Expression and characterization of splice variants of PYK2 ... - CiteSeerX

2 downloads 0 Views 397KB Size Report
kinase-related protein. Wen-Cheng Xiong, Marlene Macklem and J. Thomas Parsons* ..... P. 738 781. CAG. GAG. FAT. Kinase. Fig. 1. (A) The schematic structures of PYK2 ...... Zachary, I., Sinnett-Smith, J. and Rozengurt, E. (1992). Bombesin,.
1981

Journal of Cell Science 111, 1981-1991 (1998) Printed in Great Britain © The Company of Biologists Limited 1998 JCS4551

Expression and characterization of splice variants of PYK2, a focal adhesion kinase-related protein Wen-Cheng Xiong, Marlene Macklem and J. Thomas Parsons* Department of Microbiology and Cancer Center, University of Virginia Health Sciences Center, Charlottesville, Virginia 22908, USA *Author for correspondence (e-mail: [email protected])

Accepted 4 May; published on WWW 30 June 1998

SUMMARY Focal adhesion kinase and the recently identified prolinerich tyrosine kinase 2 (PYK2), also known as cell adhesion kinase β, related adhesion focal tyrosine kinase or calciumdependent protein tyrosine kinase, define a new family of non-receptor protein tyrosine kinases. Activation of PYK2 has been implicated in multiple signaling events, including modulation of ion channels, T- and B-cell receptor signaling and cell death. Mechanisms underlying the functional diversity of PYK2 are unclear. Here, we provide evidence for two novel alternatively expressed isoforms of PYK2. One isoform, designated PYK2s (PYK2 splice form), appears to be a splice variant of PYK2 lacking 42 amino acids within the C-terminal domain. A second isoform, referred to as PRNK (PYK2-related non-kinase), appears to be specified by mRNAs that encode only part of the Cterminal domain of PYK2. Northern blot analysis indicates that the unspliced PYK2 is expressed at high levels in the brain and poorly expressed in the spleen, whereas PYK2s

and PRNK are expressed in the spleen. In situ hybridization studies of rat brain demonstrate that the unspliced PYK2 is selectively expressed at high levels in hippocampus, cerebral cortex and olfactory bulb, whereas PYK2s and PRNK are expressed at low levels in all regions of rat brain examined. Immunofluorescence analysis of ectopically expressed PRNK protein shows that PRNK, in contrast to full-length PYK2, is localized to focal adhesions by sequences within the focal adhesion targeting domain. In addition, PYK2, but not PRNK, interacts with p130cas and Graf. These results imply that PRNK may selectively regulate PYK2 function in certain cells by binding to some but not all PYK2 binding partners, and the functional diversity mediated by PYK2 may be due in part to complex alternative splicing.

INTRODUCTION

concentration, activation of protein kinase C and stimulation with angiotensin II (Lev et al., 1995; Yu et al., 1996). Activation of both FAK and PYK2 has been observed in response to multiple signalings. For example, activation of FAK by clustering of integrins through binding to the extracellular matrix has been implicated in cell spreading, migration and survival (Burridge and Chrzanowska-Wodnicka, 1996; Parsons, 1996; Frisch et al., 1996; Xu et al., 1996; Hungerford et al., 1996). Activation of PYK2 has been suggested to play a role in regulating neurotransmission or neuroplasticity by phosphorylating potassium channels (Lev et al., 1995), and in T- and B-cell signaling (Berg and Ostergaard, 1997; Qian et al., 1997; Manie et al., 1997). In addition, we recently showed that over-expression of PYK2, but not FAK, in rat and mouse fibroblasts leads to apoptotic cell death (Xiong and Parsons, 1997). These results suggest that while PYK2 and FAK are structurally similar, each has the capacity to mediate distinctly different signaling responses (Xiong and Parsons, 1997; Schaller and Sasaki, 1997). FAK interacts with multiple proteins. Phosphorylation of FAK on tyrosine 397, the major site of autophosphorylation,

Focal adhesion kinase (FAK) and proline-rich tyrosine kinase 2 (PYK2), also known as cell adhesion kinase β (CAKβ), related adhesion focal tyrosine kinase (RAFTK) or calciumdependent protein tyrosine kinase (CADTK), define a new family of non-receptor protein tyrosine kinases (Schaller et al., 1992; Lev et al., 1995; Sasaki et al., 1995; Avraham et al., 1995; Yu et al., 1996). FAK and PYK2 contain a central catalytic domain and large N- and C-terminal non-catalytic regions that are devoid of SH2 and SH3 domains (Schaller et al., 1992; Lev et al., 1995; Sasaki et al., 1995; Avraham et al., 1995). Both FAK and PYK2 are activated by many diverse cellular stimuli. FAK is activated by v-Src transformation (Schaller et al., 1992), attachment to the extracellular matrix (Guan and Shalloway, 1992; Schaller et al., 1992), and exposure to growth factors (e.g. PDGF; Rankin and Rozengurt, 1994), neuropeptides (e.g. bombesin; Zachary et al., 1992; Rozengurt, 1991) and lysophosphatidic acid (Moolenaar, 1991). PYK2 has been found to be differentially activated by other stimuli including elevation of the intracellular calcium

Key words: Proline-rich tyrosine kinase 2 (PYK2), Focal adhesion kinase, FAK-related non-kinase, Splicing

1982 W.-C. Xiong and others creates a high-affinity binding site for SH2 domains of Src and Fyn, both of which are associated with activated FAK (Schaller et al., 1994; Cobb et al., 1994). Phosphorylation of FAK at other tyrosine residues leads to the association with other SH2domain containing signaling proteins, including Grb2 and p85 subunits of PI3 kinase (Schlaepfer et al., 1994; Chen and Guan, 1994; Guinebault et al., 1995). Proline-rich regions in the Cterminal domain of FAK direct the binding to p130cas (Crkassociated substrates) and Graf (GTPase regulator associated with FAK) in an SH3 domain-dependent manner (Polte and Hanks, 1995; Harte et al., 1996; Hildebrand et al., 1996). The C-terminal domain of FAK can also associate with the cytoskeletal proteins paxillin and talin (Bellis et al., 1995; Chen et al., 1995; Hildebrand et al., 1995; Tachibana et al., 1995). PYK2 is highly homologous to FAK, sharing 45% overall sequence identity and 60% identity in the catalytic domain. Several tyrosine residues are conserved between FAK and PYK2, including the binding site for the SH2 domains of Src and Fyn (Y397 in FAK, Y402 in PYK2), and the putative binding site for the SH2 domain of Grb2 (Y925 in FAK, Y881 in PYK2) (Schaller et al., 1994; Cobb et al., 1994; Schlaepfer et al., 1994; Lev et al., 1995; Sasaki et al., 1995; Avraham et al., 1995; Dikic et al., 1996). In addition, PYK2 also contains the prolinerich sequences responsible for mediating the binding of p130cas and Graf (Lev et al., 1995; Sasaki et al., 1995; Avraham et al., 1995). The paxillin binding region in the C-terminal domain of FAK is also highly conserved within PYK2 (Lev et al., 1995; Sasaki et al., 1995; Avraham et al., 1995). Given the high degree of sequence similarity between PYK2 and FAK, it is possible that PYK2 interacts with many of the FAK binding partners. Indeed, tyrosine-phosphorylated PYK2 can bind to the SH2 domains of Src and Grb2 (Lev et al., 1995). PYK2 has been reported to interact with p130cas and paxillin in a manner similar to that of FAK (Astier et al., 1997; Manie et al., 1997; Hiregowdara et al., 1997; Li and Earp, 1997). Several splice forms of FAK have been identified, and they appear to be differentially expressed (Schaller et al., 1993; Burgaya and Girault, 1996; Derkinderen et al., 1996). FAK is enriched in the brain and expressed in most cell lines and tissues examined (Grant et al., 1995; Burgaya and Girault, 1996; Derkinderen et al., 1996). In human and rat brain, several FAK alternative splice forms that appear to be able to regulate FAK autophosphorylation have been observed (Andre and BeckerAndre, 1993; Burgaya and Girault, 1996; Burgaya et al., 1997). In chicken embryo fibroblasts, the C-terminal domain of FAK is expressed autonomously as a 41-43 kDa protein, designated as FRNK (FAK-related non-kinase) (Schaller et al., 1993). Both FAK and FRNK localize to focal adhesions when cells (e.g. fibroblasts) spread on a substratum and this localization is dependent upon the structural integrity of the focal adhesion targeting ‘FAT’ domain, a 140-amino-acid sequence in the carboxyl termini of FAK and FRNK. The FAT domain is both necessary and sufficient for the targeting of FAK and FRNK to focal adhesions (Hildebrand et al., 1993). Chicken embryo fibroblasts over-expressing FAK appear to be normal. However, over-expression of FRNK in these cells inhibits the tyrosine phosphorylation of FAK and several other focal adhesion proteins, e.g. paxillin and tensin, and leads to defects in cell spreading (Richardson and Parsons, 1996; Richardson et al., 1997). Based on these data, it has been suggested that FRNK may function as a negative regulator of FAK activity in certain cells.

