Direct binding of the proline-rich region of protein tyrosine phosphatase 1B to the Src homology 3 domain of p130(Cas).

Protein tyrosine phosphatase 1B (PTP1B) is an abundant intracellular enzyme that is thought to act as a negative regulator of certain signaling pathways. The C terminus of PTP1B contains two proline-rich regions which conform to the canonical class II Src homology 3 domain binding motif, Pro-X-X-Pro-X-Arg. In this study, we establish that PTP1B interacts with Crk, Grb2, and p130Cas in vitro and with at least one of these, p130Cas, in intact cells. The interaction of PTP1B and p130Cas is independent of tyrosine phosphorylation but can be disrupted by replacing two critical proline residues in the proline-rich domain of PTP1B between amino acids 301 and 315. When wild-type PTP1B is expressed in 3Y1-v-crk cells, p130Cas shows substantial dephosphorylation, whereas the PTP1B proline mutant does not have this effect. In 3Y1 and 3Y1 v-crk-transformed fibroblasts, almost all of the total PTP1B and about 40% of total p130Cas co-sediment with membranes composed primarily of endoplasmic reticulum. These results suggest that the proline-rich domain between amino acids 301 and 315 in PTP1B binds Src homology 3-containing proteins and that p130Cas may be a physiological target of this phosphatase in cells.

Protein tyrosine phosphatases (PTPs) 1 play a critical role in regulating a wide variety of intracellular signaling processes (1). In vitro, most PTPs display a broad substrate specificity, raising a question as to how these enzymes recognize their appropriate targets in cells. It has been suggested that differential location may account at least in part for substrate selection by PTPs (2). Many PTPs contain targeting motifs, which direct localization to particular sites within the cell, and such restricted location may limit access to substrates (2,3). In addition, some PTPs, such as Shp-1 and -2, contain Src homology 2 (SH2) domains, which direct binding to specific phosphotyrosine-containing proteins (4). However, the mechanisms by which most other PTPs select their targets remain unclear.
PTP1B is the prototype of nontransmembrane PTPs and has served as a useful model for these enzymes. It was originally identified as the major PTP activity in human placenta (5,6). Subsequent studies have shown that PTP1B is a ubiquitous and abundant enzyme, suggesting that it plays a general role in controlling cellular function (7,8). PTP1B localizes to the endoplasmic reticulum (ER) via its 35-amino acid C-terminal sequence, with its phosphatase domain oriented toward the cytoplasm (9). Although the mechanism by which PTP1B is regulated is not understood, PTP1B undergoes cell cycle-regulated serine phosphorylation as well as alternative splicing (10 -12). In addition, in some cell types, PTP1B undergoes Ca 2ϩ -dependent proteolysis, releasing a C-terminal truncated, soluble form of the enzyme, which may then act on previously inaccessible substrates (8). Overexpression of artificially generated C-terminal truncations of PTP1B 2 and the related enzyme T-cell PTP (13,14) has profound effects on cell proliferation that are not seen with the full-length molecule, again suggesting that appropriate location of PTPs is critical to their function. The physiological substrates of PTP1B have not been identified. In an attempt to identify candidate substrates, we noted that PTP1B contains two proline-rich domains, which fit the consensus sequence for class II SH3 domain binding motifs (15,16). In this study, we tested whether these proline-rich domains could direct PTP1B to recognize potential SH3-containing substrates. We found that PTP1B selectively binds to SH3 domains derived from Grb2, Crk, and p130 Cas in vitro. The binding between PTP1B and p130 Cas was confirmed in vivo and found to be mediated through one of the two proline-rich domains on PTP1B. This interaction is independent of p130 Cas tyrosine phosphorylation levels. In v-crk-transformed fibroblasts, overexpression of wild-type PTP1B, but not a mutant form unable to bind p130 Cas , results in dephosphorylation of p130 Cas , suggesting that PTP1B might use its proline-rich domain to recognize and dephosphorylate this protein. Like PTP1B, a substantial amount of p130 Cas co-sediments with ER membranes. Thus, the proline rich domains on PTP1B may direct this enzyme to some of its targets in cells.

