Binding of GEF-H1 to the Tight Junction-Associated Adaptor Cingulin Results in Inhibition of Rho Signaling and G1/S Phase Transition

Article

Reference

Binding of GEF-H1 to the Tight Junction-Associated Adaptor Cingulin Results in Inhibition of Rho Signaling and G1/S Phase Transition

AIJAZ, Saima, et al.

Abstract

The activity of Rho GTPases is carefully timed to control epithelial proliferation and differentiation. RhoA is downregulated when epithelial cells reach confluence, resulting in inhibition of signaling pathways that stimulate proliferation. Here we show that GEF-H1/Lfc, a guanine nucleotide exchange factor for RhoA, directly interacts with cingulin, a junctional adaptor. Cingulin binding inhibits RhoA activation and signaling, suggesting that the increase in cingulin expression in confluent cells causes downregulation of RhoA by inhibiting GEF-H1/Lfc. In agreement, RNA interference of GEF-H1 or transfection of GEF-H1 binding cingulin mutants inhibit G1/S phase transition of MDCK cells, and depletion of cingulin by regulated RNA interference results in irregular monolayers and RhoA activation. These results indicate that forming epithelial tight junctions contribute to the downregulation of RhoA in epithelia by inactivating GEF-H1 in a cingulin-dependent manner, providing a molecular mechanism whereby tight junction formation is linked to inhibition of RhoA signaling.

AIJAZ, Saima, et al . Binding of GEF-H1 to the Tight Junction-Associated Adaptor Cingulin Results in Inhibition of Rho Signaling and G1/S Phase Transition. Developmental Cell , 2005, vol. 8, no. 5, p. 777-786

DOI : 10.1016/j.devcel.2005.03.003

Available at:

http://archive-ouverte.unige.ch/unige:82113

Disclaimer: layout of this document may differ from the published version.

1 / 1

Developmental Cell, Vol. 8, 777–786, May, 2005, Copyright ©2005 by Elsevier Inc. DOI 10.1016/j.devcel.2005.03.003

Binding of GEF-H1 to the Tight Junction-Associated Adaptor Cingulin Results in Inhibition

of Rho Signaling and G1/S Phase Transition

Saima Aijaz,

Fabio D’Atri,

Sandra Citi,

components and the cytoskeleton (Anderson et al., 2004; Gonzalez-Mariscal et al., 2003).

Maria S. Balda,

* and Karl Matter

*

1

Division of Cell Biology Rho GTPases are molecular switches that are impor- tant components of many subcellular signaling pro- Institute of Ophthalmology

University College London cesses that govern cell proliferation and differentiation (Etienne-Manneville and Hall, 2002; Ridley, 2004). In London

United Kingdom their GTP-bound state, they can bind effector mole-

cules that activate downstream components; in their

2

Department of Molecular Biology

University of Geneva GDP-bound state, they are inactive. Activation is cata- lyzed by guanine nucleotide exchange factors (GEFs) Geneva

Switzerland that stimulate the exchange of GDP by GTP and inacti-

vation by GTPase-activating proteins that promote GTP

3

Dipartimento di Biologia

Università di Padova hydrolysis. The spatial and temporal control of signal- ing by Rho GTPases is thought to be determined by Padova

Italy regulating the localization and activation of these regu-

lators at specific subcellular sites, but our knowledge about these processes is still very limited (Etienne-

Summary Manneville and Hall, 2002).

In epithelial cells, confluence is paralleled by a reduc- The activity of Rho GTPases is carefully timed to con- tion of active RhoA levels and activation of Rac1 and trol epithelial proliferation and differentiation. RhoA Cdc42 (Braga, 2002; Fukata and Kaibuchi, 2001; Noren is downregulated when epithelial cells reach conflu- et al., 2001). Activation of Rac1 and Cdc42 is triggered ence, resulting in inhibition of signaling pathways by E-cadherin engagement and promotes formation of that stimulate proliferation. Here we show that GEF- the junctional complex. Downregulation of RhoA at cell H1/Lfc, a guanine nucleotide exchange factor for confluence is observed in different cell types, resulting RhoA, directly interacts with cingulin, a junctional in inhibition of G1/S phase progression (Coleman et al., adaptor. Cingulin binding inhibits RhoA activation 2004). One way of inhibiting RhoA signaling is activa- and signaling, suggesting that the increase in cingulin tion of p190RhoGAP, which has been observed upon expression in confluent cells causes downregulation cadherin engagement by immobilized ligands in trans- of RhoA by inhibiting GEF-H1/Lfc. In agreement, RNA fected CHO cells, which do not form a junctional com- interference of GEF-H1 or transfection of GEF-H1 plex (Noren et al., 2003). The relevance of GEFs for the binding cingulin mutants inhibit G1/S phase transi- confluence-dependent regulation of RhoA and the con- tion of MDCK cells, and depletion of cingulin by regu- tribution of mechanisms not associated with cadherins lated RNA interference results in irregular mono- is unknown.

layers and RhoA activation. These results indicate Here we focus on GEF-H1/Lfc, a GEF for Rho that that forming epithelial tight junctions contribute to associates with tight junctions in epithelial cells and the downregulation of RhoA in epithelia by inactiva- regulates paracellular permeability (Benais-Pont et al., ting GEF-H1 in a cingulin-dependent manner, provid- 2003). GEF-H1, originally cloned in mice and called Lfc, ing a molecular mechanism whereby tight junction is an oncoprotein of the Dbl family that activates RhoA formation is linked to inhibition of RhoA signaling. but not Rac1 or Cdc42 (Glaven et al., 1996; Krendel et al., 2002; Ren et al., 1998). GEF-H1 can associate with

Introduction different cytoskeletal structures, microtubules as well

as the actin cytoskeleton, and has been proposed to The epithelial junctional complex mediates adhesion mediate crosstalk between the two types of filaments and regulates cell proliferation and differentiation (Balda (Benais-Pont et al., 2003; Glaven et al., 1999; Krendel and Matter, 2003). Tight junctions are the most apical et al., 2002; Ren et al., 1998). The molecular basis for component of the junctional complex and separate the the association of GEF-H1 with different cytoskeletal apical from the basolateral membrane (Cereijido et al., structures is unknown. We now show that GEF-H1 2000; Schneeberger and Lynch, 2004; Tsukita et al., binds to the F-actin binding junctional adaptor cingulin.

