Regulated nucleocytoplasmic transport of protein kinase D in response to G protein-coupled receptor activation.

Protein kinase D (PKD)/protein kinase C mu is a serine/threonine protein kinase activated by growth factors, antigen-receptor engagement, and G protein-coupled receptor (GPCR) agonists via a phosphorylation-dependent mechanism that requires protein kinase C (PKC) activity. In order to investigate the dynamic mechanisms associated with GPCR signaling, the intracellular distribution of PKD was analyzed in live cells by imaging fluorescent protein-tagged PKD and in fixed cells by immunocytochemistry. We found that PKD shuttled between the cytoplasm and the nucleus in both fibroblasts and epithelial cells. Cell stimulation with mitogenic GPCR agonists that activate PKD induced a transient nuclear accumulation of PKD that was prevented by inhibiting PKC activity. The nuclear import of PKD requires its cys2 domain in conjunction with a nuclear import receptor, while its nuclear export requires its pleckstrin homology domain and a competent Crm1-dependent nuclear export pathway. This study thus characterizes the regulated nuclear transport of a signaling molecule in response to mitogenic GPCR agonists and positions PKD as a serine kinase whose kinase activity and intracellular localization is coordinated by PKC.

Protein kinase D (PKD) 1 /protein kinase C is a serine/threonine protein kinase with structural, enzymological, and regulatory properties different from other protein kinase C (PKC) family members (1,2). The salient features of PKD structure include the presence of a catalytic domain distantly related to Ca 2ϩ -regulated kinases, a pleckstrin homology (PH) domain that regulates PKD activity, and a highly hydrophobic stretch of amino acids in its N-terminal region (1)(2)(3)(4). The N-terminal region of PKD contains, in addition to the PH domain, a cysteine-rich domain (CRD) that confers high affinity binding to phorbol esters (5)(6)(7). The recent identification of additional cDNA clones, similar in overall structure, primary amino acid sequence, and enzymological properties to PKD (8,9), supports the notion that PKD isoenzymes constitute a separate family of serine protein kinases.
PKD can be activated in intact cells by pharmacological agents including biologically active phorbol esters and cellpermeant diacylglycerol (DAG) as well as by physiological stimuli including G protein-coupled receptors (GPCR) agonists, growth factors, and antigen-receptor engagement (10 -18). In all cases, PKD activation has been shown to be mediated by a PKC-dependent signal transduction pathway that involves the phosphorylation of Ser 744 and Ser 748 within the activation loop of the catalytic domain of PKD (19,20). These findings revealed a link between PKC and PKD in a novel signal transduction pathway activated by multiple growth-promoting factors (5,21).
PKD has been localized in the cytosol and in several intracellular compartments including Golgi, plasma membrane, and mitochondria (14,(22)(23)(24)(25). In addition, we recently found that bombesin, a mitogenic GPCR agonist, induced a rapid and reversible plasma membrane translocation of PKD by a mechanism that requires its catalytic domain and PKC activity (26). PKD has been implicated in the regulation of a variety of cellular functions including Golgi organization and function (27,28), epidermal growth factor receptor and c-Jun signaling (29,30), NF-B-mediated gene expression (31), cell migration (32), and DNA synthesis and cell proliferation (33). The role of PKD redistribution in the regulation of these cellular and molecular responses remains poorly understood.
In this study, we examined the distribution of PKD using immunocytochemistry and visualization of fluorescence-tagged PKD in dividing and G 0 -arrested (quiescent) Swiss 3T3 cells. We found that PKD shuttles between the cytoplasm and the nucleus in proliferating cells. Mitogenic GPCR agonist stimulation of quiescent Swiss 3T3 cells induced a transient nuclear accumulation of PKD that could be blocked by inhibiting PKC activity. These findings identified the nuclei as a compartment targeted by PKD and provides evidence for a novel mechanism where the activation and intracellular distribution of PKD is coordinated by PKC.
Earlier we demonstrated that the fusion of GFP tag to the N terminus of PKD did not produce any detectable effect on its basal catalytic activity, phorbol ester binding, or bombesin-mediated PKD activation (14,20,24,26). In addition, the inherent fluorescence of GFP allowed us to visualize the localization of GFP-PKD in live cells (14).
