Activation of SAPK/JNK Signaling by Protein Kinase Cδ in Response to DNA Damage*

The cellular response to genotoxic stress includes activation of protein kinase Cδ (PKCδ). The functional role of PKCδ in the DNA damage response is unknown. The present studies demonstrate that PKCδ is required in part for induction of the stress-activated protein kinase (SAPK/JNK) in cells treated with 1-β-d-arabinofuranosylcytosine (araC) and other genotoxic agents. DNA damage-induced SAPK activation was attenuated by (i) treatment with rottlerin, (ii) expression of a kinase-inactive PKCδ(K-R) mutant, and (iii) down-regulation of PKCδ by small interfering RNA (siRNA). Coexpression studies demonstrate that PKCδ activates SAPK by an MKK7-dependent, SEK1-independent mechanism. Previous work has shown that the nuclear Lyn tyrosine kinase activates the MEKK1 → MKK7 → SAPK pathway but not through a direct interaction with MEKK1. The present results extend those observations by demonstrating that Lyn activates PKCδ, and in turn, MEKK1 is activated by a PKCδ-dependent mechanism. These findings indicate that PKCδ functions in the activation of SAPK through a Lyn → PKCδ → MEKK1 → MKK7 → SAPK signaling cascade in response to DNA damage.

The mechanisms by which DNA damage is converted into intracellular signals that control the mammalian genotoxic stress response are largely unknown. Certain insights were derived from the finding that cells respond to agents that arrest DNA replication or damage DNA with induction of c-jun and other early response genes (1)(2)(3)(4)(5). Subsequent work showed that DNA damage is associated with activation of the stress-activated protein kinase/c-Jun N-terminal kinase (SAPK/JNK) 1 (6 -12). Phosphorylation of the c-Jun N terminus by SAPK activates the c-Jun transcription function and thereby autoinduction of the c-jun gene (13,14). SAPK also activates early response gene expression by phosphorylation of the ATF2 and Elk1 transcription factors (15)(16)(17). The available evidence indicates that genotoxic stress activates a nuclear complex of the c-Abl and Lyn protein-tyrosine kinases (18,19) and that this complex regulates SAPK activation (7,20). c-Abl functions directly upstream to MEKK1 and activates SAPK by a SEK1dependent mechanism (8,21). By contrast, whereas Lyn activates the MEKK1 3 MKK7 3 SAPK pathway, the direct downstream effector of Lyn is not known (20).
The protein kinase C (PKC) family of serine/threonine kinases has been subdivided into the following: (i) the conventional PKCs (␣, ␤, and ␥), which are calcium-dependent and activated by diacylglycerol, (ii) the novel PKCs (␦, ⑀, , and ), which are calcium-independent and activated by diacylglycerol, and (iii) the atypical PKCs ( and ), which are calciumindependent and not activated by diacylglycerol (22). The ubiquitously expressed novel PKC, PKC␦, is tyrosine-phosphorylated and activated by c-Abl in the response to DNA damage (23). As found for c-Abl and Lyn (24,25), PKC␦ interacts with the nuclear DNA-dependent protein kinase catalytic subunit (DNA-PKcs) (26). Phosphorylation of DNA-PKcs by PKC␦ inhibits the function of DNA-PKcs to form complexes with DNA and to phosphorylate downstream targets (26). Other studies have demonstrated that the nuclear complex of c-Abl and Lyn includes the protein-tyrosine phosphatase, SHPTP1 (27,28), and that PKC␦ phosphorylates and inactivates SHPTP1 in the response to DNA damage (29). In cells that respond to genotoxic stress with the induction of apoptosis, PKC␦ is cleaved by caspase-3 to a constitutively active catalytic fragment (PKC␦CF) (30,31). The finding that PKC␦CF induces nuclear condensation and DNA fragmentation has indicated that cleavage of PKC␦ contributes to the apoptotic response (32).
The present studies have addressed the involvement of PKC␦ in the activation of stress signals in response to arrest of DNA replication and to DNA damage. The results demonstrate that PKC␦ is required in part for activation of SAPK. The results also demonstrate that PKC␦ transduces Lyn-mediated signals to MEKK1 in a Lyn 3 PKC␦ 3 MEKK1 3 MKK7 3 SAPK pathway.