2001). They regulate paracellular permeability and re- Cingulin binding inhibits GEF-H1 and, hence, results in strict apical/basolateral intramembrane diffusion of lip- downregulation of RhoA and inhibition of G1/S phase ids. Multiple signaling components have been localized transition, providing a molecular mechanism whereby to epithelial tight junctions, some of which function in tight junction formation is linked to RhoA inactivation.

the regulation of epithelial polarization, differentiation,

and growth control (Matter and Balda, 2003). These sig- Results naling components interact with different types of

adaptor proteins that also bind junctional membrane GEF-H1 Interacts with the F-Actin Binding Protein Cingulin

GEF-H1 can associate with two different types of actin-

*Correspondence: m.balda@ucl.ac.uk (M.S.B.); k.matter@ucl.ac.uk

(K.M.)

based structures in different cell types: tight junctions

in epithelial cells and stress fibers in fibroblasts (Be- nais-Pont et al., 2003). We therefore tested whether tight junction-associated actin binding proteins copre- cipitate with endogenous GEF-H1. Figure 1A shows that GEF-H1 was precipitated by monoclonal antibody (mAb) B4/7. We then blotted the same samples with antibodies specific for cingulin, a tight junction-associ- ated F-actin binding protein (Citi et al., 1988; D’Atri and Citi, 2001). Anti-cingulin antibodies detected a band of approximately 150 kDa in B4/7 immunoprecipi- tates, suggesting that cingulin exists in a complex with GEF-H1.

We next determined the domain of GEF-H1 required for complex formation with cingulin using GEF-H1/GST fusion proteins (Figure 1B). Glutathione beads loaded with equal amounts of GEF-H1/GST fusion proteins were incubated with MDCK cell extracts. Specific pre- cipitation of full-length cingulin was only observed with constructs containing the PH domain (Figure 1C). The anti-cingulin antibody also recognized a band of ap- proximately 70 kDa that was only present in pull-downs with fusion proteins containing the PH domain. Since this antibody had been generated against a recombi- nant protein containing the rod and tail domains (Figure 1D), this suggests that the PH domain of GEF-H1 binds to either one of these two cingulin domains. This was confirmed with experiments with recombinant GST fu- sion proteins containing different regions of cingulin that were tested for pull-down of recombinant His

- tagged PH domain. The PH domain of GEF-H1 was effi- ciently precipitated by a cingulin fusion protein contain- ing residues 782 to 1025, suggesting that the GEF-H1/

cingulin interaction is due to direct binding of GEF-H1’s PH domain to the cingulin rod domain (Figures 1E and 1F). Because GEF-H1 and cingulin colocalize at inter- cellular junctions (Figure 1G), GEF-H1/cingulin com-

plexes are likely to be primarily associated with tight

(A) Association of endogenous GEF-H1 and cingulin. MDCK cell

junctions.

We next tested whether cingulin can influence the

distribution of GEF-H1. When MDCK cells were trans-

fected with myc-tagged wild-type cingulin, the trans-

body. The positions of GEF-H1, cingulin, and the antibody heavy

fected protein was found to colocalize with GEF-H1 at

chain (γ) are indicated.

cell-cell junctions (Figure 2A). In cells that stained more

(B) Domain map of GEF-H1. The C1, the Dbl homology (DH), the

brightly, myc-cingulin was not only closely associated

pleckstrin homology (PH), and the C-terminal (CTD) domain are in-

with cell-cell junctions but could also be seen in the

cytosol. This behavior was enhanced by removal of the

head domain (myc-cingulinR+T), a mutation known to

inhibit junctional recruitment of cingulin and to induce

downs were analyzed by immunoblotting for cingulin using an anti-

cytosolic aggregates (D’Atri et al., 2002). Endogenous

body generated against the C-terminal half of cingulin. The asterisk

GEF-H1 colocalized with both cingulin constructs. Ex-

marks the position of a cingulin degradation product.

pression of a construct containing cingulin residues

(D) Domain map of cingulin. Indicated are the head, the central rod,

782-1025, which include the GEF-H1 binding site,

yielded a diffuse GEF-H1 distribution. Transfection of

the cingulin head domain did not affect the distribution

of GEF-H1. Staining for GEF-H1 generally appeared to

His₆-PH domain of GEF-H1 were analyzed by immunoblotting with

be brighter in cells expressing a GEF-H1 binding cin-

an anti-His₆or anti-GST antibodies. (F) Increasing concentrations

gulin fragment. This was not due to a crossreaction of

of His6-PH domain were added to test saturation of binding.

the secondary antibodies since it was not observed

when the anti-GEF-H1 antibody was omitted (not

shown). Immunoblotting did not reveal upregulation of

GEF-H1 expression. Although this might be due to a

low transfection efficiency, it is likely that cytosolic

junctions and an xz-section to show the distribution along the

GEF-H1 was more efficiently labeled. ZO-1, another

z-axis. The colored panels represent the corresponding overlays.

tight junction protein known to associate with cingulin,

Tight Junctions and Regulation of Rho 779

cingulin can thus interact in vivo and this interaction is sufficiently strong to affect the distribution of GEF-H1.

Because the PH domain of GEF-H1 interacts with cingulin, it might be sufficient for junctional recruitment.

Although a VSV-tagged PH domain was partially tar- geted to intercellular junctions, the transfected protein also localized to cytoplasmic aggregates (Figure 2D).

Mutation of tryptophane-563 (PHW563A-VSV), a resi- due that is conserved in PH domains, or deletion of the PH domain in full-length GEF-H1 resulted in proteins that accumulated in the cytosol (Figure 2E), suggesting that the PH domain is important for junctional re- cruitment.

The N-Terminal Domains of GEF-H1 Regulate Junctional Recruitment

Since overexpressed GEF-H1 is not targeted efficiently to junctions, we expressed N- and C-terminally trun- cated GEF-H1-VSV to determine whether a specific do- main inhibits junctional recruitment. Removal of the C-terminal domain (GEF-

CTD-VSV) did not affect the distribution of the protein: This mutant protein was still associated with filaments. Cells expressing GEF-

CTD- VSV appeared more spread out and flatter; this appear- ance was paralleled by increased RhoA activation (not shown) as in other cell lines (Glaven et al., 1996; Kren- del et al., 2002). In contrast, removal of the N-terminal domain (GEF-

NTD-VSV), which contains the C1 do- main, eliminated the filamentous staining, and a signifi- cant fraction of the protein colocalized with cingulin at

junctions. Additionally, a nonfilamentous cytoplasmic

pool was observed. These data indicate that the N ter-

minus regulates junctional recruitment of GEF-H1,

which is in agreement with the observation that muta-

tion of the Zinc-fingers in the C1 domain impedes

microtubule binding (Krendel et al., 2002).

ies against GEF-H1 (A), ZO-1 (B), or the VSV-epitope to stain

transfected GEF-H1 (C). In (C), myc-cingulinR+T transfection, a

We next tested whether the N terminus is sufficient

for microtubule binding. A construct containing the C1

and the intervening domain (C1/ID-VSV) was recruited

to microtubules and induced microtubule bundling; the

latter was not observed with the full-length construct

(Figure 3B). Constructs containing either one of these

an internal deletion of the PH domain were transiently transfected

domains alone remained cytosolic and often aggre-

and localized by immunofluorescence.

gated (not shown).

