Regulated Binding of Importin-α to Protein Kinase Cδ in Response to Apoptotic Signals Facilitates Nuclear Import

Background: Tyrosine phosphorylation regulates nuclear translocation of proapoptotic protein kinase C delta (PKCδ). Results: Tyrosine phosphorylation causes a conformational change that exposes the nuclear localization sequence, allowing binding of importin-α. Conclusion: Nuclear localization of PKCδ is regulated by access of importin-α to the nuclear localization sequence. Significance: Nuclear import of PKCδ, which induces apoptosis, is tightly regulated so as to prevent inappropriate cell death. PKCδ translocates into the nucleus in response to apoptotic agents and functions as a potent cell death signal. Cytoplasmic retention of PKCδ and its transport into the nucleus are essential for cell homeostasis, but how these processes are regulated is poorly understood. We show that PKCδ resides in the cytoplasm in a conformation that precludes binding of importin-α. A structural model of PKCδ in the inactive state suggests that the nuclear localization sequence (NLS) is prevented from binding to importin-α through intramolecular contacts between the C2 and catalytic domains. We have previously shown that PKCδ is phosphorylated on specific tyrosine residues in response to apoptotic agents. Here, we show that phosphorylation of PKCδ at Tyr-64 and Tyr-155 results in a conformational change that allows exposure of the NLS and binding of importin-α. In addition, Hsp90 binds to PKCδ with similar kinetics as importin-α and is required for the interaction of importin-α with the NLS. Finally, we elucidate a role for a conserved PPxxP motif, which overlaps the NLS, in nuclear exclusion of PKCδ. Mutagenesis of the conserved prolines to alanines enhanced importin-α binding to PKCδ and induced its nuclear import in resting cells. Thus, the PPxxP motif is important for maintaining a conformation that facilitates cytosplasmic retention of PKCδ. Taken together, this study establishes a novel mechanism that retains PKCδ in the cytoplasm of resting cells and regulates its nuclear import in response to apoptotic stimuli.

PKC␦ translocates into the nucleus in response to apoptotic agents and functions as a potent cell death signal. Cytoplasmic retention of PKC␦ and its transport into the nucleus are essential for cell homeostasis, but how these processes are regulated is poorly understood. We show that PKC␦ resides in the cytoplasm in a conformation that precludes binding of importin-␣. A structural model of PKC␦ in the inactive state suggests that the nuclear localization sequence (NLS) is prevented from binding to importin-␣ through intramolecular contacts between the C2 and catalytic domains. We have previously shown that PKC␦ is phosphorylated on specific tyrosine residues in response to apoptotic agents. Here, we show that phosphorylation of PKC␦ at Tyr-64 and Tyr-155 results in a conformational change that allows exposure of the NLS and binding of importin-␣. In addition, Hsp90 binds to PKC␦ with similar kinetics as importin-␣ and is required for the interaction of importin-␣ with the NLS. Finally, we elucidate a role for a conserved PPxxP motif, which overlaps the NLS, in nuclear exclusion of PKC␦. Mutagenesis of the conserved prolines to alanines enhanced importin-␣ binding to PKC␦ and induced its nuclear import in resting cells. Thus, the PPxxP motif is important for maintaining a conformation that facilitates cytosplasmic retention of PKC␦. Taken together, this study establishes a novel mechanism that retains PKC␦ in the cytoplasm of resting cells and regulates its nuclear import in response to apoptotic stimuli.
PKC␦ is a ubiquitously expressed member of the PKC family of lipid-regulated kinases that has been implicated in diverse biological functions, including proliferation, differentiation, migration, immunity, and apoptosis (1)(2)(3)(4)(5)(6)(7)(8)(9)(10). Mice deficient for PKC␦ (␦KO) develop normally, but with age develop systemic autoimmune disease associated with hyperproliferation of B cells, suggesting that PKC␦ is important for the establishment of B-cell tolerance (2,4). Work from our lab and others has defined PKC␦ as a critical regulator of the intrinsic apoptotic pathway both in vivo and in vitro (11)(12)(13)(14)(15)(16). In addition, studies in ␦KO mice show that loss of PKC␦ protects against ␥-irradiation-induced apoptosis, suggesting that PKC␦ is required for proper induction of apoptosis in epithelial cells (9). Furthermore, we have recently shown that PKC␦ promotes proliferation and functions as a tumor promoter in lung cancer. The tumor promoter activity of PKC␦ appears to be largely due to its ability to regulate survival signaling pathways (17).