Cell Culture and Transfections-Stock cultures of Swiss 3T3 and Madin-Darby canine kidney cells (MDCK) were maintained at 37°C in Dulbecco's modified Eagle's medium, supplemented with 10% fetal bovine serum in a humidified atmosphere containing 10% CO 2 and 90% air. For live cell analysis, cells were plated onto 15-mm number 1 round glass coverslips inside 33-mm dishes at 7 ϫ 10 4 cells/dish and transfected 18 -20 h later. Cells were transfected with 1 g of DNA/33-mm dish using LipofectAMINE PLUS (Life Technologies, Inc.) according to the manufacturer's suggested conditions. For immunocytochemistry, the cells were plated in 33-mm dishes at 7 ϫ 10 4 cells/dish or in Lab-Tek chamber slides (Nalge Nunc International, Naperville, IL) (3.5 ϫ 10 4 cells/well) and transfected as indicated above with 1 g of DNA/dish or 0.5 g of DNA/well, respectively. Transfected cells were incubated for 18 -20 h before agonist analysis. Cotransfection experiments were done using 1 g of pGFP-CRD DNA and 100 ng of PKD-RFP DNA per 33-mm dish.
The Swiss 3T3-PKD.GFP cell line was established by infecting Swiss 3T3 cells with a retrovirus encoding PKD and GFP as two separated proteins but under the control of the same promoter (33). Cells were collected and sorted by fluorescence-activated cell sorting to select the GFP-positive ones. The GFP-positive cells were propagated, and multiple aliquots were frozen. A fresh batch of cells was restarted every 2 months. Cells were maintained at 37°C in Dulbecco's modified Eagle's medium, supplemented with 10% fetal bovine serum in a humidified atmosphere containing 10% CO 2 and 90% air.
Cell Imaging-Imaging of live cells expressing fluorescence-tagged proteins and immunocytochemistry of fixed cells using a rabbit polyclonal anti-PKD/PKC (C20), which recognizes an epitope mapping at the C terminus of PKD, were performed as previously described (25,26). For the experiments employing immunocytochemistry or imaging of live cells expressing GFP-or RFP-tagged proteins, 50 cells were ana-lyzed per experiment, and each experiment was performed at least in triplicate. The cells displayed in the appropriate figures were representative of 90% of the population of transfected cells.
Quantitative analysis of the cytoplasmic and nuclear fluorescence intensity was performed on images of the midsection of fixed and immunostained cells that were obtained with a Leica TCS-SP upright laser-scanning confocal microscope (LSCM) (26). Quantification was performed on a Macintosh PowerBook G4 computer using the public domain NIH Image program (developed at the National Institutes of Health and available on the Internet at rsb.info.nih.gov/nih-image/). Nuclear fluorescence intensity (N) was calculated as a percentage of the total cellular fluorescence intensity (C): (N/(N ϩ C)). Each data point represents the mean intensity fluorescence obtained from 20 randomly chosen cells unless otherwise indicated. Error is expressed as S.D.
Materials-The anti-PKD/PKC (clone C20) antibody, its corresponding blocking peptide sc-639P, and anti-GFP antibody were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Ro 81-3220 was obtained from Calbiochem. Bombesin, vasopressin, GF 109203X, and 4Ј,6-diamidino-2-phenylindole were obtained from Sigma. Texas Red-conjugated goat-anti rabbit immunoglobulins were obtained from Molecular Probes, Inc. (Eugene, OR). The plasmid pEGFP-Actin and the anti-RFP antibody were obtained from CLON-TECH Laboratories, Inc. (Palo Alto, CA). Leptomycin B was a generous gift from Dr. Minoru Yoshida (Department of Biotechnology, University of Tokyo). All of the other reagents were the highest grade commercially available.

PKD Shuttles between the Nucleus and the Cytoplasm-We
previously found that the deletion of the PH domain of PKD induced the nuclear accumulation of this GFP-tagged protein in lymphocytes, fibroblasts, and epithelial cells (24,26). In order to substantiate this observation, Swiss 3T3 fibroblasts, a model system extensively used to elucidate signaling by endogenously expressed GPCRs (34), were transiently transfected with wild type and mutant PKDs fused to GFP (Fig. 1) and examined 18 h later with an LSCM.