Since the N-terminal domain is sufficient and re- quired for microtubule binding but a construct lacking the PH domain does not bind microtubules (Figure 2E), did not coaggregate with truncated cingulin, indicating

it is possible that the two domains interact with each that the observed redistribution of GEF-H1 was not due

other and thereby modulate targeting to microtubules to a general effect on junctional proteins (Figure 2B).

and junctions. A recombinant His

-tagged PH domain To further test antibody specificity, we transfected

construct was indeed pulled down by a GST fusion pro- full-length and truncated myc-cingulin into a cell line

tein containing the C1 domain but not by GST alone overexpressing VSV-tagged GEF-H1. A considerable

(Figure 3C). Binding of the PH domain to a GST fusion fraction of overexpressed GEF-H1 was associated with

protein containing the C-terminal domain was not ob- microtubules as observed previously (Figure 2C) (Be-

served (not shown). Thus, the C1 and PH domains of nais-Pont et al., 2003; Krendel et al., 2002). In cells

GEF-H1 can interact, providing a possible explanation expressing myc-cingulin, GEF-H1-VSV was recruited

for the observed effects on each other’s targeting activ- more efficiently to intercellular junctions (middle pan-

ities.

els). The colocalization of GEF-H1-VSV and myc-cin- gulin was not due to antibody crossreaction since it

was only observed in cells expressing GEF-H1-VSV Cingulin Inhibits RhoA Activation by GEF-H1 RhoA is downregulated when cells reach confluence (Figure 2C, myc-cinR+T: GEF-H1-VSV expressing (ar-

row) and nonexpressing (asterisk) cells are shown that and cease to proliferate (Coleman et al., 2004). Since

cingulin expression increases with cell density whereas

both contain large cingulin aggregates). GEF-H1 and

loss of FRET and, hence, reduced YFP emission. Coex- pression of full-length cingulin with the FRET probe re- sulted in increased YFP emission, indicating reduced Rho-GTP levels (Figure 4D). Only mutants containing the GEF-H1 binding site (myc-cingulin R+T, myc-cin- gulin 782-1025) increased YFP emission. Although Rho inhibition was significant, the increase in YFP emission was more pronounced when Rho was inhibited by add- ing TAT-C3, a membrane permeable version of C3 transferase (Coleman et al., 2001). In contrast, transfec- tion of GEF-H1 reduced YFP emission indicative of Rho activation. Cotransfection of full-length, myc-cingulin R+T and myc-cingulin 782-1025 again counteracted Rho activation by GEF-H1, supporting the conclusion that cingulin inhibits GEF-H1.

To test whether inhibition of Rho activation depends on GEF-H1, we made use of a cell line permitting the tetracycline-regulated depletion of the exchange factor by RNA interference (Figure 4E). In GEF-H1-depleted

cells, transfection of cingulin no longer affected the

Targeting

YFP/CFP ratio, suggesting that inhibition required the

(A) MDCK cell lines expressing GEF-H1-VSV or truncated proteins

presence of GEF-H1 (Figure 4F).

lacking either the C- (GEF-⌬CTD-VSV) or the N-terminal domain

Since the above-used FRET probe can respond to

other Rho GTPases than RhoA, we used a similar probe

consisting of YFP separated from CFP by RhoA fused

to a Rho binding domain, resulting in FRET when the

either GEF-H1-VSV or a construct containing the N-terminal do-

GTPase is in the GTP-bound form (Yoshizaki et al.,

2003). Analogous probes specific for Rac1 and Cdc42

were used for comparison. RhoA inactivation was ob-

served when cingulin constructs containing the GEF-

H1 binding site were cotransfected (Figure 4G). No ef-

fects on the Rac1 probe were observed, and the Cdc42

cipitate recombinant His6-PH domain of GEF-H1. The pull-downs

probe was only affected by the head domain, which

stimulated Cdc42 activation, suggesting that the stimu-

lation of SRE-driven transcription (Figure 4C) might have been mediated by Cdc42.

expression of GEF-H1 remains comparatively constant

(Figure 4A), the interaction between cingulin and GEF- GEF-H1 Regulates G1/S Phase Transition

In fibroblasts as well as epithelial cells, RhoA signaling H1 might be part of a mechanism that inactivates RhoA

when epithelial cells reach confluence. Indeed, trans- regulates G1/S phase transition (Auer et al., 1998; Cole- man et al., 2004; Liberto et al., 2002; Olson et al., 1998).

fection of a GEF-H1 binding cingulin fragment resulted

in a reduction of stress fibers, suggesting inhibition of Since tight junctions regulate this step of the cell cycle (Balda et al., 2003), we tested whether GEF-H1 is im- RhoA (Figure 4B). To test whether increased expression

of cingulin affects GEF-H1 activity, we measured tran- portant for G1/S phase transition in MDCK cells using tetracycline-regulated depletion of GEF-H1 by RNA in- scriptional activity of SRF (Hill et al., 1995). Because

MDCK cells were found not to stimulate SRE-driven terference. Cells were plated at low density and syn- chronized in G1 phase in medium with low serum, a transcription in response to RhoA activation, we used

the retinal pigment epithelium cell line ARPE-19 as a treatment that results in efficient accumulation of MDCK cells in G1 phase (Balda et al., 2003). G1/S phase reporter system. Transfection of myc-cingulin and of

the GEF-H1 binding mutant myc-cingulinR+T inhibited transition was then stimulated by the addition of serum and cells synthesizing DNA were visualized by bromo- SRE-driven transcription (Figure 4C). Transfection of

the cingulin head domain stimulated the response deoxyuridine labeling. More than 70% of cells entered S phase in wild-type MDCK cells (Figure 5A). In con- whereas expression of the GEF-H1 binding myc-cin-

gulin 782-1025 fragment was sufficient to inhibit tran- trast, GEF-H1 depletion inhibited G1/S transition as only about 30% of the cells incorporated bromodeoxy- scription. Importantly, expression of GEF-H1 stimulated

luciferase expression, and cotransfection of full-length uridine. GEF-H1 is thus required for efficient serum- induced G1/S phase transition in MDCK cells.

cingulin or mutants that contained the GEF-H1 binding

site inhibited, indicating that cingulin can inhibit GEF- We next tested whether inhibition of GEF-H1 by cin- gulin is sufficient to inhibit G1/S phase transition. Trans- H1 function.