Here, we report the isolation of two novel PYK2-related cDNAs, one of which, designated as PYK2s (PYK2 splice form), encodes a splice variant of PYK2 lacking 42 amino acids within the C-terminal domain. A second isoform, referred to as PRNK (PYK2-related non-kinase), is derived from mRNAs that encode only part of the C-terminal domain of PYK2. Northern blot analysis indicates that whereas the mRNA of PYK2 (unspliced form) is expressed abundantly in the brain, less abundantly in kidney and lung, and poorly in spleen and other tissues (e.g. heart, liver, skeletal muscle and testis), the mRNA of PYK2s and PRNK are expressed in many tissues, including the spleen. In situ hybridization studies of rat brain demonstrate that PYK2 is selectively expressed at high levels in hippocampus, cerebral cortex and olfactory bulb, whereas PYK2s and PRNK are expressed at low levels in virtually all regions of rat brain examined. Biochemical characterizations indicate that whereas PYK2 interacts efficiently with several proteins that bind to FAK (e.g. p130cas and Graf), PRNK fails to bind to these proteins efficiently. In addition, ectopic expression of PRNK in Swiss 3T3 cells results in its localization to focal adhesions by virtue of sequences within a putative ‘FAT’ domain. These results imply that PRNK may selectively regulate PYK2 function in certain cells by binding to some but not all PYK2 binding partners, and the functional diversity mediated by PYK2 may be due in part to complex alternative splicing. MATERIALS AND METHODS Reagents Rabbit polyclonal antiserum directed to PYK2 was raised by immunization of rabbits with a GST fusion protein containing the Cterminal 400 amino acids of rat PYK2 (amino acids 587-988) as described previously (Xiong and Parsons, 1997). Monoclonal antibodies to c-myc, PYK2, paxillin and p130cas were purchased from Santa Crutz Biotechnology (c-myc) or Transduction Labs (PYK2, paxillin and p130cas). The cDNA constructs encoding different GSTPYK2 fusion proteins were generated by PCR and subcloned into pGEX-2TK vector. The cDNA constructs encoding different GSTSH3 domain fusion proteins were provided by A. Bouton (p130casSH3), Z.-H. Leu (Src-SH3, Fyn-SH3), J. Hildebrand (Graf-SH3), or generated by PCR using rat brain first strand cDNA as template (PSD95-SH3). HEK 293 cells were provided by Q.-H. Song (Columbia University, New York). PCR, isolation of cDNA clones and DNA sequence analysis Degenerate PCR primers were synthesized based on amino acid sequences conserved between FAK (amino acids P585IKWN and A1028VDAK) and PYK2 (amino acids P588IKWN and A984VDAK). RNA from rat hippocampus was purified and used as a template for the synthesis of ‘first’ strand cDNA, which was then used to amplify FAK-related sequences via PCR, using the degenerate primers indicated above. DNA fragments of the expected size (about 1.2 kb) were isolated and subcloned into Bluescript vector for sequence analysis. PRNK cDNA clones were isolated from a rat hippocampal λ-ZAP cDNA library probed with the 32P-labeled PCR fragments, generated as described above. The cDNAs were sequenced by the dideoxy chain termination method. Expression vectors The cDNAs of PYK2, PYK2s and PRNK (isolated from rat hippocampal cDNA library) were subcloned into mammalian

Characterization of splice variants of PYK2 1983 expression vectors, either downstream of a c-myc epitope tag (MEQKLISEEDL) under the control of the cytomegalovirus (CMV) promoter (pCMV-c-Myc) (Evan et al., 1985) or downstream of GST under the control of the elongation factor 1α (EF1α) promoter (pEBG). The deletion mutants in PYK2 and PRNK were generated by PCR amplification (Ho et al., 1989). The authenticities of all mutants were verified by DNA sequencing. Analysis of RNA expression For northern blot analysis, total RNA was prepared as described (Xiong and Montell, 1993), transferred to nitrocellulose and hybridized with fragments of the 32P-labeled cDNAs or oligonucleotides end-labeled with [32P]dATP. To study the mRNA distribution in different tissues, an RNA blot comprising poly(A)selected RNAs from different rat tissues (Clontech) was hybridized with specific 32P-labeled cDNAs or oligonucleotides end-labeled with [32P]dATP. For in situ hybridization, serial frozen adult rat brain sections, 10 µm thick and cut at sagittal, coronal and horizontal planes, were mounted on APS-coated glass slides and fixed with 4% paraformaldehyde. The sections were rehydrated with a graded series of ethanol, dried by cold air, and subjected to protease treatment and acetylation. After acetylation, the tissue slides were incubated with the hybridization mixture containing 106 cpm/ml 33P-labeled probes (10 pmoles) at 50°C for 24 hours. The tissue sections were washed 5 times in 1× SSC for 30 minutes at 50°C, dried by air and exposed to Hyperfilm βmax (Amersham, IL). The exposure times were as follows: 72 hours (3 days) for probe A and probe C (sagittal sections), and 10 days for probes B and probe C (coronal and horizontal sections). The probes (A, B and C) used in the in situ hybridization were 33-mer oligonucleotides complementary to different regions in PYK2/CAKβ, as indicated in Fig. 5A. The oligonucleotides were endlabeled with [33P]dATP as described previously (Talley et al., 1997). Cell culture, transfections and microinjections Swiss 3T3 cells were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10% fetal calf serum, 100 µg/ml penicillin G and 100 µg/ml Streptomycin (Gibco). HEK 293 cells were maintained in DMEM/F12 (1:1) medium with the same additives. For transfection, cells were plated for 12 hours prior to transfection at a density of 106 cells per 100 ml culture dish, and transfected using the calcium phosphate precipitation method. At 48 hours post-transfection, cells were lysed, and the extracts were subjected to analysis by SDS-PAGE and western blotting. For microinjection, cells were plated for 12 hours at a density of 104 cells per 22 mm coverslip. Purified plasmid DNA encoding PRNK or PRNK mutants (50 ng/µl) were microinjected into the nucleus. 2-4 hours later, the microinjected cells were fixed with 4% paraformaldehyde and immunostained with specific antibodies. Immunostaining was carried out as previously described (Xiong and Parsons, 1997). Briefly, fixed and permealized cells were incubated with primary antibody against PYK2 (1:500 dilution) at 37°C for 1 hour and subsequently incubated with a fluorescenceconjugated anti-rabbit or anti-mouse secondary antibody (1:300 dilution) for 1 hour at 37°C. The stained cells were visualized using a Leitz fluorescence microscope. GST ‘pull-down’ assays and immunoprecipitation For in vitro ‘pull-down’ assays, rat brain extracts (1 mg) prepared by homogenizing rat brain in modified RIPA buffer (Schaller et al., 1992) were precleared with glutathione-S-transferase (GST) bound to glutathione-Sepharose 4B (Pharmacia, Piscataway, NJ). After preclearing, these extracts were incubated with 2 µg of various GST fusion proteins coupled to glutathione-Sepharose beads at 4°C for 1 hour with constant rocking. The beads were washed three times with the modified RIPA buffer, and assayed for bound proteins by immunoblotting with different antibodies. For in vivo ‘pull-down’ assays, GST-tagged PRNK and deletion

mutants (GST-PRNKD1-95, GST-PRNKD1-131) were transfected into HEK 293 cells and the cell lysates prepared as above. The GSTtagged PRNK and deletion mutant proteins were captured by incubation of the cell lysates with glutathione-Sepharose beads at 4°C for 1 hour. The GST complexes were recovered by centrifugation, washed and assayed for the bound proteins (e.g. paxillin) as described above. Immunoprecipitation of individual proteins was carried out as previously described (Xiong and Parsons, 1997). Briefly, 1 mg of cell lysate was incubated with anti-PYK2 antibodies (1-10 mg) at 4°C for 1 hour in a final volume of 1 ml RIPA buffer. After the addition of protein A-agarose beads, the reaction was incubated at 4°C for another hour. Immunocomplexes were collected by centrifugation, washed and the bound proteins, resolved by SDS-PAGE, subjected to immunoblotting using anti-PYK2 antibodies.

