the small gtp-binding protein rhoa regulates c-jun by a rock-jnk signaling axis

Molecular Cell, Vol. 13, 29–41, April 9, 2004, Copyright 2004 by Cell Press

The Small GTP-Binding Protein RhoARegulates c-Jun by aROCK-JNK Signaling Axis

miya, 1996; Takai et al., 2001). Rho GTPases also regu-late nuclear gene expression (Coso et al., 1995; Hill etal., 1995). Indeed, Rho proteins can promote normal andaberrant cell growth, and their transforming potential isoften associated with their ability to coordinate changes

Maria Julia Marinissen,1 Mario Chiariello,1,3

Tamara Tanos,2 Ora Bernard,4 Shuh Narumiya,5

and J. Silvio Gutkind1,*1Oral and Pharyngeal Cancer BranchNational Institute of Dental and Craniofacial Research

in the actin cytoskeleton with the regulation of geneNational Institutes of Healthexpression (for review see Zohn et al., 1998). AmongBethesda, Maryland 20892the growth-promoting genes regulated by RhoA are the2 Facultad de Ciencias Exactasc-jun and c-fos protooncogenes (Hill et al., 1995; Mari-Universidad de Buenos Airesnissen et al., 2001), which are members of the AP1 (acti-Buenos Airesvating protein-1) family of transcription factors that playArgentinaa key role in human cancer progression and metastasis3 Istituto di Endocrinologia ed Oncologia(Karin et al., 1997). AP1 is composed of Fos (c-Fos,Sperimentale, CNRFos B, Fra-1, and Fra-2) and Jun (c-Jun, JunB, JunD)Via Pansini 5proteins. Whereas homo and heterodimers of Jun pro-80131 Napoliteins can bind DNA directly, Fos members require theItalyinteraction with c-Jun or other Jun protein to act as4 The Walter and Eliza Hall Institute of Medical Researchtranscriptional activators. AP1 complexes then stimu-1G Royal Parade Parkvillelate transcription by binding to AP1 sites within the pro-Victoria 3050moter/enhancer region of a wide variety of genes, in-Australiacluding those encoding proteins involved in cell motility,5 Department of Pharmacologyinvasion, and growth (Karin et al., 1997). As such, theKyoto University Faculty of Medicineexpression and activity of these transcription factorsSakyo-ku, Kyoto 606-8501are tightly regulated.Japan

In the case of c-fos, its expression is controlled by acomplex array of promoter elements. Among these, theserum response element (SRE) plays a central regulatorySummaryrole (Treisman, 1995). The c-fos SRE binds several regu-latory molecules, including the serum response factorRhoA regulates the actin cytoskeleton and the expres-(SRF) and the ternary complex factor (TCF), which aresion of genes associated with cell proliferation. Thiscontrolled by two distinct mechanisms (Treisman, 1995).includes c-fos and c-jun, which are members of theExtracellular stimuli enhance the transcriptional activityAP1 family of transcription factors that play a key roleof the TCF through the activation of the mitogen acti-in normal and aberrant cell growth. Whereas RhoAvated protein (MAPK) family of proline-targeted serinestimulates the c-fos SRE by a recently elucidatedthreonine kinases, including extracellular regulated ki-mechanism that is dependent on actin treadmilling,nases (ERKs), Jun N-terminal kinases (JNKs), and p38show RhoA regulates c-jun is still poorly understood.(Whitmarsh et al., 1995, 1997), which phosphorylate theWe found that RhoA stimulates c-jun expressionTCF transactivation domain, thereby increasing its tran-through ROCK, but independently from the ability ofscriptional activity. Instead, the SRF is activated byROCK to promote actin polymerization. Instead, weRhoA independently from the activity of these MAPKs,found that ROCK activates JNK, which then phosphor-but as a consequence of the actin reorganization in-

ylates c-Jun and ATF2 when bound to the c-jun pro-duced by this GTPase. Specifically, the pathway linking

moter. Thus, ROCK represents a point of signal diver- RhoA to the SRF involves the activation of ROCK, whichgence downstream from RhoA, as it promotes actin subsequently phosphorylates LIM kinase (LimK). LimKreorganization and the consequent expression from phosphorylates the actin depolymerizing factor cofilin,the c-fos SRE, while a parallel pathway connects resulting in the stabilization of polymerized actinROCK to JNK, thereby stimulating c-jun expression. (F-actin) (Sotiropoulos et al., 1999). In certain cells, thisUltimately, these pathways converge in the nucleus to process requires the cooperating activation of the actinregulate AP1 activity. nucleation factor Diaphanous by RhoA (Geneste et al.,

2002). The consequent reduction of the monomeric actinIntroduction (G-actin) pool is then “sensed” by the SRF, which be-

comes activated (Miralles et al., 2003; Sotiropoulos etRho-related small GTPases, including RhoA, Rac1, and al., 1999).Cdc42, transduce extracellular signals into intracellular Recent evidence indicates that the regulation of c-junevents that lead to the remodeling of the actin cytoskele- expression by RhoA involves a distinct RhoA effector,ton, ultimately controlling the cell morphology and a PKN, which does not stimulate actin polymerization orvariety of functions such as cell motility, aggregation, the SRF, but activates instead a member of the MAPKpolarity, and contraction (Bar-Sagi and Hall, 2000; Naru- superfamily, ERK6 (p38�). In turn, c-Jun function is nec-

essary for the transforming activity of RhoA (Marinissenet al., 2001). However, as the ability of RhoA to promote*Correspondence: [email protected]

Molecular Cell30

nuclear gene expression often involves a complex inte- form of RhoA and a Rho GEF, PDZ-RhoGEF, which stim-ulates endogenous RhoA (Fukuhara et al., 1999) (Figuregration of cytoskeleton changes along with cytoplasmic

and nuclear events, it is still possible that RhoA may 2A). As a complementary approach, we used a dominantinterfering mutant for ROCK, ROCK KD.IA (Itoh et al.,require the coordinated activation of the PKN-ERK6 sig-

naling pathway together with those regulating the cy- 1999). As shown in Figure 2B, expression of ROCK KD.IAsuppressed LPA-induced stress fiber formation whentoskeleton to promote c-jun expression. In this regard,

the ability of RhoA to stimulate the expression from compared to the adjacent untransfected cells that servedas controls. Interestingly, ROCK KD.IA nearly abolishedSRE-containing genes often requires the coordinated

activation of signaling pathways regulating the TCF the accumulation of c-Jun in response to LPA. Consis-tent with these findings, ROCK KD.IA also reduced thethrough MAPKs (Whitmarsh et al., 1997). These observa-

tions prompted us to investigate whether ROCK analo- activation of pSRF-Luc and the c-jun promoter by RhoA(data not shown). These results indicated that ROCKgously participates in the regulation of c-jun expression

by RhoA, and if so, whether the underlying mechanism contributes to the signaling route by which RhoA stimu-lates the c-jun promoter, and raised the possibility thatis dependent on the status of actin polymerization.