To monitor Rho activation directly, we used a FRET- fected MDCK cells were synchronized and labeled with bromodeoxyuridine. The cells were then double labeled based assay that makes use of a fusion protein in which

YFP is separated from CFP by a Rho binding domain for transfected myc-tagged cingulin and bromodeoxy-

uridine, and the fractions of cells stained with either

(Yoshizaki et al., 2003). Binding of Rho-GTP results in a

Tight Junctions and Regulation of Rho 781

Figure 4. Inhibition of GEF-H1 and Rho Activation by Cingulin

(A) MDCK cells grown to 20%, 50%, or 100% confluence were lyzed and equal amounts of protein were analyzed by immunoblotting for the expression of GEF-H1, cingulin, and, as a loading control,α-tubulin. Note the pronounced upregulation of cingulin with cell confluence.

(B) MDCK cells were transiently transfected with a truncated cingulin construct lacking the head domain (myc-cingulinR+T). Cells were stained with antibodies against myc, GEF-H1, and fluorescent phalloidin. Note the reduced appearance of stress fibers in transfected cells.

(C) Inhibition of SRE-driven transcription by cingulin. ARPE-19 cells were cotransfected with a plasmid containing a SRE driving firefly luciferase expression, one with a control promoter regulating renilla luciferase expression, and the indicated expression vectors. After 30 hr, the luciferases were assayed and the ratios of the values for firefly divided by those from renilla luciferase calculated. Shown are the means ± 1 SD of a typical experiment performed in triplicates. Both panels were normalized to plasmid controls without GEF-H1 cotransfec- tion. Asterisks indicate p values smaller than 0.05 that were calculated with two-tailed t tests comparing the single transfections to plasmid controls and the double transfections to GEF-H1 transfections.

(D) Inhibitin of RhoA activation by cingulin. MDCK cells were transfected with pRaichu-RBD, a Rho-specific FRET probe, and the indicated expression vectors. TAT-C3 labels cells that were incubated with TAT-modified C3 between the transfection and cell lysis. After cell lysis, the emission for YFP (530 nm) and CFP (475 nm) was measured with an excitation wavelength of 430 nm, and the ratios were calculated. Shown are the means ± 1 SD of a typical experiment (n = 4). Asterisks label p values smaller than 0.05 that refer to comparisons of single transfections with plasmid controls and double transfections with GEF-H1-transfected samples. Note, increased YFP emission indicates Rho inactivation.

(E) Downregulation of GEF-H1 in MDCK cells by tetracycline-regulated RNA interference. Confluent MDCK cells expressing control RNA duplexes or GEF-H1-directed RNA duplexes were treated with tetracycline for 3 days. Expression of GEF-H1 andα-tubulin was then analyzed by immunoblotting with monoclonal antibodies B4/7 and 1A2, respectively.

(F) The Rho-specific FRET probe pRaichu-RBD was transfected into tetracycline treated control or GEF-H1 RNAi cells together with empty expression vector or full-length or truncated cingulin. The extracts were then analyzed as in (D) (asterisks, p < 0.05).

(G). FRET probes specific for RhoA, Rac1, or Cdc42 were transfected into MDCK cells together with the indicated cingulin constructs (asterisks, p < 0.05). Note, decreased YFP emission indicates inactivation.

one or both antibodies were determined. Myc-cingulin These observations suggest that increased expression of cingulin inhibits G1/S phase transition and that this efficiently inhibited G1/S phase transition and so did

the expression of the construct containing the rod and inhibitory function maps to the GEF-H1 binding site.

the tail domain (Figure 5B). Whereas expression of the

head domain alone did not have an effect, transfection Cingulin Depletion Activates RhoA Signaling To test whether endogenous cingulin is indeed impor- of myc-cingulin 782-1025, which binds and inhibits

GEF-H1, inhibited bromodeoxyuridine incorporation. tant for the regulation of RhoA signaling, we depleted

depletion thus results in activation of RhoA and stimu- lation of RhoA signaling.

Discussion

We present evidence that the RhoA exchange factor GEF-H1 interacts with cingulin, resulting in inhibition of the GEF. Because cingulin expression increases with increasing cell confluence (Figure 4), the cingulin/GEF- H1 interaction provides a mechanism that links regula- tion of RhoA signaling to cell confluence.

Regulation of GEF-H1 and RhoA Signaling by Cingulin

Cingulin binds to the PH domain of GEF-H1, a domain important for GEF-H1’s transforming activity (White- head et al., 1995). PH domain-mediated interactions are important for the targeting of several Dbl family mem- bers to specific subcellular sites, including the actin cy- toskeleton, and mediate interactions with proteins and

lipids (Bellanger et al., 2000; Hoffman and Cerione,

2002; Olson et al., 1997; Zheng, 2001). PH domains of

some GEFs bind phosphoinositides; however, binding

affinities are relatively low and phosphoinositides may

not be required for membrane targeting (Baumeister et

al., 2003; Snyder et al., 2001). The PH domain of GEF-

Hoechst 33258. The percentages of cells were then determined by

H1 does not bind to the main plasma membrane phos-

phoinositides in vitro (not shown).

12 fields per experiment and sample. Shown are means ± 1 SD of

Junctional targeting of GEF-H1 is regulated by the

N-terminal domain, which binds microtubules as well

as the PH domain. In the full-length protein, the PH do-

main is required for microtubule binding, suggesting

labeled with bromodeoxyuridine as in (A). Shown are means ± 1 SD

that the interaction between the PH domain and the C1

domain is part of a mechanism that regulates recruit- ment to different cytoskeletal structures. The C1 do- main might also have other interaction partners such as diacylglycerol. However, we have not been able to cingulin expression by RNA interference using a tetra-

cycline regulated promoter. Incubation with tetracycline detect binding of this lipid (not shown).

Binding to cingulin as well as microtubules inhibits resulted in a strong reduction of cingulin expression in

cells expressing cingulin-directed RNA duplexes but GEF-H1 function (Krendel et al., 2002). It will thus be important to understand how the C1/PH domain in- not in cells expressing control RNA duplexes (Figure

6A). The expression level of GEF-H1 was not signifi- teraction is modulated to determine the sizes of these pools as well as the amount of free GEF. Our data indi- cantly affected by the tetracycline treatment. Cingulin-

depleted cells had an irregular morphology (Figure 6B), cate that the level of cingulin expression is an important determinant of GEF-H1-mediated Rho regulation. It is reminiscent of cells expressing constitutively active

RhoA (Jou and Nelson, 1998). The remaining cingulin possible that other parameters affect GEF-H1 activity as well. GEF-H1 can be phosphorylated by Pak1, re- still associated with cell-cell junctions, and the GEF-H1

staining became more cytoplasmic (Figure 6C). The sulting in binding to 14-3-3 (Zenke et al., 2004). How- ever, Pak1 phosphorylation and 14-3-3 binding do not tight junction proteins ZO-1 and occludin as well as the

adherens junction protein

-catenin remained junction appear to affect the activity or localization of GEF-H1 (Zenke et al., 2004).

associated (Figure 6D).