RESULTS Identification of variant isoforms of PYK2 To identify FAK family tyrosine kinases in rat brain, we employed a PCR approach using a rat hippocampal cDNA template and degenerate oligonucleotide primers based on conserved amino acids between FAK and PYK2 (see Materials and methods). PCR amplification yielded DNA fragments encoding portions of both FAK and PYK2. DNA sequence analysis of a cloned PCR fragment revealed a novel splice variant of PYK2, containing a 42-amino-acid deletion of residues 738-780 within the C-terminal domain of PYK2 (Fig. 1A). A rat brain cDNA library was subsequently screened using the mixed PCR fragments as probes. In addition to clones that were identical to FAK or PYK2, two novel cDNA clones were isolated. One cDNA contained an open reading frame encoding a portion of the PYK2 C-terminal domain fused with a unique 5′ untranslated sequence. Since this putative PYK2related protein lacked both N-terminal and catalytic domains, it was called ‘PRNK’ (PYK2-related non-kinase) (Fig. 1A). A second cDNA clone contained an open reading frame encoding a PYK2-related protein with a 42-amino-acid deletion (e.g. deletion of residues 738-780 of wild-type PYK2), which was identical to the sequence generated from the cloned PCR product. This form was referred to as PYK2s (PYK2 splice form) (Fig. 1A). Interestingly, the C-terminal 228 residues of PRNK and PYK2s were identical (corresponding to residues 780-1009 in PYK2) (Fig. 1A), suggesting that both cDNAs may be derived by splicing from different upstream exons to a common 3′ exon. In the case of PRNK, the 5′ exon encoded an unique sequence containing multiple stop codons, indicating that it is an untranslated leader sequence. The nucleotide and amino acid sequences of PRNK are shown in Fig. 1B. The PRNK cDNA (3.6 kb) contains a 2114-nucleotide 5′ untranslated region and a single long open reading frame encoding a 238-amino-acid protein with a predicted molecular mass of 29 kDa. Comparison of the deduced amino acid sequence of PRNK with that of PYK2 reveals the presence of two PxxPxR/K SH3 domain binding motifs and a putative FAT domain. Expression of PYK2, PYK2s and PRNK To determine whether PRNK is expressed in vivo, RT-PCR analysis was carried out using a pair of PRNK-specific primers. PCR amplification using a 5′ primer unique to PRNK

1984 W.-C. Xiong and others (nucleotides 1770-1790 within the 5′ untranslated region), and the 3′ primer (20 bp) complementary to sequences common to both PYK2 and PRNK (nucleotides 2474-2494), yielded a fragment of the expected size (0.7 kb) when rat brain cDNA was used as a template. No detectable PCR product was observed when PYK2 cDNA was used as a template (Fig. 2A). These results indicated the presence of a PRNK-like RNA in rat brain. We then characterized the mRNA expression of

A 5'

1

PYK2, PYK2s and PRNK using several sequence-specific probes (Fig. 1). Probe A is a 33-bp oligonucleotide probe specific for the exon present in PYK2 and deleted in PRNK and PYK2s (Fig. 1A). Probe B is a 34-bp oligonucleotide specific for PYK2s encompassing a contiguous sequence specific to PYK2s (Fig. 1A). Probe C is a 0.5 kb fragment specific for the 5′ untranslated region of PRNK (Fig. 1A). Probe D is a 0.7-kb DNA fragment that is common to PYK2, PYK2s and PRNK

AAA AAA AAA AAA AAA AAA AAA AAA 418

680 738 781

Kinase

PYK2 (CAKβ)

P

868

PP

CAG

1009

FAT

3' 3787 bp

GAG

A

738 781

Kinase

PYK2s 5' PRNK

B

Fig. 1. (A) The schematic structures of PYK2, PYK2s and PRNK. The boxes denote the open reading frame, and the catalytic domain of the PYK2-encoded protein tyrosine kinase (kinase) and focal adhesion targeting domain (FAT) are indicated. The positions of the putative initiation and termination codons of the PYK2 and PRNK cDNAs and the splice junctions of PYK2s and PRNK are shown. The positions of four oligonucleotides or DNA fragments used as probes (AD) for northern blot analysis and in situ hybridization analysis are also indicated. (B) Nucleotide sequence of the PRNK cDNA and deduced amino acid sequence of the putative PRNK protein (GenBank accession number AF063890). The predicted amino acid sequence is shown under the nucleotide sequence. The sequence which is different from PYK2 is underlined. The two PxxPxR/K motifs for the SH3 domain binding are boxed.

P

CAG

B

1

PP

B

9(781)

PP

C

CAG

FAT

GAG

GAG

238(1009)

FAT

3'

3586 bp

D

TGCGGCCGCGGCATACATTGCTTGTAACTTGTGGCAATCTCCTAGGACCTACAAGAATATATAGTAGATTTTGGAGTTCAGTAACTCATCCT AACCGTATTTAATATTCCTCACTCTCTATGTTTTCTTTTATTTCTTTTTTTTTTCCGGAGCTGGGGACTGAACCCAGGGCCTTGCGCTTG CTAGGCAAGCACTCTACCACTGAGCTAAATCCCCAACCATATGTTTTCTTTTATATTTATGACCCTGTATTTAAAATTTTTGACACTTAG TTCTAACATTTATTAAGAGCCCAAGGTGACCATAGACCGCCATTGAGTTTAACTGTTTTTTTTTTTAACTTTAAAATGCTTACCTTCATT TTTGGAGCATTTGTGTCTTAGTCTGTTTTTCCCTTCTATAACAAAATGCCTGAGGCTAGCATTTCACACAAAGAAGAAGTATAGTTCAGT TTTGAGTGCTAAGAACATGGACAGGCAGCCACTGGGGTCCACTGAAGACCTTCTGGCCACATTATACTATAGTAGATAGCATTGTGTTAG GGTGCATAAATAACTCTATTAGTCACTTGGCACTCTGAAGACAGAAACACGGTGGTTTTTAAAACTGCCATTCTCCCAGCATCCAGCTCT ATGCCTACCACACAGTATGCTTAATCAATAGTTACTGAATGGGTTGGATAAAGAGTACTATAAGAAATAAACACAGAGAGTTGAATCTAT ATTAGTGATGGGCAATGTAACCCAATACAACTATTTTAAAGCTATATTGTAATGATCAAGCCTGTTTTACAGGGCTTAGTGGGGAACTCA TCAGAGGCCTGGCGGAGACATGACTAGTTAAATGTAGATATAATATTTTTTATTTAGCATTTTCTTTAAGTGACCTATAAGCAGGCACAT CAAAAGAAAGGGACCGTGTCACTTGCCTTGGGCTTCTGTAACAGGATACCTGAGACTAGGTGCTTTTTACTGGACAGAAGTTTGTTGCTT CAGTTTGGGGGACCGGGAGTCCATGATTGAAGGCCACTTCTTTTTTTTTTTTTTTGAAGGCCACTTCTAACAAGGGCTTTCCTGCTGTTT TTTGGCACGGCAGAAGAAACAGTGGTTGAGAGGCAAAACAGATGCGAACAGAAAGGCAAATTTTCCCTTTTATCAGGGCCCTTATTTTTT AACTTGTTAACAAGGTCAGCGCCTCCATGATTTGCAAACCACCCCAACTCCCAATGCTGTTGTTCTGGAATTCAAATTCCTGACACATGA GCCTTACAGTGGGGCCACAATCAAGCCTTAACAGGACTTCACAAAAGCACCCTATCGCTAGCAAAGGGTTCCCCAATCAAGGAAATACAT GCAACCTTATTTGTCCATTAGATCCAAGAGGCCAAGGCGATCAATCATAGTAGACCACAGCCAAAGAGTTCAGAGGGACTGAGAAAGTAA TTTTGCAAATGCACCCTGGTAAGACTGAGGTAAATAGGGAAACACCAACACCTACTTAAGCAACAAAACTAAGCACAAGTCCAAAGCCTC ACTGGTATGAGAAGCCCAGGCTTAGGGCTGTGAAGGAGGAAGGAGCATCCATCTTGTGTCTATAGGGTTATGCTTCTGGGACTTTGATTA TTCATTTTCTCTTCTTAGCTCATTCATTTCTTTTAACCTTCCGCTGTGTCATTCAAGCTGTCATCCTAACCAAGCTTGATCGATGGGCAC AAACGATCCCGTACACCTGGAGCTCAGGGTAGCCCCACTGTCTGTCGACGTCTCCAGACCTGTCACCATAGTGCTCAGATGAAACAAGTT CCGAGGGTTTGGAGCCAGCTCTGTATTTCTAAGCCTAAGAAGTGTCAGCAGCCATCACAAATGCCAGCCAAGAAAAATCGGTGGGTGGTT GAAAATTGGATGTCATTGGCAGACTGTGTGACAGTTTGAACCCAGTGCTCTGGGGTTTGGCATCTGGTTACAGCCAGGCCTGCGGATGAG GCTGAGACCAGAGTAAAACTGTCTGCATTTGTGCCTCTGGCTTTAGGAACAGCCAAGAAAAGAGCAGTGGGCAATGCCAAGGTTTGGCTC CCTGAGGCTCTTCAGCAGAACCTCTCTGATCCAGGCCTGAGGATGGGGCTCATTGTGTTGTCCTCACAGGAGGAGGACTTCATCCGGCCC M G L I V L S S Q E E D F I R P

