the small gtp-binding proteins rho and rac induce t cell adhesion to the mucosal addressin madcam-1...

0014-2980/99/0909-2875$17.50+.50/0© WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999

The small GTP-binding proteins Rho and Racinduce T cell adhesion to the mucosal addressinMAdCAM-1 in a hierarchical fashion

Wei V. Zhang1, Yi Yang1, Randal W. Berg1, Euphemia Leung1 and Geoffrey W.Krissansen1

1 Department of Molecular Medicine, School of Medicine and Health Science, University ofAuckland, Auckland, New Zealand

Here we report that an activator (AIF4–) of heterotrimeric GTP-binding proteins (G-proteins)

and inhibitors (lovastatin and C3 exoenzyme) of small GTP-binding proteins regulate theinduction of § 4 g 7-mediated adhesion of TK-1 T lymphoma cells ( § 4+

g 7+g 1–) to the mucosal

addressin cell adhesion molecule MAdCAM-1. Activation of cell adhesion by AIF4– was abro-

gated by lovastatin, thereby establishing a link between heterotrimeric G-proteins and smallGTP-binding proteins in the regulation of § 4 g 7-mediated cell adhesion. Increased numbersof cells bound MAdCAM-1-coated microspheres following activation with AIF4

–, discountingan obligatory role for cell spreading in § 4 g 7-mediated cell adhesion. MAdCAM-1-Fc dimerstriggered ligand-induced clustering of § 4 g 7 in response to AIF4

– and Mn2+-induced activationof integrins. Hence § 4 g 7 cluster formation may be responsible, at least in part, for inducingcell adhesion in response to both extracellular and intracellular signals that impact on integ-rin function. Electroporation of constitutively active V14RhoA and V12Rac1 recombinantproteins into TK-1 cells revealed that both RhoA and Rac1 induce § 4 g 7 adhesion toMAdCAM-1. Activation is hierarchical since Rac1 is unable to directly activate § 4 g 7, butinduces cell adhesion via RhoA, whereas the transient induction of cell adhesion mediatedby RhoA is dependent on the activities of protein tyrosine kinases and protein kinase(s) C.

Key words: Integrin activation / MAdCAM-1 / RhoA / Rac1 / Integrin § 4 g 7

Received 10/9/98Revised 8/6/99Accepted 8/6/99

[I 18828]

Abbreviations: C3: C3 transferase exoenzyme G-protein:GTP-binding protein ICAM-1: Intercellular cell adhesionmolecule-1 LFA-1: Lymphocyte function-associated antigen1 MAdCAM-1: Mucosal addressin cell adhesion molecule-1PI-3K: Phosphatidylinositol 3-kinase VCAM-1: Vascular celladhesion molecule-1 Tiam-1: T lymphoma invasion andmetastatis-1 antigen GALT: Gut-associated lymphoid tissueSLC: Secondary lymphoid tissue chemokine GST: Glutathi-one S-transferase

1 Introduction

The control of leukocyte adhesiveness is critical toimmunity in maintaining effective homeostasis, lympho-cyte homing and recirculation, the localization of leuko-cytes at sites of inflammation, antigen presentation andother immune responses. The g 7 integrin subfamily ofleukocyte cell adhesion molecules consists of § 4 g 7 [1, 2]and § E g 7 [3, 4]. § 4 g 7 mediates the adherence of lym-phocytes to high endothelial venules (HEV) via its pre-

ferred ligand, mucosal addressin cell adhesion molecule(MAdCAM)-1 [5, 6], whereas § E g 7 mediates the adhe-sion of intraepithelial lymphocytes of the intestine to theintestinal epithelium by an interaction with E-cadherin [7,8]. In gene knockout mice deficient in the g 7 integrinsubfamily, formation of the gut-associated lymphoid tis-sue (GALT) is severely impaired due to a failure of g 7–/–

lymphocytes to adhere to blood vessel walls at the site oftransmigration into the GALT [9].

Intracellular signaling pathways that impinge on integrinsubunit cytoplasmic domains trigger changes in integrinconformation, clustering [10], adhesiveness/affinity [11,12], and cell spreading [13] which can all contribute toincreased cell adhesion. Chemoattractant chemokinesstimulate integrin-mediated cell adhesion by interactingwith serpentine receptors that are coupled to heterotri-meric GTP-binding proteins (G-proteins). Recently, sec-ondary lymphoid tissue chemokine (SLC), expressed inPeyer’s patches and lymph nodes, was shown to induce§ 4 g 7-mediated adhesion of T cells to MAdCAM-1 [14].

Pretreatment of leukocytes with pertussis toxin, which

Eur. J. Immunol. 1999. 29: 2875–2885 Hierarchical activation of T cell adhesion to MAdCAM-1 by Rho and Rac 2875

ADP-ribosylates and inactivates G-protein § i subunits,inhibits § 4- and g 2-integrin-mediated emigration of leu-kocytes into inflammatory sites [15], and lymphocyterecirculation and transmigration into peripheral lymphnodes or Peyer’s patches [16].

AIF4–, which mimics the + phosphate of GTP by promot-

ing the active conformation of heterotrimeric G-proteins[17], induces integrin lymphocyte function-associatedantigen (LFA)-1 ( § L g 2)-mediated adhesion of T cellhybridoma cells to intercellular cell adhesion molecule(ICAM)-1 [18]. Laudanna et al. [19] elegantly showed thatchemoattractant stimulation of transfected formyl pep-tide and IL-8 G-protein-coupled receptors led to nucleo-tide exchange on the small GTP-binding protein RhoA,and triggered lymphocyte adhesion to vascular celladhesion-molecule (VCAM)-1. This and a number ofother studies have provided a tentative link between G-coupled receptors and small GTP-binding proteins in theinduction of cell adhesion. Chemoattractant-induced§ 4 g 1-mediated adhesion of lymphocytes to VCAM-1, g 2

integrin-mediated adhesion of neutrophils to fibrinogen,and PMA-induced LFA-1-dependent lymphocyte aggre-gation were blocked by the C3 transferase exoenzyme ofClostridium botulinum (C3) [19, 20] which specificallyinhibits Rho by ADP-ribosylation. Elegant studies in 3T3fibroblasts suggest that cell motility is coordinately con-trolled by the sequential activation of different Rho pro-teins [21]. Here we have explored the roles of GTP-binding proteins in integrin § 4 g 7-mediated adhesion ofT cells to the mucosal addressin MAdCAM-1.

