Hyaluronan-CD44 interaction with Rac1-dependent protein kinase N-gamma promotes phospholipase Cgamma1 activation, Ca(2+) signaling, and cortactin-cytoskeleton function leading to keratinocyte adhesion and differentiation.

In this study we have investigated hyaluronan (HA)-CD44 interaction with protein kinase N-gamma (PKNgamma), a small GTPase (Rac1)-activated serine/threonine kinase in human keratinocytes. By using a variety of biochemical and molecular biological techniques, we have determined that CD44 and PKNgamma kinase (molecular mass approximately 120 kDa) are physically linked in vivo. The binding of HA to keratinocytes promotes PKNgamma kinase recruitment into a complex with CD44 and subsequently stimulates Rac1-mediated PKNgamma kinase activity. The Rac1-activated PKNgamma in turn increases threonine (but not serine) phosphorylation of phospholipase C (PLC) gamma1 and up-regulates PLCgamma1 activity leading to the onset of intracellular Ca(2+) mobilization. HA/CD44-activated Rac1-PKNgamma also phosphorylates the cytoskeletal protein, cortactin, at serine/threonine residues. The phosphorylation of cortactin by Rac1-PKNgamma attenuates its ability to cross-link filamentous actin in vitro. Further analyses indicate that the N-terminal antiparallel coiled-coil (ACC) domains of PKNgamma interact directly with Rac1 in a GTP-dependent manner. The binding of HA to CD44 induces PKNgamma association with endogenous Rac1 and its activity in keratinocytes. Transfection of keratinocytes with PKNgamma-ACCcDNA reduces HA-mediated recruitment of endogenous Rac1 to PKNgamma and blocks PKNgamma activity. These findings suggest that the PKNgamma-ACC fragment acts as a potent competitive inhibitor of endogenous Rac1 binding to PKNgamma in vivo. Most important, the PKNgamma-ACC fragment functions as a strong dominant-negative mutant that effectively inhibits HA/CD44-mediated PKNgamma phosphorylation of PLCgamma1 and cortactin as well as keratinocyte signaling (e.g. Ca(2+) mobilization and cortactin-actin binding) and cellular functioning (e.g. cell-cell adhesion and differentiation). Taken together, these findings strongly suggest that hyaluronan-CD44 interaction with Rac1-PKNgamma plays a pivotal role in PLCgamma1-regulated Ca(2+) signaling and cortactin-cytoskeleton function required for keratinocyte cell-cell adhesion and differentiation.

In the epidermis, extracellular matrix (ECM) 1 components form an integral part of hemidesmosomes and mediate keratinocyte attachment to the underlying basement membrane. Hyaluronan (HA) is the major glycosaminoglycan in the extracellular matrix (ECM) of most mammalian tissues including epidermis and dermis (1)(2)(3) and HA has been implicated in skin epidermal function (1)(2)(3). However, the cellular and molecular mechanism by which keratinocytes respond to HA is not understood.
The binding of HA to CD44 causes cells to adhere to extracellular matrix (ECM) components and has also been implicated in the stimulation of several different biological activities (16 -19). The intracellular domain of CD44 binds to signaling proteins, such as RhoA-and Rac1-specific guanine nucleotide exchange factors (20 -23), and cytoskeletal proteins including ankyrin (16 -19) and the ERM proteins (ezrin, radixin, and moesin) (24). In addition, the cytoplasmic domain of CD44 is known to be tightly coupled with c-Src kinase, which promotes tyrosine phosphorylation of cortactin (a filamentous actin (Factin)-associated protein) and cytoskeleton function resulting in specific structural changes in the plasma membrane and cell migration (18,19,61). These findings strongly suggest that the CD44 molecule provides a direct linkage between the ECM and the cytoskeleton.
Several enzymes have been identified as possible downstream targets for Rac1 signaling. One such enzyme is protein kinase N-␥ (PKN␥) (also called PRK2), which belongs to a family of serine-threonine kinases known to interact with Rac1 in a GTP-dependent manner, and it shares a great deal of sequence homology with protein kinase C in the C-terminal region (29 -31). The N-terminal region of PKN contains three homologous stretches of ϳ70 amino acids (relatively rich in charged residues) and forms an antiparallel coiled-coil fold (ACC domain) (29). This ACC domain has been shown to interact with RhoGTPases such as RhoA and Rac1 (and to a lesser extent with Cdc42) (29 -31). The C-terminal region contains the C2-like region, which functions as an auto-inhibitory domain (32). The ACC and the C2-like domains, together with the catalytic domain, are conserved among the PKN family members (29 -32). In keratinocytes, RhoA-activated PKN␥ has been found to be involved in Fyn/Src kinase-regulated cell-cell adhesion during Ca 2ϩ -induced differentiation (33). The question of whether HA/CD44-induced Rac1 activation and PKN␥-targeted downstream effectors are involved in regulating keratinocyte cell-cell adhesion and differentiation has not been clearly addressed.
HA binding to CD44-expressing cells also stimulates intracellular Ca 2ϩ mobilization, which is a prerequisite for the onset of a variety of biological activities (34 -36). Although the cellular and molecular mechanisms involved in HA/CD44-mediated Ca 2ϩ signaling are currently not well understood, one of the likely pathways involves the phosphatidylinositol cascade that leads to Ca 2ϩ release from intracellular stores (37). In response to different stimuli (including HA), phosphoinositide-specific phospholipase Cs (PLCs) hydrolyze phosphatidyl-4,5-bisphosphate into diacylglycerol and inositol 1,4,5-trisphosphate (IP 3 ) (35). IP 3 , as a second messenger, binds to the IP 3 receptor and induces the release of Ca 2ϩ from intracellular stores (34 -37). At least four major PLC families exist, PLC-␤, PLC-␥, PLC-␦, and PLC-⑀, and each family type has a number of subtypes (38,39). One of the best characterized PLCs in keratinocytes is PLC-␥1, which is involved in inositol lipid signaling and keratinocyte differentiation (40,41). Specifically, Xie and Bikle (40) have shown that blockage of PLC-␥1 expression with antisense PLC-␥1cDNA or down-regulation of PLC-␥1 activity with a specific inhibitor (41) can effectively impair Ca 2ϩ -induced keratinocyte differentiation. These findings suggest that intracel-lular PLC-␥1-mediated Ca 2ϩ mobilization plays a critical role in regulating keratinocyte differentiation.

MATERIALS AND METHODS
Cell Culture Model Systems-Normal human keratinocytes were isolated from neonatal human foreskins and grown in serum-free keratinocyte growth medium (KGM, Clonetics, San Diego, CA) as described previously (45). Briefly, keratinocytes were isolated from newborn human foreskins by trypsinization (0.25% trypsin, 4°C, overnight), and primary cultures were established in KGM containing 0.07 mM calcium. After the first and second passages, keratinocytes were incubated with KGM containing 0.03 mM calcium and used in the experiments described below.
Antibodies and Reagents-Monoclonal rat anti-human CD44 antibody (clone, 020; isotype, IgG 2b ; obtained from CMB-TECH, Inc., San Francisco) used in this study recognizes a common determinant of the CD44 class of glycoproteins including CD44s (the standard form), CD44 variant isoforms, and CD44 3-10 (Epican). Rabbit anti-phosphothreonine antibody and rabbit anti-phosphoserine antibody were obtained from Zymed Laboratories Inc.. Polyclonal mouse anti-involucrin and polyclonal mouse anti-transglutaminase were purchased from Covance Inc. (Princeton, NJ) and Neomarkers (Fremont, CA), respectively. Rabbit anti-PKN␥, rabbit anti-PLC␥1, mouse anti-RhoA, and mouse anti-Rac1 were obtained from Santa Cruz Biotechnology. Mouse anti-His and mouse anti-cortactin antibody (clone 4F11) were purchased from Invitrogen and Upstate Biotechnology, Inc. (Lake Placid, NY), respectively. Several other reagents including GST-tagged RhoA, GST-tagged Rac1, and inhibitors (e.g. U73122, Xestospongin C, and BAPTA) were obtained from Calbiochem. Rooster comb hyaluronan (HA) and cytochalasin D were purchased from Sigma. High molecular weight HA polymers (ϳ10 6 daltons) were purified by gel filtration column chromatography using Sephacryl S1000 column. The purity of high molecular weight HA polymers used in our experiments was further verified by anion exchange high performance liquid chromatography. No small HA fragments were detected in these preparations.