MATERIALS AND METHODS
Cell Culture-Human U-937 myeloid leukemia cells were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM L-glutamine. MCF-7 cells, HeLa cells, and 293T embryonal kidney cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and antibiotics. Cells were treated with araC (Sigma-Aldrich), rottlerin (Calbiochem), or Gö6976 (Calbiochem). Irradiation was performed at room temperature with a Gammacell 1000 (Atomic Energy of Canada) and a 137 Cs source emitting at a fixed rate of 0.21 gray/min.
Glycerol Gradient Sedimentation Analysis-Cosedimentation analysis of PKC␦ and Lyn by glycerol gradient ultracentrifugation was performed as described (35).
Direct Binding Assays-Assays for direct binding of PKC␦ and Lyn in vitro were performed as described previously (35).

PKC␦ Functions in Activation of SAPK in Response to DNA
Damage-To investigate whether PKC␦ functions as an upstream effector of the SAPK response to genotoxic agents, U-937 cells were pretreated with PKC␦ inhibitor, rottlerin, followed by exposure to araC. Anti-SAPK immunoprecipitates were subjected to in vitro kinase assays by phosphorylating GST-c-Jun. The results demonstrate that pretreatment with rottlerin attenuates araC-induced activation of SAPK (Fig. 1A). By contrast, there was no detectable effect of pretreatment with the PKC␣ and PKC␤ inhibitor, Gö6976 (Fig. 1A). Similar results were obtained with U-937 cells exposed to ionizing radiation (IR) (Fig. 1B). To determine whether PKC␦ is required for the induction of other MAPK family members, ERK and p38 MAPK, cells exposed to rottlerin and then araC or IR were subjected to immunoprecipitation with anti-ERK2 and anti-p38 MAPK. Analysis of the immunoprecipitates for phosphorylation of myelin basic protein (MBP) or ATF-2 demonstrated that rottlerin has less of an inhibitory effect on araC-or IR-induced p38 MAPK activity and no apparent effect on ERK activation ( Fig. 1, C and D and data not shown). These results suggest that activation of SAPK is at least in part dependent on PKC␦ in the response to genotoxic agents.
Attenuation of SAPK Activation in Cells Expressing a Kinaseinactive PKC␦ Mutant-To further assess involvement of PKC␦ in SAPK activation, a Myc-tagged kinase-inactive PKC␦(K-R) mutant was introduced stably in U-937 cells. In contrast to U-937 cells expressing the empty vector, two separate clones of transfected cells expressed Myc-PKC␦(K-R) ( Fig. 2A). Anti-PKC␦ immunoprecipitates from control and araC-treated U-937/neo and U-937/PKC␦(K-R) cells were analyzed for phosphorylation of histone H1. The results demonstrate that activation of PKC␦ by araC is abrogated in U-937 cells expressing PKC␦(K-R) (Fig. 2B). To confirm the involvement of PKC␦ in DNA damage-induced activation of SAPK, U-937/neo and U-937/PKC␦(K-R) cells exposed to araC were analyzed for SAPK activity. The results demonstrate that activation of SAPK is attenuated, but not completely inhibited, in araCtreated cells expressing kinase-negative PKC␦(K-R) (Fig. 2C). Similar results were obtained in IR-treated cells (Fig. 2D). By contrast, DNA damage-induced ERK and p38 MAPK activation were unaffected in U-937/PKC␦(K-R) cells, as compared with that in U-937/neo cells (data not shown). These findings indi- GST-c-Jun phosphorylation was assessed by SDS-PAGE and autoradiography (upper panels). The lysates were also subjected to immunoblot analysis (IB) with anti-SAPK (lower panel). C and D, cells were pretreated with rottlerin and then exposed 10 M araC. Anti-ERK (C) or anti-p38 MAPK (D) immunoprecipitates were subjected to immune complex kinase assays with MBP or GST-ATF2, respectively, as substrate (upper panels). The lysates were also subjected to immunoblot analysis with anti-ERK (C) or anti-p38 MAPK (D) (lower panels).
cate that PKC␦ is involved in SAPK activation in the cellular response to diverse genotoxic agents.
