Oncogene (2005) 24, 4591–4596 & 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00 www.nature.com/onc An essential role of Pak1 phosphorylation of SHARP in Notch signaling Ratna K Vadlamudi1, Bramanandam Manavathi1,2, Rajesh R Singh1,2, Diep Nguyen1, Feng Li1 and Rakesh Kumar*,1 Department of Molecular and Cellular Oncology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030, USA The p21-activated kinases (Paks), an evolutionarily conserved family of serine/threonine kinases, play an important role in cytoskeletal reorganization in mammalian cells The Notch signaling pathway plays an important role in the determination of cell fate/differentiation in a number of organs Notch signaling is a complex process, and the mechanism by which Notch regulates multiple cellular processes is intriguing The expression of both Notch and Pak1 has been shown to be deregulated in several human cancers Using yeast two-hybrid screening, we identified SHARP, one of the Notch signaling components, as a Pak1-interacting protein We found that SHARP is a physiologic interacting substrate of Pak1, and that this interaction enhances SHARPmediated repression of Notch target genes Pak1 phosphorylation sites in SHARP were mapped to Ser3486 and Thr3568 within the SHARP repression domain Mutation of Pak1 phosphorylation sites in SHARP, inhibition of Pak1 functions by a Pak1-autoinhibitory fragment (amino acids 83–149), or expression of Pak1-specific siRNA interfered with SHARP-mediated repression of Notch target reporter gene activation These results demonstrate that Pak1–SHARP interaction plays an essential role in enhancing the corepressor functions of SHARP, thereby modulating Notch signaling in human cancer cells Oncogene (2005) 24, 4591–4596 doi:10.1038/sj.onc.1208672 Published online 11 April 2005 Keywords: Pak1 signaling; SHARP; Notch; repression The Notch signaling pathway plays an important role in the determination of cell fate (Schweisguth, 2004), and influences cell proliferation, differentiation, and apoptosis in a variety of cell types (Miele and Osborne, 1999) Initial studies with Notch signaling components suggested its involvement in neurogenesis (Beatus and Lendahl, 1998), but subsequent work showed its involvement in most organs (Iso et al., 2003) Binding of specific ligand to Notch receptors triggers cleavage of the transmembrane receptor, giving rise to the Notch intracellular domain, which translocates to the nuclear *Correspondence: R Kumar; E-mail: rkumar@mdanderson.org Both these authors contributed equally to this work Received 10 September 2004; revised February 2005; accepted 11 February 2005; published online 11 April 2005 compartment, interacts with transcription factor RBPJk, and activates transcription of Notch target genes such as HES (Hairy enhancer of Split) (Iso et al., 2003) In the absence of Notch intracellular interaction, RBPJk acts as a corepressor of Notch target genes by recruiting corepressors, including SMRT/NCOR and HDACs (Kao et al., 1998; Hsieh et al., 1999; Zhou and Hayward, 2001) Recent studies have identified SHARP as novel component of the Notch-RBP-Jk signaling pathway and implicated SHARP in repressing Notch target genes in the absence of activated Notch (Oswald et al., 2002) The molecular mechanisms by which Notch promotes RBP-Jk activation and the pathways that enhance RBP-Jk repression are not clear at the moment and are an area of active investigation The p21-activated kinase (Pak1), an evolutionarily conserved family of serine/threonine kinase, was initially identified as effectors of the Rho family of GTPases (Bokoch, 2003; Vadlamudi and Kumar, 2003) Pak1 was initially identified in the rat brain as a serine/threonine kinase activated by Rac1 or Cdc42 (Manser et al., 1994) Pak1 has been shown to play an important role in a wide variety of functions, including cytoskeletal