Reconstitution of Regulated Phosphorylation of FcϵRI by a Lipid Raft-excluded Protein-tyrosine Phosphatase*

To examine the exquisite regulation of IgE-FcϵRI tyrosine phosphorylation by Lyn kinase that is stimulated by antigen-mediated cross-linking, we utilized co-expression of FcϵRI and Lyn in Chinese hamster ovary cells, which results in high basal levels of Lyn kinase activity and spontaneous phosphorylation of FcϵRI. We found that co-expression of a lipid raft-excluded transmembrane tyrosine phosphatase, PTPα, suppresses Lyn kinase activity and markedly reduces the level of spontaneous phosphorylation of FcϵRI, while facilitating its antigen-stimulated phosphorylation. Other tyrosine phosphatases, including SHP-1, CD45, and a lipid raft-preferring chimeric version of PTPα fail to reconstitute antigen-dependent FcϵRI phosphorylation. We concluded that both substrate specificity and submembrane location are critical to phosphatase-mediated regulation of Lyn kinase activity that supports activation of FcϵRI.

Fc⑀RI, the high affinity receptor for IgE, is a member of the family of multichain immune recognition receptors including T and B cell receptors for antigen, certain NK cell receptors, and other Fc receptors on various hematopoietic cells. All of these receptors contain at least one ITAM 1 or ITIM sequence in the cytoplasmic segment of one or more subunit (1,2). In every case, receptor-mediated signaling is initiated when tyrosine residues in these sequences are phosphorylated by a Src family tyrosine kinase. For Fc⑀RI in mast cells, ITAM phosphorylation of both its ␤ and ␥ 2 subunits is catalyzed principally by the Src kinase Lyn, and this is stimulated by antigen-mediated crosslinking of two or more IgE-receptor complexes (3,4). ITAM phosphorylation of this and other multichain immune recognition receptor family members recruits and activates the tyrosine kinase Syk (or Zap70 in T cells) leading to a cascade of downstream signaling and cell activation processes (5,6).
The mechanism by which IgE-Fc⑀RI cross-linking initiates receptor tyrosine phosphorylation has been extensively studied.
The ␤ subunit of Fc⑀RI has been shown to bind Lyn weakly in the absence of ITAM phosphorylation (7), and a transphosphorylation model has been developed from this and other studies (3). In addition, Fc⑀RI ␤ has been shown to serve an amplifying role for Fc⑀RI ␥ 2 ITAM phosphorylation and consequent Syk activation, most probably by the binding of the SH2 domain of Lyn to phosphorylated Fc⑀RI ␤ (4). However, Fc⑀RI ␤ is dispensable for signal initiation and downstream cell activation (8,9), and a role has been demonstrated for ordered regions of the plasma membrane, commonly called lipid rafts, in facilitating signal initiation by this receptor (10,11) and other multichain immune recognition receptor family members (12).
To understand the role of lipid rafts in Fc⑀RI phosphorylation by Lyn, we investigated the effects of this ordered lipid environment on Lyn tyrosine kinase activity in RBL mast cells. We found that rafts protect the active site tyrosine residue of Lyn from dephosphorylation, thereby enhancing the specific activity of Lyn kinase in lipid rafts relative to that in more disordered regions of the plasma membrane (13). As crosslinking of IgE-Fc⑀RI causes association of these receptors with lipids rafts (10,14), we postulated that the raft environment serves principally to permit receptor phosphorylation by active Lyn in the same environment and to protect these proteins from dephosphorylation by a phosphatase that is excluded from lipid rafts (13).
To test this hypothesis, we utilized Fc⑀RI stably expressed in a CHO cell line that exhibits an unusually high level of basal (unstimulated) Fc⑀RI tyrosine phosphorylation because of coexpressed Lyn (15). We compared the capacity of different tyrosine phosphatases to regulate both basal and stimulated Fc⑀RI tyrosine phosphorylation in this situation. We find that the expression of the transmembrane phosphatase PTP␣ is highly effective in this role. Furthermore, we showed that the conversion of PTP␣ to a lipid-anchored, raft-preferring phosphatase abolishes its capacity to facilitate antigen-stimulated Fc⑀RI tyrosine phosphorylation. Our results are consistent with a general role for protein segregation by lipid rafts in the regulation of enzymatic and other signaling processes.