To test whether depletion of cingulin stimulates Rho Cingulin binds to several TJ proteins and to F-actin (Cordenonsi et al., 1999; D’Atri and Citi, 2001). Cingulin activation, we used the FRET probe. Depletion of cin-

gulin decreased YFP emission, suggesting stimulation expression in several cell types increases upon inhibi- tion of histone deacetylase, a treatment that can induce of Rho (Figure 7A). If activation of RhoA was monitored

with a pull-down experiment using a fusion protein con- differentiation (Bordin et al., 2004). This is supported by the observation that cingulin expression increases with taining the Rho binding domain of rhotekin, increased

levels of RhoA were detected in precipitates derived cell density (Figure 4). Targeted deletion of the head domain of cingulin results in expression of a truncated from cingulin-depleted cells (Figure 7B), indicating in-

creased activation of RhoA. RhoA signaling inhibits form of cingulin and altered expression of genes regu- lating endodermal differentiation (Guillemot et al., myosin light chain phosphatase by phosphorylation.

Cingulin depletion indeed stimulated enhanced phos- 2004). Although this was paralleled by altered expres-

sion of GEF-H1, it is not clear whether this was caused

phorylation of the phosphatase (Figure 7C). Cingulin

Tight Junctions and Regulation of Rho 783

Figure 6. Regulated Depletion of Cingulin Af- fects Cell Morphology and GEF-H1 Local- ization

(A) Stable MDCK cells expressing control or cingulin-directed RNA duplexes under the control of a tetracycline-inducible promoter were induced for 40 hr. Expression of cin- gulin,α-tubulin, and GEF-H1 was then ana- lyzed by immunoblotting.

(B) Depletion of cingulin and cell morphol- ogy. Control- and Cingulin-RNAi cells were cultured at 40% confluence with or without tetracycline. After 2 days, the cell morphol- ogy was analyzed by phase contrast micro- scopy.

(C and D) Cingulin-RNAi cells were cultured without or with tetracycline for 40 hr before processing for immunofluorescence with anti-GEF-H1 and anti-cingulin antibodies, or antibodies specific for ZO-1, occludin, or β-catenin. Note, cingulin depletion results in increased cytoplasmic staining of GEF-H1 but not ZO-1, occludin, andβ-catenin.

by the expression of the truncated form of cingulin, kuyama et al., 2004; Irie et al., 2004; Kawakatsu et al., 2005; Sander et al., 1998). Our observations now link which contains the GEF-H1 binding site, or an adaptive

response that occurred during the selection. However, tight-junction formation with inactivation of GEF-H1 and inhibition of RhoA signaling, suggesting that the the present results suggest that the effects of cingulin

mutation on gene expression may in part be due to ef- changes in Rho GTPase signaling that occur at conflu- ence are orchestrated by the recruitment of different fects on Rho signaling.

GEFs to the forming junctional complex.

Inactivation of RhoA signaling is thought to be impor- Junction Formation and RhoA Inactivation

tant for the inhibition of signaling pathways that pro- Increased expression of cingulin results in lower levels

mote proliferation in various cell types including epithe- of active RhoA, whereas reduced cingulin expression

lial cells (Coleman et al., 2004). Our results indicate that causes higher levels of active RhoA. These observa-

GEF-H1 promotes G1/S phase transition and that its tions suggest a molecular mechanism for the inhibition

inhibition by cingulin binding regulates cell-cycle pro- of RhoA signaling with cell confluence that is linked to

gression. Tight junctions recruit other signaling compo- the formation of tight junctions: in low confluent cells

nents that have been linked to the downregulation of expression of cingulin is low and GEF-H1 is primarily

signaling pathways that promote proliferation and, in cytoplasmic; with increasing cell density, cingulin accu-

particular, G1/S phase transition. These include the tu- mulates at forming tight junctions resulting in seque-

mor suppressor PTEN and a complex formed by the stration of free GEF-H1 at tight junctions and inhibition

transcription factor ZONAB and CDK4 (Balda et al., of RhoA signaling.

2003; Matter and Balda, 2003; Wu et al., 2000). It is not Engaging cadherins by plating transfected CHO cells

known whether these signaling systems are regulated on immobilized ligands stimulates the activity of

by RhoA. Nevertheless, the data described here link p190RhoGAP (Noren et al., 2003). Although CHO cells

formation of tight junctions to inactivation of RhoA sig- do not form a junctional complex, this suggests that

naling and inhibition of cell-cycle progression, support- different types of intercellular junctions contribute to

ing a model according to which tight-junction assembly the regulation of RhoA signaling: adherens junctions by

serves as an indicator of epithelial cell density that pro- activating a Rho GAP and tight junctions by inhibiting

gressively inhibits different proliferation promoting sig- a Rho GEF. Our results indicate that the regulation of

naling pathways with increasing cell density.

GEF-H1 by cingulin makes a significant contribution to the regulation of RhoA in epithelial cells because total

Experimental Procedures

cellular levels of active RhoA were found to change sig-

nificantly in response to changes in cingulin expression.

The canine GEF-H1 sequence was used for all cDNA constructs

When epithelial cells reach confluence, RhoA is inac-

(Benais-Pont et al., 2003). A C-terminally VSV-tagged full-length

tivated, whereas Rac1 and Cdc42 are activated (Braga,

protein was cloned into pcDNA4/TO (Invitrogen). All mutants were

2002). Strikingly, both processes appear to be medi-

generated by PCR and confirmed by sequencing. GEF-⌬CTD-VSV

ated by GEFs that are recruited to the forming junc-

was generated by inserting the VSV-tag after the codon for amino

tional complex. Adherens junctions recruit GEFs for

Cdc42 and Rac1, resulting in stimulation of Rac1 and

Cdc42, and stabilization of the junctional complex (Fu-

residues 217 to 600; PH, residues 447 to 600; CTD, residues 577 to 986. For cingulin fusion proteins, the indicated sequences were cloned into pGEX-4T-1 and pcDNA3.1myc-His to generate C-ter- minally tagged molecules. For regulated RNAi, the 3#-end of the mouse U6 promoter in mU6pro was changed into a tetracycline operator, and annealed oligonucleotides coding for RNA duplexes were inserted (van de Wetering et al., 2003; Yu et al., 2002). The sequence 5#-AAGGCCACCATCTATGGCATC-3# of cingulin was targeted. For GEF-H1, the sequences were as previously described (Benais-Pont et al., 2003). The RNAi vectors were cotransfected with pcDNA6/TR and pCB6, and clones were selected with blastici- din and G418.

