Inhibition of the calcium-dependent tyrosine kinase (CADTK) blocks monocyte spreading and motility.

Freshly isolated peripheral blood monocytes lack focal adhesion kinase (p125(FAK)) but activate a second member of this kinase family, calcium-dependent tyrosine kinase (CADTK; also known as Pyk2/CAKbeta/RAFTK/FAK2), upon adhesion or stimulation with chemokines. To study the role of CADTK in monocyte adherence and motility, we performed immunocytochemical localization that showed CADTK at the leading edge and ruffling lamellipodial structures in freshly isolated, adhered human monocytes. We next introduced CADTK/CAKbeta-related non-kinase (CRNK), the C-terminal noncatalytic domain of CADTK, into monocytes by electroporation and showed that it inhibited CADTK autophosphorylation. Introduction of the fusion protein glutathione S-transferase (GST)-CRNK also reduced (i) cell spreading, as reflected in a reduced cell area 30 min after adhesion, (ii) adhesion-induced phosphotyrosine increases and redistribution into lamellipodia, and (iii) adhesion-induced extracellular signal-regulated protein kinase (ERK) activation. In control experiments, introduction of GST or GST-C3 transferase (an inhibitor of RhoA GTPase activity) by electroporation did not affect these parameters. Monocytes adhered in the presence of autologous serum were highly motile even after introduction of GST (83% motile cells). However, only 26% of monocytes with introduced GST-CRNK were motile. In contrast, GST-CRNK-treated monocytes were fully capable of phagocytosis and adhesion-induced cytokine gene induction, suggesting that CADTK is not involved in these cellular activities and that GST-CRNK introduction does not inhibit global monocyte functions. These results suggest that CADTK is crucial for the in vitro monocyte cytoskeletal reorganization necessary for cell motility and is likely to be required in vivo for recruitment to sites of inflammation.

Integrin-mediated monocyte adhesion to extracellular matrix or endothelial monolayers results in a dramatic alteration of the cytoskeleton and rapid transcriptional activation of numerous genes required in the inflammatory response (1)(2)(3)(4)(5)(6). Monocytes adhered in the absence of any chemotactic stimuli fail to polarize and maintain a radial symmetry while exhibit-ing continuous but moderate membrane ruffling (7). 1 In contrast, monocytes adhered in the presence of either chemotactic stimuli or autologous serum will rapidly polarize and migrate (9,10). Migration requires both contractile and protrusive events, mediated by changes in the actin cytoskeleton (7). Although several lamellipodia are extended simultaneously, directional movement is obtained by the development of a dominant lamellipodium protruding in response to a concentration gradient (7).
Focal adhesion kinase (p125 FAK ) 2 often coordinates integrinmediated cell migration in nonhematopoetic cells (10,11). After adhesion, p125 FAK is targeted to focal adhesions, where it colocalizes with integrins and becomes activated and phosphorylated on multiple tyrosine residues. p125 FAK serves as both a kinase and as a structural protein capable of binding other proteins in a phosphotyrosine-dependent (e.g. src or Grb2) and -independent (e.g. paxillin) manner (11,12). In an alternatively spliced form of p125 FAK , FRNK, the C-terminal noncatalytic domain has been used to identify the role of p125 FAK in motility, presumably because its experimental overexpression displaces p125 FAK from its site of action. For example, FRNK expression in endothelial cells blocked migration (13). In addition, fibroblasts isolated from p125 FAK Ϫ/Ϫ knockout mice had reduced motility (14), whereas ectopic expression of p125 FAK in p125 FAK -null cells dramatically enhanced cell spreading and migration (15,16). The mechanism by which p125 FAK promotes cell migration is, however, still not clear.
