Receptor-type Protein-tyrosine Phosphatase-κ Regulates Epidermal Growth Factor Receptor Function*

Epidermal growth factor receptor (EGFR), the prototypic receptor protein tyrosine kinase, is a major regulator of growth and survival for many epithelial cell types. We report here that receptor-type protein-tyrosine phosphatase-κ (RPTP-κ) dephosphorylates EGFR and thereby regulates its function in human keratinocytes. Protein-tyrosine phosphatase (PTP) inhibitors induced EGFR tyrosine phosphorylation in intact primary human keratinocytes and cell-free membrane preparations. Five highly expressed RPTPs (RPTP-β, δ, κ, μ, and ξ) were functionally analyzed in a Chinese hamster ovary (CHO) cell-based expression system. Full-length human EGFR expressed in CHO cells, which lack endogenous EGFR, displayed high basal (i.e. in the absence of ligand) tyrosine phosphorylation. Co-expression of RPTP-κ, but not other RPTPs, specifically reduced basal EGFR tyrosine phosphorylation. RPTP-κ also reduced epidermal growth factor-dependent EGFR tyrosine phosphorylation in CHO cells. Purified RPTP-κ preferentially dephosphorylated EGFR tyrosines 1068 and 1173 in vitro. Overexpression of wild-type or catalytically inactive RPTP-κ reduced or enhanced, respectively, basal and EGF-induced EGFR tyrosine phosphorylation in human keratinocytes. Furthermore, siRNA-mediated knockdown of RPTP-κ increased basal and EGF-stimulated EGFR tyrosine phosphorylation and augmented downstream Erk activation in human keratinocytes. RPTP-κ levels increased in keratinocytes as cells reached confluency, and overexpression of RPTP-κ in subconfluent keratinocytes reduced keratinocyte proliferation. Taken together, the above data indicate that RPTP-κ is a key regulator of EGFR tyrosine phosphorylation and function in human keratinocytes.

PTPs and active PTKs has recently been estimated to be very similar (3,4). Emerging evidence indicates that, depending on the particular pathway, protein tyrosine dephosphorylation can be of equal or greater importance than protein tyrosine phosphorylation for the regulation of cellular function (5).
Epidermal growth factor receptor (EGFR, ErbB1) belongs to the receptor protein-tyrosine kinase (RPTK) superfamily. It is composed of an extracellular ligand binding domain, a single transmembrane domain, and an intracellular domain possessing PTK activity. Ligand binding to the extracellular domain of EGFR stabilizes homodimerization and heterodimerization with other ErbB members, which promotes trans tyrosine phosphorylation of the intracellular C-terminal domain. EGFR activation is synonymous with increased phosphorylation of specific tyrosine residues within its intracellular C-terminal domain. These phosphorylated tyrosines function as docking sites for a variety of signaling molecules that regulate membrane-proximal steps of signal transduction cascades that ultimately bring about cellular responses to EGFR ligands (6). Recent data suggest that EGFR not only participates in cognate ligand-induced signal transduction pathways but also plays important roles in diverse signal transduction pathways initiated by G protein-coupled receptors, cytokine receptors, integrins, ion channels, and stress responses (7)(8)(9).
Aberrant regulation of EGFR has been shown to promote multiple tumorigenic processes by stimulating proliferation, angiogenesis, and metastasis (10). EGFR and/or its ligands have a critical role in most common human epithelial cancers and many different types of solid tumors (11). The central role of EGFR in diverse signal transduction pathways dictates that its tyrosine phosphorylation must be strictly regulated. One potential mechanism for such regulation is through PTPcatalyzed dephosphorylation.
The subfamily of "classical," strictly tyrosine-specific PTPs contains 38 members, 21 of which are transmembrane receptor types and 17 of which are intracellular, non-receptor types (4). All classical PTPs contain a signature motif (HVCXXXXXR(S/T)) within a catalytic domain of 250 amino acid residues (12). The cysteine residue in the PTP signature motif is absolutely required for catalytic activity (13). The receptor-type PTPs (RPTPs) are integral membrane proteins composed of extracellular adhesion molecule-like domains, a single transmembrane domain, and a cytoplasmic domain containing one or two catalytic domains.