2 Results

2.1 Effects of regulators of GTP-binding proteinson > 4 I 7 activation

AIF4– induced the transient adhesion of TK-1 cells to

MAdCAM-1, where adhesion was optimal after 10 to15 min, but returned to basal levels after 60 min (Fig. 1A).Adhesion was specifically blocked by the DATK32 mAb,indicating that § 4 g 7 was specifically responsible forAIF4

–-induced cell binding (Fig. 1B). Adhesion was alsospecifically blocked by MAdCAM-1-Fc, indicating thatsoluble ligand was capable of binding to § 4 g 7, and sug-gesting that the ligand binding affinity of § 4 g 7 may beincreased in response to AIF4

– (Fig. 1C). Increased num-bers of cells bound soluble MAdCAM-1-Fc dimers fol-lowing Mn2+ treatment, however, we were not able todemonstrate an increase following AIF4

– treatment(Fig. 1D). Nevertheless, there was an increase in thenumber of cells binding MAdCAM-1-coated micro-spheres after both AIF4

– (24 % increase), and Mn2+ (31 %increase) treatment (Fig. 1E), versus control Mg2+-treated

(12 % increase) cells. Elegant experiments by Yauch etal. [10] revealed that binding of § 4 g 1 to VCAM-1 wasregulated through integrin diffusion/clustering. Treatmentof TK-1 cells either with Mg2+, AIF4

–, or Mn2+ alonecaused no change in the diffuse cellular distribution of§ 4 g 7 (Fig. 2G–I). In contrast, addition of MAdCAM-1-Fc

dimers caused § 4 g 7 cluster formation, which was partic-ularly noticeable in the case of Mn2+ treated cells(Fig. 2A–C). AIF4

– facilitated the formation of small clus-ters, whereas cluster formation with Mg2+ was neverconvincing. § 4 g 7 cluster formation was more obvious onaddition of MAdCAM-1-Fc-coated microspheres, suchthat large patches of fluorescence could be readilyobserved on both AIF4

–- and Mn2+-treated cells(Fig. 2D–F).

Lovastatin, an inhibitor of small GTP-binding proteins[22], abrogated AIF4

–-induced (Fig. 3A), basal, and PMA-induced TK-1 cell adhesion to MAdCAM-1 (Fig. 3B).Addition of mevalonic acid to bypass the blockade ofthe isoprenyl group synthesis by lovastatin completelyreversed the inhibition (Fig. 3B). Hence small GTP-binding proteins appear to be critical for § 4 g 7-mediatedcell adhesion induced by both heterotrimeric G-proteinsand PMA.

2.2 Effects of C3 on PMA-induced > 4 I 7activation

Pretreatment of TK-1 cells with C3 led to the ADP-ribosylation of RhoA at 23 kDa (Fig. 4A, lane 1). Ribosyla-tion of RhoA is at least 100-fold more efficient than onRac1 [23] (Fig. 4B). To determine whether we couldcompletely ADP-ribosylate endogenous cellular RhoA,TK-1 cells were first pretreated with different concentra-tions of C3 (1.25–50 ? g/ml, Fig. 4A, lanes 2–4). Cellswere then lysed and lysates subjected to an in vitro ADP-ribosylation with C3 using [32P] NAD. As shown in lanes 3and 4, cellular Rho A was completely ADP-ribosylated bypretreatment of unactivated TK-1 cells with 12.5–50 ? g/ml C3, since it was no longer able to be labelled in vitroby ADP-ribosylation (Fig. 4A). Treatment of cells with thephorbol ester PMA induced dose-dependent cell bindingto MAdCAM-1, where binding in response to 50 ng/mlPMA was optimal and equivalent to that obtained withMn2+ ions which induces maximal activation of § 4 g 7(Fig. 4C). C3 (25 ? g/ml) inhibited PMA-induced binding ofcells to MAdCAM-1 at low PMA concentrations, but wasless effective at higher concentrations of PMA, suggest-ing that PMA and C3 may compete with each other. Analternative explanation is that PKC works both upstreamand downstream of small GTP-binding proteins, henceincreased concentrations of PMA are able to bypass the

2876 W. V. Zhang et al. Eur. J. Immunol. 1999. 29: 2875–2885

Figure 1. Effect of AIF4– and Mn2+ on § 4 g 7 ligand binding and cell adhesion. (A) The trimeric G-protein activator AIF4

– induces theadhesion of TK-1 T cells to MAdCAM-1. TK-1 cells were cultured in the absence ( 1 ) and presence ( Æ ) of AIF4

– for different timeperiods, and added to glass slides coated with MAdCAM-1-Fc. (B) AIF4

–-induced TK-1 cell adhesion to MAdCAM-1 is § 4 g 7dependent. TK-1 cells were treated for 30 min at 4 °C with anti- § 4 g 7 mAb DATK32 and control mAb GK1.5 (anti-mouse CD4) andOKT3 (anti-human CD3). Cells were activated for 15 min with AIF4

–, and added to MAdCAM-1-coated slides. (C) Soluble dimericMAdCAM-1-Fc blocks AIF4

