A novel functional target for tumor-promoting phorbol esters and ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY Val. 268, No. 15,Issue of May 25, pp. 10709-10712.1993 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Communication

Printed in U.S.A.

PKC and if administered to cells for prolonged periods cause it tobe down-regulated (4,5). These characteristicsof phorbol esters have been exploited to dissect the role of PKC in signal transduction pathways (6). Lysophosphatidic acid (LPA) can act as a mitogen (7) and affects the shape of PC12 cells (8) and neuronal cells (9). In addition, LPA has recently been THE p2lrac-GTPase ACTIVATING PROTEIN shown to be the component of serum that activates actin n-CHIMAERIN* stress fiber formation, possibly via the rm-related protein rho (Received for publication, January 8, 1993, and in revised form, (10). To date, no functional target for LPA is known, although February 24, 1993) a 38-40-kDa protein that binds LPA has recently been idenSohail Ahmed$§ll, Joel Lee$§, Robert Kozma$§, tified using a photoaffinity label for detection (9). Anthony Bests, Clinton Monfriess, and Louis Lim$§ The PKC family of enzymes consist of regulatory (Cl/C2) and catalytic (C3/C4) domains. The C1 cysteine-rich region From the SZnstitute of Molecular and Cell Biology, National Uniuersity of Singapore, Kent Ridge Crescent, of PKC is responsible for phospholipid-dependent phorbol Singapore 051 1 and the SZnstitute of Neurology, 1 ester binding (11).The N-terminal region of n-chimaerin has Wakefield Street, London WClN lPJ, United Kingdom approximately 50% amino acid sequence identity with the cysteine-rich phorbol ester binding domain of PKC (12, 13). Phorbol esters are potent tumor promoters widely The characteristics of phorbol ester binding to n-chimaerin used for investigating mechanisms of cell transforma- and PKC are similar, including phospholipid and zinc detion with protein kinase C (PKC) generally considered pendence, stereospecificity, high affinity, and competition as being their only protein target. Lysophosphatidic acid (LPA) can act as a mitogen, affecting cell shape with diacylglycerol (13, 14).Proteins that also possess a and the actin cytoskeleton. There is no identified func- cysteine-rich domain with sequence identity to PKC include tional target for LPA. We have isolated a cDNA encod- the Caenorhabditis elegans UNC-13 gene product, 80-kDa ing a protein n-chimaerin that is a high affinityphorbol diacylglycerol kinase, andthe oncogene products RAF (a ester receptor and a p2lrac-GTPase activatingprotein serine/threonine kinase) and VAV (a potential GTP/GDP (rue-GAP). p2 lrac isa memberof the ras superfamily exchange protein for the rm-related rho family). Only UNCof small molecular weight GTP-binding proteins, 13, in addition to n-chimaerin, has been shown to be a specific which stimulates actin microfilament formation in phorbol ester receptor (15). From work on PKC, the overall Swiss 3T3 cells and superoxide production by the neu- function of the cysteine-rich domain appears to be to confer trophil oxidase. We now show that the ruc-GAP activ- phospholipid/phorbol ester sensitivity on enzyme activity. ity of n-chimaerin is stimulated by phosphatidylserine The C-terminal region (catalytic region) of n-chimaerin has (PS) and phosphatidic acid (PA) and that phorbol esters sequence similarity to Bcr (product of breakpoint cluster can synergize with PS and PA. LPA, in contrast, was region gene involved in Philadelphia chromosome translocafound to inhibit n-chimaerin. The phospholipid/phor- tion), rho-GTPase activating protein (rho-GAP), phosphatibo1 ester modulation of the ruc-GAP activity requires dylinositol3-kinase (p85 a / @subunits), abl-SH3-binding prothe PKC-like cysteine-rich domainof n-chimaerin. tein (3BP-1), and theras-GAP-binding protein p190 (12, 16Thus, n-chimaerin is a novel functional target (distinct 19). n-Chimaerin, Bcr, rho-GAP, and p190 are GAPS for the from PKC) for both phorbol esters and LPA. These ras-related protein rac (16, 20). data suggest that the physiological role of n-chimaerin The domain structure of n-chimaerin resembles that of the is to link events initiating the at cell surface/membrane PKC family, suggesting that n-chimaerin may also be a phoswith p2lrac effector pathways. pholipid/phorbol ester-sensitive enzyme. Here we show that the PKC-like cysteine-rich domain of n-chimaerin allows regulation of its rac-GAP activity. More specifically, we find that phosphatidylserine (PS) and phosphatidic acid (PA) Tumor-promoting phorbol esters, analogs of the naturally occurring lipid diacylglycerol (DG),’ cause a variety of phys- activate n-chimaerin, phorbol esters can synergize with these iological changes when administered to cells and tissues (I), phospholipids, and that LPA inhibits n-chimaerin. Phosphawith the phospholipid-dependent protein kinase C (PKC) the tidylinositol lipids (PI, PIP, PIP21 and arachidonic acid (aa) only known functional target (2, 3). Phorbol esters activate also inhibit n-chimaerin. These results suggest that n-chimaerin is a novel functional receptor for both phorbol esters *This work was supported in part by the Glaxo-Singapore Re- and certain phospholipids that link changes in the levels of search Fund and the Brain Research Trust. The costs of publication these molecules with rac effector pathways.

