Protein Kinase Cν/Protein Kinase D3 Nuclear Localization, Catalytic Activation, and Intracellular Redistribution in Response to G Protein-coupled Receptor Agonists*

The protein kinase D (PKD) family consists of three serine/threonine kinases: PKCμ/PKD, PKD2, and PKCν/PKD3. Whereas PKD has been the focus of most studies, virtually nothing is known about the effect of G protein-coupled receptor agonists (GPCR) on the regulatory properties and intracellular distribution of PKD3. Consequently, we examined the mechanism that mediates its activation and intracellular distribution. GPCR agonists induced a rapid activation of PKD3 by a protein kinase C (PKC)-dependent pathway that leads to the phosphorylation of the activation loop of PKD3. Comparison of the steady-state distribution of endogenous or tagged PKD3 versus PKD and PKD2 in unstimulated cells indicated that whereas PKD and PKD2 are predominantly cytoplasmic, PKD3 is present both in the nucleus and cytoplasm. This distribution of PKD3 results from its continuous shuttling between both compartments by a mechanism that requires a nuclear import receptor and a competent CRM1-nuclear export pathway. Cell stimulation with the GPCR agonist neurotensin induced a rapid and reversible plasma membrane translocation of PKD3 that is PKC-dependent. Interestingly, the nuclear accumulation of PKD3 can be dramatically enhanced in response to its activation. Thus, this study demonstrates that the intracellular distribution of PKD isoenzymes are distinct, and suggests that their signaling properties are regulated by differential localization.

The protein kinase C (PKC) 1 family links receptor-mediated signaling events to a extensive variety of cellular functions (1)(2)(3)(4)(5). There are multiple related PKC isoforms (1,3,4), which can be classified into three distinct subgroups on the basis of structural and regulatory differences: the conventional PKCs (␣, ␤ I , ␤ II , and ␥), which are regulated by calcium, DAG, and phospholipids; the novel PKCs (␦, ⑀, , and ), which are regulated by diacylglycerol (DAG) and phospholipids; and the atypical PKCs ( and ), whose regulation is less characterized, but that have been proposed to be regulated by D-3 phosphoinositides (6). The DAG-regulated PKC isoforms all bind phorbol esters and are major cellular targets for this class of tumor promoter (7). It is increasingly recognized that each isoform has specific functions in vivo (8). However, the mechanisms by which PKC-mediated signals in the plasma membrane are propagated to critical downstream targets remain largely undefined.
Protein kinase D (PKD)/protein kinase C (9, 10) and two recently identified serine protein kinases termed PKD2 and PKC/PKD3, which are similar in overall structure and primary amino acid sequence to PKD (11,12), constitute a new protein kinase subfamily separate from the previously identified PKCs. The salient features of the PKD structure include the presence of a catalytic domain distantly related to Ca 2ϩregulated kinases, a pleckstrin homology (PH) domain that regulates PKD activity, and a highly hydrophobic stretch of amino acids in its NH 2 -terminal region (9,10,13,14). The NH 2 -terminal region of PKD contains in addition to the PH domain, a cysteine-rich domain (CRD) that confers high affinity binding to phorbol esters and DAG (9, 14 -16).
Much less information is available on the regulation, biological roles, and intracellular distribution of other members of the PKD family. PKD2 can be activated by phorbol esters and the GPCR agonist gastrin (12,50) but the precise mechanism mediating its activation is still under investigation (50). As for its intracellular distribution, we recently showed that PKD2 is present only in the cytosol of unstimulated or NT-stimulated cells (51). Regarding PKD3, virtually nothing is known about its regulatory properties or intracellular distribution in response to GPCR agonists, although its cDNA sequence obtained from expressed sequence tag deduction and rapid amplification of cDNA ends was published before the cloning of PKD2 (11).
The experiments presented in this paper were designed to examine the mechanisms regulating the intracellular distribution and catalytic activity of PKD3. In striking contrast to the predominantly cytoplasmic localization of PKD and PKD2 shown in our previous studies, we demonstrate here that PKD3 is present both in the cytoplasm and in the nucleus of nonstimulated cell. Several lines of evidence, including studies using fluorescence loss in photobleaching (FLIP), indicate that PKD3 continuously shuttles between the nucleus and the cytoplasm by a mechanism that requires a nuclear import receptor and a competent CRM1-nuclear export pathway. Our results also show that GPCR agonists induce a striking translocation of PKD3 from the cytosol to the plasma membrane that leads to its catalytic activation and rapid plasma membrane dissociation through PKC-mediated phosphorylation of the activation loop of PKD3 (Ser 731 and Ser 735 ). Although PKD and PKD3 can be activated by PKC, we found that only PKD forms a stable molecular complex with this novel PKC isoform. Upon its activation, PKD3 rapidly dissociates from the plasma membrane, moves into the cytosol, and is subsequently imported into the nucleus at a faster rate than in non-stimulate cells, which results in its further nuclear accumulation. These findings reveal a connection between PKD3 and PKC and support a novel mechanism by which GPCRmediated activation of the closely related serine/threonine kinases of the PKD family can potentially generate diverse physiological responses based on their unique subcellular distribution.

Cell Culture and Transfections
Panc-1 cells, obtained from American Type Cell Collection, are epithelioid cells originally obtained from human ductal pancreatic adenocarcinoma. Panc-1 cells were grown at 37°C in Dulbecco's modified Eagle's medium, supplemented with 4 mM glutamine, 1 mM sodium pyruvate, and 10% fetal bovine serum in a humidified atmosphere containing 10% CO 2 and 90% air. COS-7 and Madin-Darby canine kidney (MDCK) cells were obtained from American Type Cell Collection and maintained 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 biochemical analysis, cells were subcultured to 80 -90% confluence in 10-cm dishes and transfections or cotransfections were carried out with equivalent amounts of DNA (10 g/dish). Transfections were carried out in Opti-MEM using Lipofectin according to the manufacturer's suggested conditions (Invitrogen). Transfected cells were incubated for 48 h before analysis. 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, a time that we previously found optimum to detect very low levels of expression of green fluorescent protein (GFP)-tagged proteins (41). Cells were transfected with 1 g of DNA/33-mm dish using LipofectAMINE PLUS (Invitrogen) according to the manufacturer's suggested conditions. For immunocytochemistry, the cells were plated in 33-mm dishes at 7 ϫ 10 4 cells/dish and transfected as indicated above with 1 g/DNA dish. Transfected cells were incubated for 18 h before live or fixed cell imaging.

