- Email: [email protected]
Raf-1 and B-Raf promote protein kinase C θ interaction with BAD
Cellular Signalling 19 (2007) 547 – 555 www.elsevier.com/locate/cellsig
Raf-1 and B-Raf promote protein kinase C θ interaction with BAD Alison Hindley a , Walter Kolch a,b,⁎ a
Signalling and Proteomics Laboratory, The Beatson Institute for Cancer Research, Garscube Estate, Switchback Road, Glasgow G61 1BD, UK b Institute of Biomedical and Life Sciences, University of Glasgow, Glasgow G12 8QQ, UK Received 22 July 2006; received in revised form 13 August 2006; accepted 13 August 2006 Available online 23 August 2006
Abstract PKCθ regulates the proliferation, survival and differentiation of T-cells. Here we show that PKCθ interacts with Raf-1 and B-Raf kinases. Raf-1 enhanced the kinase activity of associated PKCθ, while PKCθ reduced the catalytic activity of associated Raf-1. In contrast, B-Raf binding did not affect PKCθ kinase activity, and PKCθ did not change B-Raf activity. Coexpression of mutationally activated Raf-1 in cells enhanced the phosphorylation of T538 in the PKCθ activation loop. PKCθ and Raf cooperated in terms of binding to BAD, a pro-apoptotic Bcl-2 family protein that is inactivated by phosphorylation. While neither Raf-1 nor B-Raf could phosphorylate BAD, they enhanced the ability of PKCθ to interact with BAD and to phosphorylate BAD in vitro and in vivo, suggesting a new role for Raf proteins in T-cells by targeting PKCθ to interact with and phosphorylate BAD. © 2006 Elsevier Inc. All rights reserved. Keywords: T-cells; Raf-1; B-Raf; PKCθ; BAD; Phosphorylation; Signal transduction
1. Introduction Raf kinases are the entry point to the ERK/MAPK pathway, a three tiered kinase cascade where Raf phosphorylates and activates MEK which in turn phosphorylates and activates ERK. This pathway is involved in the regulation of many fundamental cellular processes including cell survival, proliferation, transformation and differentiation [1,2]. The Raf family has three members, A-Raf, Raf-1 and B-Raf. Raf-1 is ubiquitously expressed, while A-Raf and B-Raf expression appears more restricted [3]. A-Raf is mainly found in urogenital tissues, and B-Raf in haematopoietic and neuronal cells, although newer data suggest that they are more widely expressed [4]. Recently, B-Raf has received major attention as it is frequently mutated in melanoma and other cancers [5]. All three Raf kinases are activated by binding to Ras, but there are salient differences in the detailed mode of activation [6,7]. Raf-1 undergoes a multistep activation sequence that ensues with membrane translocation as a result of Ras binding. A-Raf activation probably is ⁎ Corresponding author. The Beatson Institute for Cancer Research, Garscube Estate, Switchback Road, Glasgow G61 1BD, UK. Tel.: +44 141 330 3983; fax: +44 141 942 6521. E-mail address: [email protected] (W. Kolch). 0898-6568/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.cellsig.2006.08.004
similar, whereas B-Raf activation is much simpler. Raf kinases share MEK as a common substrate, but differ in their specific activity with B-Raf being the most active and A-Raf the poorest MEK kinase. Despite many attempts to identify other substrates, MEK is hitherto the only commonly accepted substrate for Raf kinases. However, accumulating evidence mainly from knock-out studies in mice has strongly suggested that Raf isozymes also serve isoform specific functions that are different from their well characterised role in the ERK pathway [8]. So far, Raf-1 has been convincingly demonstrated to function independently of the ERK pathway in apoptosis protection by inhibiting the pro-apoptotic kinases MST2 [9] and ASK1 [10], and in cell migration by regulating Rho kinase α (ROKα) [11]. Surprisingly, these roles are also independent of Raf-1 kinase activity and could be explained by Raf-1 acting as a scaffold or adaptor protein. Nevertheless, alternative Raf kinase substrates may exist. One of the proposed Raf-1 substrates includes BAD [12], a proapoptotic BH3 family member, which is inactivated by phosphorylation [13,14]. However, later reports could not confirm the evidence for a direct phosphorylation of BAD by Raf-1 [15] indicating that it may be a Raf-1 associated kinase that phosphorylates BAD. We have previously engineered Raf-1 to accept orthogonal ATP analogues and shown that novel
548
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555 12,000 rpm for 10 min at 4 °C. Supernatants were added to proteinA or proteinG (for mAbs) agarose beads, and incubated with the appropriate antibody overnight at 4 °C. Immunoprecipitates were washed three times with ice cold lysis buffer. GST pulldowns were carried out by incubating cell lysates with glutathione sepharose beads (Sigma) and processing the beads as described above for immunoprecipitations.
phosphorylations could be induced in cell lysates suggesting the existence of further Raf-1 substrates [16]. Here we have employed a bioinformatics approach to predict alternative Raf substrates. For this we used Scansite [17] to assemble a weighted matrix of a Raf consensus phosphorylation sites based on an alignment from MEK1 and MEK2 proteins from different species across all eukaryotic taxa. Besides MEK proteins the search revealed only two other main candidates, Adenylate Cyclase VI (ACVI) and PKCθ. ACVI has recently been described as a putative Raf-1 substrate [18,19], and here we characterise the relationship between PKCθ and Raf proteins in more detail. The phosphorylation site in the PKCθ activation loop was detected by a commonly used phospho-MEK antibody attesting the similarity of the phosphorylation motifs. However, while we could not obtain evidence that Raf kinases phosphorylate PKCθ, we found that Raf kinases enhance the ability of PKCθ to interact with BAD and phosphorylate BAD.
His-PKCθ, GST-B-Raf and GST-Raf-1 virus were used at a multiplicity-ofinfection of 10 to infect Sf-9 cells plated on dishes at 0.5 × 106 cells/cm2. 42–44 h post infection cells were harvested by centrifugation and lysed in 50 mM Tris pH 7.5, 150 mM NaCl, 1 mM EDTA and 1% Triton-X100 supplemented with protease and phosphatase inhibitors (1 mM PMSF, 1 mM Na3V04, 10 mM βglycerolphosphate, 2 mM sodium pyrophosphate, 5 μg/ml leupeptin and 2μg/ml aprotinin). Lysates were incubated on ice for 20 min, then centrifuged at 12,000 rpm for 10 min at 4 °C. Supernatants were added to glutathione sepharose or nickel sepharose as appropriate and incubated overnight at 4 °C. Then beads were washed three times in lysis buffer and two times with PBS containing 50% glycerol. Beads were stored in the 50% glycerol buffer at − 20 °C.
2. Materials and methods
2.5. Kinase assays
2.1. Cell culture
Raf kinase assays were using purified GST-MEK1 from Sf-9 cells as substrate. Briefly, Raf immunoprecipitates were washed three times in lysis buffer, then twice in kinase assay buffer (50 mM Tris pH 7.5, 150 mM NaCl, 5 mM EGTA, 5 mM MgCl2) and adjusted to equal volumes. In vitro kinase assays were performed incubating the Raf immunoprecipitates, GST-MEK1, 100 μM ATP and 10 mM MgCl2 in kinase assay buffer at 32 °C for 20 min. Reactions were resolved on 7.5% SDS-polyacrylamide gels and blotted. MEK1 phosphorylation was detected by Western blotting with a phospho-specific MEK antibody detecting the Raf specific phosphorylation sites S217/221. PKCè kinase assays were performed using myelin basic protein (MBP; Invitrogen) as substrate. Immunoprecipitates were prepared as described above, and in vitro kinase assays were performed by incubating the immunoprecipitates, 2 μg MBP, 100 μM ATP, 10 mM MgCl2 and 0.08 MBq [32P]-γ-ATP in kinase assay buffer at 30 °C for 10 min. Reactions were resolved on 10% SDS-polyacrylamide gels and blotted. MBP phosphorylation was detected by autoradiography and phosphoimager. BAD kinase assays were performed using purified mouse BAD (Upstate) as substrate. For some assays GST-tagged mouse BAD was used which was overexpressed in COS cells and purified by adsorption to glutathione sepharose beads as described above. PKCθ immunoprecipitates from mammalian cells, and His-PKCθ, GST-B-Raf or GST-Raf-1 proteins expressed in Sf-9 insect cells were prepared as described above. In vitro kinase assays were performed incubating kinase beads, 2 μg BAD, 100 μM ATP, 10 mM MgCl2 in kinase assay buffer at 30 °C, for 30 min. Reactions were resolved on 10% SDS-polyacrylamide gels and blotted. BAD phosphorylation was detected by Western blotting with phospho-specific BAD antibodies to phospho-S112 and S136.
COS-1 cells were cultured in Dulbecco's minimal essential medium (DMEM; Invitrogen) supplemented with 10% foetal calf serum (FCS) and 1% glutamine, and grown at 37 °C in 5% CO2. For serum starvation cells were washed twice with PBS and incubated in DMEM supplemented with 0.2% FCS and 1% glutamine overnight. COS-1 cells were transfected with Effectene reagent (Qiagen) following the manufacturers instructions. T-cell lines Jurkat, CCRF-CEM and Molt-4 were maintained in RPMI (Invitrogen) supplemented with 10% FCS and 1% glutamine and grown at 37 °C in 5% CO2. Sf-9 cells were maintained in TC100 medium (Invitrogen) supplemented with 10% FCS, 1% L-glutamine, 0.1% pluronic solution (Sigma) and 0.1% amphotericin B (Sigma) and grown at 27 °C.
2.2. Antibodies, reagents and plasmids Raf-1 and PKCθ monoclonal antibodies (mAbs) were from BD Transduction Laboratories; B-Raf H145 polyclonal antibody was from Santa Cruz; MEK1/2 and phospho-MEK1/2 polyclonals, BAD polyclonal (detecting mouse BAD), phospho-BAD Ser136 polyclonal and phospho-BAD Ser112 mAb were from Cell Signalling Technology; FLAG M2 agarose was from Sigma; BAD antibody detecting human BAD was from AbCam. Rottlerin was from Calbiochem; TPA from Sigma. Raf-1, B-Raf and PKCθ mutants were created using the Quick Change kit from Stratagene. A mammalian PKCθ expression vector and the His-PKCθ virus were generously provided by Dr. A Altman. The mammalian expression vectors for B-Raf [20] and FLAG-tagged Raf-1 [21] were described previously. The GST-tagged expression vector for mouse BAD (pEBG-mBAD) was from Cell Signalling Technology. The GST-Raf-1 virus was described previously [22]. The GST-B-Raf virus was made as follows. Full length human B-Raf [20] was cloned as HindIII-NdeI into pSL1180 (Pharmacia). The GST-baculovirus vector was made by cutting out a 2.4 kb fragment of the Braf cDNA from pSL1180 with Bpu1102I (an internal B-raf site that cuts 3′ of the B-raf ATG) + XbaI (from polylinker). An NcoI-Bpu1102I oligoadapter (5′-CATGGCGGCG C-3′; 3′-CGCCGCGACTC-5′) was made. The pGST/AcC5 baculovirus vector was cut with NcoI + XbaI, and ligated to the oligoadapter and B-Raf cDNA fragment. Baculoviruses were generated as described previously [22].