In agreement with our recent results, we found that GFP-PKD and the majority of the PKD fusion mutant proteins expressed in Swiss 3T3 cells were distributed throughout the cytosol (Fig. 2). Some cells showed a more pronounced signal at the perinuclear area, consistent with the partial localization of PKD to the Golgi compartment, as we and others described previously (23,25,27,28). No fluorescence or very little fluorescence was detected in the nuclei of these cells. However, we found that cells expressing GFP-PKD-⌬PH showed a dramatic nuclear accumulation of this protein (Fig. 2). The same distri-

FIG. 1. Schematic representation of the expressed wild type and mutant PKD fusion proteins and summary of their nuclear transport properties.
Swiss 3T3 cells were transfected with constructs encoding the wild type and mutant PKD proteins and maintained at 37°C for 18 h. The cultures were incubated with LMB (10 ng/ml) for 1 h, and the accumulation of the different fusion proteins within the nuclei was analyzed by epifluorescence microscopy of live cells expressing the fluorescence-tagged proteins. The cells transiently overexpressing a nonfluorescent tagged PKD protein were fixed and stained to detect PKD as described under "Experimental Procedures," using an epifluorescence microscope. The results summarized are representative of at least four independent experiments. The employed plasmids were constructed as described under "Experimental Procedures." bution was detected with another chimeric protein between RFP from Discosoma sp. fused to the C terminus of PKD-⌬PH in Swiss 3T3 cells (PKD-⌬PH-RFP) (data not shown).
The molecular mass of GFP-PKD-⌬PH or PKD-⌬PH-RFP (ϳ125 kDa) far exceeds the size limit for passive nuclear diffusion (60 kDa) (35)(36)(37)(38), suggesting that the deletion of the PH domain of PKD could promote its nuclear localization by unmasking a normally silent nuclear localization signal (NLS). Alternatively, the deletion of the PH domain could prevent the nuclear export of PKD and cause its nuclear accumulation. If this second interpretation is correct, interference with the nuclear export machinery should also cause the nuclear accumulation of PKD.
Leptomycin B (LMB) is an antifungal antibiotic (39) that inhibits the formation of complexes consisting of Crm1, RanGTP, and proteins containing a leucine-rich nuclear export signal (NES), thereby blocking nuclear export (40 -43). In order to test whether LMB had any effect on the cellular distribution of PKD, Swiss 3T3 cells transiently transfected with GFP-PKD were incubated with LMB (10 ng/ml) for 1 h, and the distribution of GFP-PKD was analyzed by imaging live cells with a LSCM. As shown in Fig. 2, GFP-PKD accumulated in the nuclei of Swiss 3T3 cells incubated with LMB. In contrast, LMB had no effect on the nuclear accumulation of GFP-PKD-⌬PH (Fig.  2).
In order to determine whether the nucleocytoplasmic transport of PKD is restricted to Swiss 3T3 fibroblasts, we also analyzed the nuclear transport of GFP-PKD in MDCK epithelial cells transfected with pGFP-PKD. MDCK cells are one of the best studied epithelial cell model systems (44). As shown in Fig. 2, GFP-PKD also accumulated in the nuclei of proliferating MDCK cells incubated with LMB, indicating that the nuclear transport of PKD was not restricted to fibroblasts. GFP-PKD expression or LMB incubation up to 3 h had no effect on the morphology of Swiss 3T3 or MDCK cells (data not shown).
The nuclear accumulation of PKD was not due to the fluorescent tag or protein overexpression as revealed by immunocytochemistry of nontagged PKD. Exogenous nontagged (ex-PKD) expressed by transient transfection and endogenous PKD (enPKD) accumulated in the nuclei of dividing Swiss 3T3 incu-bated with LMB (Fig. 2). The localization of endogenous PKD in the cells treated with LMB corresponded to the nuclear compartment as revealed by 4Ј,6-diamidino-2-phenylindole (DAPI) costaining of those cells (Fig. 2). Inclusion of the immunizing peptide encompassing the C terminus of PKD completely prevented the staining of the endogenous PKD (data not shown).