In this study, we found that inhibition of ROCK nearly the formation of stress fibers, and the consequent deple-tion of the G-actin pool, may represent a common mech-abolishes the ability of RhoA to stimulate the c-jun pro-

moter and c-Jun expression. However, these effects anism by which RhoA and ROCK may control in parallelboth the SRF and the activity of the c-jun promoter.were independent from the ability of ROCK to promote

actin polymerization. Instead, we present evidence ofthe existence of a distinct signaling route by which RhoA An Activated Form of ROCK, ROCK��4,stimulates c-jun expression through ROCK by activating Stimulates the c-jun PromoterJNK, which leads to the phosphorylation and activation We next investigated whether an activated C-terminalof c-Jun and ATF2 that are bound to the jAP1 site on deletion mutant of ROCK, ROCK��4 (Sahai et al., 1999),the c-jun promoter. Thus, ROCK may represent a point could stimulate the c-jun promoter (Figure 3A). GFP andof signal divergence downstream from RhoA, as it medi- HA-tagged ROCK��4 and wild-type HA-ROCK fullates the reorganization of actin and expression from the length were readily expressed (Figure 3B). Both GFPSRE, while an independent pathway connects ROCK to (not shown) and HA-tagged ROCK��4 (Figure 3C) wereJNK, thereby stimulating c-jun expression. These two able to induce the c-jun promoter potently, in a dose-pathways ultimately converge in the nucleus to regulate dependent manner, and to an extent similar to that of theAP1 activity. pSRF-Luc reporter, whereas ROCK wild-type exerted a

much more limited activation. These results indicatedthat activated ROCK might be sufficient to stimulate theResultstranscriptional activity of the c-jun promoter.

Regulation of c-jun Expression by LPAand RhoA Requires ROCK ROCK Induces the Activity of the c-jun Promoter

through a Mechanism Independent of ActinTo evaluate the contribution of ROCK to the inductionof c-jun expression mediated by RhoA, we examined Cytoskeleton Rearrangements

To investigate whether LimK acts downstream fromwhether a highly selective ROCK inhibitor, Y-27632 (Na-rumiya et al., 2000), affects the expression of c-jun in ROCK in the pathway leading to c-jun expression, we

asked whether an activated form of LimK, LimK1.C,response to lysophosphatidic acid (LPA), a potent acti-vator of RhoA (Ridley and Hall, 1992). LPA elevated the could stimulate the c-jun promoter. LimK1.C induced

F-actin accumulation and potently stimulated the c-fosexpression of c-jun mRNA in NIH 3T3 cells and stimu-lated the activity of reporter plasmids under the control SRF (Figure 4A), as reported (Sotiropoulos et al., 1999;

data not shown). Surprisingly, however, it was unableof the c-jun promoter and the SRF, pJun-Luc, and pSRF-Luc, respectively (Figures 1A and 1B). The activation of to stimulate the c-jun promoter (Figure 4A). Moreover,

a dominant- negative form of LimK, LimK DN, blockedthe c-jun promoter and the consequent induction ofc-jun mRNA resulted in increased c-Jun protein levels stress fiber formation in LPA treated cells (Figure 4B),

but not the activation of the c-jun promoter (not shown)(Figure 1C). Surprisingly, pretreatment with Y-27632drastically diminished the accumulation of c-jun mRNA and c-Jun expression in response to LPA or an active

ROCK (Figure 4B), even if LimK DN blocked their abilityprovoked by LPA in a dose-dependent manner (Figure1A). However, it did not affect the stimulation of c-jun to stimulate the SRF-driven promoter (results not

shown). To confirm these findings, we used a dominant-expression induced by anisomycin, which served as aspecificity control. Y-27632 also inhibited the activation negative mutant of cofilin, Cofilin S3A, which blocks the

ability of RhoA to stimulate actin polymerization andof pSRF-Luc by LPA, as a control, and diminished theactivation of the c-jun promoter by LPA but not when the SRF (Sotiropoulos et al., 1999). Indeed, Cofilin S3A

effectively inhibited LPA-induced stress fiber formationprovoked by MEKK that acts directly upstream of JNK(Davis, 2000) (Figure 1B). In line with these observations, and the activity of pSRF-Luc, but failed to affect the

expression of c-Jun in response to LPA and ROCK (Fig-Y-27632 abolished the accumulation of c-Jun elicitedby LPA (Figure 1C), together supporting a critical role ures 4C and 4D; data not shown). Similarly, inhibition

of stress fiber formation by an actin depolymerizatingfor ROCK as part of the pathway by which LPA promotesc-Jun expression. agent, latrunculin B, that prevents the activation of the

c-fos SRE (Sotiropoulos et al., 1999), failed to inhibit theThe ROCK inhibitor also reduced the activation ofpSRF-Luc and pJun-Luc caused by both an activated accumulation of c-Jun in response to LPA (not shown).

RhoA Regulates c-jun Expression through ROCK-JNK31

Figure 1. The ROCK Inhibitor Y-27632 Prevents c-jun Expression by LPA

(A) Serum starved NIH 3T3 cells were left untreated or incubated for 2 hr with different concentrations of Y-27632, followed by 45 min treatmentwith 10 �M LPA or 10 �g/ml anisomycin. Total RNA was assessed to be equivalent in each lane by ethidium bromide staining of ribosomalRNAs, and analyzed by Northern blotting using 32P-labeled murine c-jun cDNA as a probe.(B) NIH 3T3 cells were transfected with pJun-Luc or pSRF-Luc together with pRL null and pCEFL-MEKK, as indicated, and treated withY-27632 followed by 10 �M LPA. Lysates were assayed for dual luciferase activities. The data represent firefly luciferase activity normalizedby Renilla luciferase activity present in each sample, expressed as percentage of the activity elicited by LPA or MEKK in cells treated withvehicle control (-).(C) Cells were serum starved, treated with vehicle (control) or Y-27632 and LPA, and analyzed by immunofluorescence for c-Jun (anti c-Jun)and nuclear staining with DAPI.