The ability of PKC␦ to regulate such diverse cellular functions as apoptosis and proliferation is dictated in part by tight regulation of its subcellular localization (18 -20). Inappropriate targeting of PKC␦ is associated with tumor progression in bladder and endometrial cancer and with the development of autoimmune disease in mouse models (21,22). We have previously shown that PKC␦ translocates to the nucleus in response to apoptotic signals, and that nuclear accumulation of PKC␦ is necessary and sufficient for induction of apoptosis (23,24). This suggests that conversely, the cytoplasmic retention of PKC␦ may be essential for cell survival. Proteins that are Ͼ40 kDa must be actively transported through the nuclear pore complex. Most protein transport through the nuclear pore complex is facilitated through binding of importins to a nuclear localization sequence (NLS) 3 found on cargo proteins (25). Our lab has previously identified a bipartite NLS in the catalytic domain of PKC␦, and we have shown that mutations of specific residues in this NLS abolish the nuclear localization of PKC␦ in response to apoptotic signals (23). However, in resting cells, PKC␦ is predominantly localized to the cytoplasm, suggesting that additional regulatory steps may be involved in mediating nuclear import of this kinase. Notably, tyrosine phosphorylation at specific residues in the regulatory domain of PKC␦, and caspase cleavage of PKC␦ in the hinge region are permissive for nuclear import, suggesting that the regulatory domain of PKC␦ plays a role in its cytoplasmic retention (9,23,24). How these events are coordinated to facilitate nuclear import of PKC␦ in response to apoptotic signals is not known. In the current studies, we show that translocation of PKC␦ from the cytoplasm into the nucleus is regulated by access of importin-␣ to the NLS. Our studies indicate that nuclear translocation of PKC␦ involves a series of specific and coordinated events, thus assuring tight control of the apoptotic response.
Immunofluorescent Microscopy-ParC5 cells were grown on glass coverslips and transfected with the indicated DNA constructs. 18 h following transfection, cells were either left untreated or were treated with 5 mM H 2 O 2 (Sigma, H1009) for 1 h. Immediately following treatment, cells were rinsed once with PBS and were air-dried. Cells were fixed with 2% paraformaldehyde for 15 min followed by three 15-min washes with PBS. Coverslips containing fixed cells were mounted on slides using Vectashield with DAPI mounting medium (Vector Laboratories, H-1200). Subcellular localization of GFP-PKC␦ proteins was analyzed by fluorescent microscopy. Nuclear localization was quantified as the number of cells with predominantly nuclear localized GFP over the total number of GFP positive cells per field. More than 250 cells were counted for each variable per experiment.
Subcellular Fractionation-293T cells were transfected with the indicated DNA constructs. 18 h following transfection cells were harvested, and whole cell lysate (WCL) or nuclear fractions were prepared. Nuclear fractions were isolated using a nuclear/cytosol fractionation kit (BioVision, Inc., K266-100) according to the manufacturer's protocol except that Triton X-100 was added to the nuclear extraction buffer at a final concentration of 1%. Protein concentration was determined using the DC Protein Assay Kit (Bio-Rad, 500-0111). Nuclear fractions and WCL were resolved on SDS-PAGE, and Western blots were probed with the following antibodies: anti-GFP (Invitrogen, 33-2600); anti-human PKC␦ C-20 (Santa Cruz Biotechnology, sc-937); and anti-lamin B (Santa Cruz Biotechnology, sc-6217).
PKC Kinase Assay-ParC5 or COS-7 cells were transfected with the indicated constructs. 18 h following transfection cells were either left untreated or were treated with 5 mM H 2 O 2 for 30 min. Cells were lysed with JNK lysis buffer (25 mM HEPES, pH 7.7, 20 mM ␤-glycerophosphate, 0.1% Triton X-100, 0.3 M NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA, 0.5 mM DTT, 1ϫ Complete Protease Inhibitor (Roche Applied Science, 11697498001), and 1ϫ phosphatase Inhibitor Mixture (Roche Applied Science, 04906837001)). The GFP-tagged proteins were immunoprecipitated as described above, and the immunocomplexes were washed at least three times with JNK lysis buffer. Kinase assays were performed using the PepTag® Non-Radioactive Protein Kinase C assay system (Promega Corp., V5330), following the manufacturer's protocol, except no lipid activator was added. Kinase activity was calculated as the phosphorylated/non-phosphorylated substrate ratio, normalized to the amount of GFP-tagged protein that was immunoprecipitated, and expressed as relative units.