EXPERIMENTAL PROCEDURES
Materials-3Y1 and 3Y1 v-crk-transformed cells, as well as glutathione S-transferase (GST) fusion proteins containing SH3 domains from Abl, Arg, Crk1, Crk2, Eps8, Fyn, Gap, Grb2C, Grb2N, Nck, phospholipase C␥, p130 Cas , and Sprk, respectively, were provided by Gary Kruh (Fox Chase Cancer Center). 3Y1 and 3Y1 v-crk-transformed cells were maintained in Dulbecco's modified Eagle's medium plus 10% fetal bovine serum. Wild-type and mutant forms of PTP1B were subcloned as BamHI-EcoRI fragments into pGEX-2T, and GST-PTP1B fusion proteins were made and purified by standard methods (17 anti-Fak antibodies were purchased from Transduction Laboratories, and polyclonal anti-p130 Cas antibody was from Santa Cruz Biotechnology, Inc. Expression Plasmids-A catalytically inactive, C215S (CS) mutant of PTP1B was constructed by site-directed mutagenesis using a standard technique (18). Truncated and internally deleted forms of PTP1B (PTP1B-321 and PTP1B-403 ⌬300 -320 ) were constructed by polymerase chain reaction mutagenesis. A P309A/P310A (PA) mutant of PTP1B was made by the unique site elimination method of Deng and Nickoloff (19). Mutations were confirmed by sequence analysis. pJ3H-PTP1B constructs were made as described previously (20). These plasmids express an N-terminal HA-tagged PTP1B.
Transient Transfection-COS1 cells were grown to 80% confluence in Dulbecco's modified Eagle's medium plus 10% fetal bovine serum and transfected with pJ3H-PTP expression plasmids using LipofectAMINE (Life Technologies, Inc) according to the manufacturer's recommendations. Forty-eight hours after transfection, the cells were harvested for analysis. 3Y1 and 3Y1-v-crk cells were grown to 40% confluence in Dulbecco's modified Eagle's medium plus 10% fetal bovine serum and transfected with expression plasmids using a calcium phosphate precipitation method (21). Forty-eight hours after transfection, the cells were harvested for analysis.
In Vitro Binding Assays-Transfected COS1 cells were lysed in Nonidet P-40 lysis buffer (50 mM Tris-HCl, pH 8.0, 137 mM NaCl, 10% glycerol, 1% Nonidet P-40, 50 mM NaF, and 10 mM ␤-glycerol-phosphate) containing 1 mM sodium vanadate, 1 mM phenylmethysulfonyl fluoride, and 10 g/ml aprotinin. Lysate protein concentrations were measured using BCA (Pierce). 500 g of total cell lysates were incubated with 5 l of GST SH3 domain fusion protein beads at 4°C for 2 h. The beads were washed three times with Nonidet P-40 lysis buffer and then boiled in SDS sample buffer. The samples were fractionated by 10% SDS-PAGE, transferred to nitrocellulose membranes, and probed with anti-HA antibody 12CA5. Signals were developed by chemiluminescence (Pierce). For 3Y1 and 3Y1 v-crk-transformed cells, 500 g of cell lysate were incubated with 10 l of GST PTP1B fusion protein beads. The samples were separated by 6% SDS-PAGE and analyzed by immunoblot using anti-p130 Cas antibodies.
Immunoprecipitation and Immunoblot-3Y1 and 3Y1-v-crk cells were transiently transfected with either pJ3H alone or pJ3H bearing PTP1B, CS-PTP1B, or PA-PTP1B. Cells were lysed in Nonidet P-40 lysis buffer. For immunoprecipitation, 1 mg of cell lysates was immunoprecipitated with 2 g of anti-HA antibody or anti-PTP1B antibody, or 250 g of cell lysates were immunoprecipitated with 2 g of anti-p130 Cas (polyclonal) antibody at 4°C for 2 h. Immunocomplexes were washed three times with Nonidet P-40 lysis buffer and boiled for 5 min in SDS-PAGE sample buffer. The samples were fractionated by 7% SDS-PAGE and transferred to nitrocellulose membranes. The membranes were probed with antiphosphotyrosine antibodies. The blots were then stripped and reprobed with anti-p130 Cas antibodies.