EXPERIMENTAL PROCEDURES
DNA Constructs-Lyn was amplified by PCR from Lyn-pcDM8, provided by Dr. Henry Metzger (National Institutes of Health, Bethesda, MD) with primers (forward sequence) 5Ј-TATTGTGGATCC-GCCACCATGGGATGTATTAAATCAAA-3Ј and (reverse sequence) 5Ј-ATTGTTCTCGAGCTATGGCTGCTGCTGATACTG-3Ј and subcloned into pcDNA3 with BamHI and XhoI. Tyr to Phe Lyn mutants were generated using the QuikChange site-directed mutagenesis kit (Stratagene). PTP␣-pcDNA3 was supplied by Dr. David Shalloway (Cornell University, Ithaca, NY) and SHP-1-pRc-CMV by Dr. Katherine Siminovitch (University of Toronto, Toronto, ON). PM-PTP␣ is a construct of the first 21 amino acids of Lyn attached to the cytoplasmic portion of PTP␣ with a 2 amino acid linker (KL), created using PCR to generate a fragment containing the first 63 nucleotides of Lyn with a 5Ј BamHI and a 3Ј-Hind III site using the same forward primer as above * This work was supported by Grants R01-AI22449 and T32-GM07273 from the National Institutes of Health (to R. M. Y.). 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.
Transfections-CHO cells stably transfected with Fc⑀RI (CHO-Fc⑀RI cells) (16) were transiently transfected in 6-well plates using Geneporter (GTS, San Diego, CA). For each well, 10 l of Geneporter and 1 g of Lyn-pcDNA3 and 1 g of phosphatase construct DNA (or 0.5 g of PM-PTP␣ and 0.5 g of pcDNA3), or equivalent amounts of empty pcDNA3 vector for controls, were added to 100 l of Opti-MEM (Invitrogen), incubated for 15 min, and added to cells plated over with 1 ml of Opti-MEM. The mixture was then incubated for 6 h at 37°C with 5% CO 2 and, finally, washed and plated with CHO medium containing 1 g/ml mouse IgE.
Antibodies-Mouse IgE specific for DNP-BSA was prepared as previously described (43). Antibodies used include mouse anti-Lyn H6 mAb and polyclonal rabbit anti-Lyn 44 (Santa Cruz Biotechnology, Santa Cruz, CA), mouse anti-phospho-tyrosine mAb (4G10) and rabbit anti-SHP-1 (Upstate Biotechnology, Lake Placid, NY), mouse anti-CD45 mAb (GAP8.3 (ATCC)), rabbit anti-Tyr(P)-508 (anti-phospho-Lyn (Tyr-507)) (Cell Signaling Technologies, Beverly, MA), rabbit anti-IgE (17), and fluorescein isothiocyanate-goat anti-mouse IgG2a (Southern Biotechnology, Birmingham, AL) Examination of Phosphorylation-To examine whole cell tyrosine phosphorylation, transfected CHO-Fc⑀RI cells or RBL-2H3 cells (18) were suspended at 2 ϫ 10 6 cells/ml in buffered saline solution (11) containing bovine serum albumin (Sigma) and equilibrated to 37°C. Cells were lysed with non-reducing SDS sample buffer before and after 2 or 5 min of stimulation with 0.9 g/ml DNP-BSA (14). Samples were separated by electrophoresis on 12% acrylamide SDS gels, and quantitative immunoblotting was performed as previously described (13). For immunoprecipitation of Fc⑀RI, cells were suspended at 8 ϫ 10 6 cells/ml and lysed in radioimmune precipitation assay buffer (0.5% TX-100, 0.5% deoxycholate, 0.05% SDS), which has been shown to solubilize all membrane components (13). Anti-IgE immunoprecipitations used 10 g of anti-IgE for 0.5 ml of cell lysate, and immune complexes were pulled down with 35 l of protein A beads (Pierce). A determination of Lyn in vitro kinase activity from Lyn was immunoprecipitated with mouse anti-Lyn mAb from radioimmune precipitation assay buffer or TX-100 post nuclear supernatants of CHO-Fc⑀RI cells using ␣ casein as the substrate was carried out as described previously (13). The measurement of in vitro phosphatase activity for expressed PTP␣ and PM-PTP␣, anti-hemagglutinin immunoprecipitated from CHO cell lysates, was carried out as described previously (41). (15) reported high levels of spontaneous Fc⑀RI tyrosine phosphorylation in CHO cells stably expressing Fc⑀RI and the Src family kinase Lyn. In their study, spontaneous Fc⑀RI ␤ subunit phosphorylation was ϳ17 times greater than in RBL-2H3 mast cells. Antigen stimulation only modestly increased Fc⑀RI phosphorylation in CHO cells. Similarly, we find high levels of basal tyrosine phosphorylation on Fc⑀RI stably expressed in CHO cells when Lyn kinase is transiently expressed. As shown in Fig. 1A (left panel) anti-phosphotyrosine Western blotting of whole cell lysates (WCL) reveals high levels of spontaneous Fc⑀RI ␤ and ␥ 2 subunit phosphorylation for these transfected CHO cells. Additionally, there are elevated levels of phosphorylation of Lyn and several other proteins. In contrast, the same number of RBL cell equivalents exhibit no detectable tyrosine phosphorylation in the absence of antigen stimulation, but there is a robust increase in phosphorylation on multiple proteins, including Fc⑀RI, in response to antigen. We verified that the bands of ϳ35 and ϳ25-30 kDa on anti-phosphotyrosine Western blots of CHO-Fc⑀RI whole cell lysates are Fc⑀RI ␤ and ␥ 2 , respectively, by confirming their co-migration with those bands precipitated by anti-IgE antibody (Fig. 1A, right panel), and by selective immunodepletion of these bands from whole cell lysates (data not shown). Note that the lower level of stimulated phosphorylation of Fc⑀RI ␤ and ␥ 2 in CHO-Fc⑀RI cells, compared with that in RBL cells, is expected because of lower levels of Fc⑀RI expression in the CHO cells (Ref. 15 and data not shown).

CHO Cells Expressing Fc⑀RI and Lyn Exhibit Poorly Regulated Lyn Kinase Activity-Torigoe and Metzger
To investigate the basis for the high basal tyrosine phosphorylation observed in Lyn-transfected CHO cells, we more closely examined Lyn kinase. As a first step we immunoprecipitated Lyn kinase from these cells and from RBL cells and compared their specific activities using ␣ casein as described previously (13). Fig. 1B shows that Lyn kinase expressed in CHO cells has an ϳ3-fold higher specific activity than Lyn from RBL cells, consistent with the higher levels of basal phosphorylation in the CHO-Fc⑀RI cells.
Lyn kinase activity is positively regulated by phosphorylation near its active site, Tyr-397 (Lyn A notation), and negatively regulated by phosphorylation on its C-terminal tail, Tyr-508. Our previous study in RBL cells showed that Lyn activity is largely determined by phosphorylation at Tyr-397, whereas C-terminal phosphorylation is very low in resting cells (13). To assess the relative phosphorylation of these sites for Lyn in the CHO cells, we compared various Lyn mutants, using both 4G10 anti-phosphotyrosine and an antibody specific for the phosphorylated C terminus of Lyn (anti-Tyr(P)-508) by Western blot analysis. As shown in Fig. 1C, we found that wt Lyn expressed in CHO cells has several times more net tyrosine phosphorylation than Lyn from RBL-2H3 cells (top panel), whereas C- terminal phosphorylation of Lyn from these two different cell types is more similar (middle panel). The mutation of Tyr-397 to Phe results in a dramatic reduction of tyrosine phosphorylation detected by 4G10 on Lyn expressed in CHO cells (Fig. 1C, top panel). Conversely, the mutation of Tyr-508 to Phe results in only a modest reduction in 4G10-detected phosphorylation (Fig. 1C, top panel) and complete elimination of the anti-Tyr(P)-508 phosphorylation (middle panel). Taken together, these results indicate that Lyn expressed in CHO cells is hyperphosphorylated at Tyr-397, and this is likely to account for its high specific activity.
