Cloning and Characterization of Dfak56, a Homolog of Focal Adhesion Kinase, in Drosophila melanogaster*

The focal adhesion kinase (FAK) protein-tyrosine kinase plays important roles in cell adhesion in vertebrates. Using polymerase chain reaction-based cloning strategy, we cloned aDrosophila gene that is homologous to the vertebrate FAK family of protein-tyrosine kinases. We designated this geneDfak56 and characterized its gene product. The overall protein structure and deduced amino acid sequence of Dfak56 show significant similarity to those of FAK and PYK2. Dfak56 has in vitro autophosphorylation activity at tyrosine residues. Expression of the Dfak56 mRNA and the protein was observed in the central nervous system and the muscle-epidermis attachment site in the embryo, where Drosophilaposition-specific integrins are localized. The results suggest that like FAK in vertebrates, Dfak56 functions downstream of integrins. Dfak56 was tyrosine-phosphorylated upon integrin-dependent attachment of the cell to the extracellular matrix. We conclude that the Dfak56 tyrosine kinase is involved in integrin-mediated cell adhesion signaling and thus is a functional homolog of vertebrate FAK.

Focal adhesion kinase (FAK) 1 is a member of a growing family of non-receptor protein-tyrosine kinases (1) and was originally identified as a putative substrate for the oncogenic protein-tyrosine kinase pp60 v-src (2). Accumulating data, however, show that FAK tyrosine phosphorylation is induced upon adhesion of cells to the extracellular matrix through the surface integrin (3) and upon stimulation of cells by a variety of other extracellular factors including those for receptor tyrosine kinases and for G-protein-coupled receptors (4 -6). FAK associates with multiple cellular components including other focal adhesion-associated proteins and signaling molecules (7)(8)(9)(10)(11).
FAK is localized to focal adhesions and is centrally implicated in the regulation of cell motility and adhesion. PYK2, another member of the FAK family of protein-tyrosine kinases (also named CAK␤/CadTK/RAFTK) (12)(13)(14)(15), is isolated from mammals. The function of PYK2 seems to be different from that of FAK because tyrosine phosphorylation of PYK2 is independent of cell adhesion and is not localized to focal adhesions (13). Since calcium ionophores stimulate PYK2 and chelation of extracellular calcium ion abrogates tyrosine phosphorylation of PYK2 in response to various stimuli, PYK2 is thought to be regulated by calcium (12,14).
The roles of FAK in development and cell adhesion are directly revealed by gene targeting experiment (16,17). Targeted disruption of the FAK gene in mice results in early embryonic lethality. Cells from mutant embryos have reduced mobility and increased number of focal adhesions, suggesting that FAK is involved in the turnover of focal adhesion contacts during cell migration. Disruption of the mouse ␤ 1 integrin subunit, one of the well known upstream factors of FAK, also causes embryonic lethality in a very early stage of development (18). However, because of the anomalous phenotypes, further biological analysis of knockout mice remains difficult.
In Drosophila, study of cell adhesion was started by the identification of position-specific (PS) antigens as homologs of vertebrate integrins (19). Several Drosophila genes encoding integrins have been identified, including myospheroid (mys), multiple edematous wings (mew), and inflated (if), which encode the integrin ␤ chain (␤ PS integrin) (20), ␣ PS1 integrin (21), and ␣ PS2 integrin (19), respectively. Genetic analyses of these mutations demonstrate that integrins are required for a wide variety of developmental processes, including germ-band retraction, muscle attachment, and morphogenesis of the wing and eye (22). Furthermore, genetic approaches to isolate the genes encoding possible downstream molecules of integrins have been performed (23).
In this study, we isolated the cDNA for the Drosophila homolog of FAK. Characterization of the protein product strongly suggested that Drosophila FAK is an important downstream factor of integrin-mediated signaling pathways, as is mammalian FAK.