A Novel Functional Target for Tumor-promoting PhorbolEsters and Lysophosphatidic Acid

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. n To whom correspondence should be addressed Dept. of Neurochemistry, Institute of Neurology, 1 Wakefield St., London WClN l P J , United Kingdom. Tel.: 071-278-1552; Fax: 071-278-7045. ’The abbreviations used are: DG, diacylglycerol; PKC,protein kinase C; LPA, lysophosphatidic acid; GAP, GTPase activating protein; PS, phosphatidylserine; PA, phosphatidic acid; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PI, phosphatidylinositol; PIP, phosphatidylinositol 4-phosphate; PIP2, phosphatidylinositol 4,5-bisphosphate; aa, arachidonic acid PDBu, phorbol12,13-dibutyrate; PMA, phorbol 12-myristate 13-acetate.

MATERIALS AND METHODS

Subcloning, Expression, and Purification of Proteins-The HindIIIl Sal1 fragment of racl cDNA was blunt-ended and subcloned into the Klenow-repaired SpeI site of p265 (a pGEX-2Tderivative containing an extended polylinker; a gift from Kees Vink and Dick Schaap, The Netherlands Cancer Institute). Rat n-chimaerin cDNA was ligated as two EcoRI fragments (21) and then a FokIIBalI fragment (nucleotides 416-1534) was cloned into the Klenow-repaired EcoRI site of pGEX-2T vector to generate glutathione S-transferase fusion proteins. The C-terminal construct has been previously described (14). Escherichia coli (E. coli) Y1091’”“-cells transformed with one of these

10709

10710

Phorbol Esters and Phospho'lipidsModulate n-Chimaerin

plasmids was grown overnight, and proteins were purified as previ- cofactors for activity butdoes reduce the specific activity. ously described (14, 16). The rucl protein was used throughout this In the absence of lipidcofactors, PKC is inactive. One study. possible explanation for this difference between n-chimaerin Measurement of GTP HydrolysisRates of rad-The GTPase activfully ity of rucl was measured with [r-32P]GTP (30 Ci/mmol; 5 p ~ ) and PKC is that the recombinant n-chimaerin is not folding essentially as described previously (16). All rates measured up to 50% coupled allowing leakyactivity, due either to incorrect GTP hydrolysis were linear. Proteins used were greater than 90% of E. coli expressed protein or the lack of eukaryotic postpure. translational modifications. To address this possibility, nLiposome Sedimentation Assay-2pgof n-chimaerin protein in chimaerin cDNAwas subcloned into the pMT-2 vector, transTris-HCI, pH 7.5, 100 mM KCl, 2 mMMgC12 were incubated with liposomes at 200 pg/ml for 1 h at 37 "C. Liposomes of LPA, PS, PA, fected into COS-7 cells, and cell extracts prepared. Protein PC, and PEwere freshly prepared as follows; chloroform solutions of obtained from extracts of mammalian cells is more likely to made phospholipids (1 mg/ml) were dried under vacuum, rehydrated in 20 have the characteristicsof native material. n-Chimaerin mM Tris-HCI, pH 7.5 (1 mg/ml final concentration), and sonicated up approximately 1%of total protein from such extracts' and until the solution was clear. To separate the liposomes the suspension at this level, if endogenously active, would be expected to was centrifuged at 95,000 rpm for 20 min in a Beckman TL-100 increase the total GAP activityof cell extracts. Phorbol ester ultracentrifuge. 20 r l of the pellet and supernatantwere then analyzed by SDS-polyacrylamide gel electrophoresis followedby Coomassie binding to n-chimaerin transfected COS-7 cells was compaBlue staining. The amount of protein present was estimated by image rable to COS-7 cells transfected with PKC-E and 4-5-fold higher than mock-transfected cells (data not shown). Howanalysis as described in Ref. 14.