Reduced phosphorylation of EGFR has been associated with several different PTP activities (14 -18). However, identification of RPTP(s) that directly dephosphorylate(s), and thereby regulate(s), EGFR function is lacking. Using an expression strategy in EGFR-lacking Chinese hamster ovary (CHO) cells, we have identified RPTP-as a specific EGFR PTP. We have further demonstrated that RPTP-regulates both basal and ligand-induced EGFR tyrosine phosphorylation and function.
Cell Culture-Adult human primary keratinocytes were expanded in modified MCDB153 medium (EpiLife, Cascade Biologics, Inc.). CHO cells were cultured in Ham's F-12 medium with 1.5 g/liter sodium bicarbonate supplemented with 10% fetal bovine serum.
Preparation of Keratinocyte Membranes and EGFR Activation Assay-Human keratinocytes were washed twice with ice-cold hypotonic buffer (20 mM Tris-HCl, pH 7.6 with 10 mM NaCl) supplemented with 10 g/ml aprotinin, 10 g/ml leupeptin, 10 g/ml pepstatin A, and 1 mM phenylmethylsulfonyl fluoride. Cells were disrupted in a Dounce homogenizer. Lysates were centrifuged at 500 ϫ g for 10 min, and the supernatant was centrifuged at 20,000 ϫ g for 30 min. Membranes were extracted with 0.5 M NaCl to remove loosely associated proteins. Membrane suspension was then supplemented with 100 M ATP, 0.2% ␤-mercaptoethanol and 30 mM MgCl 2 . Incubation with EGF or phosphatase inhibitors was performed at room temperature. SDS sample buffer was added to stop the reactions. EGFR tyrosine phosphorylation was analyzed by Western blot using phospho-EGFR (pY-1068) antibody.
Generation of RPTP-Polyclonal Antibody-A peptide with a unique sequence derived from the intracellular domain (RGHNESKADCLD-MDP KAPQH) with predicted high probability of surface exposure and high antigenic index was synthesized, conjugated to keyhole limpet hemocyanin, and injected into New Zealand White rabbits (Bethyl Laboratories, Inc., Montgomery, TX). After two booster injections, anti-RPTP-antibody was affinity-purified from hyperimmune serum. The antibody was tested for its performance in enzyme-linked immunosorbent assay, Western blot, and immunoprecipitation.
In Vitro Dephosphorylation of Purified EGFR-Purified full-length EGFR was purchased from BIOMOL (Plymouth Meeting, PA). EGFR was tyrosine-phosphorylated according to the manufacturer's protocol and used as substrate for the RPTPintracellular region GST fusion protein. Dephosphorylation reactions were terminated by the addition of SDS sample loading buffer, and the level of EGFR tyrosine phosphorylation was measured by Western analysis probed with phospho-EGFR antibody.
Western Analysis Detection and Quantitation-Western blots were developed and quantified using an enhanced chemifluorescence detection system (Amersham Biosciences). Immunoreactive fluorescent protein bands were detected by the STORM phosphorimaging device using ImageQuant software (Molecular Dynamics, Sunnyvale, CA). Sample loads, antibody concentration, and incubation times were adjusted to yield fluorescent signals within the linear range of detection.
Transient Transfection of CHO Cells-Mammalian expression vectors for EGFR (pRK5 EGFR) and RPTP (pShuttle RPTP) coding sequences were transiently transfected into CHO cells using Lipofectamine 2000 (Invitrogen). Expression of EGFR and RPTP mRNA and protein was confirmed by real-time reverse transcription (RT)-PCR (7700 Taqman Sequence Detector, Applied Biosystems, Foster City, CA) and Western analysis, respectively, 24 h after transfection.
