 ) mutant stimulated PI3-kinase activity 24-fold above the control GST alone (lane 1 ), while GDP-bound GST-Cdc42Hs stimulated the PI3-kinase activity by 1.5-fold (lane 2). In summary, the results of these studies indicate that the 85-kDa regulatory domain of PI 3-kinase is a candidate effector for the GTP-bound forms of the Cdc42Hs and Racl. Although p85 was originally thought to be a GAP for a Rho-subtype protein (because it shares homology with a number of known GAPS including the Cdc42Hs-GAP, Bcr, chimerin, and p190), thus far, there noisindication that thisprotein aloneis capable of influencing GTPase activity. However, it is possible that in some cases, the GAP homologydomain will be utilized for binding theGTP-bound forms of Rho-subtype proteins and thereby may serve as a target site. We have shown that this interaction lead to a stimulation of PI 3-kinase activity. Furthermore, it has been shown that either binding of phosphotyrosine peptides to the Src homology domain 2 (Backer et al., 1992; Carpenter et al., 1993) or binding of Src homology domain 3 to the proline-rich region of p85 (Pleiman et al., 1994) results in the activation of PI 3-kinase activity. These data suggest that p85 serves as a point of convergence for diverse signals toregulate PI 3-kinase. We did not observe any detectable complex formation between p85 from NIH 3T3 cells and RhoA, however, previous reports using C3 botulinum exoenzyme have suggested that RhoA is involved in theactivation of PI 3-kinase in Swiss 3T3 cells (Kumagai et al., 1993) and platelets (Zhang et al., 1993). This suggests thatRhoA may be further upstream in a pathway leading to PI3-kinase. At the present time, the specific physiological function of the PI 3-kinase is not known. Various studies have demonstrated that PI 3-kinase activity is correlated with cell growth and

Activation of PI 3-Kinase Activity byCdc42Hs Binding top85

18730

determining the extent to which these interactions occur in vivo and areregulated in different cell types, and if there is a relationship between the abilities of Cdc42Hs and Racl to bind to p85 and stimulate PI 3-kinase activity and the abilities of these GTP-binding proteins to bind to other protein kinases (PAK and ACK).

A “PIP

Acknowledgments-We thank Steve Taylor for helpful discussion, Lewis Cantley for providing the p85a cDNA, Alan Hall for providing the Racl and RhoA cDNAs, and Jeff Settleman for the p190 monoclonal antibody. We also thank Cindy Westmiller for expertsecretarial assistance. REFERENCES

B -PIP

1

0

0

e

e

2

3

4

5

-0

FIG.4. A , association of PI 3-kinase activity from PC12 cells with immobilized wild type and mutant Cdc42Hs. Lysates from PC12 cells were incubated with immobilized GST (lane I ) , GST-Cdc42Hs-GDP (lane 2 ) or GTPyS (lane 3 ) , GST-Cdc42HsT35A-GTPyS(lane 4 ) , and GST-Cdc42HsQGlLGTP (lane 5),and PI 3-kinaseactivity was assayed as described under “Experimental Procedures.” The lipids were extracted by chloroform and separated by TLC. The spotting origin is indicated by 0. Results are representative of three independent experiments. B , stimulation of PI 3-kinase activity from PC12 cells with wild type and mutant Cdc42Hs. p85 was immunoprecipitated from PC12 lysates as described under “ExperimentalProcedures.” Immunoprecipitates were mixed with GST (lane I ) , GST-Cdc42Hs-GDP (lane 2 ) or GTPyS (lane 3 ) , GST-Cdc42HsT35A-GTPyS (lane 4 ) . and GSTCdc42HsQGlGGTP (lane 5),and PI 3-kinase activity was assayed as described under “ExperimentalProcedures.” Results are representative of three independent experiments.

transformation (Cantley et al., 1991), as well as with cytoskeleta1 assembly (Eberle et al., 1990; Zhang et al., 1992).Likewise, the Rho-subtype proteins have been implicated in cytoskeletal organization, with RhoA being essential for stress fiber formation (Ridley and Hall, 19921, Racl for membrane ruffling (Ridley et al., 1992), and Cdc42Sc for bud site assembly in S. cerevisiae (Drubin, 1991). Therefore, it is tempting to speculate that theinteractions of the GTP-bound Cdc42Hsor Racl with p85 influences cytoskeletal events. It is also interesting that the PI 3-kinase has been suggested to participate in vesicular trafficking (Schu et al., 1993) and intracellular transport activities (Kapeller et al., 1993) and recently, mammalian Cdc42 has been found to accumulate in the Golgi in a brefeldin Adependent manner, thus suggesting a further possible connection between Cdc42 and ~ 8 5Future . ~ studies will bedirected a t J. Erickson and R. Cerione, unpublished results.