92 182 272 362 452 542 632 722 812 902 992 1082 1172 1262 1352 1442 1532 1622 1712 1802 1892 1982 2072 2162 16

AGTAGTCGAGAAGAGGCCCAGCAGCTCTGGGAGGCAGAGAAGATCAAGATGAGGCAGTTCCTAGACAGACAGCAGAAGCAAATGGTGGAA S S R E E A Q Q L W E A E K I K M R Q F L D R Q Q K Q M V E

2252 46

GATTCCCAGTGGCTGAGGCGGGAGGAAAGATGCTTGGACCCGATGGTTTATATGAATGACAAGTCCCCACTGACCCCAGAGAAGGAGGCC D S Q W L R R E E R C L D P M V Y M N D K S P L T P E K E A

2342 76

GGCTACACGGAGTTCACAGGGCCCCCCCAGAAGCCACCACGGCTCGGTGCACAGTCCATCCAGCCTACAGCCAACTTGGACAGGACTGAT G Y T E F T G P P Q K P P R L G A Q S I Q P T A N L D R T D

2432 106

GACCTCGTGTACCACAATGTCATGACCCTGGTGGAGGCCGTGCTGGAACTCAAGAACAAGCTCAGCCAGCTGCCTCCTGAGGAGTATGTG D L V Y H N V M T L V E A V L E L K N K L S Q L P P E E Y V

2522 136

GTGGTGGTGAAGAATGTGGGGCTGAACCTGCGGAAGCTCATCGGTAGCGTGGACGATCTCCTACCCTCCTTGCCAGCATCTTCTAGAACA V V V K N V G L N L R K L I G S V D D L L P S L P A S S R T

2612 166

GAGATTGAAGGGACCCAGAAACTGCTCAACAAAGACCTGGCAGAGCTCATCAACAAGATGAGGTTGGCTCAGCAGAACGCTGTGACATCC E I E G T Q K L L N K D L A E L I N K M R L A Q Q N A V T S

2702 196

CTAAGTGAGGACTGCAAGCGGCAGATGCTCACAGCATCCCATACCCTGGCTGTGGATGCCAAGAACCTCCTGGATGCTGTAGACCAAGCC L S E D C K R Q M L T A S H T L A V D A K N L L D A V D Q A

2792 226

AAGGTTGTGGCTAATCTGGCCCACCCGCCTGCAGAGTGATCAAGGGAGGGACCACCTGCCTGCATCTTCTGCCCCAACCTTTCCTGCCTT K V V A N L A H P P A E * GCCTTTGGTTATTGGTCTTCTAGGGAAAGCTGAGAAGAGTCCTCCCCTTACCACTTTGCACGACCCCCCTTTCCCCAACCCACCCCAGAC TGTGCTACTTGGGCTACATCTGGACAGAAAGGACTCTGGGCACAGACACGGGGGTGGGGTGACATAGTTCATAGGGGAACTTAGGGCCAG CCATTCCCTCGTACCCCAGGGTGGGTTGGCTGGAGCATGATTGGGGTAAATAACTGTACCCCTACCAGCCAAAGATGGCTTTATGCATGG ACATTTGAGAGCCAGTATTTCTCTGTCCTCTTTAGCCCTCAGGGACCCCTGATACAGAGGGGACAGAGAGGGGTTTTATTTGTAGAGAAG CTGGTGAGATGAGAGCTGGGCATGGCTCTCTTGTACAGTGTACATTGGAATTTATTTAATGTGAGTTTGGCCTGGACAGACAGCCATGGG CCACAGTCCAGGAACAAGCTAATCCAGCCACAGGGAAGAAGCAGAGTCAGGGGTCAGAATGGGACTTCATGTCCTCCCTGCGTTTCTCTT CTCCCTCTTTCCTCCTCCCCTCTTTTCTTAGTCCTCCTTTTTCTCTTCCCCCTTTTCACATCTGTTCCTTTCCTCTCTCACATTTGTGGA GAACATCCATTTACCTTTTCTCTTTTTGATCTGTGGTTGAATTAAAATCATTAGCATTCGCAAAAAAAAAAAAA

2882 238 2972 3062 3152 3242 3332 3422 3512 3586

Characterization of splice variants of PYK2 1985

Fig. 2. Expression of PRNK and PYK2s. (A) PCR products of PRNK were amplified from the templates of first strand cDNA synthesized from rat hippocampal RNA (lane 1), the cDNA of PYK2 (lane 2), and the cDNA of PRNK isolated from the rat brain cDNA library (lane 3) using a PRNK-specific primer and another antisense primer in the coding region (see Materials and methods). The DNA markers are shown on the left. (B) Northern blots of total RNAs from adult rat brain and spleen hybridized with probe A (specific for the unspliced PYK2), B (PYK2s), C (PRNK) and D (the common probe). Sequences contained in the individual probes are indicated in Fig. 1A.

(Fig. 1A). Northern blot analysis was carried out using total RNAs from rat brain and spleen. Each of the above probes hybridized to a mRNA in the brain or spleen of about 4.3-4.4 kb, similar in size to that of PYK2 (4.4 kb) (Fig. 2B). Probe D, the common probe, detected a mRNA about 4.3-4.4 kb in both brain and spleen (Fig. 2B). Probe A, the specific probe for the unspliced PYK2, hybridized to an mRNA about 4.3-4.4 kb in the brain, but failed to detect RNA in the spleen, indicating the absence of PYK2 (unspliced form) and the expression of PYK2s or PRNK in the spleen (Fig. 2B). The expression of PYK2s and PRNK in the spleen was confirmed using the PYK2s-specific probe B and PRNK 5′ leader-specific probe C (Fig. 2B). The distribution of PRNK, PYK2s and PYK2 was further analyzed by northern blot analysis using mRNAs from various rat tissues including the heart, brain, spleen, lung, liver, skeletal muscle, kidney and testis (Fig. 3). Total PYK2 RNA, identified by probe D, the common probe, was present in the brain, spleen, kidney and lung. The unspliced PYK2 RNA, identified by probe A, was present in the brain, less abundant in the kidney and lung, and nearly undetectable in the spleen and other tissues (e.g. heart, liver, skeletal muscles and testis), again supporting the expression of PYK2s or PRNK in the spleen. On the other hand, the PRNK RNA identified by probe C was expressed in all tissues in which PYK2, identified by the common probe D, was expressed (Fig. 3). We further examined the expression of the unspliced PYK2 (PYK2u), PYK2s and PRNK at the mRNA level by in situ hybridization in adult rat brain, since these cDNAs were initially isolated from an adult rat hippocampal cDNA library and PYK2-related RNAs were most abundant in the brain (Fig. 3). In situ hybridization was carried out using a series of adult rat brain sections cut at sagittal, coronal, and horizontal planes. Three sets of anti-sense and sense oligonucleotides (A, B and

Fig. 3. Differential expression of PYK2 and PRNK in various tissues. Poly A containing RNAs from rat heart, brain, spleen, lung, liver, skeletal muscle, kidney and testis were analyzed by northern blot and hybridized with probe A (unspliced PYK2), C (PRNK) and D (the common probe) (see Figs 1A and 2). The relative size of the RNA species is indicated.