EXPERIMENTAL PROCEDURES
Isolation of cDNA and Genomic DNA Clones-Polymerase chain reaction was performed using fully degenerate oligonucleotides corresponding to the amino acid sequences HRDIAAR (for the forward primer) and DVWAFG (for the reverse primer) as primers and 10 8 plaque-forming units of Drosophila embryo cDNA library (a gift from T. Todo, Radiation Biology Center, Kyoto University, Kyoto, Japan) as * This work was supported by a grant for advanced cancer research from the Ministry of Education, Science, and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence reported in this paper has been submitted to the DDBJ/GenBank TM /EBI Data Bank with accession number D88898.
template. The amino acid sequences were located in conserved motifs of the kinase domain and are unique for FAK and PYK2. Amplified cDNAs (ϳ200 base pairs) were subcloned into the pBluescript vector (Stratagene) and sequenced with Thermosequenase (Amersham Pharmacia Biotech). Replicate filters from a Drosophila embryo cDNA library were prehybridized for 3 h at 65°C in 7% SDS and 0.5 M sodium phosphate buffer (pH 8.0) containing 100 g/ml salmon sperm DNA (Sigma) and then hybridized to the 32 P-labeled cDNA fragment for 16 h at 65°C. After hybridization, the filters were washed with 0.2ϫ SSC and 0.1% SDS. Positive clones were isolated and sequenced on both strands. Other molecular biological techniques were performed as described previously (24). DNA and protein sequence analyses were performed using Genetics Computer Group software packages and ClustalW and BLAST programs employing the DDBJ/GenBank TM /EBI Data Bank running at the Human Genome Center of the University of Tokyo (Tokyo, Japan).
Southern and Northern Blotting-Genomic DNA from wild-type (Oregon-R strain) Drosophila melanogaster (1 g/sample) digested with suitable restriction enzymes was electrophoresed on 1.0% agarose gel. mRNA samples from cultured cells and Drosophila tissues were electrophoresed on formaldehyde-containing 1.2% agarose gel. The fractionated RNAs were then transferred to nylon membranes (Hybond, Amersham Pharmacia Biotech) and subjected to hybridization using full-length Dfak56 cDNA and the 5Ј-portion (SacII fragment, corresponding to nucleotides 1-2743) of the Dfak56 cDNA as probes. Hybridizations were performed employing the same procedure as that used for cDNA screening.
Cells and Antibodies-Drosophila Schneider cells (S2) and derivatives of their integrin-expressing cells were a gift from D. L. Brower (Arizona University). BG2-c6 and 7E10 cells were described previously (25,26). The cells were cultured at 27°C under normal atmospheric conditions in Shields and Sang M3 medium (Sigma) supplemented with 10% of fetal calf serum. Simian COS cells were cultured at 37°C under 5% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum.
To generate rabbit anti-Dfak56 polyclonal antibodies, glutathione S-transferase (GST)-Dfak56 was constructed by inserting the cDNA fragment encoding amino acid residues 1-280 of the Dfak56 protein into the pGEX-2T expression vector (Amersham Pharmacia Biotech). Expression of the GST fusion proteins in Escherichia coli BL21 was induced with isopropyl-␤-D-thiogalactopyranoside, and the proteins were purified with glutathione-Sepharose (Amersham Pharmacia Biotech) as described previously (27). A New Zealand White rabbit was immunized with the purified GST-Dfak56 fusion protein. Antiserum raised against Dfak56 were purified using GST and the GST-Dfak56 protein bound to a HiTrap NHS-activated affinity column (Amersham Pharmacia Biotech). Anti-phosphotyrosine monoclonal antibody 4G10 was purchased from Transduction Laboratories, Inc.
In Situ Hybridization and Immunohistochemistry-Whole-mount in situ hybridization was conducted using digoxigenin-labeled RNA probes as described previously (28). The digoxigenin-labeled RNA probes (sense and antisense) were prepared by a standard procedure with a digoxigenin-RNA labeling kit (Roche Molecular Biochemicals) and fulllength Dfak56 cDNA as a template.