Immunoprecipitations, Immunoblotting, and Pull-Down Assays For immunoprecipitations, MDCK cells were extracted with 10 mM Hepes (pH 7.4), containing 150 mM NaCl, 1% Triton X-100, 0.5%

sodium deoxycholate, 0.2% sodium dodecylsulfate, 1% Empigen BB, and a cocktail of protease inhibitors. GEF-H1 was immuno- precipitated with antibody B4/7 conjugated to Protein G-sepharose (Benais-Pont et al., 2003). GST and His6-fusion proteins were gen- erated as described (Balda and Matter, 2000; Balda et al., 1996).

For GST pull-down assays with GEF-H1 fusion proteins, cells were extracted with 1% Triton X-100 and extracts were preadsorbed with inactive beads for 15 min prior to incubation with glutathione- sepharose beads coated with 15␮g of fusion proteins. To monitor interactions between recombinant proteins, 0.5␮M GST-cingulin proteins were bound to beads, washed three times with buffer S (PBS, 1% Triton X-100, 1 mM DTT, and protease inhibitors), incu- bated with 0.5␮M His6-PH domain for 3 hr at 4°C in the presence of 3% BSA, and then washed twice with buffer S and once with PBS. Saturation of binding was tested using 0.5, 1, and 1.5␮M His₆-PH domain. For immunoblotting, samples were separated on 6 to 15% gradient gels and then transferred to nitrocellulose mem- branes (Balda et al., 1996). GEF-H1 was detected with mAb B4/7, α-tubulin with mAb 1A2, cingulin with a rabbit anti-cingulin anti- body, the VSV-epitope with mAb P5D4, the myc-epitope with mAb 9E10 (Benais-Pont et al., 2003; Cordenonsi et al., 1999; Kreis, 1987). RhoA pull-down assays were performed as described (Be- nais-Pont et al., 2003). Antibodies specific for myosin light chain phosphatase and phosphorylated (T696) MYPT-1 were obtained form Upstate Biotechnology (Lake Placid, NY).

Immunofluorescence and Bromodeoxyuridine Incorporation Cells were fixed with methanol at −20°C or 3% paraformaldehyde and then processed for double immunofluorescence using FITC- and Cy3-conjugated donkey secondary antibodies (Balda et al., 1996). For triple labeling with phalloidin, FITC-phalloidin was used together with Cy3- and Cy-5-conjugated secondary antibodies.

The following antibodies were used for immunofluorescence: GEF- Figure 7. Depletion of Cingulin and RhoA Activation H1, mAb B4/7; cingulin, rabbit polyclonal antibody;α-tubulin, mAb (A) MDCK cells were transfected with the Rho-specific FRET probe 1A2; the VSV-epitope, mAb P5D4 or a rabbit anti-peptide antibody;

and the indicated vectors. After 30 hr, emission for YFP and CFP the myc-epitope, a rabbit polyclonal antibody (MBL Laboratories).

was measured (excitation, 430 nm) and the ratios were calculated. For bromodeoxyuridine incorporation, cells were plated on cover- Shown are the means ± 1 SD of a typical experiment (n = 6). slips and incubated in 0.1% serum, resulting in accumulation of (B) Wild-type, two different clones of cingulin RNAi cells (CinRiTc1/ about 75% of the cells in Go/G1 phase (Balda et al., 2003). If cells c2), as well as control RNAi cells (ConRiT) were cultured without or were double labeled for transfected cingulin, anti-myc antibody in- with tetracycline for 40 hr. Cell extracts were then incubated with cubations were added after the initial methanol fixation. After GST with or without the Rho binding domain of rhotekin. Samples washing, the samples were fixed with 3% paraformaldehyde for 30 of total cell extracts and pellets were analyzed by immunoblotting min and then denatured with hydrochloric acid followed by staining using anti-RhoA antibodies. with anti-bromodeoxyuridine antibodies (Balda et al., 2003). Epiflu- (C) GEF-H1-VSV expressing cells, control RNAi cells (ConRiT), and orescence images were taken with a Leica DM1 RB and confocal two different clones of cingulin RNAi cells (CinRiTc1/c2) were cul- images with a Zeiss LSM 510 or a Leica LCS SP2 using 63× oil tured without or with tetracycline for 40 hr, lysed, and analyzed by immersion objectives.

immunoblotting with antibodies against phosphorylated (p-MYP) or

total myosin light chain phosphatase (MYP). The graph shows SRE Reporter and FRET Assays

quantification of the immunoblots by densitometric scanning. Reporter assays were done by cotransfecting ARPE-19 cells with Shown are the means ± 1 SD (n = 4). a plasmid containing a SRE-containing promoter driving firefly lu- ciferase expression (Clonetech), a control plasmid for renilla lucifer- ase expression, and the indicated expression vectors (Balda and Matter, 2000). After 30 hr, firefly luciferase was measured and stan- TR was cotransfected and cells were selected with blasticidin and

zeocin. For GEF-H1 domain constructs, the sequences encoding dardized using the renilla luciferase values. For FRET assays, pRaichu-RBD, pRaichu-RhoA, pRaichu-Rac1, and pRaichu-Cdc42 the following residues were inserted into pcDNA4/TO, pGEX-4T-3,

or pRSET-C: C1, residues 1 to 96; C1/ID, residues 1 to 231; DH/PH, were cotransfected with the indicated expression and RNAi plas-

Tight Junctions and Regulation of Rho 785

mids into 50% confluent MDCK cells cultured in 12-well plates (Itoh Cereijido, M., Shoshani, L., and Contreras, R.G. (2000). Molecular physiology and pathophysiology of tight junctions. I. Biogenesis of et al., 2002; Yoshizaki et al., 2003). 30 hr after the transfection, the

cells were washed with cold PBS and lyzed with 300␮l/well of 20 tight junctions and epithelial polarity. Am. J. Physiol.279, G477–

G482.

mM Tris (pH 7.5), 100 mM NaCl, 5 mM MgCl2, 0.5% Triton X-100,

and 20␮g/ml PMSF. The lysates were transferred to 96-well plates Citi, S., Sabanay, H., Jakes, R., Geiger, B., and Kendrick Jones, J.

for centrifugation at 2000 g for 5 min. Fluorescence was measured (1988). Cingulin, a new peripheral component of tight junctions. Na- with a Perkin Elmer LS 50B fluorometer (excitation, 430 nm; emis- ture333, 272–276.

sion, 475 and 530 nm). The ratios of emission at 530 nm/475 nm

Coleman, M.L., Marshall, C.J., and Olson, M.F. (2004). RAS and were calculated. Significance of the results in both types of assay

RHO GTPases in G1-phase cell-cycle regulation. Nat. Rev. Mol. Cell was determined with two-tailed t tests.