reorganization and cell survival (Vadlamudi and Kumar, 2003) Evidence also exists that Pak1 has an essential role in the dendrite formation and neurite outgrowth (Daniels et al., 1998; Hayashi et al., 2002) Emerging data indicate that Pak1 phosphorylates histone H3 (Li et al., 2002), undergoes phosphorylation during mitosis (Thiel et al., 2002), and localizes to the nuclear compartment suggesting that Pak1 also plays an important role in nuclear signaling In the present study, we used yeast two-hybrid screening of a mammary gland cDNA library to identify SHARP, one of the Notch signaling components, as a Pak1-interacting protein To confirm this SHARP and Pak1 interaction, we cotransfected purified SHARP and Pak1 cDNAs into a yeast strain that stably expresses nutrient reporter genes Yeast transformed with both Pak1 and SHARP but not with either one alone showed the ability to grow on medium that selects for one-on-one protein– protein interactions (Figure 1a), confirming our initial yeast two-hybrid screen results To show the in vivo interaction of Pak1 and SHARP, we cotransfected a T7-tagged C-terminal SHARP fragment and myctagged Pak1 into 293 cells Immunoprecipitation of T7-tagged SHARP revealed the presence of Pak1 in the precipitates (Figure 1b, left panel) Similarly, Role of Pak1 phosphorylation in Notch signaling RK Vadlamudi et al 4592 Figure Identification of SHARP as a Pak1-binding protein (a) Yeast cells were cotransfected with GAD-SHARP and GBD vector, or GBD-Pak1 N-ter (aa 1–270), or GBD-Pak1 C-ter (aa 270–545) Cotransformants were plated on selection plates lacking leucine and tryptophan (LT) or adenosine, histidine, leucine, and tryptophan (AHLT) Growth was recorded after 72 h Growth on AHLT plates indicates protein–protein interaction between SHARP and Pak1 (b) In vivo interaction of Pak1 with SHARP HEK 293 cells were cotransfected with myc-tagged Pak1 and T7-tagged SHARP C-terminal regions (aa 3281–3661) and cell lysates were immunoprecipitated with control IgG, myc, or T7 epitope tag antibody The presence of Pak1 and SHARP in the precipitates was analysed by Western blotting (c) Pak1 and SHARP interaction in the GST pull-down assays SHARP (aa 3281–3661) was translated in vitro using a transcription and translation system (Promega), and 35S-labeled SHARP was incubated with either GST-Pak1 or GSTPak1 deletions Binding was analysed by GST pull-down assays followed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and autoradiography immunoprecipitation of myc-Pak1 with an mycepitope antibody also co-immunoprecipitated SHARP (Figure 1b, right panel), demonstrating protein–protein interaction between Pak1 and SHARP in vivo Moreover, GST-Pak1 efficiently interacted with SHARP in the GST pull-down assays (Figure 1c) Using serial deletions of Pak1, we next identified the segment of Pak1 comprising amino acids (aa) 75–132 as the SHARP binding site (Figure 1c) Collectively, these results provide evidence that Pak1 interacts with SHARP both in vitro and in vivo Since Pak1 is a serine/threonine kinase, we next examined whether SHARP is a substrate of Pak1 An in vitro kinase assay using purified GST-SHARP and Pak1 enzyme showed that Pak1 could phosphorylate SHARP (Figure 2a) Further, SHARP deletion analysis determined that Pak1 phosphorylation sites were localized in the C-terminus of SHARP (aa 3471–3661, Figure 2b) This region of SHARP contains two potential Pak1 consensus sites, serine 3486 and threonine 3568 Singlepoint mutation of either serine 3486 or threonine 3568 to alanine partially reduced Pak1 phosphorylation of SHARP (Figure 2c, lanes and 3, respectively) Double mutation of both serine and threonine to alanine completely abolished SHARP phosphorylation in vitro (Figure 2c, lane 4) To demonstrate phosphorylation of SHARP in vivo, we used metabolic labeling of 293 cells with [32P]orthophosphoric acid and transient transfection of the C-terminal SHARP, which contained both Pak1 phosphorylation and binding sites Immunoprecipitation of T7-SHARP from the cells showed that SHARP is indeed a phosphoprotein (Figure 2d) MutaOncogene tion of both Pak1 phosphorylation sites (Figure 2e) or transfection of Pak1-siRNA substantially reduced the phosphorylation status of SHARP (Figure 2f) We have also observed some basal phosphorylation of SHARP, which is not affected by mutation of Pak1 sites or inhibition of Pak1 expression (Figure 2e and f, lower band) These findings suggest that SHARP is phosphorylated by Pak1 under physiological conditions, and that SHARP is an interacting substrate of Pak1 Recent studies showed that SHARP is a novel component of the Notch/RBO-Jk signaling pathway Since Pak1 is abundantly expressed in the neuronal system (Manser et al., 1994), we examined whether Pak1 could modulate Notch/RBP-Jk signaling via its interactions with SHARP To test this possibility, we used a previously described and widely used transient cotransfection assay utilizing an RBP-Jk-responsive luciferase reporter gene, which contains six repeats of EBNA2responsive element (pGa981/6-luc) and RBP fused VP16 (RBP-VP16) to enable readout of RBP recruitment to the target promoters in 293 model cells (Oswald et al., 2002) Cotransfection of RBP-VP16 activated the reporter gene and SHARP repressed the activity of the reporter gene, confirming the earlier reported finding that SHARP represses Notch target genes (Oswald et al., 2002) Interestingly, cotransfection of Pak1 along with SHARP further enhanced the SHARP-mediated repression, suggesting that Pak1 phosphorylation of SHARP may promote its repression functions (Figure 3a) To examine the possibility that Pak1 phosphorylation sites play an important role in SHARP functions, we mutated both Pak1 phosphorylation sites Role of Pak1 phosphorylation in Notch signaling RK Vadlamudi et al 4593 Figure SHARP is a substrate of Pak1 (a) In vitro Pak1 kinase reaction was performed using purified Pak1 enzyme and GST or GST SHARP (aa 3281–3661) as a substrate as described elsewhere (Li et al., 2002) Phosphorylation of SHARP was analysed by SDS– PAGE followed by autoradiography (b) In vitro Pak1 kinase assay was performed by using various deletions of SHARP as a substrate (c) In vitro Pak1 kinase assay using wild-type or single or double mutants of SHARP proteins as substrates (d) HEK 293 cells were transfected with T7-SHARP and metabolically labeled with [32P]orthophosphoric acid After 48 h, cell lysates were immunoprecipitated with control IgG or T7 monoclonal antibody and the phosphorylation status of T7-SHARP was analysed by autoradiography (e) HEK 293 cells were transfected with wild-type or double mutant of SHARP (S3568AT3568A) After 48 h, cells were labeled with [32P]orthophosphoric acid and the phosphorylation status was analysed by autoradiography (f) HEK 293 cells were transfected with T7-SHARP (aa 3281–3661) with or without Pak1 siRNA vector Cells were metabolically labeled with [32P]orthophosphoric acid, and after 48 h, the phosphorylation of T7-SHARP (aa 3281–3661) was analysed by autoradiography (g) Schematic diagram showing the localization of Pak1 phosphorylation sites in the SHARP repression domain in the context of full-length SHARP (SHARP-S3486A, T3568A) Cotransfection of SHARP or SHARPS3486A, T3568A substantially reduced the ability of SHARP to reduce the reporter gene activity (Figure 3b) Similarly, inhibition of endogenous Pak1 activity by transfection of the Pak1-autoinhibitory domain (Pak1 aa 83–149) also reduced the SHARP-mediated repression of RBP-Jk-responsive reporter gene These results suggested that Pak1 phosphorylation modulates SHARP repression functions To further examine the role of the Pak1–SHARP interactions in Notch signaling, we used