C) were used as probes. Probes A and B, specific for the unspliced PYK2 (PYK2u) and PYK2s RNAs, are identical probes used in northern blots as described above (Fig. 4A). Probe C is a 33-bp oligonucleotide specific for PRNK (comlementary to PRNK untranslated region) (Fig. 4A). The distribution of the mRNAs for PYK2u, PYK2s and PRNK in rat brain is shown in Fig. 5 and summarized in Fig. 4B. The PYK2u RNA, identified with probe A, was readily detected in hippocampus (e.g. dentate gyrus (DG), CA1, CA2 and CA3 neurons), cerebral cortex (Cx) and olfactory bulb (OB), less abundant in thalamus (Th), and nearly undetectable in cerebellum (Cb) and choroid plexus in the third ventricle (D3V) (Fig. 5A,A′,A′′). Interestingly, PYK2s and PRNK were readily detected in cerebellum where the PYK2u was poorly expressed (Fig. 5B,B′′,C,C′′ and A,A′′). In addition, the PYK2s and PRNK RNAs, identified with probe B and C, respectively, were expressed relatively ubiquitously, albeit at low levels, in virtually all regions in the brain examined (Fig. 5B,B′,B′′,C,C′,C′′). The corresponding sense oligonucleotides hybridized on the adjacent sections failed to yield detectable signals (data not shown). These data indicate that the differential expression of PYK2 and its isoforms may be important for the differential PYK2 activity in different parts of the brain. Binding of PRNK and PYK2 with different proteins Since PRNK shares sequence homology with FRNK, we examined whether PRNK interacted with proteins that normally bind FRNK, e.g. paxillin, p130cas and Graf (Harte et al., 1996; Hildebrand et al., 1996). In vitro ‘pull-down’ assays were carried out using immobilized GST-PRNK proteins and extracts from rat brain. GST-PRNK efficiently pulled down paxillin from rat brain extracts (Fig. 6A). The region responsible for paxillin binding was defined using several variant forms of PYK2 and PRNK (expressed as GST-fusion proteins). GST-PYK2-Cterm contains the C-terminal domain of PYK2 (amino acids 680-1009 in PYK2); GST-PRNK∆1-95 retains the putative ‘FAT’ domain and is missing the prolinerich motifs (containing amino acids 96-238 in PRNK or 8681009 in PYK2); and GST-PRNK∆1-131 is missing a portion

1986 W.-C. Xiong and others of the ‘FAT’ domain (containing amino acids 132-238 in PRNK or 902-1009 in PYK2). While GST-PYK2-Cterm and GST-PRNK∆1-95 bound paxillin with an efficiency similar to that of GST-FAK-Cterm, GST-PRNK∆1-131 failed to bind paxillin (Fig. 6A). These data indicate that PRNK associates with paxillin through a region within amino acids 96-238 in PRNK (or amino acids 868-1009 in PYK2). This region contains a sequence similar to the paxillin binding domain in FAK (Hildebrand et al., 1995). To determine whether PRNK interacts with paxillin in vivo, we employed an in vivo GST ‘pull-down’ assay. GST-tagged PRNK (GST-PRNK) and the deletion mutants GST-PRNK∆195 and GST-PRNK∆1-131 were transiently transfected into HEK 293 cells. The GST fusion proteins were collected on glutathione beads, the associated proteins resolved by SDSPAGE, and subjected to immunoblotting using antibodies against GST or paxillin (Fig. 6B). As shown in Fig. 6B, paxillin co-precipitated with GST-PRNK and GST-PRNK∆1-95, whereas GST-PRNK∆1-131 failed to associate with paxillin. These data confirmed the in vivo binding of PRNK with

A

A

PYK2(CAKβ)

C

KD

N

B B

PYK2s

C

PRNK

B PYK2u (A)

PYK2s (B)

PRNK (C)

Cerabral Cortex (Cx)

+++

++

++

Hippocampal Formation (DG, CA1, CA2, CA3)

+++

++

++

Thalamus (Th)

+++

++

++

Olfactory Bulb (OB)

+++

++

++

Piriform Area (Pir)

+++

++

++

Pons (Pn)

-

++

++

Hypothalamus (Ht)

-

++

++

Medulla (Me)

-

++

++

Cerabellum (Cb)

-

++

++

Choroid plexus in the third ventricle (D3V)

-

++

++

Fig. 4. (A) The schematic diagram for the positions of the probes (A-C) used in the in situ hybridization analysis. (B) Summary of the distribution of RNAs of the unspliced PYK2 (PYK2u), PYK2s and PRNK in rat brain. The expression levels of these RNAs are represented by +++ for high level, ++ for moderate level and – for an undetectable level of expression. Estimates for expression were determined by the exposure time needed to observe comparable signals when the same amount of radioactive probes were hybridized with the sections. The exposure times are indicated in Materials and methods.

paxillin via a domain similar to that observed for FAK (Hildebrand et al., 1995). Two proline-rich regions in the C-terminal domain of FAK (and FRNK) have been identified as ligands for SH3 domains of p130cas (Crk-associated substrate) and Graf (GTPase regulator associated with FAK) (Hildebrand et al., 1996). These proline-rich regions are conserved within the Cterminal domain of unspliced PYK2 and spliced PYK2 (PYK2s); however, the first PxxPxR/K motif (P714PKPSR) is not contained within PRNK. As shown in Fig. 6A, GSTPYK2-Cterm bound p130cas, albeit not quite as efficiently as GST-FAK-Cterm. In a parallel experiment, GST-PYK2s (residues 587-946) also bound p130cas, whereas GST-PRNK failed to pull down p130cas from rat brain lysates. To further examine these interactions, we compared the binding of overexpressed unspliced PYK2 and PRNK with GST fusion proteins containing the SH3 domains of p130cas and Graf using in vitro GST pull-down assays. Cell lysates prepared from HEK 293 cells transiently transfected with c-myc-tagged unspliced PYK2 (PYK2), kinase-inactive PYK2 (PYK2-KD), PYK2−Cterm or PRNK were incubated with GST-SH3 domains of p130cas or Graf. The GST beads and their associated proteins were recovered, analyzed by SDS-PAGE, and subjected to immunoblotting using antibodies against cmyc epitope. As shown in Fig. 7A, PYK2-WT, PYK2-KD and PYK2-Cterm were all efficiently precipitated by GST-SH3 domains of p130cas and Graf. However, PRNK failed to stably interact with either p130cas or Graf GST-SH3 domains, although PRNK was expressed at comparable levels (Fig. 7A). These results are consistent with the data shown in Fig. 6A, in that GST-PYK2-Cterm, but not GST-PRNK, was able to ‘pull-down’ p130cas from rat brain cell lysates. The interactions between PYK2 and SH3 domains of p130cas and Graf were selective, since PYK2 and FAK did not bind to PSD95-SH3 (95 kDa post-synaptic density protein), Src-SH3 or Fyn-SH3 domains (Fig. 7B). These results indicate that PRNK binds to a subset of proteins that normally interact with PYK2. Localization of PRNK within focal adhesions Both PRNK and PYK2 contain a putative ‘FAT’ domain, a region similar in sequence to the focal adhesion targeting domain (FAT) in FAK. Ectopic expression of full-length PYK2 in fibroblasts results in accumulation of PYK2 in the cytoplasm and the induction of apoptosis (Xiong and Parsons, 1997). To examine whether the putative ‘FAT’ domain present in PRNK will functionally target PRNK into focal adhesions, the subcellular localization of PRNK was examined in Swiss 3T3 cells. Swiss 3T3 cells were microinjected with the cDNAs encoding GST-PRNK, GSTPRNK∆1-95 and GST-PRNK∆1-131. The injected cells were immunostained using antibodies against both GST-PRNK and paxillin. The GST-PRNK expressed in Swiss 3T3 cells efficiently localized to focal adhesions, colocalizing with paxillin (Fig. 8A,B). The PRNK mutant missing the prolinerich region but retaining the ‘FAT’ domain (GST-PRNK∆195) also efficiently localized to focal adhesions (Fig. 8C,D). In contrast, GST-PRNK∆1-131 failed to localize to focal adhesions (Fig. 8E,F). Hence these data confirm that PRNK contains a functional ‘FAT’ domain and localizes to focal adhesions when expressed in fibroblasts.

