- Email: [email protected]
A possible role for p190RhoGAP in PKCε-induced morphological effects
Cellular Signalling 16 (2004) 245 – 252 www.elsevier.com/locate/cellsig
A possible role for p190RhoGAP in PKCq-induced morphological effects Ulrika Trolle´r, Arathi Raghunath, Christer Larsson * Lund University, Department of Laboratory Medicine, Molecular Medicine, Entrance 78, 3rd Floor, Malmo¨ University Hospital, UMAS, SE-205 02 Malmo¨, Sweden Received 27 May 2003; received in revised form 28 July 2003; accepted 28 July 2003
Abstract We have previously shown that protein kinase C (PKC) q induces neurite outgrowth via its regulatory domain. This is accompanied by PKC-induced stress fibre loss. Here, we show that the regulatory domain (RD) of PKCq induces processes also in NIH-3T3 fibroblasts, similar to what has been observed with p190RhoGAP. This study also shows that p190RhoGAP induces neurite outgrowth in SK-N-BE(2) neuroblastoma cells. We therefore investigated whether p190RhoGAP may be downstream of PKCq. We could detect a co-localization of p190RhoGAP and PKCq at membrane ruffles and an increased association between the proteins in fibroblasts treated with 12-Otetradecanoylphorbol-13-acetate (TPA). The association is also seen in neuroblastoma cells, and nerve growth factor (NGF) treatment of SHSYSY/TrkA cells decreases the interaction. However, overexpressed PKCq did not coprecipitate overexpressed p190RhoGAP in CHO cells, indicating that the proteins do not interact directly. This raises the possibility that p190RhoGAP is involved in mediating the morphological effects of PKCq. D 2003 Elsevier Inc. All rights reserved. Keywords: Protein kinase C; p190RhoGAP; Neurite; Neuroblastoma; Fibroblasts
1. Introduction The outgrowth and guidance of neurites are regulated at the growing end of the neurite, the growth cone, which senses extracellular cues from the surrounding cells and the extracellular matrix. The intracellular pathways conveying the external signals involve an interplay between several proteins. One family of proteins that is likely to be of key importance for the control of the growth cone advance and guidance is the Rho family of small GTPases. The members of the Rho family regulate a variety of cellular processes including the rearrangement of the actin cytoskeleton, which is necessary for the dynamics in the neuronal growth cone. Active Rac1 and Cdc42 have been shown to be important for the growth cone advance by inducing the formation of lamellipodia and filopodia, respectively, while active RhoA, which regulates the formation of stress fibers and focal adhesions, has been shown to have the opposite effect and
Abbreviations: CD, catalytic domain; EGFP, enhanced green fluorescent protein; FL, full-length; NGF, nerve growth factor; PKC, protein kinase C; RD, regulatory domain; TPA, 12-O-tetradecanoylphorbol-13acetate. * Corresponding author. Tel.: +46-40-337404; fax: +46-40-337322. E-mail address: [email protected] (C. Larsson). 0898-6568/$ - see front matter D 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0898-6568(03)00135-9
cause neurites to retract [1 – 3]. The Rho proteins cycle between an active GTP-bound state and an inactive GDPbound state, and the switch between these two states is mediated by three classes of proteins. Guanine nucleotide exchange factors (GEFs) stimulate the dissociation of GDP which allows GTP to bind and activate the protein, guanine nucleotide dissociation inhibitors (GDIs) bind Rho-GDP and thereby maintain it inactive, and GTPase-activating proteins (GAPs) enhance the intrinsic GTPase activity of the Rho GTPases and thereby inactivate them [4]. p190RhoGAP is a GTPase-activating protein for the Rho family and acts specifically on RhoA [5]. p190RhoGAP has been shown to be expressed at high levels in the developing nervous system, and p190RhoGAP / mice have defects in axon outgrowth, guidance, and fasciculation and exhibit impaired neuronal development [6]. The resulting defects include a complete lack of corpus callosum and other forebrain midline structures, disorganized layering of the cerebral cortex, and defective closure of the neural tube [7]. Protein kinase C (PKC) is another family involved in a multitude of cellular processes, including the regulation of neurite outgrowth [8 – 12]. It is a family of serine/threonine kinases consisting of at least 10 isoforms, which are grouped in three subfamilies based on their requirements for activation. Conventional PKCs (a, hI, hII, and g) are
246
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
activated by Ca2 +, diacylglycerol, and phosphatidylserine. Novel PKCs (y, q, D, and u) are Ca2 +-independent and atypical PKCs (L/E and ~) are independent of both Ca2 + and diacylglycerol [13,14]. PKC has been shown to directly phosphorylate p190RhoGAP in vitro, and treatment of fibroblasts with the PKC activator 12-O-tetradecanoylphorbol-13-acetate (TPA) leads to increased phosphorylation of p190RhoGAP and a translocation of p190RhoGAP from the cytoplasm to the membrane [7]. We have previously shown that overexpression of PKCq induces the formation of neurites in neural cells [11]. This is independent of the catalytic activity, and the regulatory domain (RD) is sufficient for the effect. The PKC-induced formation of neurites has also been shown to be accompanied by a loss of stress fibres, and the neurite outgrowth could be blocked by overexpressing constitutively active RhoA, indicating a suppression of RhoA activity by PKC [15]. This led us to speculate that p190RhoGAP may be one component that mediates PKC effects on RhoA activity and on cellular morphology. This study was designed to elucidate a putative role for p190RhoGAP in PKCq-induced morphological changes.
2. Material and methods 2.1. Plasmids Expression vectors encoding PKC domains fused to the enhanced green fluorescent protein (EGFP) have previously been described [11]. Full-length PKCq and PKCq-PSC1V3 cDNA were cut out from the pEGFP-N1 vector (Clontech) by digestion with BamHI and XhoI and ligated into pcDNA4/ myc-His vector (Invitrogen). pKH3 expression vector encoding HA-tagged p190RhoGAP was kindly provided by Dr. I. Macara [16].
between the plasmids in the co-transfection studies were in SK-N-BE(2) cells 1 (pEGFP-N1) to 5 (pKH3 with insert), and in the CHO cells, equal amounts of plasmids were used. When indicated, 12-O-tetradecanoylphorbol-13-acetate (TPA; Sigma) was used at a concentration of 16 nM. For differentiation studies, cells were treated with 100 ng/ml nerve growth factor (NGF; Promega). 2.3. Immunofluorescence After transfection NIH-3T3 and SK-N-BE(2) cells were incubated for 16 h before fixation in 4% paraformaldehyde in PBS for 4 min. For detection of F-actin, cells were incubated for 20 min with Alexa Fluor 546-conjugated phalloidin (Molecular Probes) diluted 1:100 in TBS. For detection of p190RhoGAP and PKCq, cells were permeabilized and blocked with 5% goat serum and 0.3% Triton X100 in TBS for 30 min and thereafter incubated for 1 h with primary mouse anti-p190RhoGAP (Transduction Laboratories) antibody or primary rabbit anti-PKCq (Santa Cruz) antibody, diluted 1:50 in TBS. Following washes, cells were incubated for 1 h with secondary antibodies Alexa Fluor 546- or 488-conjugated goat anti-mouse IgG or Alexa Fluor 546 goat anti-rabbit IgG (Molecular Probes), diluted 1:400 in TBS. Cover slips were mounted on object slides using 20 Al PVA-DABCO (9.6% polyvinyl alcohol, 24% glycerol, and 2.5% 1,4-diazabicyclo[2.2.2]octane in 67 mM Tris – HCl, pH 8.0). 2.4. Morphology studies Transfected cells were examined with fluorescence microscopy and were considered to have long processes if the length of the process was more than the length of two cell bodies. Two hundred transfected cells were counted in each experiment. Digital images were captured with a Sony DKC 5000 camera system.
