Regulation of DNA-dependent Protein Kinase by the Lyn Tyrosine Kinase*

The Src-like protein-tyrosine kinase Lyn is activated by ionizing radiation and certain other DNA-damaging agents, whereas the DNA-dependent protein kinase (DNA-PK), consisting of the catalytic subunits (DNA-PKcs) and Ku DNA-binding components, requires DNA double-stranded breaks for activation. Here we demonstrate that Lyn associates constitutively with DNA-PKcs. The SH3 domain of Lyn interacts directly with DNA-PKcs near a leucine zipper homology domain. We also show that Lyn phosphorylates DNA-PKcs but not Ku in vitro. The interaction between Lyn and DNA-PKcs inhibits DNA-PKcs activity and the ability of DNA-PKcsto form a complex with Ku/DNA. These results support the hypothesis that there are functional interactions between Lyn and DNA-PKcs in the response to DNA damage.

Mammalian cells respond to DNA damage with cell cycle arrest, activation of DNA repair, and, in the event of irreparable lesions, the induction of apoptosis. The signals controlling responses to genotoxic stress, while of importance to mutagenesis and treatment with certain anti-cancer agents, remain unclear. Certain insights, however, have been derived from the finding that DNA-damaging agents activate the c-Abl proteintyrosine kinase (PTK) (1)(2)(3)(4). 1 c-Abl is detectable in a nuclear complex with the DNA-dependent protein kinase (DNA-PK) and is activated in part in the response to ionizing radiation (IR) by a DNA-PK-dependent mechanism (5). Activation of c-Abl is associated with binding to the p53 tumor suppressor and the induction of growth arrest in G 1 phase through downregulation of the cyclin-dependent kinase 2 (Cdk2) (6,7). Other studies have demonstrated that c-Abl contributes to DNA damage-induced apoptosis (8,9). These findings have supported a role for c-Abl in regulating the growth arrest and apoptotic responses to genotoxic stress. c-Abl is present in a nuclear complex that includes the Srclike Lyn PTK (10). Lyn, like c-Abl, is activated by IR and other DNA damaging agents (11)(12)(13)(14). The activation of nuclear Lyn by DNA damage is associated with binding of Lyn to Cdc2 (11)(12)(13)(14). Lyn phosphorylates Cdc2 on Tyr-15 and thereby inhibits Cdc2 activity (11)(12)(13)(14). Whereas activation of Cdc2 in a complex with cyclin B is required for the transition of cells from G 2 to M phase (15), inhibition of Cdc2 by Lyn could contribute in part to the arrest at G 2 /M phase following exposure to DNA damaging agents. Alternatively, binding of Lyn to Cdc2 may prevent the interaction of Cdc2 with proteins such as c-Src that play a functional role in mitosis (16). Other studies have indicated that the activation of Lyn is not restricted to cells in G 2 . In this context, arrest of cells in G 1 /S phase by 1-␤-D-arabinofuranosylcytosine is associated with activation of Lyn and binding of Lyn to Cdk2 (17). Thus, the available evidence suggests that the Lyn PTK, like c-Abl, plays a role in the cell cycle arrest response to DNA damaging agents.
DNA-PK, a complex of three proteins, is involved in the repair of DNA double-stranded breaks, V(D)J recombination, and transcription (18 -22). The 470-kDa catalytic subunit of DNA-PK (DNA-PK cs ) is activated by binding with the 70-and 80-kDa Ku heterodimer to sites of DNA damage (23)(24)(25)(26). Under some conditions, DNA-PK is a self-contained kinase that is activated by direct interaction with double-stranded DNA, whereas Ku stabilizes binding of DNA-PK to DNA ends (21,27). Among the many substrates of activated DNA-PK cs are p53, c-Abl, and Cdc2 (5,19,26). Autophosphorylation of DNA-PK cs inhibits its activity by inducing the dissociation of DNA-PK cs from DNA (28). Phosphorylation of DNA-PK cs by c-Abl also inhibits the ability of DNA-PK to form a complex with DNA (5). c-Abl, like Ku, associates with DNA-PK cs near the kinase homology region (29). c-Abl phosphorylates DNA-PK cs in the C-terminal region and thereby dissociates DNA-PK cs from the Ku-DNA complex (29). Otherwise, little is known about the regulation of DNA-PK activity.
