p190-B, a new member of the Rho GAP family, and Rho are induced to cluster after integrin cross-linking.

p120GAP forms distinct complexes with two phosphoproteins, p62 and p190. Here we have cloned a cDNA encoding a protein with 51% amino acid identity to p190 (hereafter designated p190-A) and have designated it p190-B. The N-terminal portion of p190-B contained several motifs characteristic of a GTPase domain, while its C terminus contained a Rho GAP domain. A recombinant Rho GAP domain polypeptide showed GAP activity for RhoA, Rac1, and G25K/CDC42Hs. Immunoprecipitation and immunofluorescence studies demonstrated that p190-B protein was expressed in a variety of cells and was localized diffusely in the cytoplasm and in fibrillar patterns that co-localized with the alpha 5 beta 1 integrin receptor for fibronectin. Adhesion of fibronectin-coated latex beads to cells resulted in recruitment of significant amounts of p190-B and Rho to the plasma membrane beneath the site of bead binding. In contrast, beads coated with polylysine or concanavalin A were unable to recruit p190-B or Rho. Additionally, anti-beta 1 or anti-alpha 5 integrin antibody-coated beads were also able to recruit large amounts of p190-B and Rho. These results identify a novel second member of the p190 family and establish the existence of a novel transmembrane link between integrins and a new protein p190-B and Rho.

Extracellular matrix components have marked effects on cellular morphology, growth, and differentiation, suggesting the existence of transmembrane linkages and signal transduction pathways that can relay information from the extracellular matrix to the nucleus. The cell surface receptors that mediate interactions with many extracellular matrix components are the integrin family of proteins (1). Integrins are transmembrane molecules consisting of ␣␤ heterodimers in which each ␣ and ␤ subunit has a large extracellular domain and usually a short cytoplasmic domain. Integrin receptors for extracellular matrix ligands often cluster in adhesive junctions between cells and the substratum called focal contacts (2). Actin-containing stress fibers emanate from the focal contact sites. Analysis of both the ␣ and ␤ integrin subunits has revealed that the ␤ integrin subunit is required for localization to focal contacts (3)(4)(5). Further investigations with chimeric receptors have revealed that the intracellular cytoplasmic domain of the ␤ 1 integrin subunit is sufficient for localization to focal contacts (6,7). The ␤ 1 and ␤ 3 integrin cytoplasmic domains also modulate ligand binding affinity (8), induce tyrosine phosphorylation (9,10), and function in cell spreading and migration and matrix assembly (11).
The exact interactions of the ␤ 1 integrin intracellular domain with proteins in focal contacts and stress fibers is not known. Although the intracellular integrin domain lacks any enzymatic activity, it does bind directly to talin and ␣-actinin in vitro (12,13). An increase in tyrosine phosphorylation of certain proteins is also observed following cellular attachment, which may mediate protein-protein interactions and be partially responsible for focal adhesion formation (14 -16). Consistent with this hypothesis, the cross-linking of ␣ 3 ␤ 1 integrin subunits also induces the tyrosine phosphorylation of several proteins (17). A 125-kDa tyrosine kinase, designated focal adhesion kinase (pp125 FAK ), has been identified as one of the tyrosine-phosphorylated proteins within focal contacts (18). The phosphorylation of pp125 FAK has been shown to increase markedly following cellular adhesion to fibronectin, laminin, vitronectin, and collagen IV (19). Expression and clustering of chimeras containing only the ␤ cytoplasmic domains of several integrins can mediate pp125 FAK phosphorylation (9,10). Treatment of cells with bombesin and other neuropeptides, which induces marked changes in the actin cytoskeleton, have also been shown to activate pp125 FAK (20,21). The exact role of pp125 FAK , however, in the signal transduction pathway leading to changes in the cytoskeleton is not known.