2.2. Cell culture 2.5. Confocal microscopy Human neuroblastoma SK-N-BE(2) and SH-SY5Y/TrkA [17] cells were maintained in Minimal Essential Medium (Gibco), and mouse NIH-3T3 fibroblasts and Chinese Hamster Ovarian (CHO) cells were cultured in RPMI 1640 (Gibco). All mediums were supplemented with 10% foetal bovine serum, 100 IU/ml penicillin, and 100 Ag/ml streptomycin (Gibco). For morphology studies, cells were trypsinized and seeded at a density of 100,000 cells per 35-mm cell culture dish on glass cover slips. For immunoprecipitation studies, cells were trypsinized and seeded at a density of 1 106 cells per 100-mm cell culture dish or 500,000 cells per 60-mm cell culture dish. After 24-h incubation, cells were transfected in serum-free medium. For 35-mm dishes, 2 Ag DNA and 4 Al Lipofectamine (Gibco) in 1 ml medium were used, for 60-mm dishes, 4 Ag DNA and 8 Al Lipofectamine in 2 ml medium were used, and for 100-mm dishes, 6 Ag DNA and 12 Al Lipofectamine in 3 ml medium were used. The ratio
The subcellular localization of endogenous proteins was examined using a Bio-Rad Radiance 2000 confocal system fitted on a Nikon microscope with a 60 /NA 1.40 oil lens. Excitation wavelengths were 488 (Alexa Flour 488) and 543 nm (Alexa Flour 546), and the emission filters used were HQ515/30 (Alexa Flour 488) and 600LP (Alexa Flour 546). Co-localization was analysed with the LaserPix software (Bio-Rad). 2.6. Immunoprecipitation Cells were washed twice in PBS and lysed in RIPA buffer (10 mM Tris –HCl, pH 7.2, 1% Triton X-100, 1% sodium deoxycholate, 160 mM NaCl, 0.1% sodium dodecyl sulphate, 1 mM EGTA, 1 mM EDTA, 1 mM sodium orthovanadate, complete protease inhibitor cocktail [Roche]) for 10 min on
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
247
ice. The lysates were clarified by centrifugation at 18,000 g for 10 min at 4 jC, and incubated with 2 Ag of antibodies toward the HA epitope (Clontech), p190RhoGAP (Upstate Biotechnology), or the myc epitope (Oncogene) for 2 h with rotation at 4 jC. A 20-Al sample of GammaBind Plus Sepharose (Amersham Pharmacia) beads was added prior to an additional 1-h incubation. The immune complexes were recovered by centrifugation at 500 g for 30 s at 4 jC and washed four times in lysis buffer without protease inhibitors and resuspended in sample buffer. 2.7. Ni-NTA purification Cells were washed twice in PBS and lysed in lysis buffer (50 mM NaH2PO4, 300 mM NaCl, 10 mM imidazole, 0.05% Tween-20, 1% Triton X-100, 1 mM sodium orthovanadate, complete protease inhibitor cocktail [Roche]) on
Fig. 2. p190RhoGAP induces neurites in neuroblastoma cells. SK-N-BE(2) cells were co-transfected with empty pEGFP-N1 vector and expression vector encoding HA-tagged p190RhoGAP or empty vector encoding the HA-tag only. Sixteen hours after transfection, cells were fixed and mounted on object slides. (A) Fluorescent images of SK-N-BE(2) cells, visualized with EGFP fluorescence, transfected with empty vector or p190RhoGAP expression vector, showing a cell with neurites. (B) The percentage of transfected cells with neurites was quantified in cells expressing the HA-tag only or p190RhoGAP. Data (mean F S.E.M., n = 3) are percentage of transfected cells with neurites. (C) SK-N-BE(2) cells were transfected with either empty vector or vector encoding HA-tagged p190RhoGAP. Sixteen hours after transfection, cells were lysed and immunoprecipitated with antibodies towards the HA-tag. The immune complexes were analysed with Western blotting using antibodies toward the HA-tag.
ice for 30 min with gentle rocking. The lysates were clarified by centrifugation at 18,000 g for 10 min at 4 jC. A 450-Al sample of cleared lysate was incubated with 20 Al of nitrilotriacetic acid (Ni-NTA) agarose bead suspension (Qiagen) for 1 h with rotation at 4 jC. The Ni-NTA beads were collected by centrifugation at 500 g for 1 min at 4 jC, washed three times in wash buffer (50 mM NaH2PO 4, 300 mM NaCl, 20 mM imidazole, 0.05% Tween-20, 1 mM sodium orthovanadate, complete protease inhibitor cocktail [Roche]), and eluted by incubation for 15 min in elution buffer (50 mM NaH2PO4, 300 mM NaCl, 250 mM imidazole, 0.05% Tween-20) at room temperature. Fig. 1. Overexpression of PKCq induces beaded extensions via its regulatory domain in NIH-3T3 fibroblasts. NIH-3T3 fibroblasts were transfected with empty pEGFP-N1 vector or vectors encoding full-length PKCq (PKCq-FL), the PKCq regulatory domain (PKCq-RD), the PKCq catalytic domain (PKCq-CD) or full-length PKCy fused to EGFP. Sixteen hours after transfection, cells were fixed and mounted on object slides. (A) Fluorescent images of NIH-3T3 cells expressing EGFP, PKCq-FL, PKCqRD, and PKCq-CD showing the beaded extensions induced by PKCq-FL and PKCq-RD. (B) Quantification of the number of transfected cells with beaded extensions longer than two cell bodies. Data (mean F S.E.M., n = 3) are percentage of transfected cells with beaded extensions.
2.8. Immunoblotting Proteins were electrophoretically separated with SDSPAGE and transferred to a PVDF membrane (Millipore). The membranes were immunoblotted with polyclonal antibodies towards PKCa, hII, y, and q (Santa Cruz), the HA epitope (Clontech); monoclonal antibodies toward p190RhoGAP (Transduction Laboratories), the myc epitope (Oncogene), or the EGFP-tag (BabCO). Membranes were
248
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
thereafter incubated with horseradish peroxidase-linked secondary antibodies (Amersham), and detection was done with the SuperSignal system (Pierce) as substrate. The chemiluminescence was detected with a CCD camera (Fujifilm). Quantification of the bands was performed with the ImageGauge software (Fujifilm). The values were obtained by quantifying the intensity of the bands, subtracting the background, and normalizing to the values obtained from the loading control.
3. Results 3.1. PKCe induces a beaded extension phenotype in NIH3T3 fibroblasts We have previously shown that PKCq, via its regulatory domain, induces neurites in neuronal cells [11,18]. It has also been demonstrated that PKCq can trigger the outgrowth of beaded extensions from NIH-3T3 fibroblasts [19]. To investigate whether these effects can be related, we tested if PKCq, through its regulatory domain, could induce the extensions in fibroblasts. Accordingly, NIH-3T3 fibroblasts were transfected with vectors coding for EGFP-tagged fulllength PKCq (PKCq-FL), the PKCq regulatory domain (PKCq-RD), the PKCq catalytic domain (PKCq-CD), or
full-length PKCy (Fig. 1). We found that not only overexpression of the full-length PKCq induced profound morphological effects with the formation of beaded extensions (Fig. 1A), but also that the regulatory domain of PKCq is sufficient for this effect. This is in line with effects of PKCq on the induction of neurites seen in neural cell lines [11]. Quantification of the morphological effects by counting cells with processes longer than the length of two cell bodies (Fig. 1B) showed that PKCq-FL and PKCq-RD induced beaded extensions in 61% and 74%, respectively, of the transfected cells, while PKCq-CD had no processinducing effect. Overexpression of PKCy also induced beaded extensions, but not to the same degree as seen with PKCq.