A second member of the p125 FAK family has been identified, the calcium-dependent tyrosine kinase CADTK (17), also known as Pyk2, CAK␤, RAFTK, and FAK2 (18 -21). CADTK and p125 FAK are coexpressed in many epithelial, some mesenchymal, and hematopoetic cells and presumably have complementary functions. In contrast to fibroblasts and epithelial cells, freshly isolated peripheral blood monocytes do not form typical focal adhesions and totally lack p125 FAK transcripts and protein (22,23). Monocytes do express CADTK, (23) a cytoskeletal associated kinase of this family, as a splice variant that deletes 42 amino acids in the first proline-rich domain, reducing the approximate molecular mass from 115 to 110 kDa. This variant has been observed in other cell types as well (24,25). Unlike p125 FAK , which is constitutively phosphorylated in adherent cells, in resting adherent epithelial and mesenchymal cells CADTK is predominantly unphosphorylated. Agonists, which raise intracellular calcium, activators of G-protein-coupled receptors, and other signals such as protein kinase C induce CADTK tyrosine autophosphorylation (17, 18, 26 -28). We and others (23,29) show that adhesion or other stimuli activate monocyte CADTK tyrosine phosphorylation. There are complexities to adherence-induced activation of CADTK, as adherence to tissue culture dishes provides a stronger stimulus than adhesion to fibronectin-coated surfaces (23). CADTK associates with cytoskeletal proteins such as paxillin, Hic-5 (31), and leupaxin (32) through conserved "LD" motifs, (33), found in all three proteins. CADTK activation results in paxillin tyrosine phosphorylation in rat hepatic epithelial (30) and other cells (34,35) and may well be involved in tyrosine phosphorylation of leupaxin (32). CADTK immunolocalization varies by cell type. Immunolocalization demonstrated CADTK in a concentric pattern, like that of actin in rat hepatic epithelial cells, 3 along actin stress fibers and in focal adhesions in rat smooth muscle cells (27) and diffusely throughout the cell when overexpressed in fibroblasts (36).
A number of recent studies report CADTK/Pyk2 tyrosine phosphorylation in cells of hematopoetic origin, e.g. T and B lymphocytes (37)(38)(39), monocytes (23,29), NK cells (34,40,41), granulocytes (42), bone marrow progenitors (43), mast cells (44), megakaryocytes (45), and platelets (46,47). The stimuli that trigger CADTK in these cells are often associated with cell motility or at least cytoskeletal rearrangement. However, the downstream targets of CADTK activation and its role in hematopoetic cell cytoskeletal rearrangement are unknown. In other cell types, CADTK has been implicated in the activation of ERK, JNK, and p38 MAP kinases (18,26,48). Monocytes constitutively express a detectable level of activated p38 MAP kinase and rapidly activate ERK and JNK kinases upon adhesion (6). However these activations are transient in motile monocytes adhered in the presence of autologous serum. 1 The present work supports a role for CADTK in monocyte motility first by showing that CADTK immunolocalized to the leading edge ruffling lamellipodia. Taking a cue from studies of the role of p125 FAK in adhesion and motility and its inhibition by FRNK (36), we inhibited CADTK function by introducing a homologous protein referred to as CRNK (36,49), beginning at an initiating methionine (CADTK Met-685) in a position similar to that of FRNK (36). This encompasses areas of the prolinerich region to the C terminus. In freshly isolated monocytes, GST-CRNK blocked adhesion-induced CADTK autophosphorylation and also resulted in dramatic inhibition of cell spreading, a loss of cell motility, and blockade of tyrosine phosphorylation redistribution. Although GST-CADTK had no effect on JNK activation, GST-CRNK introduction did block adhesioninduced ERK activation. Importantly, introduction of other GST proteins did not produce the same results, and GST-CRNK inhibited monocyte motility without altering monocyte phagocytosis or adhesion-dependent transcriptional activation of cytokine genes. These results suggest that CADTK regulates cytoskeletal reorganization and motility but is not required for all adhesion-dependent functions.

Isolation of Monocytes and Culture
Conditions-Human monocytes were isolated from healthy volunteer donors as described previously (5) by centrifugation through Ficoll/Histopaque 1077 (Sigma) and a Percoll (Amersham Pharmacia Biotech) step gradient (50). Monocytes were adherently cultured in endotoxin-free RPMI 1640 medium at 37°C and 5% CO 2 . To produce polarized cells, monocytes were cultured in 5% autologous serum (9). Monocyte motility was assessed using 20-s frametime lapse microphotometry with a 40ϫ phase objective and the NIH Image Movie program. Monocytes were studied for 70 frames.
Immunostaining-Cells were adhered for 30 min, fixed in 4% paraformaldehyde in PBS (pH 7.5) for 10 min, and permeabilized with 0.2% Triton X-100 for 5 min at room temperature. Cells were blocked with 25% normal goat serum in PBS for 30 min at room temperature, washed in PBS, and incubated with Bodipy FL -phallacidin (Molecular Probes) diluted 1:20 with 25% goat serum in PBS for 30 min at room temperature. For the detection of CADTK, cells were stained with purified rabbit polyclonal anti-CADTK antibody (1:50 -1:200). The anti-CADTK #72 antiserum was raised against amino acids 660 -880 of CADTK expressed as a GST fusion protein. Cells were washed in PBS and incubated with a rhodamine (tetramethylrhodamine B isothiocyanate)-conjugated affinity-purified goat anti-rabbit IgG (Molecular Probes) 1:400 dilution with 25% goat serum in PBS. After extensive washing in PBS, cells were mounted in 50% glycerol in PBS. Specificity of the anti-CADTK antisera was determined by blocking with purified GST-CRNK for 1 h at room temperature before use.