Attenuation of SAPK Activation in Cells Transfected with PKC␦ siRNAs or a Kinase-inactive MKK7(K-R) Mutant-To further assess the role of PKC␦ in SAPK activation, 293T cells were treated with siRNA duplexes that target PKC␦. The results demonstrate that the PKC␦siRNA1 down-regulates PKC␦ expression (Fig. 3A). Less pronounced results were obtained with PKC␦siRNA2 (Fig. 3A). As a control, treatment with a GFPsiRNA had little if any effect (Fig. 3A). Importantly, treatment with PKC␦siRNA, but not GFPsiRNA, attenuated araCinduced SAPK activation (Fig. 3B). To determine whether the kinase function of PKC␦ is sufficient for SAPK activation, 293T cells were transfected with PKC␦CF. Analysis of anti-SAPK immunoprecipitates for phosphorylation of GST-c-Jun demonstrated that expression of PKC␦CF induces SAPK activity (Fig.  3C). By contrast, expression of kinase-inactive PKC␦CF(K-R) had little effect on SAPK activity (Fig. 3C). Similar results were obtained in HeLa cells expressing PKC␦CF or PKC␦CF(K-R) (data not shown). To define the effectors downstream to PKC␦ in the SAPK pathway, 293T cells were co-transfected with PKC␦CF and kinase-inactive SEK1(K-R) or MKK7(K-R). Analysis of anti-SAPK immunoprecipitates demonstrated that expression of MKK7(K-R), but not SEK1(K-R), is associated with attenuation of SAPK activity by PKC␦ (Fig. 3D). These results indicate that PKC␦ induces SAPK activation by an MKK7-dependent, SEK1/MKK4-independent mechanism.
Lyn Tyrosine Kinase Is an Upstream Effector of PKC␦ in Response to Genotoxic Stress-Previous studies have demonstrated that the Lyn tyrosine kinase is activated in the re-sponse to genotoxic stress and that Lyn regulates SAPK activation by an MKK7-dependent pathway (20). Whereas the present results indicate that the PKC␦-mediated SAPK signaling pathway is also MKK7-dependent, we asked whether Lyn is involved in PKC␦ 3 SAPK signaling. To determine whether PKC␦ binds to Lyn, we incubated glutathione beads containing GST or GST-PKC␦ with purified Lyn. Analysis of the adsorbates demonstrated binding of Lyn with GST-PKC␦ and not GST (Fig. 4A). To determine whether PKC␦ phosphorylates Lyn in vitro, heat-inactivated purified Lyn was incubated with recombinant PKC␦. Analysis of the reaction products demonstrated no detectable phosphorylation of Lyn by PKC␦ (Fig. 4B,  lane 2). By contrast, incubation of heat-inactivated PKC␦ with purified Lyn showed that Lyn phosphorylates PKC␦ (Fig. 4B,  lane 5). These findings provide support for a direct interaction between PKC␦ and Lyn.
To determine whether Lyn contributes to the regulation of PKC␦ in vivo, anti-Lyn immunoprecipitates from control and araC-treated cells were analyzed by immunoblotting with anti-PKC␦. The results demonstrate that Lyn associates constitutively with PKC␦ and that araC treatment has little if any effect on the extent of the interaction (Fig. 5A). These results were confirmed when anti-PKC␦ immunoprecipitates were subjected to immunoblotting with anti-Lyn (Fig. 5A). A constitutive association between Lyn and PKC␦ was also observed in MCF-7 and HeLa cells (data not shown). To confirm these findings with another approach, lysates from U-937 cells were separated by sedimentation in a glycerol gradient. Analysis of the gradient fractions by immunoblotting with anti-Lyn and anti-PKC␦ demonstrated cosedimentation of Lyn and PKC␦ ( Fig. 5B). To extend the analysis, anti-PKC␦ immunoprecipitates from araC-treated U-937/neo and U-937/Lyn(K-R) cells were analyzed for tyrosine phosphorylation of PKC␦ by immu-noblotting with anti-phospho-Tyr. The results demonstrate that expression of Lyn(K-R) blocks tyrosine phosphorylation of PKC␦ in response to araC (Fig. 5C). In concert with these findings and the demonstration that PKC␦ is activated by tyrosine phosphorylation (23), Lyn(K-R) also attenuated the induction of PKC␦ activity by araC (Fig. 5D, left). Conversely, to determine whether PKC␦ regulates Lyn in the DNA damage response, anti-Lyn immunoprecipitates from U-937/neo and U-937/PKC␦(K-R) cells were analyzed for Lyn activity by assessing autophosphorylation and transphosphorylation of enolase. The finding that araC-induced activation of Lyn is comparable in U-937/neo and U-937/PKC␦(K-R) cells (Fig. 5D,  right) suggests that Lyn may function as an upstream effector of PKC␦.