Biol.5, 355–366.

Coleman, M.L., Sahai, E.A., Yeo, M., Bosch, M., Dewar, A., and Olson, M.F. (2001). Membrane blebbing during apoptosis results Acknowledgments

from caspase-mediated activation of ROCK I. Nat. Cell Biol. 3, 339–345.

We thank Dr. Matsuda for the FRET probes and the members of

Cordenonsi, M., D’Atri, F., Hammar, E., Parry, D.A., Kendrick-Jones, our laboratories for critical reading of the manuscript. This research

J., Shore, D., and Citi, S. (1999). Cingulin contains globular and was supported by Cancer Research UK (CUK) grant number

coiled-coil domains and interacts with ZO-1, ZO-2, ZO-3, and myo- C1483/A2632, The Wellcome Trust (063661 and 066100), and the

sin. J. Cell Biol.147, 1569–1582.

Medical Research Council. The Citi team was supported by the

D’Atri, F., and Citi, S. (2001). Cingulin interacts with F-actin in vitro.

Swiss National Science Foundation and the Swiss Cancer League.

FEBS Lett.507, 21–24.

D’Atri, F., Nadalutti, F., and Citi, S. (2002). Evidence for a functional Received: June 15, 2004

interaction between cingulin and ZO-1 in cultured cells. J. Biol.

Revised: January 5, 2005

Chem.277, 27757–27764.

Accepted: March 3, 2005

Etienne-Manneville, S., and Hall, A. (2002). Rho GTPases in cell Published: April 3, 2005

biology. Nature420, 629–635.

Fukata, M., and Kaibuchi, K. (2001). Rho-family GTPases in cad- References

herin-mediated cell-cell adhesion. Nat. Rev. Mol. Cell Biol.2, 887–

Anderson, J.M., Van Itallie, C.M., and Fanning, A.S. (2004). Setting 897.

up a selective barrier at the apical junction complex. Curr. Opin. Fukuyama, T., Ogita, H., Kawakatsu, T., Fukuhara, T., Yamada, T.,

Cell Biol.16, 140–145. Sato, T., Shimizu, K., Nakamura, T., Matsuda, M., and Takai, Y.

(2004). Involvement of the c-Src-Crk-C3G-Rap1 signaling in the Auer, K.L., Park, J.S., Seth, P., Coffey, R.J., Darlington, G., Abo,

nectin-induced activation of Cdc42 and formation of adherens A., McMahon, M., Depinho, R.A., Fisher, P.B., and Dent, P. (1998).

junctions. J. Biol. Chem.280, 815–825.

Prolonged activation of the mitogen-activated protein kinase path-

way promotes DNA synthesis in primary hepatocytes from p21Cip- Glaven, J.A., Whitehead, I.P., Nomanbhoy, T., Kay, R., and Cerione, 1/WAF1-null mice, but not in hepatocytes from p16INK4a-null mice. R.A. (1996). Lfc and Lsc oncoproteins represent two new guanine Biochem. J.336, 551–560. nucleotide exchange factors for the Rho GTP-binding protein. J.

Biol. Chem.271, 27374–27381.

Balda, M.S., and Matter, K. (2000). The tight junction protein ZO-1

and an interacting transcription factor regulate ErbB-2 expression. Glaven, J.A., Whitehead, I., Bagrodia, S., Kay, R., and Cerione, R.A.

EMBO J.19, 2024–2033. (1999). The Dbl-related protein, Lfc, localizes to microtubules and mediates the activation of Rac signaling pathways in cells. J. Biol.

Balda, M.S., and Matter, K. (2003). Epithelial cell adhesion and the

Chem.274, 2279–2285.

regulation of gene expression. Trends Cell Biol.13, 310–318.

Gonzalez-Mariscal, L., Betanzos, A., Nava, P., and Jaramillo, B.E.

Balda, M.S., Garrett, M.D., and Matter, K. (2003). The ZO-1 associ-

(2003). Tight junction proteins. Prog. Biophys. Mol. Biol.81, 1–44.

ated Y-box factor ZONAB regulates epithelial cell proliferation and

Guillemot, L., Hammar, E., Kaister, C., Ritz, J., Caille, D., Jond, L., cell density. J. Cell Biol.160, 423–432.

Bauer, C., Meda, P., and Citi, S. (2004). Disruption of the cingulin Balda, M.S., Whitney, J.A., Flores, C., González, S., Cereijido, M.,

gene does not prevent tight junction formation but alters gene ex- and Matter, K. (1996). Functional dissociation of paracellular per-

pression. J. Cell Sci.117, 5245–5256.

meability and transepithelial electrical resistance and disruption of

Hill, C.S., Wynne, J., and Treisman, R. (1995). The Rho family the apical-basolateral intramembrane diffusion barrier by expres-

GTPases RhoA, Rac1, and CDC42Hs regulate transcriptional acti- sion of a mutant tight junction membrane protein. J. Cell Biol.134,

vation by SRF. Cell81, 1159–1170.

1031–1049.

Hoffman, G.R., and Cerione, R.A. (2002). Signaling to the Rho Baumeister, M.A., Martinu, L., Rossman, K.L., Sondek, J., Lemmon,

GTPases: networking with the DH domain. FEBS Lett.513, 85–91.

M.A., and Chou, M.M. (2003). Loss of phosphatidylinositol 3-phos-

phate binding by the C-terminal Tiam-1 pleckstrin homology do- Irie, K., Shimizu, K., Sakisaka, T., Ikeda, W., and Takai, Y. (2004).

main prevents in vivo Rac1 activation without affecting membrane Roles and modes of action of nectins in cell-cell adhesion. Semin.

targeting. J. Biol. Chem.278, 11457–11464. Cell Dev. Biol.15, 643–656.