Characterization of splice variants of PYK2 1987 DISCUSSION In this report, we characterize two novel isoforms of PYK2, the FAK-related protein tyrosine kinase. One such isoform (designated PYK2s), lacking 42 amino acids within the Cterminal domain, appears to be a splice variant of PYK2. The other isoform of PYK2 (designated PRNK) is derived from mRNAs comprising a unique 5′ leader fused to the PYK2 sequence encoding the C-terminal 238 amino acids of PYK2. PYK2 (unspliced form) appears to be restricted in its expression in the brain, and is poorly expressed in the spleen. However, PYK2s and PRNK are expressed in the spleen. In situ hybridization analysis of rat brain shows that the unspliced PYK2 is expressed in hippocampus, cerebral cortex, olfactory bulb and poorly expressed in cerebellum, whereas PYK2s and

PRNK are expressed in cerebellum and also at low levels in nearly all regions of the brain examined. Both PYK2 and PRNK bind to paxillin. However, each appears to have the capacity to interact with a different set of proteins. PYK2, but not PRNK, interacts with both p130cas and Graf. PRNK when expressed in fibroblasts is also able to localize to focal adhesions by sequences within ‘FAT’ domain. These results suggest that PYK2 and its isoforms may participate in distinct sets of protein interactions and that the complex alternative splicing may provide one of the mechanisms for the functional diversity mediated by PYK2. The PYK2s cDNA characterized in this study was identical to the PYK2 cDNA, except for a 126-base-pair deletion resulting in a PYK2 isoform missing 42 amino acids within the C-terminal domain. PYK2s appears to be generated by

Fig. 5. In situ hybridization analysis using adult rat brain sections hybridized with 33P-labeled antisense probes specific for the unspliced PYK2 (PYK2u, A), PYK2s (B) and PRNK (C). Rat brain sections at the sagittal (A,B,C), coronal (A′,B′,C′) and horizontal (A′′,B′′,C′′) planes were hybridized with 106 cpm/ml 32P-labeled probes for PYK2u (A,A′,A′′), PYK2s (B,B′,B′′) and PRNK (C,C′,C′′). Exposure times were as follows: sections hybridized with PYK2u (A) and sagittal sections hybridized with PRNK (C)-specific oligonucleotides were exposed for 72 hours; all other sections were exposed for 10 days. The cerebral cortex (Cx), hippocampal formation (DG, CA1, CA2 and CA3), thalamus (Th), olfactory bulb (OB), piriform area (Pir), pons (Pn), hypothalamus (Ht), medulla (Me), cerebellum (Cb) and the third ventricle (D3V) are indicated in A,A′,A′′.

1988 W.-C. Xiong and others

Fig. 6. Interaction of paxillin with PRNK. (A) Interaction of paxillin with PRNK identified by in vitro GST ‘pull-down’ assays. GSTfusion proteins (GST-FAK-Cterm, GST-PYK2-Cterm, GST-PRNK, GST-PRNK∆1-95, GST-PRNK∆1-131, GST-PYK2-Nterm, GSTPYK2 (587-988), GST-PYK2s (587-946) and GST alone) were used for the in vitro ‘pull-down’ assay (see Materials and methods). Antibodies that were used for the immunoblotting are indicated. (B) Interaction of paxillin with PRNK identified by in vivo GST ‘pull down’ assays. Cell lysates from HEK 293 cells transfected with GST-tagged PRNK, GST-PRNK∆1-95 and GST-PRNK∆1-131 were incubated with glutathione-Sepharose beads for 1 hour. The binding complexes were subjected to immunoblotting with antibodies against GST and paxillin.

alternative RNA splicing, since the nucleic acid sequence analysis of PYK2s reveals potential splice donor sites at the junctions of the deletion. Based on northern blot analysis with a probe specific for the spliced exon, PYK2 (the unspliced form) mRNA is readily detected in the brain, kidney and lung. Interestingly, the unspliced PYK2 is poorly expressed in the spleen relative to PYK2s expression, suggesting that the 42amino-acid insertion may mediate a function of PYK2 required in neuronal and epithelial cells but not in hematopoietic cells in the spleen. In situ hybridization analysis demonstrate that PYK2 is expressed in certain neuronal cells (e.g. hippocampal, cortex and olfactory bulb), whereas PYK2s is expressed in all regions of rat brain examined, albeit less abundantly, suggesting that the 42-amino-acid insertion may be critical for a function of PYK2 in some, but not all, of the neuronal cells. Thus, alternative RNA splicing of PYK2 may be one of the mechanisms for the regulation of its functional diversity. In support of this hypothesis is the observation of the regulation of FAK function by alternative RNA splicing (e.g. several FAK splicing isoforms exhibit increased autophosphorylation activity) (Andre and Becker-Andre, 1993; Burgaya and Girault, 1996; Burgaya et al., 1997; Schaller et al., 1993).

Fig. 7. Interactions of SH3 domain-containing proteins (p130cas and Graf) with PYK2 and PRNK. (A) PYK2, but not PRNK, is able to interact with the SH3 domains of p130cas and Graf. Cell lysates from HEK 293 cells over-expressing c-myc-tagged wild-type PYK2 (PYK2), kinase-inactive PYK2 (PYK2-KD), PYK2-Cterm and PRNK were incubated with GST-SH3 domains of p130cas or Graf, and the protein complexes associated with the GST beads were subjected to immunoblotting using antibodies against c-myc epitope. (B) The interactions between PYK2 and SH3 domains of p130cas and Graf are specific. The cell lysates from Cos-1 cells over-expressing c-myc-tagged wild-type PYK2 were incubated with different GSTSH3 domain fusion proteins (p130cas-SH3, Graf-SH3, PSD95-SH3, Src-SH3 and Fyn-SH3), and the protein complexes associated with the beads were subjected to immunoblotting with antibodies against c-myc (mAb 9e10), FAK (mAb 2A7) and p130cas (mAb from Transduction Labs).

PRNK cDNA has a long 5′ noncoding region containing multiple poly(T) tracts, which make good candidates for translational regulation (Tzamarias et al., 1986; Roussou et al., 1988). The mechanism by which PRNK transcripts are generated is not clear. The most probable mechanisms are alternative splicing and/or utilization of alternative transcriptional promoters. A number of examples of genes regulated by alternative transcriptional promoters (also called ‘a gene within a gene’) have been described. These include: calmodulin kinase IV and calspermin (Ohmstede et al., 1991; Means et al., 1991); myosin light-chain kinase (MLCK) and a calmodulin-binding protein (KRP) (Collinge et al., 1992). Our laboratory has recently demonstrated that FRNK expression is

Characterization of splice variants of PYK2 1989

Fig. 8. Subcellular localization of PRNK. The cDNAs of GST-PRNK (A, B), GST-PRNK∆1-95 (C,D) and GST-PRNK∆1-131 (E,F) were microinjected into the nuclei of Swiss 3T3 cells. 2 hours after microinjection, the cells were fixed and doubly immunostained using both antibodies against PYK2 (green) and paxillin (red).

regulated by promoter elements, which map to the intron between the exons encoding the kinase domain and the beginning to the C-terminal domain (K. Nolan and J. T. Parsons, unpublished work). Thus, FRNK is an example of ‘a gene within a gene’. Based on the structural similarities between FRNK and PRNK, we speculate that PRNK and PYK2 may be another example for ‘a gene within a gene’. Both PYK2 and PRNK bind to paxillin in vitro and in vivo. These findings are consistent with predictions from the sequence of both PYK2 and PRNK, indicating a region similar to the paxillin binding domain in FAK. Several published reports document the interaction of PYK2 with paxillin (Salgia et al., 1996; Li and Earp, 1997; Schaller and Sasaki, 1997). In contrast to paxillin, PYK2 and PRNK appear to interact differentially with two FAK binding proteins p130cas and Graf. PYK2 efficiently interacts with p130cas and Graf, presumably via two conserved PxxP motifs present in the C-terminal domain of PYK2. The inability of PRNK to interact with Graf and p130cas may be due to the lack of the first proline-rich sequence (PPKPSR) in PRNK, suggesting that the absence of the first PxxP motif may cause a structural change in PRNK that decreases the affinity of PRNK for the SH3 domains of p130cas and Graf. Harte et al. have suggested a cooperative interaction between the first and second PxxP motifs in FAK (Harte et al., 1996). Immunostaining of microinjected Swiss 3T3 cells indicates that PRNK is localized to focal adhesions, and this localization