Preparation GST Fusion Proteins-GST fusion proteins were grown and isolated from the expression plasmids pGEX3X (Amersham Pharmacia Biotech), pGEX3X-C3 (generously provided by Dr. K. Burridge, University of North Carolina), or pGEX3X-CRNK as described (6). Protein concentrations were determined by the Bradford method (Bio-Rad). GST-CRNK was labeled with Oregon green 488 (Molecular Probes) according to the manufacturer's recommendations.
Immunoprecipitation-Cells, 5-8 ϫ 10 6 , were adhered, rinsed, and lysed on ice for 30 min in 50 mM Tris (pH 7.5), 150 mM NaCl, 1% Triton X-100, 50 mM NaF, 1 mM sodium vanadate, 10 g/ml leupeptin, 1 g/ml aprotinin, and 10 g/ml phenylmethylsulfonyl fluoride. Lysates were clarified at 10,000 rpm for 30 min at 4°C. Equivalent protein concentrations were incubated with 5 l of anti-CADTK antibody for 1 h at 4°C in the presence of 20 l of protein-A/protein G-agarose beads. Beads were washed 3 times in lysis buffer and resuspended in 30 l of 2ϫ Laemmli sample buffer, boiled, and frozen until use.
Western Transfer Analysis-Whole cell extracts were prepared from 2 ϫ 10 6 cells by direct lysis into Laemmli buffer. Proteins were resolved by SDS-10% polyacrylamide gel electrophoresis and analyzed by immunoblotting using antibodies that detect the active, phosphorylated forms of the ERK, JNK, and p38 MAP kinases. The p38 MAP kinase antibody was obtained from New England Biolabs. Phospho-specific ERK and JNK antibodies were from Promega and used according to the manufacturer's recommendations. Horseradish peroxidase-RC20 (antiphosphotyrosine) was obtained from Transduction Labs.
Phagocytosis Assays-Monocytes, 1 ϫ 10 6 , were adhered in the presence of autologous serum for 30 min and then fed 3 ϫ 10 5 /ml FITCconjugated zymosan A (Saccharomyces cerevisiae) bioparticles (Molecular Probes) for 45 min. Extracellular fluorescence was quenched by the addition of 0.4% trypan blue. The percentage of phagocytosing cells was determined as the fraction of monocytes with internalized fluorescent particles from 500 monocytes counted by fluorescence microscopy. Experiments were performed in triplicate and repeated at least twice.

CADTK Concentrates toward the Leading Edge in Migrating,
Polarized Monocytes-To begin tests of the role of CADTK in monocyte motility, we immunolocalized CADTK under several conditions. When monocytes were adhered under serum-free conditions, they failed to polarize and maintained radial symmetry (Fig. 1A, M-NM). In contrast, monocytes adhered in the presence of 5% autologous serum rapidly polarized, generated multiple small lamellipodia, but then migrated after development of a single large lamellipodium (Fig. 1A, M-M). Retraction fibers and filapodia were only observed in motile monocytes. Indirect immunofluorescence revealed that CADTK localized in the leading edge of motile monocytes and to the actin ruffles of both nonpolarized and polarized monocytes ( Fig. 1, B and E). CADTK was detected diffusely throughout the monocyte cytoplasm in nonmotile cells, although it also concentrated in the nuclear cleft. Regions of intense actin staining showed increased localization of CADTK. Although some monocytes demonstrated increased CADTK localization at the adhered surfaces, CADTK immunoreactivity was most prominent in the cortical ruffles of both polarized and nonpolarized cells (Fig. 1, B and E). CADTK did not apparently localize to podosomes or points of focal contact, retraction fibers, or filapodia. In contrast, as previously shown, (28), CADTK did localize to points of contact in the large enucleated structures, which are presumably derived from the platelet precursor, megakaryocytes (Fig. 1B). CADTK localization was highly specific, as immunostaining with antisera that was absorbed with purified, baculoviral-expressed and purified CRNK, which contains the amino acids used to construct the immunizing GST fusion peptide, failed to detect any immunoreactivity within the ruffles (data not shown). Another group has also recently published data showing that CADTK/Pyk2 localized in the ruffles of lipopolysaccharide-stimulated human monocytes and murine macrophages (51).
