Cloning and characterization of cell adhesion kinase beta, a novel protein-tyrosine kinase of the focal adhesion kinase subfamily.

A second protein-tyrosine kinase (PTK) of the focal adhesion kinase (FAK) subfamily, cell adhesion kinase β (CAKβ), was identified by cDNA cloning. The rat CAKβ is a 115.7-kDa PTK that contains N- and C-terminal domains of 418 and 330 amino acid residues besides the central kinase domain. The rat CAKβ has a homology with mouse FAK over their entire lengths except for the extreme N-terminal 88 residues and shares 45% overall sequence identity (60% identical in the catalytic domain), which indicates that CAKβ is a protein structurally related to but different from FAK. The CAKβ gene is less evenly expressed in a variety of rat organs than the FAK gene. Anti-CAKβ antibody immunoprecipitated a 113-kDa protein from rat brain, 3Y1 fibroblasts, and COS-7 cells transfected with CAKβ cDNA. The tyrosine-phosphorylated state of CAKβ was not reduced on trypsinization, nor enhanced in response to plating 3Y1 cells onto fibronectin. CAKβ localized to sites of cell-to-cell contact in COS-7 transfected with CAKβ cDNA, in which FAK was found at the bottom of the cells. Thus, CAKβ is a PTK possibly participating in the signal transduction regulated by cell-to-cell contacts.

Protein-tyrosine kinases (PTKs) 1 that do not span the plasma membranes (so-called nonreceptor PTKs) have been classified into different subclasses (subfamilies) based on the sequence similarity and distinct structural characteristics (1). Many nonreceptor PTKs participate in cellular signal transduction by associating with the intracellular portions of transmembrane receptors which do not themselves have PTK activity. Different nonreceptor PTKs play diverse and specific roles in mediating the signal transduction by different nonkinase receptors (2)(3)(4).
Focal adhesion kinase (FAK) has been proposed as the prototype (and hitherto the sole member) of a new subfamily of nonreceptor PTK, represented by proteins with large N-and C-terminal domains flanking the catalytic domain but without Src homology 2 and 3 (SH-2 and SH-3) domains (5)(6)(7)(8)(9). FAK is concentrated in focal adhesions (5,6), and its phosphorylation and activation are triggered by the ligand binding to integrins and by the stimulation of certain growth factor and neuropeptide receptors (6, 10 -24). The N-and C-terminal domains of FAK mediate its interactions with integrins, the Src-family kinases and paxillin, a focal adhesion associated protein (8,9,(25)(26)(27)(28). By these and other yet to be characterized interactions, FAK regulates signaling via different receptors. Because only one member of the FAK subfamily is known to date, we sought to identify a second PTK of the FAK subfamily by a homologybased cDNA cloning strategy. We describe here an isolation and characterization of a cDNA coding for a new member of the FAK family. The novel PTK described here is the second member, to our knowledge, of the FAK subfamily whose cDNA has been cloned and sequenced and is designated CAK␤ for cell adhesion kinase ␤.

Amplification of PTK Catalytic Domain cDNA Fragments by PCR-
PTK cDNAs were amplified from adult rat brain RNA by reverse transcriptase-directed PCR. PCR primers were designed to recognize conserved regions in PTK catalytic domains: upstream "EcoRI- where N ϭ (A/C/G/T). RNA extracted from rat brain was reverse transcribed with the downstream primers and the Rous associated virus 2 (RAV-2) reverse transcriptase following the conditions of the manufacturer (Perkin-Elmer) in a 20-l reaction. PCR was performed on the reverse transcriptase reaction product in a 50-l reaction containing 50 mM KCl, 10 mM Tris-HCl (pH 8.3), 1.5 mM MgCl 2 , 100 g/ml gelatin, 0.2 mM dNTPs, 1.25 units of Taq polymerase (Perkin-Elmer), and 50 pmol each of the upstream and downstream primers. The thermocycling parameters used in PCR were as follows: annealing, 2 min at 55°C; extension, 2 min at 72°C; denaturation, 1 min at 94°C. After 30 cycles, amplified cDNA products were digested with EcoRI and BamHI and electrophoretically separated on a 3% low melting agarose gel. An ethidium bromide-stained band at about 210 base pairs was cut out. The DNA was extracted from the gel and subcloned into pBluescriptII SK(ϩ). Nucleotide sequences were determined for 100 inserts by the dideoxynucleotide chain termination method (29) using the BcaBEST dideoxy sequencing kit (Takara Shuzo, Otsu, Japan) and the Sequenase 2.0 kit (U. S. Biochemical Corp.), and compared with those in Gen-Bank TM data base by the BLASTx program of NCBI (National Center for Biotechnology Information, Bethesda, MD).