Co-expression of Protein-tyrosine Phosphatase PTP␣ with Lyn in CHO Cells Reconstitutes Antigen-dependent Fc⑀RI Phosphorylation-The large amount of Lyn-dependent basal phosphorylation in CHO cells suggests that these cells lack sufficient phosphatase activity to regulate exogenously expressed Lyn kinase. PTP␣ is a ubiquitously expressed transmembrane tyrosine phosphatase that has been found to regulate Src kinase activity (19). Fig. 2A shows that transient co-expression of this phosphatase with Lyn in CHO-Fc⑀RI cells dramatically reduces basal levels of Lyn phosphorylation and the accompanying spontaneous phosphorylation of Fc⑀RI ␤ and ␥ 2 subunits. Notably, co-expression of PTP␣ permits antigen-stimulated tyrosine phosphorylation of both receptor subunits, which is maximal between 2 and 5 min and declines thereafter, as observed for stimulation in RBL-2H3 cells (20). As reported by Vonakis et al. (16), CHO-Fc⑀RI cells have a poorly detectable endogenous Srcfamily kinase activity, which mediates variable amounts of Fc⑀RI ␤ and ␥ 2 phosphorylation in response to antigen, and this is apparent in the vector control in Fig. 2A, lanes 1 and 2. The quantification of basal and stimulated Fc⑀RI ␤ and ␥ 2 phosphorylation from multiple experiments is summarized in Fig. 2B. For Fc⑀RI ␤, Lyn expression causes a significant increase in both basal and stimulated phosphorylation. Co-expression of PTP␣ reduces basal phosphorylation of Fc⑀RI ␤ to very low levels while maintaining elevated levels of stimulated phosphorylation. For Fc⑀RI ␥ 2 , Lyn expression causes substantial increases in basal and stimulated phosphorylation. Coexpression of PTP␣ reduces basal phosphorylation by ϳ10-fold while reducing stimulated phosphorylation by Ͻ3-fold. These results demonstrate that co-expression of PTP␣ with Lyn in CHO-Fc⑀RI cells controls spontaneous phosphorylation while permitting antigen-stimulated phosphorylation of Fc⑀RI by Lyn.
Reconstitution of Signaling in CHO Cells Is Dependent on the Location of PTP Activity-We showed in Fig. 1 that Lyn expressed in CHO cells is highly active, most likely because it is strongly phosphorylated at its active site. Likewise, the activity of endogenous Lyn in RBL cells is controlled by phosphoryla-tion at the active site (13). In that study, we found that Lynspecific activity is substantially higher in lipid rafts than in disordered regions of the plasma membrane. We hypothesized that the rafts exclude transmembrane phosphatases that dephosphorylate and thereby inactivate Lyn located outside of lipid rafts. In Fig. 3A, a sucrose gradient analysis of TX-100lysed CHO-Fc⑀RI cells transfected with PTP␣ shows that greater than 99% of this transmembrane protein is found in the soluble, non-raft portion of the gradient. To test the role of raft exclusion in the reconstitution of stimulated phosphorylation, we created a raft-associating form of PTP␣, named PM-PTP␣, by fusing the cytoplasmic portion of PTP␣ with the N-terminal 21 amino acids of Lyn, which target Lyn to rafts by virtue of post-translationally added palmitoylation and myristoylation (21). PM-PTP␣ and Lyn have a similar localization to lipid rafts in transfected CHO cells, with Ͼ60% of these proteins in the low density raft fractions (Fig. 3A). To evaluate the role of this differential raft targeting for PTP␣ and PM-PTP␣, we compared the effects of similar co-expression of each phosphatase with Lyn. As shown in Fig. 3B (top panel), co-expression of PM-PTP␣ leads to a further decrease in Lyn phosphorylation than for PTP␣ co-expression. Importantly, little antigen-stimulated phosphorylation of Fc⑀RI is detected in the presence of PM-PTP␣ in this and in three other experiments of similar design. In contrast, PTP␣ suppresses spontaneous Fc⑀RI phosphorylation but allows stimulated Fc⑀RI phosphorylation, similar to results in Fig. 2.
In these experiments, quantitative analysis of anti-PTP␣ blots indicates that PM-PTP␣ was expressed at 1.9 Ϯ 0.3-fold (S.E., n ϭ 6) higher levels than PTP␣. To attempt to evaluate whether this modest difference in expression could account for the more profound suppression of stimulated Fc⑀RI phosphorylation because of PM-PTP␣, we compared in vitro phosphatase activities of PTP␣ and PM-PTP␣ immunoprecipitated from equal numbers of transfected CHO cells. In two separate experiments, the ratios of PM-PTP␣/PTP␣ phosphatase activities were 1.22 and 0.67. These results suggest that net phosphatase activity/cell equivalent due to PM-PTP␣ is not significantly different from that due to PTP␣, and this is not likely to account for their qualitatively different effects on stimulated Fc⑀RI phosphorylation.