Thus, ROCK may promote gene expression through two et al., 1995; Minden et al., 1995), it was noticed in subse-quent reports that RhoA can stimulate JNK potently indivergent pathways, one dependent on cytoskeletal re-

arrangements impinging on the SRE, and another LimK/ certain cell types (Alberts and Treisman, 1998; Nagaoet al., 1999; Teramoto et al., 1996), albeit by a yet tocofilin independent that regulates the activity of the c-jun

promoter and c-Jun expression. be elucidated molecular mechanism. Based on thesefindings, we explored whether ROCK mediates the acti-vation of JNK by RhoA. Indeed, Y-27632 diminishedROCK Mediates RhoA-Induced JNK Activity

As the activation of the c-jun promoter by ROCK ap- the activation of JNK provoked by RhoA, which wasparticularly remarkable at the low concentrations of thispeared to be independent of changes in the actin-based

machinery, we asked whether activated ROCK was able GTPase (Figure 5B). In contrast, this inhibitor did notaffect the activation of JNK by Cdc42. Furthermore, JNKto stimulate any of the MAPKs that are known to regulate

the activity of this promoter (Marinissen et al., 1999). was phosphorylated in cells transfected with ROCK��4or RhoA QL (Figure 5C), thus confirming their ability toWe transfected NIH 3T3 and HEK293T cells with HA-

tagged forms of p38�, ERK6 (p38�), ERK5, and JNK activate JNK in vivo. As SEK1 (MKK4) and MKK7 areupstream activators for JNK (Davis, 2000), we also askedalong with ROCK��4 and performed kinase assays us-

ing specific substrates. Neither p38�, ERK5, nor ERK6 whether the RhoA/ROCK pathway promotes the phos-phorylation of these JNKKs within their activation loop.was affected by ROCK (data not shown). Surprisingly,

however, JNK was potently activated by ROCK��4 in SEK1 was phosphorylated upon contransfection withRhoA QL or ROCK��4. In contrast, MKK7 was not phos-both cell types, and to an extent similar to MEKK, which

was used as a positive control (Figure 5A). phorylated under identical conditions, although phos-pho-MKK7 was readily detectable in the presence ofAlthough we and others have shown that Rac and

Cdc42 are stronger activators of JNK than RhoA (Coso Cot, a MAPKKK that served as a positive control. Simi-

Molecular Cell32

Figure 2. ROCK Mediates the Expression of c-Jun by RhoA and LPA

(A) NIH 3T3 cells were transfected with pRL null, pCEFL-AU5 RhoA QL, PDZ RhoGEF, and pSRF-Luc or pJun-Luc, as indicated, and treatedwith vehicle control (-) or Y-27632. Lysates assayed for dual luciferase activities. The data are expressed as percentage of the activity elicitedby RhoA QL or PDZ RhoGEF in control cells.(B) Cells were transfected with pCAG-myc-ROCK KD.IA, serum starved, treated with LPA, and analyzed by immunofluorescence for the myctag and c-Jun staining using anti-Myc (��myc) and c-Jun (�-c-Jun) specific antibodies. Stress fibers were labeled with Texas Red phalloidin(Phalloidin) and nuclei with DAPI. Transfected cells are indicated with an arrowhead. Photographs shown are representative of at least 5–10different fields.

larly, phosphorylated MEK1, a MAPKK for ERK1/2, was secondary effects leading to JNK activation throughmechanisms that are independent of ROCK, we askedobserved when cotransfected with an activated form of

c-Raf but not of ROCK (data not shown). These data whether a more physiologically relevant condition lead-ing to the activation of endogenous RhoA, such as LPAindicated that expression of activated forms of RhoA

and ROCK can elevate the enzymatic activity of JNK, treatment, was able to activate endogenous JNK, andif so whether this effect was mediated by ROCK. Indeed,likely through SEK, thus providing a biochemical route

by which RhoA can affect the expression of c-jun. stimulation of NIH 3T3 cells with LPA induced the rapidactivation of JNK, and this effect was nearly abolishedby Y-27632 (Figure 5D). However, this inhibitor did notThe Activation of JNK by LPA Is Mediated by ROCKaffect the parallel activation of endogenous ERK2. InAs the prolonged expression of high levels of active

forms of RhoA could cause cellular stress and other turn, the contribution of JNK to the activation of the


Figure 3. Activated ROCK Stimulates the c-jun Promoter

(A) Structure of ROCK and its deletion mutant. KD, kinase, PH, pleckstrin homology, and Cys, cysteine rich domains, respectively.(B) GFP and HA-tagged forms of ROCK�FL and ROCK��4 were detected by Western blot analysis of transfected HEK293T cells using a mixof anti-GFP and -HA antibodies. The position of the molecular weight markers is indicated.(C) NIH 3T3 cells were transfected with ROCK��4 and ROCK�FL along with pJun-Luc or pSRF-Luc together with pRL null, as indicated, andassayed for dual luciferase activities. The data are expressed as fold induction relative to control.

c-jun promoter by LPA, RhoA, and ROCK was examined (Marinissen et al., 1999) and ROCK (data not shown).Although the jAP1 site was previously described as anby the use of a JNK inhibitor, SP600125, which effec-

tively prevented the activation of the c-jun promoter AP1-like response element, its sequence, TGACATCA,is more similar to a consensus ATF site, TGACGTCA,by each of these stimuli (not shown). Furthermore, a

dominant-negative form of SEK also diminished the acti- than to a consensus AP1 site, TGACTCA (Han et al.,1992). Indeed, the jAP1 site binds to Jun/ATF2 dimers,vation of this promoter, but a selective inhibitor of p38�

and p38�, SB203580, had no demonstrable effect (data and in some cells ATF1/CREB, but not Jun/Fos proteins(Clarke et al., 1998; Herr et al., 1994). To explore whichnot shown). Together, these findings strongly suggest

that RhoA can stimulate the activity of JNK through transcription factors bind the c-jun promoter in NIH 3T3cells we performed gel mobility analysis. Two specificROCK, thereby promoting the transcriptional activation

of the c-jun promoter. retarded complexes were observed when nuclear ex-tracts from serum-deprived cells were incubated with ajAP1 probe (arrows on the left, Figure 6A, upper panel).c-Jun, JunD, and ATF2 Bind the jAP1 Site withinThese two bands were competed by an excess of unla-the c-jun Promoter and Are the Targetbeled oligonucleotide probe (data not shown) and su-for ROCK-Initiated Pathwayspershifted when the extracts were incubated with anti-The c-jun promoter contains a number of response ele-bodies specific for ATF2, c-Jun, and JunD, whereas thements, including sequences binding the transcriptionpresence of JunB, ATF1/CREB, or Fos protein was notfactors SP1, CTF, and MEF2, along with two GATAAevident. As a control, all Jun and Fos but not ATF2and two jAP1 elements (Han et al., 1992). The jAP1 site,antibodies supershifted the complexes formed when us-located at nt –71 to –64, and the MEF2 site are criticaling a canonical AP1 probe (Figure 6A, lower panel). Fur-elements regulating the expression from the c-jun pro-thermore, both ATF2 and ATF1/CREB antibodies super-moter by cell surface receptors and RhoA (Marinissenshifted either an upper or lower band, respectively, fromet al., 2001). In particular, the JNK pathway impingestwo gel-shifted bands observed when the same extractson this jAP1 site, as point mutations that inactivate this

site reduce the activation of the c-jun promoter by MEKK were incubated with a CRE probe. This supported that

Molecular Cell34

Figure 4. The LimK-Cofilin Pathway Participates in Stress-Fiber Formation and the Transcriptional Activation of the SRF but Not in c-jun Ex-pression