Isolation of cDNA Clones Encoding CAK␤-A clone, M9-3, isolated from the PCR library was labeled with [␣-32 P]dCTP (Amersham Corp.) using a random primer labeling system (BcaBEST labeling kit, Takara Shuzo). The labeled probe was used to screen an oligo(dT)-and randomprimed adult rat brain cDNA library constructed in the ZAP II vector (Stratagene, La Jolla, CA). Seven positive phage plaques were identified. To obtain clones covering the entire 4.0-kilobase transcript, it was necessary to rescreen the library with probes derived from the 5Ј-and 3Ј-ends of the initial cDNA isolates: 34 additional CAK␤ cDNA clones were obtained by screening about 8 ϫ 10 5 independent clones. Nucleotide sequences were determined on both strands for selected overlap-ping clones and their derivatives prepared by exonuclease III/mung bean nuclease deletions to obtain the composite sequence (see "Results"). Human CAK␤ cDNA was cloned by screening a human hippocampus cDNA library constructed in ZAP II vector (Stratagene, La Jolla, CA) with a probe derived from an ApaI/SacI fragment (nucleotides 60 -544) of rat CAK␤ cDNA.
Northern Analysis of Expression-Total RNA was extracted from the tissues of adult rat (Sprague-Dawley strain) and the indicated cell lines using ISOGEN kit (Nippon Gene, Toyama, Japan) according to the manufacturer's protocol. RNA samples were electrophoresed through a 1.0% agarose, 2% formaldehyde gel and transferred to a nitrocellulose membrane. Hybridization to 32 P-labeled fragments of CAK␤ cDNA, FAK cDNA, and actin cDNA was carried out in 50% formamide, 5 ϫ SSPE (1 ϫ SSPE ϭ 0.18 M NaCl, 10 mM sodium phosphate, pH 7.7, 1 mM EDTA), 5 ϫ Denhard's solution, 5 mM EDTA, 0.1% SDS, and 100 g/ml denatured salmon sperm DNA at 42°C for 14 -16 h. The filters were washed (final wash: 0.2 ϫ SSC and 0.1% SDS; 1 ϫ SSC ϭ 0.15 M NaCl, 15 mM sodium citrate, pH 7.6) under conditions of either high stringency (final wash at 55°C for 1 h) or low stringency (final wash at 43°C for 1 h) as indicated in the figure legends. All DNA probes were radiolabeled by random priming. After hybridization, all blots were exposed to Kodak XAR film with an intensifying screen at Ϫ80°C. cDNA probes were derived from StyI fragments of the CAK␤ cDNA (nucleotides 74 -935 and 2990 -3519, which are the 5Ј-and 3Ј-terminal regions). The expression of FAK was detected by hybridizing a probe derived from rat FAK cDNA, corresponding to the amino acid residues 342-600 of mouse and human FAKs (6,7). The rat FAK cDNA was cloned from rat brain PCR library prepared by the use of degenerated PCR primers designed from common amino acid sequences of FAK and CAK␤. The actin probe was prepared from human ␤-actin cDNA.
Production of Antiserum to CAK␤ and Affinity Purification of the Antibody-Digestion of rat CAK␤ cDNA (clone 24) with SphI and PstI restriction endonucleases generated a 688-base pair fragment encompassing nucleotides 2333 through 3020 of the CAK␤ cDNA. This fragment, encoding amino acid residues 779-1008 of CAK␤, was inserted into pATH21 vector (ATCC 37701) (30) doubly digested with SphI and PstI at the polylinker site. Escherichia coli RR1 (ATCC 31343) transformed by this constract was grown and then induced to produce a TrpE-CAK␤ fusion protein (30). The bacteria were lysed by sonication and the TrpE-CAK␤ fusion protein was purified by SDS-PAGE. The fusion protein was electroblotted onto a PVDF membrane (Immobilon, Millipore) and located by staining with Commassie Blue. The portion of the membrane where the fusion protein was located was broken to a powder in liquid nitrogen and used to prepare a water-in-oil emulsion in an adjuvant. Polyclonal antibodies directed against CAK␤ were prepared by immunization of New Zealand White male rabbits with the antigen. The antibody was affinity-purified by binding to a glutathione S-transferase fusion protein containing the CAK␤ C-terminal domain and by eluting with 0.5 M ammonium hydroxide, 3 M sodium thiocyanate (pH 11.0).