ever, COS-7 cell extracts containing n-chimaerin, at concen1pg, did notaffect the intrinsic GTPase activity trations up to of racl, suggesting that the COS-7 proteinmay require lipid The N terminus of n-chimaerin has approximately 50% cofactors for activity. amino acid sequence identity with the cysteine-rich domain Phospholipids and Arachidonic Acid Modulate n-Chimaof PKC and also binds phorbol esters. However, n-chimaerin erin-The phospholipids phosphatidylcholine (PC) and phosis not a protein kinase but rather a rac GTPase activating phatidylethanolamine (PE) did notaffect the rac-GAP activprotein (rac-GAP). This comparison suggests that the rac- ity of n-chimaerin (Fig. 2a). However, phosphatidylserine GAP activity of n-chimaerin may be regulated in a manner (PS) and phosphatidic acid (PA) clearly stimulated n-chimasimilar to thekinase activity of PKC. Two functionalpredic- erin (Fig. 2b). The substratespecificity for phospholipids PC, tions for the rac-GAP activity of n-chimaerin can be made PE, PS, and PA is similar to that reported for PKC-c (25). (24). First, the basal activityof n-chimaerin should be negli- Since LPA affects the shape of PC12 and neuronal cells (8, gible in the absence of phospholipid cofactors, and, second, 9), and neuroblastoma cells stably transfected with n-chimacertain phospholipids and phorbol esters (PDBu and PMA) erin cDNA exhibit distinct changes in morph~logy,~ we specshould regulate n-chimaerin through its cysteine-rich domain. ulated that n-chimaerinmay be regulated by LPA. Strikingly, Recombinant n-Chimaerin Purified from E. coli Has Basal LPA was found to inhibit rac-GAP activity (rather than to GAP Activity That Is Increased by Removal of the Cysteine- stimulate it). To assess more effectively this inhibition we rich Domain-To assess the role of the PKC-like cysteine- increased the concentrationof n-chimaerin to 680 ng so that rich domain in regulating the GAP domain recombinant n- the ts0%for racl was 3 min (Fig. 2c). LPA at 258 pg/ml was chimaerin full-length (330 amino acid residues) and C-tersufficient to reduce the t5,,%of rucl-GTP, in the presence of minal (200 amino acid residues)glutathioneS-transferase n-chimaerin, by 5 min. The stimulation by PS and PA and fusion proteins were compared. Fig. 1 shows the effect of inhibition by LPA was dose-dependent (Fig. 2). Thecysteinedifferent concentrations of these recombinant n-chimaerin rich domain was required for PS, PA, and LPA toregulate nproteins on the intrinsic racl GTPase activity. The results chimaerin, as these lipids did not affect the GAP activity of are presented as t50%(time taken for 50% hydrolysis of the the C-terminal protein. The intrinsic GTPase activity of racl bound GTP). Without any additions the tS0%for rac is 12.5 was also not affected by these phospholipids (data not shown). min at 15 "C. Full-length n-chimaerin has rac-GAP activity, A screen of other lipids, including some known to inhibit which is linearly related to protein concentration. The C- the catalytic domainsof ras-GAP and NF-1,was next underterminal protein has a higher specific activity, which begins taken. At 100 pg/ml,phosphatidylinositol (PI), phosphatidylto saturate above at 0.1 p ~ Thus, . the presence of the cysteine- inositol4 phosphate (PIP), phosphatidylinositol 4,5-bisphosrich domain does not lead to anabsolute requirementfor lipid phate (PIP,), and arachidonic acid (aa) inhibited the rac-GAP activity of n-chimaerin (Fig. 3). Interestingly, the phospholipids required the cysteine-rich domain for inhibition, while aa did not.Arachidonic acid was the most potent of the lipid inhibitors of n-chimaerin giving 100% inhibition at 100 pg/ ml. Arachidic acid, however, was without effect (Fig. 3). Phorbol Esters Synergize with PS and PA in Activation of n-Chimaerin-When n-chimaerin was preincubated with either phorbol 12,13-dibutyrate (PDBu)or phorbol 12-myristate 13-acetate (PMA) in the presenceof saturating concentrations of PS and PA (100 wg/ml) before the t50% for raclG T P was determined, no additional stimulationby the phorbo1 esters was seen. To optimize the conditions forsynergism, we next chose limiting PA and PS concentrations (5 wg/ml) '= I, I) 0.1 11.2 11.3 11.4 0.5 that did not affect n-chimaerin activity. Under these conditions, there is a synergism between phospholipidsand phorbol I,rolciu esters (Table I). The largest decrease in t50%was seen with FIG. 1. Protein titration of n-chimaerin rac-GAP activity. the PA/PMA couple. PMA stimulation was dependent on PA RESULTS