CADTK Is Required for Monocyte Spreading, Adhesion-induced Phosphorylation and ERK Activation-Localization of CADTK in the leading edge ruffles reinforced a possible role in migration and monocyte recruitment. To test this hypothesis, a GST fusion protein was generated by cloning the C-terminal region of CADTK, CRNK, (nucleotides 2050 -3027) into the expression vector pGEX3X (36,49). Purified GST-CRNK has a molecular mass of ϳ42 kDa. Coexpression of CADTK and CRNK in 293T cells (49) or expression in rat liver epithelial or rat smooth muscle cells inhibits CADTK activation. 4 Thus, CRNK has dominant negative activity.
Monocytes cannot be transfected by conventional methods. However, high efficiency introduction of recombinant proteins into cells can be achieved by electroporation (6,50). We have recently shown that electroporation is adaptable for primary monocytes with proteins retaining functional activity over time (6).
Monocytes were electroporated with ϳ300 g of fluorescently labeled GST-CRNK to assess the uptake efficiency. Electroporated monocytes were extensively washed in cold medium and cytospin preparations made immediately or cells held at 37°C for various time periods. Although uptake distribution was heterogenous, greater than 90% of monocytes were positive ( Fig. 2A). Monocytes remained viable and retained expression of the fusion protein even after incubation for 90 min. Similar results were obtained with other FITC-labeled proteins including GST-C3 or with FITC-conjugated dextran (data not shown). Western blot analysis of monocytes lysed 30 min after electroporation and washing revealed comparable protein uptake for multiple GST fusion proteins (Fig. 2B), GST at ϳ21 kDa, GST-CRNK at ϳ42 kDa, and the GST fusion protein including the RhoA inhibitor, C3 transferase, at ϳ46 kDa. Both GST-CRNK and GST-C3 show some protein degradation after introduction and adhesion. This was not surprising, as we noted intense lysosomal staining after the monocytes were electroporated with the FITC-labeled GST-CRNK.
To test the functionality of GST-CRNK, we assessed both a direct effect on CADTK phosphorylation and a downstream consequence of CADTK activation. As previously shown (23), CADTK is not phosphorylated in freshly isolated nonadherent monocytes. Adhesion rapidly induced tyrosine phosphorylation of several proteins including CADTK (Fig. 2C). Electroporation per se did not interfere with adhesion-induced CADTK tyrosine phosphorylation; monocytes electroporated with GST alone showed CADTK phosphorylation comparable with that of unelectroporated monocytes (compare adherence, ADH, with GST, Fig. 2D). Electroporation of GST-CRNK, however, inhibited both total adhesion-induced tyrosine phosphorylation as well as tyrosine phosphorylation of endogenous CADTK without affecting total CADTK levels (Fig. 2, C and D). The inhibition of tyrosine phosphorylation was similar to that seen with 20 M genistein.
We recently showed that adhesion rapidly activated ERK and JNK kinases in monocytes (6). As CADTK/Pyk2 has been shown to activate ERK, JNK, and p38 MAP kinases in various cell types (18,26,48), we assessed the effects of GST-CRNK on adhesion-induced MAP kinase activation. GST-CRNK clearly blocked adhesion-induced ERK activation (Fig. 2E). Although GST-CRNK had a slight but inconsistent effect on p38 MAP kinase activation, there was no effect on activated JNK expression (Fig. 2E). Similar results were obtained whether the inhibitory CRNK protein electroporated was from bacteria (GST or GST-CRNK) or from baculovirus-infected insect cells (GFP or CRNK) (Fig. 2E). The effect was specific for CRNK as we detected no modulation of ERK, JNK, or p38 MAP kinase activation in monocytes electroporated with GST-C3 despite profound differences in cell morphology and cell spreading (see below). We recently showed that inhibition of adhesion-induced tyrosine phosphorylation with either genistein or the cellular phosphatase, PTP1B, introduced as GST-PTP1B, can inhibit monocyte spreading (6). Because GST-CRNK also inhibited total tyrosine phosphorylation, we assessed the effects on monocyte adhesion and spreading. Monocytes were adhered and immunostained for tyrosine phosphorylation (Fig. 3A). In control GST electroporated monocytes, tyrosine phosphorylation was seen as a pattern of intense staining around the nucleus and a distinct continuous ring of peripheral staining in the ruffling lamellipodia. Phosphotyrosine staining correlated with that of filamentous actin. In contrast, tyrosine phosphorylation appeared random and diffuse in GST-CRNK monocytes, and these cells lacked both the continuous peripheral edge staining seen in control monocytes and the intense perinuclear phosphotyrosine staining (Fig. 3A). There was little correlation between phosphotyrosine staining and actin localization.