Enzyme Assay-GST-PTP1B and GST-PA-PTP1B activity was measured in a reaction mixture (1 ml) containing 10 mM para-nitrophenol phosphate in 100 mM sodium acetate (pH 5.5), 1 mM EDTA, with ionic strength adjusted to 0.15 M with NaCl. The reaction mixture was placed in a 30°C water bath for 5 min prior to the addition of PTP1B (5-l protein beads). For K m determinations, the amount of substrate was varied from 0.25 to 20 mM. Incubations were carried out for 5 min at 30°C, and the reaction was terminated with 0.5 ml of 1 M NaOH. The amount of product (p-nitrophenol) produced was measured from the increase in absorbance at 405 nm at 30-s intervals. The nonenzymatic hydrolysis of para-nitrophenol phosphate was corrected by measuring the increase in absorbance at 405 nm obtained in the absence of enzyme.
Subcellular Fractionation-Cell fractionation was performed as described previously (9). Briefly, cells were lysed by Dounce homogenization, nuclei and unbroken cells were pelleted at 1,000 ϫ g (P1 fraction), and the supernatant was centrifuged at 100,000 ϫ g, yielding a supernatant (S100, mostly cytosol) and pellet (P100, cell membranes). Cellular membranes in the P100 fraction were further separated by isopycnic centrifugation such that low density membranes (plasma membrane and Golgi) fractionate to the interface between 0.25 and 1.0 M sucrose, whereas high density membranes (mostly rough endoplasmic reticulum) fractionate to the interface between 1.2 and 2.0 M sucrose. The extent of enrichment for various subcellular compartments was assessed by assaying specific markers as follows; protein concentration was assayed using BCA, and DNA (22), 5Ј nucleotidase (23), NADPH cytochrome c reductase (24), and lactate dehydrogenase (25) were assayed by standard methods. All chemical reagents were purchased from Sigma.

PTP1B Binds Selectively to SH3 Domain Fusion Proteins in
Vitro-PTP1B contains two proline-rich motifs that fit the con- Lysates from 3Y1 and 3Y1-v-crk cells were incubated with glutathione-agarose beads bound to either GST or GST-PTP1B. The beads were separated by 7% SDS-PAGE and immunoblotted with monoclonal anti-p130 Cas antibodies (A) and antiphosphotyrosine antibodies (B), and GST-PTP1B and GST-alone protein beads were separated by 10% SDS-PAGE and stained with Coomassie Blue (C). p130 Cas -P is tyrosine-phosphorylated p130 Cas . sensus sequence for SH3 binding (15,16). In an attempt to determine whether PTP1B binds to proteins containing SH3 domains, we assessed the ability of this phosphatase to associate with a variety of SH3-containing proteins in vitro. Purified GST-SH3 domain fusion proteins were immobilized on glutathione-Sepharose beads and then incubated with lysates from COS cells expressing HA-tagged PTP1B. The proteins adsorbed by the GST fusion proteins were analyzed by anti-HA immunoblot (Fig. 1A). Among the constructs tested, only the SH3 domains derived from Grb2 (both N-and C-terminal SH3 domains), Crk, and p130 Cas bound detectable amounts of PTP1B. Similar results were obtained with nontagged PTP1B (data not shown). The SH3 domains from Crk and Grb2 are known to bind class II SH3 ligands (Pro-X-X-Pro-X-Arg) (26), whereas the preferred ligands for p130 Cas are unknown. We also identified Crk as a PTP1B-binding protein using a yeast interaction trap screen (27), with PTP1B as bait and a HeLa cDNA library as source of interacting proteins. 3 As the SH3 domain derived from p130 Cas consistently bound more PTP1B than any other SH3 domain tested in vitro (as assessed by densitometry of immunoblots), we investigated this interaction in more detail.