+ IU~CTIIIIBII~.~~

n-Chimaerin full fusion (m) and C-terminal fusion (0)protein were added at various concentrations between 0 and 0.45 p M and the rate of rucl mediated GTP hydrolysis measured (see "Materials and Methods"). Similar results were obtained inthree other experiments.

* R. Kozma and S. Ahmed, unpublished data. R. Kozma, A. Best, and S. Ahmed, manuscript in preparation.

Phorbol Esters and Phospholipids Modulate n-Chimaerin

10711

E

FIG. 2. Effect of phospholipids on

the rac-GAP activity of n-chimaerin. a, PC (A) andPE (A);b, PA (B) and P S (0);c, LPA (O),GTP hydrolysis rates ofrac were measured as described under“Materials and Methods.” n-Chimaerin protein was incubated with different concentrations of phospholipid prior to the assay. n-Chimaerin was present in a and b at 250 ng and in c at 680 ng. Similar results were obtained in three other experiments.

g,,,

.‘“It

z

E

9

W

,

$x

2 9

3 f

~

111

tt

ill1

60

10

1011

I20

phospholipids microgramlml

I

1211,

,

2

Kc,.

4

1.P.i

5

FIG. 3. Effect of PI, PIP, PIPz, aa, and arachidic acid on rac-GAP activity of n-chimaerin. The GAP activity was measured with either full-length (B) or C-terminal n-chimaerin (@). Additions (100 pg/ml) were as follows: I , PI; 2, PIP; 3, PIP,; 4, aa; 5, arachidic acid. Full-length and C-terminal n-chimaerin were incubated at 500 and 230 ng, respectively. GAP activity was measured as described under “Materials and Methods.” Similar results were obtained in two other experiments.

TABLEI Synergism between phospholipids and phorbol esters in actiuating n-chimaerin When present, concentrationwas as follows:PS and PA (5 /*g/ml), PMA and PDBu (both at 100 nM), n-chimaerin (125 ng).These reagents were preincubated for 30 min before measuring the rate of GTP hydrolysis of racl as described in Fig. 1. Addition of the solvent used to dissolve phorbol esters (acetonitrile/water, 2:3, v/v) to either PS or PA did not affect rac-GAP activity of n-chimaerin. Values are averages f S.D. from three separate experiments. Addition

f60%

rnin

None n-Chimaerin n-Chimaerin/PA n-Chimaerin/PA/PDBu n-ChirnaerinlPAIPMA n-ChimaerinlPS n-ChimaerinlPSIPDBu n-ChimaerinJPSJPMA

PA

PS

PC

PE

,,l,ci,l’l,olipid

1111111111.r,

11.9 f 0.1 10.0 f 0.0 10.3 f 0.4 8.7 f 0.1 7.3 f 0.5 9.9 f 0.1 8.8 & 0.6 8.6 f 0.2