–-induced cell adhesion. TK-1 cells were suspended in buffer containing either Mg2+, AIF4–, or Mn2+,

and were either left untreated or treated for 1 h at 4 °C with MAdCAM-1-Fc (10 ? g/ml) or the control recombinant Fc molecule E-cadherin-Fc (10 ? g/ml). Cells were added to MAdCAM-1-coated slides. (D) Mn2+-induced activation of § 4 g 7 leads to increasedbinding of soluble MAdCAM-1. TK-1 cells were treated with Mg2+ (control for both panels), AIF4

– (left panel), and Mn2+ (rightpanel), and incubated with dimeric MAdCAM-1-Fc. Cells were fixed, and ligand binding measured by staining with anti-humanFc and FITC-labeled rat anti-mouse, followed by flow cytometry. Illustrated is the fluorescence intensity in logarithmic scale forcells activated with Mg2+ (shaded), and AIF4

– and Mn2+ (dark lines). (E) Both AIF4– and Mn2+ induce the binding of MAdCAM-1-

coated microspheres. TK-1 cells were treated with Mg2+ (control for both panels), AIF4– (left panel), and Mn2+ (right panel), and

incubated with MAdCAM-1-Fc-coated microspheres. Ligand binding was analyzed as above.


Figure 2. Activators of cell adhesion potentiate ligand-induced clustering of § 4 g 7. TK-1 cells were treated with AIF4– (A, D, G),

Mn2+ (B, E, H), and Mg2+ (C, F. I), and subsequently incubated with dimeric MAdCAM-1-Fc (A–C), MAdCAM-1-Fc-coated micro-spheres (D–F), or media (G–I). To visualize § 4 g 7 clustering, cells were stained with FITC-conjugated M293 mAb, followed byconfocal microscopy. The data are representative of two separate experiments.

requirement for RhoA and overcome the effect of C3 (seebelow).

2.3 Electroporation of RhoA and Rac1 into TK-1cells induces cell binding to MAdCAM-1

V14RhoA and V12Rac1 glutathione S-transferase (GST)fusion proteins were electroporated into TK-1 cells, lead-ing to enhanced (four- to fivefold) binding of TK-1 cells toMAdCAM-1, whereas GST protein alone had no effect(Fig. 5A, B). Cell binding was similar to that achieved withMn2+ ions, and was transient. Thus maximal cell bindingwas obtained within 2 h of electroporation, and slowlydeclined to background level by 24 h (Fig. 5C). V14RhoAinactivated ex vivo by pretreatment with C3 wascompletely unable to enhance cell binding (Fig. 5D).

2.4 Hierarchical activation of > 4 I 7: Rac1 doesnot independently activate > 4 I 7, butfunctions via RhoA

GST-V14RhoA was able to enhance TK-1 cell binding toMAdCAM-1, where endogenous cellular RhoA activityhad first been inhibited by pretreatment of cells with C3(Fig. 6A). In contrast, GST-V12Rac1 was not able to over-come the blockade imposed by C3 (Fig. 6B), suggestingthat the effects of Rac1 are mediated through RhoA.

2.5 Protein kinase inhibitors block RhoA-inducedcell adhesion

Given that the tyrosine kinase inhibitor genistein inhibitsthe Rho-induced formation of stress fibres in 3T3 fibro-blasts [24], we treated TK-1 cells with genistein followingthe introduction of V14RhoA by electroporation. Genis-tein completely inhibited basal and RhoA-induced celladhesion to MAdCAM-1 (Fig. 7A). Likewise the PKC-specific inhibitor bisindolymaleimide, and the proteinkinase A (PKA) and PKC inhibitor staurosporine, both


Figure 3. Effect of activators and inhibitors of small GTP-binding proteins on § 4 g 7-mediated cell adhesion. (A) AIF4

–-induced TK-1 cell adhesion to MAdCAM-1 is dependent onsmall GTP-binding proteins. TK-1 cells were cultured with orwithout (control) 40 ? M lovastatin for 24 h, and activated withAIF4

– 15 min prior to binding to MAdCAM-1. (B) Lovasta-tin, an inhibitor of small GTP-binding proteins, inhibitsPMA-induced § 4 g 7-mediated adhesion of TK-1 cells toMAdCAM-1. TK-1 cells were cultured with or without 40 ? Mlovastatin, and 1 mM mevalonate for 24 h, and activated with10 ng/ml PMA (immediately prior to binding), or left unacti-vated with Mg2+ to measure basal binding to MAdCAM-Fc.The experiment was repeated five times.

Figure 4. PMA-induced adhesion of TK-1 T cells toMAdCAM-1 is dependent on RhoA. (A) C3 preferentially[32P]ADP-ribosylates the 23-kDa RhoA protein in TK-1 cells.TK-1 cells were incubated overnight with 0 (lane 1), 1.25(lane 2), 12.5 (lane 3), and 50 (lane 4) ? g/ml C3. To confirmwhether ADP-ribosylation of cellular RhoA was complete,cells were lysed by freeze-thawing and an aliquot of thesupernatant labeled by ADP-ribosylation in vitro with 0.5 ? g/ml C3 and [32P]NAD. Labeled proteins were resolved byelectrophoresis on a 12.5 % acrylamide SDS-gel, and thegel autoradiographed. In vivo treatment of cells with 12.5 ? g/ml C3 (lane 3) caused complete ADP-ribosylation of endog-enous cellular RhoA (arrowed), since RhoA in cell lysateswas not able to be labeled in vitro by ADP-ribosylation. (B)C3 preferentially ADP-ribosylates RhoA. Purified GST-V14RhoA (lanes 1 and 3) and GST-V12Rac1 (lanes 2 and 4),were ADP-ribosylated by incubating 2 ? g of each GST fusionprotein for 60 min at 37 °C with 0–3 ? g/ml C3. The extent ofADP-ribosylation was determined by electrophoresis oftreated proteins on 12 % acrylamide SDS-gels, followed byautoradiography for 15 min (lanes 1 and 2) and 4 h (lanes 3and 4). (C) C3 inhibits PMA-induced adhesion of lympho-cytes to MAdCAM-1. TK-1 cells were incubated overnight inthe presence or absence of 25 ? g/ml C3, activated with 0 to50 ng/ml PMA and Mn2+ [6] as indicated, and added to glassslides coated with MAdCAM-1-Fc. The results are presentedas the mean absorbance at 595 nm, as a measure of themean number of cells bound in two wells. The experimentwas repeated twice.