2.3. Preparation of lysates, immunoprecipitations and pulldowns Cells were washed with ice cold PBS and lysed in lysis buffer (20 mM Hepes pH 7.5, 150 mM NaCl, 0.5 mM EGTA, 0.5% NP40) supplemented with protease and phosphatase inhibitors (1 mM PMSF, 1 mM Na3V04, 10 mM β-glycerolphosphate, 2 mM sodium pyrophosphate, 5 μg/ml leupeptin and 2 μg/ ml aprotinin), on ice for 20 min. Lysates were subsequently centrifuged at
2.4. Purification of Sf-9 expressed proteins
3. Results 3.1. PKCθ associates with Raf-1 and B-Raf As the Raf substrates MEK1/2 associate with Raf proteins, we tested whether the Scansite predicted Raf substrate PKCθ can be co-immunoprecipitated with Raf proteins. For this purpose we co-expressed PKCθ and FLAG-tagged Raf-1 in COS cells and performed co-immunoprecipitation experiments (Fig. 1A). We also tested whether transfected PKCθ could co-immunoprecipitate with endogenous B-Raf (Fig. 1B). Both Raf-1 and B-Raf associated with PKCθ in this assay. We further examined the effect of mitogens on the interaction. In transient transfection experiments in COS cells neither serum, EGF, PDGF or TPA affected the association of Raf-1 with PKCθ (Fig. 1A and data
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
549
Fig. 1. PKCθ co-immunoprecipitates with Raf-1 and B-Raf. (A) Co-immunoprecipitation of exogenously expressed PKCθ and Raf-1 proteins. COS-1 cells were transiently transfected with PKCθ and FLAG-tagged Raf-1. Cells were serum starved (SS) overnight and stimulated with 100 ng/ml TPA for 20 min. Lysates were immunoprecipitated using the FLAG antibody and probed for associated PKCθ. An aliquot of the lysates was used to examine the expression of the transfected expression vectors. (B) Co-immunoprecipitation of exogenously expressed PKCθ and B-Raf proteins. COS-1 cells were transiently transfected with PKCθ and HAtagged B-Raf. Cells were starved overnight and stimulated with 100 ng/ml TPA for 20 min. Lysates were immunoprecipitated using the HA antibody and probed for associated PKCθ. (C) Co-immunoprecipitation of endogenous PKCθ, Raf-1 and B-Raf proteins. Cycling CCRF-CEM and Molt4 T-cell lines were left untreated or stimulated with 100 ng/ml TPA for 30 min. Lysates were immunoprecipitated using Raf-1 or PKCθ antibodies and probed for associated PKCθ, Raf-1 and B-Raf, as indicated. (D) Association between Raf-1 and PKCθ does not require activating Raf-1 phosphorylation sites. COS-1 cells were co-transfected with PKCθ and the indicated Raf-1 mutants. PKCθ immunoprecipitates were examined for coprecipitation of Raf-1 proteins.
not shown), whereas TPA enhanced the interaction of B-Raf with PKCθ (Fig. 1B). PKCθ expression was reported to be restricted, and mainly found in haematopoietic and muscle cells [23]. With the available antibodies we could detect endogenous PKCθ expression reliably only in T-cells. These cells also express detectable levels of endogenous Raf-1 and B-Raf. In co-immunoprecipitation experiments endogenous complexes between PKCθ and Raf-1 and B-Raf proteins, respectively, were readily detectable (Fig. 1C). Such complexes were observed in all T-cell lines examined (Jurkat, MOLT-4, CCRF-CEM). Interestingly, complex formation between PKCθ and Raf-1 as well as between PKCθ and B-Raf was consistently enhanced by TPA. The inducibility of the association conveniently also serves as an internal control for antibody specificity, as spurious coprecipitation would not be expected to be TPA regulated. Mitogen induced binding often uses phosphorylation as docking sites [24]. Therefore, we tested whether the mitogen
induced Raf-1 — PKCθ binding was dependent on mitogen induced Raf-1 phosphorylations sites (Fig. 1D). Raf-1 activation comprises phosphorylation in the N-region at Y340/341 [25] and in the activation loop at T491 and S494 [60]. Raf-1 mutants where these phosphorylation sites were replaced by alanines also associated with PKCθ, even slightly better than wildtype Raf-1. Thus, we conclude that Raf-1 phosphorylation of these sites is not required for binding to PKCθ. 3.2. PKCθ and Raf-1 modulate each others kinase activity Both PKCθ and Raf kinases have important roles in T-cell activation and signalling [26,27]. The above results suggested that PKCθ may be a Raf substrate and involved in Raf signalling in Tcells. The MEK1/2 and the PKCθ activation loops share sequence homologies, which are clustered around the activating phosphorylation sites, and are significantly higher than similarities between MEK1/2 and any other PKC isoforms (Fig. 2A). Indeed, in the
550
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
vicinity of the phosphorylation sites the activation loops of PKCθ is quite divergent from other PKCs. Interestingly, the phosphoMEK antibody, which recognises the MEK activation loop phosphorylated by Raf kinases, also detected the phosphorylation
Fig. 2. Raf substrate MEK shares Raf target sequence homology with PKCθ activation loop. (A) Alignment of the Raf target sequence in MEK (including the phosphorylation sites S217 and S221 underlined and bold) with all PKC family activation loop sequences. Conserved amino acids surrounding the phosphorylation sites are highlighted in grey. (B) The phospho-MEK antibody recognises PKCθ phosphorylation at T538 in the activation loop. COS-1 cells were transiently transfected with PKCθ and the PKCθ T538A mutant. Cells were starved overnight and stimulated with 100 ng/ml TPA for 20 min. Lysates were immunoprecipitated using the PKCθ antibody and probed with the phosphoMEK antibody for PKCθ phosphorylation. In parallel, phosphorylated MEK was detected in crude lysates. (C) COS-1 cells were transfected with PKCθ and activated (YY340/1DD) or kinase negative (K375M) Raf-1 mutants. PKCθ immunoprecipitates were examined for T538 phosphorylation using the phospho-MEK antibody.
of PKCθ (Fig. 2B), but not of other PKC isoforms (data not shown). This was due to the phosphorylation of T538, which is the PKCθ activation loop phosphorylation site. This residue was predicted by Scansite [28] as a Raf phosphorylation site. Mutation of T538 abolished the reactivity with the phospho-MEK antibody, and no other cross-reacting bands were detected in cell lysates except phospho-MEK (Fig. 2B). Thus, the phospho-MEK antibody seems to be able to detect phosphorylation sites known to be targeted by Raf kinases as well as T538 in PKCθ. T538 is considered a PDK-1 phosphorylation site [29], but intriguingly PDK-1 was reported to be able to phosphorylate MEK1/2 on the activating residues [30] suggesting a functional overlap between PDK-1 and Raf kinases. Thus, we tested whether Raf-1 can regulate the phosphorylation of T538 in PKCθ (Fig. 2C). The phosphorylation of T538 was induced by co-expression of an activated Raf-1 mutant, Raf-1YY340/1DD, where the activating tyrosines 340 and 341 are replaced by aspartic acids to mimic phosphorylation [31]. Further, the co-expression of a kinase negative Raf-1 mutant, Raf-1 K375M, inhibited the phosphorylation of T538, suggesting that Raf-1 can indeed regulate PKCθ activation loop phosphorylation in vivo. We also tested whether Raf-1 could phosphorylate PKCθ in vitro. Despite performing the assay in many variations, including the use of different Raf mutants and prior de-phosphorylation of PKCθ, we could not obtain any evidence that Raf kinases phosphorylate PKCθ in vitro. Likewise, we also could not find any evidence that PKCθ could phosphorylate Raf proteins (data not shown). Taken together, these data suggest that Raf-1 can regulate PKCθ phosphorylation at T538, but that this is not by direct phosphorylation. Recently, it has been reported that Raf-1 can control the activity of other kinases, MST2 [32] and ASK1 [10], independent of its own catalytic activity. Therefore, we examined whether the kinase activity of PKCθ changes when in association with Raf proteins (Fig. 3). For this purpose we compared the kinase activities of PKCθ co-immunoprecipitating with Raf-1 or B-Raf to the kinase activities of a serial dilution of PKCθ immunoprecipitates using myelin basic protein (MBP) as substrate. The kinase activity of PKCθ bound to Raf-1 was elevated approximately fourfold compared to the activity of PKCθ immunoprecipitates (Fig. 3A). MBP also has been used as artificial substrate for Raf-1, but it is only poorly phosphorylated by Raf-1 [33]. Consistent with this previous observation the kinase activity measured in the assay shown in Fig. 3A is mainly contributed by PKCθ. Control immunoprecipitates containing only Raf-1 were largely inactive. On the other hand, B-Raf did not affect the activity of associated PKCθ (Fig. 3B) suggesting that the ability to regulate the kinase activity of associated PKCθ is a specific property of Raf-1. We also tested whether PKCθ could affect the activity of Raf proteins by comparing the kinase activities of Raf proteins co-immunoprecipitating with PKCθ with serial dilutions of Raf-1 or B-Raf immunoprecipitates using MEK as a substrate (Fig. 3C and D). PKCθ immunoprecipitates did not phosphorylate MEK showing that the assay only measures Raf kinase activity. The kinase activity of Raf-1 bound to PKCθ was severely reduced (Fig. 3C). The inhibition of Raf-1 kinase activity did not depend on the activity status of PKCθ and was similarly exerted by kinase
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
551
Fig. 3. PKCθ and Raf-1 affect each others kinase when in a complex. (A) PKCθ kinase activity in Raf-1 immunoprecipitates. COS-1 cells were transiently transfected with PKCθ and Raf-1 or with PKCθ alone. Cells were serum starved overnight and stimulated with 100 ng/ml TPA for 20 min. Where PKCθ and Raf-1 were coexpressed, lysates were immunoprecipitated using the Raf-1 antibody and assayed for PKCθ activity using myelin basic protein (MBP) as substrate. Where PKCθ alone was transfected, PKCθ was immunoprecipitated using a PKCθ antibody, and PKCθ activity assayed with MBP as substrate using a serial dilution of the immunoprecipitation. Assays were quantitated by laser densitometry. Test sample values were compared to standard curves obtained from the serial dilutions. A graphic comparison of the relative kinase activities between equal amounts of Raf-1 bound and free PKCθ protein is shown below. Activity is given as scan units. (B) PKCθ kinase activity in B-Raf immunoprecipitation. As for part (A) but B-Raf was transfected instead of Raf-1. The asterisk indicates a PKCθ antibody reactive band often observed in transfected cells. (C) Raf-1 kinase activity in PKCθ immunoprecipitation. COS-1 cells were transiently transfected with Raf-1 plus PKCθ or the indicated mutants, or with Raf-1 alone. Cells were starved overnight and stimulated with 100 ng/ml TPA for 20 min. Where PKCθ and Raf-1 were co-expressed, lysates were immunoprecipitated using the PKCθ antibody and assayed for Raf-1 activity using MEK as substrate. Where Raf-1 alone was transfected, Raf-1 was immunoprecipitated using a Raf-1 antibody, and Raf activity was assayed as above using a serial dilution of the immunoprecipitation. (D) B-Raf activity in PKCθ immunoprecipitation. As for part (C) but B-Raf was transfected instead of Raf-1.