(A and C) NIH 3T3 cells were transfected with pEF-myc-LimK1 (k1.C), pCEFL-ROCK��4, or pCEFL-AU5-Cofilin S3A along with pJun-Luc orpSRF-Luc together with pRL null, as indicated. Cells were assayed for dual luciferase activities. The data are expressed as fold inductionrelative to control.(B and D) Cells were transfected with expression vectors for myc-Limk DN or AU5-Cofilin S3A, as indicated. Serum starved cells were treatedfor 6 hr with LPA and analyzed by immunofluorescence for Myc- or AU5-tagged proteins and c-Jun using anti-Myc (�-myc), anti-AU5 (�-AU5),or c-Jun (�-c-Jun) specific antibodies. Stress fibers and nuclei were labeled with Texas Red phalloidin (Phalloidin) and DAPI. Transfectedcells are indicated with arrowheads. Photographs shown are representative of at least 5–10 different fields.

the inability of the anti-ATF1/CREB antiserum to su- JunD, ATF2, and c-Jun are the main transcription factorsthat can bind the jAP1 site in NIH 3T3 cells.pershift the jAP1 probe reflects its absence in these gel-

retarded complexes. Together, these data indicated that To examine whether c-Jun, JunD, or ATF2 bind to


Figure 5. ROCK Stimulates JNK and Medi-ates the Activation of JNK by RhoA and LPA

(A) HEK293T and NIH 3T3 cells were trans-fected with expression vectors for HA-taggedJNK along with pCEFL-GFP-ROCK��4 andMEKK, and used for immunocomplex kinasereactions. 32P-labeled ATF2 (P-ATF2) is indi-cated in the upper panels.(B) HEK293T and NIH 3T3 cells were trans-fected with expression vectors for HA-taggedJNK along with vector control (C) or differentamounts of pCEFL-AU5-RhoA QL andpCEFL-AU5-Cdc42 QL and treated withY-27632, as indicated, and used for kinasereactions. Blots are representative of threedifferent experiments.(C) HEK293T cells were cotransfected withpCEFL-GFP-ROCK��4, RhoA QL, or AU1-Cot, together with HA-tagged JNK (upperpanel), GST-SEK (middle panel) or -MKK7(lower panel), as indicated, serum starved,and tagged proteins were immunoprecipi-tated with anti-HA antibody or recovered withGlutathion-sepharose. Western blots wereperformed with phospho-JNK (P-JNK) anti-bodies and a phospho-SEK antibody that rec-ognizes a conserved phospho-threonine inboth SEK and MKK7 (P-SEK and P-MKK7,respectively). The level recovered kinaseswas assessed using anti-tag antibodies (HAand GST). Mobility of IgG is indicated.(D) Serum starved NIH 3T3 cells were incu-bated with Y-27632 for 2 hr and treated with10 �M LPA for the indicated times. Endoge-nous JNK and ERK2 kinase activities wereassayed. In each case, the upper gel shows32P-labeled substrates and the lower gelshows the amount of JNK or ERK2 presentin each sample. Autoradiograms are repre-sentative of three independent experiments.

the c-jun promoter in vivo, we carried out chromatin To investigate whether, in turn, Jun and ATF2 proteinsimmunoprecipitation (ChIP) assays with specific anti- participate in the expression of c-Jun induced by LPAbodies anti c-Jun, JunD, or ATF2, and included JunB and RhoA through ROCK, we used a dominant-negativeas control. After reversing the crosslinking, the presence form of c-Jun, c-Jun TAM67, which lacks a transactiva-of c-jun sequences in immunoprecipitates was demon- tion domain but dimerizes with Jun, Fos, and ATF pro-strated by PCR amplification using primers specific for teins, thus diminishing their transcriptional and DNAa region of the c-jun promoter (nt -87 to �419) that binding activity (Brown et al., 1993). As shown in Figureincludes the jAP1 site. Purified total genomic DNA (total 6C, c-Jun Tam67 blocked the nuclear accumulation ofinput) was used as a PCR positive control, and nuclear c-Jun induced by LPA, as judged by the use of anextract incubated without antibody served as an addi- N-terminal anti c-Jun serum. Similarly, c-Jun Tam67 re-tional negative control (Figure 6B). This approach reduced the activation of pJun-Luc by RhoA QL andvealed that c-Jun, JunD, and ATF2 are bound in vivo to ROCK��4 (Figure 6D).the amplified region from the c-jun promoter. As thec-jun promoter also exhibits a distantly related AP1-like

ROCK Induces the Phosphorylation of c-Junelement located at nt –190 to –183, we cannot rule outand ATF2 In Vivothe possibility that these factors also bind this distal siteRhoA QL and ROCK��4 induced the transcriptional ac-adjacent to the amplified region. Samples were alsotivity of the jAP1 site, as supported by their ability toamplified for a region corresponding to the c-myc pro-stimulate a jAP1-driven reporter plasmid (Marinissen etmoter that displays a canonical AP1 site (Iavarone etal., 2001) (Figure 6E), but not a similar reporter containingal., 2003). PCR reactions rendered a DNA band in c-Junan inactive point mutant of the jAP1 site (not shown).and Jun D, but not ATF2, immunoprecipitates, thus sup-To confirm that ROCK��4 can induce the in vivo phos-porting the specificity of the complexes bound to thephorylation of transcription factors that bind this jAP1c-jun promoter in vivo, which are more consistent withsite, we transfected cells with tagged c-Jun, JunD, andthose expected for the jAP1 site. Lack of nonspecificATF2 along with ROCK��4, or MEKK as a positive con-DNA pull down was confirmed by PCR amplificationtrol. As shown in Figure 6F, ROCK��4 induced the phos-of an unrelated fragment from the exon 11 of DESC1phorylation of c-Jun, as judged by the use of antibodies(Iavarone et al., 2003). Only samples containing total

genomic DNA rendered a PCR product. specific for phospho-serine 63 and 73, two target sites

Molecular Cell36

Figure 6. ROCK Induces the Phosphorylation of c-Jun and ATF2 that Are Bound to the c-jun Promoter

(A) Nuclear extracts from serum-starved NIH 3T3 cells were incubated with a 32P-labeled jAP1, canonical AP1, or CRE site-containingoligonucleotide in the presence of antibodies against ATF2, c-Jun, JunD, JunB, ATF1/CREB, C-Fos, FosB, Fra1, and Fra2, as indicated.Autoradiograms show DNA-protein complexes resolved in a nondenaturing acrylamide gel. Left arrows indicate the shifted complexes observedin control samples with no antibodies (-). Bands above these complexes correspond to supershifted complexes. The panel is representativeof three experiments with similar results.(B) Ethidium bromide staining of PCR products corresponding to a fragment of the c-jun promoter (332 bp), c-myc promoter (250 bp), and