abrogated basal and RhoA-induced cell adhesion(Fig. 7A). In contrast, low concentrations (0.5 to 1 ? M) ofwortmannin which specifically inhibit the lipid and serinekinase activities of phosphatidylinositol 3-kinase (PI-3K)[25] were without effect (Fig. 7B). However, higher con-centrations (10 ? M) capable of negating the activities ofan array of different kinases including PKC were inhibi-tory.


Figure 5. Electroporated V14RhoA and V12Rac1 proteins induce lymphocyte binding to MAdCAM-1. (A) GST-V14RhoA, GST-V12Rac1 and GST were electroporated into TK-1 cells, cells incubated for 2 h at 37 °C, and added to MAdCAM-1-Fc-coatedslides. Cells were pretreated with Mg2+ and Mn2+ to measure basal and maximal cell binding, respectively; and added to FCS-coated slides as a measure of background binding. The experiment was repeated thrice. (B) Visualization of binding of TK-1 cellsto MAdCAM-1 in response to activation with GST-V14RhoA, GST-V12Rac1, and GST. Illustrated are microscopic images of rep-resentative slides from the binding assay described in (A). (C) Transient activation of § 4 g 7 by RhoA. V14RhoA was electroporatedinto TK-1 cells, and cells were incubated for varying lengths of time, prior to adding to MAdCAM-1-Fc-coated slides. The experi-ment was repeated twice. (D) V14RhoA inactivated by C3 is unable to induce lymphocyte adhesion to MAdCAM-1. GST, GST-V14RhoA, and GST-V14RhoA that had first been ADP-ribosylated by treatment with C3, were introduced into TK-1 cells by elec-troporation, and cells added to MAdCAM-1-Fc-coated slides. Mg2+ and Mn2+-treated cells provide a measure of basal and maxi-mal cell binding, respectively. The experiment was repeated thrice.


Figure 6. Hierarchical signal transduction by Rho proteinsleads to the activation of § 4 g 7. TK-1 cells (107) were incu-bated with 25 ? g/ml C3 for 24 h at 37 °C, and GST-V14RhoA(A, rho + C3-treated TK-1 cells) and V12Rac1 (B; rac + C3-treated TK-1 cells) introduced by electroporation. Cells wereadded to slides coated with MAdCAM-1-Fc. Controlsincluded TK-1 cells that had either been treated with C3 (C3-treated TK-1 cells), or electroporated with V14RhoA (rho),V12Rac1 (rac), and GST (GST). Mg2+ and Mn2+-treated cellsprovide a measure of basal and maximal cell binding,respectively. The experiment was repeated twice.

Figure 7. RhoA-induced activation of § 4 g 7 is dependent onprotein kinases. (A) Genistein and PKC inhibitors blockRhoA-induced lymphocyte adhesion to MAdCAM-1. GST-V14RhoA was introduced into TK-1 cells by electroporation,and cells were incubated with 10 ? g/ml genistein, 100 nMstaurosporine and 10 nM bisindolylmaleimide for 3 h at37 °C. Cells were resuspended in buffer containing the inhib-itors, and added to MAdCAM-1-Fc-coated slides. Untreated(Mg2+) and GST-electroporated control cells were treatedwith inhibitors to measure effects on basal cell binding. (B)Low concentrations of wortmannin specific for PI-3K haveno effect on RhoA-induced cell adhesion. GST-V14RhoAwas introduced into TK-1 cells by electroporation, and cellswere incubated with the concentrations of wortmannin indi-cated (0 to 10 ? M) for 3 h at 37 °C. Cells were resuspended inbuffer containing the inhibitor, and added to MAdCAM-1-Fc-coated slides. The experiment was repeated twice. Thekinase inhibitors used above had no apparent effect on cellviability, as measured by trypan blue exclusion.

3 Discussion

In accord with a previous study [18], the heterotrimericG-protein activator AIF4

– induced the adhesion of TK-1T cells to immobilized MAdCAM-1. AIF4

–-induced adhe-sion was blocked by the small GTP-binding proteininhibitors lovastatin and C3, thereby linking heterotri-meric and small GTP-binding proteins in a pathway lead-ing to integrin-mediated cell adhesion [19]. A previousstudy had shown that AIF4

– activates trimeric G-proteins,but has no effect on the dissociation of GDP from ras-like monomeric GTP-binding proteins [26]. The receptorsinvolved in the G-protein-dependent activation of § 4 g 7-


mediated cell adhesion have not yet been identified, butare likely to include G-protein-coupled receptors for che-mokines such as SLC that are potentially involved in lym-phocyte traffic into the GALT. The implication is that het-erotrimeric G-proteins linked to cell surface receptorsactivate RhoA which in turn facilitates ligand-inducedcell adhesion by integrins.