inactive (PKCθ K409R, PKCθ T538A) and constitutively activated (PKCθ A124E) mutants. The PKCθ cDNA we used has a polymorphism that changes proline 330 to leucine [34]. Although no changes in PKCθ properties due to this polymorphism have been reported, we mutated this amino acid back in order to exclude any effects of this variation. The PKCθ L330P revertant behaved indistinguishable from the wildtype PKCθ in this assay (Fig. 3C and D). Interestingly, PKCθ again had no effect on B-Raf kinase activity (Fig. 3D) suggesting a specific
cross-regulation between Raf-1 and PKCθ when they are in a complex, which results in inhibition of Raf-1 and activation of PKCθ catalytic activity. 3.3. Biochemical function of the Raf/ PKCθ complex Following up on the above observations we tried to determine whether Raf-1/PKCθ complexes could affect downstream signalling. Both Raf and PKCθ signalling share some
552
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
downstream targets, such as the AP-1 and NFκB transcription factors [35,36]. Therefore, we used transient transfection reporter gene assays to assess whether co-expression of Raf-1 and PKCθ would change AP-1 and NFκB mediated transcription. In a large series of experiments including the coexpression of both activated and kinase negative mutants we could not see a clear cross-regulation of the Raf-1 and PKCθ signalling (data not shown) suggesting that Raf-1 and PKCθ operate in independent parallel pathways to regulate AP-1 and NFκB transcription factors. Another downstream target common to PKCθ and Raf-1 is the pro-apoptotic protein BAD. BAD promotes cell death by dimerising and neutralising the protective Bcl2 and BclXL proteins [37]. BAD is inactivated by phosphorylation, and a number of kinases have been reported to phosphorylate BAD including Raf-1 and PKCθ [12,38]. However, the phosphorylation by Raf-1 in vitro is very weak [15] and may be due to an associated kinase. Therefore, we re-examined this issue using recombinant kinases expressed in Sf-9 insect cells (Fig. 4). PKCθ efficiently phosphorylated recombinant BAD protein in vitro on both inactivating sites S112 and S136 (Fig. 4A). It should be noted that PKCθ produced in Sf-9 insect cells possesses a high constitutive level of kinase activity that is only marginally enhanced by TPA treatment of the Sf-9 cells. Similarly, B-Raf expressed in Sf-9 cells also exhibits a high constitutive kinase activity that is hardly enhanced by co-
expression of RasV12. In contrast, basal Raf-1 kinase activity is low and readily inducible by co-expression of RasV12 (Fig. 4B). Despite robust activation neither Raf-1 nor B-Raf exhibited detectable kinase activity against BAD although they readily phosphorylated recombinant MEK (Fig. 4B), suggesting that PKCθ is a direct BAD kinase, but Raf-1 and B-Raf are not. To test this hypothesis in cells, we treated Jurkat cells with Rottlerin, a PKC inhibitor that preferentially inhibits PKCθ over other PKC isoforms [38,39]. Only the phosphorylation of S112 could be examined as the phospho-S136 antibody did not detect human BAD. Rottlerin efficiently diminished TPA induced and to lesser extent basal BAD phosphorylation in serum starved Jurkat cells (Fig. 4C). In growing Jurkat cells Rottlerin caused a similar dose dependent inhibition of both BAD S112 phosphorylation in vivo and PKCθ kinase activity as measured in vitro by MBP phosphorylation (Fig. 4D), suggesting that PKCθ functions as BAD kinase in Jurkat T-cells. Therefore, we conclude that PKCθ is a bona fide in vitro BAD kinase, and if Raf proteins participate in BAD phosphorylation their role is likely to serve as adaptors that recruit a true BAD kinase such as PKCθ. 3.4. Raf proteins promote complex formation of BAD with PKCθ and BAD phosphorylation In order to examine the possibility that Raf proteins may promote the binding of PKCθ to BAD we co-expressed GST-
Fig. 4. PKCθ phosphorylates BAD in vitro and in vivo. (A) Recombinant PKCθ phosphorylates BAD at S112 and S136 in vitro, but recombinant Raf-1 and B-Raf do not. His-PKCθ, GST-Raf-1 and GST-B-Raf were purified from Sf-9 cells and used to phosphorylate recombinant BAD in vitro. BAD phosphorylation was detected by phospho-specific antibodies against pS112 and pS136. Asterisks indicate that recombinant PKCθ and GST-Raf proteins were activated by treatment of Sf-9 cells with TPA (100 ng/ml; 20 min) before harvest, or co-expression of a Ha-RasV12 mutant, respectively. (B) Recombinant GST-Raf-1 and GST-B-Raf are active and phosphorylate MEK in vitro. Same as part (A) except that the substrate was recombinant GST-MEK. The arrowheads indicate the phosphorylation of GST-MEK and PKCθ T538, respectively, as detected by the phospho-MEK antibody. (C) The PKCθ inhibitor Rottlerin reduces endogenous BAD phosphorylation at S112. Cycling Jurkat T-cells were incubated with the indicated concentrations of Rottlerin for 30 min, prior to TPA stimulation for 4 h. Lysates were blotted for phospho-BADS112 and BAD levels. The pS136 antibody was not used as it does not recognise human BAD. (D) Jurkat cells were treated as in (C). The in vitro kinase activity of PKCθ immunoprecipitates was measured with MBP as substrate, while the phosphorylation of endogenous BAD was examined by Western blotting with pS112 antibody.
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
553
suggest that the ability of PKCθ to phosphorylate BAD in vitro is enhanced by the co-expression of Raf-1 and B-Raf. Thus, we also examined whether Raf-1 and B-Raf could increase the phosphorylation of endogenous BAD by PKCθ in human Jurkat T-cells (Fig. 6B). As the phospho-S136 antibody did not recognise human BAD only the phosphorylation of S112 could be assayed. Overexpression of PKCθ led to a small increase in BAD phosphorylation, which was enhanced by coexpression of Raf-1 or B-Raf. Again, the strongest effect was observed when PKCθ was co-expressed with both Raf-1 and B-Raf. We also tested whether the downregulation of PKCθ, Raf-1 and B-Raf by siRNA could inhibit the phosphorylation of endogenous BAD, but did not observe clear effects (data not shown). These experiments are technically difficult as the T-cell lines which express endogenous PKCθ are difficult to transfect with the high efficiency required to make siRNA experiments conclusive. Furthermore, S112 (corresponding to S75 in human BAD) can be phosphorylated by several other kinases including
Fig. 5. Raf-1 and B-Raf promote the interaction of PKCθ with BAD. COS-1 cells were transiently transfected with GST-BAD, PKCθ, Raf-1 and B-Raf in the indicated combinations. GST-BAD was pulled down using glutathione sepharose beads and associated proteins were detected by Western blotting. Lysates were examined for the overexpression of the transfected proteins by Western blotting with the indicated antibodies.
BAD with Raf-1, B-Raf and PKCθ alone, or in combination in COS-1 cells, and assessed the formation of kinase complexes co-purifying with GST-BAD (Fig. 5). All three kinases bound to BAD, although the association of BAD with Raf-1 or B-Raf only was readily detectable when the Raf kinases were overexpressed. The association with endogenous Raf kinases was at the limit of detection in most experiments. Interestingly, the co-expression of B-Raf and PKCθ enhanced the binding of both kinases to BAD. Raf-1 co-expression had very little if any effect. Similarly, the co-expression of Raf-1 and B-Raf only had a small effect. However, the simultaneous co-expression of Raf-1, B-Raf and PKCθ strongly enhanced the binding of all three kinases to BAD, suggesting that PKCθ and Raf proteins cooperate to interact with BAD. To explore the functional consequences of this cooperative complex formation we expressed PKCθ alone or together with Raf-1 and B-Raf in COS-1 cells, and assayed the ability of PKCθ immunoprecipitates to phosphorylate BAD in vitro (Fig. 6A). GST-tagged mouse BAD expressed in COS-1 cells and purified via affinity adsorption to glutathione sepharose beads was used as substrate, hence S112 and S136 showed some basal phosphorylation. This phosphorylation was clearly enhanced by PKCθ immunoprecipitated from cells that co-expressed both Raf-1 and B-Raf, but not if none or only one of the Raf isoforms had been co-expressed. Importantly, the enhancement of BAD phosphorylation was absent when a kinase negative PKCθ (PKCθ K409R) was co-expressed showing that the BAD phosphorylation was due to PKCθ activity. These results
Fig. 6. PKCθ phosphorylation of BAD is enhanced by Raf-1 and B-Raf in vitro and in vivo. (A) PKCθ was co-expressed with Raf-1 and B-Raf in COS-1 cells as indicated. PKCθ immunoprecipitates were used to phosphorylate recombinant BAD in vitro. BAD phosphorylation was detected using antibodies against phospho-S112 and phospho-S136. The levels of PKCθ in immunoprecipitates were visualised by Western blotting. (B) PKCθ was co-expressed with Raf-1 and B-Raf in Jurkat T-cells as indicated. The phosphorylation of endogenous human BAD was detected by a phospho-S112 specific antibody, and the expression of the transfected kinases was assayed by Western blotting of crude cell lysates as indicated.
554
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
AKT [40], cAMP regulated protein kinase (PKA) [41], PAK5 [42], Pim1 and Pim2 [43], and RSK2 [44], JNK1, MSK1 [45]. This high redundancy emphasizes the regulatory importance of S112 phosphorylation, but also is likely to obscure the effects of individual kinase knockdowns on S112 phosphorylation. 4. Discussion PKCθ is mainly expressed in T-cells. Upon antigen recognition by the T-cell receptor PKCθ is recruited to the central supramolecular activation cluster of the immunological synapse, a multi-protein signalling complex induced by engagement of the T-cell receptor by antigens [46]. Thus, PKCθ is central for T-cell activation stimulating the production of cytokines required for T-cell survival and in vitro proliferation. Of particular importance is the induction of interleukin-1 (IL-2) expression via activation of the transcription factors AP-1 [47], NFκB [48] and NFAT [49]. Raf kinases, in particular B-Raf, are also crucial for induction of IL-2 expression involving similar transcriptional targets as PKCθ [35,36]. Consistent with these findings both PKCθ and Raf kinases could induce AP-1 and NFκB dependent reporter gene expression in T-cells. However, co-expression experiments with activated and dominant negative mutants indicated that Raf kinases and PKCθ activate these pathways largely independently from each other (data not shown). PKCθ is activated by diacylglycerol binding and phosphorylation of the activation loop and the hydrophobic motif. PKC activation loop phosphorylation is considered to have a more structural role required for the maturation of a kinase competent conformation [50,51]. However, some observations suggest that PKCθ activation loop phosphorylation also is involved in the acute regulation of PKCθ activity [52], localisation [53] and interaction with other proteins [54]. In MEL erythroleukaemic cells PKCθ was found phosphorylated on T538 in the detergent soluble fraction, while PKCθ residing at the Golgi apparatus was devoid of T538 phosphorylation, yet surprisingly retained activity [53]. The significance of this finding is not known, but could be related to different modes of activation and downstream signaling. In the same vein it was reported that low level TCR activation specifically activates N-Ras at the Golgi [55]. T538 phosphorylation also is acutely regulated by interferon treatment of T-cells [52] suggesting that this site is used to regulate PKCθ activity in response to extracellular signals. In response to TCR activation PKCθ is activated by PDK1 and recruits the IκB kinase complex to lipid rafts for activation, subsequently leading to IκB phosphorylation, degradation and NFκB activation [54]. We found that T538 phosphorylation is constitutive, but also can be further induced by TPA treatment or expression of an activated Raf-1 mutant (Fig. 2B and C). At present the exact role of T538 phosphorylation is not resolved [51], but it is possible that T538 phosphorylation has both a structural and signaling role as it could regulate the levels of correctly folded, kinase competent PKCθ available for signaling. The kinase responsible for PKCθ activation loop phosphorylation on T538 is commonly considered to be PDK-1 [29], however there is also evidence for autophosphophorylation [50] and phosphorylation by other PKC isoforms [56]. Our data show
that — different from all other PKC isoforms — PKCθ activation loop phosphorylation is efficiently detected by phospho-MEK specific antibodies that recognise the activating sites phosphorylated by Raf kinases in MEK. Intriguingly, MEK also has been reported to be phosphorylated and activated by PDK-1 [30], although this report has not been confirmed yet. These results indicate that the PKCθ activation loop and MEK activating phosphorylation site sequences share common features, and are consistent with our bioinformatics analysis suggesting that PKCθ also may be a Raf substrate. However, despite extensive experimentation we could not find evidence for PKCθ being a Raf kinase substrate or for Raf being a PKCθ substrate. These experiments were performed in many different ways including the use of in vitro reconstitution assays with purified enzymes, the use of pharmacological Raf kinase inhibitors, and the use of activated and kinase negative Raf and PKCθ mutants. These negative results probably reflect the fact that Raf kinases require features in addition to a consensus sequence in order to recognise a substrate. For instance, Raf kinases do not phosphorylate peptides derived from MEK or denatured MEK proteins [33]. Nevertheless, Raf-1 could regulate T538 phosphorylation of in vivo suggesting an indirect mechanism e.g. the induction of autocrine factors that can stimulate 3-phosphoinositide production and PDK1 activation. Alternatively, and more interestingly the association with Raf-1 may promote a PKCθ conformation that permits more efficient autophosphorylation of T538. Such a mechanism also could account for our observation that PKCθ associated with Raf-1 has a higher in vitro kinase activity (Fig. 3A). Thereby, Raf-1 may generate highly active PKCθ confined to Raf-1 signalling complexes in the cell. Raf-1 signalling complexes have been reported to contain BAD [12,57], a Bcl2 family protein that promotes apoptosis and can be inactivated by multiple phosphorylations [13,58]. BAD inactivation seems to be physiologically controlled by several different pathways, and a great number of kinases have been reported to phosphorylate and inactivate BAD. In addition to AKT [40], PKA [41], PAK5 [42], Pim1/2 [43], RSK2 [44], JNK1, and MSK1 [45], the list of BAD kinases reported in the literature also includes Raf-1 [12] and PKCθ [38]. Raf-1 was originally described to phosphorylate BAD directly [12], but a later publication stated that BAD is only a poor in vitro substrate for Raf-1 [15], indicating that Raf-1 acts indirectly to promote BAD phosphorylation. PKCθ was reported to induce BAD phosphorylation indirectly via activation of RSK [59] and also via direct phosphorylation of BAD [38]. This is consistent with our results showing that PKCθ, but not Raf-1 or B-Raf, can phosphorylate BAD directly (Fig. 4A). Rather our findings indicate a novel function of Raf kinases in promoting the interaction of PKCθ with BAD. B-Raf seems more efficient than Raf-1 in this respect, although B-Raf does not change PKCθ kinase activity. In contrast, Raf-1 stimulates the kinase activity of associated PKCθ. This suggests a model where the two Raf kinases cooperate in terms of B-Raf facilitating PKCθ targeting to BAD and Raf-1 enhancing PKCθ kinase activity. Our results also lend support to the emerging view that signalling is regulated by an intricate network of protein associations that determine both substrate interactions and enzymatic activities.