In some systems, the control of the actin cytoskeleton and cell morphology has been shown to be mediated by members of the Ras superfamily of GTPases. Although several reports have demonstrated that Ras itself affects the actin cytoskeleton (22,23), the mammalian Rho subfamily appears to be particularly important in mediating many cytoskeleton changes. In Saccharomyces cerevisiae, the Rho family member CDC42Sc regulates bud site formation and actin organization (24). In serumstarved Swiss 3T3 cells, microinjection of recombinant Rho protein rapidly stimulates stress fiber and focal adhesion formation (25). The stress fibers induced by Rho resemble those induced by serum, lysophosphatidic acid (LPA) 1 or bombesin. In contrast, microinjection of another Rho family member, Rac, resembles treatment of the cells with platelet-derived growth factor and induces membrane ruffles (26). Together these results suggest that the Rho GTPase family affect the actin cytoskeleton and markedly influence cell morphology.
Like other GTPases, the Rho family members bind and hydrolyze GTP. These proteins are in an active state when bound to GTP and are inactive when bound to GDP. An increasing number of proteins that stimulate the hydrolysis of GTP bound to Rho-related proteins have been identified and contain a similar Rho GAP domain (27). These proteins act as negative regulators of the Rho GTPases, and may serve as targets and/or effectors of these pathways. Consistent with this idea is the presence of additional structural/functional domains within many of these Rho GAP proteins that may mediate interactions with other signal transduction pathways (28). For example, the break point cluster region protein (Bcr), which acts as a GAP for Rac (29) contains both a Dbl-like GEF domain for CDC42 and a serine kinase domain. p190-A was originally identified as a protein that associates with p120 GAP following platelet-derived growth factor treatment (30). p190-A also has a GTPase domain and has GAP activity for Rho, Rac, and CDC 42Hs (31,32). The p85 regulatory domain of phosphatidylinositol 3-kinase is a multidomain protein containing SH2/SH3 domains and a Rho-GAP domain (33,34). 3BP-1 also contains a Rho GAP domain and binds the SH3 domain of Abl via a prolinerich binding domain (35).
In this report, we describe the cloning of a second p190 GTPase and find that it has GAP activity for several members of the Rho family. Immunofluorescence revealed that p190-B clustered at intracellular sites at regions of contact between the cells and fibronectin and RGD substrates, but not with polylysine or concanavalin A. Furthermore, beads coated with anti-␤1 integrin antibodies also clustered p190-B and RhoA. These results demonstrate a transmembrane linkage between integrins and the regulatory molecules p190-B and Rho, and they may be involved in mediating integrin-associated changes in the actin cytoskeleleton.

MATERIAL AND METHODS
Isolation and Characterization of the p190-B cDNA Clone-Poly(A) ϩ RNA from human mesangial cells was a generous gift from Dr. Jeffrey Kopp. An oligo(dT)-primed library was constructed with -ZAP-II as described by the manufacturer (Stratagene), using 4 g of poly(A) ϩ RNA, and approximately 3 ϫ 10 6 recombinants were obtained.
This human mesangial cell cDNA library was screened with a peptide antibody derived from the basic region (KRVRRKIRNKRAAQES-RKKKK) of the LZIP transcription factor (36). One of the antibodypositive clones was plaque-purified following three rounds of screening. Additional clones were obtained using various 5Ј-and 3Ј-end fragments as probes with both random and oligo(dT)-primed human keratinocyte cDNA libraries. Plasmids were banded in cesium chloride and sequenced by the dideoxy nucleotide method with Sequenase (U. S. Biochemical Corp.). Sequencing was performed using a combination of controlled unidirectional deletions and sequence-deduced oligonucleotide primers.
Recombinant p190-B GAP Protein and p190-B Antiserum Production-cDNA fragments of p190-B were prepared using the polymerase chain reaction. These fragments were subcloned into the BamHI-SalI site of the PQE-9 bacterial expression vector containing a tag of six histidine residues at the N terminus (Qiagen, Chatsworth, CA) and confirmed by DNA sequence analysis. Fusion proteins were induced in bacteria by isopropyl treatment and extracted using 6 M guanidine-HCl. These proteins were purified on a nickel-affinity column as described by the manufacturer.