PTP1B Binds p130 Cas in a Phosphotyrosine-independent Manner in Vitro-To determine whether tyrosine phosphorylation of p130 Cas affects its binding to PTP1B, we analyzed the ability of GST-PTP1B to bind p130 Cas from 3Y1 and 3Y1-v-crk fibroblasts. Although these two lines have similar amounts of p130 Cas protein, the level of p130 Cas tyrosine phosphorylation is greatly elevated in 3Y1-v-crk cells (Ref. 28 and Fig. 2B). GST-PTP1B fusion protein beads were incubated with 3Y1 and 3Y1-v-crk cell lysates, and the beads were washed and ana-lyzed by anti-p130 Cas and antiphosphotyrosine immunoblot (Fig. 2). Both non-tyrosine-phosphorylated p130 Cas from 3Y1 cells and tyrosine-phosphorylated p130 Cas from 3Y1-v-crk cells bound about equally well to the GST-PTP1B fusion protein, indicating that this association is phosphotyrosineindependent.
The p130 Cas Binding Element on PTP1B Requires Prolines 309 and 310 -There are two proline-rich domains in PTP1B. One (PPPEHIPPPPRPPKR) is located from amino acids 301 to 315, the other (SPAKGEPSLPEK) spans amino acids 386 -397 (Fig. 3A). Both of these proline-rich motifs contain the consensus for class II SH3 binding ligands (PXXPXR/K) (16). To determine whether either of these regions mediates binding to p130 Cas , we tested the binding properties of PTP1B constructs that lack one or the other proline-rich domain. PTP-321, which lacks the second (more C-terminal) proline-rich domain, and PTP-403 ⌬300 -320 , which lacks the first (more N-terminal) proline-rich domain (Fig. 3A), were expressed in COS cells, and the cell lysates were incubated with immobilized p130 Cas SH3 domain. Anti-HA immunoblotting revealed that the p130 Cas SH3 domain binds to PTP321 but not to PTP403 ⌬300 -320 , indicating that the first proline-rich domain is required for binding to p130 Cas (Fig. 3, B and C). To further characterize the binding elements, we made point mutations within this proline-rich domain, replacing proline residues 309 and 310 with alanine. Based on the crystal structure of SH3 ligands, these mutations are predicted to prevent formation of a left-handed prolinecontaining helix required for interaction with SH3 domains (15). This mutant form of PTP1B (PA-PTP) fails to bind the p130 Cas SH3 domain (Fig. 3D). These results indicate that the association of PTP1B with p130 Cas is likely to be based on an interaction between the more N-terminal of the two PTP1B proline-rich regions to the p130 Cas SH3 domain.
PTP1B Interacts with p130 Cas in Cells-To investigate potential interactions between PTP1B and p130 Cas in intact cells, we asked whether p130 Cas co-immunoprecipitates with PTP1B from 3Y1-v-crk cells. Anti-PTP1B immunoprecipitates contain a ϳ130-kDa band that co-migrates with and is recognized by anti-p130 Cas antibodies (Fig. 4). This band is not seen in control immunoprecipitates. The p130 protein that co-immunoprecipitates with PTP1B is tyrosine-phosphorylated, as is authentic p130 Cas . These data strongly suggest that endogenous PTP1B associates with p130 Cas in 3Y1-v-crk cells.
To further investigate the nature of the PTP1B-p130 Cas interaction in cells, 3Y1 and 3Y1-v-crk cells were transiently transfected with various PTP1B expression plasmids, and the cells were lysed and immunoprecipitated with anti-HA antibodies. The immunocomplexes were separated by SDS-PAGE and analyzed by immunoblot using anti-Cas antibodies. Both wild-type (WT) PTP1B and enzymatically inactive CS-PTP1B bind to p130 Cas (Fig. 5). The binding of both WT-and CS-PTP1B to p130 Cas suggests that the interaction between PTP1B and p130 Cas is phosphotyrosine-independent. As in the in vitro experiments, PA-PTP1B fails to bind to p130 Cas , indicating that prolines 309 and 310 on PTP1B are required for binding to this protein in both cell types as well.