Method for Preparing His-tagged Dominant-negative Form (Containing ACC Domains) of PKN␥ Kinase-The cDNA fragment encoding the ACC domains of PKN␥ (amino acids 98 -256) was generated by reverse transcription-PCR using two specific primers (5Ј-GGTGCAGCA-GAAATTGGATGA-3Ј and 5Ј-CTTTGACGTGGACTTAGTGTTGGTG-3Ј). The PCR product digested with EcoRI and HindIII was purified with QIAquick PCR purification kit (Qiagen). The PKN␥-ACC fragment cDNA was subsequently cloned into pcDNA3.1/HisC vector that contains Xpress epitope to create His-tagged PKN␥-ACCcDNA. The inserted ACC domain sequence was confirmed by nucleotide sequencing analyses. This His-tagged PKN␥-ACC cDNA was then used for transient expression in keratinocytes as described below.
Cell Transfection-To establish a transient expression system, keratinocytes were transfected with various plasmid DNAs (e.g. Histagged PKN␥-ACC domain cDNA or vector alone) using TransIT R -Keratinocyte Transfection Reagent (Mirus, Madison, WI) according to those procedures described previously (46). Briefly, keratinocytes were plated at a density of 2 ϫ 10 6 cells per 100-mm dish and transfected with 25 g/dish plasmid cDNA using TransIT R -Keratinocyte Transfection Reagent. Transfected cells were grown in the culture medium containing G418 for at least 3 days. Various transfectants were then analyzed for their protein expression (e.g. PKN␥-ACC-related proteins) by immunoblot, PKN␥ activity, and keratinocyte functional assays as described below.
Measurement of Rac1 Activation-Keratinocytes (ϳ5.0 ϫ 10 6 cells) were resuspended in a buffer containing 118 mM KCl, 5 mM NaCl, 0.4 mM CaCl 2 , 1 mM EGTA, 1.2 mM magnesium acetate, 1.2 mM KH 2 PO 4 , 25 mM Tris-HCl (pH 7.4), 20 mg/ml bovine serum albumin. An aliquot of the cell suspension was added to the electroporation cuvette and incubated at 4°C for 5-10 min followed by adding [ 35 S]GTP␥S (12.5 Ci). Subsequently, cells were electroporated at 25 microfarads and 2.0 kV/cm followed by incubating with 50 g/ml HA (in the presence or absence of rat anti-CD44 antibody (50 g/ml)) or without any HA treatment at 37°C for 10 min. Subsequently, [ 35 S]GTP␥S-labeled cells were washed in PBS (pH 7.4) and solubilized in 1.0% Nonidet P-40 with 1 mM GTP, 25 mM magnesium acetate, and protease inhibitors in PBS (pH 7.4). Nonidet P-40-solubilized cells were then incubated with mouse anti-Rac1 IgG (5 g/ml) plus goat anti-mouse-conjugated beads. The amount of [ 35 S]GTP␥S-Rac1 associated with anti-Rac1-conjugated immunobeads was measured by a gamma counter. The values expressed in Table I represent an average of triplicate determinations of five experiments with a standard deviation of less than 5%.
Untransfected keratinocytes (untreated or pretreated with the various inhibitors U73122 (1 M), Xestospongin C (1 M), BAPTA/AM (1 M), cytochalasin D (20 g/ml), and colchicine (1 ϫ 10 Ϫ5 M) for 1 h at 37°C) or keratinocytes transfected with PKN␥-ACCcDNA (or vector alone) were incubated with HA (50 g/ml) at 37°C for 10 min (or pretreated with anti-CD44 antibody followed by adding HA (50 g/ml) or incubated with no HA). These cells were then immunoprecipitated with anti-PLC␥1 (or anti-cortactin) antibody followed by immunoblotting with anti-phosphoserine or anti-phosphothreonine, respectively. Subsequently, these blots were then developed using the ECL reagent according to the manufacturer's instructions. During these immunological analyses, an equal amount of cellular protein (50 g/ml) immunoprecipitated with the antibody was applied to SDS-PAGE followed by immunoblot analyses.
In some experiments, keratinocytes (e.g. untransfected or transfected with His-tagged PKN␥-ACCcDNA or vector only) were incubated with HA (50 g/ml) at 37°C for 10 min (or pretreated with anti-CD44 antibody followed by adding HA (50 g/ml) or incubated with no HA). These cells were then immunoprecipitated with rabbit anti-PKN␥ IgG (or mouse anti-His IgG) followed by immunoblotting with mouse anti-Rac1 IgG/mouse anti-RhoA IgG or reblotted with anti-PKN␥ IgG (or anti-His IgG) followed by incubation with horseradish peroxidase-conjugated goat anti-mouse IgG (or goat anti-rabbit IgG) (1:10,000 dilution). Cell lysates of keratinocytes (e.g. transfected with His-tagged PKN␥cDNA or vector only) were also immunoblotted with anti-PKN␥ IgG and anti-His IgG, respectively. These blots were then developed using ECL reagent (Amersham Biosciences).
PKN␥-mediated Protein Phosphorylation Assay in Vitro-The PKN␥ kinase reaction was carried out in 50 l of the reaction mixture containing 40 mM Tris-HCl (pH 7.5), 2 mM EDTA, 1 mM dithiothreitol, 7 mM PKN␥ associated with GTP-bound Rac1 beads. C, analysis of the CD44-PKN␥ complex in keratinocytes. Keratinocytes (untreated or treated with HA) were solubilized by 1% Nonidet P-40 buffer and immunoprecipitated (IP) with anti-CD44 antibody followed by immunoblotting using anti-PKN␥ antibody (a) or anti-CD44 antibody (b), respectively. Lane 1, detection of PKN␥ in the complex isolated from untreated keratinocytes by anti-CD44-mediated immunoprecipitation followed by immunoblotting with anti-PKN␥ (a) or reblotting with anti-CD44 (b) as a loading control. Lane 2, detection of PKN␥ in the complex isolated from keratinocytes (treated with HA) by anti-CD44-mediated immunoprecipitation followed by immunoblotting with anti-PKN␥ (a) or reblotting with anti-CD44 (b) as a loading control. a Keratinocytes (ϳ5.0 ϫ 10 6 cells) were preloaded with ͓ 35 S͔GTP␥S (12.5 Ci) using electroporation methods as described under ''Materials and Methods.'' These cells were incubated with 50 g/ml HA at 37°C for 10 min (in the presence or absence of rat anti-CD44 antibody (50 g/ml) or without any HA treatment). Subsequently, ͓ 35 S͔GTP␥S-labeled cells were solubilized in 1.0% Nonidet P-40 and incubated with mouse anti-Rac1 IgG (5 g/ml) plus goat anti-mouse conjugated beads. The amount of ͓ 35 S͔GTP␥S-Rac1 associated with anti-Rac1-conjugated immunobeads was measured using a gamma counter. The values expressed represent an average of triplicate determinations of three experiments with an S.D. of less than Ϯ 5%.
or 200 M GTP). After incubation for 30 min at 30°C, the reaction mixtures were immunoprecipitated with anti-PLC␥1 beads (or anticortactin beads), boiled in SDS sample buffer, and subjected to SDS-PAGE. The protein bands associated with anti-PLC␥1 beads (or anti-cortactin beads) were revealed by anti-phosphoserine/anti-phos-phothreonine-mediated immunoblot or anti-PLC␥1 (or anti-cortactin)mediated immunoblot, respectively.