We have shown for the first time that both RhoA andRac1 can facilitate ligand-induced adhesion mediated by§ 4 g 7, where Rac1 functions through RhoA. The latter is

deduced from the finding that Rac1 was unable to stimu-late lymphocyte binding to MAdCAM-1 when endoge-nous RhoA was inactivated by treatment with C3. Ourapproach was novel in that the GTP-binding proteinswere introduced into cells by electroporation. While elec-troporation is conventionally used to introduce DNA, ithas been used to make living cells permeable to anti-bodies [27]. This approach should be useful to dissectthe roles of other intracellular components in signalingcascades.

Our results point to a difference in the kinases required toactivate LFA-1-mediated cell adhesion via heterotrimericG-proteins [18] versus activation of § 4 g 7-mediated celladhesion by recombinant RhoA. Both AIF4

– activation ofLFA-1-mediated cell adhesion and RhoA-induced § 4 g 7-mediated adhesion are inhibited by the protein tyrosinekinase inhibitor genistein. In accord, Rho-inducedassembly of focal adhesions and actin stress fibres isalso blocked by genistein [24]. In contrast, activation ofLFA-1 adhesion by AIF4

– is not blocked by inhibitors ofPKC, whereas the PKC inhibitors bisindolylmaleimideand staurosporine inhibited both basal and RhoA-induced cell adhesion mediated by § 4 g 7. Since the PKCactivator PMA can activate cell adhesion by leukocyteintegrins including § 4 g 7 [28], and C3 inhibits PMA-induced LFA-1- [20] and § 4 g 7-dependent cell adhesion,it would seem that PKC works both upstream and down-stream of RhoA. The PI-3K inhibitor wortmannin, whichinhibits RANTES-induced T lymphocyte chemotaxis [29]and AIF4

–-induced activation of LFA-1 adhesion [18], hadno effect on RhoA-induced § 4 g 7-mediated cell adhesionat low concentrations where it is specific for PI-3K.Hence PI-3K appears to function downstream of trimericG-proteins, and upstream of the small GTP-binding pro-teins. This is in accord with the finding that wortmannininhibits membrane ruffling induced by platelet-derivedgrowth factor, but not by V12Rac1 [30, 31]. Similarly, theformation of lamellipodia in murine T cells in response toIL-2 is blocked by wortmannin, whereas constitutivemembrane ruffling in Val12Rac cells is insensitive towortmannin [32]. In contrast, Cdc42 and Rac1 appar-ently can induce integrin-mediated cell motility andinvasiveness of mammary epithelial cells in a Rho-

independent fashion through PI(3)K [33]. Potentiation ofligand-induced integrin-mediated cell adhesion by Rac1may explain how genes such as Tiam-1, that encodeGDP dissociation stimulator proteins for Rho-like GTPa-ses, induce invasion by T lymphoma cells [34].

Integrin-mediated cell adhesion in response to activationof small GTP-binding proteins could arise from changesin integrin conformation, clustering, adhesiveness/affin-ity, and/or from cell spreading. Several studies havebeen able to demonstrate a direct increase in integrinaffinity by demonstrating increased binding of solubleintegrin ligand to cells in response to an activator [11,12]. Soluble MAdCAM-1 at high concentration must beable to bind § 4 g 7 on TK-1 cells as it was able to blockAIF4

–-activated TK-1 cell adhesion to immobilizedMAdCAM-1. We could detect only a very weak increasein the number of TK-1 cells binding to soluble MAdCAM-1-Fc in response to the strong activator Mn2+, and anyincrease in binding in response to the weaker activatorAIF4

– was not detectable. Explanations for the latterinclude the possibility that the affinity of § 4 g 7 forMAdCAM-1 is not altered by cellular activation withAIF4

–, or the affinity is increased but not sufficient todemonstrate binding to soluble MAdCAM-1. Neverthe-less, we could demonstrate increased cell binding ofsmall microspheres coated with MAdCAM-1-Fc follow-ing AIF4

– and Mn2+ treatment. Binding of ligand-coatedmicrospheres to AIF4

–-treated cells is presumably possi-ble because of stronger multivalent interactions, whichmight be enhanced by receptor clustering. The latternotion was supported by the finding that both AIF4

– andMn2+ potentiated ligand-induced clustering of § 4 g 7.Hence ligand-induced clustering appears to be inducedby putative changes in the extracellular conformation of§ 4 g 7 in response to Mn2+, and by changes at the cyto-

plasmic interface in response to activation of G-proteins.The degree of clustering appeared to parallel the abilityto each agent to induce cell adhesion, suggesting thatreceptor clustering may be an important component ofboth “inside-out” and “outside-in” pathways leading tocell adhesion.

In summary, the results provide evidence that Rac1 andRhoA can activate cell adhesive functions mediatedby § 4 g 7, and that the effects of Rac1 are exertedthrough RhoA. G-proteins appear to work, at least inpart, by potentiating ligand-induced clustering of inte-grins. Despite intensive effort we have been unable toidentify a direct association of RhoA and Rac1 with theintegrin g 7 subunit cytoplasmic domain, however tetras-pans such as CD9 which associate with lymphocyteintegrins have been proposed to associate with smallGTP-binding proteins [35]. Tetraspans as “molecularfacilitators” may sequester small GTP-binding proteins,


thereby assembling components of integrin signalingpathways. Further insights into integrin activation maybe gained by examining the ability of Rho targets such asprotein kinase N, myosin phosphatase, p160 Rho-associated coiled-coil-containing protein kinase (ROCK),and phosphatidylinositol-4-phosphate 5-kinase (PIP5-K)to independently activate integrin-mediated adhesive-ness. Rho kinase has already been shown to enhancethe formation of actin stress fibers and focal adhesions[36], and mediate tumor cell invasion [37]. The Na-Hexchanger NHE1 acts downstream of rhoA to regulateintegrin-mediated cell adhesion [38]. Future studies willneed to address the actual mechanism(s) by which RhoAand its effectors facilitate ligand-induced clustering of§ 4 g 7, leading to cell adhesion, and the involvement of

the growing number of integrin-associated cytoplasmicproteins such as Rack1 and WAIT-1 [39, 40] that couldpotentially act as RhoA effectors.