Backer, J. M., Myers, M. G., Shoelson, S. E., Chin, D. J., Sun, X. J., Miralpeix, M., Hu, P., Margolis, D., Skolnik, E. Y., Schlessinger, J., and White, M.F. (1992) EMBO J. 11,3469-3479 Barfod, E. T., Zheng, Y., Kuang, W. J., Hart, M. J., Evans, T., Cerione, R. A., and Ashkenazi, A. (1993) J. Bid. Chem. 268,26059-26062 Boguski, M. S., and McCormick, F. (1993)Nature 366,643-654 Cantley, L. C., Auger, K. R., Carpenter, C., Duckworth, B., Graziani, A., Kapeller, R., and Soltoff, S. (1991) Cell 64,281-302 Carpenter, C. L., Anger, K. R., Chanudhuri, M., Yoakim, M., Schafiausen, B., Shoelson, S.,and Cantley, L. C. (1993)J. B i d . Chem. 268,9478-9483 Diekmann, D., Brill, S., Garrett, M. D., Totty, N., Hsuan, J., Monfries, C., Hall, C., Lim, L.,and Hall, A. (1991) Nature 351,400-402 Drubin, D. (1991) Cell 65, 1093-1096 Eberle, M., Traynor-Kaplan, A. E., Sklar, L. A,, and Norgauer, J. (1990) J. B i d . Chem. 265,16725-16732 Fukui, Y., and Hanafusa, H. (1989) Mol. Cell. Biol. 9, 1651-1658 Hart, M. J., Shinjo,R, Hall, A,, Evans, T., and Cerione, R. A. (1991)J. Bid. Chem. 266,20840-20848 Hart, M. J., Maru, Y., Leonard, D., Witte, O., Evans, T., and Cerione, R. A. (1992) Science 258,812-815 Hart, M. J., Eva, A,, Zangrilli, D., Aaronson, S. A,, Evans, T., Cerione, R. A., and Zheng, Y. (1994) J. B i d . Chem. 269,62-65 Heisterkamp, N., Stam, K., Groffen, J., Dekline,A., and Gronveld, G. (1985)Nature 315,758-761 Johnson, D. I., and Pringle, J. R. (1990) J. Cell Biol. 111,143-152 Kapeller, R., Chakrabarti, R., Cantley, L., Fay, F., and Corvera,S.(1993) Mol. Cell. Biol. 13,6052-6063 Kumagai, N., Morii, N., Fujisawa, R,Nemoto, Y., and Narumiya,S.(1993)J. Biol. Chem. 268,24535-24538 Leonard, D., Hart, M. J., Platko, J. V.,Eva, A., Henzel, W., Evans, T., and Cerione, R. A. (1992) J. B i d . Chem. 267,22860-22868 Manser, E., Leung, T., Snlihuddin, H., Tan, L., and Lim, L. (1993) Nature 363, 364-367 Manser, E., Leung, T., Salihuddin, H., Zhao, Z.-S., and Lim, L.(1994) Nature 367, 40-46 Marshall, M. S. (1993) Denends Biochem. Sci. 18,250-254 Miller, P. J., and Johnson, D. I. (1994) Mol. Cell. B i d . 14, 1075-1083 Otsu, M., Hiles, I., Gout, I., Fry, M. J., Ruiz-Larrea, F., Panayotou, G., Thompson, A,, Dhand, R., Hsuan, J., Totty, N., Smith, A. D., Morgan, S. J., Courtneidge,S. A., Parker, P., and Waterfield, M. D. (1991) Cell 65,91-104 Pleiman, C. M., Hertz, W. M., and Cambier, J. C. (1994) Science 263, 1609-1612 Ridley, A. J., and Hall, A. (1992) Cell 70, 389-399 Ridlev, A. J., Paterson, H. F., Johnston, C. L., Diekmann, D., and Hall, A. (1992) Cell 70,401-410 Ridley, A. J., Self, A., Kasmi, F, Paterson, H. F, Hall, A,, Marshall,C. J., and Ellis, C. (1993) EMBO J. 12,5151-5160 Schu, P. V., Takegawa, IC, Fry, M. I., Stack, J. H., Waterfield, M. D., and Erm, S. D. (1993) Science 260.88-91 Settleman, J., Albhght,C.F., Foster, L. C., and Weinberg, R. A. (1992)Nature 359, 153-154 Ueda. T.. Kikuchi. IC. J.. and Takai,Y. (1990) J. B i d . Chem. . Ohza. - . N..Yamamoto, . 265,9373-9380 Zhang, J., Fry, M. J., Waterfield, M.D., Jaken, S., Liao, L., Fox, J. E. B., and Rittenhouse, S. E. (1992) J. Bid. Chem. 267,4686-4692 Zhang, J., King, W. G., Dillon, S., Hall, A., Frig, L., and Rittenhouse, S. E. (1993) J. B i d . Chem. 268,22251-22254 Zheng, Y., Hart, M. J., Shinjo, K., Hall, A,, Evans, T., Bender, A,, and Cerione, R. A. (1993) J. Biol. Chem. 268,24629-24634 Ziman, M., Preuss, D., Mulholland, J., O’Brien. J. M.. Botstein. D.. and Johnson,D. I. (1993)Mol. Biol. Cell 4, 1307-1316 ~