Fusion protein (amino acid residues 1203-1499) was used as antigen to immunize a rabbit, and antiserum was obtained after four injections. Antibodies were then affinity purified on a p190-B fusion-Sepharose affinity column.
GAP Activity Assay-The p190-B Rho GAP domain (amino acid residues 1203-1499) was expressed as a polyhistidine fusion protein and purified on a nickel affinity column as described above. Several members of the Rho family of proteins were affinity purified from Escherichia coli as glutathione S-transferase fusion proteins and released from glutathione-agarose beads using thombin (25). For GTPase assays under subsaturating conditions, the proteins (4 ng) were preloaded with ␥-32 P (DuPont NEN, 60,000 Ci/mmol) by incubating for 10 min at 30°C in a buffer containing 20 mM Tris, pH 7.5, 25 mM NaCl, 5 mM EDTA, 0.1 mM dithiothreitol. The reaction was stopped by the addition of MgCl 2 to a final concentration of 20 mM. [␥-32 P]GTP-bound proteins (4 ng) were incubated in 30 l of a solution containing 20 mM Tris, pH 7.5, 0.1 mM dithiothreitol, 1 mg/ml bovine serum albumin, 1 mM cold GTP, in the presence or absence of 0.6 l of p190-B (0.2 mg/ml). Aliquots (9 l) were removed immediately, and following a 10-min incubation at 20°C they were added to 1 ml of ice-cold buffer A (50 mM Tris, pH 7.5, 50 mM NaCl, 5 mM MgCl 2 ). Samples were filtered through nitrocellulose (BA 85, Schleicher and Schuell). Filters were washed with 10 ml of buffer A, and the amount of radioactivity bound was determined.
Immunoprecipitations-Rabbit anti-p120 GAP polyclonal and an anti-p120 GAP mouse monoclonal antibodies were obtained from Upstate Biochemicals (Lake Placid, NY). An anti-p190-A antibody was obtained from Transduction Laboratories (Lexington, KY). Normal human foreskin fibroblasts, HT-1080 fibrosarcoma cells and RD cells were cultured in Dulbecco's modified medium with 10% fetal calf serum, 1 mM glutamine. For immunoprecipitation, human foreskin fibroblasts, RD muscle cells, or HT-1080 fibrosarcoma cells were labeled with [ 35 S]methionine (ICN) in methionine-free medium for 4 h. After washing, the cells were extracted with 1% Triton X-100, 2 mM phenylmethylsulfonyl fluoride in phosphate-buffered saline on ice, followed by centrifugation at 15,000 rpm for 15 min at 4°C. All steps of the immunoprecipitation were carried out on ice or in a 4°C room. The supernatants of the centifuged lysates were incubated with preimmune, anti-p190-B, or anti-p120 GAP antibody in the presence of protein G-Sepharose (Pierce, Rockford, IL) for 10 h. The beads were washed 3 times with lysis buffer and boiled in 1 ϫ electrophoresis buffer containing 10 mM dithiothreitol. The samples were then resolved by SDS-polyacrylamide gel electrophoresis electrophoresis in a 4 -12% gel. The gel was fixed and incubated with Amplify intensifying solution (Amersham Corp.), dried, and exposed for autoradiography.
Experiments were also performed to test whether the p190-B antibody cross-reacted with p190-A protein. Immunoprecipitation experiments were performed using protein G-Sepharose alone or with the anti-p190-B antibody using [ 35 S]methionine-labeled extracts from HT-1080 cells. Duplicate samples, as well as, total cell lysate as positive control, were electrophoresed and transferred to nitrocellulose. The blot was then blocked with phosphate-buffered saline containing 5% nonfat milk, 0.1% Tween 20. The filter was then incubated for 2 h with a mouse monoclonal antibody against p190-A. After washing, an anti-rabbit or anti-mouse peroxidase-conjugated secondary antibody was used. Visualization was performed by enhanced chemiluminescence (Amersham Corp.).