Expression of PTP1B in 3Y1-v-crk Cells Causes Tyrosine Dephosphorylation of p130 Cas -The ability of PTP1B to bind to p130 Cas suggests that this protein might be a physiological target for PTP1B. To examine this possibility, 3Y1-v-crk cells were transiently transfected with expression vectors bearing either (a) no insert, (b) WT-PTP, (c) CS-PTP, or (d) PA-PTP. Anti-HA immunoblot analysis indicates that all the PTP constructs were expressed equally (Fig. 6C). Cell lysates were immunoprecipitated with anti-p130 Cas antibodies, separated by 7% SDS-PAGE, and probed with antiphosphotyrosine antibodies to determine whether PTP1B expression affected p130 Cas tyrosine phosphorylation (Fig. 6A). The blots were then stripped and reprobed with anti-p130 Cas antibodies to ensure that the immunoprecipitates contained equal levels of this protein (Fig. 6B). The level of tyrosine-phosphorylated p130 Cas is substantially reduced (about a 3-4-fold decrease from five independent experiments, as assessed by densitometry) in 3Y1v-crk cells transfected with WT-PTP compared with those transfected with either inactive PTP (CS-PTP) or the p130 Cas binding mutant (PA-PTP). Although we cannot exclude an indirect effect of PTP1B on p130 Cas tyrosine phosphorylation levels (e.g. by inactivating a tyrosine kinase that acts on p130 Cas , such as Src or Fak), these data are consistent with a direct enzyme-substrate relationship between these two proteins.
Besides p130 Cas , three other prominent phosphotyrosyl proteins, which migrate at about 120, 90, and 65 kDa on SDS-PAGE, are apparent in 3Y1-v-crk cells. Two of these (pp120 and pp65) are also partially dephosphorylated in cells expressing WT-PTP but not PA-PTP. These proteins may represent additional binding partners or substrates for PTP1B in 3Y1-v-crk cells. Although we do not know the identity of these proteins, the 120-kDa band may represent an isoform of p130 Cas (28) or the related protein Hef1 (29). The ϳ90-kDa phosphotyrosyl protein is not affected by PTP1B expression, indicating that this phosphatase does not indiscriminately dephosphorylate all potential substrates within the cell.
To exclude the possibility that PA-PTP fails to affect the tyrosine phosphorylation of p130 Cas due to a reduction in overall phosphatase activity, we compared the enzymatic activity of WT-and PA-PTP. The activities of GST fusions with WT-and PA-PTP were measured against the phosphomonoester paranitrophenol phosphate (Fig. 6D). The WT-and PA-PTP mutant have similar activity profiles, indicating that this mutation does not overtly affect the ability of this enzyme to dephospho- FIG. 4. PTP1B binds to p130 Cas in vivo. 3Y1-v-crk cells were immunoprecipitated (IP) with anti-PTP1B and anti-p130 Cas antibodies or protein A beads alone (Con). The immunocomplexes were separated by 7% SDS-PAGE and immunoblotted with polyclonal anti-p130 Cas or monoclonal antiphosphotyrosine (Anti-PY) antibodies.
FIG. 5. PTP1B binds to p130 Cas in a phosphotyrosine-independent manner in vivo. 3Y1 and 3Y1-v-crk cells were transiently transfected with expression plasmids encoding wild-type, CS-(enzymatically inactive), and PA-PTP1B. The cell lysates were immunoprecipitated (IP) with anti-HA antibodies or preimmune serum (pre). The immunocomplexes were separated by 7% SDS-PAGE and immunoblotted with polyclonal anti-p130 Cas antibodies. PTP1B expression in these cells was monitored by immunoblot (lower panels).

FIG. 6. Overexpression of wild-type but not a proline mutant PTP1B causes dephosphorylation of p130 Cas in 3Y1-v-crk cells.