3.2. Overexpression of p190RhoGAP in neuroblastoma cells induces neurites The PKCq-induced phenotype with beaded extensions and a rounded cell body has previously been seen in NIH3T3 cells overexpressing p190RhoGAP [16]. We therefore investigated if p190RhoGAP had a neurite-inducing effect in neuroblastoma cells. SK-N-BE(2) cells were co-transfected with vectors encoding HA-tagged p190RhoGAP with an empty pEGFP-N1 vector to visualize transfected cells.
Fig. 3. Co-localization of p190RhoGAP and PKCq at membrane ruffles. NIH-3T3 fibroblasts were stimulated with TPA for 15 min, fixed, and immunostained for PKCq (A and D) or p190RhoGAP (B and G). F-actin was visualised with Alexa Fluor 546-conjugated phalloidin (E and H). The cover slips were thereafter mounted on object slides and studied by confocal microscopy. The merged images show co-localization (white pixels) between PKCq and p190RhoGAP (C), PKCq and F-actin (F), and p190RhoGAP and F-actin (I).
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
249
(Fig. 3G –I). Thus, PKCq and p190RhoGAP appear to colocalize at membrane ruffles in TPA-treated fibroblasts. 3.4. Endogenous protein interaction between PKCe and p190RhoGAP To investigate if the co-localization was due to an association between PKCq and p190RhoGAP, we stimulated fibroblasts with TPA for different time periods and immunoprecipitated p190RhoGAP followed by immunoblot towards PKCq. As seen in Fig. 4, endogenous PKCq can be co-immunoprecipitated with p190RhoGAP, and this complex formation is further enhanced by stimulation with TPA. To explore if this complex had a potential role in neurite outgrowth, we investigated if the interaction also was seen in neuroblastoma cells. Cell extracts from SK-N-BE(2) cells were immunoprecipitated with antibodies towards p190RhoGAP, and the precipitates were thereafter analysed with antibodies towards different PKC isoforms. SK-N-BE(2) cells express the PKC isoforms a, hI, hII, y, and q [11], but since the hI isoform is mostly located in the nucleus [20], we
Fig. 4. Endogenous association of p190RhoGAP and PKCq in fibroblasts. (A) NIH-3T3 fibroblasts untreated or stimulated with 16 nM TPA for 15 or 60 min were lysed and immunoprecipitated for p190RhoGAP. The immune complexes were analysed with Western blotting using antibodies towards PKCq, or p190RhoGAP as control. (B) Quantification of the results from seven separate experiments, presenting the intensity of the PKCq bands normalized to the intensity of the p190RhoGAP band, expressed as percent of the value obtained in unstimulated cells.
As seen in Fig 2A, 15% of p190RhoGAP-overexpressing cells had neurites compared to 6.5% of the control cells. The expression of p190RhoGAP was very low as we could only detect the protein after immunoprecipitation towards the HA-tag (Fig. 2C). 3.3. PKCe and p190RhoGAP co-localize at membrane ruffles Since overexpression of PKCq or p190RhoGAP seems to give rise to the same phenotype, we further wanted to examine a possible interplay between the two proteins. p190RhoGAP has been shown to translocate to the membrane upon stimulation with the PKC-activating phorbol ester TPA, suggesting the PKC may regulate the subcellular localization of p190RhoGAP [7]. We stimulated fibroblasts with TPA for 15 min and studied the subcellular localization of endogenous PKCq and p190RhoGAP. PKCq and p190RhoGAP were found to co-localize at membrane ruffles (Fig. 3A – C). Staining for F-actin (Fig. 3E) showed that there is a co-localization between PKCq (Fig. 3D) and the cortical F-actin (Fig. 3F) in ruffle-like structures. p190RhoGAP also localizes to the cortical actin cytoskeleton in these structures
Fig. 5. The PKCq isoform associates with p190RhoGAP in neuroblastoma cells. SK-N-BE(2) cells were lysed and immunoprecipitated with antibodies towards endogenous p190RhoGAP. The immune complex was divided into four and analysed with Western blotting using antibodies towards PKCa, hII, y and q, or p190RhoGAP as a control.
250
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
excluded this isoform. As seen in Fig. 5, there is an interaction between endogenous p190RhoGAP and PKCq which is not seen with the other PKC isoforms, strengthening our results seen in the fibroblasts. 3.5. Long-term treatment with nerve growth factor decreases the interaction between PKCe and p190RhoGAP Thus, TPA leads to increased association of PKCq and p190RhoGAP in fibroblasts, and this interaction can also be detected in untreated neuroblastoma cells. Given the putative roles for these proteins in neurite outgrowth, we wanted to analyze what happens to this association during neuronal differentiation. SH-SY5Y/TrkA neuroblastoma cells were stimulated for 3 days with nerve growth factor (NGF), which elicits neuronal differentiation in these cells [17]. Thereafter, immunoprecipitation towards p190RhoGAP was performed, and bound PKCq was detected by immunoblotting. The result showed that the interaction between the proteins seemed to decrease upon long-term treatment with NGF (Fig. 6). 3.6. Overexpressed PKCe does not co-immunoprecipitate p190RhoGAP To get an indication of whether there is a direct binding between PKCq and p190RhoGAP, the proteins were overexpressed in CHO cells and immunoprecipitated. Cells were co-transfected with vectors encoding HA-tagged p190RhoGAP and myc/His-tagged full-length PKCq (q-FL-myc/His). Immunoprecipitation was thereafter performed against the myc-tag on PKCq. Immunoblotting towards the HA-tag on
Fig. 6. NGF treatment reduces the association of p190RhoGAP and PKCq in TrkA-expressing SH-SY5Y cells. SH-SY5Y/TrkA neuroblastoma cells were treated for 3 days with 100 ng/ml NGF. Cells were thereafter lysed and immunoprecipitated with antibodies towards p190RhoGAP. The immune complexes were analysed with Western blotting using antibodies toward PKCq or p190RhoGAP as a control. The result shown is representative of two separate experiments.
Fig. 7. Overexpressed PKCq and p190RhoGAP do not co-precipitate. (A) CHO cells were co-transfected with vectors encoding HA-tagged p190RhoGAP together with either empty myc/His vector or vector encoding full-length PKCq fused to myc/His. Sixteen hours after transfection, cells were lysed and immunoprecipitated with antibodies towards the myc epitope. The immune complexes were analysed with Western blotting using antibodies toward p190RhoGAP or myc as a control. (B) Fibroblasts were transfected with vector encoding PKCqPSC1V3 fused to myc/His and 16 h after transfection; cells were lysed and purified with Ni-NTA agarose beads. Samples were separated on SDSPAGE and analysed with Western blotting using antibodies against p190RhoGAP, or myc as a control.
p190RhoGAP did not detect proteins, neither in the precipitate nor the lysate (data not shown), indicating low transfection efficiency also in these cells in line with the result seen with the SK-N-BE(2) cells (Fig. 2C). Therefore, we also re-probed the membrane with antibodies towards endogenous p190RhoGAP, but neither in this case could any p190RhoGAP be detected in the precipitate, even though it is present in the lysate (Fig 7A). Immunoblotting towards the myc-tag shows that PKCq is precipitated. In another approach, we tried to precipitate endogenous p190RhoGAP with a short fragment of PKCq encompassing the pseudosubstrate, the C1 domains and the V3 hinge (qPSC1V3), which have been shown to be sufficient for neurite induction in neuroblastoma cells [11]. For this experiment, we used NIH-3T3 fibroblasts since our previous results indicated an interaction between endogenous PKCq and p190RhoGAP in these cells (Fig. 4). Fibroblasts transfected with vectors encoding the myc/His-tagged PKCqPSC1V3 were lysed, and the proteins were purified by
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
adding nickel nitrilotriacetic acid (Ni-NTA) resins (which binds to the His-tag) to the cell extract. The samples were separated by SDS-PAGE and immunoblotted for endogenous p190RhoGAP and the myc-tag as a control (Fig. 7B). The result shows that the Ni-NTA resins efficiently purified myc/His-tagged PKCq-PSC1V3, but that we neither in this case could co-purify p190RhoGAP.