TUNEL Assay-Primary mouse parotid cells were transduced with the indicated adenoviruses as described previously (27). Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling was performed using the In Situ Cell Death Detection TMR Red kit (Roche Applied Science, 12156792910) according to the manufacturer's protocol. Cells were stained with 5 g/ml DAPI (Sigma, D9564) to visualize the nucleus. Cells expressing the GFP fusion proteins were visualized by immunofluorescent microscopy and counted using a 40ϫ objective. For cells expressing GFP-PKC␦ proteins, GFP-positive, TUNEL-positive (TMR Red fluor) cells exhibiting chromatin condensation were scored as apoptotic; Ͼ 400 cells were counted for each variable per experiment. For cells not expressing GFP-PKC␦ proteins, TUNEL-positive (TMR Red fluor) cells exhibiting chromatin condensation were scored as apoptotic.

RESULTS
Regulated Binding of Importin-␣ to PKC␦ in Response to Apoptotic Agents-Work from our lab and others have shown that PKC␦ translocates from the cytoplasm to the nucleus in response to apoptotic stimuli such as etoposide (23,24). In the experiment shown in Fig. 1A, ParC5 cells were transfected with pGFP-WT-PKC␦ and treated with H 2 O 2 for the indicated times. H 2 O 2 induced nuclear import of PKC␦ as early as 5 min following treatment, with maximal accumulation in the nucleus at 15 min. Endogenous PKC␦ translocated into the nucleus with identical kinetics as that of the ectopically expressed protein (Fig. 1A). PKC␦ was imported into the nucleus with similar kinetics in 293T cells transfected with pGFP-WT-PKC␦ and treated with H 2 O 2 (Fig. 1C). Transport through the nuclear pore complex is mediated by a family of nuclear transport receptors that recognize and bind to the NLS on cargo proteins, the most common of which are members of the importin family (25, 28 -32). In resting cells, PKC␦ could be retained in the cytoplasm bound to importin-␣ awaiting an apoptotic signal, or the interaction of PKC␦ with importin-␣ could be induced by apoptotic signals. To address these possibilities, we analyzed the association of PKC␦ and importin-␣ by co-immunoprecipitation in cells treated with H 2 O 2 or etoposide. As shown in Fig. 1D, a small amount of PKC␦ associates with importin-␣ in untreated cells, consistent with the basal level of PKC␦ typically found in the nucleus under resting conditions (Fig. 1A). However, the association of importin-␣ with PKC␦ dramatically increased as early as 5 min after treatment with H 2 O 2 , similar to the kinetics of nuclear import of PKC␦ seen in Fig. 1A. Regulated association of impor- Upper panels, GFP-tagged proteins were immunoprecipitated from WCL using an anti-GFP antibody and assayed by Western blot for endogenous importin-␣ interaction or for the amount of GFP-tagged protein immunoprecipitated (IP). Lower panels, Western blot of WCL to show the amount of importin-␣ present in the lysates. For all panels, each experiment was repeated three or more times; a representative experiment is shown.
tin-␣ with PKC␦ also occurs in response to etoposide and was maximal at ϳ15 min post-treatment (Fig. 1E). Interaction of importin-␣ with PKC␦ occurred prior to etoposide induced nuclear import of PKC␦, which we have previously reported to be detectable at ϳ30 min following treatment with etoposide (24). As the association kinetics of PKC␦ and importin-␣ in response to both agents closely parallels the nuclear import kinetics of PKC␦, this suggests that importin-␣ binding is ratelimiting for nuclear import.
Tyrosine Phosphorylation of PKC␦ Facilitates Binding to Importin-␣-The observation that H 2 O 2 and etoposide induce binding of PKC␦ to importin-␣ suggests that regulated exposure of the NLS controls nuclear import. Our laboratory and others have shown that tyrosine phosphorylation of PKC␦ in response to various stimuli regulates its subcellular localization as well as other aspects of its function (9,27,(33)(34)(35)(36)(37)(38)(39)(40). To explore a potential role for tyrosine phosphorylation in regulating binding of importin-␣ to PKC␦ in response to H 2 O 2 , we first examined the kinetics of tyrosine phosphorylation of PKC␦ in response to H 2 O 2 using a pan phospho-tyrosine antibody (4G10). As shown in Fig. 2A, upper panels, total tyrosine phosphorylation of PKC␦ is detected as early as 5 min post-treatment, following similar kinetics as that of importin-␣ binding to PKC␦ (Fig. 1D).