A, expression plasmids C (vector alone), WT, CS (enzymatic inactive), and PA (defective in p130 Cas SH3 binding) PTP1B were transiently transfected in 3Y1-v-crk cells. The cell lysates were immunoprecipitated with anti-p130 Cas antibodies. The immunocomplexes and cell lysates were separated by 7% SDS-PAGE and immunoblotted with antiphosphotyrosine antibodies (A), anti-p130 Cas antibodies (B), or anti-HA antibodies (C). p130 Cas -P is tyrosine-phosphorylated p130 Cas . D, GST-PTP1B and GST-PA-PTP1B were purified and immobilized on glutathione-agarose beads. Aliquots (about 5 l of beads) were analyzed for tyrosine phosphatase activity toward para-nitrophenol phosphate as described under "Experimental Procedures." ࡗ, wild-type PTP1B; Ⅺ, PA-PTP1B. rylate this substrate. Kinetic analysis reveals that the K m values for this substrate are 6.6 -6.8 mM for WT-PTP and 6.1-6.4 mM for PA-PTP (two determinations).
Subcellular Fractionation of PTP1B and p130 Cas -In view of the fact that PTP1B co-precipitates with p130 Cas in both 3Y1 and 3Y1-v-crk cells, we examined the subcellular distribution of p130 Cas and PTP1B to determine whether these proteins share a common compartment. Exponentially growing 3Y1 and 3Y1-v-crk cells were lysed in hypotonic buffer and fractionated into four fractions (nuclei, cytosol, low density membranes, and high density membranes) as described under "Experimental Procedures." The results from a typical experiment are shown in Table I. As expected, the cytosolic marker lactate dehydrogenase is enriched in the S100 (cytosol) fraction more than 10-fold relative to other three fractions. The P1 (nuclear and unbroken cells) fraction contains most of the DNA. The plasma membrane 5Ј-nucleotidase marker is approximately 3.5-fold higher in the 0.25-1.2 M sucrose fraction (low density membranes) than in the 1.2-2.0 M sucrose interface (high density membranes), whereas the ER membrane marker NADPH cytochrome c reductase has nearly the reverse ratio.
The distribution of PTP1B, p130 Cas , and Fak in these fractions was monitored by immunoblot. As in HeLa cells (9), PTP1B from both 3Y1 and in 3Y1-v-crk cells is highly enriched in the ER. p130 Cas distributes primarily to the soluble fraction (S100), but substantial amounts are also apparent in high density membranes (Fig. 7, lanes 4 and 9), which is enriched in the ER, as well as the P1 fraction, which contains nuclei and unbroken cells (Fig. 7, lanes 5 and 10). The percentage of p130 Cas in the high density membrane fraction is about 2-fold higher (representing about 40% of total p130 Cas , as assessed by densitometry) in 3Y1-v-crk cells than in the parental 3Y1 cells, suggesting that this protein redistributes when tyrosine-phosphorylated. Similar results were obtained using a different (monoclonal) anti-p130 Cas antibody (data not shown). These results indicate that p130 Cas exists in multiple cellular locations, one of which corresponds to the location of PTP1B. Fak, which is located primarily in the cytosol and in focal adhesions (30,31), distributes mostly to the S100 and P1 fractions, indicating that the high density membranes are not significantly contaminated with focal adhesions, a known site of p130 Cas localization (40). DISCUSSION In this study, we demonstrate that PTP1B binds to three SH3-containing proteins in vitro and at least one of these proteins, p130 Cas , in intact cells. This binding is mediated by the proline-rich domain between residues 301 and 315 on PTP1B, and point mutations within this region abolish binding to p130 Cas and also inhibit the ability of PTP1B expression to promote p130 Cas tyrosine dephosphorylation in cells. These data suggest that p130 Cas may be a physiological substrate for PTP1B.
To our knowledge, only two other PTPs have been shown to bind to an SH3-containing protein, and neither of these is mediated by a polyproline-SH3 interaction. PTP-PEST associates with the SH3-containing adaptor protein Shc, but this association appears to be mediated by an NPLH sequence in PTP-PEST to the phosphotyrosine binding domain on Shc (32). Similarly, the PTPs Shp-1 and -2 bind to Src, but the interaction is thought to be mediated by the SH2 domain on Shp to tyrosine autophosphorylation sites on Src (33,34). T-cell-PTP, which is closely related to PTP1B and is also located in the ER (35), has no proline-rich motifs and does not bind p130 Cas . 3 It is therefore possible that these two PTPs have distinct cellular functions based on the ability to interact with different substrates.