Association of EGFR with Wild-type and Trapping Mutant RPTP--Full-length EGFR and His 6 -tagged full-length wild-type or trapping mutant (D1051A) RPTP-were co-transfected into CHO cells. One day after transfection, the cells were treated with 50 ng/ml EGF for 10 min at 37°C and subsequently lysed in TGH buffer (50 mM Hepes, pH 7.2, 20 mM NaCl, 10% glycerol, and 1% Triton X-100), and RPTP-protein complexes were purified using Dynabeads TALON (Dynal Biotech, Oslo, Norway) and analyzed by Western blot.
Adeno-X Expression Vector Construction and Adenovirus Production-Wild-type and catalytically inactive (2C/S) human RPTP-expression vectors were generated using the Adeno-X expression system (Clontech Laboratories, Inc., Palo Alto, CA). To facilitate detection of expressed RPTP-, a polyhistidine (His 6 ) tag was inserted into the C terminus of RPTP-. HEK293 cells were used for adenovirus production and purification.
Generation of RPTP--GST Fusion Protein-cDNA encoding the intracellular region of human RPTP-(RPTP--IC, corresponding to residues 754 -1414) was cloned into the XhoI (5Ј) and NotI (3Ј) site of GST fusion protein expression vector pGEX-6P-3. GST-RPTP--IC fusion protein was expressed in Escherichia coli strain BL21 and purified by GST affinity column according to the manufacturer's protocol (Amersham Biosciences). Purity was at least 90%, as judged by SDS-PAGE.
siRNA Silencing of RPTP-in Primary Human Keratinocytes-A unique 21-mer RNA sequence derived from RPTP-coding sequence (5Ј-AAG GTT TGC CGC TTC CTT CAG-3Ј) was designed using Oligoengine software (Seattle, WA). Control siRNA contained a random sequence without homology to any known human gene. Doublestranded siRNA was synthesized by Xeragon Inc, (Valencia, CA). The synthetic siRNA was transfected into primary human keratinocytes using Amaxa Biosystems Nucleofactor (Cologne, Germany). Expression of RPTP-in primary keratinocytes increased with cell confluency (see Fig. 10A). To achieve the maximal knockdown effect, siRNA-transfected keratinocytes were cultured to confluency for analysis.

Inhibition of Protein-Tyrosine-Phosphatase Activity Is Associated with Increased EGFR Tyrosine Phosphorylation in Primary Human
Keratinocytes-Treatment of intact primary adult human keratinocytes with two nonspecific PTP inhibitors, H 2 O 2 and pervanadate, caused substantial and rapid tyrosine phosphorylation of EGFR (Fig. 1A). The magnitude of EGFR tyrosine phosphorylation was similar to that induced by EGF (20 ng/ml) (Fig. 1A). For these, an antibody that specifically recognizes phosphorylated tyrosine 1068 in EGFR was used. This approach allowed us to detect phosphorylated EGFR in lysates without the need for immunoprecipitation prior to Western blotting. The finding that PTP inhibitors cause accumulation of EGFR tyrosine phosphorylation suggests the potential importance of PTP activity in the regulation of EGFR function.
This potential importance was further investigated in cell-free, EGFR-enriched membrane fractions from human keratinocytes. Treatment of membranes in the presence of Mg 2ϩ /ATP with PTP inhibitors hydrogen peroxide, pervanadate, or orthovanadate significantly induced phosphorylation of EGFR tyrosine residue 1068 (Fig. 1B). The magnitude of tyrosine 1068 phosphorylation following treatment with either PTP inhibitor was 3-5-fold greater than treatment of membranes with EGF (20 ng/ml). As expected, treatment of keratinocyte membranes with hydrogen peroxide, pervanadate, or orthovanadate (but not EGF) inhibited endogenous membrane-associated PTP activity (Fig. 1C). These data suggest the possible involvement of an integral membrane RPTP activity in the regulation of EGFR tyrosine phosphorylation.

