Calpain cleaves RhoA generating a dominant-negative form that inhibits integrin-induced actin filament assembly and cell spreading.

Integrin-induced cell adhesion results in transmission of signals that induce cytoskeletal reorganizations and resulting changes in cell behavior. The cytoskeletal reorganizations are regulated by transient activation and inactivation of Rho GTPases. Previously, we identified mu-calpain as an enzyme that is activated by signaling across beta1 and beta3 integrins. We showed that it mediates cytoskeletal reorganizations in bovine aortic endothelial (BAE) and Chinese hamster ovary (CHO) cells and does so by acting upstream of Rac1 activation. Here we show that mu-calpain is also involved in inactivating RhoA during integrin-induced signaling. Cleavage of RhoA was detectable in BAE cells plated on an integrin substrate; it did not occur in cells plated on poly-l-lysine. Cleavage was inhibited by calpain inhibitors. In vitro, mu-calpain cleaved RhoA generating a fragment of the same size as in intact cells. The cleavage site was identified, an HA-tagged construct expressing calpain-cleaved RhoA generated, and the construct expressed in BAE and CHO cells. Calpain-cleaved RhoA inhibited integrin-induced stress fiber assembly and decreased cell spreading. Together, our data show that calpain cleaves RhoA and generates a form that inhibits integrin-induced stress fiber assembly and cell spreading.

Signaling through multiple transmembrane receptors such as receptor tyrosine kinases, chemoattractant receptors, and integrins can transiently activate members of the Rho family of GTP-binding proteins (1)(2)(3)(4)(5)(6). These GTPases have been implicated in the induction of cytoskeletal reorganizations during spreading, migration, proliferation, differentiation, and cytokinesis (7,8). Activation of cdc42 leads to filopodia formation; activation of Rac1 induces the formation of submembranous networks of actin filaments required for lamellipodia extension and motility; activation of RhoA leads to the formation of stress fibers and focal adhesions that are involved in contractility and in maintaining a spread, adhesive phenotype (9,10). The coordinate regulation of these cellular responses is presumed to be brought about by selective activation of one GTPase over the other (11). For example, growth factors that activate cdc42 and Rac1 typically promote cell migration and at the same time inhibit RhoA-induced contractility and stress fiber formation (1,11,12). Signals transmitted across integrins as they bind to their ligands cause the integrins to cluster and assemble complexes of signaling molecules that lead to Rac1 activation and the extension of lamellipodia. During this early stage of integrin-induced cell spreading, the activity of RhoA is down-regulated (5,6). At a later time point during integrin-induced spreading, the activity of RhoA is increased, and focal adhesions and contractile actin-myosin stress fibers are formed (6).
The inverse relationship between activation of Rac1 and RhoA suggests the presence of mechanisms by which Rho GT-Pases are rapidly activated and probably equally rapidly inactivated in a dynamic and highly regulated manner during transmembrane receptor signaling. Although there is now considerable information available concerning the mechanisms involved in activation of Rho family members, little is known about the mechanisms involved in the dynamic inhibition of these proteins. We have been interested in signaling mechanisms involved in activating Rac1 and inactivating RhoA at the early times after integrin-induced spreading.
Several years ago, we identified -calpain as one of the signaling molecules that was activated following signaling across ␤ 3 -containing integrins in platelets (13). Calpains are ubiquitous, intracellular proteases that are active at physiological pH (14). They are heterodimers containing a common regulatory subunit of 30 kDa and an 80-kDa catalytic subunit (15)(16)(17). The two primary catalytic subunits are the one present in m-calpain, which requires millimolar levels of calcium for half-maximal activation in vitro, and the one present in -calpain, which requires micromolar levels of calcium (18). More recently, we showed that -calpain was also activated following signaling across ␤ 1 -and ␤ 3 -containing integrins in cultured cells. We have shown that calpain is required for the formation of a specific type of integrin cluster that forms early after signaling and prior to detectable cell spreading (19). These integrin clusters form upstream of Rac1 activation and appear to be the site of Rac activation (19).