4 Materials and methods

4.1 Materials

Recombinant V14RhoA and V12Rac1 cDNA in the pGEX-2Tvector were kindly donated by Prof. Alan Hall, UniversityCollege London, GB [ 21]. They were expressed in Escheri-chia coli, and the GST proteins purified on glutathione-Sepharose. C3 cDNA in the pGEX-2T vector was generouslyprovided by Dr. Simon Dillon, Tufts University, Boston, MA.It was cleaved from glutathione-Sepharose beads asdescribed [41]. Mouse MAdCAM-1-Fc was prepared in abaculovirus system [6]. MAdCAM-1-Fc-coated micro-spheres were prepared by mixing 1 ? g of recombinantMAdCAM-1-Fc with 1 ? l of a 1 % suspension of Power-bindprotein A microparticles (Seradyn, 0.979 mM diameter) in100 ? l for 30 min at 4 °C.

4.2 Cell culture

The mouse spontaneous AKR/Cum Y CD8 LPAM-1+ VLA-4–

T lymphoma cell line TK-1 [1] was cultured at 37 °C in a 5 %CO2 incubator in RPMI 1640 medium containing 50 U/mlpenicillin, 50 ? g/ml streptomycin, 10 % FCS and 0.05 mM 2-ME.

4.3 Confocal microscopy analysis of receptorclustering

TK-1 cells suspended in PBS/0.02 % sodium azide werewashed with HBSS containing 10 mM HEPES, and resus-pended in the same solution containing either AIF4

–, 2 mMMn2+, or 2 mM Mg2+. To examine whether MAdCAM-1 wouldinduce the clustering of § 4 g 7, cells were incubated as is, orwith either dimeric MAdCAM-1-Fc, or MAdCAM-1-coatedmicrospheres for 1 h on ice. Cells were washed with PBS

containing 0.02 % sodium azide, and fixed in 4 % (w/v) para-formaldehyde in PBS for an additional 1 h. They were thenstained with FITC-conjugated M293 mAb, washed and fixedas above. All steps were carried out at 4 °C in the presenceof 0.02 % sodium azide to prevent receptor internalization.Cells were mixed 1:1 with Citifluor glycerol/PBS solution,mounted on slides, and analyzed using a Leica TCS4D con-focal laser scanning microscope equipped with an exter-nal argon-krypton laser (488 nm). Images were digitallyrecorded and printed on an Epson colour printer usingMicrosoft Power Point software.

4.4 Flow cytometry measurement of ligand binding

TK-1 cells were suspended in HBSS containing 5 mM EDTA,washed, and resuspended in HBSS containing either 2 mMMg2+, 2 mM Mg2+ and AIF4

–, or 2 mM Mn2+. After 5 min, solu-ble MAdCAM-1-Fc and MAdCAM-1-Fc-coated micro-spheres were added and allowed to bind for 30 min at roomtemperature. Cells were fixed with 4 % (w/v) paraformalde-hyde, and stained with anti-human Fc (HP6001 mAb, 1:320dilution of ascites), followed by a FITC-conjugated rat anti-mouse secondary mAb. MAdCAM-1 binding was analyzedusing a FACScan flow cytometer.

4.5 Electroporation of cells

TK-1 cells were cultured until 60 to 80 % confluent, incu-bated with 20 ? g GST fusion protein on ice for 10 min, andelectroporated at 300 V, 1000 ? F and n Y . They wereallowed to stand at room temperature for 10 min, resus-pended in media, and after a 2–4-h incubation at 37 °C wereadded to MAdCAM-1-Fc-coated slides.

4.6 C3 treatment of cells and proteins

To measure ADP-ribosylation of recombinant proteins, puri-fied GST-V14RhoA and GST-V12Rac1 were ADP-ribosylatedby incubating 2 ? g of each GST fusion protein for 60 min at37 °C with 0–3 ? g/ml C3 in 90 mM Tris-HCl (pH 8.0) contain-ing 2.6 mM MgCl2, 10 mM thymidine, 1 mM DTT, 1 mM EDTA,1 mM ATP and 1 ? Ci/ml [32P]NAD. To inactivate recombinantRhoA, GST-V14RhoA-glutathione-Sepharose beads wereresuspended in buffer containing 10 ? M cold NAD+ and0.5 ? g/ml C3, and incubated at 30 °C for 1 h. ADP-ribosylatedRhoA was recovered by elution from beads with 100 mM Tris-HCl, pH 8.0, 120 mM NaCl, containing 200 mM glutathione.To inactivate endogenous RhoA, cells were treated overnightin media containing 0–50 ? g/ml C3 as indicated.

4.7 Cell adhesion assays

Cell adhesion assays were performed as described previ-ously [6]. Fixed cells were counted by light microscopy, and/or cell binding was measured by staining slides with 0.1 %


methylene blue, and recording the absorbance at 595 nm.Lymphocytes were activated with Mn2+ as described previ-ously [6], and by incubating with AIF4

– (10 mM NaF and40 ? M AlCl3) [18]. The results are presented as the meannumber of cells bound (± SEM) in six independent fields.Background binding to FCS was subtracted unless indi-cated otherwise.

Acknowledgements: We thank Dr. Dillon, Tufts University,for providing a construct encoding C3 exoenzyme. Thiswork was supported in part by grants from the WellcomeTrust (UK), the Royal Society of New Zealand, the Leukae-mia and Blood Foundation, the Auckland Medical ResearchFoundation, the Cancer Society of New Zealand, the HealthResearch Council of New Zealand, the Multiple SclerosisSociety of New Zealand, the Marsden Fund, and The LotteryGrants Board. G. W. K. is a James Cook Research Fellowfunded by the Royal Society of New Zealand. The first threeauthors contributed equally to the results.