PLC␥1 Activity Assay-The procedures for analyzing the activity of unphosphorylated or phosphorylated PLC␥1 were similar to those described previously (47). Briefly, keratinocytes were solubilized in Nonidet P-40 buffer in the presence of protease inhibitors plus 1 M NaF, 1 M sodium orthovanadate, and 1 M okadaic acid. Lysates were spun at 5000 ϫ g, and the supernatants were collected. PLC␥1 was immunoprecipitated from the supernatants using rabbit anti-PLC␥1-conjugated beads. These PLC␥1-conjugated beads were then phosphorylated with PKN␥ in the presence of GDP/GTP-bound RhoA or GDP/GTP-bound Rac1 or no PKN␥ (as described above). Subsequently, unphosphorylated or RhoA/Rac1-PKN␥-phosphorylated PLC␥1 immunobeads were incubated with vesicles containing 100 M phosphatidylethanolamine, 10 M phosphatidylserine, 10 M PIP 2 , and 0.028 M [ 3 H]PIP 2 in PLC activity buffer (25 mM HEPES, 80 mM KCl, 3 mM EGTA, 0.5 mM dithiothreitol, pH 7.0) at 37°C. In some cases, PLC␥1 isolated from keratinocytes (untransfected or transfected with PKN␥-ACCcDNA or vector alone), which were treated with HA (50 g/ml) at 37°C for 10 min (or pretreated with anti-CD44 antibody followed by adding HA (50 g/ml) or incubated with no HA), was also incubated in this enzyme assay mixture. Reactions were carried out for 15 min and terminated by adding 10% trichloroacetic acid and 2 mg/ml bovine serum albumin. Precipitated proteins were removed by centrifugation. Radioactive Double Immunofluorescence Staining-Keratinocytes (transfected with PKN␥-ACCcDNA or vector alone) were incubated with HA (50 g/ml) at 37°C for 36 h (or pretreated with anti-CD44 antibody followed by adding HA (50 g/ml) or incubated with no HA) followed by fixing with 2% paraformaldehyde. Subsequently, these cells were stained with Texas Red-labeled rat anti-CD44 IgG. Texas Red-labeled cells were then rendered permeable by ethanol treatment followed by incubating with fluorescein (FITC)-conjugated immunoreagents (e.g. anti-PKN␥ IgG, anti-PLC␥1 IgG, or anti-cortactin IgG). To detect nonspecific antibody binding, Texas Red-conjugated anti-CD44 or FITC-conjugated immunoreagent (e.g. anti-PKN␥ IgG, anti-PLC␥1 IgG, and anti-cortactin IgG)-labeled cells were incubated with FITC-conjugated normal rabbit/normal mouse or Texas Red-conjugated normal rat IgG, respectively. No labeling was observed in such control samples. These fluorescein-and Texas Red-labeled samples were examined with a confocal laser scanning microscope.

FIG. 2. Detection of PLC␥1 phosphorylation by PKN␥ in vitro (A) and in vivo (B).
A, the PKN␥ kinase reaction was carried out in the reaction mixture containing ATP and purified PKN␥ and PLC␥1 in the presence of RhoA or Rac1 (incubated with GDP or GTP). After incubation at 30°C for 30 min, the reaction mixtures were immunoprecipitated (IP) with anti-PLC␥1 beads, boiled in SDS sample buffer, and subjected to SDS-PAGE. The protein bands associated with anti-PLC␥1 beads were revealed by anti-phosphothreonine/anti-phosphoserine-mediated immunoblot or anti-PLC␥1-mediated immunoblot, respectively, as described under "Materials and Methods." Lane 1, PLC␥1 phosphorylation by GDP-RhoA-treated PKN␥ using anti-phosphothreonine (a)/ anti-phosphoserine (b)-mediated immunoblot or anti-PLC␥1-mediated immunoblot (c). Lane 2, PLC␥1 phosphorylation by GTP-RhoA-treated PKN␥ using anti-phosphothreonine (a)/anti-phosphoserine (b)-mediated immunoblot or anti-PLC␥1-mediated immunoblot (c). Lane 3, PLC␥1 phosphorylation by GDP-Rac1-treated PKN␥ using anti-phosphothreonine (a)/anti-phosphoserine (b)-mediated immunoblot or anti-PLC␥1-mediated immunoblot (c). Lane 4, PLC␥1 phosphorylation by GTP-Rac1-treated PKN␥ using anti-phosphothreonine (a)/anti-phosphoserine (b)-mediated immunoblot or anti-PLC␥1-mediated immunoblot (c). B, keratinocytes treated with HA (or pretreated with anti-CD44 plus HA or no HA) were solubilized by Nonidet P-40 and immunoprecipitated with anti-PLC␥ antibody followed by immunoblotting with anti-phosphothreonine/anti-phosphoserine, respectively. Lane 1, detection of PLC␥1 phosphorylation by anti-PLC␥-mediated immunoprecipitation followed by immunoblotting with anti-phosphothreonine (a)/antiphosphoserine (b) or (reblotting with anti-PLC␥1 as a loading control (c)) in keratinocytes treated with no HA. Lane 2, detection of PLC␥1 phosphorylation by anti-PLC␥-mediated immunoprecipitation followed by immunoblotting with anti-phosphothreonine (a)/anti-phosphoserine (b) or (reblotting with anti-PLC␥1 as a loading control (c)) in keratinocytes treated with HA. Lane 3, detection of PLC␥1 phosphorylation by anti-PLC␥-mediated immunoprecipitation followed by immunoblotting with anti-phosphothreonine (a)/anti-phosphoserine (b) or (reblotting with anti-PLC␥1 as a loading control (c)) in keratinocytes pretreated with anti-CD44 followed by adding HA treatment. 10,568 a PLC␥1 isolated from keratinocytes was immunoprecipitated from the supernatants using rabbit anti-PLC␥1-conjugated beads. This PLC␥1 was then phosphorylated by PKN␥ in the presence of GDP/GTPbound RhoA or GDP/GTP-bound Rac1 or no PKN␥. Subsequently, unphosphorylated or RhoA/Rac1-PKN␥ phosphorylated PLC␥1 immunobeads were incubated with vesicles containing 100 M phosphatidylethanolamine, 10 M phosphatidylserine, 10 M PIP 2 , and 0.028 M ͓ 3 H͔PIP 2 in PLC activity buffer as described under ''Materials and Methods.'' In some cases, PLC␥1 isolated from keratinocytes (treated with HA (50 g/ml) at 37°C for 10 min, pretreated with anti-CD44 antibody followed by adding HA (50 g/ml), or incubated with no HA) was also incubated with this enzyme assay mixture. Reactions were performed for 15 min and terminated by adding trichloroacetic acid and 2 mg/ml BSA. Precipitated proteins were removed by centrifugation. Radioactive ͓ 3 H͔IP 3 (liberated from ͓ 3 H͔PIP 2 ) in the trichloroacetic acid-soluble supernatant fraction was analyzed by using scintillation counting. Cells were subsequently washed three times with the same buffer. Cells (10 7 cells/ml) (with or without rat anti-CD44 antibody (1 g/ml)) resuspended in 0.1 M phosphate-buffered saline (pH 7.0) were incubated simultaneously with an equal volume of 0.1 M phosphate-buffered saline (pH 7.0) containing HA (50 g/ml) (Sigma) into a 20-l chamber alternately illuminated with 200-ms flashes of 340 and 380 nm every 10 ms (monitoring the emission wavelength of 510 nm) using a Dural wavelength fluorescence imaging system (Intracellular Imaging Inc., Cincinnati, OH). The concentration of intracellular Ca 2ϩ was determined by the following equation: F-actin Cross-linking Assay-The procedures for conducting F-actin cross-linking experiments were the same as those described previously (48,49) with some modifications. Unphosphorylated cortactin or PKN␥phosphorylated cortactin (as described above) (50 -100 nM) in 50 l of TKM buffer (50 mM Tris-HCl (pH 7.4), 134 mM KCl, and 1 mM MgCl 2 ) was mixed with an equal volume of 8 M of 125 I-labeled F-actin followed by a 30-min incubation at room temperature. Subsequently, the mixture was centrifuged at 25,000 ϫ g for 10 min at room temperature. The supernatant was then collected, and the radioactivity in this fraction was counted. The decrease (or loss) of radioactivity in the supernatant fraction reflects F-actin precipitation because of the cross-linking reaction (48,49). The F-actin cross-linking reaction in the presence of unphosphorylated cortactin (control) is designated as 100%. In some cases, unphosphorylated cortactin isolated from vector-transfected cells using His-conjugated Sepharose beads (control) is designated as 100%. Each assay was set up in triplicate and repeated at least three times. All data were analyzed statistically using the Student's t test, and statistical significance was set at p Ͻ 0.01.