Western Blotting and Immunocomplex Kinase Assay-Cells were lysed in TNNE buffer (10 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1% Nonidet P-40, and 2 mM EDTA) containing 1 mM sodium orthovanadate, 50 units/ml aprotinin, 0.1 mM phenylmethylsulfonyl fluoride, and 10 mM NaF. Insoluble materials were removed by centrifugation. The lysates were separated by 7.5% SDS-polyacrylamide gel electrophoresis and transferred to a polyvinylidene difluoride membrane (Trans-Blot, Bio-Rad). The proteins reacted with anti-Dfak56 antibodies were visualized with a chemiluminescence detection kit (Renaissance, NEN Life Science Products). For immunoprecipitation, the cell lysates were precleared with protein A-Sepharose 4B (Amersham Pharmacia Biotech) for 1 h at 4°C. After removing the beads, the lysates were incubated for 2 h at 4°C with anti-Dfak56 antibodies and protein-A Sepharose. The immune complex was washed several times with TNNE buffer and then with kinase buffer (20 mM HEPES (pH 7.4), 10 mM MgCl 2 , and 10 mM MnCl 2 ). Following addition of [␥-32 P]ATP (370 kBq, 110 TBq/mmol), the immunoprecipitates were incubated for 30 min at 30°C and then separated by 7.5% SDS-polyacrylamide gel electrophoresis. The gels were dried and subjected to autoradiography. Phosphoamino acid analysis of phosphorylated proteins and two-dimensional analysis of tryptic phosphopeptides were performed as described previously (31).

RESULTS
cDNA Cloning and Sequence Analysis of a Drosophila Homolog of Focal Adhesion Kinase-To identify a Drosophila FAK homolog of protein-tyrosine kinase, we used polymerase chain reaction with degenerate oligonucleotide primers to amplify cDNAs from a Drosophila embryo cDNA library. The sequences of the degenerate primers corresponded to the amino acid sequences conserved in both FAK and PYK2. Amplified cDNA fragments were subcloned into plasmids and sequenced. We searched the DDBJ/GenBank TM /EBI Data Bank against these sequences. A cDNA fragment was revealed to contain a novel sequence with homology to FAK and PYK2. The cDNA fragment was subsequently used to screen the same library. Four cDNA clones of varying sizes were isolated, and the one containing the longest insert (4.3 kilobase pairs) was sequenced on both strands. The complete nucleotide sequence of the cDNA predicted an open reading frame of 1198 amino acids (Fig. 1A). The calculated molecular mass of the predicted protein was 130 kDa. Since the cDNA hybridized at 56A or 56B on the second chromosome, which was determined by in situ hybridization with polytene chromosomes (data not shown), we named the gene Dfak56.
The deduced amino acid sequence of the putative product of the Dfak56 gene showed the presence of a protein kinase domain. Like FAK and PYK2, Dfak56 has large N-and C-terminal sequences flanking the kinase domain. Comparison of the amino acid sequence of the Dfak56 kinase domain with the protein sequence data bases revealed that the kinase domain of Dfak56 is most similar to that of FAK (59% identity to human FAK) and PYK2 (52% identity to human PYK2) (Fig. 1B). The sequences of the N-and C-terminal regions are relatively divergent. The amino acid sequence alignment of the kinase domains of these proteins is shown in Fig. 1C. Unlike the vertebrate counterparts, Dfak56 contains an additional 24amino acid sequence near the ATP-binding site of the kinase. In the C-terminal regions, FAK and PYK2 have a conserved sequence of ϳ150 amino acid residues. The sequence is called the focal adhesion targeting sequence because it is essential for localization of FAK to focal adhesions (32). Dfak56 also carries a focal adhesion targeting-like sequence (ϳ40% identity to human FAK) in the C-terminal region (Fig. 1B), suggesting that the focal adhesion targeting sequence is conserved as a functional domain.
There are tyrosine phosphorylation sites within FAK that are conserved in PYK2. A major autophosphorylation site (Y 397 AEI for human FAK and Y 402 AEI for human PYK2) that mediates binding to the SH2 domain of Src-like protein-tyrosine kinases is well conserved in Dfak56 (Y 430 AEI). Another tyrosine phosphorylation site (Y 925 ENV for FAK and Y 881 LNV for PYK2) serves for binding of the GRB2 SH2 domain. Tyrosine 954 in Dfak56 is likely to correspond to this phosphorylation site, although the sequence is slightly divergent (Y 954 CAT). The C-terminal region of FAK has two PXXP motifs, P 712 PKPSRP and P 874 PKKPPRP, which can bind to the SH3 domain. These sequences are highly conserved in PYK2, and only one of them (P 772 PSKPSR, corresponding to P 712 PKPSRP of FAK) is conserved in Dfak56. We conclude that Dfak56 is a Drosophila homolog of vertebrate FAK family protein-tyrosine kinases. The sequence data did not reveal which kinase, FAK or PYK2, is the counterpart of Dfak56.