for JNK (upper panel). Interestingly, an antibody that is required for cell transformation induced by RhoA (Mar-inissen et al., 2001). These findings and the recentlyrecognizes phospho-serine 100 in JunD revealed thataccumulated knowledge on the regulatory processesthe basal level of phospho-JunD was already elevatedby which extracellular stimuli control c-jun expression,when compared to c-Jun (middle panel), but JunD phos-provided an opportunity to explore the molecular mech-phorylation did not increase in ROCK��4 transfectedanisms by which RhoA regulates the expression ofcells. Under the same conditions, ROCK��4 inducedgrowth-promoting genes. Previously, we have shownthe accumulation of ATF2 phosphorylated in threoninethat PKN acts downstream from RhoA in a pathway that69 and 71 (lower panel), and also induced a moderateleads to ERK6 activation and the stimulation of the c-junincrease in the c-Jun and ATF2 protein levels, whichpromoter (Marinissen et al., 2001). Recent work indi-may be related to their increased stability upon phos-cates that PKN forms a scaffold that directs the activityphorylation.of an MLK-related kinase, known as MLTK or MRK,toward ERK6 (Takahashi et al., 2003). However, whetherLPA-Induced AP1 Binding and Transcriptionalthis pathway is sufficient to account for the stimulationActivity Is Inhibited by Y-27632of c-jun expression by RhoA, or whether other Rho-As LPA regulates c-fos expression through RhoA (Hilleffector molecules coordinate c-jun expression with theet al., 1995) but c-Fos strictly requires Jun proteins tostatus of actin polymerization, is still unknown. In thisbind AP1, we asked if the ability of LPA to promotestudy, we observed that inhibition of ROCK preventsc-Jun expression contributes to an enhanced AP1 DNAthe ability of LPA and RhoA to stimulate the c-jun pro-binding activity in NIH 3T3 cells. By using a consensusmoter and the consequent accumulation of c-jun mRNAAP1 site in gel mobility assays, we observed that starvedand c-Jun. Furthermore, an active mutant of ROCK wascells have very limited AP1 binding activity, and thatsufficient to stimulate the c-jun promoter to an extentLPA stimulation leads to a prolonged accumulation ofsimilar to that caused by active RhoA, together indicat-AP1 binding complexes, which was similar to that pro-ing that ROCK plays a critical role as part of the multiplevoked by serum (Figure 7A). These AP1 complexes con-converging pathways linking RhoA to c-jun expression.tained Jun proteins along with Fos proteins (see Figure

The best-understood transcriptional response to6A, lower panel). Western blot for JunD, whose expres-RhoA is the activation of expression from the c-fos SRE,

sion is not affected by LPA, was used as a loadingwhich is independent of MAPK cascades (Hill et al.,

control (Figure 7A, lower panel).1995) but instead involves the activation of the SRF by

The enhanced AP1 binding activity in response to LPA a process that includes a decrease in the cellular poolwas reflected in the activation of transcription from a of unpolymerized actin (Sotiropoulos et al., 1999). Inreporter plasmid driven by two consensus AP1 sites, this case, the stimulation of gene expression and actinpdAP1-Luc (Figure 7B). ROCK��4 also activated polymerization into stress fibers by RhoA appear to actpdAP1-Luc potently, almost to the same extent as the in a coordinated fashion (Sotiropoulos et al., 1999). Thepositive control MEKK. Furthermore, Y-27632 dimin- observation that both actin stress fiber formation andished both AP1-DNA binding and transcriptional activity the activation of the c-jun promoter and c-Jun expres-in response to LPA in a dose-dependent manner (Fig- sion caused through RhoA require the functional activityures 7C and 7D). These findings further support a key of ROCK suggested that a link may also exist betweenrole for ROCK in the pathway linking cell surface recep- actin rearrangements and the transcriptional activationtors acting on RhoA, such as LPA, to the regulation of of the c-jun promoter. Remarkably, we found that thec-fos and c-jun expression, and ultimately to the stimula- pathway linking RhoA to c-jun regulation diverges down-tion of AP1 activity. stream from ROCK from that controlling the status of

actin polymerization and SRF activity. Instead, RhoADiscussion stimulates a kinase cascade resulting in JNK activation

through ROCK, in parallel to its ability to regulate theRhoA stimulates multiple signaling events, whose rela- actin-based cytoskeleton. JNK then regulates nucleartive contribution to cell growth and transformation is events through the direct phosphorylation of transcrip-still far from being completely understood. In this regard, tion factors bound to the c-jun promoter.recent evidence indicates that RhoA can promote the The precise mechanism by which ROCK activates

JNK is still not fully understood. SEK/MKK4 and MKK7expression of c-jun, and that the activity of Jun proteins

DESC1 (250 bp) obtained using chromatin samples obtained by ChIP using the indicated antibodies. Genomic DNA was used as a positivecontrol (total input, TI) and ChIP samples without antibody was used as a negative control (-). Upper and lower arrows indicate PCR productsand free primers, respectively.(C) Cells were transfected with pCEFL-AU5 c-JunTam67 or pCEFL-GFP, serum starved, treated for 6 hr with LPA, and analyzed by immunofluo-rescence for AU5-tagged proteins and c-Jun using anti-AU5 (�-AU5) or c-Jun (�-c-Jun) specific antibodies, and GFP by direct fluorescence.Nuclei were stained with DAPI. Transfected cells are indicated with arrowheads. Photographs shown are representative of at least 5-10different fields.(D and E) NIH 3T3 cells were cotransfected with the pJun-Luc or the pjAP1-Luc reporters along with pRL null, and expression vectors forRhoA QL, ROCK��4, and c-Jun Tam67, as indicated, and analyzed for dual luciferase activities. The data are expressed as fold inductionrelative to the control.(F) Cells were transfected with pCEFL-GFP-c-Jun, pCEFL-GFP-JunD or pCGN-HA-ATF2 alone or in combination with ROCK��4 or MEKK.Lysates were immunoprecipitated with mouse monoclonal anti-GFP and anti-HA antibodies and used in Western blotting experiments.Phospho-c-Jun, phospho-JunD, and phospho-ATF2 were detected with anti-phospho-specific antibodies.

Molecular Cell38

Figure 7. LPA-Induced AP1 Activity Is Regulated by ROCK

(A) Starved NIH 3T3 cells were treated with 10 �M LPA or 10% serum (S) for the indicated time, and nuclear extracts were incubated with a32P-labeled canonical AP1 probe. Autoradiogram shows the DNA-protein complexes resolved in a nondenaturing acrylamide gel. AP1-DNAcomplexes and free probe in the top and bottom of the autoradiogram are indicated. Western blot analysis of JunD in each nuclear extractserved as a loading control.(B) NIH 3T3 cells were cotransfected with the pdAP1-Luc and the pRL null plasmid DNAs and ROCK��4 or MEKK or MEKK, or treated with10 �M LPA, as indicated, and assayed for dual luciferase activities. The data are expressed as fold induction relative to control.(C) Nuclear extracts from NIH 3T3 cells treated with Y-27632 followed by LPA were incubated with a 32P-labeled AP1 probe. Western blot forJunD served as a loading control.(D) NIH 3T3 cells were transfected as in (B), pretreated with Y-27632, and exposed to 10 �M LPA. Dual luciferase activity was assayed andresults expressed as percentage of the response to LPA in the absence of Y-27632.(E) Proposed model for RhoA signaling to the c-jun promoter through ROCK. Activated RhoA can initiate a signaling pathway inducing the


can both function as upstream activators for JNK (Davis, translational modification of ATF2 and c-Jun that areprebound to the jAP1 site from the c-jun promoter,2000). However, ROCK induced the phosphorylation of

SEK in vivo whereas no MKK7 phosphorylation was de- thereby enhancing c-jun expression.The emerging picture is that the signaling route linkingtected. Moreover, a dominant-negative SEK but not a

dominant-negative MKK7 reduced the ability of acti- RhoA to the nucleus bifurcates at the level of ROCK toinitiate two independent pathways, one controlling thevated ROCK to stimulate the c-jun promoter (not shown).