5 References

1 Holzmann, B. and Weissmann, I. L., Peyer’s patch-specific lymphocyte homing receptors consist of a VLA-4-like § chain associated with either of two integrin gchains, one of which is novel. EMBO J. 1989. 8:1735–1741.

2 Yuan, Q., Jiang, W.-M., Leung, E., Hollander, D., Wat-son, J. D. and Krissansen, G. W., Molecular cloning ofthe mouse integrin g 7 subunit. J. Biol. Chem. 1992. 267:7352–7358.

3 Parker, C. M., Cepek, K. L., Russell, G. J., Shaw, S. K.,Posnett, D. N., Schwarting, R. and Brenner, M. B., Afamily of g 7 integrins on human mucosal lymphocytes.Proc. Natl. Acad. Sci. USA 1992. 89: 1924–1928.

4 Krissansen, G. W., Print, C. G., Prestidge, R. L., Hol-lander, D., Yuan, Q., Jiang, W.-M., Jenkins, D. R.,Leung, E., Mead, P., Yong, R., Ameratunga, R. V.,Cerf-Bensussan, N. and Watson, J. D., Immunologicand structural relatedness of the integrin g 7 complexand the human intraepithelial lymphocyte antigenHML-1. FEBS Lett. 1992. 296: 25–28.

5 Berlin, C., Berg, E. L., Briskin, M. J., Andrew, D. P., Kil-shaw, P. J., Holzmann, B., Weissman, I. L., Hamann,A. and Butcher, E. C., § 4 g 7 integrin mediates lympho-cyte binding to the mucosal vascular addressinMAdCAM-1. Cell 1993. 74: 185–195.

6 Yang, Y., Sammar, M., Harrison, J. E. B., Lehnert, K.,Print, C. G., Leung, E., Prestidge, R. and Krissansen,G. W., Construction and adhesive properties of a solubleMAdCAM-1-Fc chimera expressed in a baculovirus sys-tem: phylogenetic conservation of receptor-ligand inter-action. Scand. J. Immunol. 1995. 42: 235–247.

7 Cepek, K. L., Shaw, S. K., Parker, C. M., Russell, G. J.,Morrow, J. S., Rimm, D. L. and Brenner, M. B., Adhe-sion between epithelial cells and T lymphocytes medi-ated by E-cadherin and the § 4 g 7 integrin. Nature 1994.372: 190–193.

8 Karecla, P. I., Bowden, S. J., Green, S. J. and Kilshaw,P. J., Recognition of E-cadherin on epithelial cells by themucosal T cell integrin § M290 g 7 ( § 4 g 7). Eur. J. Immunol.1995. 25: 852–856.

9 Wagner, N., Lohler, J., Kunkel, E. J., Ley, K., Leung, E.,Krissansen, G., Rajewsky, K. and Mueller, W., Criticalrole for g 7 integrins in the formation of the gut associ-ated lymphoid tissue. Nature 1996. 382: 366–370.

10 Yauch, R. L., Felsenfeld, D. P., Kraeft, S.-K., Chen,L. B., Sheetz, M. P. and Hemler, M. E., Mutationalevidence for control of cell adhesion through integrindiffusion/clustering, independent of ligand binding.J. Exp. Med. 1997. 186: 1347–1355.

11 Jakubowski, A., Rosa, M. D., Bixler, S., Lobb, R. andBurkly, L. C., Vascular cell adhesion molecule (VCAM)-Ig fusion protein defines distinct affinity states of theVery Late Antigen-4 (VLA-4) receptor. Cell Adhes. Com-mun. 1995. 3: 131–142.

12 Zhang, Z., Vuori, K., Wang, H.-G., Reed, J. C. andRuoslahti, E., Integrin activation by R-ras. Cell 1996. 85:61–69.

13 D’Souza-Schorey, C., Boettner, B. and van Aelst, L.,Rac regulates integrin-mediated spreading and in-creased adhesion of T lymphocytes. Mol. Cell. Biol.1998. 18: 3936–3946.

14 Pachynski, R. K., Wu, S. W., Gunn, M. D. and Erle,D. J., Secondary lymphoid-tissue chemokine (SLC)stimulates integrin alpha 4 beta 7-mediated adhesionof lymphocytes to mucosal addressin cell adhesionmolecule-1 (MAdCAM-1) under flow. J. Immunol. 1998.161: 952–956.

15 Spangrude, G. J., Sacchi, F., Hill, H. R., Van Epps,D. E. and Daynes, R. A., Inhibition of lymphocyte andneutrophil chemotaxis by pertussis toxin. J. Immunol.1985. 135: 4135–4143.

16 Spangrude, G. J., Braaten, B. A. and Daynes, R. A.,Molecular mechanisms of lymphocyte extravasation.I. Studies of two selective inhibitors of lymphocyte recir-culation. J. Immunol. 1984. 132: 354–362.

17 Coleman, D. E., Berghuis, A. M., Lee, E., Linder, M. E.,Gilman, A. G. and Sprang, S. R., Structures of activeconformations of Gi § 1 and the mechanism of GTPhydrolysis. Science 1994. 265: 1405–1412.

18 Driessens, M. H. E., Van Hulten, P. E. M., Van Rijtho-ven, E. A. M., Soede, R. D. M. and Roos, E., Activationof G-proteins with AIF4– induces LFA-1-mediated adhe-sion of T-cell hybridoma cells to ICAM-1 by signal


pathways that differ from phorbol ester- andmanganese-induced adhesion. Exp. Cell Res. 1997. 231:242–250.