4. Discussion We have previously shown that PKCq induces neurite outgrowth in neural cells via its regulatory domain, independently of its catalytic activity [11,18]. In this study, we sought to determine the mechanism by which PKCq regulates outgrowth of neurites, especially the effects on the actin cytoskeleton associated with the outgrowth. For this purpose, we have also utilized fibroblasts, which have been used extensively to study the regulation of the Rho family of small GTPases and are considered to be key regulators of the actin cytoskeleton [21]. We show here that overexpression of the regulatory domain of PKCq in NIH-3T3 fibroblasts induces a similar morphology as seen in cells overexpressing the full-length PKCq, with the outgrowth of long processes and the rounding up of the cell body [19]. The result correlates with our previous data, which demonstrated that the regulatory domain of PKCq is sufficient to induce neurite outgrowth in neural cells [11]. The fact that PKC induces the formation of long processes even in non-neural cells suggests that PKC targets cytoskeletal regulators that are common for several cell types. Similar morphological changes can also be induced by inhibiting RhoA both in fibroblasts [22] and in neuronal cells [1,3,23]. This data, together with our recent finding that PKC-induced neurite outgrowth is accompanied by a stress fibre dismantling and is counteracted by the RhoA pathway [15], led us to speculate that inactivation of RhoA may be a mechanism through which PKC induces neurites. One candidate to mediate such an effect is p190RhoGAP, since this protein has been shown to induce beaded extensions in fibroblasts [16] and to be putatively regulated by PKC and of importance for normal neuronal development [6,7]. Indeed, overexpression of p190RhoGAP does to a certain extent induce neurites in SK-N-BE(2) neuroblastoma cells. However, the neurite-inducing effect of p190RhoGAP was not as potent as with overexpressed PKCq. This may be due to the barely detectable levels of HA-tagged p190RhoGAP that was consistently obtained in transfected cells. Another explanation could be that PKC also triggers other pathways that contribute to induction of neurite outgrowth. Furthermore, our data show that in TPA-treated fibroblasts there was a co-localization of p190RhoGAP with PKCq at membrane ruffles. By using immunoprecipitation, we also detected an association between PKCq and p190RhoGAP
251
which increased after TPA treatment. This suggests that p190RhoGAP could be downstream of activated PKC in a pathway leading to cytoskeletal changes, such as the outgrowth of cellular processes. The interaction between PKCq and p190RhoGAP could also be detected in SK-N-BE(2) neuroblastoma cells. The fact that it only associated with the PKCq isoform correlates with our previous data which show that only this isoform is capable of inducing neurites in neuroblastoma cells [11,18]. We also detected an association between the proteins in SHSY5Y/TrkA cells. However, this association was decreased upon induction of differentiation with NGF treatment for 3 days. This suggests that the association of PKCq with p190RhoGAP may not be abundant once neurite outgrowth has been elicited, which is the case after NGF treatment for 3 days. The interaction may be most critical during the initiation of neurite outgrowth, which would be in line with previous findings that active RhoA inhibits neurite initiation, but is essential for neurite elongation [24]. However, we could not detect an interaction between PKCq and p190RhoGAP using overexpressed proteins. This indicates that the proteins do not directly interact with a high affinity binding. A putative interaction therefore likely involves other components. In summary, we have shown here that PKCq induces outgrowth of long cellular processes in NIH-3T3 fibroblasts via the regulatory domain, similarly to our previous results with neural cells [11]. Our data further indicate that activated PKCq indirectly interacts with p190RhoGAP at membrane ruffles, which raises the possibility that p190RhoGAP is involved in mediating PKCq-induced cytoskeletal changes.
Acknowledgements We would like to thank Dr. S. Pa˚hlman for providing SHSY5Y/TrkA cells and Dr. I. Macara for providing p190RhoGAP expression vector. This work was supported by the Swedish Cancer Society, the Children’s Cancer Foundation of Sweden, the Swedish Society for Medical Research, the Crafoord; Magnus Bergvall; Gunnar, Arvid and Elisabeth Nilsson; Ollie and Elof Ericsson; and Greta and Johan Kock Foundations, and Malmo¨ University Hospital Research Funds.
References [1] Jalink K, van Corven EJ, Hengeveld T, Morii N, Narumiya S, Moolenaar WH. J Cell Biol 1994;126:801 – 10. [2] Tigyi G, Fischer DJ, Sebok A, Marshall F, Dyer DL, Miledi R. J Neurochem 1996;66:549 – 58. [3] Kozma R, Sarner S, Ahmed S, Lim L. Mol Cell Biol 1997;17: 1201 – 11. [4] Takai Y, Sasaki T, Matozaki T. Physiol Rev 2001;81:153 – 208. [5] Ridley AJ, Self AJ, Kasmi F, Paterson HF, Hall A, Marshall CJ, et al. EMBO J 1993;12:5151 – 60.
252
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
[6] Brouns MR, Matheson SF, Settleman J. Nat Cell Biol 2001;3:361 – 7. [7] Brouns MR, Matheson SF, Hu KQ, Delalle I, Caviness VS, Silver J, et al. Development 2000;127:4891 – 903. [8] Brodie C, Bogi K, Acs P, Lazarovici P, Petrovics G, Anderson WB, et al. Cell Growth Differ 1999;10:183 – 91. [9] Fagerstro¨m S, Pa˚hlman S, Gestblom C, Na˚nberg E. Cell Growth Differ 1996;7:775 – 85. [10] Pa˚hlman S, Odelstad L, Larsson E, Grotte G, Nilsson K. Int J Cancer 1981;28:583 – 9. [11] Zeidman R, Lo¨fgren B, Pa˚hlman S, Larsson C. J Cell Biol 1999;145: 713 – 26. [12] Corbit KC, Foster DA, Rosner MR. Mol Cell Biol 1999;19:4209 – 18. [13] Newton AC. J Biol Chem 1995;270:28495 – 8. [14] Nishizuka Y. Science 1992;258:607 – 14. [15] Ling M, Trolle´r U, Zeidman R, Lundberg C, Larsson C. Exp Cell Res (in press).
[16] Tatsis N, Lannigan DA, Macara IG. J Biol Chem 1998;273:34631 – 8. [17] Lavenius E, Gestblom C, Johansson I, Na˚nberg E, Pa˚hlman S. Cell Growth Differ 1995;6:727 – 36. [18] Zeidman R, Trolle´r U, Raghunath A, Pa˚hlman S, Larsson C. Mol Biol Cell 2002;13:12 – 24. [19] Hernandez RM, Wescott GG, Mayhew MW, McJilton MA, Terrian DM. J Cell Biochem 2001;83:532 – 46. [20] Svensson K, Zeidman R, Trolle´r U, Schultz A, Larsson C. Cell Growth Differ 2000;11:641 – 8. [21] Hall A. Science 1998;279:509 – 14. [22] Miura Y, Kikuchi A, Musha T, Kuroda S, Yaku H, Sasaki T, et al. J Biol Chem 1993;268:510 – 5. [23] Gebbink MF, Kranenburg O, Poland M, van Horck FP, Houssa B, Moolenaar WH. J Cell Biol 1997;137:1603 – 13. [24] Sebok A, Nusser N, Debreceni B, Guo Z, Santos MF, Szeberenyi J, et al. J Neurochem 1999;73:949 – 60.