Southern hybridization analysis strongly suggested that Dfak56 is a single-copy gene of Drosophila ( Fig. 2A). Moreover, under less stringent washing conditions, no bands other than those for Dfak56 were found (data not shown). Therefore, we tentatively concluded that Drosophila contains no close relatives of Dfak56.
Temporal and Spatial Expression of Dfak56 -Northern hybridization analysis of Drosophila poly(A) ϩ -selected RNAs from various developmental stages (embryo, larva, pupa, and adult) and cell lines showed that two mRNA species of ϳ5 and 6 kb hybridized with the Dfak56 cDNA probe (Fig. 2B). A short transcript of ϳ1.5 kb was also detected in embryonic and larval RNA samples. This short transcript did not hybridize to the cDNA probe corresponding to the 5Ј-portion (including the cod-ing region N-terminal to the kinase domain) of the Dfak56 mRNA, suggesting that this transcript corresponds to the 3Јportion of the Dfak56 mRNA. The expression of a short transcript was also reported in the case of the avian FAK gene (33).
To determine the spatial expression of the Dfak56 gene, we performed whole-mount in situ hybridization with the embryos. In stage 16 embryos, strong expression of the Dfak56 mRNA was detected in the central nervous system and the junction of muscle and epidermis (Fig. 3, A and B). Immunohistochemical analysis also showed expression of the Dfak56 protein in the muscle attachment (Fig. 3C). A cross-sectioned Oregon-R strain) were digested by restriction enzymes, blotted, and probed with Dfak56 cDNA. Standard molecular sizes are indicated in kilobase pairs. B, poly(A) ϩ -selected RNAs (5 g each) from Drosophila cell lines (BG2-c6, derived from the central nervous system; and 7E10, derived from body fluid cells) and from embryonic, larval, pupal, and adult (male and female) flies were subjected to Northern blotting and probed with Dfak56 cDNA. Transcripts of ϳ5 and ϳ6 kb were observed in all samples, and a short ϳ1.5-kb transcript was observed only in the embryo and larva using fulllength Dfak56 cDNA as a probe (left panel). This short transcript was not detected with the 5Ј-probe (right panel).
view of the immunostained image suggests that Dfak56 is expressed predominantly in the epidermal cells, but less or not at all in the muscular cells (Fig. 3D).
Subcellular Localization of Dfak56 -We examined the subcellular localization of the Dfak56 protein using the Drosophila neuronal cell line BG2-c6 (25). The BG2-c6 cells expressed the Dfak56 mRNA (Fig. 2B). They also expressed integrins, and formation of integrin clusters was observed in cells with several focal adhesion proteins, including p21-activated kinase and ␣-actinin (30). Immunofluorescent staining of the BG2-c6 cells with anti-Dfak56 antibodies produced a focal adhesion-like pattern (Fig. 4A). The nucleus was also stained with these antibodies. In some populations, the edges of the cells were also stained. Staining with anti-phosphotyrosine antibody produced a pattern that overlapped with that of anti-Dfak56 antibodies, except for the nucleus (Fig. 4B). Staining with anti-phosphotyrosine and anti-␤ PS integrin antibodies (Fig. 4C) showed colocalization of tyrosine-phosphorylated proteins with integrin. These data suggested that Dfak56 is a component of focal adhesion in Drosophila.