The finding that Rac1 activation occurs in the initial clusters of integrins that form following integrin-ligand interactions (19) suggests that signaling molecules in these clusters may be involved in the activation and inactivation of Rho family members at these early stages of spreading. Because -calpain is present in these clusters in an active form (19) and several of the proteins in the clusters are cleaved by this protease (20, 21), we have considered the possibility that calpain may be involved * This work was supported by research grants HL30657 and HL56264 (to J. E. B. F.) from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ʈ in regulating the activity of Rho family members. In the present study, we could not detect cleavage of Rac1 or cdc42. However, we found that RhoA is cleaved in cultured cells spreading on an integrin substrate. Cleavage was more pronounced under conditions in which Rac1 activation predominates and was not detected under conditions that favored RhoA activation. Cleavage was induced by calpain, as shown by the fact that 1) both calpain activation and RhoA cleavage were absolutely dependent on integrin engagement; 2) generation of RhoA cleavage product was inhibited by known calpain inhibitors but not by a caspase inhibitor or proteasome inhibitor; and 3) RhoA was cleaved by -calpain in vitro, generating fragments of the same size as those generated in intact cells. Further, we identified the site at which -calpain cleaves RhoA, expressed calpaincleaved RhoA in bovine aortic endothelial (BAE) 1 cells, and demonstrated that the calpain-cleaved RhoA inhibited the integrin-induced formation of stress fibers and cell spreading. Our previous results show that following integrin signaling calpain activates pathways leading to Rac1 activation (13). The present results suggest that calpain simultaneously acts directly on RhoA, generating a dominant-negative form of the molecule. They are consistent with a model in which cleavage of RhoA by -calpain provides a mechanism for inhibiting RhoA activity and the subsequent formation of stress fibers and focal adhesions under conditions in which a more motile Rac1-induced phenotype is required.

EXPERIMENTAL PROCEDURES
Reagents-Lysophosphatidic acid (LPA), platelet-derived growth factor (PDGF), and poly-L-lysine were from Sigma. Monoclonal antibodies against the HA-epitope were from Roche Molecular Biochemicals. The antibodies against RhoA (monoclonal and polyclonal), Rac1 (polyclonal), and cdc42 (polyclonal) were from Santa Cruz Biotechnology, Inc. Recombinant Rho GTPases were obtained from two sources. For some experiments His-tagged human RhoA, Rac1, or cdc42 were purchased from Cytoskeleton, Inc. For other experiments, we expressed human RhoA in Escherichia coli using plasmid pGEX2T encoding RhoA as a GST fusion protein (kindly provided by Dr. Sadashiva Karnik, Cleveland Clinic Foundation). -Calpain was purified from bovine skeletal muscle as described previously (22,23), and -calpain from human erythrocytes was purchased from Calbiochem. The membrane-permeable inhibitor of calpain, calpeptin (Z-Leu-Nle-H), was from Novabiochem, human recombinant calpastatin was from Calbiochem, and calpain inhibitor I was from Roche. Proteasome inhibitor II and caspase inhibitor I were from Calbiochem.
In Vitro Cleavage of Rho GTPases-Recombinant GTPases were incubated with purified -calpain at room temperature in Tris-HCl (50 mM, pH 7.5), KCl (100 mM), EDTA (1 mM), and dithiothreitol (5 mM). Reactions were initiated by the addition of 2 mM CaCl 2 . At the indicated times, reactions were terminated by the addition of SDS sample buffer. Samples were analyzed on Western blots. In the experiment in which we determined the site of calpain cleavage, RhoA was expressed in E. coli as a GST fusion protein, purified on GST-agarose beads, and digested with -calpain in a 5-min reaction. Products were electrophoresed through a polyacrylamide gel, transferred to polyvinylidene difluoride membrane, and stained with Coomassie Brilliant Blue, and intact RhoA and the two calpain-induced fragments were excised as three individual bands. The N-terminal sequence of each band was deter-mined in the Cleveland Clinic Foundation Protein Core Facility using previously described methods (24,25).