a dominant active form of Notch (constitutively activated Notch plasmid that lacks extracellular domain, Notch-1 del E) HEK 293 cells were transiently transfected with wildtype SHARP or SHARP-S3486A, T3568A, along with the RBP-Jk reporter gene As expected, Notch-l del E activated the RBP-Jk reporter gene, while coexpression of wild-type SHARP inhibited the reporter gene activity (Figure 4a) However, cotransfection of the SHARP Oncogene Role of Pak1 phosphorylation in Notch signaling RK Vadlamudi et al 4594 a pGa981-6-Luc VP16 RBP Relative pGA981-6 luc activity EBNA2 reselement 3000000 2500000 2000000 1500000 1000000 500000 Relative pGA981-6 luc activity pGa981-6 luc RBP-vp16 SHARP Pak1 (ng) b TATA + - + + - + + + - + + + 100 + + + 200 1200000 1000000 800000 600000 400000 200000 + + + + + + pGa981-6 luc - + - + + - + + + - + + - + + + RBP-vp16 Pak1 SHARP (WT) SHARP (DMT) Pak1 Inhibitor Figure Pak1 enhances SHARP-mediated repression (a) Schematic representation of pGa9811/6 luc construct and RBP-VP16 activator used in the reporter gene assay The reporter construct contains four EBNA2 response elements, which are recognized by RBP-Jk transcription factor In the RBP-VP16 activator, RBP-Jk was fused to the VP16 activation domain and thus acts as a constitutive activator of RBP-Jk target genes HEK 293 cells were cotransfected with RBP-Jk reporter gene pGa981/6 luc (2 mg) and RBP-VP16 (100 ng), with or without increasing amounts of SHARP (100 and 200 ng) and Pak1 (100 ng) expression plasmids After 48 h, luciferase activity was determined by using a comertial luciferase assay kit (Promega) and normalized to b-gal activity (b) HEK 293 cells were cotransfected with pGA981/6 luc and RBP-VP16 together with SHARP wild type or mutant that lacks Pak1 phosphorylation sites When indicated, Pak1 wild type or Pak inhibitor (Pak1 aa 83–149) was included in the cotransfection After 48 h, luciferase activity was measured and normalized to b-gal activity mutant, which lacks Pak1 phosphorylation sites, failed to inhibit Notch-mediated activation of RBP-Jk reporter gene (Figure 4a) Similarly, cotransfection of Pak1-specific siRNA vector (Li et al., 2003), which reduces Pak1 endogenous levels or Pak1-autoinhibitory fragment (aa 83–149), also interfered with SHARP Oncogene repression of Notch-mediated transactivation of the reporter gene (Figure 4b) Together, these findings suggest that Pak1 phosphorylation of SHARP is important for a productive repression of Notch target genes by SHARP To further implicate the potential contribution of Pak1 regulation of SHARP in exerting transcriptional repressor activity of SHARP, we next examined the effects of these regulatory interaction upon the transactivation function of Notch on Hes-1, an accepted physiological target of Notch (Ohtsuka et al., 1999) In this context, NIH3T3 murine fibroblast cells were transfected with HES-1 promoter reporter, Notch-l del E and either SHARP alone or combined with wild-type Pak1 or Pak1-autoinhibitory domain expressing plasmids, and reporter activity was measured after 48 h (Figure 4c) As expected from the previous work, Notch del E activated the Hes-1 promoter-luc activity, while SHARP coexpression resulted in a distinct reduction by 50% (Figure 4c) Interestingly, we also observed a further significant repression of Hes-1 promoter-luc activity when cells were cotransfected with Pak1, whereas Pak1 inhibitor relieved the repression To validate the noted repression of Hes-1 promoter activity by coexpression of SHARP and Pak1, we next depleted the endogenous Pak1 using a Pak1-specific siRNA in NIH3T3 cells, and examine the status of Hes1 mRNA by RT–PCR analysis We found that a reduced expression of Pak1 leads to a significant and reproducible upregulation of Hes-1 mRNA as compared to the cells transfected with the control siRNA (Figure 4d) Together, these