Profile of RPTPs in EGFR-expressing Human Keratinocytes-To
investigate the above possibility, we first assessed which of 21 known RPTPs in the human genome are expressed in human keratinocytes. Specific PCR primers for each of the 21 human RPTPs were designed and tested for specificity using cloned cDNAs as templates. Each of the 21 PCR products generated from cloned templates was authenticated by DNA sequencing. RT-PCR was used to detect mRNA expression of each of the 21 RPTPs in adult human keratinocytes and adult human skin. RT-PCR reactions for 13 RPTPs yielded products of the expected size (data not shown). Each product was cloned and verified by DNA sequencing. RPTPs ␤, ␦, , , and were predominantly expressed and therefore chosen for further study.
Dephosphorylation of EGFR by RPTP-Transiently Expressed in CHO Cells-To determine whether any of the five candidate RPTPs is able to regulate EGFR tyrosine phosphorylation, we employed a transient transfection system using CHO cells, which do not express EGFR. Transient transfection of CHO cells with EGFR expression vector alone resulted in a high level of basal (i.e. in the absence of ligand) EGFR tyrosine phosphorylation (Fig. 2A). Treatment of EGFR-expressing CHO cells with EGF modestly increased tyrosine phosphorylation of EGFR ( Fig. 2A). Tyrosine phosphorylation of EGFR in CHO cells was completely blocked by treatment of cells with EGFR tyrosine kinase inhibitor PD169540, demonstrating that tyrosine phosphorylation of EGFR in CHO cells was due to intrinsic EGFR tyrosine kinase activity (data not shown).
To examine the ability of the RPTPs to dephosphorylate EGFR, cDNAs encoding human RPTP-␤, -␦, -, -, and -were co-expressed with EGFR in CHO cells. To verify that each RPTP was expressed, we determined their mRNA levels using real-time RT-PCR. In vector control-transfected CHO cells, no mRNA for any of the five RPTPs was detectable. In RPTP-transfected CHO cells, the mRNA level of each of  the five RPTPs was readily detectable and similar (Fig. 2B). We also determined that transfection resulted in detectable RPTP-␤, -, andprotein expression (we could not obtain useful antibodies for RPTP-␦ and -) (Fig. 2C). Importantly, only RPTP-(but not RPTP-␤, -␦, -, -, and -) was able to significantly reduce constitutive tyrosine phosphorylation of EGFR in CHO cells (Fig. 3A). In addition to the reduction of constitutive EGFR tyrosine phosphorylation, expression of RPTPreduced EGF-stimulated EGFR tyrosine phosphorylation in CHO cells (Fig. 3B). These results indicate that RPTP-is capable of reducing EGFR intrinsic tyrosine kinase-catalyzed phosphorylation when co-expressed in CHO cells.

RPTP-Directly Dephosphorylates EGFR in Vitro-
To determine whether RPTP-can directly dephosphorylate EGFR, we constructed, expressed, and purified catalytically active human RPTP-intracellular region GST fusion protein (GST-RPTP--IC). GST-RPTP--IC was incubated with autophosphorylated purified full-length human EGFR, and the rate of tyrosine dephosphorylation was monitored by Western analysis using antibodies specific for phosphotyrosine residues 1068, 992, and 1173. RPTPrapidly dephosphorylated EGFR tyrosine 1068 and 1173 (Fig. 4). EGFR tyrosine 992 was dephosphorylated at a substantially slower rate than tyrosine 1068 (Fig. 4). In subsequent experiments, tyrosine phosphorylation at residue 1068 was chosen to monitor EGFR dephosphorylation by RPTP-.