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
Acknowledgements We thank Dr. Amnon Altman for the PKCθ expression vectors. This research was supported by the Cancer Research UK. References [1] W. Kolch, Nat. Rev., Mol. Cell Biol. 6 (2005) 827. [2] C. Wellbrock, M. Karasarides, R. Marais, Nat. Rev., Mol. Cell Biol. 5 (2004) 875. [3] S.M. Storm, J.L. Cleveland, U.R. Rapp, Oncogene 5 (1990) 345. [4] C. Wellbrock, M. Karasarides, R. Marais, Nat. Rev., Mol. Cell Biol. 5 (2004) 875. [5] M.J. Garnett, R. Marais, Cancer Cell 6 (2004) 313. [6] C. Wellbrock, M. Karasarides, R. Marais, Nat. Rev., Mol. Cell Biol. 5 (2004) 875. [7] A.S. Dhillon, W. Kolch, Arch. Biochem. Biophys. 404 (2002) 3. [8] E. O'Neill, W. Kolch, Br. J. Cancer 90 (2004) 283. [9] E. O'Neill, L. Rushworth, M. Baccarini, W. Kolch, Science 306 (2004) 2267. [10] J. Chen, K. Fujii, L. Zhang, T. Roberts, H. Fu, Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 7783. [11] K. Ehrenreiter, D. Piazzolla, V. Velamoor, I. Sobczak, J.V. Small, J. Takeda, T. Leung, M. Baccarini, J. Cell Biol. 168 (2005) 955. [12] H.G. Wang, U.R. Rapp, J.C. Reed, Cell 87 (1996) 629. [13] J. Zha, H. Harada, E. Yang, J. Jockel, S.J. Korsmeyer, Cell 87 (1996) 619. [14] J. Downward, Nat. Cell Biol. 1 (1999) E33. [15] A. von Gise, P. Lorenz, C. Wellbrock, B. Hemmings, F. Berberich-Siebelt, U.R. Rapp, J. Troppmair, Mol. Cell. Biol. 21 (2001) 2324. [16] A.D. Hindley, S. Park, L. Wang, K. Shah, Y. Wang, X. Hu, K.M. Shokat, W. Kolch, J.M. Sedivy, K.C. Yeung, FEBS Lett. 556 (2004) 26. [17] J.C. Obenauer, L.C. Cantley, M.B. Yaffe, Nucleic Acids Res. 31 (2003) 3635. [18] M.A. Beazely, J.K. Alan, V.J. Watts, Mol. Pharmacol. 67 (2005) 250. [19] Q. Ding, R. Gros, I.D. Gray, R. Taussig, S.S. Ferguson, R.D. Feldman, Mol. Pharmacol. 66 (2004) 921. [20] E.M. Weissinger, G. Eissner, C. Grammer, S. Fackler, B. Haefner, L.S. Yoon, K.S. Lu, A. Bazarov, J.M. Sedivy, H. Mischak, W. Kolch, Mol. Cell. Biol. 17 (1997) 3229. [21] E. O'Neill, L. Rushworth, M. Baccarini, W. Kolch, Science 306 (2004) 2267. [22] S. Häfner, H.S. Adler, H. Mischak, P. Janosch, G. Heidecker, A. Wolfman, S. Pippig, M. Lohse, M. Ueffing, W. Kolch, Mol. Cell. Biol. 14 (1994) 6696. [23] A. Altman, M. Villalba, Immunol. Rev. 192 (2003) 53. [24] W. Kolch, Nat. Rev., Mol. Cell Biol. 6 (2005) 827. [25] R. Marais, Y. Light, H.F. Paterson, C.J. Marshall, EMBO J. 14 (1995) 3136. [26] G. Baier, Immunol. Rev. 192 (2003) 64. [27] M.I. Pastor, M. Woodrow, D. Cantrell, Cancer Surv. 22 (1995) 75. [28] J.C. Obenauer, L.C. Cantley, M.B. Yaffe, Nucleic Acids Res. 31 (2003) 3635. [29] E.M. Dutil, A. Toker, A.C. Newton, Curr. Biol. 8 (1998) 1366. [30] S. Sato, N. Fujita, T. Tsuruo, J. Biol. Chem. 279 (2004) 33759.
555
[31] R. Marais, Y. Light, H.F. Paterson, C.S. Mason, C.J. Marshall, J. Biol. Chem. 272 (1997) 4378. [32] E. O'Neill, L. Rushworth, M. Baccarini, W. Kolch, Science 306 (2004) 2267. [33] T. Force, J.V. Bonventre, G. Heidecker, U. Rapp, J. Avruch, J.M. Kyriakis, Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 1270. [34] J.D. Chang, Y. Xu, M.K. Raychowdhury, J.A. Ware, J. Biol. Chem. 268 (1993) 14208. [35] H. Tsukamoto, A. Irie, Y. Nishimura, J. Biol. Chem. 279 (2004) 48457. [36] Y.Q. Li, C.S. Hii, C.J. Der, A. Ferrante, Immunology 96 (1999) 524. [37] S.N. Willis, J.M. Adams, Curr. Opin. Cell Biol. 17 (2005) 617. [38] M. Villalba, P. Bushway, A. Altman, J. Immunol. 166 (2001) 5955. [39] S.P. Davies, H. Reddy, M. Caivano, P. Cohen, Biochem. J. 351 (2000) 95. [40] L. del Peso, M. Gonzalez-Garcia, C. Page, R. Herrera, G. Nunez, Science 278 (1997) 687. [41] H. Harada, B. Becknell, M. Wilm, M. Mann, L.J. Huang, S.S. Taylor, J.D. Scott, S.J. Korsmeyer, Mol. Cell 3 (1999) 413. [42] S. Cotteret, Z.M. Jaffer, A. Beeser, J. Chernoff, Mol. Cell. Biol. 23 (2003) 5526. [43] B. Yan, M. Zemskova, S. Holder, V. Chin, A. Kraft, P.J. Koskinen, M. Lilly, J. Biol. Chem. 278 (2003) 45358. [44] Y. Tan, H. Ruan, M.R. Demeter, M.J. Comb, J. Biol. Chem. 274 (1999) 34859. [45] Q.B. She, W.Y. Ma, S. Zhong, Z. Dong, J. Biol. Chem. 277 (2002) 24039. [46] C.R. Monks, H. Kupfer, I. Tamir, A. Barlow, A. Kupfer, Nature 385 (1997) 83. [47] G. Baier-Bitterlich, F. Uberall, B. Bauer, F. Fresser, H. Wachter, H. Grunicke, G. Utermann, A. Altman, G. Baier, Mol. Cell. Biol. 16 (1996) 1842. [48] Z. Sun, C.W. Arendt, W. Ellmeier, E.M. Schaeffer, M.J. Sunshine, L. Gandhi, J. Annes, D. Petrzilka, A. Kupfer, P.L. Schwartzberg, D.R. Littman, Nature 404 (2000) 402. [49] C. Pfeifhofer, K. Kofler, T. Gruber, N.G. Tabrizi, C. Lutz, K. Maly, M. Leitges, G. Baier, J. Exp. Med. 197 (2003) 1525. [50] R. Czerwinski, A. Aulabaugh, R.M. Greco, S. Olland, K. Malakian, S. Wolfrom, L. Lin, R. Kriz, M. Stahl, Y. Huang, L. Liu, D. Chaudhary, Biochemistry 44 (2005) 9563. [51] T. Gruber, M. Freeley, N. Thuille, I. Heit, S. Shaw, A. Long, G. Baier, Science 312 (2006) 55. [52] K.K. Srivastava, S. Batra, A. Sassano, Y. Li, B. Majchrzak, H. Kiyokawa, A. Altman, E.N. Fish, L.C. Platanias, J. Biol. Chem. 279 (2004) 29911. [53] B. Sparatore, M. Passalacqua, M. Pedrazzi, S. Ledda, M. Patrone, D. Gaggero, S. Pontremoli, E. Melloni, FEBS Lett. 554 (2003) 35. [54] K.Y. Lee, F. D'Acquisto, M.S. Hayden, J.H. Shim, S. Ghosh, Science 308 (2005) 114. [55] D.C. Perez, I.T.G. Bivona, M.R. Philips, A. Pellicer, Mol. Cell. Biol. 24 (2004) 3485. [56] V.O. Rybin, A. Sabri, J. Short, J.C. Braz, J.D. Molkentin, S.F. Steinberg, J. Biol. Chem. 278 (2003) 14555. [57] S. Jin, Y. Zhuo, W. Guo, J. Field, J. Biol. Chem. 280 (2005) 24698. [58] J. Downward, Nat. Cell Biol. 1 (1999) E33. [59] C. Bertolotto, L. Maulon, N. Filippa, G. Baier, P. Auberger, J. Biol. Chem. 275 (2000) 37246. [60] H. Chong, J. Lee, J., K.L. Guan, EMBO J. 20 (2001) 3716.