PKC␦ Is an Upstream Effector of MEKK1-Our previous studies showed that Lyn-induced SAPK activation is MEKK1dependent (20). To investigate the possibility that PKC␦ interacts with MEKK1, anti-MEKK1 immunoprecipitates were analyzed by immunoblotting with anti-PKC␦. The results show that complexes of PKC␦ and MEKK1 are detectable in control and araC-treated cells (Fig. 6A, left panel). araC-induced increases in the association of PKC␦ and MEKK1 were detectable at 0.5 h and maximal at 2 h of treatment (Fig. 6A, right panel). To further assess the interaction between PKC␦ and MEKK1, we first performed in vitro kinase assays by incubating recombinant PKC␦ with GST or GST-MEKK1 in the presence of  [␥-32 P]ATP. Analysis of the products demonstrated that PKC␦ phosphorylates MEKK1 in vitro (Fig. 6B). To further define whether PKC␦ activates MEKK1, GST-MKK7(K-R) was coincubated with or without recombinant PKC␦ and kinase-active MEKK1 (yeast-derived) in the presence of [␥-32 P]ATP. The results demonstrate that MEKK1 phosphorylates GST-MKK7(K-R) and that the addition of recombinant PKC␦ increases MEKK1-mediated MKK7(K-R) phosphorylation (Fig.  6C, lane 3). As a control, there was no detectable phosphorylation of MKK7(K-R) by PKC␦ (Fig. 6C, lane 1). To extend these findings, anti-SAPK immunoprecipitates from 293T cells expressing PKC␦CF and FLAG-tagged kinase-inactive MEKK1(K-M) were analyzed for phosphorylation of GST-c-Jun. The results demonstrate that expression of MEKK1(K-M) decreases PKC␦-induced SAPK activation in a dose-dependent manner (Fig. 6D). These results collectively indicate that MEKK1 is a downstream effector of PKC␦ in the SAPK signaling pathway.

PKC␦ Is Activated by Different Mechanisms in DNA Damage
Response-Involvement of PKC␦ in the genotoxic stress response has been supported by the finding that both arrest of DNA replication and induction of DNA lesions are associated with PKC␦ activation (23,29). The available evidence indicates that full-length PKC␦ is activated as an early event within 1 h of exposure to genotoxic agents (23,29). Phosphorylation of PKC␦ on tyrosine as a mechanism for PKC␦ activation by DNA-damaging agents is mediated in part by the c-Abl kinase (23,36). The results of the present study demonstrate that Lyn also phosphorylates PKC␦ and that Lyn-mediated tyrosine phosphorylation of PKC␦ contributes to PKC␦ activation.
PKC␦ is also activated as a later event (3-6 h) in the DNA damage response by caspase-3-mediated proteolytic cleavage (23,30,31,37). The resulting C-terminal 40-kDa fragment contains the catalytic domain, which, in the absence of the N-terminal regulatory domain, is constitutively active (30 -32). The finding that tyrosine phosphorylation of PKC␦ is required for activation of caspase-3 and thereby PKC␦ cleavage has supported a link between both mechanisms of PKC␦ activation (36). Whereas expression of PKC␦CF induces characteristics of apoptosis (32), the precise events responsible for this response are unknown but may involve an interaction between PKC␦CF and DNA-PKcs (26). In contrast to the activation of PKC␦ by tyrosine phosphorylation, cleavage of PKC␦ to the constitutively active catalytic fragment is irreversible and thus may function in the prolonged stimulation of multiple pro-apoptotic pathways.