FIG. 4. Physical interaction of n-chimaerin with phospholipids requires the PKC-like cysteine-rich domain. Partitioning of full-length n-chimaerin into liposomes of LPA, PA, PS, PC, and PE. B, liposomes; @, supernatant. 100% of C-terminal n-chimaerin protein remained in supernatant and therefore did not physically interact with liposomes from any phospholipid (see “Materials and Methods”). Similar results were obtained in two other experiments.

terminal n-chimaerin glutathione S-transferase fusion proteins were incubated with liposomes of different phospholipid composition (PC, PE, PS, PA, or LPA) for 1 h, and the two phases were then separated by centrifugation. Full-length nchimaerin fusion protein sedimented with liposomes formed from all these phospholipids apart from PE (Fig. 4). The Cterminalproteindidnotsedimentwith liposomes of any composition. Although PC physically interacted, it did not affect n-chimaerin activity (seeFig. 2u). Thus, physical interaction with phospholipids is not sufficient for modulation of the ruc-GAP activity of n-chimaerin and more specific interactions, determinedby the chemical structure of the phospholipid, are required. DISCUSSION

n-Chimaerin has overall structural similarity to the PKC11 (6-1) family of enzymes in that it has a cysteine-rich phorbol ester binding domain anda catalytic domain (12, 13). This comparison highlights the possibility that there is also similarity between n-chimaerin and PKC in the mechanism or PS. aPMA, a stereoisomer of PMA that does not bind t o of regulation of enzyme activity. The following findings supor regulate PKC,was ineffective.However, unlike the phorbol port this hypothesis; (i) deletion of the cysteine-rich domain esters, DG did not synergize with PS or PA at either low or increases n-chimaerin-specific activity, (ii) phospholipids (PS high concentration of phospholipid (data not shown). This and PA) activate ruc-GAP activity of n-chimaerin, (iii) phormay have resulted from use of suboptimal conditions to obtainbo1 esters can synergize with phospholipids in activating nsynergism between PS/PA and DG. Further work, including chimaerin, and (iv) the cysteine-rich domain is required for the use of purified eukaryotically expressed or native racl and phospholipids tointeractwithand regulate n-chimaerin. n-chimaerin protein, will be necessary to see if DG regulates Thus,thecysteine-richdomain of n-chimaerinandPKC n-chimaerin (see “Discussion”). LPA a t 258 wg/ml was able appear to fulfill similar roles with similar cofactor specifici(100 wg/ml) or ties. It will be interesting to see if a PKC-like regulation to reverse the stimulation seen with either PA PA/PDBu ( 5 pg/m1/100 nM). occurs in other proteins thatpossess a cysteine-rich domain, n-Chimaerin Physically Interacts with Phospholipids-To including the oncogene products RAF and VAV, UNC-13, examine whetherphysical interaction with phospholipids was and 80-kDa diacylglycerol kinase (15). If this turns out tobe sufficient for modulation of the rac-GAP activity of n-chi- the case, these proteins will represent new functional targets maerin a sedimentation assay was used. Full-length and C- for phospholipid/phorbol ester second messengers.