mRNA Amplification
To detect the expression of sequences encoding PKD, PKD2, and PKD3 in Panc-1 cells we employed reverse transcriptase-PCR (RT-PCR) with specific oligonucleotide primers designed using the public domain Primer3 program (52). 2 The cDNA sequences used to design the primers were based on the published PKD (10) and PKD2 (12) sequences. For PKD3, we designed specific primers with the Primer3 program based on the cDNA sequence obtained from expressed sequence tag deduction and rapid amplification of cDNA ends (11). The primers sequences for PKD were, sense primer 5Ј-GAGATGGCTT-GCTCCATTGT-3Ј corresponding to nucleotide positions 235-254 and the antisense primer 5Ј-TCGCCTTCCTGGATATCACT-3Ј corresponding to nucleotide positions 361-380, considering for both primers as nucleotide number 1 the first nucleotide of the ATG initiation codon. The primers sequences for PKD2 were, sense primer 5Ј-CACCTTC-GAGGACTTCCAGA-3Ј corresponding to nucleotide positions 404 -421 and the antisense primer 5Ј-CAGCGCTTGTGGTAGTTCAG-3Ј corresponding to nucleotide positions 537-556, considering for both primers as nucleotide number 1 the first nucleotide of the ATG initiation codon. The primers sequences for PKD3 were, sense primer 5Ј-TCAGCCCA-GAAGTCTGTATTACCC-3Ј corresponding to nucleotide positions 25-48 and the antisense primer 5Ј-TGGAGTTGGTGAGTGATG-GTGC-3Ј corresponding to nucleotide positions 139 -160, considering for both primers as nucleotide number 1 the first nucleotide of the ATG initiation codon. RT-PCR was performed on 1 g of total RNA extracted from Panc-1 cells using TRIzol reagent (Invitrogen). First-strand cDNA was synthesized at 45°C by using the designed PKD, PKD2, or PKD3 antisense oligonucleotides described above and ThermoScript reverse transcriptase (Invitrogen) under the conditions suggested by the manufacturer. A fraction of the obtained cDNAs were amplified by PCR using Platinum TaqDNA polymerase high fidelity (Invitrogen), as suggested by the manufacturer, and the sense and antisense primers specific for PKD, PKD2, or PKD3 are indicated above. As controls, in a set of reactions the sense primers for PKD, PKD2, or PKD3 were not added to the obtained cDNAs during the PCR amplification step. The products of PCR were resolved in 2.5% NuSieve GTG-agarose (FMC BioProducts, Rockland, ME), 1ϫ TBE buffer (100 mM Tris, 90 mM boric acid, 1 mM EDTA, pH 8.4). The gel was stained for 60 min with ethidium bromide (0.5 g/ml) in 1ϫ TBE, follow with two 15-min washes with distilled water. Gel was viewed and images captured using a Gel Doc 1000 system (Bio-Rad). The predicted sizes of the RT-PCR products for PKD, PKD2, and PKD3 were 145, 152, and 135 bp, respectively.

PKD3 Cloning and cDNA Constructs
Primers for the synthesis of the cDNA encoding PKD3 were designed with the Primer3 program using the cDNA sequence obtained from expressed sequence tag deduction and rapid amplification of cDNA ends (11). The sense primer was 5Ј-CGGGGTACCTCAGATGTCTG-CAAATAATTCCCCT-3Ј corresponding to nucleotide positions Ϫ3 to 21 and the antisense primer was 5Ј-CGGGGTACCGAAGAACAAGTTA-CACAGCA-3Ј corresponding to nucleotide positions 2725-2746, considering for both primers as nucleotide number 1 the first nucleotide of the ATG initiation codon. Total RNA (3 g) extracted from Panc-1 cells using TRIzol reagent was employed as template for the synthesis at 55°C of the cDNA encoding PKD3 using the ThermoScript reverse transcriptase and the antisense primers indicated above. A fraction of the obtained cDNAs were amplified by PCR using Platinum TaqDNA polymerase high fidelity, as suggested by the manufacturer, and the sense and antisense primers indicated above. The PCR products were digested with KpnI and resolved in a 1% agarose gel, 1ϫ TBE buffer. The band with the predicted size for PKD3 was eluted from the gel and cloned into a vector encoding the green fluorescent protein, pEGFP-C3 (BD Biosciences and Clontech, Palo Alto, CA), previously digested with KpnI. The obtained construct, pGFP-PKD3, was verified by DNA sequence analysis and the products of expression analyzed by Western blot using antibodies against GFP. The deduced amino acid sequence for the PKD3 cDNA clone obtained from Panc-1 cells was identical to the published one (11). The product of expression of pGFP-PKD3 matched the predicted protein size and did not show any degradation. The cDNA encoding PKD3 was isolated from pGFP-PKD3 using KpnI and subcloned into pFLAG-CMV-4 (Sigma), previously digested with KpnI. The obtained construct, pFLAG-PKD3, was verified by DNA sequence analysis and the product of expression analyzed by Western blot using the murine monoclonal antibodies M2 and M5 (Sigma) against the FLAG epitope. The vectors encoding the chimeric fusion protein between GFP and PKD or PKD2 were previously described (29,51). A schematic representation of the employed constructs is shown in Fig. 1B.
The constitutively active PKC clones , ⑀, and were obtained from Peter Parker, Protein Phosphorylation Laboratory, Imperial Cancer Research Fund (53,54). Dr. J. Battey (Laboratory of Molecular Biology, NIDCD, National Institutes of Health) kindly provided the vector bNR-pCD2 containing the cDNA encoding the bombesin/gastrin releasing peptide receptor.

In Vitro Kinase Assays
Immune complexes were washed twice with lysis buffer, then twice with kinase buffer consisting of 30 mM Tris-HCl, pH 7.4, 10 mM MgCl 2 , 1 mM dithiothreitol. Autophosphorylation reactions were initiated by combining 20 l of immune complexes with 5 l of a phosphorylation mixture containing 100 M [␥-32 P]ATP (specific activity, 400 -600 cpm/ pmol) in kinase buffer. Following incubation at 30°C for 10 min, the reactions were terminated by addition of 1 ml of ice-cold kinase buffer and placed on ice. Immune complexes were recovered by centrifugation, and the proteins were extracted for SDS-PAGE analysis by addition of 2ϫ SDS-PAGE sample buffer (200 mM Tris-HCl, pH 6.8, 0.1 mM sodium orthovanadate, 1 mM EDTA, 6% SDS, 2 mM EDTA, 4% 2-mercaptoethanol, 10% glycerol). Dried SDS-PAGE gels were subjected to autoradiography to visualize radiolabeled protein bands.

Western Blot Analysis
Samples of cell lysates were directly solubilized by boiling in SDS-PAGE sample buffer. Following SDS-PAGE on 8% gels, proteins were transferred to Immobilon-P membranes (Millipore Corp., Bedford, MA) and blocked by overnight incubation with 5% nonfat dried milk in phosphate-buffered saline (PBS), pH 7.2. Membranes were incubated at room temperature for 3 h with rabbit antiserum specifically recognizing GFP, PKD3 (N17), or the murine monoclonal antibodies recognizing the FLAG epitope diluted in PBS containing 3% nonfat dried milk. To determine the phosphorylation/activation state of the different PKD isoforms, the membranes were incubated with the Aloop antibody that specifically recognizes the phosphorylated state of the equivalent serines within the activation loop of the three PKD isoforms (PKD: Ser 744 -Ser 748 ; PKD2: Ser 706 -Ser 710 ; PKD3: Ser 731 -Ser 735 ) (Refs. 36, 50, and this article).