We have previously shown that phosphorylation at Tyr-64 and Tyr-155 in the regulatory domain of PKC␦ is required for its nuclear import, and in turn apoptosis, in response to etoposide treatment (27). In Fig. 2A, lower panels, we show that the kinetics of phosphorylation of PKC␦ at these residues parallels total tyrosine phosphorylation as detected with the pan-phosphotyrosine antibody, 4G10 ( Fig. 2A, upper panels). To confirm that Tyr-64 and Tyr-155 are the major sites for tyrosine phosphorylation in response to H 2 O 2 , we transfected ParC5 cells with pGFP-WT-PKC␦ (WT), or mutants of PKC␦ in which the tyrosine residues at amino acid 64 and 155 were mutated to phenylalanine: Y64F, Y155F, and Y64F/Y155F. As predicted, H 2 O 2 treatment for 10 min resulted in an increase in tyrosine phosphorylation of the wild type PKC␦ (Fig. 2B). However, the PKC␦ Y64F and Y155F mutants exhibited a dramatic reduction in tyrosine phosphorylation that was nearly undetectable in the Y64F, Y155F, and Y64F/Y155F mutants (Fig. 2B), confirming that these residues are the major phosphorylation tyrosine sites in response to H 2 O 2 . In addition, we show that these residues are required for H 2 O 2 -induced apoptosis, as PKC␦ mutated at Y64F, and the Y64F/Y155F mutants failed to reconstitute H 2 O 2 induced apoptosis in primary salivary cells isolated from ␦KO mice (Fig. 2C).
Our previous finding that a Y64D/Y155D phosphomimetic mutant of PKC␦ accumulates in the nucleus in the absence of an apoptotic signal, suggests that tyrosine phosphorylation is sufficient to relieve the constraint on importin-␣ binding to the PKC␦ NLS (27). To determine whether phosphorylation at Tyr-64 and Tyr-155 is sufficient for importin-␣ binding, 293T cells were transfected with pGFP-WT-PKC␦ (WT) or pGFP-Y64D/Y155D-PKC␦, and binding of PKC␦ to importin-␣ was assessed by co-immunoprecipitation. As predicted, the GFP-Y64D/Y155D-PKC␦ mutant had an increased ability to bind importin-␣ as compared with wild type PKC␦ (Fig. 2D), suggesting that tyrosine phosphorylation in the regulatory domain of PKC␦ is sufficient for importin-␣ binding.
To explore the mechanism by which tyrosine phosphorylation regulates the interaction of PKC␦ with importin-␣, we examined a model of PKC␦ in the inactive or "closed" conformation (Fig. 3A). This model is based on previous structural studies of inactive PKC␦ (41). In reconstructing this model, we used the crystal structure of the C2-domain of PKC␦ (41) and the catalytic domain of PKC␤II (42), which shares substantial structural homology with PKC␦. In the closed conformation, the C2-domain is folded back upon the catalytic domain and is predicted to make extensive contacts with the region that contains the NLS. This model predicts that contacts between the NLS and the C2-domain limit access of importin-␣ to the NLS. Therefore, to allow for importin-␣ binding to the NLS, PKC␦ must first undergo a conformational change to expose the NLS. We predict that the addition of a negative charge imparted by phosphorylation at Tyr-64 and Tyr-155 facilitates a conformational change in the kinase such that the contacts between the C2 and the catalytic domains are disrupted, leading to exposure of the NLS. PKC activity in vitro in the absence of exogenous lipid is indicative of a conformation of the kinase in which the regulatory domain constraints on the catalytic domain are relieved, resulting in an active and "open" conformation of the kinase (37,(43)(44)(45)(46). To determine whether H 2 O 2 induced tyrosine phosphorylation of PKC␦ results in an active conformation of PKC␦, we assayed the lipid-independent activity of PKC␦ in cells treated with H 2 O 2 . As seen in Fig. 3B, PKC␦ immunoprecipitated from H 2 O 2 treated cells exhibited a significant increase in lipid-independent activity compared with PKC␦ immunoprecipitated from untreated cells. In addition, we confirm that GFP-WT-PKC␦ indeed underwent tyrosine phosphorylation in response to H 2 O 2 treatment as detected with the pan-phosphotyrosine antibody, 4G10 (Fig. 3B, Western blots).