10712

Phorbol Esters and Phospholipids Modulate n-Chimaerin

n-Chimaerin appears tobe regulated by a PKC-like mech- as the only phorbolester receptor. Subsequently,phorbol anism. However, unlike phorbol esters, DG did notsynergize esters have beenused to modulate specifically and downwithphospholipids to stimulate its rac-GAP activity. One regulate PKC to which all their cellular effects have been possible explanation for this is that the assay conditions were attributed. We have previously shown that n-chimaerin is a not optimal for synergism. Post-translational modifications specific phospholipid-dependent phorbol ester receptor (13). of both racl and n-chimaerin may be necessary to observe The results presented here demonstrate the functional conactivation by DG. Racl is isoprenylated at the C-terminal sequence of that binding. Tumor-promoting phorbol esters CAAX motif (where A is aliphatic amino acid; Ref. 26), and and phospholipids regulate n-chimaerin and thereby control this may affect interactions with lipid environments. Indeed, its potency todown-regulate rac. post-translationally modified racI ismuch more potent in In conclusion, the physiological implication of these findactivating neutrophil oxidase than recombinant racl purified ings is that the chimaerin family of rac-GAPS may allow actin from E. coli, and is essential for racl to interact with rho- microfilament structures and other raceffectors, such as the GDP-dissociation inhibitor(27). The specific type of DG may rho-signal transduction pathway and neutrophil oxidase, to also be an important factor. Another possibility is that two sense changes occurringat the cell surface/membrane. cysteine-rich domains may be necessary for DG to activate Acknowledgments-We thank Kees Vink and Dick Schaap for enzyme activity. In supportof this is the finding that PKC-{, like n-chimaerin, possesses onlyone cysteine-rich domain andpGEX-2T derivative p265. is not activatedby DG (28). Nakanishiet al. (29) haveshown REFERENCES that phosphatidylinositol 3,4-bisphosphate and phosphatidyl-1. Diamond, L., O’Brian, T. G. & Baird, W. M. (1980)Ado. Cancer Res. 32, 1-74 inositol 3,4,5-trisphosphate are potent activators of PKC-{. 2. Nishizuka. Y. (1984)Nature 308.693-698 We are presently investigating whether these lipids can also 3. Nishizuka; Y. (1988)Nature 334;661-665 4. Ballester, R. & Rosen, 0. M. (1985)J. Biol. Chem. 260, 15194-15199 modulate n-chimaerin. 5. Young, S., Parker, P. J., Ullrich, A. & Stabel, S. (1987)Biochem. J. 244, The rac-GAP activity of n-chimaerin was inhibited by LPA, 77.5-7711 6. Stabel, S. & Parker, P. J. (1991)Pharmacol. Ther. 51, 71-95 PI, PIP, PIP2, and aa. LPA, PI, PIP, and PIPprequired the 7. van Corven, E. J., Groenink, A,, Jalink, K., Eichholtz, T. & Moolenaar, W. cysteine-rich domain to modulate n-chimaerin, whereas aa 11989) EMBO J. 59.45-54 ~, .~ .. ~~, did not.Arachidonic acid is a potent inhibitor of the catalytic 8. Tigyi, G. & Miledi, R. (1992)J.Biol. Chem. 267,21360-21367 9. van der Bend, R., Brunner, J., Jalink, K., van Corven, E. J., Moolenaar, domain of both NF-1 (neurofibromatosistype 1 gene product) W. H. & van Blitterswiik. W. J. (1992)EMBO J. 7.2495-2501 and p120 ras-GAP (30, 31). LPA was also able toreverse the 10. Ridley, A. & Hall, A. (19g2)Cell 70, 389-410 11. Ono, Y., Fujii, T., Igarashi, K., Kuno,T.,Tanaka, C., Kikkawa, U. & stimulation seen by PA and PA/PDBu, suggesting that the Nishizuka, Y. (1989)Proc. Natl. Acad. Sci. U. S. A . 