In line with the emerging view that SEK and MKK7 have actin cytoskeleton that regulates the SRF and a secondleading to the activation of a MAPK cascade promotingdistinct functions and therefore are differentially acti-

vated (Davis, 2000), our results suggest that ROCK im- the phosphorylation and activation of JNK. Togetherwith those pathways initiated by RhoA through PKNpinges preferentially on the SEK/JNK signaling module.

However, it is still unclear whether ROCK phosphory- (Marinissen et al., 2001), this results in the activation oftranscription factors prebound to the c-jun promoterlates SEK directly or through an intermediate kinase. As

purified ROCK does not phosphorylate SEK efficiently and consequently in enhanced c-Jun expression. Thesetwo independent pathways acting downstream from(unpublished data), it is more likely that the activation

of SEK-JNK by ROCK occurs through one or more inter- ROCK ultimately converge in the nucleus to control thelevels and transcriptional activity of AP1 complexes,mediate kinases, likely belonging to the large family of

JNKKKs and JNKKKKs (Davis, 2000), whose precise thus coordinating RhoA-induced cytoskeletal changeswith the transcriptional activation of genetic programsidentification will require further investigation.

The ability of ROCK to stimulate JNK provides a mo- that contribute to the numerous biological activities ofRhoA, including those promoting cell invasion and nor-lecular mechanism by which RhoA and its GEFs can

stimulate JNK, as reported in a variety of cellular sys- mal and aberrant cell growth.tems (Alberts and Treisman, 1998; Nagao et al., 1999;

Experimental ProceduresTeramoto et al., 1996). Given the importance of AP1 innormal cell proliferation and cancer, this pathway may

DNA Constructsalso provide a mechanism by which RhoA can itself The plasmid pJCX-ROCK�, provided by Louis Lim, was amplifiedpromote focus-formation or cooperate with Raf to trans- by PCR and subcloned into GFP or HA-tagged pCEFL expression

vectors to generate an activated mutant of ROCK�, ROCK��4form cells, a process that requires ROCK activity but(amino acids 1-477). The dominant-negative form of ROCK, ROCKnot its cytoskeletal effects (Sahai et al., 1998, 1999).KD.IA, has been described (Itoh et al., 1999). Activated LimK1(k1.C)The Rho-ROCK-JNK pathway may also participate inand myc-tagged dominant-negative LimK (LimK DN) were provideddevelopmental morphogenic processes, as suggestedby Richard Treisman or described previously, respectively (Arber et

by genetic epistasis studies in Drosophila indicating that al., 1998; Sotiropoulos et al., 1999). A dominant-negative Cofilin,JNK mediates the generation of tissue polarity induced Cofilin S3A, was obtained by PCR amplification replacing serine 3

by alanine, and subcloned into an AU5-tagged pCEFL vector. Toby RhoA (Strutt et al., 1997). Moreover, the regulatorygenerate pJun-Luc, the wild-type murine c-jun promoter was ampli-role of ROCK through JNK may not be restricted tofied by PCR from pJC6, a pBLCAT3 based reporter plasmid (Hanc-Jun, as RhoA also promotes the phosphorylation ofet al., 1992), and subcloned into the pGL3 vector (Promega). Thec-Myc through ROCK, a pathway that is overactive inpSRF luciferase reporter (pSRF-Luc) was from Stratagene. pGL3

human breast cancer cells (Watnick et al., 2003). Al- reporter plasmids containing two jAP1 sites from the murine c-junthough the kinase that phosphorylates c-Myc was not promoter, pdjAP1, or two mutated jAP1 sites, pdjAP1mut, have been

described (Marinissen et al., 2001). The expression vectors pCEFL-addressed, the phosphorylated sites on c-Myc, serinesAU5-RhoA QL, pCEFL HA-tagged JNK, ERK2, ERK5, p38�, ERK6,62 and 71, can be the target for JNK in vivo (NoguchipCEFL-MEKK, pCEV29-MEKEE, pCEFL-MEK AA, pCEFL-GST-et al., 1999). Similarly, as actin treadmilling alone ap-MKK6, pCEFL-GST-SEK, pEBG-MKK7, pEBG-MEK, pCGN-HA-pears not to be sufficient to induce c-fos expressionATF2, pCEFL-AU5 c-JunTAM67, and pCEFL-PDZRhoGEF have

through the SRF (Alberts et al., 1998), the finding that been described (Chiariello et al., 2000; Fukuhara et al., 1999; Marinis-ROCK stimulates JNK may provide a parallel pathway sen et al., 1999). The purification of bacterially expressed GST-ATF2

was performed as described (Marinissen et al., 1999). JunD andby which RhoA stimulates the TCF component of thec-Jun cDNAs were amplified by PCR and cloned in a pCEFL-GFP-SRE, thus together promoting c-fos expression.tagged vector.Together, our data support a model by which the

RhoA-ROCK-JNK signaling module activates the c-jun Cell Lines and Transfectionspromoter and enhances c-Jun expression (Figure 7E) NIH 3T3 fibroblasts and HEK293T cells were maintained in Dulbec-through an AP1-like site, jAP1, which binds primarily co’s modified Eagle’s medium (DMEM, Life Technologies, Inc.) sup-

plemented with 10% calf serum or 10% fetal bovine serum, respec-ATF2, JunD, and c-Jun. ROCK activation promotes thetively. Transient transfections were performed using thephosphorylation of c-Jun and ATF2 on their transactiva-Lipofectamine Plus Reagent (Life Technologies, Inc.).tion domains, resulting in their increased transcriptional

activity, and probably also enhancing their protein sta- Northern Blot and Luciferase Reporter Assaysbility. Thus, extracellular stimuli that signal through Total RNA extraction from NIH 3T3 cells and Northern blot analysis

were performed as described (Marinissen et al., 2001), using 32P-RhoA may cause a rapid ROCK-JNK-dependent post-

activity of ROCK, which in turn activates JNK through SEK, which then phosphorylates and activates transcription factors bound to the jAP1site on the c-jun promoter, thereby enhancing c-jun expression in concert with those signals regulating the jMEF2 site through PKN/ERK6.Available data suggest that ROCK mediates the activation of the JNK pathway by RhoA, and is involved in the regulation of gene expressionindependently of its ability to promote changes in the actin cytoskeleton. Other known pathways are depicted and described in the text.Unknown molecules are indicated by question marks.