Association of Substrate-trapping Mutant RPTP-and EGFR in Intact Cells-To further investigate dephosphorylation of EGFR by RPTP-, we performed substrate-trapping studies. Mutation of an aspartic acid (Asp-1051 for RPTP-), which is conserved in the active site of protein-tyrosine phosphatases, to alanine prevents completion of phosphate ester hydrolysis and therefore causes the formation of a stable enzyme-substrate complex (14). We co-expressed full-length human EGFR with His-tagged full-length wildtype or D1051A mutant RPTPin CHO cells. Following expression, RPTP-was captured on nickel-coated beads and analyzed for its association with EGFR by Western analysis. Although wild-type RPTP-catalyzes the dephosphorylation of EGFR in vitro (Fig. 4), it does not appear to form a stable complex with EGFR, as expected, when co-expressed in cells (Fig. 5). In contrast, D1051A RPTPdoes form a stable complex with EGFR, which can be readily detected by Western analysis (Fig. 5). These data demonstrate that EGFR is a substrate for RPTP-in intact cells.

Dephosphorylation of Endogenous EGFR by RPTP-in Primary Human
Keratinocytes-We next examined the effect of RPTP-on tyrosine phosphorylation of endogenous EGFR in primary human keratinocytes. Adenovirus-mediated overexpression of RPTP-in primary human keratinocytes significantly reduced EGF-induced EGFR tyrosine 1068 phosphorylation (Fig. 6). Expression of catalytically inactive 2C/S-RPTP-, with cysteine to serine mutation in both PTP catalytic domains, increased both basal and EGF-induced EGFR tyrosine 1068 phosphorylation (Fig. 6). This increased EGFR tyrosine phosphoryla-    DECEMBER 30, 2005 • VOLUME 280 • NUMBER 52 tion likely reflects dominant negative activity of catalytically inactive 2C/S-RPTP- (12,13). Reduction of EGFR tyrosine phosphorylation observed with expression of RPTP-was specific, because adenovirusmediated overexpression of RPTP-, which is most structurally related to RPTP-, did not reduce EGF-induced EGFR tyrosine phosphorylation in human keratinocytes (Fig. 7).

RPTP-Knockdown Potentiates EGF-induced Erk Activation in Primary Human
Keratinocytes-One of the major downstream effectors of EGFR is Erk mitogen-activated protein kinase. Because reduction of endogenous RPTPpotentiates basal and EGF-induced EGFR activation, we investigated the effect of RPTP-knockdown on Erk activation. siRNA-mediated RPTP-knockdown resulted in a 2-fold increase of basal phosphorylation of p44 and p42 Erk, in primary human keratinocytes (Fig. 9C). Similarly, RPTP-knockdown resulted in a further 2-fold increase of EGF-induced phosphorylation of Erk (Fig. 9C).
RPTP-Inhibits Keratinocyte Growth-In culture, growth of human keratinocytes slows and eventually stops as the cells reach confluency. Because keratinocyte growth is EGFR-dependent (19,20), we examined RPTPexpression as a function of culture confluency. We found that RPTP-mRNA expression was relatively low in subconfluent keratinocytes and increased markedly when keratinocytes became confluent (Fig. 10A). These data indicate that increased expression of RPTP-is associated with reduced keratinocyte growth.
To examine whether the level of RPTP-can influence keratinocyte growth, RPTP-was overexpressed in subconfluent keratinocyte cultures with low endogenous RPTP-and the rate of proliferation determined. Both empty and RPTP-adenovirus-treated keratinocytes displayed slow growth during the first two days after treatment, which is typical for human keratinocyte seeded at low density. Empty adenovirus-treated keratinocytes exhibited accelerated growth during days three and four post-treatment, reaching 80 -90% confluency at day four (Fig. 10B). In contrast, RPTPadenovirus-treated keratinocytes showed no significant proliferation during days three and four posttreatment (Fig. 10B). During the course of the experiment, keratinocyte   viability, determined by trypan blue exclusion, was Ͼ95% for both empty and RPTP-adenovirus-treated keratinocytes.