Lyn expressed in CHO cells is highly active (Fig. 1B), and we next examined whether the effects of PTP␣ on whole cell phosphorylation are mediated through control of Lyn activity. Fig.  3C compares the specific activities of Lyn solubilized and immunoprecipitated from CHO cells with or without PTP␣ or PM-PTP␣ co-expressed. PTP␣ causes a substantial decrease in specific activity of Lyn. This reduction in Lyn kinase activity may be sufficient to account for the control of basal Fc⑀RI phosphorylation in CHO cells expressing exogenous Lyn. The specific activity of Lyn solubilized from CHO-Fc⑀RI cells coexpressing Lyn and PM-PTP␣ is more severely reduced than that with PTP␣ (Fig. 3C). This further reduction in activity correlates with a reduction of Lyn activity regardless of its raft environment in PM-PTP␣ transfected cells, consistent with the dephosphorylation of Lyn at its active site tyrosine by PM-PTP␣ in both the raft and non-raft environment. Because PM-PTP␣ co-localizes with Lyn in lipid rafts, it has continual access to this substrate, whereas transmembrane PTP␣ is restricted to non-raft portions of the plasma membrane and is unable to inactivate raft-associated Lyn.
Effects of Other PTPs on Fc⑀RI Phosphorylation by Lyn in CHO Cells-To ascertain the selectivity of PTP␣ regulation of Fc⑀RI phosphorylation by Lyn, we compared its effects to those of two other tyrosine phosphatases, SHP-1 and CD45. SHP-1 is a hematopoietic phosphatase shown to negatively regulate multichain immune recognition receptor signaling, including Fc⑀RI signaling in mast cells (22,23). It is located primarily in the cytoplasm, but it is recruited to the plasma membrane during antigen stimulation (22). Fig. 4A shows a representative experiment in which co-expression of SHP-1 with Lyn in CHO-Fc⑀RI cells is compared with co-expression of PTP␣. SHP-1 efficiently inhibits Lyn autophosphorylation and suppresses Fc⑀RI stimulation, both before and after stimulation, reminiscent of the effects of PM-PTP␣ (Fig. 3B). Active SHP-1 binds to protein substrates via its SH2 domains and is probably not restricted from Lyn in lipid rafts. In contrast, expression of CD45 in CHO cells causes small increases in tyrosine phosphorylation levels of multiple cellular proteins, including Lyn and Fc⑀RI, both before and after stimulation (Fig. 4B). Because CD45 is difficult to detect by immunoblotting, we used flow cytometry analysis that showed relatively low levels of CD45 in transfected CHO cells compared with the highly abundant expression of this phosphatase on Jurkat T cells (data not shown). Like PTP␣, CD45 is a type 1 transmembrane protein that is excluded from lipid rafts (24), so it is unlikely that its plasma membrane location distinguishes its effects from those of PTP␣. CD45 is often implicated in the positive regulation of immune cell signaling by dephosphorylation of the C-terminal negative regulatory tyrosine of Src family kinases such as Lck (25), and our results are consistent with this role.
To investigate the mechanism by which CD45 expression leads to an increase in basal phosphorylation in CHO-Fc⑀RI cells we compared tyrosine phosphorylation of Lyn expressed in the presence or absence of PTP␣, SHP-1, or CD45. Fig. 4C summarizes the quantitative comparisons for both total tyrosine phosphorylation and Tyr-508 phosphorylation normalized separately to equivalent amounts of Lyn. We found that PTP␣ and SHP-1 both dephosphorylate Lyn exclusively at Tyr-397 and that SHP-1 expression actually leads to an increase in Tyr-508 phosphorylation through an unknown mechanism. In contrast, CD45 expression results in a small net dephosphorylation of Lyn at Tyr-508 that can account for the modest increase in cellular tyrosine phosphorylation we observe in CD45 expressing CHO cells (Fig. 4B). Thus, among these three phosphatases, only PTP␣ has the necessary combination of plasma membrane location (exclusion from lipid rafts) and substrate specificity (Lyn Tyr-397) that allow it to reconstitute effective regulation of Lyn kinase activity and the early events of Fc⑀RI signaling in CHO cells.