One striking difference between GST-, GST-C3-, and GST-CRNK-electroporated monocytes was the effect on cell spreading. Although GST-CRNK electroporated cells adhered with similar efficiency to those electroporated with GST, spreading was markedly compromised (Fig. 3B). Using the NIH Image software, we quantitated spreading by measuring the total surface area occupied by each cell. Approximately 50% of monocytes electroporated with GST occupied a surface area of at least 2000 pixels. In contrast, less than 20% of GST-CRNK monocytes were well spread (Ͼ2000 pixels), with the majority of GST-CRNK monocytes occupying a surface area of less than 2000 pixels. As a positive control for modifying cell spreading, we electroporated monocytes with the RhoA inhibitor C3 transferase (52). Monocytes electroporated with GST-C3 demonstrated profound cell flattening and increased cell spreading, with almost 70% of cells having an area of Ͼ2000 pixels (Fig.  3B). Thus, electroporation of a GST protein does not, by itself,

FIG. 2. High efficiency uptake of GST-CRNK inhibits adhesioninduced CADTK autophosphorylation and ERK activation.
A, monocytes were electroporated with 300 g of purified FITC-labeled GST-CRNK protein and sequentially washed 4ϫ in cold serum-free media, and cytospin preparations were made. Electroporation of FITClabeled GST-CRNK resulted in Ͼ90% positive cells detectable within minutes. A representative experiment is shown (n ϭ 3). B, monocytes were electroporated with GST, GST-CRNK, or GST-C3 and adhered for 30 min. Western blot analysis was performed on whole cell lysates for expression of the GST-tag. CTRL, control. C, monocytes were electroporated, and Western blot analysis was performed on both nonadhered (NAD) and adhered monocytes (GST-CRNK, Genistein) for total phosphotyrosine. As a control, monocytes were treated with genistein (20 M) for 20 min before adhesion. D, monocytes were electroporated with GST or GST-CRNK and adhered for 30 min, and CADTK was immunoprecipitated (IP) as described previously (20). Equivalent protein concentrations were loaded, and proteins were separated by 10% SDSpolyacrylamide gel electrophoresis. Blots were sequentially probed for phosphotyrosine and CADTK. NAD, nonadherent monocytes; ADH, adherent monocytes; GN/A, GN4 rat smooth muscle cells stimulated with angiotensin II. pTyr, phosphorylated tyrosine. E, adhesion-induced ERK activation was inhibited by GST-CRNK. Monocytes were electroporated with either GST fusion proteins (GST, GST-CRNK, or GST-C3), or with baculoviral-produced recombinant proteins, adhered for 30 min in serum-free medium. Western blot analysis was performed with phospho-specific antibodies for ERK, p38, or JNK MAP kinases. E. coli, Escherichia coli.

FIG. 3. GST-CRNK alters phosphotyrosine localization and cell spreading.
A, GST-CRNK alters phosphotyrosine localization in monocytes away from the peripheral edge. Monocytes were electroporated with GST or GST-CRNK, adhered for 30 min, fixed in 3% paraformaldehyde-sucrose, and immunostained for phosphotyrosine as described under "Materials and Methods." B, monocytes electroporated with either GST, GST-C3, or GST-CRNK were adhered for 30 min, fixed in 3% PFA-S, and stained with hematoxylin-eosin. The area was determined by phase light microscopy using a 40ϫ lens and the NIH Image inhibit cell spreading. In other experiments we showed that cell spreading was neither inhibited by the MEK1/2 (mitogen-activated protein kinase/extracellular signal-regulated kinase kinase 1/2) inhibitor PD 98059 nor the p38 MAP kinase inhibitor SB 203580, suggesting that the absence of activation of these kinases is not directly responsible for the alteration in the cytoskeletal organization seen in GST-CRNK monocytes (data not shown).