A possible role for p190RhoGAP in PKCq-induced morphological effects Ulrika Trolle´r, Arathi Raghunath, Christer Larsson * Lund University, Department of Laboratory Medicine, Molecular Medicine, Entrance 78, 3rd Floor, Malmo¨ University Hospital, UMAS, SE-205 02 Malmo¨, Sweden Received 27 May 2003; received in revised form 28 July 2003; accepted 28 July 2003
Abstract We have previously shown that protein kinase C (PKC) q induces neurite outgrowth via its regulatory domain. This is accompanied by PKC-induced stress fibre loss. Here, we show that the regulatory domain (RD) of PKCq induces processes also in NIH-3T3 fibroblasts, similar to what has been observed with p190RhoGAP. This study also shows that p190RhoGAP induces neurite outgrowth in SK-N-BE(2) neuroblastoma cells. We therefore investigated whether p190RhoGAP may be downstream of PKCq. We could detect a co-localization of p190RhoGAP and PKCq at membrane ruffles and an increased association between the proteins in fibroblasts treated with 12-Otetradecanoylphorbol-13-acetate (TPA). The association is also seen in neuroblastoma cells, and nerve growth factor (NGF) treatment of SHSYSY/TrkA cells decreases the interaction. However, overexpressed PKCq did not coprecipitate overexpressed p190RhoGAP in CHO cells, indicating that the proteins do not interact directly. This raises the possibility that p190RhoGAP is involved in mediating the morphological effects of PKCq. D 2003 Elsevier Inc. All rights reserved. Keywords: Protein kinase C; p190RhoGAP; Neurite; Neuroblastoma; Fibroblasts
1. Introduction The outgrowth and guidance of neurites are regulated at the growing end of the neurite, the growth cone, which senses extracellular cues from the surrounding cells and the extracellular matrix. The intracellular pathways conveying the external signals involve an interplay between several proteins. One family of proteins that is likely to be of key importance for the control of the growth cone advance and guidance is the Rho family of small GTPases. The members of the Rho family regulate a variety of cellular processes including the rearrangement of the actin cytoskeleton, which is necessary for the dynamics in the neuronal growth cone. Active Rac1 and Cdc42 have been shown to be important for the growth cone advance by inducing the formation of lamellipodia and filopodia, respectively, while active RhoA, which regulates the formation of stress fibers and focal adhesions, has been shown to have the opposite effect and
Abbreviations: CD, catalytic domain; EGFP, enhanced green fluorescent protein; FL, full-length; NGF, nerve growth factor; PKC, protein kinase C; RD, regulatory domain; TPA, 12-O-tetradecanoylphorbol-13acetate. * Corresponding author. Tel.: +46-40-337404; fax: +46-40-337322. E-mail address: [email protected] (C. Larsson). 0898-6568/$ - see front matter D 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0898-6568(03)00135-9
cause neurites to retract [1 – 3]. The Rho proteins cycle between an active GTP-bound state and an inactive GDPbound state, and the switch between these two states is mediated by three classes of proteins. Guanine nucleotide exchange factors (GEFs) stimulate the dissociation of GDP which allows GTP to bind and activate the protein, guanine nucleotide dissociation inhibitors (GDIs) bind Rho-GDP and thereby maintain it inactive, and GTPase-activating proteins (GAPs) enhance the intrinsic GTPase activity of the Rho GTPases and thereby inactivate them [4]. p190RhoGAP is a GTPase-activating protein for the Rho family and acts specifically on RhoA [5]. p190RhoGAP has been shown to be expressed at high levels in the developing nervous system, and p190RhoGAP / mice have defects in axon outgrowth, guidance, and fasciculation and exhibit impaired neuronal development [6]. The resulting defects include a complete lack of corpus callosum and other forebrain midline structures, disorganized layering of the cerebral cortex, and defective closure of the neural tube [7]. Protein kinase C (PKC) is another family involved in a multitude of cellular processes, including the regulation of neurite outgrowth [8 – 12]. It is a family of serine/threonine kinases consisting of at least 10 isoforms, which are grouped in three subfamilies based on their requirements for activation. Conventional PKCs (a, hI, hII, and g) are
246
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
activated by Ca2 +, diacylglycerol, and phosphatidylserine. Novel PKCs (y, q, D, and u) are Ca2 +-independent and atypical PKCs (L/E and ~) are independent of both Ca2 + and diacylglycerol [13,14]. PKC has been shown to directly phosphorylate p190RhoGAP in vitro, and treatment of fibroblasts with the PKC activator 12-O-tetradecanoylphorbol-13-acetate (TPA) leads to increased phosphorylation of p190RhoGAP and a translocation of p190RhoGAP from the cytoplasm to the membrane [7]. We have previously shown that overexpression of PKCq induces the formation of neurites in neural cells [11]. This is independent of the catalytic activity, and the regulatory domain (RD) is sufficient for the effect. The PKC-induced formation of neurites has also been shown to be accompanied by a loss of stress fibres, and the neurite outgrowth could be blocked by overexpressing constitutively active RhoA, indicating a suppression of RhoA activity by PKC [15]. This led us to speculate that p190RhoGAP may be one component that mediates PKC effects on RhoA activity and on cellular morphology. This study was designed to elucidate a putative role for p190RhoGAP in PKCq-induced morphological changes.
2. Material and methods 2.1. Plasmids Expression vectors encoding PKC domains fused to the enhanced green fluorescent protein (EGFP) have previously been described [11]. Full-length PKCq and PKCq-PSC1V3 cDNA were cut out from the pEGFP-N1 vector (Clontech) by digestion with BamHI and XhoI and ligated into pcDNA4/ myc-His vector (Invitrogen). pKH3 expression vector encoding HA-tagged p190RhoGAP was kindly provided by Dr. I. Macara [16].
between the plasmids in the co-transfection studies were in SK-N-BE(2) cells 1 (pEGFP-N1) to 5 (pKH3 with insert), and in the CHO cells, equal amounts of plasmids were used. When indicated, 12-O-tetradecanoylphorbol-13-acetate (TPA; Sigma) was used at a concentration of 16 nM. For differentiation studies, cells were treated with 100 ng/ml nerve growth factor (NGF; Promega). 2.3. Immunofluorescence After transfection NIH-3T3 and SK-N-BE(2) cells were incubated for 16 h before fixation in 4% paraformaldehyde in PBS for 4 min. For detection of F-actin, cells were incubated for 20 min with Alexa Fluor 546-conjugated phalloidin (Molecular Probes) diluted 1:100 in TBS. For detection of p190RhoGAP and PKCq, cells were permeabilized and blocked with 5% goat serum and 0.3% Triton X100 in TBS for 30 min and thereafter incubated for 1 h with primary mouse anti-p190RhoGAP (Transduction Laboratories) antibody or primary rabbit anti-PKCq (Santa Cruz) antibody, diluted 1:50 in TBS. Following washes, cells were incubated for 1 h with secondary antibodies Alexa Fluor 546- or 488-conjugated goat anti-mouse IgG or Alexa Fluor 546 goat anti-rabbit IgG (Molecular Probes), diluted 1:400 in TBS. Cover slips were mounted on object slides using 20 Al PVA-DABCO (9.6% polyvinyl alcohol, 24% glycerol, and 2.5% 1,4-diazabicyclo[2.2.2]octane in 67 mM Tris – HCl, pH 8.0). 2.4. Morphology studies Transfected cells were examined with fluorescence microscopy and were considered to have long processes if the length of the process was more than the length of two cell bodies. Two hundred transfected cells were counted in each experiment. Digital images were captured with a Sony DKC 5000 camera system.