To rule out any effect of the fluorescent tag on the nuclear transport of PKD, we analyzed the intracellular distribution of GFP-tagged actin in Swiss 3T3 in the presence of LMB. GFPactin was detected in the cytoplasm of either LMB-treated or -untreated cells as a diffuse immunofluorescent signal (G-actin) or associated with microfilaments (F-actin). LMB did not induce any detectable nuclear accumulation of GFP-actin. The results in Fig. 2 demonstrate that PKD shuttles between the cytoplasm and nuclei and that the nuclear export of PKD can be prevented by blocking the Crm1-dependent nuclear export pathway with LMB.
The PH Domain Mediates the Nuclear Export of PKD-In order to identify the domain(s) of PKD responsible for its nuclear import and export, different GFP-tagged PKD mutants were expressed transiently in Swiss 3T3 cells, and the distribution of these molecules was monitored with an epifluorescence microscope in live cells incubated with or without LMB. Nuclear accumulation of any mutated PKD, in the absence of LMB, would indicate a defect in its nuclear export. Conversely, lack of nuclear accumulation of any mutated PKD in the presence of LMB would indicate a defect in its nuclear import.
As showed in Fig. 2, deletion of the complete PH domain of PKD caused the nuclear accumulation of GFP-PKD-⌬PH in the absence of LMB, suggesting that this domain was involved in the nuclear export but not the import of PKD. Further support for this conclusion was obtained by analyzing the intracellular distribution of two different PH domain mutant proteins in Swiss 3T3 cells. The first one, GFP-PKD⌬1-4␤, lacks the four ␤-sheets of the ␤-barrel of the PH domain encompassing amino acids 429 -474, whereas the second one, GFP-PKD⌬␣, lacks the carboxyl-terminal ␣-helix encompassing amino acids 535-557. We found that both mutant proteins accumulated in the nuclei of Swiss 3T3 cells in the absence of LMB (Fig. 3). Since both mutations target different regions of the PH domain and they involve different size deletions (45 and 22 amino acid residues in GFP-PKD⌬1-4␤ and GFP-PKD⌬␣, respectively) our results suggested that the integrity of this domain, rather than a particular signal, was critical for the nuclear export of PKD. We obtained further support for this conclusion by mutating a stretch of amino acids similar to a leucine-rich NES in the PH domain of PKD (amino acids 474 -480). Similar mutations within the NES in other proteins abrogate the NES function (45)(46)(47)(48)(49). The simultaneous mutation to alanine of the leucine residues 478 and 480 within the putative NES in the PH domain of PKD did not prevent the nuclear export of GFP-PKD-L478A/L480A (data not shown), confirming that the integrity of the PH domain is critical for the nuclear export of PKD. Our results also implied that the nuclear export of PKD may occur via interaction, very likely mediated by its PH domain, with an adaptor protein(s) that is exported from the nucleus by a Crm1-dependent nuclear export pathway.
The cys2 Domain Mediates the Nuclear Import of PKD-The N-terminal regulatory region of PKD, in addition to its PH domain, contains a phorbol ester/DAG-binding CRD (1). We identified within the CRD of PKD a region encompassing amino acid residues 184 -201 with homology to known classical bipartite NLS (50) that could be involved in the nuclear import of PKD. Consequently, we analyzed the intracellular distribution of another set of GFP-tagged PKD proteins with mutations in the CRD of PKD (see Fig. 1). The GFP-PKD-⌬CRD mutant contains a deletion of the entire CRD domain (encompassing amino acid residues 145-353), while the GFP-PKD-P287G mutant contains a proline to glycine substitution within the second cysteine-rich motif of the CRD. While the CRD deletion prevented the binding of phorbol esters/DAG to PKD, the P287G mutation significantly diminished it (7). As shown in Fig. 4, GFP-PKD-P287G was imported into the nuclei to the same extent as wild type PKD as revealed by its accumulation in the presence of LMB. Contrary to GFP-PKD-P287G, GFP-PKD-⌬CRD was excluded from the nuclei despite the presence of LMB (Fig. 4), suggesting that the nuclear import of PKD depends on its CRD but very unlikely on phorbol ester/DAG binding.