Cell Imaging
Live Cell Imaging-To maintain a constant temperature of 37°C during the experimental procedures, cells were grown on the 15-mm glass coverslips and subsequently mounted in a perfusion chamber (RC-25 Warner Instrument Corp., Hamden, CT) and perfused with medium preheated at 37°C by a SH-27B solution in-line heater (Warner Instrument Corp.). The medium was supplemented with 10 mM HEPES, pH 7.2. The perfusion chamber was mounted on an epifluorescence microscope (Zeiss Axioskop) and cells were imaged with a Zeiss water immersion objective (Achroplan ϫ40/0.75w, Carl Zeiss Inc., Jena, Germany). Images were captured as uncompressed 24-bit TIFF files with a SPOT cooled (Ϫ12°C) single CCD color digital camera (three pass method) driven by SPOT version 2.1 software (Diagnostic Instruments, Inc., Sterling Heights, MI). GFP fluorescence was observed with a HI Q filter set for fluorescein isothiocyanate (Chroma Technology, Brattleboro, VT). Fifty cells were analyzed per experiment and each experiment was performed in triplicate. The selected cells displayed in the appropriate figures were representative of 90% of the population of positive cells.
Indirect Immunofluorescence-Indirect immunofluorescence was performed in 10% buffered formalin phosphate fixed, 0.3% Triton X-100/PBS permeabilized cultures. FLIP-FLIP experiments were performed on a Zeiss LSM310 confocal microscope (Carl Zeiss Inc.) using a ϫ100/1.0 water immersion lens. First, a control image was obtained at focal plane at a depth approximately midway through the cell using the 488-nm line of the Argon laser for excitation of GFP with an attenuation of 90% and stored as 8-bit TIFF file. Then a small rectangular region of interest was defined within the cytoplasm or nuclei, the laser attenuation was completely removed and the region of interest (noted by white outlined areas in the Fig. 6) was continuously scanned at 488 and 514 nm at full laser power for 2 min (cytoplasm) or 30 s (nuclei) by using the LSM software provided with the Zeiss 310 microscope, resulting in continuous photobleaching (loss of fluorescence) of any molecule within that region. In time, freely moving molecules located previously outside of the region of interest will diffuse into that region and become photobleached, resulting in a time-dependent loss of fluorescence within the larger area (cytoplasm or nucleus). At the end of the bleaching period, laser attenuation and excitation wavelengths were restored to those of the control image, and the fluorescence throughout the cell was monitored at various time intervals. During experiments, a constant temperature of 37°C was maintained following the methodology described under "Live Cell Imaging."

Materials
[␥-32 P]ATP (370 MBq/ml) and horseradish peroxidase-conjugated donkey anti-rabbit IgG were from Amersham Biosciences. The rabbit anti-PKC, anti-PKC⑀, anti-PKC, and anti-GFP antibodies were obtained from Santa Cruz Biotechnology, Inc., Santa Cruz, CA. The Aloop antibody (phospho-PKD/PKC Ser 744 -Ser 748 antibody), which specifically recognizes the phosphorylated state of the equivalent serines within the activation loop of the three PKD isoforms (PKD: Ser 744 -Ser 748 ; PKD2: Ser 706 -Ser 710 ; PKD3: Ser 731 -Ser 735 ), was obtained from Cell Signaling Technology, Beverly, MA. The polyclonal anti-PKD antibody PA-1 was previously described (13). Fluorescein or Texas Redconjugated goat anti-mouse or anti-rabbit immunoglobulins were obtained from Molecular Probes, Eugene, OR. The polyclonal anti-PKD3 N17 antibody was raised by immunizing rabbits with a peptide corresponding to the human PKD3 17 amino-terminal residues SANNSPP-SAQKSVLPTA using standard procedures. Neurotensin, DAPI, and the murine monoclonal anti-FLAG antibodies M2 and M5 were obtained from Sigma. Leptomycin B (LMB) was a generous gift from Dr. Minoru Yoshida, Department of Biotechnology, University of Tokyo. All the other reagents were the highest grade commercially available.

PKD3 Cloning, Structural Analysis, and Functional
Expression-Panc-1 cells, which express both PKD and PKD2 (22,51), is an extensively studied cell model of human ductal pancreatic cancer (22,(55)(56)(57)(58)(59)(60)(61). To examine whether Panc-1 cells also express PKD3, total RNA isolated from these cells was subjected to RT-PCR using specific primers designed to amplify PKD3 transcripts. We also designed specific primers to amplify transcripts encoding for PKD and PKD2 to be used as positive controls. A major PCR product of the predicted size for PKD3 (135 bp), PKD2 (152 bp), and PKD (145 bp) was obtained for each set of primers (Fig. 1A, lanes 2, 4, and 6), indicating that transcripts for PKD3, as well as PKD and PKD2, are present in Panc-1 cells. No DNA migrating in the vicinity of the positive bands was visible when the sense primer was omitted in the PCR amplification step (Fig. 1A, lanes 1, 3, and 5).
Once established that putative transcripts encoding for PKD3 were present in total RNA isolated from Panc-1 cells, we used that same total RNA to synthesize a full-length cDNA for PKD3 by RT-PCR. The deduced amino acid sequence of the obtained PKD3 cDNA (890 amino acids with a predicted molecular mass of 100.5 kDa) matched 100% the one obtained from expressed sequence tag deduction and rapid amplification of cDNA ends for PKD3 (11). PKD3 has 71 and 66% overall homology to PKD and PKD2, respectively. The catalytic domain is the most conserved structural domain between the different members of the PKD family. In contrast, the regulatory domain is the most divergent region of this family (Table  I). In particular, the PH domain of PKD3 has the lowest level of homology to the PH domains of PKD and PKD2 (Table I). The PH domain of PKD mediates regulation of its catalytic activity, complex formation with other signaling molecules, and its nuclear export (13,62,63). Therefore, it is possible that these sequence differences are important for the regulation and intracellular distribution of PKD3. In addition, hydropathy analysis revealed that PKD3 does not have a hydrophobic domain in its NH 2 terminus as the one present in PKD and PKD2. Overall, PKD3 is the most distinct member of the PKD family.
To facilitate the unambiguous detection of PKD3, we constructed a plasmid encoding a chimeric PKD3 protein by fusing the GFP of Aequorea victoria to 5Ј of the cDNA encoding PKD3 (Fig. 1B). The obtained construct, pGFP-PKD3, was transfected into Panc-1 cells and the product of its expression analyzed by Western blot using antibodies against GFP (Fig. 1C, left panel). The product of expression of pGFP-PKD3 matched the predicted size for the encoded protein (Ϸ130 kDa) and did not show any degradation.
Differential Intracellular Distribution of PKD Isoforms-In view of the differences between PKD and PKD3 in their regulatory region (Table I), we consider the possibility that PKD3 is a kinase with an intracellular distribution distinct from PKD. To test this hypothesis, we transiently transfected pGFP-PKD3 into Panc-1 cells and its intracellular distribution was examined in living cells. A construct encoding a fusion protein between GFP and PKD was also used for comparison because the distribution of the encoded GFP-PKD protein has already been described in several cell types (40,41,63). In addition, we also compared the distribution of GFP-PKD3 to GFP-PKD2, because we recently demonstrated that PKD2 is principally localized in the cytoplasm in Panc-1 cells (51).