requires the ‘FAT’ domain (amino acids 868-1009 in PYK2). As the ‘FAT’ domain in FAK is both necessary and sufficient for recruitment of FAK to focal adhesions (Hildebrand et al., 1993), we suggest a similar function for the ‘FAT’ domain in PRNK (Fig. 6). However, PYK2 and PYK2s, both of which contain an intacted ‘FAT’ domain, fail to localize to focal adhesions when expressed in Swiss 3T3 cells (data not shown), consistent with a recent report by Schaller and Sasaki (1997). These data argue that other domains within PYK2 and PYK2s may play a dominant role in localizing PYK2 to specific compartments within the cell. In contrast, expression of PRNK should result in the preferential localization of this protein to focal adhesions. The physiological significance of the alternative splice variant PYK2s remains to be examined. The 42 amino acid sequences deleted in PYK2s are rich in serine and threonine residues, characteristic of protein phosphorylation motifs. Whether deletion of the 42 amino acids may change the susceptibility of PYK2 to phosphorylation by serine/threonine kinases, influence the local tertiary structure or change the localization of PYK2 remains to be examined. The deletion in PYK2s also alters the distance between the two SH3 domain binding motifs, suggesting that PYK2s might have different affinity for the interactions with SH3 domain containing proteins (e.g. p130cas and Graf). The structural organization of PRNK is reminiscent of that of FRNK, the autonomously expressed C-terminal domain of FAK. Whether PRNK functions similarly to FRNK, which acts as an inhibitor of FAK and regulates cell spreading (Richardson and Parsons, 1996; Richardson et al., 1997) remains to be tested. It is interesting to note that several other kinases (e.g. MLCK) also exhibit independent expression of C-terminal domains from a single gene (Means et al., 1991; Ohmstede et al., 1991; Collinge et al., 1992). The independently expressed domains of MLCK (termed KRP) regulate its corresponding full-length protein (Means et al., 1991). Drawing upon these precedents, we speculate that PRNK may function to regulate the activity of PYK2 in some cells. Alternatively, the targeting of PRNK to focal adhesions suggests the possibility that it may serve as a co-regulator of FAK in cells where both FAK and PRNK/PYK2 are expressed, thus providing a means for crosstalk between members of the family kinases. In conclusion, the multiplicity of isoforms, combining with the diversity of possible protein-protein interactions, suggests that PYK2, PYK2s and PRNK may have specific and discrete functions. Future studies will be directed toward understanding the functions of PYK2s and PRNK, and gaining insight into the functions of PYK2. We are grateful to E. M. Talley, Z.-H. Wang and C. Borgman for excellent technical assistance. We also thank L. Mei, D. A. Bayliss, K. Martin, A. Ma and Y. Du for helpful advice and comments on the manuscript. These studies were supported by DHHS grants CA 40042 and CA 29243 to J. T. P; W. C. X is supported by NIH NRSA fellowship NS 09918.

REFERENCES Andre, E. and Becker-Andre, M. (1993). Expression of an N-terminally truncated form of human focal adhesion kinase in brain. Biochem. Biophysic. Res. Comm. 190, 140-147.

1990 W.-C. Xiong and others Astier, A., Avraham, H., Manie, S. N., Groopman, J., Canty, T., Avraham, S. and Freedman, A. S. (1997). The related adhesion focal tyrosine kinase is tyrosine-phosphorylated after β1-integrin stimulation in B cells and binds to p130cas. J. Biol. Chem. 272, 228-232. Avraham, S., London, R., Fu, Y., Ota, S., Hiregowdara, D., Li, J., Jiang, S., Pasztor, L. M., White, R. A., Groopman, J. E. and Avraham, H. (1995). Identification and characterization of a novel related adhesion focal tyrosine kinase (RAFTK) from megakaryocytes and brain. J. Biol. Chem. 270, 27742-27751. Bellis, S. L., Miller, J. T. and Turner, C. E. (1995). Characterization of tyrosine phosphorylation of paxillin in vitro by focal adhesion kinase. J. Biol. Chem. 270, 17437-17441. Berg, N. N. and Ostergaard, H. L. (1997). T cell receptor engagement induces tyrosine phosphorylation of FAK and PYK2 and their association with Lck. J. Immunol 159, 1753-1757. Burgaya, F. and Girault, J. A. (1996). Cloning of focal adhesion kinase, pp125FAK, from rat brain reveals multiple transcripts with different patterns of expression. Mol. Brain. Res. 37, 63-73. Burgaya, F., Toutant, M., Studler, J. M., Costa, A., Bert, M. L., Gelman, M. and Girault, J. A. (1997). Alternatively spliced focal adhesion kinase in rat brain with increased autophosphorylation activity. J. Biol. Chem. 272, 28720-28725. Burridge, K. and Chrzanowska-Wodnicka, M. (1996). Focal adhesions, contractility and signaling. Annu. Rev. Cell Dev. Biol. 12, 463-519. Chen, H.-C. and Guan, J. L. (1994). Association of focal adhesion kinase with its potential substrate phosphatidylinositol 3-kinase. Proc. Nat. Acad. Sci. USA 91, 10148-10152. Chen, H.-C., Appedda, P. A., Parsons, J. T., Hidebrand, J. O., Schaller, M. D. and Guan, J. L. (1995). Interaction of focal adhesion kinase with cytoskeletal protein talin. J. Biol. Chem. 270, 16995-16999. Cobb, B. S., Schaller, M. D., Leu, T. H. and Parsons, J. T. (1994). Stable association of pp60src and pp59fyn with the focal adhesion-associated protein tyrosine kinase, pp125FAK. Mol. Cell Biol. 14, 147-155. Collinge, M., Matrisian, P. E., Zimmer, W. E., Shattuck, R. L., Lukas, T. J., Van Eldik, L. J. and Watterson, D. M. (1992). Structure and expression of a calcium-binding protein gene contained within a calmodulin-regulated protein kinase gene. Mol. Cell. Biol. 12, 2359-2371. Derkinderen, P., Toutant, M., Burgaya, F., Bert, M. L., Siciliano, J. C., Franciscis, V. d., Gelman, M. and Girault, J.-A. (1996). Regulation of neuronal form of focal adhesion kinase by anandamide. Science 273, 17191722. Dikic, I., Tokiwa, G., Lev, S., Courtneidge, S. A. and Schlessinger, J. (1996). A role for Pyk2 and Src in Linking G-protein-coupled receptors with MAP kinase activation. Nature 383, 547-550. Evan, G. I., Lewis, G. K., Ramsay, G. and Bishop, J. M. (1985). Isolation of mono-clonal antibodies specific for human c-myc proto-oncogene product. Mol. Cell Biol. 5, 3610-3616. Frisch, S. M., Vuori, K., Ruoslahti, E. and Chan-Hui, P. Y. (1996). Control of adhesion-dependent cell survival by focal adhesion kinase. J. Cell Biol. 134, 793-799. Grant, S. G., Karl, K. A., Kiebler, M. A. and Kandel, E. R. (1995). Focal adhesion kinase in the brain: novel subcellular localization and specific regulation by Fyn tyrosine kinase in mutant mice. Genes Dev. 9, 19091921. Guan, J. L. and Shalloway, D. (1992). Regulation of pp125FAK both by cellular adhesion and by oncogenic transformation. Nature 358, 690692. Guinebault, C., Payratre, B., Racaud-Sultan, C., Mazarguil, H., Breton, M., Mauco, G., Plantavid, M. and Chap, H. (1995). Integrin-dependent translocation of phosphoinositide 3-kinase to the cytoskeleton of thrmbinactivated platelets involves specific interactions of p85a with actin filaments and focal adhesion kinase. J. Cell Biol. 129, 831-842. Harte, M. T., Hildebrand, J. D., Burnham, M. R., Bouton, A. H. and Parsons, J. T. (1996). p130cas, a substrate associated with v-Src and c-Crk, localizes to focal adhesions and binds to focal adhesion kinase. J. Biol. Chem. 271, 13649-13655. Hildebrand, J. D., Schaller, M. D. and Parsons, J. T. (1993). Identification of sequences required for the efficient localization of the focal adhesion kinase, pp125FAK to cellular focal adhesions. J. Cell Biol. 123, 993-1005. Hildebrand, J. D., Schaller, M. D. and Parsons, J. T. (1995). Paxillin, a tyrosine phosphorylated focal adhesion-associated protein binds to the carboxyl terminal domain of focal adhesion kinase. Mol. Biol. Cell. 6, 637647. Hildebrand, J. D., Taylor, J. M. and Parsons, J. T. (1996). An SH3 domain-