Immunoprecipitation of CAK␤ and FAK-Confluent monolayer cul-tures of cells in 9-cm dishes were washed twice with phosphate-buffered saline (PBS) and then lysed on ice in 0.5 ml per dish of a lysis buffer (20 mM Tris-Cl (pH 7.4), 150 mM NaCl, 2.5 mM EDTA, 1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS, 10% glycerol, 1% Trasylol, 20 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, 50 mM NaF, 1 mM Na 3 VO 4 , 20 mM Na 4 P 2 O 4 ). A 2.5% rat brain lysate was prepared in the lysis buffer by the use of a Teflon pestle in a glass homogenizer. The lysates were subjected to centrifugation at 15,000 ϫ g for 20 min at 4°C to obtain clarified lysates. CAK␤ was immunoprecipitated by mixing anti-CAK␤ bound to protein A-Sepharose with 1 mg of protein of clarified lysates and incubating for 2 h at 4°C on a rotating platform. The anti-CAK␤ beads were prepared for each assay by mixing 2 g of affinity-purified anti-CAK␤ protein with 10 l (packed volume) of protein A-Sepharose and washing the Sepharose beads with the lysis buffer. As a control, preimmune serum beads were prepared for each assay by mixing 10 l of preimmune serum with 10 l of protein A-Sepharose. Four g protein of anti-FAK monoclonal antibody, 2A7, bound to 10 l (packed volume) of anti-mouse IgG-agarose were used to immunoprecipitate FAK from 1 mg of protein of clarified lysates. Two g of protein of anti-epitope tag monoclonal antibody bound to 10 l (packed volume) of anti-mouse IgG-agarose were used to immunoprecipitate epitope-tagged CAK␤ from 1 mg of clarified lysate protein.
Immunoprecipitates were washed three times with the lysis buffer, and proteins were separated by SDS-PAGE according to the method of Laemmli and Favre (37). The separated proteins were blotted onto PVDF membranes (Immobilon-P, Millipore, Bedford, MA). The membranes were blocked with 3% bovine serum albumin in TBST (25 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 0.05% Tween 20) for 30 min at 60°C and then probed with the indicated primary antibody in TBST containing 1% bovine serum albumin for 1 h at room temperature. Affinity-purified anti-CAK␤ antibody and anti-FAK polyclonal antibody were used at 1 g of protein per ml, and anti-epitope tag antibody was used at 0.4 g of protein per ml. The membranes were washed with TBST three times and probed again with a second antibody conjugated with alkaline phosphatase in TBST for 1 h, followed by washing three times in TBST. Positive bands were detected by incubating in nitro blue tetrazolium (Sigma) and 5-bromo-4-chloro-3-indolyl phosphate (Sigma).

Transient Expression of CAK␤ in COS-7 Cells-
The CAK␤ cDNA clone, 17N, was digested with EcoRI and subcloned into the simian virus 40-based expression vector pSRE (38), which was derived by a modification of the original pcDL-SR␣-296 vector (39). The resulting construct, pCAK␤(S), and a control construct, pCAK␤(AS), in which the CAK␤ cDNA was subcloned in an inverted, antisense direction, were transfected into subconfluent COS-7 cells at 10 g/9-cm dish using DEAE-dextran (40). After 3 days, the cells were harvested and lysed on ice in the lysis buffer and the lysate was analyzed by immunoprecipitation.
The epitope-tagging vector, pHSV-Tag, was created by ligating a 50 mer oligonucleotide (Novagen, Madison, WI) encoding the 11 amino acid peptide (QPELAPEDPED) derived from herpes simplex virus glycoprotein D, followed by a termination codon, into pT7Blue-T (Novagen, Madison, WI) in a sense direction. The CAK␤ cDNA clone 17N was subcloned into pHSV-Tag by using a strategy that resulted in the epitope with additional N-terminal three amino acid residues, YGL, replacing the C-terminal residue, E, of CAK␤. For expression in vivo, the derivative was subcloned into pSRE to obtain pCAK␤Tag. The 11-residue epitope tag is specifically recognized by the anti-epitope tag monoclonal antibody.
Immune Complex Kinase Assay-CAK␤, epitope-tagged CAK␤, and FAK were immunoprecipitated from the clarified lysates (0.4 mg of protein) of COS-7 cells, transfected with cDNA constructs, and 3Y1 cells as described above. The immune complexes were washed twice with 0.5 ml of the lysis buffer, once with 20 mM Tris-HCl (pH 7.4) containing 0.5 M LiCl, once with 20 mM Tris-HCl (pH 7.4) containing 1 mM EDTA, and once with a kinase assay buffer (20 mM Tris-HCl (pH 7.4), 10 mM MnCl 2 , 1 mM dithiothreitol) and suspended in 20 l of the kinase assay buffer containing 5 Ci of [␥-32 P]ATP (4500 Ci/mmol; ICN Pharmaceuticals, Inc., Irvine, CA). After incubation for 20 min at 20°C, the incubation was terminated by the addition of 20 l of 2 ϫ SDS-PAGE sample buffer. A 20-l portion of the 32 P-labeled immune complexes was subjected to SDS-PAGE in a 7.5% gel. The gel was dried, and the 32 Plabeled proteins were made visible by autoradiography.