Bead Binding Assay and Immunofluorescence-Extracellular matrix proteins or antibodies were immobilized on latex beads as described previously (37). Polystyrene latex beads (11 m average diameter, Sigma) were coated with either 50 g/ml fibronectin, 100 g/ml concanavalin A, or 100 g/ml polylysine. Goat-IgG-GRGDSPC peptide conjugates were synthesized as described previously (38) and were used at 50 g/ml to coat the beads. Two anti-functional integrin monoclonal antibodies, ␣ 5 -M16 and ␤ 1 -M13 (39) were also used to coat the beads at 50 g/ml. After coating, each batch of beads was incubated with 10 mg/ml bovine serum albumin in phosphate-buffered saline for 60 min at 22°C.
Acid-washed 22 ϫ 22-mm glass coverslips were coated with 20 g/ml collagen I (Vitrogen 100, Collagen Corp., Palo Alto, CA) overnight at 4°C and then blocked for 1 h with 10 mg/ml heat-denatured bovine serum albumin. To ensure an initially diffuse distribution of fibronectin receptors, fibronectin was depleted from the cultures. This was accomplished by blocking fibronectin synthesis by preincubating the cells for 2 h in Dulbecco's modified Eagle's medium containing 25 g/ml cycloheximide and using fibronectin-depleted serum. After 2 h, the cells were washed twice with Hanks' balanced salt solution and then briefly detached with 0.05% trypsin, 0.53 mM EDTA, and 25 g/ml cycloheximide. The cells were washed with Dulbecco's modified Eagle's medium containing fibronectin-depleted serum and cycloheximide. Cells (1 ϫ 10 5 ) were subsequently plated on the collagen-coated coverslips for 60 min at 37°C. Next, each different type of bead (approximately 2 ϫ 10 6 ) was incubated with the cells for 20 min. Typically, 40 -60% of the cells retained the beads within the outline of the spreading cell. The cells were fixed and stained using primary and secondary antibodies as described previously (7). For immunostaining of ␣-tubulin, mouse monoclonal anti-␣-tubulin was used (Sigma Immunochemicals). A rabbit anti-RhoA polyclonal antibody was used as described previously (40). Primary antibodies were visualized using fluorescein isothiocyanate-conjugated goat anti-rabbit, anti-mouse, and anti-rat IgG antibodies and rhodamine-conjugated goat anti-rat IgG antibody (Biosource International). For double-label immunofluorescence, the distribution of p190-B protein was analyzed using rabbit polyclonal anti-p190-B antibody with Texas red-labeled goat anti-rabbit IgG second antibody (Molecular Probes) and rat monoclonal antibody 11 to the ␣5 integrin subunit with fluorescein isothiocyanate-conjugated anti-rat IgG antibody (Biosource International). Immunofluorescence microscopy was carried out using an Optiphot-2 microscope (Nikon, Japan) or an Axiophot microscope (Zeiss, Germany), equipped for fluorescein and rhodamine labeling.
The assessment of clustering of RhoA and p190-B was calculated as the ratio of A/B, where A is the number of beads showing clustering of the antigen adjacent to a bead and B is the total number of beads located completely within the outline of the spreading cell. 50 beads (B) were counted in each experiment, and each experiment was performed 6 times.

RESULTS
Cloning of Human p190-B-In order to identify new transcription factors in the kidney, a human mesangial cell cDNA library was screened with an anti-peptide antibody that recognizes the bipartite DNA-binding domain and predicted nuclear targeting domain of the Lzip-1 and Lzip-2 transcription factors (36). Strong antibody cross-reactivity identified several clones from this kidney library. One of these human clones serendipitously encoded part of a long open reading frame with significant similarity to the rat RhoGAP-containing protein, p190 (32). Using this cDNA clone as probe, we isolated several overlapping cDNAs from a human keratinocyte library. The composite nucleotide sequence of the cDNAs coding for this protein, designated p190-B, is 4,993 base pairs long and contains an open reading frame coding for 1,499 amino acid residues (Fig.  1). The nucleotides surrounding the initiation methionine matched the consensus ATG start codon (41). The antibody used to clone p190-B most likely recognized a cluster of basic amino acid residues beginning at amino acid residue 1224.