86,4868-4871 overall rac-GAP activityof n-chimaerin is determinedby the 12. Hall, C:, Monfries, C., Smith, P., Lim, H. H., Kozma, R., Ahmed, S., Vanmasingham, V., Leung, T. & L m , L. (1990)J. Mol. Biol. 211, 11level of competing regulators. Thus, additional complexity __16 overlays the PKC-like mechanism of lipid modulation of n- 13. Ahmed, S., Kozma, R., Monfries, C., Hall, C., Lim, H. H., Smith, P. & Lim, L. (1990)Biochem. J. 272, 767-773 chimaerin. A large number of GAPS of different substrate 14. Ahmed, S., Kozma, R., Lee, J., Monfries, C., Harden, N. & Lim, L. (1991) Biochem. J. 280,233-241 specificities have now been described (32).One reason for this 15. Ahmed, S. Marayuma, I. N., Kozma, R., Lee, J., Brenner, S. & Lim, L. diversity could be to allow different forms of regulation. (1992)Biochem. J. 287,995-999 n-Chimaerin cDNA was isolated from a retinal library and 16. Diekmann, D., Brill, S., Garrett, M. D., Totty, N., Hsuan, J., Monfries, C., Hall, C., Lim, L. & Hall, A. (1991)Nature 361,400-402 shown to have brain and testes specificity in expression (12). 17. Otsu, M., Hiles, I., Gout, I., Fry, M. J., Ruiz-Larrea, F., Panaytou, G., Thom son, A.., Dhand, R., Hsuan, J., Totty, N., Smith, A. D., Morgan, More recent data suggest that there exists a family of chiS. J., Eourtneldge, S. A., Parker, P. J. & Waterfield, M. (1991)Cell 66, maerin molecules having different expression patterns. Iso91-104 forms /3 and y, expressed in the testes and lymphocytes,4 18. Cicchetti, P., Mayer, B. J., Thiel, G. & Baltimore, D. (1992)Science 257, 803-806 respectively, have been cloned (33). Moreover, further North- 19. Settleman, J., Narasimhan, V., Foster, L. C. & Weinberg, R. A. (1992)Cell 69,539-549 ern analysis and in situ hybridizationhave revealedchimaerin 20. Settleman, J., Albright, C. F., Foster, L. C. & Weinberg, R. A. (1992)Nature signals in lung and intestine.s The finding that the rac-GAP 359, 153-154 activity of n-chimaerin (chimaerin-a) is modulatedby lipids 21. Lim, H. H., Michael, G. J., Smith, P., Lim, L. & Hall, C. (1992)Biochern. J. 287,415-422 may therefore have relevance to non-brain cells and tissues. 22. Deleted in uroof It has been speculated that the ras-related rho subfamily 23. Deleted in proof 24. Bell, R. M. & Burns, D. J. (1991)J . Biol. Chem. 266,4661-4664 (34). 25. Koide, H., Ogita, K., Kikkawa, U. & Nishizuka, Y.(1992)Proc. Natl. Acad. are involved in regulating actin cytoskeletal organization Sci. U. S. A . 89, 1149-1153 Recent work has shown that microinjection of rac and rho J. R.,,Uhing, R. J. & Synderman, R. (1990)Biochem. Biophys. proteins into Swiss 3T3 cells leads to changes in actin micro- 26. Didsbu? Res. ornrnun 171,804-812 27. Ando, S., Kaibuchi, K., Sasaki, T., Hiraoka, K., Nishiyama, Y., Mizuno, filaments associated with membrane ruffling and stress fiber T., Asada, M., Nunoi, H., Matsuda, I., Matsuura, Y., Polakis, P.,Mcformation, respectively (10, 35). In particular, LPA has been Cormick, F. & Takai, Y. (1992)J. Biol. Chem. 267, 25709-25713 suggested to activate“rho-signal transduction pathways.” Our 28. Ono, Y., Fujii, T., Ogita, K.,, Kikkawa, U., Igarashi, K. & Nishizuka, Y. (1989)Proc. Natl. Acad. SCL.U. S . A. 86,3099-3103 data support theidea that the chimaerinfamily of rac-GAPS 29. Nakanishi,. N.,. Brewer. K. A. & Exton, J. H. (1993)J. Biol. Chern. 268, 13-16 provide the link between LPA/second messengers and actin 30. Golubic, M., Tanaka, K., Dobrowolski, S., Wood, D.,Tsai, M.-W., Marshall, organization throughtheprotein rac (36).Since rac can M., Tamanoi, F. & Stacey, D. (1991)EMBO J. 10, 2897-2903 activate the rhu-signal transduction pathway (35), the effect 31. Bollas G. & McCormlck. F. (1991)Nature 351.576-579 32. Fry, a.J. (1992)Curr. Bid. 2 , 78-80 of LPA on actin stressfibers might be rac-GAP-mediatedby 33. Leune. T..How. B.-E.. Manser. E. & Lim. L. (1993)J. Biol. Chem. 268. a macromolecule such as n-chimaerin. 34. Paterson, H. F., Self, A. J., Garrett, M. D., Just, I., Aktories, K. & Hall, A. Since the discovery that phorbol ester binding co-purified (1990)J. Cell Biol. 11 1, 1001-1007 with PKC activity (37), PKC has been viewed in most studies 35. Ridley, A,, Paterson, H. F., Johnston, C.L., Diekmann, D. & Hall, A. ~

T. Leung, E. Manser, and L. Lim, unpublished data. G. Micheal, C. Hall, and L. Lim, unpublished data.

~~

(1992)Cell 70,411-418 36. Downward, J. (1992)Nature 359,273-274 37. Castagna, M., Takai, Y., Kaibuchi, K., Sano, K., Kikkawa, U. & Nishizuka, N. (1982)J. Btol. Chern. 257,7847-7851