For assays of protein-tyrosine kinase activity using poly(Glu,Tyr) photo-stimulated luminescence were corrected to reflect the incorporated radioactivity by subtracting the count of an appropriate control set in each series of kinase assays.
Analysis of CAK␤ Phosphotyrosine Content-To study changes in phosphotyrosine content of CAK␤ and FAK in response to trypsinization and plating cells on fibronectin and poly-L-lysine, confluent 3Y1 cells were cultured overnight in a medium containing 0.5% fetal calf serum and harvested by trypsinization, which was stopped by adding soybean trypsin inhibitor (Sigma). The cells were suspended in serumfree Iscove's modified Dulbecco's medium and plated on 9-cm tissue culture dishes coated with either bovine plasma fibronectin (Biomedical Technologies, Stoughton, MA) (0.1 mg/dish) or poly-L-lysine (0.2 mg/ dish). The dishes were incubated at 37°C for 50 min, and cells attached on dishes were analyzed. Cell lysates were prepared and clarified by centrifugation at 15,000 ϫ g for 20 min. The clarified lysates (0.6 mg protein per assay) were subjected to the immunoprecipitation with either anti-CAK␤ beads or anti-FAK beads. After SDS-PAGE and blotting to PVDF membranes, CAK␤, FAK, and phosphotyrosine were detected by probing blots with anti-CAK␤, anti-FAK polyclonal antibody, and anti-phosphotyrosine, 4G10.
Confocal Laser Scanning Microscopy of Immunostained COS-7 Cells-COS-7 cells, grown overnight on glass coverslips, were transfected with the indicated plasmid. After 3 days, the cells were rinsed in PBS and fixed for 15 min at 4°C in 95% ethanol. Fixed cells were preincubated with fetal calf serum for 30 min at 20°C and then incubated with primary antibodies for 1 h at 20°C. The cells were then washed three times for 10 min each in PBS. The secondary antibodies were then applied for 1 h at 20°C. After washing as above, the coverslips were mounted in PBS containing 50% glycerol and 0.02% 1,4diazobicyclo-(2,2,2)-octane (Aldrich), which was added to delay fading of immunofluorescence during microscopy. The primary antibodies used were affinity-purified anti-CAK␤ (used at 20 g of protein/ml), antiepitope tag (used at 40 g of protein/ml), and anti-FAK polyclonal antibody (used at 67 g protein/ml). Secondary antibodies were fluorescein-conjugated goat anti-mouse and swine anti-rabbit immunoglobulins (DAKO-Japan, Tokyo) (used at a 100-fold dilution). Immunofluorescence was imaged with a Bio-Rad MRC-500 confocal laser scanning microscope. The microscope is fitted with 60 ϫ (numerical aperture, 1.4) objectives in connection with a Nikon Optiphot-2 upright fluorescence microscope. Digitalized fluorescence images obtained by illuminating with a 25-milliwatt multiline argon ion laser were filed in a 768 ϫ 512 pixel frame memory. A series of optical sections through each cell was taken at vertical steps of 1 m. Digital image files were stored on an optomagnetic disc and were subsequently recorded on 35-mm film.

RESULTS AND DISCUSSION
Isolation of cDNA Clones Encoding a Second FAK-To identify novel members of the PTK gene family, a PCR-based approach was used. Reverse transcriptase-directed PCR was performed using RNA from adult rat brain and degenerated primers. A mixture of oligonucleotide primers coding for a PTK hallmark sequences allowed us to isolate eight different PTK catalytic domain cDNA fragments. Six of these PCR products coded for already known members of receptor PTKs. A computer-assisted sequence analysis of one other PCR fragment, M9 -3, showed that it encoded a novel amino acid sequence with the conserved residues characteristic of PTKs. Among the known PTKs, the highest homology was found with the PTK catalytic domain of FAK (76% identity in the translated amino acid sequence). M9 -3 is clearly related to but distinct from the cDNA of the mouse, human, and chicken FAKs (5-7); it was evident that M9 -3 does not code for a rat homologue of FAK. M9 -3, a 201-base cDNA fragment, was subsequently used as a probe for screening a commercially available rat brain cDNA library made by random and oligodeoxythymidylic acid (oligo(dT)) primers. The screening of about 6 ϫ 10 5 independent clones allowed us to obtain seven overlapping cDNAs covering the major portion of the mRNA (Fig. 1). The cDNA clones covering 5Ј-and 3Ј-terminal regions of the mRNA were obtained by screening the library with the 5Ј-and 3Ј-portions of the initial overlapping cDNAs as probes. We present in Fig. 2 a composite sequence of 4.05 kilobases deduced from these cDNA fragments, together with the sequence of the protein, which we call CAK␤.