Northern blot analysis of various rat and human tissues with the p190-B cDNA identified two transcripts of 4.4-kilobase mRNA and 6.4 kilobases in a variety of tissues including kidney, brain, liver, and lung (data not shown). These results suggest that p190-B mRNA is commonly expressed in most tissues.
Homology of P190-B with Other Proteins-To determine whether the p190-B gene product showed sequence similarity to any other proteins, a BLAST search was performed on the GenBank protein data base (August, 1994). p190-B showed the strongest similarity with the putative N-terminal GTPase domain and C-terminal Rho GAP domain of p190-A. Direct comparison of human p190-B with the rat p190-A revealed 51% amino acid identity over 1499 amino acid residues (Fig. 1). Although the full-length amino acid sequence of human p190-A is not known, the region of p190-B corresponding to the known human sequence, lacking the GTPase and Rho GAP domains (42), showed 45% identity. These results confirm that p190-B is indeed a homologue of p190-A.
p190-B was also homologous with other GTPases and Rho GAP-containing proteins. The N terminus of p190-B has a predicted GTPase domain (Fig. 2) containing the various conserved GTPase motifs (43). p190-B shares some features in common with several members of the Ras, Rab, Ral, and Rho family of small GTPases, but it is most closely related to p190-A. Unlike the small GTPases, however, p190-B showed a 60-amino acid residue insertion between the G2 and G3 domains. The exact functional significance of this region is unclear, but it is reminiscent of the amino acid insertion seen between G1 and G2 in the G protein-coupled receptors.
The C terminus of p190-B shows considerable identity with other Rho GAP-containing proteins (44) such as p190-A (58%), n-chimerin (41%), Bcr (31%), p85 (27%), and RhoGAP (18% identity). This area of homology comprising approximately 150 amino acid residues has now been found in a number of different proteins and is responsible for specifically accelerating the hydrolysis of GTP bound to different Rho members. More recently, the GAP domains of Bcr, p190-A, and RhoGAP have been shown to be active when micoinjected into cells and to have specificity for different Rho proteins (45).
A systematic search for additional functional domains/protein motifs in p190-B yielded two findings. First, p190-B may contain several potential regions involved in interacting with SH2-and SH3-containing proteins. p190-B contains four distinct regions rich in proline residues homologous to the SH3 binding domains of 3BP-1, 3BP-2, and formin (46). Two of these proline-rich sequences are located in the middle of the p190-B protein (amino acid residues 984 -997 and 1033-046), and one of these sequences occurs in a similar location in p190-A. Two other potential proline-rich SH3 binding domains are located at the C-terminal end of p190-B (amino acid residues 1457-1470 and 1475-1488). Interestingly, there are no such sequences found in the C terminus of p190-A. p190-B also contains a repeated amino acid sequence EEYIN (amino acid residues 228 -233 and 276 -281) separated by 28 amino acid residues, which is similar to a high affinity binding site for the SH2 domain of SEM-5/GRB2 (47). The presence of these additional regions in p190-B suggests that p190-A and p190-B may differ in their ability to interact with GAP and/or other SH2-and SH3-containing proteins. Second, both p-190-A and p190-B contain a cluster of basic amino acid residues that precedes the Rho GAP domain. Although similar basic amino acid stretches occur as nuclear targeting regions in a number of proteins, the proximity of this region to the Rho GAP domain may suggest that it might also be important in dictating the conformation of the Rho GAP domain of p190-B.