Keratinocyte Cell-Cell Adhesion Assay-Keratinocytes (10 7 cells/ml) (untreated or pretreated with the various inhibitors U73122 (1 M), Xestospongin C (1 M), BAPTA/AM (1 M), cytochalasin D (20 g/ml), and colchicine (1 ϫ 10 Ϫ5 M) at 37°C for 1 h or transfected with PKN␥-ACCcDNA or vector alone) grown in 0.03 mM Ca 2ϩ were incubated with no HA, with HA (50 g/ml), or pretreated with anti-CD44 antibody followed by adding HA (50 g/ml) at 37°C for various intervals (e.g. 24, 36, or 48 h). The degree of cell-cell adhesion was calculated as the cell-cell adhesion index (N o Ϫ N a )/N o . Specifically, N o represents the total number of cells (free cells plus aggregated cells) before HA treatment, and N a is the number of free cells after HA treatment. The number of cell units was counted using a hemocytometer as described previously (50). The values expressed in Tables V-VII represent an average of triplicate determinations of five experiments with a standard deviation of less than 5%.

HA-stimulated Rac1
Signaling and CD44-PKN␥ Kinase Association in Keratinocytes-The binding of HA to CD44 is known to induce important changes in certain RhoGTPases such as Rac1 (22,23,26). By using an in vitro [ 35 S]GTP␥S binding assay, we have determined that Rac1, isolated from human keratinocytes, displays specific guanine nucleotide binding activity (Table I). In particular, we have demonstrated that the addition of HA to CD44-containing keratinocytes causes almost a 3-fold increase in the binding of [ 35 S]GTP␥S to Rac1 as compared with the amount of binding present in untreated keratinocytes (Table I) or in keratinocytes pretreated with anti-CD44 antibody followed by HA treatment (Table I). These findings suggest that HA and CD44 are directly involved in the activation of Rac1 in human keratinocytes.
Eukaryotic cells contain a large number of protein serine/ threonine kinases, which appear to play important roles in signal transduction required for various cell functions. One of the known downstream targets for the GTP-bound (activated) form of Rac1 is PKN␥ kinase (also called PRK2 kinase) (29 -31). By using a specific anti-PKN␥-mediated immunoblot technique, we have found significant levels of the PKN␥ (molecular mass ϳ120 kDa) are expressed in keratinocytes (Fig. 1A, lane 2). We believe that the PKN␥ detected in keratinocytes revealed by anti-PKN␥-mediated immunoblot is specific because no protein is detected in these cells using preimmune rabbit IgG (Fig. 1A, lane 1). Several lines of evidence indicate that PKN␥ binds to RhoG-TPases (e.g. Rac1 and RhoA) and that the activity of PKN␥ is up-regulated in the presence of the GTP-bound form of Rac1 or RhoA (29 -31). In this study we have incubated purified PKN␥ (isolated from keratinocytes) with GDP-or GTP-loaded forms of either Rac1-or RhoA-GST conjugated beads. Proteins associated with Rac1 or RhoA-GST beads were then analyzed by immunoblotting with anti-PKN␥ (Fig. 1B). Our results indicate that a large amount of PKN␥ is associated with GTP-bound Rac1-GST beads (Fig. 1B, lane 4), whereas very little PKN␥ is bound to GDP-bound Rac1 (Fig. 1B, lane 3). We have also noted that PKN␥ binds GDP-bound RhoA (Fig. 1B, lane 1) and GTPbound RhoA (Fig. 1B, lane 2) equally well. These results are consistent with previous findings that determined that PKN␥-Rac1 interaction is GTP-dependent but PKN␥-RhoA binding is GTP-independent (30).
In addition, we have addressed the question of whether there is an interaction between CD44 and PKN␥ in human keratinocytes. To this end we first carried out anti-CD44-mediated immunoprecipitation followed by anti-PKN␥ immunoblot (Fig.  1C, a) or anti-CD44 immunoblot (Fig. 1C, b), respectively, using untreated keratinocytes. Our results indicate that a low level of PKN␥ (Fig. 1C, a, lane 1) is present in the anti-CD44immunoprecipitated materials (Fig. 1C, b, lane 1). Subsequently, we have determined that HA treatment causes the recruitment of a significant amount of PKN␥ (Fig. 1C, a, lane 2) into the CD44-PKN␥ complex (Fig. 1C, b, lane 2). These findings clearly establish that CD44 and PKN␥ are closely associated with each other in vivo, particularly following HA treatment of the keratinocytes. a PLC␥1 isolated from keratinocytes was immunoprecipitated from the supernatants by using rabbit anti-PLC␥1-conjugated beads. This PLC␥1 was then phosphorylated by PKN␥ in the presence of GDP/GTPbound RhoA or GDP/GTP-bound Rac1 or no PKN␥. Subsequently, unphosphorylated or RhoA/Rac1-PKN␥-phosphorylated PLC␥1 immunobeads were incubated with vesicles containing 100 M phosphatidylethanolamine, 10 M phosphatidylserine, 10 M PIP 2 , and 0.028 M ͓ 3 H͔PIP 2 in PLC activity buffer as described under ''Materials and Methods.'' In some cases, PLC␥1 isolated from keratinocytes (treated with HA (50 g/ml) at 37°C for 10 min, pretreated with anti-CD44 antibody followed by adding HA (50 g/ml), or incubated with no HA) was also incubated with this enzyme assay mixture. Reactions were performed for 15 min and terminated by adding trichloroacetic acid and 2 mg/ml BSA. Precipitated proteins were removed by centrifugation. Radioactive ͓ 3 H͔IP 3 (liberated from ͓ 3 H͔PIP 2 ) in the trichloroacetic acid-soluble supernatant fraction was analyzed by using scintillation counting. PIP 2 hydrolysis was linear for 15 min under the conditions used. The values expressed represent an average of triplicate determinations of 3-5 experiments with an S.D. of less than Ϯ5%.
Next, we examined the potential impact of PKN␥-mediated phosphorylation on the regulation of PLC␥1 activity (measured by PLC␥1-mediated IP 3 production). Our results indicate that the level of IP 3 production is slightly elevated using PLC␥1 phosphorylated with GDP-RhoA or GTP-RhoA-activated PKN␥ (Table II). PLC␥1 activity is greatly stimulated using PLC␥1 treated with GTP-bound Rac1-activated PKN␥ (Table II) (but not PKN␥ in the presence of GDP-bound Rac1 (Table II)). The amount of IP 3 production induced by PLC␥1 activated by GTP-Rac1-PKN␥ is also significantly higher than PLC␥1-activated by GDP-RhoA-PKN␥ or GTP-RhoA-PKN␥ (Table II). In control experiments, we found that unphosphorylated PLC␥1 (in the absence of PKN␥) (Table II) can induce only a basal level or very minimal amounts of IP 3 production. Therefore, we conclude that Rac1-activated (and to a lesser extent RhoA-activated) PKN␥ phosphorylation of PLC␥1 is directly involved in the activation of PLC␥1 activity. Moreover, we have determined that the stimulation of PLC␥1 activity occurs immediately after HA addition to CD44-expressing keratinocytes (Table III). Preincubation with anti-CD44 antibody blocks PLC␥1 activation in keratinocytes treated with HA (Table III) indicating that HA induces PLC␥1 activity in a CD44-dependent manner in keratinocytes.