In agreement with our recent results (41, 51), we found that GFP-PKD and GFP-PKD2 expressed in Panc-1 cells are predominantly in the cytoplasm (Fig. 1C). In striking contrast, GFP-PKD3 accumulates in the nucleus, in addition to its cytosolic localization (Fig. 1C). The morphology of GFP-, GFP-PKD-, GFP-PKD2-, or GFP-PKD3-transfected cells was indistinguishable from that of non-transfected Panc-1 cells (data not shown). A similar nuclear accumulation of GFP-PKD3 was observed when this fusion protein was transiently expressed in other cell types, including COS-7 or MDCK epithelial cells (Fig.  1C), NIH and Swiss 3T3 fibroblasts, and in the intestinal epithelial cell line IEC-18 (data not shown).
The molecular weight of GFP-PKD3 (ϳ130,000) far exceeds the size limit for passive nuclear diffusion (45,000 -60,000) (64 -67). Nevertheless, its fluorescent tag could interfere with the normal distribution of PKD3. Therefore, we constructed a plasmid encoding a different chimeric PKD3 protein by fusing the small FLAG epitope (8 amino acid residues) to the 5Ј of the cDNA encoding PKD3 (Fig. 1B). The obtained construct, pFLAG-PKD3, was transfected into Panc-1 cells and its expression was analyzed by Western blot using monoclonal antibodies against the FLAG epitope (Fig. 1D, left panel). The product of expression of pFLAG-PKD3 matched the predicted size for the encoded protein (Ϸ100 kDa) and did not show any degradation.
We then analyzed the localization of FLAG-PKD3 by indirect immunofluorescence of fixed cells. No difference was detected between the distributions of FLAG-PKD3 and GFP-PKD3 in different cell types, including Panc-1 or MDCK cells (Fig. 1D, right panels), demonstrating that the unique intracellular distribution of GFP-PKD3 was not because of the fluorescent tag. The localization of FLAG-PKD3 in the nucleus of Panc-1 and MDCK cells was corroborated by DAPI costaining of those cells (Fig. 1D).
Intracellular Distribution of Endogenous PKD3-Our results indicated that PKD3 displayed an intracellular distribution that was unique within the PKD family. However, this conclusion was based on the observed distribution of ectopically expressed PKD3 (GFP-PKD3 and FLAG-PKD3). To corroborate this conclusion, we examined the intracellular distribution of endogenous PKD3 using a specific antibody that recognizes PKD3.
A polyclonal anti-PKD3 antibody (N17) was raised by immunizing rabbits with the peptide SANNSPPSAQKSVLPTA corresponding to 17 amino acids in the NH 2 terminus of human PKD3. The N17 antibody specificity was tested against different members of the PKD family using Western blot analysis. Panc-1 cells transiently transfected with pGFP-PKD, pGFP-PKD2, or pGFP-PKD3 were lysed 18 h post-transfection and the solubilized proteins were resolved by SDS-PAGE and Western blotted using the N17 antibody. As shown in Fig. 2A, top panel, the N17 antibody only recognized GFP-PKD3. The reactivity of N17 against GFP-PKD3 was eliminated by preincubating the N17 antiserum with the immunizing peptide ( Fig.  2A, top panel). Parallel controls were stained with an anti-GFP antibody to verify the expression and equal loading of the three PKD isoforms (Fig. 2A, bottom panel).
Next, we examined the intracellular distribution of endogenous PKD3 by indirect immunofluorescence in unstimulated fixed Panc-1 cells using the N17 antibody. Contrasting with the cytoplasmic localization of endogenous PKD (Fig. 2B, left panel), endogenous PKD3, in addition to its cytoplasmic localization, showed a very prominent nuclear localization (Fig. 2B,  center panel). The same distribution of endogenous PKD3 was detected in MiaPaca-2 cells, another human pancreatic cancer cell line (data not shown). No signal was detected when the cells were stained with the N17 antibody previously preincubated with the immunizing peptide (Fig. 2B, right panel). A similar intracellular distribution of endogenous PKD3 was observed in Panc-1 cells using a polyclonal antibody recognizing the carboxyl-terminal 16 residues (HFIMAPNPDDMEEDP) of human PKD3 (data not shown). Thus, the intracellular distribution of endogenous PKD3 is identical to that of ectopically expressed FLAG-PKD3 or GFP-PKD3. Our results indicate that in contrast to PKD (41) or PKD2 (51), PKD3 localizes not only to the cytosol but also to the nucleus in unstimulated cells. Phorbol Esters and GPCR Agonists Activate PKD3-The distinct intracellular distribution of PKD3 prompted us to examine the pathways and mechanisms leading to its catalytic activation. Panc-1 cultures transiently transfected with pGFP-PKD, pGFP-PKD3, or pEGFP-C3 (empty vector) (see Fig. 3A) were stimulated with 200 nM PDBu for 10 min, lysed, and the different chimeric proteins immunoprecipitated from the extracts with an antibody against GFP. The immune complexes were incubated with [␥-32 P]ATP, subjected to SDS-PAGE, and analyzed by autoradiography to detect the prominent 140 -130-kDa bands corresponding to the autophosphorylated PKD moieties of GFP-PKD and GFP-PKD3, respectively. GFP-PKD3 isolated from unstimulated Panc-1 cells had low catalytic activity that was markedly activated by PDBu stimulation of intact cells. This rapid and striking increase in PKD3 kinase activity was maintained during cell lysis and immunoprecipitation. PDBu treatment of Panc-1 cells expressing GFP-PKD also induced a rapid and striking increase in PKD kinase activity shown for comparison (Fig. 3A).
Recently we reported that NT binding to its heptahelical GPCR induces PKC-dependent PKD activation and DNA synthesis in Panc-1 cells (22). Interestingly, the expression of the NT receptor is markedly increased in pancreatic cancer samples as compared with normal controls (68 -70). It is therefore important to characterize NT-mediated signal transduction pathways in human pancreatic cancers and in this study we examined whether NT induces PKD3 activation in Panc-1 cells.
Cultures of these cells transiently transfected with pGFP-PKD, pGFP-PKD3, or pEGFP-C3 (empty vector) were stimulated for 10 min with 50 nM NT, lysed, and the catalytic activity of the different chimeric proteins was assayed as described above. The results presented in Fig. 3A show that stimulation of Panc-1 cells with NT induced a marked increase in the catalytic activity of the PKD3 moiety of GFP-PKD3 as well as in the PKD moiety of GFP-PKD. No catalytic basal activity was detected in the control Panc-1 cells expressing GFP in the absence or presence of NT or PDBu.