DISCUSSION
Our present findings showed that a transmembrane tyrosine phosphatase, PTP␣, plays an essential role in the regulation of Fc⑀RI phosphorylation by Lyn when this kinase and receptor are co-expressed in CHO cells. Lyn expressed in CHO cells has a substantially higher specific activity than Lyn in RBL cells because of high levels of phosphorylation at its active site Tyr-397 (Fig. 1). Enhanced Lyn kinase activity in the CHO cells leads to spontaneous phosphorylation of multiple cellular proteins, including Fc⑀RI (Fig. 1). Co-expression of PTP␣ results in marked dephosphorylation of Lyn in these cells, thereby suppressing basal phosphorylation of several substrates while enabling stimulated phosphorylation of Fc⑀RI (Fig. 2).
Our previous studies reveal that active Lyn in RBL mast cells is largely sequestered in lipid rafts, and this Lyn is active because of phosphorylation at its active site (13). In light of this, we investigated whether segregation of Lyn from a transmembrane phosphatase by lipid rafts is important for the regulation of Lyn activity. CHO-Fc⑀RI cells provide a useful vehicle to test this possibility, because Lyn in these cells is less effectively regulated than in RBL cells. We determined that PTP␣ is excluded from lipid rafts, and thus is unable to gain access to and inactivate Lyn within the ordered raft environment. In clear contrast, targeting chimeric PM-PTP␣ to lipid rafts abolishes stimulation-dependent Fc⑀RI phosphorylation by inactivating Lyn both within and outside of the raft environment (Fig. 3). These results support a model in which lipid rafts act both to protect Lyn from phosphatases, which keeps raft-associated Lyn active, and to inhibit productive interactions of Fc⑀RI with active Lyn until cross-linking by multivalent antigen brings them together by driving raft association of Fc⑀RI. Cross-linking Fc⑀RI in this manner may also serve to stabilize ordered membrane domains containing active Lyn.
Metzger and colleagues (26,27) have argued against a role for lipid rafts in protecting Fc⑀RI from dephosphorylation in studies that utilized the reversal of antigen cross-linking by a monovalent hapten to observe rapid dephosphorylation of Fc⑀RI and other phosphorylation substrates. However, the reversal of cross-linking causes a rapid reversal of Fc⑀RI lateral immobilization (28) and the loss of its lipid raft association (10), so that the dephosphorylation of Fc⑀RI due to hapten addition probably occurs outside of the lipid rafts and leads to dephosphorylation of more downstream substrates. Furthermore, the role of lipid rafts in protecting Lyn kinase from dephosphorylation was not evaluated by Metzger and colleagues (26,27), and this is probably at least as important as protection of Fc⑀RI from dephosphorylation based on the current results.
Torigoe and Metzger (15) originally described spontaneous phosphorylation of stably expressed Fc⑀RI by stably co-ex-pressed Lyn in CHO cells. These authors attributed this to differential glycosylation of Fc⑀RI between CHO and RBL cells after detecting no other differences in physical and biochemical properties of Fc⑀RI or Lyn in these two cell types. Using a different assay for Lyn kinase activity, we found substantial differences in immunoprecipitated Lyn activity (Fig. 1B). In addition, we have characterized tyrosine phosphorylation in whole cell lysates, which shows high basal phosphorylation of multiple proteins independent of Fc⑀RI expression (Figs. 1A, 2A, and data not shown).
Previous studies on PTP␣ largely focused on its positive regulation of Src kinase activity in fibroblasts because of dephosphorylation of Src C-terminal tyrosine (19). Our present findings demonstrating negative regulation of Lyn kinase activity by dephosphorylation of its active site tyrosine were unanticipated by these studies. Ng et al. (29) reported little phosphatase activity of PTP␣ relative to CD45 toward a phosphorylated peptide mimic of the C terminus of Fyn in vitro. They found that PTP␣ had a lower K m toward phosphorylated peptide mimics of the active site versus the C terminus of Src in vitro, suggesting that PTP␣ is not generally restricted to the dephosphorylation of the negative regulatory site. Our preliminary results indicated that PTP⑀, which is highly homologous to PTP␣, also reconstitutes antigen-stimulated signaling in CHO cells by negatively regulating Lyn (data not shown). Other transmembrane phosphatases may likewise negatively regulate Lyn kinase outside of lipid rafts in a redundant fashion, suppressing spontaneous signaling and complementing the negative role of C-terminal tyrosine phosphorylation in different cell types and for different Src family members (30).