It has been well documented that PLC␥1-mediated IP 3 production triggers IP 3 receptor-regulated intracellular Ca 2ϩ mobilization (37)(38)(39). In this study, we have used the fluorescence indicator, Fura-2, to measure the intracellular free Ca 2ϩ concentration after HA binding to CD44-containing keratinocytes. The ratio of the fluorescence signal from Fura-2 at 340 and 380 nm excitation was monitored and used to determine the intracellular Ca 2ϩ concentration. Our results clearly show that the intracellular Ca 2ϩ concentration is elevated after the addition of HA to keratinocytes followed by a period of continuous Ca 2ϩ influx (Fig. 3A). These data indicate that intracellular Ca 2ϩ mobilization is one of the early signaling events to occur following HA binding to keratinocytes. Pretreatment of keratinocytes with anti-CD44 antibody (Fig. 3B) or various inhibitors (e.g. U73122 (a PLC inhibitor) (Fig. 3C), Xestospongin C (a membrane-permeable blocker of IP 3 receptor-mediated Ca 2ϩ release) (Fig. 3D), or BAPTA (a membrane-permeable Ca 2ϩ chelator) (Fig. 3E)) effectively blocks HA-mediated intracellular Ca 2ϩ elevation. These findings strongly suggest that Ca 2ϩ signaling in keratinocytes involves HA/CD44-dependent and PLC-regulated processes.
Cortactin Serves as Another New Cellular Substrate for Rac1dependent PKN␥ Kinase and Interacts with F-actin-Although GTPase-activated PKN␥ has been shown to regulate actin cytoskeleton organization (29 -31), the specific cytoskeletal components regulated by PKN␥ in HA-mediated CD44 signaling in keratinocytes have yet to be clearly defined. In this study, we have observed that PKN␥ kinase isolated from keratinocytes is capable of phosphorylating cortactin at both threonine (Fig. 4A,  a) and serine residues (Fig. 4A, b) (as detected by anti-phosphothreonine and anti-phosphoserine-mediated immunoblot, respectively, followed by reblotting with anti-cortactin anti- in vivo (B). A, the PKN␥ kinase reaction was carried out in the reaction mixture containing ATP, purified PKN␥, and cortactin in the presence of RhoA or Rac1 (incubated with GDP or GTP). After incubation for various times (e.g. 0, 10, 20, 30, 60, and 120 min) at 30°C, the reaction mixtures were immunoprecipitated (IP) with anti-cortactin beads and boiled in SDS sample buffer and subjected to SDS-PAGE. The protein bands associated with anti-cortactin beads were revealed by anti-phosphoserine/anti-phosphothreonine-mediated immunoblot or anti-cortactin-mediated immunoblot, respectively, as described under "Materials and Methods." Lane 1, cortactin phosphorylation by GDP-RhoA-treated PKN␥ using anti-phosphothreonine (a)/anti-phosphoserine (b)-mediated immunoblot or anti-cortactin-mediated immunoblot (c). Lane 2, cortactin phosphorylation by GTP-RhoA-treated PKN␥ using anti-phosphothreonine (a)/anti-phosphoserine (b)-mediated immunoblot or anti-cortactin-mediated immunoblot (c). Lane 3, cortactin phosphorylation by GDP-Rac1-treated PKN␥ using anti-phosphothreonine (a)/anti-phosphoserine (b)-mediated immunoblot or anticortactin-mediated immunoblot (c). Lane 4, cortactin phosphorylation by GTP-Rac1-treated PKN␥ using anti-phosphothreonine (a)/antiphosphoserine (b)-mediated immunoblot or anti-cortactin-mediated immunoblot (c). B, keratinocytes treated with HA (or pretreated with anti-CD44 plus HA or no HA) were solubilized by Nonidet P-40 and immunoprecipitated with anti-cortactin antibody followed by immunoblotting with anti-phosphothreonine/anti-phosphoserine, respectively. Lane 1, detection of cortactin phosphorylation by anti-cortactin-mediated immunoprecipitation followed by immunoblotting with anti-phosphothreonine (a)/anti-phosphoserine (b) or (reblotting with anti-cortactin as a loading control) (c) in keratinocytes treated with no HA. Lane 2, detection of cortactin phosphorylation by anti-cortactin-mediated immunoprecipitation followed by immunoblotting with anti-phosphothreonine (a)/anti-phosphoserine (b) or (reblotting with anti-cortactin as a loading control) (c) in keratinocytes treated with HA. Lane 3, detection of cortactin phosphorylation by anti-cortactin-mediated immunoprecipitation followed by immunoblotting with anti-phosphothreonine (a)/anti-phosphoserine (b) or (reblotting with anti-cortactin as a loading control) (c) in keratinocytes pretreated with anti-CD44 followed by HA treatment. body (Fig. 4A, c)) in the presence of activated Rac1 (GTP-bound Rac1) (Fig. 4A, a, lane 4, and A, b, lane 4) (and to a lesser extent GDP-Rac1 (Fig. 4A, a, lane 3, and A, b, lane 3)). The level of threonine/serine phosphorylation of cortactin appears to be relatively low when PKN␥ was incubated with GDP-bound RhoA (Fig. 4A, a, lane 1, and A, b, lane 1) or GTP-bound RhoA (Fig. 4A, a, lane 2, and A, b, lane 2). These results suggest that cortactin phosphorylation by PKN␥ requires Rac1 in a GTP-dependent manner and involves RhoA in a GTP-independent fashion in vitro.

FIG. 4. Detection of cortactin phosphorylation by PKN␥ in vitro (A) and
Further analyses indicate that the level of threonine and serine phosphorylation of cortactin in vivo (as detected by anticortactin-mediated immunoprecipitation followed by immunoblotting with anti-threonine (Fig. 4B, a), anti-serine (Fig. 4B,  a), or anti-cortactin (Fig. 4B, c), respectively) is significantly enhanced in keratinocytes treated with HA ( Fig. 4B, a, lane 2,  and B, b, lane 2). In contrast, cortactin threonine and serine phosphorylation is relatively low in keratinocytes without any HA treatment (Fig. 4B, a, lane 1, and B, b, lane 1) or those keratinocytes pretreated with anti-CD44 antibody followed by HA treatment (Fig. 4B, a, lane 3, and B, b, lane 3). These observations strongly support the conclusion that HA-mediated cortactin phosphorylation is CD44-dependent.
Effects of PKN␥-ACC Domains on PLC␥1/Cortactin Phosphorylation, Ca 2ϩ Signaling, Cytoskeleton Interaction, Keratinocyte Cell-Cell Adhesion, and Differentiation-Previous studies (29 -31) have indicated that the N-terminal ACC domains of PKN␥ (Fig. 6A, a) bind to RhoGTPases (e.g. Rac1) and upregulate the activity of the PKN␥ in the presence of the GTPbound form of Rac1. These findings suggest that PKN␥-ACC domains play an important role in regulating the activation of Rac1-dependent PKN␥ kinase. In order to demonstrate the role of PKN␥-ACC domains in regulating keratinocyte signaling, PKN␥-ACC domain cDNA was cloned into a His-tagged expression vector pcDNA 3.1 (Fig. 6A, b) followed by a transient transfection of PKN␥-ACCcDNA into keratinocytes. First, we incubated purified His-tagged PKN␥-ACC domains with GDPand GTP-loaded forms of either Rac1-or RhoA-GST-conjugated beads. Bound proteins were detected by immunoblotting with anti-His antibody (Fig. 6B). Our results indicate that PKN␥-ACC domains interact with GDP-bound RhoA (Fig. 6B, lane 1) and GTP-bound RhoA (Fig. 6B, lane 2) equally well. However, PKN␥-ACC domains appear to preferentially bind to GTPbound Rac1 (Fig. 6B, lane 4) but not GDP-bound Rac1 (Fig. 6B,  lane 3). These results have confirmed the hypothesis that PKN␥-ACC domains are interacting with Rac1 in a GTP-dependent manner but are associating with RhoA in a GTPindependent manner. We have also observed that both endogenous Rac1 (Fig. 6C, a, lane 2) and RhoA (Fig. 6C, b, lane 2) are bound to His-tagged PKN␥-ACC fragments in keratinocytes transfected with PKN␥-ACCcDNA, but no association between Rac1 (Fig. 6C, a, lane 1)/RhoA (Fig. 6C, b, lane 1) and His-PKN␥-ACC fragment is detected in vector-transfected cells using anti-His-mediated immunoprecipitation followed by anti-Rac1 or anti-RhoA-mediated immunoblot, respectively. These results establish the existence of a close association between RhoGTPases (e.g. Rac1 and RhoA) and PKN␥-ACC fragments in vivo.