CAK␤ Is a New Member of the FAK Subfamily of PTKs-The combined 4048-base pair cDNA contained a long open reading frame encoding a protein of 1009 amino acid residues with calculated molecular mass of 115,724 Da, which has all the characteristics of a nonreceptor PTK. The open reading frame is flanked by a 5Ј-untranslated sequence of 261 base pairs and a 3Ј-untranslated sequence of 757 base pairs. The 3Ј-extremity of the cDNA contains the polyadenylation signal 5Ј-ATTAAA-3Ј, followed closely by 12 consecutive terminal adenosine residues. The proposed initiation codon at nucleotides 1-3 is of the form 5Ј-GAGAGGATGTCC-3Ј, which represents a suboptimal primary sequence context for initiation of translation with a purine at position Ϫ3, but without a G at position ϩ4 (41). The assignment of ATG at nucleotides 1-3 as the initiation codon is confirmed by the 5Ј-terminal sequence analysis of a human CAK␤ cDNA, which we cloned from human hippocampus cDNA library. The nucleotide sequences are well conserved in the putative coding regions of rat and human CAK␤ cDNAs, resulting in the almost identical N-terminal amino acid sequences of rat and human CAK␤s (Figs. 3 and 4). The 5Јnoncoding regions of rat and human CAK␤ cDNAs contain an in-frame TAG stop codon at positions Ϫ18 to Ϫ16 and the rat and human cDNA sequences clearly diverge from the positions of about Ϫ60 to the 5Ј-extremity (Figs. 3 and 4).
A protein kinase catalytic domain typical of the PTKs (1) and including the sequence identical to that encoded by the M9 -3 PCR fragment encompasses amino acids 419 -679 (Fig. 2). The deduced protein contains a 418 amino acid N-terminal and a 330 amino acid C-terminal noncatalytic domains. A comparison of the CAK␤ cDNA sequence by the BLASTx program of NCBI with those in GenBank TM data base revealed homology of CAK␤ and FAK over their entire lengths, and did not detect any sequence more closely related to CAK␤. Comparison of the deduced amino acid sequence of the encoded protein with those of mouse, human, and chicken FAKs revealed that this cDNA encoded a FAK-related but distinct PTK (Fig. 5). The amino acid sequence and the structural organization of CAK␤ clearly indicate that CAK␤ is a PTK of the FAK subfamily. The unique overall architecture of FAK that the catalytic domain is flanked by large N-terminal and C-terminal domains (5,9) is also found in CAK␤. The amino acid sequence of the catalytic domain of CAK␤ is 60% identical with the catalytic domains of mouse and human FAKs (Fig. 6). The amino acid sequences of the N-and C-terminal domains of CAK␤ are 39 and 40% identical with those of mouse FAK (Fig. 5). As in FAK, CAK␤ contains neither SH-2 nor SH-3 domains. FAK are highly conserved evolutionary between species; human FAK shares 97% amino acid identity with mouse FAK and 95% identity with chicken FAK (7). The result that rat CAK␤ shares only 45% amino acid identity with mouse FAK indicates that CAK␤ is the second PTK of the FAK subfamily. Indeed, we have amplified rat FAK in addition to CAK␤ from rat brain RNA by RT/PCR using degenerated oligonucleotide primers designed from the amino acid sequences common to both FAK and CAK␤. 2 Homology of CAK␤ and FAK-The predicted amino acid sequence of CAK␤ contains isoleucine in one of the PTK-specific peptide sequences, Asp 549 -Ile-Ala-Val-Arg-Asn 554 (Figs. 5 and 6). The isoleucine at residue 550 is characteristic to FAK (5); leucine is found at the analogous position in other PTKs. The valine at residue 552 is unusual since alanine is found at the analogous position in most of the other PTKs including FAK (Fig. 6), with exceptions of several PTKs, which contain threonine or serine. CAK␤ contains other PTK-specific peptide sequences, Pro 588 -Ile-Lys-Trp-Met 592 and Ser 606 -Asp-Val-Trp 609 , and the structural motifs conserved in all protein kinases (1) including an ATP-binding site, three residues predicted to interact with the ␥-phosphate group of the bound ATP, and the catalytic site Asp 549 . In addition to the replacement of conserved leucine by isoleucine at the residue 550 of CAK␤, three other residues highly conserved in most of the other PTK catalytic domains are not conserved in CAK␤ (Fig.  6). Two of these at the residues 536 and 626 of CAK␤ are not conserved in FAK as well. The other one at the residue 612 of CAK␤ is alanine. The corresponding residue in FAK is glycine, the conserved amino acid of this position in the PTK catalytic domain. Conversely, Met 537 of CAK␤ is the residue highly conserved in the PTK catalytic domains but is replaced to leucine in FAK (Fig. 6).