The CRD of PKD contains a tandem repeat of cysteine-rich zinc finger-like motifs, termed cys1 and cys2, that are not functionally equivalent (7) and that could be differentially involved in the nuclear import of PKD. Consequently, we analyzed the intracellular distribution of two other CRD mutants that were generated by deleting either the complete first (amino acid residues 145-223) or second (amino acid residues 277-353) cysteine-rich zinc finger-like motif of PKD. These deleted PKD mutants were fused to GFP and expressed in Swiss 3T3 cells. We found that the deletion of cys1 did not interfere with the nuclear import of GFP-PKD-⌬cys1, as revealed by its nuclear accumulation in the presence of LMB. Furthermore, this mutation rather induced a slight nuclear accumulation of GFP-PKD-⌬cys1, even in the absence of LMB (Fig. 4). In striking contrast, the deletion of cys2 prevented the nuclear import of GFP-PKD-⌬cys2 in the presence of LMB (Fig. 4). In other experiments, we found that either longer incubation of the cultures with LMB (up to 2 h) or the addition of higher LMB concentrations (up to 30 ng/ml) did not induce the nuclear accumulation of GFP-PKD-⌬cys2 (data not shown).
A Nuclear Transport Receptor Mediates the Nuclear Import of PKD-One of the functional characteristics of nuclear transport receptors is the saturability of signal recognition (51). We used this criterion to further assess the NLS specificity found in PKD by blocking its nuclear import via competition of its nuclear import receptor with the CRD of PKD, since this domain of PKD mediated its nuclear import. A protein consisting of GFP fused to the N terminus of the CRD of PKD (GFP-CRD) was coexpressed with another protein consisting of RFP fused to the C terminus of the full-length PKD (PKD-RFP) in an attempt to block the nuclear import of PKD. GFP-CRD was distributed throughout the cytosol and the nuclei with a distinct accumulation in the plasma membrane and the Golgi compartment (Fig. 5). The addition of LMB to the cultures expressing GFP-CRD did not modify its distribution (Fig. 5). The distribution of PKD-RFP was identical to that described for GFP-PKD under basal conditions. In the absence of LMB, PKD-RFP was distributed throughout the cytosol with very little fluorescence in the cell nuclei, whereas LMB induced its nuclear accumulation (Fig. 5).
The coexpression of GFP-CRD and PKD-RFP (10:1 ratio) did not produce any major changes in their distributions in the absence of LMB, with the exception of a slight accumulation of PKD-RFP in the Golgi compartment that colocalized with GFP-CRD (see superimposed images in Fig. 5). The inclusion of LMB in the cultures coexpressing GFP-CRD and PKD-RFP had no effect on the distribution of GFP-CRD. However, the nuclear import of PKD-RFP was prevented, as revealed by its lack of nuclear accumulation in the presence of LMB (see superimposed images in Fig. 5). These results suggested that GFP-CRD successfully competed with PKD-RFP for the nuclear import receptor in charge of delivering PKD to nucleus, and they support the conclusion that the nuclear entry of PKD is mediated by its CRD in conjunction with a nuclear import receptor.
Coexpression of GFP, which uniformly distributes in the cytoplasm and nuclei of Swiss 3T3 cells (26), did not interfere with the nuclear accumulation of PKD-RFP in the presence of LMB (data not shown). This result further supported the con- clusion that the nuclear import block induced by the CRD domain of PKD was specific and not due to the presence of GFP in the nuclear compartment. A summary of the nuclear import and export properties of the different mutants analyzed is shown in Fig. 1.
Although we cannot rule out the possibility that other domains of PKD might also be involved in its nuclear shuttling, these results demonstrate that the critical regions of PKD mediating its nuclear import and export are the cys2 and PH domains, respectively.