CADTK Is Critical for Monocyte Motility but Not Phagocytosis nor Adhesion-induced Cytokine Gene Expression-Mono-
cytes exhibit chemokinetic motility when cultured in the presence of autologous serum (9,10). Because CADTK localized to the leading edge lamellipodia and membrane ruffles and is activated by chemokines (23), we assessed the motility by timelapse photography of monocytes electroporated with either GST, GST-CRNK, or GST-C3 proteins and adhered to glass slides in the presence of 5% autologous serum. Data from an experiment superimposing the first and last frames over 25 min of monocyte motility are displayed in Fig. 4. GST-electroporated cells exhibited motility for greater than 20 min and covered distances of several cell diameters during this time. In contrast, although continuing to ruffle and bleb, GST-CRNKelectroporated monocytes failed to migrate. Quantification of 5 independent experiments revealed that only an average 26.0 Ϯ 17.9% (range 10 -50%) of monocytes electroporated with GST-CRNK exhibited motility. Motility in monocytes electroporated with GST was demonstrable in 82.3 Ϯ 12.5% (range 70 -100%) of cells. Although the leading edges of GST-C3-electroporated monocytes were highly motile, allowing the majority of the cell to move up to several diameters, the tails failed to retract. This indicates that GST proteins that have distinct effects on monocyte cytoskeletal organization also have distinct effects on motility.
To determine whether GST-CRNK was inhibiting other monocyte functions, we examined the role of CADTK in phagocytosis and adhesion-induced gene induction. Electroporated cells were adhered in the presence of autologous serum for 1 h and then fed fluorescent zymosan yeast particles for 45 min, fixed, and analyzed. Random fields were chosen by light microscopy and then subjected to fluorescence to avoid bias in data collection. Although GST-CRNK had marked effects on monocyte spreading and motility, phagocytosis was unimpaired by the electroporated protein (Fig. 5). There was no significant difference between GST-, GST-C3-, and GST-CRNK-electroporated monocytes in the percentage of cells capable of phagocytosis.
Monocyte adhesion and spreading are associated with the rapid (within 5 min) transcriptional activation of a variety of cytokine genes (4 -6). Because these events occur consequential to profound changes in cytoskeletal organization, we were interested in determining whether GST-CRNK modulation of monocyte spreading also interferes with adhesion-induced cytokine gene induction. As previously reported (5, 6), nonadhered monocytes do not spontaneously express interleukin-1␤ mRNA transcripts (Fig. 5B). Adhesion resulted in the induction of interleukin-1␤ mRNA after electroporation with GST, GST-CRNK, or GST-C3. These results suggest that CADTK, whereas important in monocyte cytoskeletal organization and motility, plays a limited, if any, role in the functional activities of phagocytosis or adhesion-induced gene induction.

FIG. 4. GST-CRNK inhibits monocyte motility.
Monocytes were electroporated either with GST, GST-C3, or GST-CRNK and adhered in the presence of 5% autologous serum. Motility was monitored by light microscopy using a 40ϫ objective and the NIH Image 1.6 Movie software. Monocytes were followed over 70 frames. The first and last frames were combined to show cell movement. The arrow denotes direction of motility from the first frame to the last. The asterisk denotes nonmotile cell. A representative of five experiments is shown.

FIG. 5. GST-CRNK does not inhibit monocyte phagocytosis or adhesion-induced gene expression.
A, monocytes were adhered in the presence of 5% autologous serum for 45 min and then fed FITClabeled zymosan bioparticles for 45 min. Cells were washed twice in PBS and incubated with 0.4% trypan blue to quench extracellular fluorescence from unphagocytosed particles. At least 100 cells were counted per sample. Bars represent mean Ϯ S.D. of triplicate samples. A representative of three experiments is shown. B, monocytes were electroporated and adhered in the presence of autologous serum for 45 min, and total RNA was isolated. Northern analysis was performed on 1 g of total RNA/lane, and the blot was probed sequentially for interleukin-1␤ and ␤ -actin.

DISCUSSION
Freshly isolated peripheral blood monocytes do not make typical focal adhesions or actin stress fibers nor do they express p125 FAK . Monocytes do express the closest known orthologue tyrosine kinase, CADTK. These highly related signaling molecules are expressed together in many cell types and presumably have distinct but perhaps complementary functions. The specific biology of monocytes apparently allow it to dispose with p125 FAK and rely on a single cytoskeletal-associated kinase from this family, CADTK. In monocytes, CADTK does not localize to contact points upon adherence but, instead, localizes to the ruffling lamellipodial structures. As shown in Fig. 1 and as recently published by Williams and Ridley (51), CADTK/ Pyk2 localized to the leading lamellipodia and ruffling membranes of adhered motile monocytes (Fig. 1) and lipopolysaccharide-stimulated monocytes and macrophages.