Raf-1 and B-Raf promote protein kinase C θ interaction with BAD Alison Hindley a , Walter Kolch a,b,⁎ a
Signalling and Proteomics Laboratory, The Beatson Institute for Cancer Research, Garscube Estate, Switchback Road, Glasgow G61 1BD, UK b Institute of Biomedical and Life Sciences, University of Glasgow, Glasgow G12 8QQ, UK Received 22 July 2006; received in revised form 13 August 2006; accepted 13 August 2006 Available online 23 August 2006
Abstract PKCθ regulates the proliferation, survival and differentiation of T-cells. Here we show that PKCθ interacts with Raf-1 and B-Raf kinases. Raf-1 enhanced the kinase activity of associated PKCθ, while PKCθ reduced the catalytic activity of associated Raf-1. In contrast, B-Raf binding did not affect PKCθ kinase activity, and PKCθ did not change B-Raf activity. Coexpression of mutationally activated Raf-1 in cells enhanced the phosphorylation of T538 in the PKCθ activation loop. PKCθ and Raf cooperated in terms of binding to BAD, a pro-apoptotic Bcl-2 family protein that is inactivated by phosphorylation. While neither Raf-1 nor B-Raf could phosphorylate BAD, they enhanced the ability of PKCθ to interact with BAD and to phosphorylate BAD in vitro and in vivo, suggesting a new role for Raf proteins in T-cells by targeting PKCθ to interact with and phosphorylate BAD. © 2006 Elsevier Inc. All rights reserved. Keywords: T-cells; Raf-1; B-Raf; PKCθ; BAD; Phosphorylation; Signal transduction
1. Introduction Raf kinases are the entry point to the ERK/MAPK pathway, a three tiered kinase cascade where Raf phosphorylates and activates MEK which in turn phosphorylates and activates ERK. This pathway is involved in the regulation of many fundamental cellular processes including cell survival, proliferation, transformation and differentiation [1,2]. The Raf family has three members, A-Raf, Raf-1 and B-Raf. Raf-1 is ubiquitously expressed, while A-Raf and B-Raf expression appears more restricted [3]. A-Raf is mainly found in urogenital tissues, and B-Raf in haematopoietic and neuronal cells, although newer data suggest that they are more widely expressed [4]. Recently, B-Raf has received major attention as it is frequently mutated in melanoma and other cancers [5]. All three Raf kinases are activated by binding to Ras, but there are salient differences in the detailed mode of activation [6,7]. Raf-1 undergoes a multistep activation sequence that ensues with membrane translocation as a result of Ras binding. A-Raf activation probably is ⁎ Corresponding author. The Beatson Institute for Cancer Research, Garscube Estate, Switchback Road, Glasgow G61 1BD, UK. Tel.: +44 141 330 3983; fax: +44 141 942 6521. E-mail address: [email protected] (W. Kolch). 0898-6568/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.cellsig.2006.08.004
similar, whereas B-Raf activation is much simpler. Raf kinases share MEK as a common substrate, but differ in their specific activity with B-Raf being the most active and A-Raf the poorest MEK kinase. Despite many attempts to identify other substrates, MEK is hitherto the only commonly accepted substrate for Raf kinases. However, accumulating evidence mainly from knock-out studies in mice has strongly suggested that Raf isozymes also serve isoform specific functions that are different from their well characterised role in the ERK pathway [8]. So far, Raf-1 has been convincingly demonstrated to function independently of the ERK pathway in apoptosis protection by inhibiting the pro-apoptotic kinases MST2 [9] and ASK1 [10], and in cell migration by regulating Rho kinase α (ROKα) [11]. Surprisingly, these roles are also independent of Raf-1 kinase activity and could be explained by Raf-1 acting as a scaffold or adaptor protein. Nevertheless, alternative Raf kinase substrates may exist. One of the proposed Raf-1 substrates includes BAD [12], a proapoptotic BH3 family member, which is inactivated by phosphorylation [13,14]. However, later reports could not confirm the evidence for a direct phosphorylation of BAD by Raf-1 [15] indicating that it may be a Raf-1 associated kinase that phosphorylates BAD. We have previously engineered Raf-1 to accept orthogonal ATP analogues and shown that novel
548
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555 12,000 rpm for 10 min at 4 °C. Supernatants were added to proteinA or proteinG (for mAbs) agarose beads, and incubated with the appropriate antibody overnight at 4 °C. Immunoprecipitates were washed three times with ice cold lysis buffer. GST pulldowns were carried out by incubating cell lysates with glutathione sepharose beads (Sigma) and processing the beads as described above for immunoprecipitations.
phosphorylations could be induced in cell lysates suggesting the existence of further Raf-1 substrates [16]. Here we have employed a bioinformatics approach to predict alternative Raf substrates. For this we used Scansite [17] to assemble a weighted matrix of a Raf consensus phosphorylation sites based on an alignment from MEK1 and MEK2 proteins from different species across all eukaryotic taxa. Besides MEK proteins the search revealed only two other main candidates, Adenylate Cyclase VI (ACVI) and PKCθ. ACVI has recently been described as a putative Raf-1 substrate [18,19], and here we characterise the relationship between PKCθ and Raf proteins in more detail. The phosphorylation site in the PKCθ activation loop was detected by a commonly used phospho-MEK antibody attesting the similarity of the phosphorylation motifs. However, while we could not obtain evidence that Raf kinases phosphorylate PKCθ, we found that Raf kinases enhance the ability of PKCθ to interact with BAD and phosphorylate BAD.
His-PKCθ, GST-B-Raf and GST-Raf-1 virus were used at a multiplicity-ofinfection of 10 to infect Sf-9 cells plated on dishes at 0.5 × 106 cells/cm2. 42–44 h post infection cells were harvested by centrifugation and lysed in 50 mM Tris pH 7.5, 150 mM NaCl, 1 mM EDTA and 1% Triton-X100 supplemented with protease and phosphatase inhibitors (1 mM PMSF, 1 mM Na3V04, 10 mM βglycerolphosphate, 2 mM sodium pyrophosphate, 5 μg/ml leupeptin and 2μg/ml aprotinin). Lysates were incubated on ice for 20 min, then centrifuged at 12,000 rpm for 10 min at 4 °C. Supernatants were added to glutathione sepharose or nickel sepharose as appropriate and incubated overnight at 4 °C. Then beads were washed three times in lysis buffer and two times with PBS containing 50% glycerol. Beads were stored in the 50% glycerol buffer at − 20 °C.
2. Materials and methods
2.5. Kinase assays
2.1. Cell culture
Raf kinase assays were using purified GST-MEK1 from Sf-9 cells as substrate. Briefly, Raf immunoprecipitates were washed three times in lysis buffer, then twice in kinase assay buffer (50 mM Tris pH 7.5, 150 mM NaCl, 5 mM EGTA, 5 mM MgCl2) and adjusted to equal volumes. In vitro kinase assays were performed incubating the Raf immunoprecipitates, GST-MEK1, 100 μM ATP and 10 mM MgCl2 in kinase assay buffer at 32 °C for 20 min. Reactions were resolved on 7.5% SDS-polyacrylamide gels and blotted. MEK1 phosphorylation was detected by Western blotting with a phospho-specific MEK antibody detecting the Raf specific phosphorylation sites S217/221. PKCè kinase assays were performed using myelin basic protein (MBP; Invitrogen) as substrate. Immunoprecipitates were prepared as described above, and in vitro kinase assays were performed by incubating the immunoprecipitates, 2 μg MBP, 100 μM ATP, 10 mM MgCl2 and 0.08 MBq [32P]-γ-ATP in kinase assay buffer at 30 °C for 10 min. Reactions were resolved on 10% SDS-polyacrylamide gels and blotted. MBP phosphorylation was detected by autoradiography and phosphoimager. BAD kinase assays were performed using purified mouse BAD (Upstate) as substrate. For some assays GST-tagged mouse BAD was used which was overexpressed in COS cells and purified by adsorption to glutathione sepharose beads as described above. PKCθ immunoprecipitates from mammalian cells, and His-PKCθ, GST-B-Raf or GST-Raf-1 proteins expressed in Sf-9 insect cells were prepared as described above. In vitro kinase assays were performed incubating kinase beads, 2 μg BAD, 100 μM ATP, 10 mM MgCl2 in kinase assay buffer at 30 °C, for 30 min. Reactions were resolved on 10% SDS-polyacrylamide gels and blotted. BAD phosphorylation was detected by Western blotting with phospho-specific BAD antibodies to phospho-S112 and S136.
COS-1 cells were cultured in Dulbecco's minimal essential medium (DMEM; Invitrogen) supplemented with 10% foetal calf serum (FCS) and 1% glutamine, and grown at 37 °C in 5% CO2. For serum starvation cells were washed twice with PBS and incubated in DMEM supplemented with 0.2% FCS and 1% glutamine overnight. COS-1 cells were transfected with Effectene reagent (Qiagen) following the manufacturers instructions. T-cell lines Jurkat, CCRF-CEM and Molt-4 were maintained in RPMI (Invitrogen) supplemented with 10% FCS and 1% glutamine and grown at 37 °C in 5% CO2. Sf-9 cells were maintained in TC100 medium (Invitrogen) supplemented with 10% FCS, 1% L-glutamine, 0.1% pluronic solution (Sigma) and 0.1% amphotericin B (Sigma) and grown at 27 °C.
2.2. Antibodies, reagents and plasmids Raf-1 and PKCθ monoclonal antibodies (mAbs) were from BD Transduction Laboratories; B-Raf H145 polyclonal antibody was from Santa Cruz; MEK1/2 and phospho-MEK1/2 polyclonals, BAD polyclonal (detecting mouse BAD), phospho-BAD Ser136 polyclonal and phospho-BAD Ser112 mAb were from Cell Signalling Technology; FLAG M2 agarose was from Sigma; BAD antibody detecting human BAD was from AbCam. Rottlerin was from Calbiochem; TPA from Sigma. Raf-1, B-Raf and PKCθ mutants were created using the Quick Change kit from Stratagene. A mammalian PKCθ expression vector and the His-PKCθ virus were generously provided by Dr. A Altman. The mammalian expression vectors for B-Raf [20] and FLAG-tagged Raf-1 [21] were described previously. The GST-tagged expression vector for mouse BAD (pEBG-mBAD) was from Cell Signalling Technology. The GST-Raf-1 virus was described previously [22]. The GST-B-Raf virus was made as follows. Full length human B-Raf [20] was cloned as HindIII-NdeI into pSL1180 (Pharmacia). The GST-baculovirus vector was made by cutting out a 2.4 kb fragment of the Braf cDNA from pSL1180 with Bpu1102I (an internal B-raf site that cuts 3′ of the B-raf ATG) + XbaI (from polylinker). An NcoI-Bpu1102I oligoadapter (5′-CATGGCGGCG C-3′; 3′-CGCCGCGACTC-5′) was made. The pGST/AcC5 baculovirus vector was cut with NcoI + XbaI, and ligated to the oligoadapter and B-Raf cDNA fragment. Baculoviruses were generated as described previously [22].