The findings that c-Abl forms a nuclear complex with Lyn (10) and that c-Abl interacts with DNA-PK prompted studies on a potential interaction between Lyn and DNA-PK. The present results demonstrate that Lyn binds directly to DNA-PK. We also show that Lyn phosphorylates DNA-PK cs and inhibits DNA-PK cs activity.

MATERIALS AND METHODS
Cell Culture-U-937 monoblastic leukemia cells (ATCC, Rockville, MD) were grown in RPMI 1640 media containing 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin, 100 mg/ml streptomycin, and 2 mM L-glutamine. Irradiation was performed at room temperature using a Gammacell-1000 (Atomic Energy of Canada, Ottawa, ON, Canada) under aerobic conditions with 137 Cs source emitting at a fixed dose rate of 0.76 gray/min.
Immunoprecipitation and Immunoblot Analysis-Preparation of cell lysates and immunoprecipitations were performed as described (3). Soluble proteins were incubated with anti-Lyn (Upstate Biotechnology Inc., Lake Placid, NY) or anti-DNA-PK (Upstate Biotechnology Inc.) for 1 h and precipitated with protein A-Sepharose for an additional 1 h. The * This investigation was supported by Public Health Service Grants CA75216 (to S. K.) and CA55241 (to D. K.) awarded by the National Cancer Institute, Department of Health and Human Services. 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.
Production of GST-Lyn Fusion Protein-The pGEX plasmid encoding a GST-Lyn (amino acids 1-243) fusion protein (30) was transfected into Escherichia coli DH5␣, and the fusion protein was prepared by inducing log phase cells with isopropyl-␤-D-thiogalactopyranoside. The cell pellets were lysed by sonication. The fusion proteins were purified by affinity chromatography using glutathione-Sepharose beads (Amersham Pharmacia Biotech) as described (30) and equilibrated in lysis buffer.
Fusion Protein Binding Assays and Immunoblotting-Total cell lysates were incubated with 5 g of immobilized GST or the indicated GST fusion proteins for 2 h at 4°C (31). Protein complexes were washed three times with lysis buffer and boiled for 5 min in SDS sample buffer. The complexes were separated by electrophoresis in 5% SDS-polyacrylamide gels and then transferred to nitrocellulose paper. After blocking in 5% dry milk in phosphate-buffered saline and 0.05% Tween 20 for 1 h at room temperature, the filters were incubated for 1 h with anti-DNA-PK antibody and analyzed as described above.
In Vitro Transcription/Translation of DNA-PK Fragments-DNA-PK cs polypeptides were prepared using a coupled in vitro transcription/ translation method (Promega) with templates generated from the DNA-PK cs cDNA by polymerase chain reaction as described (29). Lyn binding to in vitro translated DNA-PK cs polypeptides was assayed by incubating GST-Lyn SH3 (5 g) in 20 l of MB (10 mM Tris, pH 7.4, 150 mM NaCl, 1% Triton X-100, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, pepstatin, and aprotinin) with equal amounts of each in vitro translated [ 35 S]DNA-PK cs product for 1-2 h at 4°C. A separate incubation of the [ 35 S]DNA-PK cs product with GST was used as a negative control. The beads were washed four times in 1 ml of MB at 4°C, and proteins were eluted by boiling in 2ϫ SDS sample buffer (100 mM Tris-HCl, pH 7.0, 4% SDS, 720 nM 2-mercaptoethanol, 5 mg/ml bromphenol blue). Samples were analyzed by SDS-PAGE and autoradiography.
Direct Interaction of DNA-PK cs with Lyn-Purified DNA-PK (Promega) was incubated with purified Lyn (Upstate Biotechnology Inc.) in lysis buffer for 1 h at 4°C. After incubation, proteins were subjected to immunoprecipitation with anti-Lyn or PIRS, and the precipitates were analyzed by immunoblotting with anti-DNA-PK antibody.