Role for PKC␦ in SAPK Activation-SAPK is activated in diverse cell types by agents, such as araC, that block DNA replication and by others, such as IR, that induce DNA lesions (7)(8)(9)(10)(11)(12). Like SAPK, PKC␦ is also activated by both arrest of DNA replication and DNA damage (23,29). The present results demonstrate that treatment of cells with the PKC␦ inhibitor, rottlerin, attenuates activation of SAPK in response to genotoxic stress. In concert with these findings, expression of the kinase-inactive PKC␦(K-R) mutant attenuated activation of both PKC␦ and SAPK in response to araC and IR. Moreover, down-regulation of PKC␦ expression by siRNA was associated with attenuation of DNA damage-induced SAPK activation. These findings provided support for the involvement of PKC␦ as an upstream effector in the regulation of SAPK activation. The results also demonstrate that inhibition of PKC␦ signaling attenuates the early (Ͻ1 h) and later periods of SAPK activation. Thus, SAPK is activated by signals transduced by PKC␦ and possibly, after activation of caspase-3, by PKC␦CF. In this context, the results show that expression of PKC␦CF is also associated with SAPK activation.
Previous studies have shown that nuclear c-Abl is an upstream effector of the SAPK response to both arrest of DNA replication and DNA damage (7,8). Other work has demonstrated that activated forms of Abl induce SAPK activity (5,38,39). Lyn forms a nuclear complex with c-Abl and, like c-Abl, also contributes to SAPK activation in response to genotoxic stress (20). Whereas c-Abl activates SAPK by a SEK1-dependent mechanism, Lyn 3 SAPK signaling is mediated by MKK7 (8,20). The respective roles of the c-Abl 3 SEK1 3 SAPK and the Lyn 3 MKK7 3 SAPK pathways in DNA damage response may vary in different cell types or under different growth conditions. Nonetheless, the finding that inhibition of PKC␦ attenuates SAPK activation in response to araC and IR indicates that PKC␦ affects one or both of the pathways.
Evidence for a Lyn 3 PKC␦ 3 MEKK1 3 MKK7 3 SAPK Cascade-c-Abl interacts directly with MEKK1 and activates MEKK1 in response to DNA damage (21). In turn, MEKK1 transduces c-Abl-mediated signals to SEK1 and SAPK (8, 21) (Fig. 7). The finding that kinase-inactive MEKK1(K-M) blocks Lyn-mediated activation of SAPK provided support for the involvement of MEKK1 in both the c-Abl and Lyn pathways (Fig. 7). However, unlike studies with c-Abl (21), the available evidence failed to support a direct interaction between Lyn and MEKK1 (20). The present findings demonstrate that Lyn associates with PKC␦ and that tyrosine phosphorylation and activation of PKC␦ is mediated in part by a Lyn-dependent mechanism. The results further demonstrate that PKC␦ associates with MEKK1 and that the interaction between PKC␦ and MEKK1 is increased in response to DNA damage. Moreover, phosphorylation of MEKK1 by PKC␦ stimulated MEKK1 activity. These findings thus collectively support a model in which PKC␦ functions downstream to Lyn and as an upstream effector of the MEKK1 3 MKK7 3 SAPK pathway (Fig. 7).
SAPK is activated in the response of cells to environmental stress (40). MKK7 is a specific activator of SAPK, whereas SEK1 (also known as MKK4) activates both SAPK and p38 MAPK. In turn and depending on the type of stress, upstream effectors of MKK7 and SEK1 include MEKK1-4 and the mixed-lineage protein kinases, ASK1-2, TAK1, and TPL2 (40). The available evidence indicates that MEKK1 is activated by the Rho family GTPases (41), TRAF family members (42), the ECSIT adapter protein (43), protein kinase G (44), and c-Abl (21). The results of the present studies provide support for PKC␦ as yet another effector of MEKK1 activation. PKC␦ is activated in response to growth factor receptor stimulation (45,46), DNA-damage (23), and oxidative stress (47). PKC␦ is also activated by caspase-3-mediated cleavage in the apoptotic response. Thus, although the present work has focused on involvement of PKC␦ in DNA damage-induced signaling, the findings do not exclude the possibility that PKC␦ or PKC␦CF contributes to SAPK activation in response to other types of stress.