Molecular Cell40

labeled murine c-jun cDNA as a probe. For reporter assays, cells 1 �g of anti c-Jun (H-79), JunD (329), JunB (N-17), ATF2 (C-19),ATF1/CREB-1 (C-21), c-Fos (A-2), FosB (102)-G, Fra1 (N-17), andwere transfected, per well, with expression plasmids together with

0.1 �g of the indicated reporter plasmid and 0.01 �g of pRL null, Fra2 (L-15) antibodies (Santa Cruz Biotechnology, Inc) was addedto the binding reaction prior to the addition of radiolabeled probe.which expresses Renilla luciferase from Renilla reniformis as an

internal control. The total amount of plasmid DNA was adjusted withpcDNA3-�-Galactosidase. When indicated, cells were pretreated Acknowledgmentsfor 2 hr with Y-27632 (Calbiochem), SP600125 (AG Scientific), orSB203580 (Calbiochem), and treated with LPA (10 �M) for an addi- We thank Akrit Sodhi for insightful discussions and comments dur-tional 6–8 hr. Firefly and Renilla luciferase activities present in cellu- ing the manuscript preparation.lar lysates were assayed using the Dual-Luciferase Reporter System(Promega). Light emission was quantitated using a Monolight 2010 Received: November 6, 2003luminometer (Analytical Luminescence Laboratory). Data were rep- Revised: February 23, 2004resented as firefly luciferase activity normalized by Renilla luciferase Accepted: February 25, 2004activity present in each sample, and the values plotted were the Published: April 8, 2004average � S.E. of triplicate samples from typical experiments, whichwere repeated at least 3–5 times with nearly identical results.

References

Kinase Assays and Western BlotAlberts, A.S., Geneste, O., and Treisman, R. (1998). Activation of

Cells were cotransfected with expression vectors for HA-taggedSRF-regulated chromosomal templates by Rho-family GTPases re-

kinases and different upstream molecules, cultured for 16–20 hr,quires a signal that also induces H4 hyperacetylation. Cell 92,

and then serum starved overnight for ERK2 and ERK5, and 2 hr for475–487.

JNK, p38�, and ERK6. Immunocomplex kinase assays were per-Alberts, A.S., and Treisman, R. (1998). Activation of RhoA and SAPK/formed essentially as described (Marinissen et al., 1999). The endog-JNK signalling pathways by the RhoA-specific exchange factorenous JNK and MAPK activity was also assessed after immunopre-mNET1. EMBO J. 17, 4075–4085.cipitation with anti-JNK monoclonal (PharMingen) and anti-MAPK

polyclonal rabbit antibodies (Santa Cruz Biotechnology, Inc). Arber, S., Barbayannis, F.A., Hanser, H., Schneider, C., Stanyon,Tagged molecules were detected by Western blotting using anti-HA C.A., Bernard, O., and Caroni, P. (1998). Regulation of actin dynam-monoclonal (HA.11, Covance, Inc.), anti-GFP monoclonal (Covance, ics through phosphorylation of cofilin by LIM-kinase. Nature 393,Inc.), or anti-GST rabbit polyclonal antibodies. Anti-phospho-SEK, 805–809.-MEK1, -JNK, -ATF2 and -c-Jun antibodies were from Cell Signaling Bar-Sagi, D., and Hall, A. (2000). Ras and Rho GTPases: a familyTechnlogy. Anti-JunD antibody was from Santa Cruz Biotechnology, reunion. Cell 103, 227–238.Inc. Proteins were visualized by enhanced chemiluminescence de-

Brown, P.H., Alani, R., Preis, L.H., Szabo, E., and Birrer, M.J. (1993).tection (Amersham Corp.) using goat anti-mouse and anti-rabbitSuppression of oncogene-induced transformation by a deletion mu-IgGs coupled to horseradish peroxidase as the secondary anti-tant of c-jun. Oncogene 8, 877–886.body (Cappel).Chiariello, M., Marinissen, M.J., and Gutkind, J.S. (2000). Multiplemitogen-activated protein kinase signaling pathways connect theIndirect Immunofluorescencecot oncoprotein to the c-jun promoter and to cellular transformation.NIH 3T3 cells were seeded on glass coverslips and transfected asMol. Cell. Biol. 20, 1747–1758.above, serum starved for 16–20 hr, washed, fixed, and permeabilizedClarke, N., Arenzana, N., Hai, T., Minden, A., and Prywes, R. (1998).with 4% formaldehyde and 0.5% Triton X-100 in 1 PBS. Cells wereEpidermal growth factor induction of the c-jun promoter by a Racwashed and blocked with 1% bovine serum albumin, and incubatedpathway. Mol. Cell. Biol. 18, 1065–1073.with anti c-Jun (H-79) (Santa Cruz Biotechnology, Inc), anti AU5, or

anti c-Myc (9E10) (Covance, Inc.) for 1 hr, washed, and incubated Coso, O.A., Chiariello, M., Yu, J.C., Teramoto, H., Crespo, P., Xu,with Phalloidin (Molecular Probes) (1:200) to stain F-actin, or the N., Miki, T., and Gutkind, J.S. (1995). The small GTP-binding proteinscorresponding secondary antibodies (1:200) conjugated with tetra- Rac1 and Cdc42 regulate the activity of the JNK/SAPK signalingmethylrhodamine B isothiocyanate, fluorescein isothiocyanate, or pathway. Cell 81, 1137–1146.Texas Red (Jackson ImmunoResearch Laboratories, Inc.). Cov- Davis, R.J. (2000). Signal transduction by the JNK group of MAPerslips were washed and mounted in Vectashield mounting medium kinases. Cell 103, 239–252.with DAPI (Vector Laboratories, Inc) and viewed using a Zeiss Axio-

Fukuhara, S., Murga, C., Zohar, M., Igishi, T., and Gutkind, J.S.phot photomicroscope equipped with epifluorescence.(1999). A novel PDZ domain containing guanine nucleotide ex-change factor links heterotrimeric G proteins to Rho. J. Biol. Chem.Chromatin Immunoprecipitation Assay274, 5868–5879.ChIP assay was performed using the ChIP Assay Kit (Upstate Bio-Geneste, O., Copeland, J.W., and Treisman, R. (2002). LIM kinasetechnology). In brief, chromatin from NIH 3T3 cells was fixed withand Diaphanous cooperate to regulate serum response factor andformaldehyde (1% final), and nuclear extracts were isolated andactin dynamics. J. Cell Biol. 157, 831–838.sonicated. Transcription factors bound to chromatin were immuno-

precipitated with specific antibodies anti c-Jun (H-79), JunD (329), Han, T.H., Lamph, W.W., and Prywes, R. (1992). Mapping of epider-JunB (N-17), and ATF2 (C-19) (Santa Cruz Biotechnology, Inc), pro- mal growth factor-, serum-, and phorbol ester-responsive sequencetein-DNA crosslinking was reversed and the isolated genomic DNA elements in the c-jun promoter. Mol. Cell. Biol. 12, 4472–4477.was amplified by PCR, using specific primers encompassing the Herr, I., van Dam, H., and Angel, P. (1994). Binding of promoter-region �87 to �419 from the murine c-jun promoter (GenBank Ac- associated AP-1 is not altered during induction and subsequentcession U03709), a fragment from the c-myc promoter, or a 250 bp repression of the c-jun promoter by TPA and UV irradiation. Carcino-fragment from the exon 11 of DESC1 (Iavarone et al., 2003). PCR genesis 15, 1105–1113.products were separated on a 2% agarose gel and visualized by

Hill, C.S., Wynne, J., and Treisman, R. (1995). The Rho familyethidium bromide staining.GTPases RhoA, Rac1, and CDC42Hs regulate transcriptional activa-tion by SRF. Cell 81, 1159–1170.