2.2. Cell culture 2.5. Confocal microscopy Human neuroblastoma SK-N-BE(2) and SH-SY5Y/TrkA [17] cells were maintained in Minimal Essential Medium (Gibco), and mouse NIH-3T3 fibroblasts and Chinese Hamster Ovarian (CHO) cells were cultured in RPMI 1640 (Gibco). All mediums were supplemented with 10% foetal bovine serum, 100 IU/ml penicillin, and 100 Ag/ml streptomycin (Gibco). For morphology studies, cells were trypsinized and seeded at a density of 100,000 cells per 35-mm cell culture dish on glass cover slips. For immunoprecipitation studies, cells were trypsinized and seeded at a density of 1 106 cells per 100-mm cell culture dish or 500,000 cells per 60-mm cell culture dish. After 24-h incubation, cells were transfected in serum-free medium. For 35-mm dishes, 2 Ag DNA and 4 Al Lipofectamine (Gibco) in 1 ml medium were used, for 60-mm dishes, 4 Ag DNA and 8 Al Lipofectamine in 2 ml medium were used, and for 100-mm dishes, 6 Ag DNA and 12 Al Lipofectamine in 3 ml medium were used. The ratio
The subcellular localization of endogenous proteins was examined using a Bio-Rad Radiance 2000 confocal system fitted on a Nikon microscope with a 60 /NA 1.40 oil lens. Excitation wavelengths were 488 (Alexa Flour 488) and 543 nm (Alexa Flour 546), and the emission filters used were HQ515/30 (Alexa Flour 488) and 600LP (Alexa Flour 546). Co-localization was analysed with the LaserPix software (Bio-Rad). 2.6. Immunoprecipitation Cells were washed twice in PBS and lysed in RIPA buffer (10 mM Tris –HCl, pH 7.2, 1% Triton X-100, 1% sodium deoxycholate, 160 mM NaCl, 0.1% sodium dodecyl sulphate, 1 mM EGTA, 1 mM EDTA, 1 mM sodium orthovanadate, complete protease inhibitor cocktail [Roche]) for 10 min on
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
247
ice. The lysates were clarified by centrifugation at 18,000 g for 10 min at 4 jC, and incubated with 2 Ag of antibodies toward the HA epitope (Clontech), p190RhoGAP (Upstate Biotechnology), or the myc epitope (Oncogene) for 2 h with rotation at 4 jC. A 20-Al sample of GammaBind Plus Sepharose (Amersham Pharmacia) beads was added prior to an additional 1-h incubation. The immune complexes were recovered by centrifugation at 500 g for 30 s at 4 jC and washed four times in lysis buffer without protease inhibitors and resuspended in sample buffer. 2.7. Ni-NTA purification Cells were washed twice in PBS and lysed in lysis buffer (50 mM NaH2PO4, 300 mM NaCl, 10 mM imidazole, 0.05% Tween-20, 1% Triton X-100, 1 mM sodium orthovanadate, complete protease inhibitor cocktail [Roche]) on
Fig. 2. p190RhoGAP induces neurites in neuroblastoma cells. SK-N-BE(2) cells were co-transfected with empty pEGFP-N1 vector and expression vector encoding HA-tagged p190RhoGAP or empty vector encoding the HA-tag only. Sixteen hours after transfection, cells were fixed and mounted on object slides. (A) Fluorescent images of SK-N-BE(2) cells, visualized with EGFP fluorescence, transfected with empty vector or p190RhoGAP expression vector, showing a cell with neurites. (B) The percentage of transfected cells with neurites was quantified in cells expressing the HA-tag only or p190RhoGAP. Data (mean F S.E.M., n = 3) are percentage of transfected cells with neurites. (C) SK-N-BE(2) cells were transfected with either empty vector or vector encoding HA-tagged p190RhoGAP. Sixteen hours after transfection, cells were lysed and immunoprecipitated with antibodies towards the HA-tag. The immune complexes were analysed with Western blotting using antibodies toward the HA-tag.
ice for 30 min with gentle rocking. The lysates were clarified by centrifugation at 18,000 g for 10 min at 4 jC. A 450-Al sample of cleared lysate was incubated with 20 Al of nitrilotriacetic acid (Ni-NTA) agarose bead suspension (Qiagen) for 1 h with rotation at 4 jC. The Ni-NTA beads were collected by centrifugation at 500 g for 1 min at 4 jC, washed three times in wash buffer (50 mM NaH2PO 4, 300 mM NaCl, 20 mM imidazole, 0.05% Tween-20, 1 mM sodium orthovanadate, complete protease inhibitor cocktail [Roche]), and eluted by incubation for 15 min in elution buffer (50 mM NaH2PO4, 300 mM NaCl, 250 mM imidazole, 0.05% Tween-20) at room temperature. Fig. 1. Overexpression of PKCq induces beaded extensions via its regulatory domain in NIH-3T3 fibroblasts. NIH-3T3 fibroblasts were transfected with empty pEGFP-N1 vector or vectors encoding full-length PKCq (PKCq-FL), the PKCq regulatory domain (PKCq-RD), the PKCq catalytic domain (PKCq-CD) or full-length PKCy fused to EGFP. Sixteen hours after transfection, cells were fixed and mounted on object slides. (A) Fluorescent images of NIH-3T3 cells expressing EGFP, PKCq-FL, PKCqRD, and PKCq-CD showing the beaded extensions induced by PKCq-FL and PKCq-RD. (B) Quantification of the number of transfected cells with beaded extensions longer than two cell bodies. Data (mean F S.E.M., n = 3) are percentage of transfected cells with beaded extensions.
2.8. Immunoblotting Proteins were electrophoretically separated with SDSPAGE and transferred to a PVDF membrane (Millipore). The membranes were immunoblotted with polyclonal antibodies towards PKCa, hII, y, and q (Santa Cruz), the HA epitope (Clontech); monoclonal antibodies toward p190RhoGAP (Transduction Laboratories), the myc epitope (Oncogene), or the EGFP-tag (BabCO). Membranes were
248
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
thereafter incubated with horseradish peroxidase-linked secondary antibodies (Amersham), and detection was done with the SuperSignal system (Pierce) as substrate. The chemiluminescence was detected with a CCD camera (Fujifilm). Quantification of the bands was performed with the ImageGauge software (Fujifilm). The values were obtained by quantifying the intensity of the bands, subtracting the background, and normalizing to the values obtained from the loading control.
3. Results 3.1. PKCe induces a beaded extension phenotype in NIH3T3 fibroblasts We have previously shown that PKCq, via its regulatory domain, induces neurites in neuronal cells [11,18]. It has also been demonstrated that PKCq can trigger the outgrowth of beaded extensions from NIH-3T3 fibroblasts [19]. To investigate whether these effects can be related, we tested if PKCq, through its regulatory domain, could induce the extensions in fibroblasts. Accordingly, NIH-3T3 fibroblasts were transfected with vectors coding for EGFP-tagged fulllength PKCq (PKCq-FL), the PKCq regulatory domain (PKCq-RD), the PKCq catalytic domain (PKCq-CD), or
full-length PKCy (Fig. 1). We found that not only overexpression of the full-length PKCq induced profound morphological effects with the formation of beaded extensions (Fig. 1A), but also that the regulatory domain of PKCq is sufficient for this effect. This is in line with effects of PKCq on the induction of neurites seen in neural cell lines [11]. Quantification of the morphological effects by counting cells with processes longer than the length of two cell bodies (Fig. 1B) showed that PKCq-FL and PKCq-RD induced beaded extensions in 61% and 74%, respectively, of the transfected cells, while PKCq-CD had no processinducing effect. Overexpression of PKCy also induced beaded extensions, but not to the same degree as seen with PKCq.