Kinase Activity of Dfak56 -We tested for protein kinase activity and phosphorylation of Dfak56 with mutational analysis. By site-directed mutagenesis, an amino acid substitution at residue 430 (tyrosine to phenylalanine) was introduced (Y430F mutant). This residue corresponds to the major autophosphorylation site in vertebrate FAK (tyrosine 397). We also substituted methionine for lysine at residue 513, a possible ATP-binding site essential for tyrosine kinase activity (K513M mutant). The cDNAs for the wild-type Dfak56 protein and the mutant proteins constructed in expression plasmids were then introduced in simian COS cells. Cell lysates were immunoprecipitated with anti-Dfak56 antibodies, and the precipitates were subjected to in vitro kinase assay.
Autophosphorylation of the Dfak56 protein was observed in the wild-type protein and the Y430F mutant, but not in the K513M mutant (Fig. 5A). The Dfak56 kinase activity observed was not attributable to co-immunoprecipitated kinases because the lysine at residue 513 of Dfak56 was essential for the kinase activity. Phosphoamino acid analysis of in vitro phosphorylated Dfak56 detected phosphorylated tyrosine only (Fig. 5B), confirming that Dfak56 is a protein-tyrosine kinase. Furthermore, the incorporated radioactivity of the phosphorylated Y430F mutant was approximately half that of wild-type Dfak56 (Fig.  5A), suggesting that the tyrosine at residue 430 is a major autophosphorylation site. This was confirmed by two-dimen-sional phosphopeptide mapping of the wild-type and Y430F mutant proteins (Fig. 5C). A major spot in the peptide map of wild-type Dfak56 (arrow) was not observed in that of the Y430F mutant.
Phosphorylation of Dfak56 upon Cell Attachment to the Extracellular Matrix-Drosophila S2 cells and stable transfectants (34) expressing Drosophila PS integrin chains of ␣ PS1 and ␤ PS (called PS1 cells) or ␣ PS2 and ␤ PS (called PS2 cells) were cultured on plastic dishes. Among them, only PS2 cells could attach to a plastic dish in serum-containing medium because of the cross-reactivity of ␣ PS2 ␤ PS integrin to vertebrate vitronectin and fibronectin (34). Using these cell lines, the levels of phosphotyrosine in the Dfak56 protein were examined following integrin-dependent cell attachment to the extracellular matrix. As shown in Fig. 6, the Dfak56 protein was tyrosinephosphorylated in PS2 cells that attached to the surface of the dish. This indicated that Dfak56 is tyrosine-phosphorylated upon integrin-dependent cell attachment to the extracellular matrix. Note that tyrosine-phosphorylated proteins of 50 -60 kDa were associated with the Dfak56 protein (Fig. 6). These phosphoproteins were also present in the anti-Dfak56 precipitates of PS1 cells or suspended PS2 cells, but not in precipitates of parental S2 cells. This suggests that the interaction between Dfak56 and these phosphoproteins requires the presence of integrins and is enhanced by cell attachment to the extracellular matrix. DISCUSSION We have identified a Drosophila homolog of vertebrate focal adhesion kinase and termed it Dfak56. Dfak56 shares both sequence and structural homology with the vertebrate FAK family of protein-tyrosine kinases (FAK and PYK2). We demonstrated that Dfak56 is tyrosine-phosphorylated following integrin-dependent cell attachment to the extracellular matrix. This property is the unique functional feature that distinguishes FAK from PYK2. Thus, we conclude that Dfak56 is a functional homolog of vertebrate FAK rather than PYK2. However, tyrosine phosphorylation of FAK and PYK2 occurs in response to a number of stimuli, and some of them overlap (stimuli for the G-protein-coupled receptor, e.g. bradykinin) (35)(36)(37), suggesting that the biochemical function of FAK and PYK2 is partly interchangeable. Furthermore, the similarity of the kinase domains between FAK and PYK2 (50 -60% amino acid identity) is almost the same as that between Dfak56 and FAK or PYK2. Therefore, we strongly suspect that Drosophila has only one FAK-related gene and that the ancestor gene duplicated and diverged in vertebrates during evolution.
In chicken embryonic fibroblasts, the FAK gene is alternatively transcribed to generate another short mRNA that encodes a C-terminal portion of the FAK protein named FAKrelated non-kinase (33). Ectopic expression of FAK-related nonkinase in fibroblasts blocks the formation of focal adhesions on fibronectin, acting as an inhibitor of FAK (38). We observed a relatively short, 1.5-kb transcript corresponding to the 3Ј-portion of the Dfak56 mRNA only in embryo and larva (Fig. 2B). This mRNA might encode a FAK-related non-kinase-like protein. If so, Dfak56-mediated cell adhesion processes may be regulated by the expression of this FAK-related non-kinaselike protein in embryogenesis.