2.3. Preparation of lysates, immunoprecipitations and pulldowns Cells were washed with ice cold PBS and lysed in lysis buffer (20 mM Hepes pH 7.5, 150 mM NaCl, 0.5 mM EGTA, 0.5% NP40) supplemented with protease and phosphatase inhibitors (1 mM PMSF, 1 mM Na3V04, 10 mM β-glycerolphosphate, 2 mM sodium pyrophosphate, 5 μg/ml leupeptin and 2 μg/ ml aprotinin), on ice for 20 min. Lysates were subsequently centrifuged at
2.4. Purification of Sf-9 expressed proteins
3. Results 3.1. PKCθ associates with Raf-1 and B-Raf As the Raf substrates MEK1/2 associate with Raf proteins, we tested whether the Scansite predicted Raf substrate PKCθ can be co-immunoprecipitated with Raf proteins. For this purpose we co-expressed PKCθ and FLAG-tagged Raf-1 in COS cells and performed co-immunoprecipitation experiments (Fig. 1A). We also tested whether transfected PKCθ could co-immunoprecipitate with endogenous B-Raf (Fig. 1B). Both Raf-1 and B-Raf associated with PKCθ in this assay. We further examined the effect of mitogens on the interaction. In transient transfection experiments in COS cells neither serum, EGF, PDGF or TPA affected the association of Raf-1 with PKCθ (Fig. 1A and data
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
549
Fig. 1. PKCθ co-immunoprecipitates with Raf-1 and B-Raf. (A) Co-immunoprecipitation of exogenously expressed PKCθ and Raf-1 proteins. COS-1 cells were transiently transfected with PKCθ and FLAG-tagged Raf-1. Cells were serum starved (SS) overnight and stimulated with 100 ng/ml TPA for 20 min. Lysates were immunoprecipitated using the FLAG antibody and probed for associated PKCθ. An aliquot of the lysates was used to examine the expression of the transfected expression vectors. (B) Co-immunoprecipitation of exogenously expressed PKCθ and B-Raf proteins. COS-1 cells were transiently transfected with PKCθ and HAtagged B-Raf. Cells were starved overnight and stimulated with 100 ng/ml TPA for 20 min. Lysates were immunoprecipitated using the HA antibody and probed for associated PKCθ. (C) Co-immunoprecipitation of endogenous PKCθ, Raf-1 and B-Raf proteins. Cycling CCRF-CEM and Molt4 T-cell lines were left untreated or stimulated with 100 ng/ml TPA for 30 min. Lysates were immunoprecipitated using Raf-1 or PKCθ antibodies and probed for associated PKCθ, Raf-1 and B-Raf, as indicated. (D) Association between Raf-1 and PKCθ does not require activating Raf-1 phosphorylation sites. COS-1 cells were co-transfected with PKCθ and the indicated Raf-1 mutants. PKCθ immunoprecipitates were examined for coprecipitation of Raf-1 proteins.
not shown), whereas TPA enhanced the interaction of B-Raf with PKCθ (Fig. 1B). PKCθ expression was reported to be restricted, and mainly found in haematopoietic and muscle cells [23]. With the available antibodies we could detect endogenous PKCθ expression reliably only in T-cells. These cells also express detectable levels of endogenous Raf-1 and B-Raf. In co-immunoprecipitation experiments endogenous complexes between PKCθ and Raf-1 and B-Raf proteins, respectively, were readily detectable (Fig. 1C). Such complexes were observed in all T-cell lines examined (Jurkat, MOLT-4, CCRF-CEM). Interestingly, complex formation between PKCθ and Raf-1 as well as between PKCθ and B-Raf was consistently enhanced by TPA. The inducibility of the association conveniently also serves as an internal control for antibody specificity, as spurious coprecipitation would not be expected to be TPA regulated. Mitogen induced binding often uses phosphorylation as docking sites [24]. Therefore, we tested whether the mitogen
induced Raf-1 — PKCθ binding was dependent on mitogen induced Raf-1 phosphorylations sites (Fig. 1D). Raf-1 activation comprises phosphorylation in the N-region at Y340/341 [25] and in the activation loop at T491 and S494 [60]. Raf-1 mutants where these phosphorylation sites were replaced by alanines also associated with PKCθ, even slightly better than wildtype Raf-1. Thus, we conclude that Raf-1 phosphorylation of these sites is not required for binding to PKCθ. 3.2. PKCθ and Raf-1 modulate each others kinase activity Both PKCθ and Raf kinases have important roles in T-cell activation and signalling [26,27]. The above results suggested that PKCθ may be a Raf substrate and involved in Raf signalling in Tcells. The MEK1/2 and the PKCθ activation loops share sequence homologies, which are clustered around the activating phosphorylation sites, and are significantly higher than similarities between MEK1/2 and any other PKC isoforms (Fig. 2A). Indeed, in the
550
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
vicinity of the phosphorylation sites the activation loops of PKCθ is quite divergent from other PKCs. Interestingly, the phosphoMEK antibody, which recognises the MEK activation loop phosphorylated by Raf kinases, also detected the phosphorylation
Fig. 2. Raf substrate MEK shares Raf target sequence homology with PKCθ activation loop. (A) Alignment of the Raf target sequence in MEK (including the phosphorylation sites S217 and S221 underlined and bold) with all PKC family activation loop sequences. Conserved amino acids surrounding the phosphorylation sites are highlighted in grey. (B) The phospho-MEK antibody recognises PKCθ phosphorylation at T538 in the activation loop. COS-1 cells were transiently transfected with PKCθ and the PKCθ T538A mutant. Cells were starved overnight and stimulated with 100 ng/ml TPA for 20 min. Lysates were immunoprecipitated using the PKCθ antibody and probed with the phosphoMEK antibody for PKCθ phosphorylation. In parallel, phosphorylated MEK was detected in crude lysates. (C) COS-1 cells were transfected with PKCθ and activated (YY340/1DD) or kinase negative (K375M) Raf-1 mutants. PKCθ immunoprecipitates were examined for T538 phosphorylation using the phospho-MEK antibody.
of PKCθ (Fig. 2B), but not of other PKC isoforms (data not shown). This was due to the phosphorylation of T538, which is the PKCθ activation loop phosphorylation site. This residue was predicted by Scansite [28] as a Raf phosphorylation site. Mutation of T538 abolished the reactivity with the phospho-MEK antibody, and no other cross-reacting bands were detected in cell lysates except phospho-MEK (Fig. 2B). Thus, the phospho-MEK antibody seems to be able to detect phosphorylation sites known to be targeted by Raf kinases as well as T538 in PKCθ. T538 is considered a PDK-1 phosphorylation site [29], but intriguingly PDK-1 was reported to be able to phosphorylate MEK1/2 on the activating residues [30] suggesting a functional overlap between PDK-1 and Raf kinases. Thus, we tested whether Raf-1 can regulate the phosphorylation of T538 in PKCθ (Fig. 2C). The phosphorylation of T538 was induced by co-expression of an activated Raf-1 mutant, Raf-1YY340/1DD, where the activating tyrosines 340 and 341 are replaced by aspartic acids to mimic phosphorylation [31]. Further, the co-expression of a kinase negative Raf-1 mutant, Raf-1 K375M, inhibited the phosphorylation of T538, suggesting that Raf-1 can indeed regulate PKCθ activation loop phosphorylation in vivo. We also tested whether Raf-1 could phosphorylate PKCθ in vitro. Despite performing the assay in many variations, including the use of different Raf mutants and prior de-phosphorylation of PKCθ, we could not obtain any evidence that Raf kinases phosphorylate PKCθ in vitro. Likewise, we also could not find any evidence that PKCθ could phosphorylate Raf proteins (data not shown). Taken together, these data suggest that Raf-1 can regulate PKCθ phosphorylation at T538, but that this is not by direct phosphorylation. Recently, it has been reported that Raf-1 can control the activity of other kinases, MST2 [32] and ASK1 [10], independent of its own catalytic activity. Therefore, we examined whether the kinase activity of PKCθ changes when in association with Raf proteins (Fig. 3). For this purpose we compared the kinase activities of PKCθ co-immunoprecipitating with Raf-1 or B-Raf to the kinase activities of a serial dilution of PKCθ immunoprecipitates using myelin basic protein (MBP) as substrate. The kinase activity of PKCθ bound to Raf-1 was elevated approximately fourfold compared to the activity of PKCθ immunoprecipitates (Fig. 3A). MBP also has been used as artificial substrate for Raf-1, but it is only poorly phosphorylated by Raf-1 [33]. Consistent with this previous observation the kinase activity measured in the assay shown in Fig. 3A is mainly contributed by PKCθ. Control immunoprecipitates containing only Raf-1 were largely inactive. On the other hand, B-Raf did not affect the activity of associated PKCθ (Fig. 3B) suggesting that the ability to regulate the kinase activity of associated PKCθ is a specific property of Raf-1. We also tested whether PKCθ could affect the activity of Raf proteins by comparing the kinase activities of Raf proteins co-immunoprecipitating with PKCθ with serial dilutions of Raf-1 or B-Raf immunoprecipitates using MEK as a substrate (Fig. 3C and D). PKCθ immunoprecipitates did not phosphorylate MEK showing that the assay only measures Raf kinase activity. The kinase activity of Raf-1 bound to PKCθ was severely reduced (Fig. 3C). The inhibition of Raf-1 kinase activity did not depend on the activity status of PKCθ and was similarly exerted by kinase
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
551
Fig. 3. PKCθ and Raf-1 affect each others kinase when in a complex. (A) PKCθ kinase activity in Raf-1 immunoprecipitates. COS-1 cells were transiently transfected with PKCθ and Raf-1 or with PKCθ alone. Cells were serum starved overnight and stimulated with 100 ng/ml TPA for 20 min. Where PKCθ and Raf-1 were coexpressed, lysates were immunoprecipitated using the Raf-1 antibody and assayed for PKCθ activity using myelin basic protein (MBP) as substrate. Where PKCθ alone was transfected, PKCθ was immunoprecipitated using a PKCθ antibody, and PKCθ activity assayed with MBP as substrate using a serial dilution of the immunoprecipitation. Assays were quantitated by laser densitometry. Test sample values were compared to standard curves obtained from the serial dilutions. A graphic comparison of the relative kinase activities between equal amounts of Raf-1 bound and free PKCθ protein is shown below. Activity is given as scan units. (B) PKCθ kinase activity in B-Raf immunoprecipitation. As for part (A) but B-Raf was transfected instead of Raf-1. The asterisk indicates a PKCθ antibody reactive band often observed in transfected cells. (C) Raf-1 kinase activity in PKCθ immunoprecipitation. COS-1 cells were transiently transfected with Raf-1 plus PKCθ or the indicated mutants, or with Raf-1 alone. Cells were starved overnight and stimulated with 100 ng/ml TPA for 20 min. Where PKCθ and Raf-1 were co-expressed, lysates were immunoprecipitated using the PKCθ antibody and assayed for Raf-1 activity using MEK as substrate. Where Raf-1 alone was transfected, Raf-1 was immunoprecipitated using a Raf-1 antibody, and Raf activity was assayed as above using a serial dilution of the immunoprecipitation. (D) B-Raf activity in PKCθ immunoprecipitation. As for part (C) but B-Raf was transfected instead of Raf-1.