Detection of RhoA Cleavage in Cultured Cells-To minimize the contribution from the effects of growth factors, cells were withdrawn from serum-containing media 24 h prior to experiments. Cells were plated on fibronectin-coated dishes (35 100 mm, Becton Dickinson, San Jose, CA) in the presence or absence of reagents for the indicated time. Cells were solubilized in cold radioimmune precipitation buffer containing Tris-HCl, pH 7.5 (20 mM), NaCl (150 mM), sodium deoxycholate (0.1%), Triton X-100 (1%), SDS (O.1%), and the inhibitors phenylmethylsulfonyl fluoride (1 mM), aprotinin (10 g/ml), leupeptin (10 g/ml), EDTA (2 mM), sodium fluoride (50 mM), and Na 3 VO4 (1 mM). Total protein was estimated using bovine serum albumin as the standard protein in a colorimetric assay (Bio-Rad microassay kit). In other experiments cells were transfected with the plasmids encoding HA-tagged proteins and allowed to recover for 24 h prior to extract preparation. Proteins were denatured by the addition of SDS sample buffer and electrophoresed through SDS gels containing 3.5% polyacrylamide in a stacking gel and 12.5% polyacrylamide in a resolving gel. Western blotting was performed by the method of Towbin et al. (26).
Generation of Constructs Encoding the 20-kDa Calpain-cleaved RhoA Fragment-The plasmid expressing the calpain-cleaved fragment of RhoA was generated by polymerase chain reaction (PCR) using plasmid pGEX2T encoding wild-type human RhoA as a template. The forward primer was 5Ј-TATTTAAGCTTATGTATGATGTTCCTGATTATGCAA-GTTTAGCTGCCATCCGGAAG-3Ј encoding an HA epitope (YDVPDYA-SL) from hemagglutinin virus, and the reverse primer was 5Ј-TTTAT-GGATCCTCATTGCAGAGCAGCTCT-3Ј. The resulting PCR product encoding HA-tagged RhoA truncated at the calpain cleavage site (lacking the last 39 nucleotides corresponding to 13 amino acids) was subcloned into pcDNA3 for expression into cultured cells. The primers were also used with Val-14 RhoA and T19N RhoA (both plasmids were kindly provided by Dr. M. A. Schwartz, The Scripps Research Institute) to generate expression vectors for the HA-tagged constitutively active RhoA truncated at the calpain cleavage site or HA-tagged dominantnegative RhoA truncated at the calpain cleavage site. The sequences were verified by nucleotide sequencing for all plasmids.
Transfections and Immunofluorescence-Transient transfections with the plasmids encoding HA-tagged RhoA mutants truncated at the calpain cleavage site, wild-type RhoA (provided by Dr. M. A. Schwartz), the active Val-14 RhoA, or the inactive T19N RhoA were carried out in BAE or CHO cells. Cells were cultured in 35-60-mm dishes and transfected with 4 g of DNA using LipofectAMINE Plus reagent (Invitrogen) as described earlier. Transfected cells were allowed to recover for 24 h, trypsinized, and plated on fibronectin-coated coverslips. After 2 h, the cells were fixed with 1.4% formaldehyde for 10 min, permeabilized with 0.5% Triton X-100 for 5 min, and blocked with 4% horse serum. Cells were incubated with anti-HA antibody (0.5 g/ml) to detect transfected cells and TRITC-labeled phalloidin (Sigma, 0.5 g/ml) to detect actin filaments. Images were collected using a Zeiss immunofluorescence microscope or Leica confocal microscope.