To determine whether phosphorylation at Tyr-64 and Tyr-155 is sufficient to induce lipid independent activity of PKC␦, we assayed the lipid-independent activity of phosphomimetic mutants of PKC␦ at these sites. As seen in Fig. 3C, in the absence of added lipids, the activity of the Y64D and Y155D mutants is increased 2.5-to 3-fold compared with WT. No additional increase in the lipid independent activity was observed with the Y64D/Y155D mutant (data not shown). Taken together, these studies suggest that tyrosine phosphorylation of PKC␦ in the regulatory domain, downstream of H 2 O 2 PPxxP Motif Facilitates Retention of PKC␦ in the Cytoplasm-Our studies indicate that intramolecular contacts between the C2 and the catalytic domains regulate access of importin-␣ to the PKC␦ NLS and, in turn, nuclear import of PKC␦. Curiously, the region of the catalytic domain where the C2-domain makes extensive contacts corresponds to a highly conserved PPxxP motif at the C-terminal tail of PKC␦ (Fig. 4A). In PKC␦, this motif overlaps the bipartite NLS, suggesting that it may play a role in regulating access of importin-␣ to the NLS. Based on its unique location in the NLS and its proposed role in intramolecular interactions, we hypothesize that the PPxxP motif may contribute to the conformational constraints that limit importin-␣ binding to PKC␦ in resting cells (47). To test this directly, we mutated the conserved proline residues to alanine and assessed accumulation of the GFP-tagged proteins in nuclear fractions of transfected cells. As shown in Fig. 4B, very little GFP-WT-PKC␦ is found in the nucleus of unstimulated cells; however, upon stimulation with H 2 O 2 , GFP-WT-PKC␦ accu-mulation in the nucleus increases substantially. In contrast, both GFP-PAxxA-PKC␦ and GFP-AAxxA-PKC␦ accumulated in the nucleus in resting cells, and H 2 O 2 treatment resulted in little or no increase in the nuclear targeting of these mutants compared with the wild type protein, suggesting that the integrity of the PPxxP motif is required for the cytoplasmic retention of PKC␦. Nuclear targeting of the PPxxP mutant proteins was confirmed by immunofluorescence microscopy in ParC5 cells transfected with GFP-tagged proteins (Fig. 4C).
To determine whether nuclear accumulation of the PPxxP mutants results from increased binding to importin-␣, GFPtagged PPxxP mutants of PKC␦ were immunoprecipitated from transfected cells and probed for importin-␣ binding. Fig. 4D shows that the increased nuclear accumulation of the PKC␦ PPxxP mutants correlates with increased importin-␣ binding in resting cells compared with the wild type protein, and H 2 O 2 treatment resulted in little or no increase in importin-␣ binding to these mutants. To explore whether constraints imposed by the PPxxP motif are dependent on the conformation of the full-length kinase, we analyzed importin-␣ binding to a 13-a- D and E, upper panels, the GFP-tagged proteins were immunoprecipitated from WCL using an anti-GFP antibody and assayed by Western blot for endogenous importin-␣ interaction and the amount of GFP-tagged proteins immunoprecipitated (IP). Lower panels, WCL was assayed by Western blot to show the amount of importin-␣ present in the lysate. F, 293T cells were transfected with pGFP-WT-PKC␦ (WT), pGFP-PAxxA-PKC␦ (PAxxA), or pGFP-AAxxA-PKC␦ (AAxxA). The GFP-tagged proteins were immunoprecipitated from WCL using an anti-GFP antibody and assayed by Western blot for total PKC␦ tyrosine phosphorylation (4G10) and the amount of GFP-tagged immunoprecipitated. The graph represents densitometric analysis of the above Western blots quantifying the amount of PKC␦ total tyrosine phosphorylation relative to protein levels. For all panels, each experiment was repeated three or more times; a representative experiment is shown. mino acid peptide of the PKC␦ NLS that is either wild type or mutated at one or more prolines in the PPxxP tagged with two GFP proteins in tandem (GFP2). As shown in Fig. 4E, mutation of single or multiple proline residues in the context of the NLS peptide results in little or no change in importin-␣ binding, suggesting that the constraints imposed by this region are dependent on its participation in intramolecular contacts with other regions of the kinase. Notably, mutation of the PPxxP motif also results in a 20-fold increase in tyrosine phosphorylation of PKC␦ in the absence of apoptotic stimuli (Fig. 4F), presumably due to exposure of tyrosine residues for phosphorylation. We conclude that the PPxxP motif per se does not regulate importin-␣ binding, but rather contacts between this motif and the C2-domain may be important for stabilizing a conformational state of PKC␦ in which binding of importin-␣ to the NLS is inhibited.