p190 B Has GAP Activity for Rho and Rac-The strong homology of p190-B with other proteins that accelerate the hydrolysis of GTP bound to Rho proteins suggested that p190-B might also have such activity. To test this hypothesis, a fusion protein spanning the Rho GAP domain of p190-B was expressed in Escherichia coli using the PQE-9 bacterial expression system. This fusion protein contained a 6 ϫ histidine tag, which allows affinity purification by chromatography on nickel-Sepharose. The GAP activity of this fusion protein was demonstrated for three Rho family proteins including RhoA, Rac1, and G25K (CDC42), enhancing the release of 32 P from each of the proteins (Fig. 3). Smaller fusion proteins lacking the full Rho GAP domain were inactive in this assay (data not shown). A Val 12 mutant of G25K, similar to the analogous Ras Val 12 mutant protein (48), is not susceptible to the Rho-GAP-accelerated GTP hydrolysis (49). The p190-B fusion protein was unable to stimulate the GTPase activity of the G25K Val 12 mutated fusion protein (Fig. 3). These results confirm that in vitro, p190-B has GAP activity for the Rho family of proteins.
p190-B Is Expressed in a Variety of Cell Types-An antibody was generated against the p190-B fusion protein containing amino acid residues 1100 -1499. Antisera from the immunized rabbit were affinity-purified on a p190-B fusion protein affinity column. Immunoprecipitation experiments utilizing this antibody were performed on [ 35 S]methionine-labeled cellular extracts from normal human foreskin fibroblasts, RD muscle cells, and HT-1080 cells. Using the anti-p190-B antibody, a protein of approximately 195 kDa was observed in all three cell extracts (Fig. 4A). The reason for the difference in size between the calculated (172 kDa) and observed value (195 kDa) is unclear, but it is known that many proteins, including p190-A, migrate anomalously on SDS-polyacrylamide gels (32).
Immunoprecipitation with an anti-p120 GAP polyclonal antibody using extracts from the three different cell types revealed FIG. 1. Homology between human p190-B and rat p190-A. Amino acid sequence comparision is shown between p190-B and p190-A using the program BESTFIT. Identical amino acid residues between the two p190 forms is denoted by the dashed line. The GTPase motifs are underlined. The Rho GAP domain is boxed.
strong reactivity with the 120-kDa Ras-GAP species (Fig. 4A). In addition, the HT-1080 cell extracts showed a large amount of 190-kDa associated protein (Fig. 4), which was not observed in the other two cell types. Preliminary experiments using the anti-p190-B antibody indicate that p190-B also interacts with p120 Ras-GAP. 2 Additional immunoprecipitation experiments with HT-1080 cell extracts were used to determine whether the p190-B antibody could react with the p190-A protein. Significant amounts of p190-B protein were again detected using the anti-p190-B antibody (lane 2). Additionally, Western blotting with the anti-p190-A antibody detected a large amount of p190-A immunoreactivity with HT-1080 crude cell lysate (lane 3). However, no p190-A reactivity was detected in duplicate samples from protein G-Sepharose (lane 4) or anti-p190-B immunoprecipitates (lane 5). Thus, these results demonstrate that the p190-B antibody does not recognize p190-A protein.
Contact of Cells with Fibronectin-coated Beads Induces the Aggregation of p190-B and RhoA-Immunofluorescence staining of human foreskin fibroblasts revealed that p190-B protein was localized in a weak fibrillar pattern, with both diffuse and distinctly fibrillar components (Fig. 5). The later pattern resembles the previously described localization of the ␣ 5 ␤ 1 fibronectin receptors associated with fibronectin fibrils in extracellular matrix contacts rather than in focal contacts (7). Double labeling experiments using antibodies against p190-B and the ␣ 5 integrin subunit revealed co-localization of these molecules in streak-like patterns at the cell periphery (Fig. 5).
Because of the p190-B staining pattern and a potential role of p190-A in cytoskeletal structure and cell adhesion (50), we next examined the transmembrane control of distribution of p190-B in normal human fibroblasts using beads coated with fibronectin or anti-integrin antibodies. This approach has been useful for studying early integrin interactions with cytoskeletal 2 P. D. Burbelo, unpublished results. FIG. 2. p190-B and p190-A contain a novel GTPase domain. The N-terminal region of p190-B contains several motifs that have been shown to be required for guanine nucleotide binding (43). The potential GTPase motifs in p190-A are also shown (31).