Experiments to Determine whether Rho Proteins Are Cleaved by Calpain in Cultured
Cells-Previously, we have shown that calpain is activated as a consequence of integrin-induced signaling in spreading cells. Thus, as a first step in determining whether Rho proteins were cleaved by calpain following integrin-ligand interactions, we determined whether cleaved forms of Rho proteins were generated as a consequence of integrininduced signaling. Using Western blot analysis, we were unable to detect cleaved forms of Rac1 or cdc42 in BAE cells spreading on an integrin substrate (data not shown). However, as shown in Fig. 1, a 20-kDa fragment that was not detectable To investigate the possibility that the RhoA fragment was generated by calpain, the calpain inhibitors calpeptin and calpain inhibitor I were included in the incubations prior to plating the BAE cells on an integrin substrate. As shown in Fig Experiments to Determine whether -Calpain Cleaves Rho Family Members in Vitro-The finding that the RhoA fragment was generated under conditions in which -calpain is known to be activated and was inhibited under conditions known to inhibit calpain suggested that it might be generated as a consequence of cleavage by calpain. To determine whether Rho family members were substrates of calpain, recombinant human His-tagged cdc42, Rac1, and RhoA were incubated with -calpain purified from human erythrocytes (Fig. 3). Reactions were terminated by the addition of SDS sample buffer, and samples were electrophoresed through SDS polyacrylamide gels and transferred to nitrocellulose. Rac1 was detected with polyclonal antibodies (Fig. 3A), cdc42 was detected with polyclonal antibodies (Fig. 3B), and RhoA was detected with the same monoclonal antibody that was used in Figs. 1 and 2 (Fig.  3C). Cleavage of Rac1 or cdc42 by -calpain was not detected. However, -calpain cleaved RhoA, generating a fragment of a molecular mass of 20 kDa that co-migrated with the 20-kDa fragment detected in cultured cells. To determine whether we could detect additional RhoA fragments generated by calpain, the Western blots of calpain-cleaved RhoA were probed with polyclonal antibodies. As shown in Fig. 3D (lane 2), the polyclonal antibodies detected the 20-kDa fragment identified with the monoclonal antibody in intact cells (Figs. 1 and 2) and in vitro incubations (Fig. 3C), but it also detected a 14-kDa fragment. As in cultured cells spreading on an integrin substrate ( Fig. 2A), calpeptin and calpastatin inhibited the -calpaininduced cleavage of recombinant Rho A (Fig. 3D, lanes 3 and 4). Similar results were obtained with RhoA that was expressed as a GST fusion protein in E. coli, purified, and incubated with either -calpain from human erythrocytes or with -calpain purified from bovine skeletal muscle (data not shown).
Identification of the Site of Cleavage in RhoA and Its Consequence on Integrin-induced Cell Spreading-To gain insight into the functional consequence of RhoA cleavage, we identified the site of cleavage in recombinant RhoA and transfected cells with cDNA encoding RhoA that was terminated at this site. First, to identify the site of cleavage, recombinant RhoA was expressed in E. coli, purified, and incubated with purified -calpain for 5 min. The reaction products were separated on a polyacrylamide gel, transferred to a polyvinylidene difluoride membrane, and stained with Coomassie Brilliant Blue. The band corresponding to intact RhoA and the fragments of 20 and 14 kDa were excised and subjected to protein microsequencing. The N-terminal sequence of all three bands was determined. The intact RhoA protein and the 20-kDa fragment had the same N-terminal amino acids. The fragment that migrated with an apparent molecular mass of 14 kDa contained the last 13 amino acid residues of RhoA. Thus, -calpain cleaved RhoA between Gln-180 and Ala-181 (Fig. 4), generating RhoA lacking the C-terminal 13 amino acids and a fragment containing only the last 13 amino acids. The reason why a peptide containing only 13 amino acids migrates with an apparent molecular mass of 14 kDa is not known, but it suggests some form of oligomerization that persists in SDS gels and is consistent with the known presence of structural determinants for dimerization in the C-terminal polybasic region of Rho GTPases (27).