Electrophoretic Mobility Shift AssaysIavarone, C., Catania, A., Marinissen, M.J., Visconti, R., Acunzo, M.,Nuclear extracts from NIH 3T3 cells were prepared as describedTarantino, C., Carlomagno, M.S., Bruni, C.B., Gutkind, J.S., and(Iavarone et al., 2003). Oligonucleotides containing a canonical AP1,Chiariello, M. (2003). The platelet-derived growth factor controlsthe jAP1 site, or CRE were labeled with [�32P]-ATP using T4 poly-c-myc expression through a JNK- and AP-1-dependent signalingnucleotide kinase (Invitrogen), purified, and added to the reactionspathway. J. Biol. Chem. 278, 50024–50030.(20,000 cpm/ reaction) for 15 min. Complexes were analyzed on

nondenaturing (4.5%) polyacrylamide gels. For super-shift assays, Itoh, K., Yoshioka, K., Akedo, H., Uehata, M., Ishizaki, T., and Naru-


miya, S. (1999). An essential part for Rho-associated kinase in the Zohn, I.M., Campbell, S.L., Khosravi-Far, R., Rossman, K.L., andDer, C.J. (1998). Rho family proteins and Ras transformation: thetranscellular invasion of tumor cells. Nat. Med. 5, 221–225.RHOad less traveled gets congested. Oncogene 17, 1415–1438.Karin, M., Liu, Z., and Zandi, E. (1997). AP-1 function and regulation.

Curr. Opin. Cell Biol. 9, 240–246.

Marinissen, M.J., Chiariello, M., and Gutkind, J.S. (2001). Regulationof gene expression by the small GTPase Rho through the ERK6 (p38gamma) MAP kinase pathway. Genes Dev. 15, 535–553.

Marinissen, M.J., Chiariello, M., Pallante, M., and Gutkind, J.S.(1999). A network of mitogen-activated protein kinases links G pro-tein-coupled receptors to the c-jun promoter: a role for c-Jun NH2-terminal kinase, p38s, and extracellular signal-regulated kinase 5.Mol. Cell. Biol. 19, 4289–4301.

Minden, A., Lin, A., Claret, F.X., Abo, A., and Karin, M. (1995). Selec-tive activation of the JNK signaling cascade and c-Jun transcrip-tional activity by the small GTPases Rac and Cdc42Hs. Cell 81,1147–1157.

Miralles, F., Posern, G., Zaromytidou, A.I., and Treisman, R. (2003).Actin dynamics control SRF activity by regulation of its coactivatorMAL. Cell 113, 329–342.

Nagao, M., Kaziro, Y., and Itoh, H. (1999). The Src family tyrosinekinase is involved in Rho-dependent activation of c-Jun N-terminalkinase by Galpha12. Oncogene 18, 4425–4434.

Narumiya, S. (1996). The small GTPase Rho: cellular functions andsignal transduction. J. Biochem. (Tokyo) 120, 215–228.

Narumiya, S., Ishizaki, T., and Uehata, M. (2000). Use and propertiesof ROCK-specific inhibitor Y-27632. Methods Enzymol. 325,273–284.

Noguchi, K., Kitanaka, C., Yamana, H., Kokubu, A., Mochizuki, T.,and Kuchino, Y. (1999). Regulation of c-Myc through phosphoryla-tion at Ser-62 and Ser-71 by c-Jun N-terminal kinase. J. Biol. Chem.274, 32580–32587.

Ridley, A.J., and Hall, A. (1992). The small GTP-binding protein rhoregulates the assembly of focal adhesions and actin stress fibersin response to growth factors. Cell 70, 389–399.

Sahai, E., Alberts, A.S., and Treisman, R. (1998). RhoA effector mu-tants reveal distinct effector pathways for cytoskeletal reorganiza-tion, SRF activation and transformation. EMBO J. 17, 1350–1361.

Sahai, E., Ishizaki, T., Narumiya, S., and Treisman, R. (1999). Trans-formation mediated by RhoA requires activity of ROCK kinases.Curr. Biol. 9, 136–145.

Sotiropoulos, A., Gineitis, D., Copeland, J., and Treisman, R. (1999).Signal-regulated activation of serum response factor is mediatedby changes in actin dynamics. Cell 98, 159–169.

Strutt, D.I., Weber, U., and Mlodzik, M. (1997). The role of RhoA intissue polarity and Frizzled signalling. Nature 387, 292–295.

Takahashi, M., Gotoh, Y., Isagawa, T., Nishimura, T., Goyama, E.,Kim, H.S., Mukai, H., and Ono, Y. (2003). Regulation of a Mitogen-Activated Protein Kinase Kinase Kinase, MLTK by PKN. J. Biochem.(Tokyo) 133, 181–187.

Takai, Y., Sasaki, T., and Matozaki, T. (2001). Small GTP-bindingproteins. Physiol. Rev. 81, 153–208.

Teramoto, H., Crespo, P., Coso, O.A., Igishi, T., Xu, N., and Gutkind,J.S. (1996). The small GTP-binding protein rho activates c-JunN-terminal kinases/stress-activated protein kinases in human kid-ney 293T cells. Evidence for a Pak-independent signaling pathway.J. Biol. Chem. 271, 25731–25734.

Treisman, R. (1995). Journey to the surface of the cell: Fos regulationand the SRE. EMBO J. 14, 4905–4913.

Watnick, R.S., Cheng, Y.N., Rangarajan, A., Ince, T.A., and Weinberg,R.A. (2003). Ras modulates Myc activity to repress thrombospondin-1expression and increase tumor angiogenesis. Cancer Cell 3,219–231.

Whitmarsh, A.J., Shore, P., Sharrocks, A.D., and Davis, R.J. (1995).Integration of MAP kinase signal transduction pathways at the serumresponse element. Science 269, 403–407.

Whitmarsh, A.J., Yang, S.H., Su, M.S., Sharrocks, A.D., and Davis,R.J. (1997). Role of p38 and JNK mitogen-activated protein kinasesin the activation of ternary complex factors. Mol. Cell. Biol. 17, 2360–2371.

the small gtp-binding protein rhoa regulates c-jun by a rock-jnk signaling axis

Documents