findings suggest that Pak1 signaling indeed regulates the levels of Notch targets in vivo Phosphorylation-dependent signaling has been proposed as a potential mechanism for regulation of deacetylase-catalysed transcriptional repression (Galasinski et al., 2002) Serine phosphorylation of SMART by CK2 (Zhou et al., 2001) or by protein kinase C (PKC) (Ishaq et al., 2003) has been shown to enhance the repression function of SMRT Likewise, Pak1 phosphorylates corepressor CtBP and modulates its corepressor functions (Barnes et al., 2003) The results from the present study suggest that Pak1 interacts with and phosphorylates SHARP in its C-terminal repression domain, which has been previously shown to interacts with SMART (Shi et al., 2001) Cotransfection of Pak1 also promoted the repression activity of SHARP, while downregulation of Pak1 function or mutation of Pak1 phosphorylation sites in SHARP interfered with SHARP repression Since SHARP is a SMRT/ HDAC1-associated repressor protein (Shi et al., 2001), Pak1 phosphorylation of SHARP may promote its repressor functions in a way similar to PKC phosphorylation of SMART by enhancing its interactions with SMART Alternatively, it is also possible that SHARP– Pak1 interaction may allow Pak1 recruitment to the corepressor complex, where Pak1 may phosphorylate other components of the complex SHARP was initially identified as a SMART/HDACassociated protein and was implicated in nuclear receptor signaling as corepressor (Shi et al., 2001) Role of Pak1 phosphorylation in Notch signaling RK Vadlamudi et al 4595 Relative pGa981-6-Luc activity Notch E RBP 45000000 90000000 80000000 70000000 60000000 50000000 40000000 30000000 20000000 10000000 pGa981-6 luc Notch E SHARP (WT) Pak1 Inhibitor Pak1 siRNA 40000000 35000000 30000000 + - + + - + + + - + + + + - + + + + 6000000 5000000 4000000 3000000 2000000 1000000 Hes1-Luc Notch.E SHARP Pak1 Pak1-inhibitor + - + + - + + + - + + + + - + + + + 25000000 20000000 d 15000000 10000000 5000000 pGa981-6 luc Notch E SHARP (WT) SHARP (DMT) Con RNAi Pak1 RNAi RTPCR Relative pGA981-6-Luc activity TATA EBNA2 responsive elements c Relative Hes-1 Luc activity b pGa981-6-Luc + - + + - + + + - + + + WB a + - e + Notch Activation HES1 GAPDH RBP/ CBF-1 Pak1 Repression Vinculin Pak1 SHARP Figure Pak1 interferes with Notch-mediated activation of the RBP-Jk target gene (a) Schematic representation of the reporter gene assay used pGa981/6 luc reporter gene contains four RBP-Jk binding sites A dominant active form of Notch that lacks extracellular domain was used to activate RBP-Jk target genes by its constitutive interaction with endogenous RBP-Jk HEK 293 cells were cotransfected with a dominant active form of Notch (Notch del E, 100 ng) and pGa981/6 luc (2 mg), along with SHARP or SHARP (200 ng) ỵ Pak1 siRNA (200 ng) or SHARP (200 ng) ỵ Pak1-autoinhibitory domain aa 83149 (200 ng) expression vectors Luciferase reporter activity was measured after 48 h (b) HEK 293 cells were transfected with an activated form of Notch and pGa981/6 luc together with either wild-type SHARP or mutant SHARP that lacks Pak1 phosphorylation sites Luciferase activity was measured after 48 h (c) NIH3T3 cells were cotransfected with Hes-1 luciferase reporter (250 ng), a dominant active form of Notch (Notch del E, mg) and either SHARP alone or with Pak1 (500 ng) or Pak1-autoinhibitory domain aa 83–149 (500 ng) expression plasmids Luciferase activity was measured following 48 h after transfection (d) Pak1 was knock down in NIH3T3 using Pak1-specific RNAi Following 48 h after siRNA transfection, total RNA was extracted and the expression levels of HES-1 was monitored by RT–PCR analysis using HES-1-specific primers GAPDH was used as a control Pak1 expression was confirmed by Western blotting Vinculin was used as control (e) A working model showing Pak1 regulation of Notch signaling