SHP-1 and CD45, two common hematopoietic phosphatases, do not reconstitute regulated Fc⑀RI phosphorylation in CHO cells (Fig. 4). Both phosphatases are often implicated for regulating signaling through Fc⑀RI and other members of the multichain immune recognition receptor family, and SHP-1 is reported to negatively regulate Fc⑀RI signaling in RBL mast cells (23). Similar to PTP␣, we found that SHP-1 dephosphorylates the active site tyrosine of Lyn expressed in CHO-Fc⑀RI cells (Fig. 4) consistent with results previously reported in B cells  (31). Unlike PTP␣, SHP-1 is a cytosolic phosphatase and is not excluded from lipid rafts, allowing it to dephosphorylate Lyn regardless of its membrane environment, thus preventing stimulated Fc⑀RI phosphorylation in the CHO-Fc⑀RI cells, similar to results with raft-targeted PM-PTP␣ (Fig. 3). These results are consistent with a model in which SHP-1 generally suppresses spontaneous receptor phosphorylation by Lyn in resting cells (32). Following cell activation, SHP-1 is recruited to phosphorylated ITIM sequences by its SH2 domains, where it has a more targeted role in dephosphorylating ITAMs on Fc⑀RI or at the active site of Lyn (33). Its strong suppression of stimulated Fc⑀RI phosphorylation in the present experiments may be the result of relatively high expression levels, but reduced expression of this phosphatase also did not confer a differential effect on stimulated versus unstimulated Fc⑀RI phosphorylation as seen with PTP␣ (data not shown).
In contrast, CD45 is largely excluded from lipid rafts (24), like PTP␣, but its expression in CHO cells causes small increases in spontaneous and stimulated Fc⑀RI phosphorylation (Fig. 4). We found that CD45 preferentially dephosphorylates Lyn at its negative regulatory site, Tyr-508, which likely leads to an increase in Lyn activity, causing enhanced Fc⑀RI phosphorylation. This suggests that CD45 is primarily a positive regulator of Lyn activity, in agreement with evidence showing that high levels of CD45 expression in RBL cells leads to increased degranulation following antigen stimulation (34). Previous studies showed that CD45 is not required for Fc⑀RI signaling in RBL mast cells (35,36), and our results are consistent with these findings.
In other studies, the expression of lipid raft-targeted chimeric phosphatases from either SHP-1 or CD45 were used to investigate the role of tyrosine phosphatase segregation from lipid rafts in T cell receptor signaling (37)(38)(39). All of these studies reported inhibition of downstream signaling because of the raft-targeted chimeric phosphatases, and, in one study, an inhibition of stimulated receptor phosphorylation was observed (39). These studies, together with our results, suggest that the exclusion of membrane-associated tyrosine phosphatases from the ordered lipid environment of lipid rafts may be a general mechanism for the regulation of immunoreceptor signaling.
Scharenberg et al. (40) showed that regulated Fc⑀RI phosphorylation could be reconstituted in NIH3T3 cells upon coexpression of Lyn. These authors concluded that non-hematopoietic phosphatases present in NIH3T3 cells sufficiently regulate Fc⑀RI signaling. PTP␣ is ubiquitously expressed and may be relevant to this regulation in NIH3T3 cells (41). We have detected endogenous PTP␣ in both the CHO-Fc⑀RI and RBL-2H3 cells by immunoblotting; its endogenous expression in the CHO-Fc⑀RI cells is generally less than that in RBL cells or when this phosphatase is transfected into CHO cells (Fig.  3B, and data not shown). Although it is possible that transfected PTP␣ is expressed to levels that overcome normal limits to its substrate specificity, we found that transfected PTP␣ functions consistently in reconstituting regulated Fc⑀RI phosphorylation over a wide range of expression levels in the CHO-Fc⑀RI cells, including levels approaching those of endogenous PTP␣ in these cells (data not shown). Regulation of PTP␣ activity by dimerization (42) or other mechanisms may also influence its functional capacity in addition to effects of differential expression.
The role of PTP␣, or related transmembrane tyrosine phosphatases, in Fc⑀RI signaling in mast cells remains to be established. From this study it is clear that PTP␣ can regulate Lyn kinase in a manner that enables robust tyrosine phosphorylation of Fc⑀RI in response to cross-linking by an antigen. Our results demonstrate that phosphatase location in the plasma membrane and its substrate specificity are both important to the control of signal initiation via Fc⑀RI. Unlike SHP-1 and CD45, PTP␣ fulfills both criteria, but it may be only one of several such phosphatases responsible for regulating Lyn kinase activity in mast cells and other hematopoietic cells in which Lyn plays important physiological roles.