Similarly, we also analyzed the catalytic activation of FLAG-PKD3 in Panc-1 cells. Cultures of these cells transiently transfected with pFLAG-CMV-4 (empty vector) or pFLAG-PKD3 (see Fig. 1B) were stimulated for 10 min with 50 nM NT or 200 nM PDBu, lysed, and FLAG-PKD3 was immunoprecipitated from the cell extracts with a mixture of murine monoclonal antibodies against the FLAG epitope. The immune complexes were incubated with [␥-32 P]ATP, subjected to SDS-PAGE, and analyzed by autoradiography to detect a prominent Ϸ100-kDa band corresponding to the autophosphorylated PKD3. The results presented in Fig. 3B show that treatment of Panc-1 cells  , lane 4). The production of the N17 antibody is described under "Experimental Procedures." Parallel controls were stained with an anti-GFP antibody to verify the expression and equal loading of the three PKD isoforms (lower panel). B, Panc-1 were fixed and processed for immunocytochemistry using the anti-PKD (left panel) or anti-PKD3 N17 antibody non-preabsorbed (center panel) or preabsorbed (right panel) with the immunizing N17 peptide as described above. Fluorescein isothiocyanate-conjugated goat anti-rabbit immunoglobulins were used as secondary antibodies. An epifluorescence microscope was used to visualize the distribution of endogenous PKD and PKD3. Representative images were captured as described under "Experimental Procedures." Bar, 10 m. with either PDBu or NT also induced a striking increase in the catalytic activity of FLAG-PKD3 that was comparable with the one detected, under the same conditions, with GFP-tagged PKD3. No catalytic basal activity was detected in control Panc-1 cells transfected with pFLAG-CMV-4 in the absence or presence of NT or PDBu.
To substantiate the results shown in Fig. 3A with tagged PKD3, we also examined the catalytic activation of endogenous PKD3 in Panc-1 cells stimulated with NT or PDBu. Panc-1 cells were treated for 10 min with 50 nM NT or 200 nM PDBu, lysed, and PKD3 was immunoprecipitated from the cell extracts with the N17 antibody. The immune complexes were incubated with [␥-32 P]ATP, subjected to SDS-PAGE, and analyzed by autoradiography. A prominent Ϸ100-kDa band corresponding to the autophosphorylated PKD3 was detected in the Panc-1 cells treated with either PDBu or NT (Fig. 3C). No catalytic basal activity was detected in control Panc-1 cells in the absence of NT or PDBu.
Next, we examined whether an increase in PKD3 catalytic activity can be demonstrated in another cell system by stimulation with phorbol ester or through another GPCR-mediated pathway. COS-7 cells transiently cotransfected with the bombesin GPCR and pGFP-PKD, pGFP-PKD3, or pEGFP-C3 (empty vector) were stimulated for 10 min with 10 nM bombesin or 200 nM PDBu, lysed, and the catalytic activity of the different chimeric proteins assayed by in vitro kinase assays as described above. As shown in Fig. 3D, bombesin or PDBu stimulation of COS-7 cells transiently cotransfected with GFP-PKD3 together with the bombesin GPCR induced a rapid increase in the catalytic activity of PKD3 as well as in its control expressing the bombesin GPCR and GFP-PKD. No catalytic basal activity was detected in COS-7 cells expressing GFP and the bombesin GPCR in the absence or presence of bombesin or PDBu (data not shown). The results presented in Fig. 3 demonstrate, for the first time, that PKD3 can be converted from inactive to an active state by cell treatment with either a biologically active phorbol ester or the GPCR agonists NT or bombesin.
PKD3 Activation Correlates with the Phosphorylation of Its Activation Loop-The signal-dependent phosphorylation of amino acid residues within the activation loop of many protein kinases plays a critical role in their activation. Here we examined whether the activation of PKD3 by treatment of intact cells with PDBu, NT, or bombesin also induces the phosphorylation of Ser 731 and Ser 735 located in the activation loop of PKD3. These residues in PKD3 are equivalent to Ser 744 and Ser 748 within the activation loop of PKD.
Panc-1 cells transiently transfected with pEGFP-C3 (empty vector), pGFP-PKD, or pGFP-PKD3 were treated for 10 min with NT (50 nM) or PDBu (200 nM) and lysed. The solubilized proteins were resolved by SDS-PAGE and Western blotted using the Aloop antibody that recognizes the phosphorylated serines within the PKDs activation loop sequence IGEKS-FRRSVVGT. As shown in Fig. 4A, top panel, NT or PDBu induced the activation loop phosphorylation of the PKD3 moiety of GFP-PKD3, which correspond to Ser 731 and Ser 735 , as well as the activation loop phosphorylation of the PKD moiety of GFP-PKD, which correspond to Ser 744 and Ser 748 . Several bands in the vicinity of the 100-kDa marker were detected in PDBu and NT challenged control Panc-1 cells expressing GFP (Fig. 4A, top panel). These bands, which were also detectable in Panc-1 cells expressing GFP-PKD and GFP-PKD3, correspond to the endogenously expressed PKDs upon their activation mediated by PDBu or NT (Fig. 4A, top panel). The same membrane was stripped and subsequently probed with an antibody against GFP to verify the expression and equal loading of GFP-PKDs (Fig. 4A, bottom panel). Similar results were also obtained with COS-7 cells expressing the bombesin receptor and GFP-PKD3 when they were challenged with PDBu or bombesin (Fig. 4B) or in Panc-1 cells expressing FLAG-PKD3 when the cultures were challenged with NT or PDBu (Fig. 4C). The results displayed in Figs. 3 and 4 demonstrate, for the first time, that PDBu and GPCR agonists induce a rapid and persistent increase in the catalytic activity of PKD3 as measured by autokinase activity and direct activation loop phosphorylation.
Activation of Neurotensin-GPCR Induces Transient Translocation of GFP-PKD3 from the Cytosol to the Plasma Membrane-Taking into consideration the similarity in the activation but the differential intracellular distribution of PKD and PKD3, we decided to compare their spatial and temporal distribution in Panc-1 cells in response to NT. As illustrated by the image presented in Fig. 4D, GFP-PKD expressed in unstimulated Panc-1 cells was distributed throughout the cytosol, whereas GFP-PKD3 was present in both cytosol and nucleus in agreement with the results presented in Fig. 1C. In each case, very little fluorescent signal was localized to the plasma membrane. Real time imaging of living cells revealed that NT stimulation of Panc-1 cells induced a rapid translocation of GFP-PKD3 to the plasma membrane (Fig. 4D, arrows). Translocation of GFP-PKD3 occurred within 30 s and reached a maximum between 2 and 3 min. The association of this chimeric protein with the plasma membrane was transient. The reverse translocation of GFP-PKD3 from the plasma membrane to the cytosol was complete within 5 min of NT stimulation (Fig. 4D). The rapid and transient translocation of GFP-PKD3 to the plasma membrane is comparable with that of GFP-PKD (Fig. 4D). The distribution of GFP, detected in the nuclei as well as throughout the cytosol, was not affected by NT stimulation (data not shown).
Inhibition of PKC Prevents the Activation and Rapid Plasma Membrane Dissociation of PKD3-To determine whether GPCR agonists induce PKD3 activity and activation loop phosphorylation through a PKC-mediated pathway, we used the inhibitors of phorbol ester-sensitive isoforms of PKC, GF I (also known as GF 109203X or bisindolylmaleimide I), or Ro 31-8220. Panc-1 cells transiently transfected with pGFP-PKD3 were treated for 1 h with GF I or Ro 31-8220 prior to stimulation with 50 nM NT or 200 nM PDBu. As shown in Fig. 5A, exposure to either GF I or Ro 31-8220 prevented the activation of PKD3 as well as the phosphorylation of its activation loop induced by NT or PDBu. In contrast, the compound GF V, which is related to GF I but biologically inactive, did not prevent PKD3 phosphorylation in response to NT or PDBu in Panc-1 cells. Inhibition of the activation loop phosphorylation of PKD3 as well as inhibition of its catalytic activity was also observed in GF I or Ro 31-8220 pre-treated COS-7 cells expressing the bombesin receptor and GFP-PKD3 when these cells were challenged with either PDBu or bombesin (Fig. 5B).