Subsequent studies showed that SHARP is an essential component of Notch signaling and plays a role in rescuing from Notch-dependent inhibition of primary neurogenesis (Oswald et al., 2002) In this context, Pak1 activation promotes neuronal dendrite initiation (Hayashi et al., 2002) Our finding that Pak1 phosphorylates SHARP and enhances its repressor functions suggests that Pak1-mediated phosphorylation of SHARP may constitute an important mechanism by which Pak1 may promote dendrite formation Our findings also raise a possibility that deactivation of the Pak1 pathway may be necessary for Notch-mediated productive neuronal differentiation It is therefore possible that Notch activation somehow downregulates Pak1 activity as a feedback mechanism CDK5/p35 kinase is widely expressed in neuronal cells and is shown to phosphorylate Pak1 and downregulate Pak1 activity in neuronal cells (Nikolic et al., 1998) Notch-mediated activation of CDK5/p35 kinase either directly or indirectly via activation of Abl kinase (Giniger, 1998) may constitute one possible mechanism by which Notch relieves Pak1SHARP-mediated repression In summary, we discovered and report here that SHARP is a Pak1-interacting protein and is a physiologic substrate of Pak1 These findings suggest that both Pak1 and SHARP may be critical molecules for repression of Notch target genes such as HES-1, and that Pak1 phosphorylation of SHARP might be essential for the productive repression function of SHARP Acknowledgements We thank Dr D Wu for providing pSuper-EGFP-Pak1siRNA, Dr Ronald Evans for SHARP expression plasmid, Dr Ronald M Schmid for RBP-Jk-VP16, pGa981/6 luc, Notch del E and Dr Ryoichiro Kageyama for Hes-1 promoter-luc reporter This study was supported by NIH Grants 90970 and 80066 (RK) References Barnes CJ, Vadlamudi RK, Mishra SK, Jacobson RH, Li F and Kumar R (2003) Nat Struct Biol., 10, 622–628 Beatus P and Lendahl U (1998) J Neurosci Res., 54, 125–136 Oncogene Role of Pak1 phosphorylation in Notch signaling RK Vadlamudi et al 4596 Bokoch GM (2003) Annu Rev Biochem., 72, 743–781 Daniels RH, Hall PS and Bokoch GM (1998) EMBO J., 17, 754–764 Galasinski SC, Resing KA, Goodrich JA and Ahn NG (2002) J Biol Chem., 277, 19618–19626 Giniger E (1998) Neuron, 20, 667–681 Hayashi K, Ohshima T and Mikoshiba K (2002) Mol Cell Neurosci., 20, 579–594 Hsieh JJ, Zhou S, Chen L, Young DB and Hayward SD (1999) Proc Natl Acad Sci USA, 96, 23–28 Ishaq M, DeGray G and Natarajan V (2003) J Biol Chem., 278, 39296–39302 Iso T, Kedes L and Hamamori Y (2003) J Cell Physiol., 194, 237–255 Kao HY, Ordentlich P, Koyano-Nakagawa N, Tang Z, Downes M, Kintner CR, Evans RM and Kadesch T (1998) Genes Dev., 12, 2269–2277 Li F, Adam L, Vadlamudi RK, Zhou H, Sen S, Chernoff J, Mandal M and Kumar R (2002) EMBO Rep., 3, 767–773 Li Z, Hannigan M, Mo Z, Liu B, Lu W, Wu Y, Smrcka AV, Wu G, Li L, Liu M, Huang CK and Wu D (2003) Cell, 114, 215–227 Oncogene Manser E, Leung T, Salihuddin H, Zhao ZS and Lim L (1994) Nature, 367, 40–46 Miele L and Osborne B (1999) J Cell Physiol., 181, 393–409 Nikolic M, Chou MM, Lu W, Mayer BJ and Tsai LH (1998) Nature, 395, 194–198 Ohtsuka T, Ishibashi M, Gradwohl G, Nakanishi S, Guillemot F and Kageyama F (1999) The EMBO J, 18, 2196–2207 Oswald F, Kostezka U, Astrahantseff K, Bourteele S, Dillinger K, Zechner U, Ludwig L, Wilda M, Hameister H, Knochel W, Liptay S and Schmid RM (2002) EMBO J., 21, 5417–5426 Schweisguth F (2004) Curr Biol., 14, R129–R138 Shi Y, Downes M, Xie W, Kao HY, Ordentlich P, Tsai CC, Hon M and Evans RM (2001) Genes Dev., 15, 1140–1151 Thiel DA, Reeder MK, Pfaff A, Coleman TR, Sells MA and Chernoff J (2002) Curr Biol., 12, 1227–1232 Vadlamudi RK and Kumar R (2003) Cancer Metast Rev., 22, 385–393 Zhou S and Hayward SD (2001) Mol Cell Biol., 21, 6222–6232 Zhou Y, Gross W, Hong SH and Privalsky ML (2001) Mol Cell Biochem., 220, 1–13