In Vitro Phosphorylation of DNA-PK by Lyn-Purified DNA-PK (Promega, 0.5 g) was incubated in the absence of DNA in kinase buffer containing [␥-32 P]ATP with purified kinase-active Lyn (Upstate Biotechnology Inc.) for 20 min at 30°C. The reaction was terminated by the addition of SDS-PAGE sample buffer, and reaction products were analyzed by SDS-PAGE and autoradiography. In vitro kinase reactions containing active or heat-inactivated (HI) Lyn and DNA-PK cs were also performed with cold ATP. Proteins were separated by SDS-PAGE, To ensure that phosphorylation was not due to DNA-PK, 20 M wortmannin (Sigma) was added to inhibit DNA-PK activity (5). The supernatant fraction was obtained by sedimentation of the beads. Following washing of the beads with kinase buffer, the beads and supernatant fraction were boiled in SDS sample buffer. Proteins were separated by 5% or 10% SDS-PAGE and analyzed by immunoblotting with anti-DNA-PK or anti-Ku antibodies. Signal intensities were determined by densitometric analysis (UltroScan).
Inactivation of DNA-PK by Phosphorylation with Lyn-Purified DNA-PK was incubated with purified kinase-active Lyn or HI Lyn in kinase buffer containing [␥ 32 P]ATP for 15 min at 30°C. The reactions containing the phosphorylated DNA-PK were then incubated with GST-p53 and [␥ 32 P]ATP in KB for an additional 15 min at 30°C. As controls, purified Lyn was also incubated with GST-p53 in the presence and absence of DNA. The kinase reactions were stopped by boiling in 2ϫ SDS sample buffer. Eluted proteins were analyzed by SDS-PAGE and autoradiography.

RESULTS AND DISCUSSION
To determine whether nuclear Lyn associates with DNA-PK, we subjected anti-Lyn immunoprecipitates to immunoblot analysis with anti-DNA-PK antibodies. Whereas immunoprecipitates obtained with PIRS had no detectable DNA-PK cs , immunoprecipitation with the anti-Lyn antibody revealed the presence of complexes containing DNA-PK cs (Fig. 1A). To evaluate the stoichiometry of interaction between DNA-PK cs and Lyn, we subjected U-937 cell lysates to immunoprecipitation with anti-Lyn and analyzed the precipitates by immunoblotting with anti-DNA-PK. Signal intensities from before and after anti-Lyn immunoprecipitation were compared by laser densitometric scanning. The results demonstrate that approximately 25% of DNA-PK cs is associated with Lyn (data not shown). Because exposure of cells to IR is associated with activation of both Lyn (11) and DNA-PK (24), we also investigated whether IR affects the interaction between these proteins. IR treatment was associated with a reproducible increase in the association of Lyn with DNA-PK cs to some extent ( Fig.   FIG. 1. Association of DNA-PK and Lyn. A, lysates from U-937 cells were subjected to immunoprecipitation with PIRS or anti-Lyn. Immunoprecipitates were analyzed by immunoblotting with anti-DNA-PK. Cell lysate was also subjected to immunoprecipitations with anti-DNA-PK as a positive control. B, U-937 cells were treated with 20-gray IR and harvested at the indicated times. Lysates were then subjected to immunoprecipitation with anti-Lyn. Lysate was also used directly as a positive control. Immunoprecipitates were analyzed by immunoblotting with anti-DNA-PK. C, schematic diagrams of various Lyn constructions. 1B). To confirm binding of Lyn and DNA-PK cs , cell lysates were incubated with a GST fusion protein prepared from the 1-243 amino acid fragment of Lyn that includes a unique N-terminal region, Src homology 3 (SH3) and SH2 domains but not kinase domains (30) (Fig. 1C). Adsorbates obtained with GST-Lyn 1-243 but not with GST demonstrated binding of DNA-PK cs ( Fig. 1D and data not shown). Adsorbates obtained with GST-Lyn 1-131 and GST-Lyn 27-131 but not with GST-Lyn 131-243 also revealed specific binding of DNA-PK cs to the Lyn SH3 domain (Fig. 1D). These findings indicate that the SH3 domain of Lyn contributes to the association with DNA-PK.