In order to examine the effects of PKN␥-ACC fragment overexpression on Rac1-RhoA interaction with endogenous PKN␥, we have transfected keratinocytes with His-tagged PKN␥-ACCcDNA (or vector alone). Our data indicate that endogenous PKN␥ is expressed at comparable levels in vector-transfected (Fig. 6D (I), lanes 1-3) and PKN␥-ACCcDNA-transfected keratinocytes (Fig. 6D (I), lanes 4 -6) treated with HA, pretreated with anti-CD44 antibody plus HA, or no HA treatment. Furthermore, we have observed that a low level of endogenous Rac1 and PKN␥ is co-precipi- . Unphosphorylated or phosphorylated cortactin was then mixed with 125 I-labeled F-actin followed by a 30-min incubation at room temperature. Subsequently, the mixture was centrifuged at 25,000 ϫ g for 10 min at room temperature. The supernatant was then collected, and the radioactivity in this fraction was counted. The decrease (or loss) of radioactivity in the supernatant fraction reflects F-actin precipitation because of the cross-linking reaction. The F-actin cross-linking reaction in the presence of unphosphorylated cortactin (in the absence of PKN␥) (control) is designated as 100%. Each assay was set up in triplicate and repeated at least three times. All data were analyzed statistically using the Student's t test, and statistical significance was set at p Ͻ 0.01.  . 6D  (II), a, lane 1, and (II), c, lane 1)) or pretreated with anti-CD44 antibody followed by HA treatment (( Fig. 6D (II), a,  lane 3, and (II), c, lane 3)). Moreover, we have demonstrated that HA is capable of promoting an additional recruitment of endogenous Rac1 (Fig. 6D (II), a, lane 2) into a complex with PKN␥ ( Fig. 6D (II), c, lane 2) in vector-transfected keratinocytes. Although endogenous RhoA is found to be complexed with PKN␥ in vector-transfected keratinocytes treated with no HA (Fig. 6D (II), b, lane 1) or pretreated with anti-CD44 antibody followed by adding HA (Fig. 6D (II), b, lane 3)), very little HA-induced recruitment of endogenous RhoA (Fig. 6D  (II), b, lane 2) into PKN␥ (Fig. 6D (II), c, lane 2) is detected in these cells. These results suggest that the recruitment of endogenous Rac1 (and to a lesser extent RhoA) into PKN␥ in keratinocytes is HA-dependent and CD44-specific. Moreover, our observations indicate that transfection of keratinocytes with PKN␥-ACCcDNA not only causes a significant inhibition of HA/CD44-mediated recruitment of endogenous Rac1 (Fig. 6D (II), a, lane 5) to PKN␥ (Fig. 6D (II), c, lane 5) but also blocks the basal level of endogenous Rac1/RhoA association with PKN␥ ( Fig. 6D (II), a and b, lane 4, and (II), a and  b, lane 6). These findings established the fact that the PKN␥ fragment containing the ACC domains acts as a potent competitive inhibitor for endogenous Rac1 binding to PKN␥ in vivo during HA-mediated CD44 signaling.
Furthermore, we have demonstrated that threonine (but not serine) phosphorylation of PLC␥1 (Fig. 7, a-c, lane 2) is greatly enhanced in vector-transfected keratinocytes treated with HA as compared with those observed in untreated vector-transfected cells (Fig. 7, a-c, lane 1) or pretreated with anti-CD44 antibody followed by HA treatment (Fig. 7, a-c, lane 3). Both PLC␥1 activity (measured by the production of IP 3 ) and intracellular Ca 2ϩ mobilization are also up-regulated in vectortransfected keratinocytes treated with HA (Table IV). In contrast, the level of PLC␥1 phosphorylation, PLC␥1-mediated IP 3 production, and Ca 2ϩ mobilization is relatively low in untreated vector-transfected keratinocytes (Table IV) or pretreated with anti-CD44 antibody plus HA (data not shown). However, overexpression of the PKN␥-ACC fragment by transfecting keratinocytes with PKN␥-ACCcDNA not only significantly reduces the ability of PLC␥1 to respond to HA/CD44stimulated phosphorylation (Fig. 7, a-c, lanes 4 -6) but also down-regulates enzymatic activity (measured by PLC␥1-mediated IP 3 production) and Ca 2ϩ mobilization (Table IV). Therefore, we believe that the ACC fragment of PKN␥ (by interfering endogenous Rac1 association with PKN␥) (Fig. 6D (II), a, lanes  4 -6) functions as a potent dominant-negative mutant that effectively inhibits HA/CD44-induced Rac1-PKN␥ activation and PLC␥1 phosphorylation required for Ca 2ϩ signaling.
In addition, we have determined that both threonine and serine phosphorylation of cortactin (Fig. 7, d-f, lane 2) are stimulated in vector-transfected keratinocytes treated with HA. In contrast, very little cortactin phosphorylation is observed in untreated vector-transfected cells (Fig. 7, d-f, lane 1) or pretreated with anti-CD44 antibody followed by HA treatment (Fig. 7, d-f, lane 3). Transfection of keratinocytes with the His-tagged PKN␥-ACC fragment (Fig. 7g, lanes 4 -6) effectively inhibits HA-mediated and CD44-dependent cortactin phosphorylation (Fig. 7, d-f, lanes 4 -6). Further analyses indicate that unphosphorylated cortactin (treated by PKN␥ kinase isolated from vector-transfected cells in the absence of HA) promotes F-actin cross-linking activity in vitro (Table IV). In contrast, PKN␥ kinase-phosphorylated cortactin (treated by PKN␥ kinase isolated from vector-transfected cells in the presence of HA) significantly reduces its ability to cross-link F-actin (Table  IV). Moreover, we have shown that unphosphorylated cortactin (isolated from keratinocytes transfected with PKN␥-ACC fragment cDNA in the presence or absence of HA) retains its F-actin cross-linking properties (Table IV). These results suggest that the ACC fragment of PKN␥ (by blocking Rac1 binding to PKN␥) (Fig. 6D (II), a, lanes 4 -6) acts as a dominantnegative mutant that down-regulates HA/CD44-induced Rac1-PKN␥ activation and cortactin phosphorylation required for cytoskeleton function.