GPCR Activation Induces a Transient Nuclear Accumulation of PKD-Having established that PKD shuttles between the cytoplasm and the nucleus in proliferating Swiss 3T3 and MDCK cells, our next objective was to determine whether a regulated transport of PKD occurred in response to mitogenic GPCR agonists. This was important because regulated transport mechanisms are crucial in modulating signaling information by targeting proteins to different cellular compartments in response to specific cell stimuli (54 -56). Quiescent cultures of Swiss 3T3 cells are an excellent cell system model extensively used to elucidate the signal transduction pathways activated by mitogenic agonists that act through endogenously expressed GPCRs (34,52). Because the tissue culture conditions (medium containing 10% fetal bovine serum) that we used to transiently express the fluorescence-tagged PKDs prevented us from inducing G 0 /G 1 cell cycle arrest in Swiss 3T3 cells, we instead used a cell line termed Swiss 3T3-PKD.GFP. This cell line has been recently established in our laboratory by infecting Swiss 3T3 cells with a retrovirus encoding PKD linked via an internal ribosome entry site to GFP (33). This bicistronic retroviral vector drives the expression of PKD and GFP as two separate proteins under the control of the long terminal repeat of the murine stem cell virus. The PKD protein that is constitutively overexpressed in Swiss 3T3-PKD.GFP cells retains PKC-dependent regulation by multisite phosphorylation, resulting in enhanced kinase catalytic activity like endogenous PKD (33). Overexpression of PKD potentiates DNA synthesis and cell proliferation induced by bombesin and vasopressin (33). Of special interest in the context of the present study, PKD can be readily detected by immunocytochemistry in Swiss 3T3-GFP-.PKD cells even after they became quiescent.
Consequently, we employed quiescent Swiss 3T3-PKD.GFP cells to determine whether bombesin or vasopressin, two mitogenic GPCR agonists that activate PKD by a PKC-dependent signal transduction pathway, induced any changes in the nuclear transport of PKD. Fig. 6A shows that PKD was predominantly present in the cytoplasm and perinuclear region of quiescent Swiss 3T3-PKD.GFP cells, with very little nuclear PKD, as observed by indirect immunofluorescence. The addition of LMB to the quiescent Swiss 3T3-PKD.GFP cells induced, after 30 min of incubation, a nuclear accumulation of FIG. 5. Overexpression of the CRD of PKD blocks the nuclear import of full-length PKD. Swiss 3T3 cells were transfected with either constructs encoding GFP-CRD or PKD-RFP or cotransfected simultaneously with pGFP-CRD (1 g of DNA) and pPKD-RFP (100 ng of DNA) and incubated at 37°C for 18 h. LMB (10 ng/ml) was added to the cultures, and the cells were further incubated at 37°C for 1 h. Epifluorescence microscopy was used to visualize the distribution of the fluorescence-tagged full-length or truncated PKD in live cells. Representative images were captured as described under "Experimental Procedures." Bar, 10 m.

FIG. 6. Bombesin and vasopressin induce a transient redistribution of PKD from the cytoplasm to the nucleus of quiescent cells.
A and B, quiescent Swiss 3T3-GFP.PKD cells were incubated at 37°C with LMB (Ⅺ; 10 ng/ml) for 30 or 90 min and processed immediately for immunocytochemistry. Quiescent Swiss 3T3-GFP.PKD cells incubated at 37°C were stimulated for 10 min with 10 nM bombesin (f; 50 nM) vasopressin (u), or 1 g/ml insulin (o) and processed for immunocytochemistry 30 or 90 min after stimulation. C, quiescent Swiss 3T3-GFP.PKD cells were incubated at 37°C with LMB (Ⅺ; 10 ng/ml) for 10, 20, or 30 min and processed immediately for immunocytochemistry. Quiescent Swiss 3T3-GFP.PKD cells at 37°C incubated with LMB (10 ng/ml) and bombesin (10 nM) (f) simultaneously for 10, 20, or 30 min and processed for immunocytochemistry at the end of those incubation times. PKD detection and quantification by immunocytochemistry and a LSCM was performed as described under "Experimental Procedures." Representative images were captured as described under "Experimental Procedures." Nuclear fluorescence intensity was calculated as a percentage of the total cellular fluorescence intensity (N/(N ϩ C)). Each data point represents the mean intensity fluorescence obtained from 20 (B) or 10 (C) randomly chosen cells. Error is expressed as S.D. *, significantly different from control (p Ͻ 0.0001). Bom, bombesin; Vas, vasopressin; Ins, insulin; Bar, 10 m.
PKD that increased only minimally after 90 min of LMB treatment (Fig. 6A). In contrast, the addition of bombesin to quiescent Swiss 3T3-PKD.GFP cells induced a striking nuclear accumulation of PKD in the absence of LMB (Fig. 6A). A salient feature of this result is that the nuclear accumulation of PKD after 30 min of bombesin stimulation was more prominent than that induced by LMB, even after 90 min (Fig. 6A). The nuclear accumulation of PKD induced by bombesin was transient. The majority of PKD returned to the cytoplasm 90 min after bombesin stimulation (Fig. 6A). PKD activation in response to bombesin stimulation was confirmed in parallel cultures of Swiss 3T3-PKD.GFP cells using an antiserum that specifically recognizes a C-terminal residue (serine 916) when it is autophosphorylated by active PKD (53) (data not shown).