Hsp90 Binding Is Required for PKC␦ Binding to Importin-␣-Recent work by Gould et al. (47) has implicated regions C-and N-terminal of the PPxxP motif in mediating binding of Hsp90 to PKC␤II. To explore a potential role for Hsp90 in regulating nuclear import of PKC␦, we assessed whether binding of Hsp90 to PKC␦ is regulated in response to H 2 O 2 treatment. H 2 O 2 treatment induced binding of Hsp90 to PKC␦ with the same kinetics as importin-␣ binding, suggesting that Hsp90 may facilitate binding of importin-␣ to PKC␦ (Fig. 5A). To address this directly, we treated cells with Celasterol or 17-AAG, both of which inhibit Hsp90 binding to its client proteins (48 -50).
Both Celastrol and 17-AAG disrupted Hsp90 binding to PKC␦ as well as the interaction of importin-␣ with PKC␦ (Fig. 5B). As determined by densitometry, treatment with 17-AAG, Celastrol, or both agents resulted in loss of a similar amount of both Hsp90 and importin-␣ binding to PKC␦ (Fig. 5B). Furthermore, the proline-to-alanine mutation of the PPxxP, which enhances the association with importin-␣ and nuclear import of PKC␦ (Fig. 4D), also results in increased interaction of PKC␦ with Hsp90 (Fig. 5C). Taken together, these studies suggest that Hsp90 interaction with PKC␦ may be important for stabilizing a conformation that facilitates exposure of the NLS, thus allowing its interaction with importin-␣.

DISCUSSION
We have previously demonstrated that nuclear localization of PKC␦ in response to apoptotic stimuli is both required and sufficient for its proapoptotic function (23,24). Here, we explored the mechanism that retains PKC␦ in the cytoplasm under resting conditions and regulates its nuclear import in response to apoptotic stimuli. We show that in resting cells, PKC␦ is kept in a conformation that prevents binding of its NLS to importin-␣. However, upon stimulation with H 2 O 2 or etoposide, PKC␦ binds to importin-␣ with kinetics similar to that of the nuclear import of PKC␦. Our studies indicate that tyrosine phosphorylation of PKC facilitates binding of importin-␣ to PKC␦, and that binding of importin-␣ to the NLS is stabilized by Hsp90 (see model in Fig. 6). The tight regulation of nuclear A-C, upper panels, the GFP-tagged proteins were immunoprecipitated from WCL using an anti-GFP antibody, and the immunocomplexes were assayed by Western blot for endogenous importin-␣ and Hsp90 interaction, and the amount of GFP-tagged protein immunoprecipitated (IP). Lower panels, WCL was assayed by Western blot to show the amount of importin-␣ and Hsp90 present in the lysates. B, the graph represents densitometric analysis of the above Western blots quantifying the amount of Hsp90 (black bars) and the amount of importin-␣ (gray bars) bound to PKC␦ relative to the amount of GFP-tagged protein immunoprecipitated. For all panels, each experiment was repeated three or more times; a representative experiment is shown.
import assures cytoplasmic retention of PKC␦ under basal conditions and nuclear import and induction of apoptosis only when appropriate.
Maintaining a balance between nuclear and cytoplasmic levels of a given protein can be crucial for determining cell fate; therefore nucleocytoplasmic shuttling is a highly orchestrated process and is often coupled with other levels of regulation such as protein phosphorylation. In most cases, phosphorylation enhances nuclear import of cargo proteins, although abrogation of nuclear import has been reported (51). Currently, only a few studies have described the molecular mechanism by which phosphorylation either enhances or abrogates nuclear import. STAT1 (signal transducers and activators of transcription) phosphorylation induces its homo-dimerization, forming a dimer-specific NLS that is recognized by importin-␣ (52). Phosphorylation of 14-3-3 leads to its dissociation from c-Abl, exposing the NLS of the latter and inducing its nuclear import (53). Here, we report that in the context of PKC␦, phosphorylation regulates exposure of a hidden NLS through remodeling of intramolecular interactions.