DISCUSSION
We demonstrated that RPTP-directly dephosphorylates EGFR in vitro and increased RPTP-expression reduces basal and ligand-stimulated EGFR tyrosine phosphorylation, whereas reduced RPTPexpression increases EGFR tyrosine phosphorylation. Regulation of EGFR tyrosine phosphorylation is complex and occurs at multiple levels, including ligand activation, internalization, recycling, biosynthesis, and dephosphorylation (21,22). Several different PTPs have been demonstrated to reduce EGFR phosphorylation at various stages following ligand activation (14 -18). To our knowledge, however, RPTPis unique among RPTPs in its ability to regulate both basal (i.e. in the absence of exogenous ligand) and ligand-activated EGFR tyrosine phosphorylation. In addition, RPTPis unique in its ability to regulate downstream EGFR function, including Erk activation and cellular proliferation.
Human keratinocytes were found to express 13 of 21 known RPTPs. We tested the five most abundantly expressed RPTPs for their ability to reduce basal EGFR tyrosine phosphorylation. Among the five RPTPs tested, only RPTPwas able to significantly reduce basal EGFR tyrosine phosphorylation. It is possible that other RPTPs, which we did not study, act in concert with RPTP-to reduce EGFR phosphorylation. We found that RPTP-preferentially dephosphorylated tyrosines 1068 and 1173, compared with tyrosine 992. This observation further supports the concept that EGFR is acted upon by multiple RPTPs. In this regard, both RPTP LAR, and RPTPhave been shown to reduce EGFR phosphorylation (18,23). However, whether either RPTP acts directly on EGFR, or which phosphotyrosine residue(s) either RPTP affects remains to be determined.
Human keratinocytes, similar to many other cell types in culture, cease proliferation in response to cell-cell contacts. Although the detailed mechanism of contact inhibition remains elusive, interactions among adhesion molecules on the surface of adjoining cells plays a pivotal role. The extracellular domains of many RPTPs contain adhesion molecule-like sequences, leading to the proposal that RPTP functions may be regulated, at least in part, by cell-cell contacts (24). In fact, RPTPand RPTP-have been shown to possess adhesion properties that can mediate cell-cell and cell-matrix communication (25,26). Interestingly, membrane-associated PTP activity is increased up to 10-fold in contacted-inhibited cells and harvested at high density, compared with proliferating cells at low density, whereas tumor cells, which are not subjected to contact-inhibition, do not show density-dependent increase in membrane-associated PTP activity (27,28). Furthermore, RPTP-, RPTP-␤, and DEP-1 have been shown to be up-regulated as a function of increased cell density in culture (24,28,29). In the current study, we demonstrated that RPTP-levels also increase as a function of confluence in cultured human keratinocytes. This finding is consistent FIGURE 9. siRNA-mediated knockdown of RPTP-increases basal and potentiates EGF-induced EGFR tyrosine phosphorylation and Erk activation in human keratinocytes. A, human keratinocytes were transfected with the indicated concentrations of control (CTRL) or RPTP-siRNA. Two days post-transfection, whole cell lysates were analyzed by Western blot for EGFR phosphotyrosine 1068 (pY-1068) and total EGFR. Results are means Ϯ S.E. of fluorescent band intensities quantified by STORM as described under "Experimental Procedures." n ϭ 2; *, p Ͻ 0.05 versus control siRNA. Inset shows representative Western blot. B, human keratinocytes were transfected with 90 nM control or RPTP-siRNA. Two days post-transfection, the cells were treated with vehicle (Veh) or EGF (5 ng/ml) for 10 min at 37°C. Whole cell lysates were analyzed by Western blot for EGFR phosphotyrosine 1068 (pY-1068) and total EGFR. Data are means Ϯ S.E. of fluorescent band intensities quantified by STORM as described under "Experimental Procedures." n ϭ 3; *, p Ͻ 0.05 versus control siRNA. Inset shows representative Western blot. C, human keratinocytes were transfected with 90 nM control or RPTP-siRNA. Two days post-transfection, the cells were treated with vehicle (Veh) or EGF (5 ng/ml) at 37°C for 30 min. Whole cell lysates were analyzed by Western blot for phospho-and total Erk. Results are means Ϯ S.E. of fluorescent band intensities quantified by STORM as described under "Experimental Procedures." n ϭ 3; *, p Ͻ 0.05 versus control siRNA. Inset shows representative Western blot.