containing GTPase-activating protein for Rho and Cdc42 associates with focal adhesion kinase. Mol. Cell. Biol. 16, 3169-3178. Hiregowdara, D., Avraham, H., Fu, Y., London, R. and Avraham, S. (1997). Tyrosine phosphorylation of the related adhesion focal tyrosine kinase in megakaryocytes upon stem cell factor and phorbol myristate acetate stimulation and its association with paxillin. J. Biol. Chem. 272, 10804-10810. Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, J. K. and Pease, L. R. (1989). Site-directed mutagenesis by overlap extension using polymerase chain reaction. Gene 7, 51-59. Hungerford, J. E., Campton, M. T., Matter, M. L., Hoffstrom, B. G. and Otey, C. A. (1996). Inhibition of pp125FAK in cultured fibroblasts results in apoptosis. J. Cell Biol. 135, 1383-1390. Lev, S., Moreno, H., Martinez, R., Canoll, P., Peles, E., Musacchio, J. M., Plowman, G. D., Rudy, B. and Schlessinger, J. (1995). Protein tyrosine kinase PYK2 involved in Ca2+-induced regulation of ino channel and MAP kinase functions. Nature 376, 737-745. Li, X. and Earp, H. S. (1997). Paxillin is tyrosine-phosphorylated by and preferentially associates with the calcium-dependent tyrosine kinase in rat liver epithelial cells. J. Biol. Chem. 272, 14341-14348. Manie, S. N., Beck, A. R. P., Astier, A., Law, S., Canty, T., Hirai, H., Druker, B. J., Avraham, H., Haghayeghi, N., Sattler, M., et al. (1997). Involvement of p130cas and p105HEF1, a novel cas-like docking protein, in a cytoskeleton-dependent signaling pathway initiated by ligation of integrin or antigen receptor on human B cells. J. Biol. Chem. 272, 4 230-4236. Means, A. R., Cruzalegui, F., Lemagueresse, B., Needleman, D. S., Slaughter, G. R. and Ono, T. (1991). A novel Ca2+/calmodulin-dependent protein kinase and a male germ cell-specific calmodulin-binding protein are derived from the same gene. Mol. Cell. Biol. 11, 3960-3971. Moolenaar, W. H. (1991). Mitogenic action of lysophosphatidic acid. Advan. Can. Res. 57, 87-102. Ohmstede, C. A., Bland, M. M., Merrill, B. M. and Sahyoun, N. (1991). Relationship of genes encoding Ca2+/calmodulin-dependent protein kinase Gr and calspermin: A gene within a gene. Proc. Nat. Acad. Sci. USA 88, 5784-5788. Parsons, J. T. (1996). Integrin-mediated signalling: regulation by protein tyrosine kinases and small GTP-binding proteins. Curr. Opin. Cell Biol. 8, 146-152. Polte, T. R. and Hanks, S. K. (1995). Interaction between focal adhesion kinase and Crk-associated tyrosine kinase substrate p130cas. Proc. Nat. Acad. Sci. USA 92, 10678-10682. Qian, D., Lev, S., van Oers, N. S. C., Dikic, I., Schlessinger, J. and Weiss, A. (1997). Tyrosine phosphorylation of PYK2 is selectively regulated by Fyn during TCR signaling. J. Exp. Med. 185, 1253-1259. Rankin, S. and Rozengurt, E. (1994). Platelet-derived growth factor modulation of focal adhesion kinase and paxillin tyrosine phosphorylation in Swiss 3T3 cells: Bell-shaped dose response and cross-talk with bombesin. J. Biol. Chem. 269, 704-710. Richardson, A., Malik, R. K., Hildebrand, J. D. and Parsons, J. T. (1997). Inhibition of cell spreading by expression of the C-terminal domain of focal adhesion kinase (FAK) is rescued by coexpression of src or catalytically inactive FAK: a role for paxillin tyrosine phosphorylation. Mol. Cell. Biol. 17, 6906-6914. Richardson, A. and Parsons, J. T. (1996). A mechanism for regulation of the adhesion-associated protein tyrosine kinase pp125FAK. Nature 380, 538-540. Roussou, I., Thireos, G. and Hauge, B. M. (1988). Transcriptionaltranslational regulatory circuit in Saccharomyces cerevisiae which involves the GCN4 transcriptional activator and the GCN2 protein kinase. Mol. Cell. Biol. 8, 2132-2139. Rozengurt, E. (1991). Neuropeptides as cellular growth factors: role of multiple signaling pathways. Eur. J. Clin. Invest. 21, 123-134. Salgia, R., Avraham, S., Pisick, E., Li, J.-L., Raja, S., Greenfield, E. A., Sattler, M., Avraham, H. and Griffin, J. D. (1996). The related adhesion focal tyrosine kinase forms a complex with paxillin in hematopoietic cells. J. Biol. Chem. 271, 31222-31226. Sasaki, H., Nagura, K., Ishino, M., Tobioka, H., Kotani, K. and Sasaki, T. (1995). Cloning and characterization of cell adhesion kinase β, a novel protein-tyrosine kinase of the focal adhesion kinase subfamily. J. Biol. Chem. 270, 21206-21219. Schaller, M. D., Borgman, C. A., Cobb, B. S., Vines, R. R., Reynolds, A. B. and Parsons, J. T. (1992). pp125FAK, a structurally distinctive proteintyrosine kinase associated with focal adhesions. Proc. Nat. Acad. Sci. USA 89, 5192-5196. Schaller, M. D., Borgman, C. A. and Parsons, J. T. (1993). Autonomous

Characterization of splice variants of PYK2 1991 expression of a noncatalytic domain of the focal adhesion-associated protein tyrosine kinase pp125FAK. Mol. Cell. Biol. 13, 785-791. Schaller, M. D., Hildebrand, J. D., Shannon, J. D., Fox, J. W., Vines, R. R. and Parsons, J. T. (1994). Autophosphorylation of the focal adhesion kinase, pp125FAK, directs SH2-dependent binding of pp60src. Mol. Cell Biol. 14, 1680-1688. Schaller, M. D. and Sasaki, T. (1997). Differential signaling by the focal adhesion kinase and cell adhesion kinase β. J. Biol. Chem. 272, 2531925325. Schlaepfer, D. D., Hanks, S. K., Hunter, T. and van der Geer, P. (1994). Integrin-mediated signal transduction linked to Ras pathway by Grb2 binding to focal adhesion kinase. Nature 372, 786-791. Tachibana, K., Sato, T., D’Avirro, N. and Morimoto, C. J. (1995). Direct association of pp125FAK with paxillin, the focal adhesion targeting mechanism of pp125FAK. J. Exp. Med. 182, 1089-1099. Talley, E. M., Negar, N. S. and Bayliss, D. A. (1997). Postnatal development of serotonergic innervation, 5-HT1A receptor expression and 5-HT responses in rat motoneurons. J. Neurosci. 17, 4473-4485. Tzamarias, D., Alexandraki, D. and Thireos, G. (1986). Multiple cis-acting

elements modulate the translational efficiency of GCN4 mRNA in yeast. Proc. Nat. Acad. Sci. USA 83, 4849-4853. Xiong, W. C. and Montell, C. (1993). Tramtrack is a transcriptional repressor required for cell fate determination in the drosophila eye. Genes Dev. 7, 1085-1096. Xiong, W. C. and Parsons, J. T. (1997). Induction of apoptosis after expression of PYK2, a tyrosine kinase structurally related to focal adhesion kinase. J. Cell Biol. 139, 529-539. Xu, L. H., Iwens, L. V., Sturge, G. C., Yang, X., Liu, E. T., Craven, R. J. and Cance, W. G. (1996). Attenuation of the expression of the focal adhesion kinase induces apoptosis in tumor cells. Cell Growth Diff. 7, 413418. Yu, H., Li, X., Marchetto, G. S., Dy, R., Hunter, D., Calvo, B., Dawson, T. L., Wilm, M., Anderegg, R. J., Graves, L. M. and Earp, H. S. (1996). Activation of a novel calcium-dependent protein-tyrosine kinase. J. Biol. Chem. 271, 29993-29998. Zachary, I., Sinnett-Smith, J. and Rozengurt, E. (1992). Bombesin, vasapressin and endothelin rapidly stimulation of tyrosine phosphorylation in Swiss 3T3 cells. J. Biol. Chem. 267, 19031-19034.