3.2. Overexpression of p190RhoGAP in neuroblastoma cells induces neurites The PKCq-induced phenotype with beaded extensions and a rounded cell body has previously been seen in NIH3T3 cells overexpressing p190RhoGAP [16]. We therefore investigated if p190RhoGAP had a neurite-inducing effect in neuroblastoma cells. SK-N-BE(2) cells were co-transfected with vectors encoding HA-tagged p190RhoGAP with an empty pEGFP-N1 vector to visualize transfected cells.
Fig. 3. Co-localization of p190RhoGAP and PKCq at membrane ruffles. NIH-3T3 fibroblasts were stimulated with TPA for 15 min, fixed, and immunostained for PKCq (A and D) or p190RhoGAP (B and G). F-actin was visualised with Alexa Fluor 546-conjugated phalloidin (E and H). The cover slips were thereafter mounted on object slides and studied by confocal microscopy. The merged images show co-localization (white pixels) between PKCq and p190RhoGAP (C), PKCq and F-actin (F), and p190RhoGAP and F-actin (I).
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
249
(Fig. 3G –I). Thus, PKCq and p190RhoGAP appear to colocalize at membrane ruffles in TPA-treated fibroblasts. 3.4. Endogenous protein interaction between PKCe and p190RhoGAP To investigate if the co-localization was due to an association between PKCq and p190RhoGAP, we stimulated fibroblasts with TPA for different time periods and immunoprecipitated p190RhoGAP followed by immunoblot towards PKCq. As seen in Fig. 4, endogenous PKCq can be co-immunoprecipitated with p190RhoGAP, and this complex formation is further enhanced by stimulation with TPA. To explore if this complex had a potential role in neurite outgrowth, we investigated if the interaction also was seen in neuroblastoma cells. Cell extracts from SK-N-BE(2) cells were immunoprecipitated with antibodies towards p190RhoGAP, and the precipitates were thereafter analysed with antibodies towards different PKC isoforms. SK-N-BE(2) cells express the PKC isoforms a, hI, hII, y, and q [11], but since the hI isoform is mostly located in the nucleus [20], we
Fig. 4. Endogenous association of p190RhoGAP and PKCq in fibroblasts. (A) NIH-3T3 fibroblasts untreated or stimulated with 16 nM TPA for 15 or 60 min were lysed and immunoprecipitated for p190RhoGAP. The immune complexes were analysed with Western blotting using antibodies towards PKCq, or p190RhoGAP as control. (B) Quantification of the results from seven separate experiments, presenting the intensity of the PKCq bands normalized to the intensity of the p190RhoGAP band, expressed as percent of the value obtained in unstimulated cells.
As seen in Fig 2A, 15% of p190RhoGAP-overexpressing cells had neurites compared to 6.5% of the control cells. The expression of p190RhoGAP was very low as we could only detect the protein after immunoprecipitation towards the HA-tag (Fig. 2C). 3.3. PKCe and p190RhoGAP co-localize at membrane ruffles Since overexpression of PKCq or p190RhoGAP seems to give rise to the same phenotype, we further wanted to examine a possible interplay between the two proteins. p190RhoGAP has been shown to translocate to the membrane upon stimulation with the PKC-activating phorbol ester TPA, suggesting the PKC may regulate the subcellular localization of p190RhoGAP [7]. We stimulated fibroblasts with TPA for 15 min and studied the subcellular localization of endogenous PKCq and p190RhoGAP. PKCq and p190RhoGAP were found to co-localize at membrane ruffles (Fig. 3A – C). Staining for F-actin (Fig. 3E) showed that there is a co-localization between PKCq (Fig. 3D) and the cortical F-actin (Fig. 3F) in ruffle-like structures. p190RhoGAP also localizes to the cortical actin cytoskeleton in these structures
Fig. 5. The PKCq isoform associates with p190RhoGAP in neuroblastoma cells. SK-N-BE(2) cells were lysed and immunoprecipitated with antibodies towards endogenous p190RhoGAP. The immune complex was divided into four and analysed with Western blotting using antibodies towards PKCa, hII, y and q, or p190RhoGAP as a control.
250
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
excluded this isoform. As seen in Fig. 5, there is an interaction between endogenous p190RhoGAP and PKCq which is not seen with the other PKC isoforms, strengthening our results seen in the fibroblasts. 3.5. Long-term treatment with nerve growth factor decreases the interaction between PKCe and p190RhoGAP Thus, TPA leads to increased association of PKCq and p190RhoGAP in fibroblasts, and this interaction can also be detected in untreated neuroblastoma cells. Given the putative roles for these proteins in neurite outgrowth, we wanted to analyze what happens to this association during neuronal differentiation. SH-SY5Y/TrkA neuroblastoma cells were stimulated for 3 days with nerve growth factor (NGF), which elicits neuronal differentiation in these cells [17]. Thereafter, immunoprecipitation towards p190RhoGAP was performed, and bound PKCq was detected by immunoblotting. The result showed that the interaction between the proteins seemed to decrease upon long-term treatment with NGF (Fig. 6). 3.6. Overexpressed PKCe does not co-immunoprecipitate p190RhoGAP To get an indication of whether there is a direct binding between PKCq and p190RhoGAP, the proteins were overexpressed in CHO cells and immunoprecipitated. Cells were co-transfected with vectors encoding HA-tagged p190RhoGAP and myc/His-tagged full-length PKCq (q-FL-myc/His). Immunoprecipitation was thereafter performed against the myc-tag on PKCq. Immunoblotting towards the HA-tag on
Fig. 6. NGF treatment reduces the association of p190RhoGAP and PKCq in TrkA-expressing SH-SY5Y cells. SH-SY5Y/TrkA neuroblastoma cells were treated for 3 days with 100 ng/ml NGF. Cells were thereafter lysed and immunoprecipitated with antibodies towards p190RhoGAP. The immune complexes were analysed with Western blotting using antibodies toward PKCq or p190RhoGAP as a control. The result shown is representative of two separate experiments.
Fig. 7. Overexpressed PKCq and p190RhoGAP do not co-precipitate. (A) CHO cells were co-transfected with vectors encoding HA-tagged p190RhoGAP together with either empty myc/His vector or vector encoding full-length PKCq fused to myc/His. Sixteen hours after transfection, cells were lysed and immunoprecipitated with antibodies towards the myc epitope. The immune complexes were analysed with Western blotting using antibodies toward p190RhoGAP or myc as a control. (B) Fibroblasts were transfected with vector encoding PKCqPSC1V3 fused to myc/His and 16 h after transfection; cells were lysed and purified with Ni-NTA agarose beads. Samples were separated on SDSPAGE and analysed with Western blotting using antibodies against p190RhoGAP, or myc as a control.
p190RhoGAP did not detect proteins, neither in the precipitate nor the lysate (data not shown), indicating low transfection efficiency also in these cells in line with the result seen with the SK-N-BE(2) cells (Fig. 2C). Therefore, we also re-probed the membrane with antibodies towards endogenous p190RhoGAP, but neither in this case could any p190RhoGAP be detected in the precipitate, even though it is present in the lysate (Fig 7A). Immunoblotting towards the myc-tag shows that PKCq is precipitated. In another approach, we tried to precipitate endogenous p190RhoGAP with a short fragment of PKCq encompassing the pseudosubstrate, the C1 domains and the V3 hinge (qPSC1V3), which have been shown to be sufficient for neurite induction in neuroblastoma cells [11]. For this experiment, we used NIH-3T3 fibroblasts since our previous results indicated an interaction between endogenous PKCq and p190RhoGAP in these cells (Fig. 4). Fibroblasts transfected with vectors encoding the myc/His-tagged PKCqPSC1V3 were lysed, and the proteins were purified by
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
adding nickel nitrilotriacetic acid (Ni-NTA) resins (which binds to the His-tag) to the cell extract. The samples were separated by SDS-PAGE and immunoblotted for endogenous p190RhoGAP and the myc-tag as a control (Fig. 7B). The result shows that the Ni-NTA resins efficiently purified myc/His-tagged PKCq-PSC1V3, but that we neither in this case could co-purify p190RhoGAP.