We then examined whether PKC inhibitors affect the rapid translocation of PKD3 from the plasma membrane to the cytosol. Treatment of cultures expressing GFP-PKD3 with GF I prior to stimulation with NT did not prevent their translocation from the cytosol to the plasma membrane, indicating that classical and novel PKCs are not necessary for this step (Fig. 5 C,  bottom panel). However, addition of GF I dramatically delayed the plasma membrane dissociation of GFP-PKD3 (Fig. 5C, bottom panel, arrows) that we detected in cells untreated with PKC inhibitors (Fig. 5C, top panel; see also Fig. 4D). Similar results were obtained using Ro 31-8220, another PKC inhibitor (data not shown). Therefore, these results demonstrate that agonist stimulation of the NT GPCR induces a striking and transient translocation of GFP-PKD3 to the plasma membrane of Panc-1 cells and that its rapid reverse translocation can be delayed by PKC catalytic activity inhibitors.
As demonstrated above, treatment of intact cells with the inhibitors GF I or Ro 31-8220 inhibited PKD3 activation induced by subsequent addition of PDBu or NT. In striking contrast, GF I or Ro 31-8220 added directly to the in vitro kinase assays at identical or higher concentrations to those used previously in intact cells did not inhibit PKD3 activity stimulated by PDBu in intact cells (Fig. 5D). Thus, the observed in vivo inhibition of PKD3 catalytic activity in response to GF I and Ro 31-8220 was not direct but rather because of the inhibition of the catalytic activity of PKC.
PKD3 Activation by Novel PKCs-Next, we examined whether novel PKC isoforms activate PKD3 by contransfecting COS-7 cells with plasmids encoding GFP-PKD3 and the constitutively active PKC⑀* or PKC*. The cells were lysed 48 h post-transfection and GFP-PKD3 was immunoprecipitated from the lysates using the GFP antiserum. The immunocomplexes were subjected to in vitro kinase assays to measure PKD3 activity by autophosphorylation. As shown in Fig. 6A, upper panel, the constitutively active PKC⑀* or PKC* induced the activation of PKD3 as measured by its autokinase activity. This activation is comparable with the one induced by PDBu alone. The addition of PDBu to the cultures expressing GFP-PKD3 and the constitutively active PKC* or PKC⑀* further enhanced the catalytic activity of PKD3. In contrast, no PKD3 activation was induced in the cells cotransfected with GFP-PKD3 and constitutively active PKC* (see Fig. 6A, upper panel). No basal kinase activity was detected in control COS-7 cells expressing GFP and constitutively active PKC* or PKC⑀* (data not shown).
We also found that in COS-7 cells, constitutively active PKC* or PKC⑀* mediated the activation loop phosphorylation of PKD3 as revealed by Western blot analysis using the Aloop antibody (Fig. 6A, center panel). The same membrane was stripped and subsequently probed with an antibody against GFP to verify the expression and equal loading of GFP-PKD3 (Fig. 6A, bottom panel).
Previous results from our laboratory demonstrated that PKC, in addition to mediating the activation of PKD, forms a complex with PKD via the regulatory region of PKD (13). In view of the differences between the PH domains of PKD and PKD3 (Table I), we examined whether PKD3 interacts with PKC. Panc-1 cells were cotransfected with pGFP-PKD3 and pCO2 (empty vector) or the plasmid encoding the constitutively active form of the novel PKC*. Panc-1 cells were also cotransfected with pGFP-PKD and pCO2 (empty vector) or the plasmid encoding the constitutively active PKC* for comparison because the interaction between PKD and PKC has been previously characterized. The transfected cells were lysed 48 h posttransfection and GFP-PKD or GFP-PKD3 in the extracts were immunoprecipitated using antiserum specific for GFP, PKD, or PKD3. The immunocomplexes were subjected to in vitro kinase assays to measure PKD and PKD3 activity by autophosphorylation (Fig. 6B). GFP or PKD immunoprecipitates from GFP-PKD/PKC*-cotransfected cells showed two prominent phosphorylated bands: one corresponding to GFP-PKD (Ϸ140 kDa) and another corresponding to PKC* (Ϸ80 kDa), in agreement with previous results (13). In striking contrast, GFP or PKD3 immunoprecipitates from PKD3/PKC*-cotransfected cells did not show any other band in addition to that corresponding to phosphorylated GFP-PKD3 (Ϸ130 kDa) (Fig. 6B). Western blot analysis was used to corroborate that the 80-kDa band corresponded to PKC* and that the levels of expression of this protein and the GFP-tagged PKDs were similar in all conditions (data not shown). These results identify an important difference between PKD and PKD3 in their ability to form a molecular complex with PKC*.
PKD3 Continuously Shuttles between the Cytoplasm and the Nucleus-The observed intracellular distribution of PKD3 could result from its continuous transport between the nucleus and the cytoplasm. If this interpretation is correct, interference with the nuclear export machinery should further increase the nuclear accumulation of GFP-PKD3.
LMB is an antifungal antibiotic (71) that inhibits the formation of complexes consisting of CRM1, RanGTP, and proteins containing a leucine-rich nuclear export signal, thereby blocking CRM1-dependent nuclear export (72)(73)(74)(75). Consequently, we decided to use LMB to determine whether the nuclear localization of GFP-PKD3 was CRM1-dependent and therefore the result of its continuous shuttling between the cytoplasm and nucleus. Panc-1 cells transiently transfected with pGFP-PKD3 were incubated with LMB (10 ng/ml) for 1 h and the distribution of GFP-PKD3 analyzed by imaging live cells. As shown in Fig. 7A, an increase in the nuclear accumulation of GFP-PKD3 was detected in response to LMB. This result strongly supports the conclusion that GFP-PKD3 is continuously shuttling between the nucleus and the cytoplasm and that this shuttling requires a competent CRM1 nuclear export pathway. LMB did not induce any detectable nuclear accumulation of GFP (Fig. 7A).
Effect of PKD3 Activation in Its Nuclear Transport-To determine whether NT-mediated activation of PKD3 increases the rate of its nuclear import, we blocked the CRM1-dependent nuclear export pathway with LMB and compared the relative nuclear accumulation of PKD3 in non-stimulated versus NTstimulated Panc-1 cells. The addition of NT to LMB-treated Panc-1 cells did not induce any significant change in the intracellular distribution of GFP (Fig. 7A). However, NT induced an increase in the rate of nuclear import of GFP-PKD3 as indicated by its more rapid nuclear accumulation when compared with non-stimulated cells in the presence of LMB (Fig. 7A). The results imply that the nuclear transport of PKD3 is a regulated process associated to its activation. Therefore, we hypothesized that constitutively active PKCs that activate PKD3 should also modify its nuclear transport without cell stimulation.