By using double immunofluorescence staining and confocal microscopic analyses, we have determined that CD44 is expressed on the cell surface of vector-transfected keratinocytes (Fig. 8A, a) and that PKN␥ kinase (Fig. 8A, c) is diffusely distributed in both the cell membrane and the cytoplasm of untreated vector-transfected keratinocytes. A low level of CD44-PKN␥ co-localization is detected in these keratinocytes in the absence of HA (Fig. 8A, e). We have also observed that both PLC␥1 (Fig. 8B, c) and cortactin (Fig. 8C, c) are distributed in the cytosol and not complexed with CD44 (Fig. 8, B, a,  and C, a) in the plasma membrane in untreated vector-transfected keratinocytes. Very little CD44 co-localization with PLC␥1 (Fig. 8B, e) and cortactin (Fig. 8C, e) was observed in these untreated cells. Upon HA addition, keratinocyte cell-cell adhesion occurs (Fig. 8A, b, d, and f) and there is a dramatic recruitment of PKN␥ (Fig. 8A, d) from the cytosol to CD44containing plasma membranes (Fig. 8A, b and f) in vectortransfected keratinocytes. These results demonstrate that HA induces a close association between CD44 and PKN␥ in the plasma membrane of vector-transfected keratinocytes. Both PLC␥1 (Fig. 8B, d) and cortactin (Fig. 8C, d) are also cocolocalized with CD44 (Fig. 8, B, b, d, and f, and C, b, d, and f) in the plasma membranes and cell-cell adhesion junction in vector-transfected keratinocytes treated with HA. In contrast, we have noted that overexpression of PKN␥-ACC by transfecting keratinocytes with PKN␥-ACCcDNA significantly reduces HA-mediated cell-cell adhesion (Fig. 8, D-F; Tables V-VII) and PKN␥ (Fig. 8D, c and d) co-localization with CD44 (Fig. 8, D, a  and b, and D, e and f). Similarly, the interaction between CD44 (Fig. 8, E, a and b, and F, a and b) and PLC␥1 (Fig. 8E, c and d)/cortactin (Fig. 8F, c and d) at the cell-cell adhesion site (Fig.  8, E, e and f, and F, e and f) is greatly inhibited in these transfectants in the absence or the presence of HA. Treatment of keratinocytes with rat anti-CD44 antibody (but not normal rat IgG) or various agents such as U73122 (a PLC inhibitor), Xestospongin C (an IP 3 receptor blocker), BAPTA (a membrane-permeable Ca 2ϩ chelator), or cytochalasin D (a microfilament disrupting agent known to prevent actin polymerization) effectively inhibits HA/CD44-induced cell-cell adhesion (Tables V-VII). Transfection of keratinocytes with PKN␥-ACCcDNA also readily impairs HA-dependent cell-cell adhesion of keratinocytes (Tables V-VII). These observations strongly suggest that CD44-mediated PKN␥ activation together with PLC␥1-regulated Ca 2ϩ signaling and cytoskeleton function are involved in HA-dependent cell-cell adhesion in keratinocytes.
Expression of both involucrin and transglutaminase has been closely associated with the progression of keratinocyte differentiation (1,2,(51)(52)(53). Immunoblot analyses using antiinvolucrin and anti-transglutaminase antibody indicate that the binding of HA to CD44 promotes the expression of early differentiation markers such as involucrin (Fig. 9A, a, lanes 1  and 2) and a terminal differentiation marker such as transglutaminase (Fig. 9A, b, lanes 1 and 2). The fact that treatment of keratinocytes with rat anti-CD44 antibody (but not normal rat IgG) (Fig. 9A, a and b, lanes 3 and 4) or various drugs (e.g. the PLC inhibitor U73122 (Fig. 9A, a and b, lanes 5 and 6), the IP 3 receptor blocker Xestospongin C (Fig. 9, a and b, lanes 7 and 8), Ca 2ϩ chelator BAPTA (Fig. 9, a and b, lanes 9 and 10), and cytoskeleton drug cytochalasin D (Fig. 9, a and b, lanes 11  and 12)) also effectively blocks HA and CD44-mediated keratinocyte differentiation (Fig. 9A). These results indicate that PLC, IP 3 receptor-mediated Ca 2ϩ activity, and cytoskeletal a The procedures for measuring PLC␥1-mediated IP 3 production, intracellular Ca 2ϩ mobilization, and cortactin-F-actin binding using keratinocytes transfected with PKN␥-ACCcNA or vector alone were described under ''Materials and Methods.'' Each assay was set up in triplicate and repeated at least three times. All data were analyzed statistically using the Student's t test, and statistical significance was set at p Ͻ 0.01. function are all required for HA/CD44-dependent keratinocyte differentiation. Finally, we have observed that HA stimulates the expression of both involucrin (Fig. 9B, a, lanes 1  and 2) and transglutaminase (Fig. 9B, b, lanes 1 and 2) in vector-transfected keratinocytes. Transfection of keratinocytes with PKN␥-ACCcDNA inhibits HA-dependent and CD44-specific expression of differentiation markers such as involucrin (Fig. 9B, a, lanes 3 and 4) and transglutaminase (Fig. 9B, b, lanes 3 and 4). Taken together, we conclude that the ACC fragment of PKN␥ (by interfering with the endogenous Rac1 association with PKN␥) (Fig. 6D (II), a, lanes 4 -6) functions as a potent dominant-negative mutant that effectively inhibits HA/CD44-induced Rac1-PKN␥ activation and PLC␥1/cortactin phosphorylation required for Ca 2ϩ and cytoskeleton-regulated keratinocyte cell-cell adhesion and differentiation. DISCUSSION HA is a well known constituent of connective tissue extracellular matrices and, in particular, is abundant in stratified squamous epithelia including the epidermis of skin (1)(2)(3). HA is often anchored to CD44, a structurally/functionally important surface receptor that contains HA-binding site(s) (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12). Transgenic mice expressing an antisense CD44 construct in their skin showed a significant reduction in endogenous HA on the keratinocyte cell surface and loss of keratinocyte functions (54). These findings suggest that both HA and CD44 play an important role in normal epidermal physiology and keratinocyte differentiation. However, current knowledge concerning the mechanism of HA-CD44 interaction in epidermal keratinocytes and the role that this interaction may have in keratinocyte functions and epidermal physiology remains not well understood.
Previous studies by Bikle and co-workers (40) have shown that a multiphasic increase in intracellular Ca 2ϩ concentration following the addition of exogenous 1-2 mM Ca 2ϩ to keratinocytes is required for Ca 2ϩ -induced keratinocyte differentiation. Inhibition of intracellular Ca 2ϩ mobilization by treatment with the Ca 2ϩ chelator BAPTA prevents Ca 2ϩinduced differentiation (41). These findings suggest that intracellular Ca 2ϩ mobilization plays a critical role in regulating keratinocyte differentiation (40,41). The binding of HA to CD44 increases intracellular Ca 2ϩ levels in many cell types including keratinocytes (Fig. 3). The mechanism by which keratinocytes respond to HA/CD44-induced Ca 2ϩ signaling is not fully established. PLC␥1 is known to generate 2-s messengers from the hydrolysis of phosphatidylinositol 4,5bisphosphate, diacylglycerol, and IP 3 . Diacylglycerol is an activator of protein kinase C, and IP 3 interacts with intracellular membrane receptors (e.g. IP 3 receptors) leading to an increased release of stored Ca 2ϩ ions (37)(38)(39). Most important, targeted deletion of PLC␥1 causes embryonic lethality in mice (55). These findings suggest that PLC␥1 plays a critical role in Ca 2ϩ signaling and developmental processes. In fact, keratinocytes are known to possess functionally active inositol lipid signaling systems, one of which involves PLC␥1 activation (40). During growth factor signaling, PLC␥1 binds to tyrosine autophosphorylation sites of the receptor-linked tyrosine kinases leading to tyrosine phosphorylation and stimulation of PLC␥1 activity (56). Because tyrosine phosphorylation of PLC␥1 and PLC␥1 activity (measured by PLC␥1-mediated IP 3 production) do not correlate well in certain case(s) (38), we were prompted to explore additional mechanisms including post-translational modifications of PLC␥1 by certain serine/threonine kinase(s), such as the PKN family of kinases, during HA/CD44-mediated activation in keratinocytes.