Comparisons of the N-and C-terminal nonkinase domains between CAK␤ and FAK are shown in Fig. 5. Although the amino acid residues 89 -418 of CAK␤ are highly homologous 2 T. Sasaki, K. Nagura, and H. Sasaki, manuscript in preparation.

FIG. 7. Northern hybridization analysis of CAK␤ transcripts in rat tissues.
A 10-g aliquot of total RNA from various rat tissues was fractionated by electrophoresis, transferred to nitrocellulose membranes, and hybridized with rat CAK␤ 5Ј-coding region cDNA probe (top), rat FAK cDNA probe (middle), and ␤-actin cDNA probe (bottom) under conditions of high stringency. The positions of 28 S and 18 S ribosomal RNAs are indicated on the right.

FIG. 8. Northern hybridization analysis of CAK␤ transcripts in cell lines.
A 10-g aliquot of total RNA from the indicated cells was fractionated by electrophoresis, transferred to nitrocellulose membranes, and hybridized with the same probes as in Fig. 7 under conditions of low stringency. Total RNA from rat brain was also analyzed as a control (leftmost lane of each blot). The positions of RNA size markers along with those of 28 S and 18 S ribosomal RNAs are indicated on the right.  (lanes 2 and 4 -8), the preimmune serum (pre) beads (lanes 1 and 3), or the anti-FAK, 2A7 (␣FAK), beads (lane 9). Bound proteins were washed, and two-thirds of each immune complex was subjected to SDS-PAGE in a 7.5% gel. The resolved proteins were transferred to a PVDF membrane and probed with affinity-purified anti-CAK␤ (lanes 1-5), anti-phosphotyrosine antibody, 4G10 (lanes 6 -8), and anti-FAK rabbit antibody (lane 9). Positions of molecular mass markers are indicated on the right. i.p., immunoprecipitation.
(47.6% identity) with the corresponding N-terminal domain of FAK, the sequence of the extreme N-terminal 88 residues of CAK␤ is entirely different from any portion of FAK (Fig. 5). This difference may imply specific binding of CAK␤ to the cytoplasmic domain of some receptors other than integrins. The binding site of FAK to integrins has been identified in the N-terminal domain (8).
In FAK, the tyrosine 397 at the juncture of the N-terminal and catalytic domains is the site of autophosphorylation and is the major in vivo and in vitro site of tyrosine phosphorylation (28). This phosphorylated tyrosine and the sequence around it are the binding site for a SH-2 domain of the Src family PTKs to FAK (27). The sequence around the Tyr 397 is Glu-Thr-Asp-Asp-Tyr 397 -Ala-Glu-Ile in chicken, mouse, and human FAKs (5-7). A homologous sequence, Glu-Ser-Asp-Ile-Tyr 402 -Ala-Glu-Ile, is found in CAK␤ at the juncture of the N-terminal and catalytic domains (Figs. 2 and 5). The sequence, Tyr-Ala-Glu-Ile, common to both FAK (Tyr 397 ) and CAK␤ (Tyr 402 ) conforms to a consensus high affinity binding site for the SH-2 domains of the Src family of PTKs (42).
The sequence of the C-terminal domain, a region following the kinase domain, is 46 amino acids shorter in CAK␤ as compared with FAK. In the sequence comparison presented in Fig. 5, three gaps were introduced in the C-terminal domain of CAK␤ to maximize the homology with FAK. As shown in Fig. 5, the C-terminal domain of CAK␤ immediately after the PTK catalytic domain (residues 699 -720, 747-777, and 778 -799) has local homologies with three C-terminal domain stretches of the FAK sequence (residues 861-882, 711-741, and 684 -705) in a reverse order; more C-terminal sequences of FAK are homologous with more N-terminal sequences of CAK␤. It should be noted that the residues 711-741 and 861-882 of FAK are the two most proline-rich stretches in the FAK sequence (5-7). The C-terminal nonkinase region of CAK␤ contains two proline-rich stretches, residues 701-767 and 831-869, where the proline content exceeds 20%. The presence of proline-rich stretches has been recognized as a characteristic element of the FAK C-terminal domain (5). The proline-rich stretches of CAK␤ may possibly function as ligands to the SH-3 domains of some proteins involved in the signal transduction. There is also a proline-rich stretch in the extreme N-terminal region of CAK␤ (residues 18 -30).