Panc-1 cells were cotransfected with plasmids encoding GFP-PKD3 and active PKC⑀* or PKC* and the distribution of GFP-PKD3 was analyzed in fixed cells. As our results shown, the constitutively active PKC⑀* or PKC* induced a dramatic nuclear accumulation of GFP-PKD3 (Fig. 7B). No further nuclear accumulation of GFP-PKD3 than the one detected in cells cotransfected with the empty vector pCO2 (data not shown) FIG. 7. PKD3 nuclear transport regulation. A, Panc-1 cells were transfected with constructs encoding GFP or GFP-PKD3 and incubated at 37°C for 18 h. LMB (10 ng/ml) was added to one set of cultures that were further incubated at 37°C. The other set of cultures was challenged with NT (50 nM) for 10 min in the presence of LMB (10 ng/ml) and after replacing the media containing LMB and NT with media containing only LMB, the cultures were further incubated at 37°C. Representative images of living cells were captured 1 h later, for both conditions, as described under "Experimental Procedures." Bar, 10 m. B, Panc-1 cells were transiently transfected with pCO2 containing the cDNAs encoding the constitutively active PKC⑀*, PKC*, or PKC* (left panels) or simultaneously transfected with those vectors and pGFP-PKD3 as indicated (center and right panels). After 48 h, an epifluorescent microscope was used to visualize the distribution of the constitutively active PKCs and GFP-PKD3 in fixed cells stained with polyclonal antibodies against PKC⑀, PKC, or PKC and Texas Red-conjugated goat anti-rabbit immunoglobulins. Representative images were captured as described under "Experimental Procedures." C, Panc-1 cells were transfected with constructs encoding GFP-CRD, FLAG-PKD3, or cotransfected simultaneously with pGFP-CRD (1 g of DNA) and pFLAG-PKD3 (100 ng of DNA) and incubated at 37°C for 18 h. An epifluorescent microscope was used to visualize the distribution of GFP-CRD and FLAG-PKD3 in fixed Panc-1 cells stained with a monoclonal anti-FLAG antibody and Texas Red-conjugated goat anti-mouse immunoglobulins. Arrows indicate nuclei. Representative images were captured as described under "Experimental Procedures." Bar, 10 m. was observed in the cells expressing the constitutively active atypical PKC*, despite its nuclear presence (Fig. 7B). The expression of GFP-PKD3 did not affect the intracellular distribution of any of the constitutively active PKCs (compare left and center panels in Fig. 7B). Similar results were obtained in living or fixed COS-7 cells coexpressing GFP-PKD3 and PKC⑀* or PKC* (data not shown). No change in the intracellular distribution of GFP was detected in control Panc-1 cells expressing GFP and PKC* or PKC⑀* (data not shown). These results demonstrated that the rate of nuclear transport for PKD3 is dramatically modified as a result of its activation via NT or constitutively active novel PKCs.
A Nuclear Transport Receptor Mediates the Nuclear Import of PKD3-One of the functional characteristics of nuclear transport receptors is the saturability of signal recognition (76). We previously employed this criterion to block the nuclear import of PKD via its CRD, because this PKD domain is responsible for its nuclear import (63). Taking into consideration that the CRD of PKD and PKD3 are highly homologous (see Table I), we hypothesized that the CRD of PKD should block the nuclear import of PKD3 by interfering with its nuclear import receptor.
Consequently, a protein consisting of GFP fused to the NH 2 terminus of the CRD of PKD (GFP-CRD) (63) was coexpressed with FLAG-PKD3 to block the nuclear import of PKD3. GFP-CRD was distributed throughout the cytosol and the nuclei of Panc-1 cells with a distinct accumulation in the plasma membrane and Golgi, in agreement with our previous observations (63), whereas FLAG-PKD3 was predominantly located in the nuclear compartment (Fig. 7C, arrows). The coexpression of GFP-CRD and FLAG-PKD3 (ratio 10:1) did not produce any major changes in the distribution of GFP-CRD. However, the nuclear import of FLAG-PKD3 was prevented, as revealed by its absence in the nuclear compartment (Fig. 7C). These results suggested that the CRD of PKD successfully competed with FLAG-PKD3 for the nuclear import receptor that delivers PKD3 to the nucleus and they support the conclusion that the nuclear import of PKD3 utilizes the same import machinery as PKD. Coexpression of GFP, which uniformly distributes in the cytoplasm and nuclei of Panc-1 cells did not interfere with the nuclear accumulation of Flag-PKD3 (data not shown).
Intracellular Dynamic Properties of PKD3-Our preceding results strongly suggested that PKD3 shuttles between the cytoplasm and the nucleus. To further support this conclusion and to study the dynamic behavior of PKD3 in the nucleus and in the cytoplasm we used FLIP.
Panc-1 cells were transiently transfected with the plasmid encoding GFP-PKD3 and analyzed by FLIP in live cells as indicated under "Experimental Procedures." As illustrated by the images presented in Fig. 8, GFP-PKD3 was distributed throughout the cytosol and the nucleus before photobleaching. FLIP was performed either in the cytoplasmic or nuclear compartments, within the restricted areas indicated by rectangles. After FLIP, the total cell fluorescence was followed over time.
Bleaching the GFP-PKD3 present in the cytoplasm (Fig. 8, top panel) resulted in almost complete disappearance of fluorescent signal from this compartment by 2 min. This result showed that most of the GFP-PKD3 cytoplasmic molecules passed through the area being bleached within 2 min suggesting a rapid and random movement. The recovery of fluorescent signal over the cytoplasm was relatively slow, very likely because of the free diffusion barrier imposed by the nuclear envelope to the GFP-PKD3 molecules inside the nuclear compartment. By 12 min a significant amount of GFP-PKD3 was detectable in the cytoplasm, which coincided with a drop in the intensity of the fluorescent signal over the non-bleached nu-clear compartment, indicating transfer of GFP-PKD3 from nuclei into the cytoplasm.