The neuropeptide vasopressin, which acts in Swiss 3T3 cells via an endogenously expressed V1 receptor subtype, also induced a marked and transient nuclear accumulation of PKD in the absence of LMB (Fig. 6A). In contrast, the addition of insulin, which does not stimulate a significant increase in either PKC or PKD activity in 3T3-PKD.GFP cells (33) did not induce any significant nuclear accumulation of PKD in Swiss 3T3-PKD.GFP cells (Fig. 6A).
To achieve a quantitative understanding of nuclear accumulation of PKD in quiescent cells after GPCR activation, we analyzed the fluorescence signal of images obtained using an LSCM. The method we employed allowed us to compile data from several individual cells and then to calculate the average fluorescence intensity in discrete cellular regions (see "Experimental Procedures"). This quantitative analysis showed that although LMB induced the nuclear accumulation of PKD in quiescent cells, this accumulation was considerably less than that induced by bombesin or vasopressin in the same length of time (Fig. 6B). This result suggested that the nuclear transport of PKD could be regulated in response to GPCR agonists that activate PKD. Insulin did not induce any significant change in the intracellular distribution of PKD.
Several proteins that shuttle continuously between the cytoplasm and the nucleus are also subjected to regulated nuclear transport (54 -56). In response to specific cellular signals, a change occurs in the rate of nuclear import and/or export, causing these proteins to become concentrated in the cytosol or in the nucleus. To determine whether mitogenic GPCR agonists induced a change in the rate of nuclear import of PKD, we blocked its nuclear export with LMB and compared its relative nuclear accumulation in nonstimulated versus bombesin-stimulated Swiss 3T3-PKD.GFP cells. We found that bombesin induced a 2-fold increase in the rate of nuclear import of PKD as indicated by its faster nuclear accumulation when compared with nonstimulated cells in the presence of LMB (Fig. 6C).
Inhibition of PKC Kinase Activity Prevents the Nuclear Accumulation of PKD in Response to GPCR Activation-Since PKD has been identified as a downstream target of PKCs (10,11,13,16,(57)(58)(59)(60), we hypothesized that the kinase activity of PKC could play a role in the nuclear accumulation of PKD. To test this hypothesis, we examined whether PKC inhibition interfered with the nuclear transport of PKD in bombesinstimulated Swiss 3T3.GFP.PKD cells. We used two different PKC inhibitors (Ro 31-8220 or GF 109203X) that, as we previously demonstrated, inhibit the phosphorylation of the activation loop in the kinase domain of PKD and its kinase activity (11,33).
Quiescent cultures of Swiss 3T3.GFP.PKD cells were preincubated with either Ro 31-8220 or GF 109203X and then stimulated with bombesin. The distribution of PKD in these cells was analyzed by immunocytochemistry at 30 and 90 min after bombesin stimulation. As illustrated in Fig. 7, the intra-cellular distribution of PKD in quiescent Swiss 3T3.GFP.PKD cells treated with either Ro 31-8220 or GF 109203X was indistinguishable from that in untreated Swiss 3T3.GFP.PKD cells (see Fig. 6). In agreement with our previous findings (26), treatment with the PKC inhibitors delayed the plasma membrane dissociation of PKD in bombesin-stimulated cells. Unexpectedly, when PKD dissociated from the plasma membrane, it did not accumulate in the nuclei of the cells (Fig. 7). These results corroborated our previous demonstration that the rapid plasma membrane dissociation of PKD in bombesin-stimulated cells is regulated by PKC, and, surprisingly, they indicated that the transient nuclear accumulation of PKD is also mediated by PKC.

DISCUSSION
The targeting of signaling molecules to different cellular compartments in response to specific stimuli is a fundamental process in the regulation of their activity (61)(62)(63). In recent years, an increasing number of regulatory protein kinases including PKC, Bruton's tyrosine kinase, Abl, and mitogenactivated protein kinase/extracellular signal-regulated kinase, have been shown to shuttle between the nucleus and the cytoplasm (38, 64 -69). In the present study, we used visualization of fluorescence-tagged PKD chimeras in live cells and immunocytochemistry of exogenous nontagged as well as endogenous PKD in fixed cells to demonstrate that PKD shuttles between the cytoplasm and the nucleus.