inactive (PKCθ K409R, PKCθ T538A) and constitutively activated (PKCθ A124E) mutants. The PKCθ cDNA we used has a polymorphism that changes proline 330 to leucine [34]. Although no changes in PKCθ properties due to this polymorphism have been reported, we mutated this amino acid back in order to exclude any effects of this variation. The PKCθ L330P revertant behaved indistinguishable from the wildtype PKCθ in this assay (Fig. 3C and D). Interestingly, PKCθ again had no effect on B-Raf kinase activity (Fig. 3D) suggesting a specific
cross-regulation between Raf-1 and PKCθ when they are in a complex, which results in inhibition of Raf-1 and activation of PKCθ catalytic activity. 3.3. Biochemical function of the Raf/ PKCθ complex Following up on the above observations we tried to determine whether Raf-1/PKCθ complexes could affect downstream signalling. Both Raf and PKCθ signalling share some
552
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
downstream targets, such as the AP-1 and NFκB transcription factors [35,36]. Therefore, we used transient transfection reporter gene assays to assess whether co-expression of Raf-1 and PKCθ would change AP-1 and NFκB mediated transcription. In a large series of experiments including the coexpression of both activated and kinase negative mutants we could not see a clear cross-regulation of the Raf-1 and PKCθ signalling (data not shown) suggesting that Raf-1 and PKCθ operate in independent parallel pathways to regulate AP-1 and NFκB transcription factors. Another downstream target common to PKCθ and Raf-1 is the pro-apoptotic protein BAD. BAD promotes cell death by dimerising and neutralising the protective Bcl2 and BclXL proteins [37]. BAD is inactivated by phosphorylation, and a number of kinases have been reported to phosphorylate BAD including Raf-1 and PKCθ [12,38]. However, the phosphorylation by Raf-1 in vitro is very weak [15] and may be due to an associated kinase. Therefore, we re-examined this issue using recombinant kinases expressed in Sf-9 insect cells (Fig. 4). PKCθ efficiently phosphorylated recombinant BAD protein in vitro on both inactivating sites S112 and S136 (Fig. 4A). It should be noted that PKCθ produced in Sf-9 insect cells possesses a high constitutive level of kinase activity that is only marginally enhanced by TPA treatment of the Sf-9 cells. Similarly, B-Raf expressed in Sf-9 cells also exhibits a high constitutive kinase activity that is hardly enhanced by co-
expression of RasV12. In contrast, basal Raf-1 kinase activity is low and readily inducible by co-expression of RasV12 (Fig. 4B). Despite robust activation neither Raf-1 nor B-Raf exhibited detectable kinase activity against BAD although they readily phosphorylated recombinant MEK (Fig. 4B), suggesting that PKCθ is a direct BAD kinase, but Raf-1 and B-Raf are not. To test this hypothesis in cells, we treated Jurkat cells with Rottlerin, a PKC inhibitor that preferentially inhibits PKCθ over other PKC isoforms [38,39]. Only the phosphorylation of S112 could be examined as the phospho-S136 antibody did not detect human BAD. Rottlerin efficiently diminished TPA induced and to lesser extent basal BAD phosphorylation in serum starved Jurkat cells (Fig. 4C). In growing Jurkat cells Rottlerin caused a similar dose dependent inhibition of both BAD S112 phosphorylation in vivo and PKCθ kinase activity as measured in vitro by MBP phosphorylation (Fig. 4D), suggesting that PKCθ functions as BAD kinase in Jurkat T-cells. Therefore, we conclude that PKCθ is a bona fide in vitro BAD kinase, and if Raf proteins participate in BAD phosphorylation their role is likely to serve as adaptors that recruit a true BAD kinase such as PKCθ. 3.4. Raf proteins promote complex formation of BAD with PKCθ and BAD phosphorylation In order to examine the possibility that Raf proteins may promote the binding of PKCθ to BAD we co-expressed GST-
Fig. 4. PKCθ phosphorylates BAD in vitro and in vivo. (A) Recombinant PKCθ phosphorylates BAD at S112 and S136 in vitro, but recombinant Raf-1 and B-Raf do not. His-PKCθ, GST-Raf-1 and GST-B-Raf were purified from Sf-9 cells and used to phosphorylate recombinant BAD in vitro. BAD phosphorylation was detected by phospho-specific antibodies against pS112 and pS136. Asterisks indicate that recombinant PKCθ and GST-Raf proteins were activated by treatment of Sf-9 cells with TPA (100 ng/ml; 20 min) before harvest, or co-expression of a Ha-RasV12 mutant, respectively. (B) Recombinant GST-Raf-1 and GST-B-Raf are active and phosphorylate MEK in vitro. Same as part (A) except that the substrate was recombinant GST-MEK. The arrowheads indicate the phosphorylation of GST-MEK and PKCθ T538, respectively, as detected by the phospho-MEK antibody. (C) The PKCθ inhibitor Rottlerin reduces endogenous BAD phosphorylation at S112. Cycling Jurkat T-cells were incubated with the indicated concentrations of Rottlerin for 30 min, prior to TPA stimulation for 4 h. Lysates were blotted for phospho-BADS112 and BAD levels. The pS136 antibody was not used as it does not recognise human BAD. (D) Jurkat cells were treated as in (C). The in vitro kinase activity of PKCθ immunoprecipitates was measured with MBP as substrate, while the phosphorylation of endogenous BAD was examined by Western blotting with pS112 antibody.
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
553
suggest that the ability of PKCθ to phosphorylate BAD in vitro is enhanced by the co-expression of Raf-1 and B-Raf. Thus, we also examined whether Raf-1 and B-Raf could increase the phosphorylation of endogenous BAD by PKCθ in human Jurkat T-cells (Fig. 6B). As the phospho-S136 antibody did not recognise human BAD only the phosphorylation of S112 could be assayed. Overexpression of PKCθ led to a small increase in BAD phosphorylation, which was enhanced by coexpression of Raf-1 or B-Raf. Again, the strongest effect was observed when PKCθ was co-expressed with both Raf-1 and B-Raf. We also tested whether the downregulation of PKCθ, Raf-1 and B-Raf by siRNA could inhibit the phosphorylation of endogenous BAD, but did not observe clear effects (data not shown). These experiments are technically difficult as the T-cell lines which express endogenous PKCθ are difficult to transfect with the high efficiency required to make siRNA experiments conclusive. Furthermore, S112 (corresponding to S75 in human BAD) can be phosphorylated by several other kinases including
Fig. 5. Raf-1 and B-Raf promote the interaction of PKCθ with BAD. COS-1 cells were transiently transfected with GST-BAD, PKCθ, Raf-1 and B-Raf in the indicated combinations. GST-BAD was pulled down using glutathione sepharose beads and associated proteins were detected by Western blotting. Lysates were examined for the overexpression of the transfected proteins by Western blotting with the indicated antibodies.
BAD with Raf-1, B-Raf and PKCθ alone, or in combination in COS-1 cells, and assessed the formation of kinase complexes co-purifying with GST-BAD (Fig. 5). All three kinases bound to BAD, although the association of BAD with Raf-1 or B-Raf only was readily detectable when the Raf kinases were overexpressed. The association with endogenous Raf kinases was at the limit of detection in most experiments. Interestingly, the co-expression of B-Raf and PKCθ enhanced the binding of both kinases to BAD. Raf-1 co-expression had very little if any effect. Similarly, the co-expression of Raf-1 and B-Raf only had a small effect. However, the simultaneous co-expression of Raf-1, B-Raf and PKCθ strongly enhanced the binding of all three kinases to BAD, suggesting that PKCθ and Raf proteins cooperate to interact with BAD. To explore the functional consequences of this cooperative complex formation we expressed PKCθ alone or together with Raf-1 and B-Raf in COS-1 cells, and assayed the ability of PKCθ immunoprecipitates to phosphorylate BAD in vitro (Fig. 6A). GST-tagged mouse BAD expressed in COS-1 cells and purified via affinity adsorption to glutathione sepharose beads was used as substrate, hence S112 and S136 showed some basal phosphorylation. This phosphorylation was clearly enhanced by PKCθ immunoprecipitated from cells that co-expressed both Raf-1 and B-Raf, but not if none or only one of the Raf isoforms had been co-expressed. Importantly, the enhancement of BAD phosphorylation was absent when a kinase negative PKCθ (PKCθ K409R) was co-expressed showing that the BAD phosphorylation was due to PKCθ activity. These results
Fig. 6. PKCθ phosphorylation of BAD is enhanced by Raf-1 and B-Raf in vitro and in vivo. (A) PKCθ was co-expressed with Raf-1 and B-Raf in COS-1 cells as indicated. PKCθ immunoprecipitates were used to phosphorylate recombinant BAD in vitro. BAD phosphorylation was detected using antibodies against phospho-S112 and phospho-S136. The levels of PKCθ in immunoprecipitates were visualised by Western blotting. (B) PKCθ was co-expressed with Raf-1 and B-Raf in Jurkat T-cells as indicated. The phosphorylation of endogenous human BAD was detected by a phospho-S112 specific antibody, and the expression of the transfected kinases was assayed by Western blotting of crude cell lysates as indicated.
554
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
AKT [40], cAMP regulated protein kinase (PKA) [41], PAK5 [42], Pim1 and Pim2 [43], and RSK2 [44], JNK1, MSK1 [45]. This high redundancy emphasizes the regulatory importance of S112 phosphorylation, but also is likely to obscure the effects of individual kinase knockdowns on S112 phosphorylation. 4. Discussion PKCθ is mainly expressed in T-cells. Upon antigen recognition by the T-cell receptor PKCθ is recruited to the central supramolecular activation cluster of the immunological synapse, a multi-protein signalling complex induced by engagement of the T-cell receptor by antigens [46]. Thus, PKCθ is central for T-cell activation stimulating the production of cytokines required for T-cell survival and in vitro proliferation. Of particular importance is the induction of interleukin-1 (IL-2) expression via activation of the transcription factors AP-1 [47], NFκB [48] and NFAT [49]. Raf kinases, in particular B-Raf, are also crucial for induction of IL-2 expression involving similar transcriptional targets as PKCθ [35,36]. Consistent with these findings both PKCθ and Raf kinases could induce AP-1 and NFκB dependent reporter gene expression in T-cells. However, co-expression experiments with activated and dominant negative mutants indicated that Raf kinases and PKCθ activate these pathways largely independently from each other (data not shown). PKCθ is activated by diacylglycerol binding and phosphorylation of the activation loop and the hydrophobic motif. PKC activation loop phosphorylation is considered to have a more structural role required for the maturation of a kinase competent conformation [50,51]. However, some observations suggest that PKCθ activation loop phosphorylation also is involved in the acute regulation of PKCθ activity [52], localisation [53] and interaction with other proteins [54]. In MEL erythroleukaemic cells PKCθ was found phosphorylated on T538 in the detergent soluble fraction, while PKCθ residing at the Golgi apparatus was devoid of T538 phosphorylation, yet surprisingly retained activity [53]. The significance of this finding is not known, but could be related to different modes of activation and downstream signaling. In the same vein it was reported that low level TCR activation specifically activates N-Ras at the Golgi [55]. T538 phosphorylation also is acutely regulated by interferon treatment of T-cells [52] suggesting that this site is used to regulate PKCθ activity in response to extracellular signals. In response to TCR activation PKCθ is activated by PDK1 and recruits the IκB kinase complex to lipid rafts for activation, subsequently leading to IκB phosphorylation, degradation and NFκB activation [54]. We found that T538 phosphorylation is constitutive, but also can be further induced by TPA treatment or expression of an activated Raf-1 mutant (Fig. 2B and C). At present the exact role of T538 phosphorylation is not resolved [51], but it is possible that T538 phosphorylation has both a structural and signaling role as it could regulate the levels of correctly folded, kinase competent PKCθ available for signaling. The kinase responsible for PKCθ activation loop phosphorylation on T538 is commonly considered to be PDK-1 [29], however there is also evidence for autophosphophorylation [50] and phosphorylation by other PKC isoforms [56]. Our data show
that — different from all other PKC isoforms — PKCθ activation loop phosphorylation is efficiently detected by phospho-MEK specific antibodies that recognise the activating sites phosphorylated by Raf kinases in MEK. Intriguingly, MEK also has been reported to be phosphorylated and activated by PDK-1 [30], although this report has not been confirmed yet. These results indicate that the PKCθ activation loop and MEK activating phosphorylation site sequences share common features, and are consistent with our bioinformatics analysis suggesting that PKCθ also may be a Raf substrate. However, despite extensive experimentation we could not find evidence for PKCθ being a Raf kinase substrate or for Raf being a PKCθ substrate. These experiments were performed in many different ways including the use of in vitro reconstitution assays with purified enzymes, the use of pharmacological Raf kinase inhibitors, and the use of activated and kinase negative Raf and PKCθ mutants. These negative results probably reflect the fact that Raf kinases require features in addition to a consensus sequence in order to recognise a substrate. For instance, Raf kinases do not phosphorylate peptides derived from MEK or denatured MEK proteins [33]. Nevertheless, Raf-1 could regulate T538 phosphorylation of in vivo suggesting an indirect mechanism e.g. the induction of autocrine factors that can stimulate 3-phosphoinositide production and PDK1 activation. Alternatively, and more interestingly the association with Raf-1 may promote a PKCθ conformation that permits more efficient autophosphorylation of T538. Such a mechanism also could account for our observation that PKCθ associated with Raf-1 has a higher in vitro kinase activity (Fig. 3A). Thereby, Raf-1 may generate highly active PKCθ confined to Raf-1 signalling complexes in the cell. Raf-1 signalling complexes have been reported to contain BAD [12,57], a Bcl2 family protein that promotes apoptosis and can be inactivated by multiple phosphorylations [13,58]. BAD inactivation seems to be physiologically controlled by several different pathways, and a great number of kinases have been reported to phosphorylate and inactivate BAD. In addition to AKT [40], PKA [41], PAK5 [42], Pim1/2 [43], RSK2 [44], JNK1, and MSK1 [45], the list of BAD kinases reported in the literature also includes Raf-1 [12] and PKCθ [38]. Raf-1 was originally described to phosphorylate BAD directly [12], but a later publication stated that BAD is only a poor in vitro substrate for Raf-1 [15], indicating that Raf-1 acts indirectly to promote BAD phosphorylation. PKCθ was reported to induce BAD phosphorylation indirectly via activation of RSK [59] and also via direct phosphorylation of BAD [38]. This is consistent with our results showing that PKCθ, but not Raf-1 or B-Raf, can phosphorylate BAD directly (Fig. 4A). Rather our findings indicate a novel function of Raf kinases in promoting the interaction of PKCθ with BAD. B-Raf seems more efficient than Raf-1 in this respect, although B-Raf does not change PKCθ kinase activity. In contrast, Raf-1 stimulates the kinase activity of associated PKCθ. This suggests a model where the two Raf kinases cooperate in terms of B-Raf facilitating PKCθ targeting to BAD and Raf-1 enhancing PKCθ kinase activity. Our results also lend support to the emerging view that signalling is regulated by an intricate network of protein associations that determine both substrate interactions and enzymatic activities.