Dfak56 expression is mainly observed in the embryonic central nervous system and the muscle attachment process. In muscle attachment, PS integrins and tiggrin, the ligand for ␣ PS2 ␤ PS integrin, are expressed. Mutants defective for these proteins show improper formation of muscle attachment, resulting in embryonic lethality (39,40). In vertebrate neurons, FAK is highly expressed and localized to a growth cone (41,42), suggesting that FAK plays a role in neural development and axon guidance. Recently, the role of integrins in axon guidance in Drosophila was determined (43). Null mutations in either the ␣ PS1 or ␣ PS2 subunit gene caused widespread axon pathfinding errors that could be rescued by supplying the wild-type integrin subunit to the mutant nervous system. Another novel ␣ integrin subunit gene, Volado, was identified to mediate olfactory learning of Drosophila (44). These data suggest that the integrin-mediated signal transduction pathway is important for the function of the central nervous system. Given the similar expression patterns and the close functional relation between integrins and FAK in vertebrates, Dfak56 may also participate in neural functions of Drosophila.
One of the most effective strategies for the study of signal transduction pathways is to identify novel upstream or downstream factors of the known players in the signaling cascade. The genetics of Drosophila are suited for the large-scale screen- were precipitated with anti-Dfak56 antibodies, and washed immunoprecipitates were subjected to in vitro kinase assay. To assess the level of expression, 0.02 volumes of the lysates used for immunoprecipitation were subjected to Western blotting and probed with the same antibodies. Standard molecular masses are indicated in kilodaltons. B, shown are the results from the phosphoamino acid analysis of the phosphorylated Dfak56 protein. The autophosphorylated wild-type Dfak56 protein was extracted from the gel and hydrolyzed. The resulting phosphoamino acids were subjected to two-dimensional electrophoresis at pH 1.9/3.5 and autoradiographed. The positions of phosphoserine (pS), phosphothreonine (pT), and phosphotyrosine (pY) are indicated. C, shown are the results from phosphopeptide mapping of the Dfak56 proteins. The autophosphorylated wild-type and Y430F mutant Dfak56 proteins were extracted from the gel, digested with trypsin, and subjected to two-dimensional phosphopeptide mapping. Asterisks show the origin of electrophoresis.
FIG. 6. Tyrosine phosphorylation of the Dfak56 protein upon integrin-dependent cell attachment to the extracellular matrix. S2 cells (10 6 ) and derivatives of their transfectants were seeded on 6-cm plastic dishes and cultured overnight. The next day, the suspended cells were shaken off and corrected. The cells attached to the dish (only for PS2 cells) were washed and lysed on the dish. The cell lysates (100 g of protein) were immunoprecipitated (IP) with anti-Dfak56 antibodies, subjected to Western blotting, and probed with anti-phosphotyrosine antibody (pY). To assess the amount of Dfak56 protein in the cell lysates, 10 g of proteins were blotted and probed with anti-Dfak56 antibodies. Standard molecular masses are indicated in kilodaltons.
ing to isolate the suppressor or enhancer if a visible mutant of the interested gene is available. Using this system, a number of the novel molecules implicated in the receptor protein-tyrosine kinase-mediated signal transduction pathways have been identified (45)(46)(47). In the locus of the Dfak56 gene, polytene band 56A to 56B, we could not find mutant alleles for the Dfak56 gene in the FlyBase data base. However, if we could isolate the mutants of Dfak56, the genetics of Drosophila would allow analysis of the cell adhesion signaling, and the findings would relate back to studies of higher organisms.
In conclusion, we cloned Dfak56, a Drosophila homolog of the FAK gene, and characterized its product by determining its expression, kinase activity, and signaling through integrinmediated cell attachment. Future genetic studies would facilitate the analysis of the roles of FAK family proteins in vivo.