The residues 869 -999 of CAK␤ continuous with the C-terminal end of the proline-rich cluster are highly homologous (61.83% identity) with the residues 913-1043 of mouse FAK (Fig. 5). The region targeting FAK to focal adhesion was located to reside within the 159 residues of chicken FAK between amino acid positions 853 and 1012, which correspond residues 851-1011 of mouse FAK (26). Thus the sequence of CAK␤  3, 4, and 7). After 3 days, the cells were washed and then lysed in the lysis buffer. The 3Y1 cell lysate was prepared from a confluent culture (lanes 5 and 6). Proteins were immunoprecipitated from 1 mg protein of the lysates with either preimmune (pre) serum beads (lane 3) or with anti-CAK␤ (␣CAK␤) beads (lanes 1,  2, and 4 -7). Two-third of each immune complex was subjected to SDS-PAGE in a 7.5% gel, and the resolved proteins were transferred to a PVDF membrane and immunoblotted with affinity-purified anti-CAK␤ (lanes 1-5) or with anti-phosphotyrosine, 4G10 (lanes 6 and 7). Positions of molecular mass markers are indicated on the left. Arrow indicates CAK␤. i.p., immunoprecipitation. B, COS-7 cells (two 9-cm dishes) were transfected with either pCAK␤(S) (lanes 1, 2, 5, and 6) or epitopetagged CAK␤ cDNA in pSRE (pCAK␤Tag) (lanes 3, 4, 7, and 8). After 3 days, the cells were lysed and CAK␤ was immunoprecipitated from 1 mg of protein of the lysates with either ␣CAK␤ beads (lanes 1, 3, 5, and 7) or anti-epitope tag (␣Tag) beads (lanes 2, 4, 6, and 8). SDS-PAGE and transfer to a PVDF membrane were done as described above. The membrane were immunoblotted with affinity-purified anti-CAK␤ (lanes 1-4) or with anti-epitope tag (␣Tag) (lanes [5][6][7][8]. Arrow indicates CAK␤. The lower band in lanes 6 and 8 represents the heavy chain of anti-tag antibody.  1, 3, and 5), and from the 3Y1 cell lysate (lane 8). CAK␤ was also immunoprecipitated with the anti-epitope tag (␣Tag) beads from the lysates of COS-7 cells transfected with pCAK␤Tag and pCAK␤(S) (lanes 6 and 4). Immunoprecipitates with preimmune (pre) serum bound to protein A-Sepharose were used as controls (lanes 2 and 7). FAK was immunoprecipitated from the 3Y1 cell lysate with the anti-FAK (␣FAK) beads (lane 9). 0.4 mg of protein of the COS-7 or 3Y1 cell lysate was used for each assay. The immune complexes were subjected to the kinase assay with [␥-32 P]ATP as the substrate. The labeled proteins in the immune complexes were separated by SDS-PAGE and made visible by exposing the gel to a XAR film as described under "Materials and Methods." Positions of molecular mass markers are indicated on the right. i.p., immunoprecipitation; S, sense; AS, antisense.
between positions 845 and 967 may possibly contain the targeting sequence of CAK␤ to a certain submembranous site. On the other hand, the CAK␤ sequence of the extreme C-terminal 10 amino acids, residues 1000 -1009, is not homologous with the C terminus of FAK. It has been reported that a replacement of the extreme C-terminal 13 residues of FAK with an epitope tag blocks paxillin binding to FAK (8). Therefore, CAK␤ may bind not to paxillin but to some other proteins associated with the cytoplasmic side of the surface membrane.
Expression of the CAK␤ Gene Transcripts-We have searched for CAK␤ gene expression in rat tissues by hybridization of the cDNA fragments, 5Ј-coding region and 3Ј-coding/ noncoding regions of the CAK␤ cDNA, to a Northern blot carrying RNA from the following adult rat tissues: whole brain without cerebellum, cerebellum, lung, liver, kidney, spleen, intestine, testis, epididymis, adrenal gland, pancreas, and skeletal muscle. The Northern blots were also probed with a rat FAK probe as a reference. Transcripts of about 4.4 kilobases, almost the same size as FAK mRNA, were detected in whole brain, intestine, kidney, spleen and epididymis (Figs. 7 and 8). The same results were obtained with the CAK␤ 5Ј-and 3Ј-cDNA probes. CAK␤ mRNA is particularly abundant in whole brain without cerebellum. The transcripts are scanty in cerebellum, testis, and adrenal gland; in these organs the FAK gene transcripts are abundant (Fig. 7). The 4.4-kilobase transcripts were also detected in rat fibroblast lines, WFB and 3Y1 (Fig. 8). These cell lines also express mRNA for FAK (Fig. 8). In a human T cell leukemia line, Jurkat, transcripts of 4.6 kilobases were detected with both CAK␤ and FAK probes (Fig. 8). No significant CAK␤ gene transcript was found in mouse fibroblast lines (Fig. 8), BALB/3T3, Swiss/3T3, or NIH/3T3, a monkey cell line, COS-7 (Fig. 8), or rat and mouse neural cell lines (data not shown), PC12, NIE115, and NG108-15. In 3T3 lines and COS-7, the expression of the FAK gene was confirmed (Fig. 8).