FIG. 3. p190-B has GAP activity for several Rho GTPases. Each
GTPases was bound to [␥-32 P]GTP and incubated in the absence or presence of purified p190-B. The radioactivity remaining on the GTPases was determined in a filter binding assay as described under "Materials and Methods." The percentage of radioactivity remaining is shown relative to the initial radioactivity bound to each GTPase.

FIG. 4. Immunoprecipitation experiments with anti-p190-B
and anti-p120 GAP antibodies. A, [ 35 S]methionine-labeled extracts from normal foreskin fibroblast, HT-1080, or RD cells were subjected to immunoprecipitation with either preimmune, affinity-purified anti-p190-B antiserum, or anti-p120 GAP antiserum. The proteins were then resolved by 4 -12% SDS-polyacrylamide gel electrophoresis, fluorographed, and exposed for autoradiography. B, HT-1080 cell extracts were either immunoprecipitated (I.P.) with either protein G-Sepharose alone (lanes 1 and 4) or polyclonal anti-p190-B antibody (lanes 2 and 5). Total cell (T.C.) lysate was also used (lane 3). The proteins were then resolved by 4 -12% SDS-polyacrylamide gel electrophoresis, and lanes 3, 4, and 5 were transferred to Immobilon P. The filter was then subjected to Western blot analysis (W.B.) with the anti-p190-A monoclonal antibody (lanes 3, 4, and 5) as described under "Materials and Methods." Molecular weight markers of 205, 118, 85 and 47 kDa are denoted by the rectangle symbols. molecules (37,51,52). RhoA, a potential target of p190-B within focal adhesions was also examined using this assay. Several different substrates were attached to the latex beads including fibronectin, RGD peptide, polylysine, and concanavalin A. Normal human fibroblasts were treated with cycloheximide and then plated in fibronectin-depleted serum on collagen I-coated coverslips, which produces cells with a nearly homogenous, diffuse pattern of integrin distribution.
Addition of the beads resulted in binding to approximately 40 -60% of the fibroblasts for each of the different beads after 20 min (beads that bound to a lateral edge were not counted). Immunofluorescence with anti-p190-B and anti-RhoA antibodies was used to evaluate the distribution of these proteins in the region of contact between the cells and the extracellular matrix-coated beads. In the case of fibronectin-coated beads, both p190-B (Fig. 6A) and RhoA (Fig. 6E) showed marked accumulation and staining around the beads. Under these conditions, integrin ␤ 1 staining was also strongly observed as expected (data not shown). In contrast, the polylysine-coated beads did not show p190-B (Fig. 6B) or RhoA (Fig. 6F) staining, even though they attached equally well to the fibroblasts. Neither the fibronectin-nor polylysine-coated beads induced clustering of ␣-tubulin (Fig. 6, D and E). These results suggest that integrin-mediated attachment to fibronectin induces the clustering of both p190-B and RhoA.
To test directly whether integrin antibodies could induce clustering of p190-B or RhoA, latex beads coated with two different monoclonal antibodies to integrin subunits were used. Immunofluorescence with the anti-p190-B antibody again revealed a marked clustering of p190-B around beads coated with antibodies that bind to extracellular epitopes of the ␤ 1 (Fig. 7A) or ␣ 5 integrin subunit (Fig. 7C). Additionally, p190-A also showed clustering to below the plasma membrane following integrin engagement (data not shown). A similar pattern of staining was observed with the anti-RhoA antibody for both the anti-␤ 1 (Fig. 7B) or anti-␣ 5 integrin (Fig. 7D) subunit antibody. No clustering of p190-B (Fig. 7E) or RhoA (Fig. 7F) was observed around beads coated with concanavalin A or with anti-␤ 1 integrin antibodies that bind to intracellular epitopes (data not shown). Quantitation of the number of beads positive for clustering of p190-B and RhoA around the different substrates and antibodies is shown in Fig. 8. A similar clustering pattern was observed for p190-B and RhoA with each of the different beads. An immobilized GRGDS conjugate was as effective as fibronectin in recruiting both p190-B and RhoA.  