Bellanger, J.M., Astier, C., Sardet, C., Ohta, Y., Stossel, T.P., and Itoh, R.E., Kurokawa, K., Ohba, Y., Yoshizaki, H., Mochizuki, N., and Debant, A. (2000). The Rac1- and RhoG-specific GEF domain of Matsuda, M. (2002). Activation of rac and cdc42 video imaged by Trio targets filamin to remodel cytoskeletal actin. Nat. Cell Biol.2, fluorescent resonance energy transfer-based single-molecule

888–892. probes in the membrane of living cells. Mol. Cell. Biol.22, 6582–

6591.

Benais-Pont, G., Punn, A., Flores-Maldonado, C., Eckert, J., Ra-

poso, G., Fleming, T.P., Cereijido, M., Balda, M.S., and Matter, K. Jou, T.S., and Nelson, W.J. (1998). Effects of regulated expression (2003). Identification of a tight junction-associated guanine nucleo- of mutant RhoA and Rac1 small GTPases on the development of tide exchange factor that activates Rho and regulates paracellular epithelial (MDCK) cell polarity. J. Cell Biol.142, 85–100.

permeability. J. Cell Biol.160, 729–740. Kawakatsu, T., Ogita, H., Fukuhara, T., Fukuyama, T., Minami, Y., Bordin, M., D’Atri, F., Guillemot, L., and Citi, S. (2004). Histone de- Shimizu, K., and Takai, Y. (2005). Vav2 as a Rac-GEF responsible acetylase inhibitors up-regulate the expression of tight junction for the nectin-induced, c-Src- and Cdc42-mediated activation of proteins. Mol. Cancer Res.2, 692–701. Rac. J. Biol. Chem.280, 4940–4947.

Kreis, T.E. (1987). Microtubules containing detyrosinated tubulin are Braga, V.M. (2002). Cell-cell adhesion and signalling. Curr. Opin.

Cell Biol.14, 546–556. less dynamic. EMBO J.6, 2597–2606.

Krendel, M., Zenke, F.T., and Bokoch, G.M. (2002). Nucleotide ex- change factor GEF-H1 mediates cross-talk between microtubules and the actin cytoskeleton. Nat. Cell Biol.4, 294–301.

Liberto, M., Cobrinik, D., and Minden, A. (2002). Rho regulates p21(CIP1), cyclin D1, and checkpoint control in mammary epithelial cells. Oncogene21, 1590–1599.

Matter, K., and Balda, M.S. (2003). Signalling to and from tight junc- tions. Nat. Rev. Mol. Cell Biol.4, 225–236.

Noren, N.K., Niessen, C.M., Gumbiner, B.M., and Burridge, K.

(2001). Cadherin engagement regulates Rho family GTPases. J.

Biol. Chem.276, 33305–33308.

Noren, N.K., Arthur, W.T., and Burridge, K. (2003). Cadherin engage- ment inhibits RhoA via p190RhoGAP. J. Biol. Chem.278, 13615–

13618.

Olson, M.F., Sterpetti, P., Nagata, K., Toksoz, D., and Hall, A. (1997).

Distinct roles for DH and PH domains in the Lbc oncogene. Onco- gene15, 2827–2831.

Olson, M.F., Paterson, H.F., and Marshall, C.J. (1998). Signals from Ras and Rho GTPases interact to regulate expression of p21Waf1/

Cip1. Nature394, 295–299.

Ren, Y., Li, R., Zheng, Y., and Busch, H. (1998). Cloning and charac- terization of GEF-H1, a microtubule-associated guanine nucleotide exchange factor for Rac and Rho GTPases. J. Biol. Chem.273, 34954–34960.

Ridley, A.J. (2004). Rho proteins and cancer. Breast Cancer Res.

Treat.84, 13–19.

Sander, E.E., van Delft, S., ten Klooster, J.P., Reid, T., van der Kam- men, R.A., Michiels, F., and Collard, J.G. (1998). Matrix-dependent Tiam1/Rac signaling in epithelial cells promotes either cell-cell ad- hesion or cell migration and is regulated by phosphatidylinositol 3-kinase. J. Cell Biol.143, 1385–1398.

Schneeberger, E.E., and Lynch, R.D. (2004). The tight junction: a multifunctional complex. Am. J. Physiol.286, C1213–C1228.

Snyder, J.T., Rossman, K.L., Baumeister, M.A., Pruitt, W.M., Sider- ovski, D.P., Der, C.J., Lemmon, M.A., and Sondek, J. (2001). Quanti- tative analysis of the effect of phosphoinositide interactions on the function of Dbl family proteins. J. Biol. Chem.276, 45868–45875.

Tsukita, S., Furuse, M., and Itoh, M. (2001). Multifunctional strands in tight junctions. Nat. Rev. Mol. Cell Biol.2, 286–293.

van de Wetering, M., Oving, I., Muncan, V., Pon Fong, M.T., Brantjes, H., van Leenen, D., Holstege, F.C., Brummelkamp, T.R., Agami, R., and Clevers, H. (2003). Specific inhibition of gene ex- pression using a stably integrated, inducible small-interfering-RNA vector. EMBO Rep.4, 609–615.

Whitehead, I., Kirk, H., Tognon, C., Trigo-Gonzalez, G., and Kay, R.

(1995). Expression cloning of lfc, a novel oncogene with structural similarities to guanine nucleotide exchange factors and to the regu- latory region of protein kinase C. J. Biol. Chem.270, 18388–18395.

Wu, X., Hepner, K., Castelino-Prabhu, S., Do, D., Kaye, M.B., Yuan, X.J., Wood, J., Ross, C., Sawyers, C.L., and Whang, Y.E. (2000).

Evidence for regulation of the PTEN tumor suppressor by a mem- brane-localized multi-PDZ domain containing scaffold protein MAGI-2. Proc. Natl. Acad. Sci. USA97, 4233–4238.

Yoshizaki, H., Ohba, Y., Kurokawa, K., Itoh, R.E., Nakamura, T., Mochizuki, N., Nagashima, K., and Matsuda, M. (2003). Activity of Rho-family GTPases during cell division as visualized with FRET- based probes. J. Cell Biol.162, 223–232.

Yu, J.Y., DeRuiter, S.L., and Turner, D.L. (2002). RNA interference by expression of short-interfering RNAs and hairpin RNAs in mamma- lian cells. Proc. Natl. Acad. Sci. USA99, 6047–6052.

Zenke, F.T., Krendel, M., DerMardirossian, C., King, C.C., Bohl, B.P., and Bokoch, G.M. (2004). p21-activated kinase 1 phosphorylates and regulates 14–3-3 binding to GEF-H1, a microtubule-localized Rho exchange factor. J. Biol. Chem.279, 18392–18400.

Zheng, Y. (2001). Dbl family guanine nucleotide exchange factors.

Trends Biochem. Sci.26, 724–732.