We show that in addition to etoposide, phosphorylation at Tyr-64 and Tyr-155 is also required for H 2 O 2 -induced apoptosis (27). Our studies suggest that Tyr-64 and Tyr-155 are the primary sites for tyrosine phosphorylation of PKC␦ in H 2 O 2treated ParC5 cells. In contrast, Konishi et al. (37,38), using COS-7 cells, have shown that PKC␦ is phosphorylated on additional residues in response to H 2 O 2 , most of which are in the catalytic domain. To show that tyrosine phosphorylation at Tyr-64 and Tyr-155 is functionally significant, we have mimicked the negative charge imparted on these residues by phosphorylation and have shown that the resulting mutant (Y64D/Y155D) has enhanced binding to importin-␣. This suggests that phosphorylation of PKC␦ at Tyr-64 and Tyr-155 may regulate access of importin-␣ to PKC␦. Consistent with a conformational change induced by tyrosine phosphorylation, we show that H 2 O 2 induces lipid-independent activity of PKC␦ and that the Y64D and Y155D phosphomimetic mutants of PKC␦ have a significantly higher lipid independent activity compared with the wild type protein. Furthermore, a model of the closed conformation of PKC␦, suggests that the C2-domain makes extensive contacts with the region of the catalytic domain containing the NLS, sterically hindering access of importin-␣ to the NLS. Likewise, in the crystal structure of full-length PKC␤II, the C2-domain is also positioned so that it partially occludes access to the same region of the protein that contains the NLS in PKC␦, suggesting that these structural contacts are likely conserved, at least in the novel and conventional family of PKCs (supplemental Fig. 1).
Next, we explored the role of the highly conserved PPxxP motif that overlaps the NLS of PKC␦. Gould et al. (47) has recently shown that this motif is required for proper priming site phosphorylation of PKC␤II, which we confirm is also true for PKC␦ (data not shown). We showed that proline-to-alanine mutagenesis of the PPxxP motif results in enhanced binding to importin-␣ and nuclear import of PKC␦. This suggests that the integrity of this motif is required for the cytoplasmic retention of PKC␦ by preventing access of the NLS to importin-␣. However, a 13-amino acid peptide of the region containing the PPxxP revealed no difference between the proline-to-alanine mutants and wild type PPxxP motif in terms of binding to importin-␣. This suggests that the role of the PPxxP motif in limiting access of importin-␣ to the NLS is unique to the fulllength protein. This is consistent with our earlier work showing that the catalytic fragment of PKC␦ translocates into the nucleus without apoptotic stimuli (23), further suggesting that the molecular contacts that confer cytoplasmic retention of PKC␦ require the C2-domain.
H 2 O 2 induced association of Hsp90 with PKC␦ with kinetics similar to that of importin-␣. Furthermore, we show that Hsp90 binding is required for the association of PKC␦ with importin-␣. In addition, we show that proline-to-alanine mutagenesis of the PPxxP motif enhanced the association of Hsp90 with PKC␦. This suggests that both H 2 O 2 treatment and PPxxP mutagenesis facilitate the availability of Hsp90-interacting regions on PKC␦. These observations are in agreement with the well established role of Hsp90 in binding to and stabilizing hydrophobic regions of proteins that become exposed in the open active conformation (54,55). In the study by Gould et al. (47) the sites for Hsp90 binding were mapped to regions C and N terminus to the PPxxP motif, but not the PPxxP motif itself. In PKC␦, the PPxxP overlaps the NLS, the site for importin-␣ binding. This suggests that Hsp90 binding is likely involved in stabilizing a conformation that favors importin-␣ binding to the NLS of PKC␦, rather than directly mediating this interaction. This is consistent with the role of Hsp90 in unmasking the NLS of the dioxin receptor, allowing importin-␣ binding and nuclear import of the dioxin receptor upon ligand binding (56).
In summary, our studies elucidate a novel mechanism for regulated nuclear import of PKC␦ as shown in the model in Fig.  6. We describe the steps and structural components involved in retaining PKC␦ in the cytosol of resting cells and in regulating its nuclear import in response to apoptotic agents. These findings provide insight for the design of rational therapeutic strategies aimed at manipulating the proapoptotic function of PKC␦ by control of its subcellular localization.