Bleaching GFP-PKD3 present in the nucleus resulted in the complete disappearance of fluorescent signal from this compartment by 30 s, in accordance with the smaller volume of the nucleus, followed by partial recovery of fluorescent signal over the bleached area by 9 min (Fig. 8, bottom panels). The spontaneous recovery after bleaching of the whole cell was less than 3% of total initial fluorescence (data not shown), which coincides with previous reports indicating that spontaneous recovery of bleached GFP did not contribute to signals (77)(78)(79). Unbleached control cells, adjacent to the photobleached ones, showed no fluorescence changes during photobleaching, indicating that there was no generalized bleaching effect during imaging (see Fig. 8). Thus, any fluorescent signal increase detected after the restricted photobleaching within a compartment was the result of non-bleached GFP-PKD3 molecules entering the other non-bleached compartment. These results demonstrate that in non-stimulated cells GFP-tagged PKD3 continuously shuttles between the cytoplasm and nucleus and indicates that PKD3 moves rapidly within each compartment. DISCUSSION The results presented in this study demonstrate that the catalytic activity of PKD3 can be rapidly increased by treatment of intact cells with tumor promoting phorbol esters as well as GPCR agonists. The activation of PKD3 coincided with the phosphorylation of Ser 731 and Ser 735 located in its activation loop, a region within the kinase domain that is identical in all PKD isoenzymes (36). Several lines of evidence indicate that PKD3 activation is mediated through a PKC-dependent pathway. The catalytic activity of PKD3 as well as its activation loop phosphorylation was inhibited in intact cells by the PKC inhibitors GF I or Ro 318220. However, the same inhibitors failed to suppress the in vitro kinase activity of PKD3, indicating that the inhibition of PKD3 activation induced by these compounds in intact cells was not direct but rather because of the inhibition of the catalytic activity of PKC. Interestingly, GF I or Ro 318220 have been shown to inhibit the in vitro kinase activity of PKD2 (50) but not PKD (31), which suggest a closer functional homology between the kinase domains of PKD and PKD3. In addition, the activation of PKD3 was prevented by down-regulating the expression of PKC. 3 Evidence indicating that novel isoforms of PKC can act as upstream activators of PKD3 was obtained by activating PKD3 by coexpression of this enzyme with constitutively active PKC* or PKC⑀* in the absence of cell stimulation. Therefore, PKC not only can mediate the activation of PKD (13) but also PKD3. However, PKD3 does not form a stable interaction with PKC such as PKD, a result that indicates an interesting difference between these enzymes. The lack of a lasting interaction between PKD3 and PKC very likely is related to the differences between the PH domains of PKD and PKD3, because the PKD PH domain mediates the interaction with PKC (13). Thus, several different lines of experimental evidence support the conclusion that PKD3 activation in response to phorbol esters or GPCR agonists is mediated through a PKC-mediated signal transduction pathway.
There are multiple related PKC isoforms, which are increasingly recognized to exhibit distinct expression patterns and spatio-temporal distribution in response to signaling events. An important role of PKCs in PKD activation and intracellular distribution has been documented in different cell types stimulated by tumor promoting phorbol esters, neuropeptides, and growth factors (27, 29 -33, 35, 41, 63). It is tempting to speculate that the differential intracellular distribution of PKCs may confer separate signaling properties on them, such as the ability to mediate the activation of differently localized PKD isoenzymes. On this regard, the present study revealed a notable difference between the intracellular distribution of PKD, PKD2, and PKD3.
The steady-state distribution of PKD3 in non-stimulated fibroblasts or epithelial cells was to a large extent nuclear, in contrast to the predominantly cytoplasmic localization of PKD (41) and PKD2 (51). A report published when this article was under review indicated that PKD3 was predominantly cytoplasmic in B-cells (80). This difference suggests that the preferential localization of PKD3 to cellular compartments, especially to the nucleus, may be dependent on the cell type.
GPCR agonists that activate PKDs induced a rapid and reversible plasma membrane translocation of PKD3. PKC inhibitors did not prevent the translocation from the cytosol to the plasma membrane of PKD3 in response to NT. However, they substantially retarded its rapid reverse translocation from the plasma membrane to the cytosol, indicating that the plasma membrane dissociation of PKD3 is coupled to their PKC-mediated activation. We conclude based on these and our previous results (41,51) that PKCs not only play a critical role in the activation of all members of the PKD family but also are required to trigger their rapid reverse translocation from the plasma membrane.
Our FLIP results indicates that all GFP-PKD3 molecules move through all locations in the cytoplasm or nucleus within minutes suggesting that no "hard-wired" association exist for PKD3 in non-stimulated cells. In addition to these movements, PKD3 is continuously shuttling between the cytoplasm and the nucleus by a mechanism that requires a nuclear import receptor and a competent CRM1 nuclear export pathway. However, the constant presence of PKD3 in the nucleus of non-stimulated cells indicates that the regulation of its intracellular distribution is different from those of PKD and PKD2, considering that both of these kinases are predominantly cytoplasmic (41,51).
Several proteins that continuously shuttle between the cytoplasm and the nucleus are subject to regulated nuclear trans-port in response to specific cellular signals, including phosphorylation (81)(82)(83). These signals can modify the relative rates of nuclear import and export of these proteins and consequently, their subcellular localization. Interestingly, the activation of PKD3 mediated by NT or the constitutively active PKC⑀* or PKC* was associated with a dramatic increase in the nuclear accumulation of GFP-PKD3 and therefore, to a distinct cellular redistribution of this kinase.
The results presented in this study suggest an attractive model that explains the regulation of the catalytic activity and intracellular distribution of PKD3 in response to GPCR agonists (Fig. 9). In this model, the steady-state distribution of inactive PKD3 results from a rate of nuclear import that exceeds its rate of nuclear export. Upon GPCR agonist binding, cytoplasmic PKD3 moves rapidly to the plasma membrane in response to an increase in the DAG concentration in that compartment. Concurrently, PKC is activated and recruited to the plasma membrane. Now, active PKC simultaneously mediates the activation/phosphorylation and plasma membrane dissociation of PKD3 (Fig. 9). Active PKD3 then migrates first to the cytosol and subsequently into the nucleus with a nuclear transport rate that favors its import over its export. In this manner, a signal originating at the plasma membrane is transduced into the nucleus, which might be important for coordinating nuclear and cytoplasmic events in a simple, reversible and rapid mode. Considering the now compelling evidence that intranuclear lipid pathways exist (84 -89) and that the PKD3 can be stimulated by DAG, this model of PKD3 regulation can include, in certain cell types, the participation of intranuclear lipid pathways.
It is tempting to speculate that the differential intracellular distribution and propensity of PKD isoenzymes to form lasting 3 O. Rey and E. Rozengurt, unpublished results.
FIG. 9. Model of PKD3 activation and intracellular distribution regulation. In unstimulated cells, inactive PKD3 is continuously shuttling between the cytoplasm and the nucleus at different rates, i.e. faster nuclear import than export. After agonist binding to its GPCR, phospholipase C-mediated hydrolysis of phosphatidylinositol 4,5bisphosphate produced DAG at the plasma membrane, which in turn mediates the translocation of inactive PKD3 from the cytosol to that cellular compartment. DAG activates PKC, which now mediates the activation and phosphorylation of PKD3. Active PKD3 dissociates from the plasma membrane and migrates first to the cytosol and then into the nucleus at a faster rate than in non-stimulated cells. The activation and dynamic intracellular distribution of PKD3 can be interrupted by pharmacological agents that inhibit PKC activation or CRM1-mediated nuclear export. Arrows representing potential pathways leading to the inactivation of PKD3, via phosphatase activity, were not included for clarity purposes. PKC*, active PKC; PIP 2 , phosphatidylinositol 4,5biphosphate; PKD3 c , inactive cytosolic PKD3; PKD3 n , inactive nuclear PKD3; PKD3 pm , inactive plasma membrane PKD3; PKD3* c , active phosphorylated cytosolic PKD3; PKD3* n , active phosphorylated nuclear PKD3; PLC, phospholipase C. Arrow direction and thickness represent the directionality and differential rates of transport of PKD3. complexes with specific PKC isoforms are characteristics associated with the ability of each PKD to execute unique functions at different subcellular locations. In this way, a similar mechanism of activation can potentially generate diverse physiological responses based on the unique redistribution of closely related signaling molecules.