The nuclear import of PKD required its cys2 domain and a nuclear import receptor. Classical NLS are characterized by short amino acid stretches that are enriched in basic amino acids that can be arranged as mono-or bipartite signals (37). Generally, proteins carrying classical NLS bind a cytoplasmic receptor, importin ␣, which in turn associates with importin ␤, forming a complex that docks in the nuclear pore from which the cargo is translocated into the cell nucleus (70,71). However, several other proteins that shuttle between the cytoplasm and the nucleus, like ribosome proteins, histones, ␤-catenin, the human immunodeficiency virus-1 Vpr protein, employ other importins that recognize distinct NLS (72). Although we found a sequence of amino acid residues with homology to classical bipartite NLS in the cys1 of PKD, the deletion of this putative NLS did not prevent the nuclear import of GFP-PKD-⌬cys1. However, the deletion of the cys2 domain blocked the nuclear import of GFP-PKD-⌬cys2 despite the lack of a classical NLS. These findings suggest that a nonclassical NLS may be mediating the nuclear import of PKD and further support the notion that cys1 and cys2 domains, despite their amino acid homology, perform nonredundant functions in the regulation of PKD.
The nuclear export of PKD required its PH domain and a competent Crm1-dependent nuclear export pathway. This is a novel function for the PH domain of PKD, which, as we previously demonstrated, plays a negative role in the regulation of PKD activity (4) and mediates the formation of a PKD⅐PKC complex (3). Interestingly, the nuclear export of Bruton's tyrosine kinase also involves its PH domain and the Crm1-dependent nuclear export pathway (66), suggesting that the involvement of the PH domains in nuclear transport may be more common than has previously been recognized.
Several proteins that continuously shuttle between the cytoplasm and the nucleus are subject to regulated nuclear transport in response to specific cellular signals (54 -56). These signals determine the relative rates of nuclear import and export of these proteins and, therefore, their subcellular localization. The stimulation induced by mitogenic GPCR agonists caused a transient nuclear accumulation of PKD that correlated with its faster rate of nuclear import. One of the signals associated with regulated nuclear transport is phosphorylation (55). Remarkably, the inhibition of PKC, which mediates the phosphorylation of the activation loop of PKD after bombesin stimulation (20), prevented the nuclear accumulation of PKD, suggesting that the activation of PKD is associated with increases in its rate of nuclear import. Interestingly, the nuclear accumulation of PKD in response to the mitogenic signals evoked by bombesin and vasopressin in Swiss 3T3 cells has a close similarity to the distinct spatial and temporal localization of mitogen-activated protein kinase/extracellular signal-regulated kinase in response to growth factors. In quiescent cells, the mitogenic stimulation elicited by growth factors also induces the transient nuclear accumulation of mitogen-activated protein kinase/extracellular signal-regulated kinase (69,(73)(74)(75)(76)(77). Overall, our results demonstrate for the first time that PKD shuttles between the cytoplasm and the nucleus and that the nuclear transport of PKD is regulated in response to GPCR agonists that promote G q -mediated activation of PKD.
Taken together with our recent results (26), the dynamic changes in the intracellular distribution of PKD in response to GPCR stimulation suggest an attractive model for the regulation of the activity and multiple site phosphorylation of this enzyme. In this model, GPCR stimulation induces a rapid translocation of PKD from the cytosol to the plasma membrane that is accompanied by its activation by a PKC-mediated process that leads to the phosphorylation of its activation loop and subsequent plasma membrane dissociation (20,26). As a result, active PKD is rapidly dispatched from the plasma membrane and transiently accumulates inside the nucleus. In this way, PKD is likely to transmit a signal(s) initiated at the cell plasma membrane to the cell nucleus. Given that PKD potentiates the reinitiation of DNA synthesis and cell proliferation induced by bombesin and vasopressin (33), it is tempting to speculate that the nuclear translocation of PKD contributes to the transduction of the mitogenic response induced by these GPCR agonists.