19 Laudanna, C., Campbell, J. J. and Butcher, E. C., Roleof rho in chemoattractant-activated leukocyte adhesionthrough integrins. Science 1996. 271: 981–983.

20 Tominaga, T., Sugie, K., Hirata, M., Morii, N., Fukata,J., Uchida, A., Imura, H. and Narumiya, S., Inhibition ofPMA-induced, LFA-1-dependent lymphocyte aggrega-tion by ADP ribosylation of the small molecular weightGTP binding protein, rho. J. Cell Biol. 1993. 120:1529–1537.

21 Nobes, C. D. and Hall, A., Rho, Rac, and Cdc42GTPases regulate the assembly of multimolecular focalcomplexes associated with actin stress fibers, lamellipo-dia, and filopodia. Cell 1995. 81: 53–62.

22 Bifulco, M., Laezza, C., Aloi, S. M. and Garbi, C.,Mevalonate controls cytoskeleton organization and cellmorphology in thyroid epithelial cells. J. Cell Physiol.1992. 155: 340–348.

23 Ridley, A. J. and Hall, A., The small GTP-binding proteinrho regulates the assembly of focal adhesions and actinstress fibers in response to growth factors. Cell 1992.70: 389–399.

24 Ridley, A. J. and Hall, A., Signal transduction pathwaysregulating Rho-mediated stress fibre formation: require-ment for a tyrosine kinase. EMBO J. 1994. 13:2600–2610.

25 Ward, S. G., June, C. H. and Olive, D., PI 3-kinase: apivotal pathway in T-cell activation? Immunol. Today1996. 17: 187–196.

26 Kahn, R. A., Fluoride is not an activator of the smaller(20–25 kDa) GTP-binding proteins. J. Biol. Chem. 1991.266: 15595–15597.

27 Chakrabati, R., Wylie, D. E. and Schuster, S. M.,Transfer of monoclonal antibodies into mammaliancells by electroporation. J. Biol. Chem. 1989. 264:15494–15500.

28 Yang, Y., Cardarelli, P. M., Lehnert, K., Rowland, S.and Krissansen, G. W., LPAM-1 (integrin § 4 g 7)-ligandbinding: overlapping binding sites recognizing VCAM-1,MAdCAM-1 and CS-1 are blocked by fibrinogen, afibronectin-like polymer and RGD-like cyclic peptides.Eur. J. Immunol. 1998. 28: 995–1004.

29 Turner, L., Ward, S. G. and Westwick, J., RANTES-activated human T lymphocytes. A role for phosphoino-sitide 3-kinase. J. Immunol. 1995. 155: 2437–2444.

30 Nobes, C., Hawkins, P., Stephens, L. and Hall, A.,Activation of the small GTP-binding proteins Rho andRac by growth factor receptors. J. Cell Sci. 1995. 108:225–233.

31 Wennstrom, S., Hawkins, P., Cooke, F., Hara, K.,Yonezawa, K., Kasuga, M., Jackson, T., Claesson-Welsh, L. and Stephens, L., Activation of phosphoino-sitide 3-kinase is required for PDGF-stimulated mem-brane ruffling. Curr. Biol. 1994. 4: 385–393.

32 Arrieumerlou, C., Donnadieu, E., Brennan, P., Keryer,G., Bismuth, G., Cantrell, D. and Trautmann, A.,Involvement of phosphoinositide 3-kinase and rac inmembrane ruffling induced by IL-2 in T cells. Eur. J.Immunol. 1998. 28: 1877–1885.

33 Parise, L. V., Cdc42 and Rac1 induce integrin-mediatedcell motility and invasiveness through PI(3)K. Nature1997. 390: 632–636.

34 Michiels, F., Habets, G. G. M., Stam, J. C., van derKammen, R. A. and Collard, J. G., A role for Rac inTiam1-induced membrane ruffling and invasion. Nature1995. 375: 338–340.

35 Seehafer, J. G. and Shaw, A. R., Evidence that thesignal-initiating membrane protein CD9 is associatedwith small GTP-binding proteins. Biochem. Biophys.Res. Commun. 179: 401–406.

36 Amano, M., Chihara, K., Kimura, K., Fukata, Y., Naka-mura, N., Matsuura, Y. and Kaibuchi, K., Formation ofactin stress fibers and focal adhesions enhanced byRho-kinase. Science 1997. 275: 1308–1311.

37 Itoh, K., Yoshioka, K., Akedo, H., Uehata, M., Ishizaki,T. and Narumiya, S., An essential part for rho-associated kinase in the transcellular invasion of tumorcells. Nature Med. 1999. 5: 221–225.

38 Tominaga, T. and Barber, D. L., Na-H exchange actsdownstream of RhoA to regulate integrin-induced celladhesion and spreading. Mol. Biol. Cell 1998. 9:2287–2303.

39 Lilienthal, J. and Chang, D. D., Rack1, a receptor foractivated protein kinase C, interacts with integrin g sub-unit. J. Biol. Chem. 1998. 273: 2379–2383.

40 Rietzler, M., Bittner, M., Kolanus, W., Schuster, A. andHolzmann, B., The human WD repeat protein WAIT-1specifically interacts with the cytoplasmic tails of beta7-integrins. J. Biol. Chem. 1998. 273: 27459–27466.

41 Dillon, S. T. and Feig, L. A., Purification and assay ofrecombinant C3 transferase. Methods Enzymol. 1995.256: 174–184.

Correspondence: Geoffrey W. Krissansen, Departmentof Molecular Medicine, School of Medicine and HealthScience, University of Auckland, Auckland, New ZealandFax: +64-93737674e-mail: gw.krissansen — auckland.ac.nz


the small gtp-binding proteins rho and rac induce t cell adhesion to the mucosal addressin madcam-1...

Documents