A. Hindley, W. Kolch / Cellular Signalling 19 (2007) 547–555
Acknowledgements We thank Dr. Amnon Altman for the PKCθ expression vectors. This research was supported by the Cancer Research UK. References [1] W. Kolch, Nat. Rev., Mol. Cell Biol. 6 (2005) 827. [2] C. Wellbrock, M. Karasarides, R. Marais, Nat. Rev., Mol. Cell Biol. 5 (2004) 875. [3] S.M. Storm, J.L. Cleveland, U.R. Rapp, Oncogene 5 (1990) 345. [4] C. Wellbrock, M. Karasarides, R. Marais, Nat. Rev., Mol. Cell Biol. 5 (2004) 875. [5] M.J. Garnett, R. Marais, Cancer Cell 6 (2004) 313. [6] C. Wellbrock, M. Karasarides, R. Marais, Nat. Rev., Mol. Cell Biol. 5 (2004) 875. [7] A.S. Dhillon, W. Kolch, Arch. Biochem. Biophys. 404 (2002) 3. [8] E. O'Neill, W. Kolch, Br. J. Cancer 90 (2004) 283. [9] E. O'Neill, L. Rushworth, M. Baccarini, W. Kolch, Science 306 (2004) 2267. [10] J. Chen, K. Fujii, L. Zhang, T. Roberts, H. Fu, Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 7783. [11] K. Ehrenreiter, D. Piazzolla, V. Velamoor, I. Sobczak, J.V. Small, J. Takeda, T. Leung, M. Baccarini, J. Cell Biol. 168 (2005) 955. [12] H.G. Wang, U.R. Rapp, J.C. Reed, Cell 87 (1996) 629. [13] J. Zha, H. Harada, E. Yang, J. Jockel, S.J. Korsmeyer, Cell 87 (1996) 619. [14] J. Downward, Nat. Cell Biol. 1 (1999) E33. [15] A. von Gise, P. Lorenz, C. Wellbrock, B. Hemmings, F. Berberich-Siebelt, U.R. Rapp, J. Troppmair, Mol. Cell. Biol. 21 (2001) 2324. [16] A.D. Hindley, S. Park, L. Wang, K. Shah, Y. Wang, X. Hu, K.M. Shokat, W. Kolch, J.M. Sedivy, K.C. Yeung, FEBS Lett. 556 (2004) 26. [17] J.C. Obenauer, L.C. Cantley, M.B. Yaffe, Nucleic Acids Res. 31 (2003) 3635. [18] M.A. Beazely, J.K. Alan, V.J. Watts, Mol. Pharmacol. 67 (2005) 250. [19] Q. Ding, R. Gros, I.D. Gray, R. Taussig, S.S. Ferguson, R.D. Feldman, Mol. Pharmacol. 66 (2004) 921. [20] E.M. Weissinger, G. Eissner, C. Grammer, S. Fackler, B. Haefner, L.S. Yoon, K.S. Lu, A. Bazarov, J.M. Sedivy, H. Mischak, W. Kolch, Mol. Cell. Biol. 17 (1997) 3229. [21] E. O'Neill, L. Rushworth, M. Baccarini, W. Kolch, Science 306 (2004) 2267. [22] S. Häfner, H.S. Adler, H. Mischak, P. Janosch, G. Heidecker, A. Wolfman, S. Pippig, M. Lohse, M. Ueffing, W. Kolch, Mol. Cell. Biol. 14 (1994) 6696. [23] A. Altman, M. Villalba, Immunol. Rev. 192 (2003) 53. [24] W. Kolch, Nat. Rev., Mol. Cell Biol. 6 (2005) 827. [25] R. Marais, Y. Light, H.F. Paterson, C.J. Marshall, EMBO J. 14 (1995) 3136. [26] G. Baier, Immunol. Rev. 192 (2003) 64. [27] M.I. Pastor, M. Woodrow, D. Cantrell, Cancer Surv. 22 (1995) 75. [28] J.C. Obenauer, L.C. Cantley, M.B. Yaffe, Nucleic Acids Res. 31 (2003) 3635. [29] E.M. Dutil, A. Toker, A.C. Newton, Curr. Biol. 8 (1998) 1366. [30] S. Sato, N. Fujita, T. Tsuruo, J. Biol. Chem. 279 (2004) 33759.
555
[31] R. Marais, Y. Light, H.F. Paterson, C.S. Mason, C.J. Marshall, J. Biol. Chem. 272 (1997) 4378. [32] E. O'Neill, L. Rushworth, M. Baccarini, W. Kolch, Science 306 (2004) 2267. [33] T. Force, J.V. Bonventre, G. Heidecker, U. Rapp, J. Avruch, J.M. Kyriakis, Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 1270. [34] J.D. Chang, Y. Xu, M.K. Raychowdhury, J.A. Ware, J. Biol. Chem. 268 (1993) 14208. [35] H. Tsukamoto, A. Irie, Y. Nishimura, J. Biol. Chem. 279 (2004) 48457. [36] Y.Q. Li, C.S. Hii, C.J. Der, A. Ferrante, Immunology 96 (1999) 524. [37] S.N. Willis, J.M. Adams, Curr. Opin. Cell Biol. 17 (2005) 617. [38] M. Villalba, P. Bushway, A. Altman, J. Immunol. 166 (2001) 5955. [39] S.P. Davies, H. Reddy, M. Caivano, P. Cohen, Biochem. J. 351 (2000) 95. [40] L. del Peso, M. Gonzalez-Garcia, C. Page, R. Herrera, G. Nunez, Science 278 (1997) 687. [41] H. Harada, B. Becknell, M. Wilm, M. Mann, L.J. Huang, S.S. Taylor, J.D. Scott, S.J. Korsmeyer, Mol. Cell 3 (1999) 413. [42] S. Cotteret, Z.M. Jaffer, A. Beeser, J. Chernoff, Mol. Cell. Biol. 23 (2003) 5526. [43] B. Yan, M. Zemskova, S. Holder, V. Chin, A. Kraft, P.J. Koskinen, M. Lilly, J. Biol. Chem. 278 (2003) 45358. [44] Y. Tan, H. Ruan, M.R. Demeter, M.J. Comb, J. Biol. Chem. 274 (1999) 34859. [45] Q.B. She, W.Y. Ma, S. Zhong, Z. Dong, J. Biol. Chem. 277 (2002) 24039. [46] C.R. Monks, H. Kupfer, I. Tamir, A. Barlow, A. Kupfer, Nature 385 (1997) 83. [47] G. Baier-Bitterlich, F. Uberall, B. Bauer, F. Fresser, H. Wachter, H. Grunicke, G. Utermann, A. Altman, G. Baier, Mol. Cell. Biol. 16 (1996) 1842. [48] Z. Sun, C.W. Arendt, W. Ellmeier, E.M. Schaeffer, M.J. Sunshine, L. Gandhi, J. Annes, D. Petrzilka, A. Kupfer, P.L. Schwartzberg, D.R. Littman, Nature 404 (2000) 402. [49] C. Pfeifhofer, K. Kofler, T. Gruber, N.G. Tabrizi, C. Lutz, K. Maly, M. Leitges, G. Baier, J. Exp. Med. 197 (2003) 1525. [50] R. Czerwinski, A. Aulabaugh, R.M. Greco, S. Olland, K. Malakian, S. Wolfrom, L. Lin, R. Kriz, M. Stahl, Y. Huang, L. Liu, D. Chaudhary, Biochemistry 44 (2005) 9563. [51] T. Gruber, M. Freeley, N. Thuille, I. Heit, S. Shaw, A. Long, G. Baier, Science 312 (2006) 55. [52] K.K. Srivastava, S. Batra, A. Sassano, Y. Li, B. Majchrzak, H. Kiyokawa, A. Altman, E.N. Fish, L.C. Platanias, J. Biol. Chem. 279 (2004) 29911. [53] B. Sparatore, M. Passalacqua, M. Pedrazzi, S. Ledda, M. Patrone, D. Gaggero, S. Pontremoli, E. Melloni, FEBS Lett. 554 (2003) 35. [54] K.Y. Lee, F. D'Acquisto, M.S. Hayden, J.H. Shim, S. Ghosh, Science 308 (2005) 114. [55] D.C. Perez, I.T.G. Bivona, M.R. Philips, A. Pellicer, Mol. Cell. Biol. 24 (2004) 3485. [56] V.O. Rybin, A. Sabri, J. Short, J.C. Braz, J.D. Molkentin, S.F. Steinberg, J. Biol. Chem. 278 (2003) 14555. [57] S. Jin, Y. Zhuo, W. Guo, J. Field, J. Biol. Chem. 280 (2005) 24698. [58] J. Downward, Nat. Cell Biol. 1 (1999) E33. [59] C. Bertolotto, L. Maulon, N. Filippa, G. Baier, P. Auberger, J. Biol. Chem. 275 (2000) 37246. [60] H. Chong, J. Lee, J., K.L. Guan, EMBO J. 20 (2001) 3716.