Detection of CAK␤ in Rat Brain and 3Y1 Cells-Anti-CAK␤ antiserum was raised by immunizing rabbits with a bacterially expressed TrpE fusion protein containing the extreme C-terminal 230 amino acids of CAK␤ (amino acid residues 779-1008). Anti-CAK␤ was affinity-purified on a column of a covalently bound glutathione S-transferase fusion protein of the CAK␤ C-domain to Sepharose. Anti-CAK␤ specifically immunoprecipitated and immunoblotted a protein of about 113-kDa (equivalent to the calculated mass of CAK␤) from the lysates of rat brain, 3Y1 cells and SR-3Y1 cells, a src-transformed line of 3Y1 (Fig. 9, lanes 2, 4, and 5). In accordance with the calculated molecular masses, the immunochemically identified CAK␤ has a faster mobility in SDS-PAGE than FAK, which was immunoprecipitated from the 3Y1 cell lysate with anti-FAK monoclonal antibody, 2A7 (36), and immunoblotted with polyclonal anti-FAK antibody (Fig. 9, lane 9). Immunoblotting with antiphosphotyrosine revealed a band at CAK␤ on the blotted membrane from a SDS-PAGE gel where the anti-CAK␤ immunoprecipitates from the lysates of rat brain, 3Y1 cells and SR-3Y1 cells were separated (Fig. 9, lanes 6 -8). CAK␤ of SR-3Y1 cells was stained more strongly with anti-phosphotyrosine than CAK␤ of 3Y1 cells, indicating higher in vivo tyrosine-phosphorylation of CAK␤ in the src-transformed cells; compare the CAK␤ band density in lane 7 divided by that in lane 4 with that in lane 8 divided by that in lane 5.
In Vitro Phosphorylation of CAK␤ in Immune Complex Kinase Assays and Demonstration of PTK Activity-Immune complexes formed by incubating cell lysates with anti-CAK␤ and anti-epitope tag were assayed for protein kinase activity with [␥-32 P]ATP as the phosphate doner without adding exogenous acceptor, and the 32 P-labeled immune complexes were analyzed by SDS-PAGE. A protein of about 113-kDa, the size of CAK␤, was found to become 32 P-phosphorylated (Fig. 11). The  3-5, 7, and 8) or to the trypsinized cells (lanes 2 and 6) (Off Dish). The lysates of cells on dishes were prepared either from cells before trypsinization (On Dish), from cells 50 min after plating on fibronectin, or from cells 50 min after plating on poly-L-lysine. A, immunoprecipitation with anti-CAK␤. Immunoblotting was either with anti-CAK␤ (lanes 1-4) or with anti-phosphotyrosine (lanes 5-8). B, immunoprecipitation with anti-FAK. Immunoblotting was either with anti-FAK (lanes 1-4) or with anti-phosphotyrosine (lanes 5-8).

TABLE I
Immune complex kinase assays with poly(Glu,Tyr) as a substrate on lysates of COS-7 cells, transfected with CAK␤ cDNA constructs, and 3Y1 cells CAK␤ was immunoprecipitated with the indicated antibody beads from the lysates of COS-7 cells transfected with pCAK␤(S), pCAK␤(AS), or pCAK␤Tag, and from the 3Y1 cell lysate. FAK was immunoprecipitated with anti-FAK beads from the 3Y1 cell lysate. Protein-tyrosine kinase activity was assayed with poly(Glu,Tyr) as an exogeneous substrate as described under "Materials and Methods." S, sense; AS, antisense.  32 P labeling of the protein was found in the kinase assays of the immunoprecipitate from the 3Y1 cell lysate with anti-CAK␤ (Fig. 11, lane 8), of that with anti-CAK␤ from the pCAK␤(S)transfected COS-7 cell lysate (Fig. 11, lane 3), of that with anti-CAK␤ from the pCAK␤(S)Tag-transfected COS-7 cell lysate (Fig. 11, lane 5), and of that with anti-epitope tag from the pCAK␤(S)Tag-transfected COS-7 cell lysate (Fig. 11, lane 6). The 32 P labeling of the 113-kDa protein was not found in the kinase assays of the control immunoprecipitate with anti-CAK␤ from the pCAK␤(AS)-transfected COS-7 cell lysate (Fig.  11, lane 1), of that with anti-epitope tag from the pCAK␤(S)transfected COS-7 cell lysate (Fig. 11, lane 4), or of the immunoprecipitates prepared by preimmune serum (Fig. 11, lanes 2  and 7). When an immunoprecipitate with anti-FAK from the 3Y1 cell lysate was subjected to the in vitro kinase assay, a 125-kDa protein was 32 P-phosphorylated and tentatively identified as the autophosphorylated FAK (Fig. 11, lane 9). These results indicate that the 113-kDa protein revealed by the in