To determine the consequence of the calpain-induced cleavage of RhoA, we generated a plasmid encoding HA-tagged RhoA truncated at the -calpain cleavage site (Ala-181) and transfected it into CHO cells. Cells were transfected with HAtagged truncated RhoA or HA-tagged full-length RhoA, and extracts were analyzed on Western blots. Fig. 5 (left panel) shows that the truncated protein was expressed and that it migrated at an appropriate position on the gels, below the HA-tagged full-length protein. The right panel of Fig. 5 shows the blot stripped of HA antibody and reprobed with RhoA antibodies to detect endogenous RhoA, providing an indication of the relative amounts of expression of endogenous and HAtagged RhoA proteins.
To investigate the functional significance of calpain-induced cleavage of RhoA, BAE cells were transfected with the HAtagged truncated RhoA, trypsinized, and plated on fibronectin for 2 h (Fig. 6). The cells expressing the truncated RhoA were identified by staining with anti-HA antibody, and the organization of the cytoskeleton was analyzed by staining stress fibers with TRITC-phalloidin. Cells that did not express the truncated RhoA construct spread and contained numerous stress fibers, whereas the cells that expressed truncated RhoA contained few stress fibers and showed decreased spreading (Fig. 6).
To ensure that the inhibitory effects of the truncated Rho construct was not a nonspecific transfection effect, cells were also transfected with HA-tagged wild-type RhoA or HA-tagged constitutively active RhoA. Cells expressing these constructs spread on fibronectin and formed stress fibers (Fig. 7). In contrast, cells expressing HA-tagged inactive RhoA (T19N RhoA) showed inhibition of spreading and stress fiber assembly comparable with that observed with the construct encoding calpain-cleaved RhoA (Fig. 7, bottom panels).
Similar results were obtained when CHO cells were transfected with the RhoA constructs. Thus, the construct encoding calpain-cleaved RhoA inhibited cell spreading and stress fiber formation (Fig. 8); HA-tagged full-length or constitutively active RhoA had no inhibitory effects on cell spreading or stress fiber formation (Fig. 9); and HA-tagged inactive RhoA inhibited spreading and stress fiber formation (Fig. 9, bottom panels). truncated at the calpain-cleavage site was expressed (data not shown).

Experiments to Determine whether Generation of the 20-kDa RhoA Fragment in Spreading Cells Is Increased under Conditions in Which Rac1
Rather than RhoA Is Expected To Be Active-Rac1 and RhoA induce very different phenotypes in cells cultured on integrin substrates. Rac1 activation leads to the formation of lamellipodia and submembranous actin filaments that favor migration, whereas RhoA induces the formation of stress fibers and focal adhesions that are thought to be important in more fully spread and stationary cells. Evidence has been provided that RhoA is down-regulated in cells exhibiting a Rac1 phenotype (5, 6, 28). The observation that calpaincleaved RhoA inhibits cell spreading suggested that calpain cleavage might provide a mechanism for inhibiting RhoA ac-    tivity. We reasoned that if this was the case, the amounts of cleaved RhoA might be high in cultured cells exhibiting a Rac1 phenotype and low in those exhibiting a RhoA phenotype. One way in which the relative activation of Rac1 and RhoA can be modulated in cells spreading on an integrin substrate is by inclusion of soluble agonists in the medium. In the absence of other agonists, cells rely absolutely on signals transmitted across ligand-occupied integrin to induce activation of Rac1 or RhoA. However, if PDGF is included in the medium, Rac1 can also be activated as a consequence of signaling across the PDGF receptor (10 -12, 29). Under these conditions, cells rapidly extend lamellipodia, become filled with Rac1-induced focal complexes (detected with phosphotyrosine antibodies in Fig.  11A), and show a motile phenotype (Fig. 11A, upper panels). In contrast, if LPA is included in the medium, RhoA is activated through G protein-coupled receptors (29,30), and the cells become filled with focal adhesions (detected with phosphotyrosine antibodies in Fig. 11A) and stress fibers and show a much more stationary, spread phenotype (Fig. 11A, lower  panels).