Many of the known posttranslation modifications of PTP1B occur at its C terminus. This region mediates binding to the ER (9), is the site of mitotic serine phosphorylations (10,11), is affected by alternative splicing (12), and is subject to proteolytic cleavage in response to certain stimuli (8). The prolinerich motif that is required for p130 Cas binding also resides in this region, just C-terminal to the PTP domain. The reported alternatively spliced and proteolytic cleavage forms of PTP1B do not alter this proline-rich motif between amino acids 301 and 315. We are currently testing whether the mitotic phosphorylation of PTP1B, which occurs at serines 352 and 386, affect its association with p130 Cas .
p130 Cas was initially described as a highly tyrosine-phosphorylated 130-kDa protein in v-crk-transformed (28,36,37) and v-src-transformed (38,39) fibroblasts. In REF52 cells, p130 Cas is concentrated at focal adhesions (40), in which it may act as a TABLE I Distribution of subcellular compartment markers after cell fractionation 3Y1-v-crk and 3Y1 cells were subjected to hypotonic lysis and sucrose gradient isopycnic centrifugation as described under "Experimental Procedures." Each subcellular fraction was analyzed for the distribution of markers specific for various organelles: lactate dehydrogenase (cytosol), 5Ј-nucleotidase (plasma membrane), NADPH cytochrome c reductase (ER), and DNA (nuclei). The specific activity of each marker in the total lysate is arbitrarily defined as 1.0. Values given for marker enzymes in each fraction represent specific activity relative to the total lysate (they are the average number of 3Y1-v-crk and 3Y1 cells). Units are protein (mg), marker enzyme activity (specific activity relative to total fraction), and DNA (percentage of total DNA in fraction/percentage of total protein in fraction). Total recovery for each assay is also displayed.  platform for the assembly of integrin-activated signaling molecules (28, 40 -44). In other cell types, p130 Cas has been reported to reside in the nucleus (38) and cytosol (27). Our data demonstrate that in 3Y1-v-crk cells, in which p130 Cas is heavily tyrosine-phosphorylated, a substantial fraction of this protein co-purifies with ER membranes. It is interesting to note that Sakai et al. (28) reported that p130 Cas moves from the cytosol to particulate fractions following tyrosine phosphorylation. As the fractionation procedure used by these authors did not distinguish plasma membrane from other membrane fractions, our results are consistent with their observations. Although most of the tyrosine phosphorylation-dependent signaling complexes described to date assemble at the plasma membrane, we observed almost no p130 Cas protein in immunoblots from the 0.25-1.2 M sucrose interface, which is enriched in the plasma membrane marker 5Ј-nucleotidase. One implication of these results is that p130 Cas may be recruited to a signaling complex that is located at the ER, rather than the plasma membrane.
In vertebrate cells, a number of signaling proteins have been localized to the ER. Among these are protein kinases, such as Ltk (45) and the and ␦ isoforms of protein kinase C (46,47), PTPs such as PTP1B (9) and T-cell-PTP (31), adaptor proteins such as Shc (48), and perhaps "docking" proteins such as p130 Cas (this study). Some or all of these proteins may regulate signaling from the ER to the nucleus (49), perhaps recapitulating the design of the signaling machinery used to transduce growth factor signals from the plasma membrane to the nucleus. Based on its effects on mitogenesis, PTP1B might be expected to play a negative role in such a signaling pathway. Alternatively, the ER might serve as a depot for certain signaling molecules, which can be recruited to other locations at appropriate times. For example, Shc redistributes from the ER to the plasma membrane and to endocytic structures following epidermal growth factor stimulation (48). A portion of p130 Cas exhibits the opposite behavior, apparently relocating from a soluble compartment to the ER when tyrosine-phosphorylated. In either event, the presence of tyrosine-phosphorylated signaling molecules in the ER membrane suggests that these may represent physiological substrates of PTP1B and/or the related enzyme T-cell-PTP.