CADTK/Pyk2/CAK␤/RAFTK/FAK2 and p125 FAK are members of a kinase family, at least one of which is expressed in virtually all cell types. For example, most fibroblasts express only p125 FAK , epithelial and neuronal cells express both p125 FAK and CADTK, and monocytes express only CADTK. CADTK is expressed throughout the hematopoetic lineages including monocytes (23,24,29), megakaryocytes (45), platelets (46 -47), NK cells (34, 40 -41), granulocytes (42), and both T and B lymphocytes (37)(38)(39). In monocytic cells, CADTK is predominately expressed as a splice variant that deletes one of the proline-rich regions, yielding a protein of ϳ110 kDa (23). CADTK shares topological similarities with p125 FAK (49), although one striking difference is the lack of localization to focal adhesions when overexpressed in chicken and other fibroblasts (36). Monocytes lack traditional focal adhesions but form focal contacts sites, termed podosomes, when cultured under conditions that stimulate motility. Whereas CADTK localized to the monocyte cortical ruffles, it did not localize to the podosome structures.
CADTK is rapidly phosphorylated upon adhesion in monocytes, megakaryocytes, and platelets and when overexpressed in COS cells (23, 28, 46 -47). Additionally, CADTK can be tyrosine-phosphorylated after engagement of integrins (23,44,53), integrin-dependent platelet aggregation (46), or by activation of either the T cell receptor complex or the B cell antigen receptor complex (37,39,53,54). Activation of CADTK in platelets results in its translocation from the cytosol to the cytoskeleton (46). In most hematopoetic cells, cytoskeletal organization is critical to CADTK tyrosine phosphorylation because treatment with cytochalasin D to disrupt the actin cytoskeleton inhibits CADTK phosphorylation (27,30,53). We have recently shown that monocyte cytoskeletal organization can be dramatically disrupted by the introduction of the phosphatase PTP1B (GST-PTP1B) or by treatment with tyrosine kinase inhibitor, genistein (6). Here we have identified adhesion-induced tyrosine phosphorylation of CADTK as having a regulatory role in monocytic cytoskeletal organization. Inhibition of adhesion-induced CADTK autophosphorylation by the introduction of GST-CRNK blocked cell spreading and disrupted the distribution of tyrosine-phosphorylated proteins in adhered monocytes, presumably by displacing CADTK. We had previously shown that cotransfection of CRNK can block CADTK activation (49). However, transfection is not suitable for introduction of genes into monocytes, and we used electroporation as a means to introduce GST-CRNK into freshly isolated monocytes. FITC-labeling of GST-CRNK and other GST proteins showed the utility and efficiency (Ͼ90% of cells) of this method. The amount of GST-CRNK electroporated into monocytes is easily detectable in lysates and exceeds the amount of CADTK. When immunofluorescence was performed on GST-CRNK electroporated cells, the protein was widely distributed and did not just localize to structures binding CADTK. However, introduction of GST, GST-C3, or GST-PTP1B (6) produced similar or greater levels of protein than GST-CADTK; all served as controls showing the specificity of the GST-CRNK biologic effect. The specific changes in motility are only seen with CRNK and are clearly not simply due to introduction of a range of proteins.
Preliminary experiments indicate that introduction of GST FRNK also blocks monocyte motility. FAK is not expressed in monocytes; however, we have shown using liver epithelial cells, rat smooth muscle cells, and human embryonic kidney cells that over expression of FRNK disrupts both FAK and CADTK signaling. Therefore, it was not a surprise that the high level of GST-FRNK obtained by electroporation blocked CADTK function. We believe that FRNK disrupts the "cytoskeletal engagement" step (23), which we have hypothesized is involved in CADTK activation in cells. Since FAK has no direct role in these cells, FRNK introduction may inhibit some processes simply by binding to CADTK-binding proteins with motifs similar to its other cell binding partners.
The present data, in addition to our previous observations with GST-PTP1B and genistein, highlight the interrelation between adhesion-dependent tyrosine phosphorylation and cytoskeletal organization in monocytes. They suggest that CADTK activation is important in triggering these tyrosine phosphorylation events but do not prove that CADTK alone is responsible. Decreases in adhesion-dependent monocyte tyrosine phosphorylation are also seen in Hck and Fgr double knockout macrophage cells (55). Thus src family members could be upstream or downstream of CADTK in monocyte adhesion-signaling events.