To determine whether generation of the RhoA fragment was increased when Rac1 activation was increased and decreased when RhoA activation was increased, cells grown on an integrin substrate in the presence or absence of PDGF or LPA were solubilized and analyzed on Western blots. The polyclonal antibodies were used to detect calpain-induced cleavage. As shown in Fig. 11B, the integrin-induced generation of the two RhoA fragments was increased in cells exposed to PDGF (Fig.   11B, compare lanes 1 and 2). Generation of the RhoA fragment was dependent on integrin signaling because the fragment was not present in cells spread on poly-L-lysine (lane 3). Because calpain is activated following integrin-ligand engagement but is not activated in cells spreading on poly-L-lysine, this observation is consistent with the fragment being generated as a consequence of integrin-induced activation of calpain. In contrast to cells spreading in the presence of PDGF, those spreading in the presence of LPA, which activates RhoA, did not show the presence of the RhoA fragment (Fig. 11B, lane 4). These results indicate that the generation of the RhoA fragment occurs under conditions favoring Rac1 activation and is downregulated under conditions favoring RhoA activation, suggesting that generation of the RhoA fragment might provide a mechanism of inhibiting RhoA function. DISCUSSION Rho GTPases are involved in the regulation of dynamic cytoskeletal reorganizations that occur following signaling across a variety of transmembrane receptors. Cdc42 is involved in inducing the formation of filopodia (31), Rac1 in the formation of the networks of submembranous actin filaments (10) that are characteristic of a motile cell, and RhoA in the formation of stress fibers and focal adhesions that stabilize adhesion in a more fully spread, stationary cell (6,9). Because of the different functions of the Rho GTPases, mechanisms must presumably exist for the selective activation or inactivation of the proteins so that only Rho GTPases involved in inducing the required cytoskeletal reorganization are active under any given situation. Although there is considerable information concerning mechanisms involved in activation of Rho proteins (7,29), little is known about mechanisms involved in their inhibition. One situation in which differential functional activity of Rho GTPases has been reported is in cultured cells spreading on an integrin substrate. Previously, we have identified the calciumdependent protease, -calpain, as being one of the signaling molecules activated during signaling across integrins (13,19). In the present study, we provide evidence that calpain cleaves RhoA in spreading cells and is involved in inhibiting the activities of RhoA under conditions favoring Rac1 activation.
Several lines of evidence indicate that RhoA is cleaved by calpain in spreading cells. First, purified -calpain cleaved recombinant RhoA, generating two fragments. Second, two RhoA fragments of comparable molecular weight were generated in cultured cells spreading on an integrin substrate. Third, the RhoA fragments were generated in cells spreading on an integrin substrate (which leads to activation of -calpain) but not in cells spreading on poly-L-lysine (which does not lead to activation of calpain). Fourth, the integrin-induced generation of RhoA fragments in cultured cells was selectively inhibited by calpain inhibitors. Taken together, these findings suggest that the RhoA fragments result from the direct action of calpain on RhoA and that the integrin-induced activation of -calpain provides a mechanism for regulating Rho GTPases in spreading cells.
When active, RhoA interacts with effectors that mediate the formation of actin filament bundles known as stress fibers (2,32). To determine whether the calpain-induced cleavage of RhoA regulates its activity, calpain-truncated RhoA (Ala-1813 stop) was expressed in cultured cells. Expression of calpaintruncated RhoA markedly inhibited cell spreading and cells expressing the truncated form assembled few stress fibers. This finding suggests that calpain cleaves RhoA into a form that does not induce its normal downstream events. The almost total inhibition of stress fiber formation suggests that cleaved RhoA can act as a dominant-negative molecule. The finding that cells expressing the truncated form of RhoA had a very similar phenotype to those expressing a known dominant-negative form of RhoA is consistent with this interpretation. Moreover, integrin-induced cleavage of RhoA was enhanced under conditions of Rac1 activation in spreading cells and was low under conditions of RhoA activation. These data are consistent with the notion that calpain cleaves RhoA during integrininduced spreading, generating a dominant-negative form of the molecule that inhibits the function of RhoA in spreading cells.