To assess whether Lyn binds directly to DNA-PK cs , we incubated purified DNA-PK cs with Lyn. Anti-Lyn immunoprecipitates were analyzed by immunoblotting with anti-DNA-PK. In contrast to PIRS, immunoprecipitation with anti-Lyn revealed the presence of DNA-PK cs ( Fig. 2A). These findings support a direct interaction between DNA-PK cs and Lyn. To assess where Lyn binds to DNA-PK cs , we prepared fragments of DNA-PK cs (Fig. 2B) by in vitro transcription/translation (29). Equal quantities of the DNA-PK cs in vitro translation products were incubated with GST-Lyn 1-243 bound to glutathione-Sepharose beads. Analysis of the adsorbates by autoradiography demonstrated binding of Lyn to DNA-PK cs fragments 8, 11, and 14 (data not shown). These results support direct binding of Lyn to a leucine zipper region in DNA-PK cs (amino acids 1503-1538). Various fragments of DNA-PKcs were also studied for binding to the Lyn SH3 domain. The Lyn SH3 domain binds to prolinerich sequences with the consensus XPPXXPX. Consistent with the presence of such a sequence (REFPPGTPRFNN, amino acids 1744 -1755; fragments 11 and 14) in DNA-PK cs , the results of Lyn SH3 binding assays with in vitro translated DNA-PK cs fragments 11 and 14 confirmed a direct interaction (Fig. 2, C and D, and data not shown). By contrast, there was no apparent binding of Lyn SH3 to other DNA-PK cs fragments (Fig. 2B). Collectively, these studies indicate that the association between Lyn and DNA-PK cs occurs by direct interaction of the Lyn SH3 domain with DNA-PK cs amino acids 1520 -1976.
To determine whether Lyn phosphorylates DNA-PK cs , we incubated purified DNA-PK with active or HI Lyn. There was no detectable phosphorylation of DNA-PK cs in the presence of HI Lyn (data not shown). By contrast, kinase reactions that included kinase-active Lyn demonstrated phosphorylation of DNA-PK cs (Fig. 3A). Lyn phosphorylation of DNA-PK cs did not require DNA-bound DNA-PK. Also, As a positive control, DNA-PK was incubated with DNA beads to show autophosphorylation of DNA-PK cs (Fig. 3A). To confirm Lyn-dependent phosphorylation of DNA-PK cs , we incubated purified DNA-PK cs with active or HI Lyn in kinase buffer containing cold ATP. The phosphorylated products were analyzed by immunoblotting with anti-P-Tyr. The results demonstrate Lyn-dependent tyrosine phosphorylation of DNA-PK cs (Fig. 3B). DNA-PK cs forms a complex with the 70-and 80-kDa Ku heterodimer (26,32), and thus we also incubated Lyn with purified Ku in a kinase reaction. In contrast to DNA-PK cs , there was no detectable phosphorylation of Ku (Fig. 3C). These findings indicate that Lyn phosphorylates DNA-PK cs and not Ku.