PKN belongs to a family of serine/threonine kinase that displays high sequence homology to that of the protein kinase C family members (29). There are at least three different isoforms of PKN (PKN␣ PAK-1/PRK-1, PKN␤, and PKN␥/PRK2) detected in mammals (29). PKN isoforms appear to be involved in a variety of cellular functions (29 -31). PKN␥ is a serinethreonine kinase known to interact with Rac1 in a GTP-dependent manner and shares a great deal of homology to protein kinase C in the C-terminal region (29). PKN␥ also has a unique regulatory region containing ACC domains that bind to the small GTPases (29 -31). In a previous study PKN␥ has been shown to be involved in Fyn kinase activity and the cell-cell adhesion in mouse keratinocytes treated with exogenously added Ca 2ϩ (33). By using a PKN␥-specific antibody, we have confirmed the presence of PKN␥ in human keratinocytes (Fig.  1). Furthermore, we have demonstrated that PKN␥ (by binding to activated Rac1) is capable of inducing marked threonine phosphorylation of PLC␥1 (isolated from keratinocytes) in vitro (Fig. 2). The ability of PKN to phosphorylate PLC␥1 in the presence of unactivated Rac1 appears to be greatly reduced (Fig. 2). These results clearly indicate that PKN acts as one of the downstream effectors of Rac1 signaling and utilizes PLC␥1 as one of its cellular targets. Both Rac1 signaling (Table I) and PLC␥ phosphorylation in vivo (Fig. 2) can be stimulated by HA and blocked by cells pretreated with anti-CD44 antibody during HA stimulation (Table I and Fig. 2).
The ACC domain of PKN␥ has been shown to be involved in the regulation of GTP-dependent PKN function (29 -31). In Fig. 6, we have detected that the ACC domain is closely associated with Rac1 as a complex. Transfection of keratinocytes with the ACC domain cDNA of PKN␥ (Fig. 6) effectively competes for endogenous activated Rac1 binding to PKN␥ (Fig. 6), inhibits the ability of PKN␥ to phosphorylate PLC␥1 (Fig. 7), blocks HA/CD44-mediated Ca 2ϩ signaling (Table IV), and down-regulates keratinocyte cell-cell adhesion and differentiation (Figs. 8 and 9; Tables V-VII). These results strongly suggest that the ACC domain of PKN␥ is a potent inhibitor of PKN␥-mediated PLC␥1 phosphorylation and Ca 2ϩ signaling required for HA/CD44-mediated keratinocyte function. Our observations are consistent with previous findings suggesting that small GTPases (e.g. Rac1) are important for Ca 2ϩ sensitization in hamster muscle resistance arteries (58). By using dominant-negative and dominant-active forms of Rac1, Hong-Geller and Cerione (44) have also suggested that these small GTPases regulate signal transduction at the level of PLC␥1 and IP 3 production.
The cytoskeletal protein cortactin contains a repeat domain and a C-terminal SH3 domain resembling neufectin, a F-actinassociated protein (59). In fact, cortactin is considered to be an actin-binding protein (48,49). Cortactin not only serves as a cellular substrate for c-Src and Fyn but also localizes with actin at cell adhesion sites during Ca 2ϩ -induced keratinocyte differentiation (60). Tyrosine phosphorylation of cortactin is greatly reduced in keratinocytes derived from Fyn knock-out mice, indicating Fyn kinase may be directly involved in cortactin phosphorylation (60). Our previous study demonstrated that HA activates Src kinase-mediated cortactin phosphorylation and modulates cortactin-F-actin binding in tumor cells (61). In this study, we have determined that cortactin is a cellular substrate for Rac1-activated PKN␥ kinase in keratinocytes (Fig. 4). Serine/threonine phosphorylation of cortactin by Rac1      (and to a lesser extent RhoA)-activated PKN␥ alters its ability to interact with filamentous actin (Fig. 5). In addition, we have confirmed that cortactin phosphorylation by the PKN␥ kinase (isolated from untransfected or vector-transfected keratinocytes in the presence of HA treatment) significantly downregulates its ability to cross-link filamentous actin (Table IV). In contrast, the reduction of cortactin phosphorylation (isolated from keratinocytes transfected with PKN␥-ACCcDNA with or without HA treatment) is sufficient to promote cross-linking of actin filaments into bundles in vitro (Table IV). These results are consistent with previous findings suggesting cortactin plays an important role as a regulator for F-actin-based cytoskeleton function.
To elucidate further the CD44 and PKN␥ interaction with cortactin and PLC␥1 in vivo, we have used immunocytochemical staining and confocal microscopy to monitor morphological changes and the intracellular distribution of various signaling related proteins (e.g. CD44, endogenous PKN␥, PLC␥1, and cortactin) in keratinocytes transfected with PKN␥-ACCcDNA or vector alone. Our results indicate that that HA promotes co-localization of cortactin and PLC␥1 with CD44 and PKN␥ at the plasma membrane region and cell-cell adhesion sites in vector-transfected keratinocytes (Fig. 8). In contrast, overexpression of the dominant-negative form of PKN␥ (PKN␥-ACC fragment) by transfecting keratinocytes with PKN␥-ACCcDNA effectively inhibits HA-mediated recruitment of cortactin, PLC␥1, or PKN␥ kinase to CD44 (Fig. 8) at the cellular mem-branes and impairs HA-induced cell-cell adhesion ( Fig. 8; Tables V-VII). These findings demonstrate that the ability of cortactin/PLC␥1 to be recruited into CD44-associated membrane region and cell-cell adhesion sites is tightly coupled with an active form of PKN␥ kinase in keratinocytes during HA signaling. Treatment of keratinocytes with various inhibitors (e.g. the PLC inhibitor (U73122), the IP 3 receptor blocker (Xestospongin C), the Ca 2ϩ chelator (BAPTA), or the microfilament inhibitor (cytochalasin D)) also displays marked inhibitory effects on HA/CD44-induced cell-cell adhesion (Tables V-VII). These data also strongly suggest that CD44-mediated PKN␥ activation is required for PLC␥1-regulated Ca 2ϩ signaling and cytoskeleton function during HA-dependent cell-cell adhesion in keratinocytes.
The early events of keratinocyte differentiation (shortly after HA/CD44-induced Ca 2ϩ signaling) are accompanied by the expression of keratinocyte differentiation markers including involucrin (precursor for the cornified envelope). Within 24 -48 h after HA binding to CD44, transglutaminase is also expressed and activated (Fig. 9). One of the mechanisms by which intracellular Ca 2ϩ triggers differentiation during HA-CD44 interaction may involve the Ca 2ϩ -dependent activation of protein kinase C isoforms (e.g. protein kinase C␣, -␦, -⑀, -, andisoforms) (62), which regulate various signal cascades and serve as regulators of keratinocyte differentiation-dependent gene expression (e.g. involucrin and/or transglutaminase gene expression) (63-67). HA/CD44-induced intracellular Ca 2ϩ may also be required for activating Ca 2ϩ -dependent enzymes such as transglutaminase. In fact, Ca 2ϩ -activated transglutaminase has been shown to catalyze the isopeptide bond between peptide-bound glutamine and lysine resulting in the cross-linking of involucrin and other proteins into the insoluble cornified envelope to establish barrier function in the skin (38, 68 -70). In addition, HA/CD44-induced Ca 2ϩ mobilization may contribute to cytoskeleton reorganization and cell-cell adhesion for the maintenance of epidermal integrity. The fact that the dominant-negative form of PKN␥ (PKN␥-ACC domains) and various inhibitors (e.g. PLC inhibitor, IP 3 receptor blocker, Ca 2ϩ chelator, and cytoskeleton inhibitor) can prevent HAand CD44-mediated keratinocyte differentiation (Fig. 9) suggests that PLC␥1 activation, Ca 2ϩ signaling, and cortactincytoskeleton function regulated by Rac1-dependent PKN␥ play important roles in HA/CD44-induced keratinocyte differentiation.
These results taken together allow us to propose the following signaling model: HA-CD44 interaction (Fig. 10, step 1) with Rac1-dependent PKN␥ kinase activation (step 2) promotes phosphorylation of PLC␥1 phosphorylation (step 3a) and cortactin (step 3b), leading to the onset of PLC␥1-mediated IP 3 production/Ca 2ϩ signaling (step 4a) and cortactin-mediated cytoskeleton function (step 4b) required for keratinocyte functions (e.g. cell-cell adhesion and differentiation) (step 5) (Fig.  10). We believe that the successful identification of these HA/ CD44-mediated signaling pathways has provided valuable new insight toward a better understanding of keratinocyte functions and normal epidermal development.