4. Discussion We have previously shown that PKCq induces neurite outgrowth in neural cells via its regulatory domain, independently of its catalytic activity [11,18]. In this study, we sought to determine the mechanism by which PKCq regulates outgrowth of neurites, especially the effects on the actin cytoskeleton associated with the outgrowth. For this purpose, we have also utilized fibroblasts, which have been used extensively to study the regulation of the Rho family of small GTPases and are considered to be key regulators of the actin cytoskeleton [21]. We show here that overexpression of the regulatory domain of PKCq in NIH-3T3 fibroblasts induces a similar morphology as seen in cells overexpressing the full-length PKCq, with the outgrowth of long processes and the rounding up of the cell body [19]. The result correlates with our previous data, which demonstrated that the regulatory domain of PKCq is sufficient to induce neurite outgrowth in neural cells [11]. The fact that PKC induces the formation of long processes even in non-neural cells suggests that PKC targets cytoskeletal regulators that are common for several cell types. Similar morphological changes can also be induced by inhibiting RhoA both in fibroblasts [22] and in neuronal cells [1,3,23]. This data, together with our recent finding that PKC-induced neurite outgrowth is accompanied by a stress fibre dismantling and is counteracted by the RhoA pathway [15], led us to speculate that inactivation of RhoA may be a mechanism through which PKC induces neurites. One candidate to mediate such an effect is p190RhoGAP, since this protein has been shown to induce beaded extensions in fibroblasts [16] and to be putatively regulated by PKC and of importance for normal neuronal development [6,7]. Indeed, overexpression of p190RhoGAP does to a certain extent induce neurites in SK-N-BE(2) neuroblastoma cells. However, the neurite-inducing effect of p190RhoGAP was not as potent as with overexpressed PKCq. This may be due to the barely detectable levels of HA-tagged p190RhoGAP that was consistently obtained in transfected cells. Another explanation could be that PKC also triggers other pathways that contribute to induction of neurite outgrowth. Furthermore, our data show that in TPA-treated fibroblasts there was a co-localization of p190RhoGAP with PKCq at membrane ruffles. By using immunoprecipitation, we also detected an association between PKCq and p190RhoGAP
251
which increased after TPA treatment. This suggests that p190RhoGAP could be downstream of activated PKC in a pathway leading to cytoskeletal changes, such as the outgrowth of cellular processes. The interaction between PKCq and p190RhoGAP could also be detected in SK-N-BE(2) neuroblastoma cells. The fact that it only associated with the PKCq isoform correlates with our previous data which show that only this isoform is capable of inducing neurites in neuroblastoma cells [11,18]. We also detected an association between the proteins in SHSY5Y/TrkA cells. However, this association was decreased upon induction of differentiation with NGF treatment for 3 days. This suggests that the association of PKCq with p190RhoGAP may not be abundant once neurite outgrowth has been elicited, which is the case after NGF treatment for 3 days. The interaction may be most critical during the initiation of neurite outgrowth, which would be in line with previous findings that active RhoA inhibits neurite initiation, but is essential for neurite elongation [24]. However, we could not detect an interaction between PKCq and p190RhoGAP using overexpressed proteins. This indicates that the proteins do not directly interact with a high affinity binding. A putative interaction therefore likely involves other components. In summary, we have shown here that PKCq induces outgrowth of long cellular processes in NIH-3T3 fibroblasts via the regulatory domain, similarly to our previous results with neural cells [11]. Our data further indicate that activated PKCq indirectly interacts with p190RhoGAP at membrane ruffles, which raises the possibility that p190RhoGAP is involved in mediating PKCq-induced cytoskeletal changes.
Acknowledgements We would like to thank Dr. S. Pa˚hlman for providing SHSY5Y/TrkA cells and Dr. I. Macara for providing p190RhoGAP expression vector. This work was supported by the Swedish Cancer Society, the Children’s Cancer Foundation of Sweden, the Swedish Society for Medical Research, the Crafoord; Magnus Bergvall; Gunnar, Arvid and Elisabeth Nilsson; Ollie and Elof Ericsson; and Greta and Johan Kock Foundations, and Malmo¨ University Hospital Research Funds.
References [1] Jalink K, van Corven EJ, Hengeveld T, Morii N, Narumiya S, Moolenaar WH. J Cell Biol 1994;126:801 – 10. [2] Tigyi G, Fischer DJ, Sebok A, Marshall F, Dyer DL, Miledi R. J Neurochem 1996;66:549 – 58. [3] Kozma R, Sarner S, Ahmed S, Lim L. Mol Cell Biol 1997;17: 1201 – 11. [4] Takai Y, Sasaki T, Matozaki T. Physiol Rev 2001;81:153 – 208. [5] Ridley AJ, Self AJ, Kasmi F, Paterson HF, Hall A, Marshall CJ, et al. EMBO J 1993;12:5151 – 60.
252
U. Trolle´r et al. / Cellular Signalling 16 (2004) 245–252
[6] Brouns MR, Matheson SF, Settleman J. Nat Cell Biol 2001;3:361 – 7. [7] Brouns MR, Matheson SF, Hu KQ, Delalle I, Caviness VS, Silver J, et al. Development 2000;127:4891 – 903. [8] Brodie C, Bogi K, Acs P, Lazarovici P, Petrovics G, Anderson WB, et al. Cell Growth Differ 1999;10:183 – 91. [9] Fagerstro¨m S, Pa˚hlman S, Gestblom C, Na˚nberg E. Cell Growth Differ 1996;7:775 – 85. [10] Pa˚hlman S, Odelstad L, Larsson E, Grotte G, Nilsson K. Int J Cancer 1981;28:583 – 9. [11] Zeidman R, Lo¨fgren B, Pa˚hlman S, Larsson C. J Cell Biol 1999;145: 713 – 26. [12] Corbit KC, Foster DA, Rosner MR. Mol Cell Biol 1999;19:4209 – 18. [13] Newton AC. J Biol Chem 1995;270:28495 – 8. [14] Nishizuka Y. Science 1992;258:607 – 14. [15] Ling M, Trolle´r U, Zeidman R, Lundberg C, Larsson C. Exp Cell Res (in press).
[16] Tatsis N, Lannigan DA, Macara IG. J Biol Chem 1998;273:34631 – 8. [17] Lavenius E, Gestblom C, Johansson I, Na˚nberg E, Pa˚hlman S. Cell Growth Differ 1995;6:727 – 36. [18] Zeidman R, Trolle´r U, Raghunath A, Pa˚hlman S, Larsson C. Mol Biol Cell 2002;13:12 – 24. [19] Hernandez RM, Wescott GG, Mayhew MW, McJilton MA, Terrian DM. J Cell Biochem 2001;83:532 – 46. [20] Svensson K, Zeidman R, Trolle´r U, Schultz A, Larsson C. Cell Growth Differ 2000;11:641 – 8. [21] Hall A. Science 1998;279:509 – 14. [22] Miura Y, Kikuchi A, Musha T, Kuroda S, Yaku H, Sasaki T, et al. J Biol Chem 1993;268:510 – 5. [23] Gebbink MF, Kranenburg O, Poland M, van Horck FP, Houssa B, Moolenaar WH. J Cell Biol 1997;137:1603 – 13. [24] Sebok A, Nusser N, Debreceni B, Guo Z, Santos MF, Szeberenyi J, et al. J Neurochem 1999;73:949 – 60.