The activation of DNA-PK by binding of DNA-PK cs to Ku/ DNA is associated with phosphorylation of p53 (19). Therefore to assess the effect of Lyn on DNA-PK activity, we incubated purified DNA-PK cs /Ku/DNA complexes with active or inactive Lyn and assessed DNA-PK cs activity using GST-p53 as a substrate. In the presence of DNA, DNA-PK cs /Ku phosphorylated GST-p53 (Fig. 4A, lane 1). By contrast, addition of either active or inactive Lyn to the reaction inhibited GST-p53 phosphorylation (Fig. 4A, lanes 2 and 3). In the absence of DNA-PK, Lyn had little effect on phosphorylation of GST-p53 (Fig. 4A, lane  4). These findings indicate that the direct interaction of Lyn with DNA-PK cs , and not necessarily the Lyn kinase function, contributes to the inactivation of DNA-PK cs . To further assess the functional significance of the interaction of Lyn and DNA-PK, the DNA-PK/Ku complex was bound to DNA beads and incubated with active or HI Lyn. The reactions included wortmannin to inhibit DNA-PK cs autophosphorylation and thereby inhibit autodissociation from DNA (28). Incubation with kinase-active or kinase-inactive Lyn resulted in release of DNA-PK cs from the beads into the supernatant (Fig. 4B). By contrast, addition of the MEK1 serine/threonine kinase had no detectable effect on release of DNA-PK cs from DNA beads (Fig.  4B). Whereas Lyn phosphorylates DNA-PK cs , but not Ku, we also asked whether Lyn affects the interaction between Ku and DNA. In the presence of Lyn, most of the Ku remained associated with the DNA beads (Fig. 4C). Similar results were obtained with the MEK1 kinase (Fig. 4C). DNA-PK cs released from Ku/DNA in the presence of Lyn as compared with release of the DNA-PK cs /Ku complex from DNA was assessed by immunoblot analysis of proteins in the supernatant and bound to the beads. The results demonstrate that approximately 40% of DNA-PK cs is released from Ku in the presence of Lyn (Fig. 4D). By contrast, only 2-4% of DNA-PK cs /Ku complexes were released from the DNA beads in the presence of Lyn (Fig. 4D). These findings demonstrate that interaction of Lyn and DNA-PK cs results in dissociation of DNA-PK cs and Ku, and thereby

FIG. 2. Direct interaction between
Lyn and DNA-PK cs . A, purified DNA-PK was incubated with Lyn in lysis buffer for 1 h at 4°C, and anti-Lyn immunoprecipitates were analyzed by immunoblotting with anti-DNA-PK. PIRS was used as a negative control. B, protein fragments from various regions of the DNA-PK cs . C, GST Lyn 1-131 bound to glutathione-Sepharose was incubated with in vitro translated fragments of [ 35 S]DNA-PK cs . Following washing, samples were separated in 10% SDS-PAGE gels and analyzed by autoradiography. D, GST-Lyn 1-131 or GST bound to glutathione-Sepharose was mixed with [ 35 S]DNA-PK cs fragments 11 or 14. Following washing, the bound proteins were analyzed by SDS-PAGE and autoradiography. Arrows indicate specific DNA-PK cs polypeptides. IB, immunoblot. directs inhibition of DNA-PK cs activity.
DNA-PK is essential in the repair of DNA double-stranded breaks that form in irradiated cells (20,33,34). Autophosphorylation inactivates DNA-PK by a mechanism in which DNA-PK cs dissociates from Ku (28). Other studies have shown that c-Abl negatively regulates DNA-PK in the response to DNA damage (5). The present studies demonstrate that DNA-PK is also regulated by Lyn. DNA-PK constitutively associates with Lyn by direct binding of the Lyn SH3 domain to an internal region of DNA-PK cs that includes a leucine zipper. Lyn also phosphorylates DNA-PK cs . The in vitro findings indicate that the direct binding of Lyn to DNA-PK cs is sufficient to inhibit DNA-PK cs activity. Thus, constitutive binding of Lyn and DNA-PK cs could regulate the accessibility of certain pools of DNA-PK cs for interaction with Ku/DNA complexes. Lyn-mediated phosphorylation of DNA-PK cs represent another level of DNA-PK cs regulation. These results are in concert with the demonstration that the interaction between DNA-PK cs and Lyn induces the dissociation of DNA-PK cs from the Ku/DNA complex and thereby inhibits DNA-PK cs activity. The activation of Lyn by IR and other DNA damaging agents contributes to the down-regulation of Cdc2 (13)(14)(15)(16), indicating that Lyn is an effector of cell cycle progression in the response to DNA damage. The present findings support a function for Lyn in the regulation of DNA repair. Accordingly, interactions between Lyn and DNA-PK cs may play a role in releasing DNA-PK cs from Ku/DNA complexes after repair to permit relocation at new sites of DNA damage.