The spreading of cells on an integrin substrate is a very dynamic process with integrin complexes and actin filament organizations changing rapidly as cells extend at the leading edge and retract at the trailing edge (33). Presumably complex mechanisms must exist to coordinate the rapid switches between the Rho GTPase at specific stages of spreading and sites within the cell. One situation in which differential activation of Rac1 and RhoA has been shown to occur is at the early stages of integrin-induced spreading. At this stage of spreading, Rac1 activity is required for the formation of lamellipodia, and mechanisms appear to exist for maintaining RhoA in an inactive form until the cells have spread to a point where RhoA-induced stress fibers and focal adhesions are required (5,6,28). Recently, we showed that one of the earliest detectable events following integrin-induced adhesion is the formation of a previously unrecognized type of integrin cluster; these clusters contain active calpain, calpain-cleaved integrin, and cytoskeletal proteins and signaling molecules not detected in the previously described focal complexes and adhesions (20,21). The clusters form upstream of Rac1 activation, and we have suggested that they allow Rac1 activation because they bring together proteins involved in integrin-induced Rac1 activation (20,21). Perhaps the presence of -calpain in these clusters provides the mechanism by which RhoA is inhibited at the early stages of integrin-induced spreading.
Another situation in which RhoA activity presumably needs to be regulated is in the breakdown of focal adhesions and stress fibers at the trailing edge of migrating cells. In this situation, the functional activity of previously activated RhoA would presumably be prevented. The finding in the present study that calpain prevents the functional activity of constitutively active Val-14 RhoA suggests that calpain might serve not only to prevent RhoA from becoming functionally active but also to prevent RhoA that has already been activated from exerting its effects.
Sequencing of the RhoA cleavage products in this study revealed that the -calpain-induced cleavage of RhoA results in a truncated protein that contains most of the molecule; only 13 amino acids are removed. It is of interest to speculate about the mechanisms by which the calpain-induced removal of 13 amino acids could inhibit RhoA functions within cells. Rho GTPases are activated by exchange factors that interact with them, accelerating the GDP/GTP exchange. The GTP-bound form of the Rho proteins can bind downstream effector molecules such as protein or lipid kinases and in this way initiate downstream cascades leading to specific cytoskeletal reorganizations (7,29). The finding that expression of RhoA truncated at the calpain cleavage site results in inhibition of stress fiber formation suggests that the cleaved molecule acts as a dominant-negative protein. Thus, the cleaved form presumably retains sites that interact with critical effectors or regulators. One possibility is that the cleaved form can no longer bind exchange factors. However, the finding that truncated, constitutively active RhoA acts as a dominant-negative molecule suggests that calpain can prevent the RhoA from exerting its effects on the cytoskeleton even after it has been activated. Similarly, the RhoA molecule remaining after removal of 13 amino acids contains the N-terminal domain that binds GDP/GTP and is responsible for GTPase activity. Thus, it would be expected to retain its ability to hydrolyze GTP.
The 13 C-terminal amino acids that are removed contain a CLVL motif (34 -36) known to be required for isoprenylation (35,37) and for targeting the GTPase to the membrane (34,36,37). Therefore, it is conceivable that calpain cleaved RhoA cannot associate with the submembranous sites at which RhoA normally acts. Thus, removal of the C-terminal 13 amino acids might provide a mechanism for ensuring that even if RhoA has been activated it would be unable to exert its functional activities once it is cleaved by calpain. In such a model, cleavage of RhoA at sites at which calpain is localized might provide a mechanism for preventing RhoA from inducing cytoskeletal reorganizations at those sites. It might also provide a mechanism for arresting the functional activities at sites at which RhoA-induced cytoskeletal organization is no longer required. Future studies will be required to investigate these possibilities.