Introduction of GST-CRNK into monocytes not only inhibited CADTK tyrosine phosphorylation but also led to a more global loss and lack of cellular redistribution of other phosphotyrosine-containing proteins. GST-CRNK but not GST interfered with monocyte cytoskeletal reorganization, as indicated by lack of spreading and motility. The identity of the CADTK tyrosine-phosphorylated proteins key to these events is unknown. However, CADTK can interact with and/or phosphorylate several signaling-associated proteins in a variety of cells; these include Grb2, Src family tyrosine kinases, p130 Cas , Crk (56), paxillin (30), and SHPS-1 (57) in numerous cell types including THP-1 promonocytic cells and epithelial cells (29,30). A particularly interesting substrate is the abundant paxillin family member leupaxin, which is present in monocytes and which has been demonstrated to bind to CADTK (32). Leupaxin as a tyrosine-phosphorylated monocyte cytoskeletal protein may be an important target disrupted by the introduction of GST CRNK.
CADTK has also been linked to other signaling pathways including activation of the MAP kinases. Monocyte adhesion resulted in the activation of ERK and JNK kinases (6). Introduction of GST-CRNK blocked ERK activation but not JNK activation, suggesting that in monocytes, CADTK is an upstream mediator of ERK activation. This has also been suggested for other cell types including NK cells (41). The consequences of inhibiting ERK activation in monocytes are not readily apparent. We have previously shown that a loss of ERK activation rapidly effects adhesion-induced cytokine mRNA stability without significantly inhibiting gene transcription (6). Here we have confirmed the observation that ERK activation is dispensable for adhesion-induced transcriptional activation of interleukin-1␤ mRNA by inhibiting its activation through CRNK introduction. Thus, CADTK activation appeared to be dispensable for adhesion-induced gene induction but not for adhesion-induced ERK activation.
We have observed that the pharmacologic inhibitor of MEK1/ 2(mitogen-activated protein kinase/extracellular signal-regulated kinase kinase 1/2), PD 98059, inhibited neither cell spreading nor cell motility despite a loss of ERK activation in primary monocytes 5 or primary bone marrow CD34 ϩ progenitor cells (8). Thus in monocytes, CADTK-induced activation of ERK is probably not a primary mechanism by which CADTK is involved in cell motility.
CADTK overexpression in fibroblasts (cells that do not normally express CADTK) does not result in cytoskeletal association and often results in cell rounding and apoptosis (25). This is despite the facts that CADTK contains a focal adhesion targeting region that closely resembles p125 FAK and that CADTK binds paxillin with similar affinity to that of p125 FAK (30) as well as Hic-5 and leupaxin (31)(32). Clearly there are other complexities to CADTK cytoskeletal association. For example, in rat smooth muscle cells, which normally express both p125 FAK and CADTK, CADTK is localized to the actin cytoskeleton. In mouse embryo fibroblasts derived from FAK Ϫ/Ϫ knockout animals, CADTK/Pyk2 protein is expressed, an unusual occurrence in fibroblasts, and yet motility is not restored by the compensatory CADTK expression (15,16); furthermore, in these cells, CADTK/Pyk2 is localized not to the cytoskeleton but to perinuclear bodies. 3 Thus the ability of CADTK to participate or govern motility appears to require a cell type in which it is capable of localizing to cytoskeletal elements, perhaps through the expression of a specific CADTK cytoskeletal linker protein.
In summary, our results implicate CADTK as a critical intermediary of monocyte motility. Our previous results that introduction of a tyrosine phosphatase PTP1B also inhibits motility (6) and the fact that macrophages derived from mice deficient in the related Src kinases, Hck and Fgr, also exhibited severely retarded motility (56) add to the evidence that tyrosine phosphorylation regulates monocyte motility. The macrophages from Hck and Fgr Ϫ/Ϫ mice failed to localize paxillin, talin, and filamentous actin to the leading lamellipodia, all contributing to reduced motility (56). CADTK is activated by adhesion, autophosphorylates on tyrosine 402 (49), a site that binds src family SH2 groups, and might regulate Hck and Fgr activity. Taken together, our results suggest that the in vivo consequences of preventing CADTK activation and/or function, like that of Hck and Fgr, would be defective localization of monocytes to sites of inflammation.