Taken together, these results demonstrate that both p190-B and RhoA are markedly recruited to sites of integrin clustering and thus may be involved in the integrin signaling pathway. DISCUSSION In this article, we describe the cloning and characterization of p190-B, a new member of the p190 GTPase family of RhoGAP proteins. The domain structure of both human p190-B and rat p190-A are similar and share 51% amino acid identity. It is also interesting to note that of all the RhoGAP proteins discovered so far, only p190-A and p190-B have potential GTPase domains. The high conservation between the two p190 forms suggests that these proteins represent a novel family of GTPases. One unusual feature within the GTPase domain of p190-A and p190-B is the presence a 60-amino acid insertion between the G2 and G3 motifs. A similar looping structure occurs in the large G proteins and is involved in regulating their GTPase activity. Recently, Settleman and co-workers (53) showed that p190-A binds to GTP, and thus one would also expect that p190-B also binds GTP. The carboxyl terminus region of p190-B contains a Rho GAP domain, and a fusion protein derived from this region was found to stimulate the hydrolysis of GTP of several Rho family members. The lack of specificity of p190-B's RhoGAP activity for any particular Rho member is similar to that of p190-A. One possibility is that p190-B's GAP does have preference for individual members of the Rho family in vivo. The GAP domain of p190-A, for example, has been shown to inhibit Rho but not Rac after micoinjection into cells (45). This specificity could be due to additional protein domains of p190 affecting different protein-protein interactions and localization within the cell. For example, specific phosphotyrosine residues within p190-B and/or proline-rich regions in the carboxyl-terminal region may regulate or target it to specific SH2-and SH3-containing proteins within the cell.
The exact role(s) of p190-A and p190-B is still unclear. One possible function of both p190 proteins is to reduce the p21 ras GTPase-activating potential of p120 GAP , resulting in higher levels of p21 ras -GTP (54,55). Alternatively, both p190 proteins may function in other signaling pathway(s). For example, McGlade et al. (50) showed that expression of a truncated amino-terminal SH2-SH3 domain of p120 GAP altered cytoskeletal structure and cell adhesion. Since this truncated protein was shown to bind p190-A (and presumably p190-B), the effects on cytoskeletal structure and cell adhesion may occur via the Rho GAP activity of both p190 proteins. These results are consistent with our observations that fibronectin-or integrincoated beads recruited p190-B and Rho proteins to intracellular sites immediately adjacent to extracellular matrix-integrin interactions. This clustering activity of Rho and p190-B appeared to be specific for integrins, since nonspecific aggregation of many glycoproteins in the membrane by concanavalin Acoated or polylysine-coated beads did not induce p190-B or Rho clustering. One possible mechanism of this integrin-induced clustering of Rho and p190-B may involve protein phosphorylation by potential kinases such as pp125 FAK or c-Src. Indeed, a similar pattern of clustering has also been observed for other cytoskeletal proteins and signaling molecules such as  Beads coated with fibronectin, goat-IgG-RGD peptide conjugates, anti-␣ 5 -m16 integrin subunit, anti-␤ 1 -m13 integrin subunit, concanavalin A, or polylysine were incubated with human foreskin fibroblasts for 20 min as described under " Materials and Methods." The percentage of beads that induced recruitment of Rho and p190-B was determined by scoring 50 cell-bound beads per experiment and scoring the number of beads that exhibited immunofluorescence staining for p190-B or RhoA proteins. The results from six independent experiments were pooled and expressed as mean Ϯ standard deviation. pp125 FAK and tensin (56). The exact role, however, of both Rho and p190-B in integrin signaling activation is still not clear. A functional link between integrins and Rho is supported by earlier observations that inhibition of Rho in T lymphocytes by C3 transferase, a specific inhibitor of Rho, could completely block LFA-1-mediated cell aggregation, while blocking actin polymerization had no effect on aggregation (57). Although future work on identifying effector proteins for Rho and p190 proteins is needed to clarify their exact function, our results support a hypothesis that extracellular signals from extracellular matrix molecules may influence the activity of p190 and Rho proteins.