RACK1: a Novel Substrate for the Src Protein-Tyrosine Kinase

RACK1: a novel substrate for the Src protein-tyrosine kinase

Betty Y Chang1, Rachel A Harte1 and Christine A Cartwright*,1

1Department of Medicine, Stanford University, Stanford, California, CA 94305, USA

RACK1 is one of a group of PKC-interacting proteins endosomal membranes it may target proteins involved collectively called RACKs (Receptors for Activated C- in intracellular trafficking or transport; or if Src is Kinases). Previously, we showed that RACK1 also localized to the perinuclear membrane, it may target interacts with the Src tyrosine kinase, and is an inhibitor proteins involved in cell cycle regulation. Thus, by of Src activity and cell growth. PKC activation induces identifying distinct substrates of Src within distinct the intracellular movement and co-localization of subcellular compartments, we will learn about specific RACK1 and Src, and the tyrosine phosphorylation of functions of Src. RACK1. To determine whether RACK1 is a Src One example of a Src substrate, whose phosphoryla- substrate, we assessed phosphorylation of RACK1 by tion by Src appears to determine specific functions of various tyrosine kinases in vitro, and by kinase-active Src, is Sam68, an RNA binding protein (reviewed in and inactive mutants of Src in vivo. We found that Thomas and Brugge, 1997; Bjorge et al., 2000). Src RACK1 is a Src substrate. Moreover, Src activity is phosphorylates Sam68 during mitosis, presumably after necessary for both the tyrosine phosphorylation of breakdown of the nuclear envelope. Src appears to be RACK1 and the binding of RACK1 to Src’s SH2 important for regulating cell cycle progression via domain that occur following PKC activation. To identify Sam68, particularly during the late states of mitosis the tyrosine(s) on RACK1 that is phosphorylated by Src, and possibly during the G1/S transition. Evidence is we generated and tested a series of RACK1 mutants. We also emerging that Src may regulate gene expression at found that Src phosphorylates RACK1 on Tyr 228 and/ the post-transcriptional level via Sam68 and that Src or Tyr 246, highly-conserved tyrosines located in the can regulate mRNA splicing and/or transport sixth WD repeat that interact with Src’s SH2 domain. (reviewed in Bjorge et al., 2000). Thus, identification We think that RACK1 is an important Src substrate that and characterization of a single Src substrate has signals downstream of growth factor receptor tyrosine provided important information about the function of kinases and is involved in the regulation of Src function Src in regulating cell cycle progression and gene and cell growth. expression. Oncogene (2002) 21, 7619 – 7629. doi:10.1038/sj.onc. RACK1 (Receptor for Activated CKinase) was the 1206002 first of a group of protein kinase C (PKC)-interacting proteins that have been identified and characterized Keywords: Src; RACK1; PKC; signal transduction (reviewed in Mochly-Rosen, 1995; Mochly-Rosen and Kauvar, 1998; Schechtman and Mochly-Rosen, 2001). Using the yeast two-hybrid assay, we identified Introduction RACK1 as a novel Src-binding protein and a novel inhibitor of Src activity and cell growth (Chang et al., The Src tyrosine kinase participates in diverse signaling 1998). While a number of interacting proteins have pathways that regulate diverse cellular functions. These been identified that upregulate Src activity, few have include proliferation, differentiation, motility and been identified that downregulate Src activity. Because adhesion (reviewed in Martin, 2001; Thomas and it is the repression of c-Src activity rather than the Brugge, 1997; Abram and Courtneidge, 2000). The elevation of v-Src activity that accounts for differences subcellular localization of Src may determine, in part, is the transforming abilities of the two kinases, it is its substrate specificity and function. For example, if important to search for cellular mechanisms that Src is localized to the plasma membrane it may inactive c-Src. In doing so, we will learn about phosphorylate proteins that function in mitogenic mechanism by which normal cells regulate their signaling via growth factor receptor-tyrosine kinases, growth. or in cell adhesion, cell migration or cell – cell Previously, we found that PKC activation induces interactions. In contrast, if Src is localized to the intracellular movement and co-localization of RACK1 and Src at the plasma membrane, and the tyrosine phosphorylation of RACK1 (Chang et al., *Correspondence: CA Cartwright, CCSR Building, Room 3115-C, 2001). Together, these findings suggested that RACK1 269 Campus Drive, Stanford University School of Medicine, is a Src substrate. However, Src is only one of many Stanford, California, CA 94305-5187, USA; E-mail: [email protected] tyrosine kinases that could potentially phosphorylate Received 3 May 2002; revised 20 August 2002; accepted 29 August RACK1 in cells. The purpose of this study was to 2002 determine whether RACK1 is a Src substrate, and if RACK1: a novel substrate for the Src kinase BY Chang et al 7620 so, to identify the site(s) on RACK1 phosphorylated by Src. Using kinase active and inactive mutants of Src, we found that RACK1 is a Src substrate. Moreover, Src activity is necessary for both the tyrosine phosphorylation and the binding of RACK1 to Src’s SH2 domain that occur following PKC activation. Using mutants of RACK1 that contained phenylalanine substitutions for tyrosines, we found that Src phosphorylates RACK1 on Tyr 228 and/or Tyr 246, highly-conserved tyrosines located in the sixth WD repeat that interact with Src’s SH2 domain.

Results

RACK1 is an in vitro substrate for Src and not for the Abl, EGFR or PDGFR tyrosine kinases Previously, we found that RACK1 is phosphorylated on tyrosine (Chang et al., 2001). We also found that PKC activation induces the intracellular movement and co-localization of RACK1 and Src, and the tyrosine phosphorylation of RACK1. These findings suggested that Src is one tyrosine kinase that phosphorylates RACK1. To identify tyrosine kinases that phosphorylate RACK1 in vitro, we incubated equivalent amounts of purified GST or GST – RACK1 with purified Src (expressed from recombinant baculovirus and purified from Sf9 cells), Abl, EGFR or PDGFR Figure 1 Phosphorylation of RACK1 by Src in vitro.(a) Phos- tyrosine kinase (normalized to equivalent amounts of phorylation of GST – RACK1 by purified receptor and non-receptor tyrosine kinases. Purified EGFR (lanes 1 and 2), Src (lanes 3 specific activity) and performed in vitro protein-kinase and 4), Abl (lanes 5 and 6) or PDGFR (lanes 7 and 8) tyrosine assays in the presence of MnCl2, MgCl2 and kinases (normalized to equivalent amounts of specific activity) [g-32P]ATP (Figure 1a). As expected, we observed were incubated with equivalent amounts of purified GST (odd 32 autophosphorylation of all of the tyrosine kinases. In lanes) or GST – RACK1 (even lanes) and [g- P]ATP, MgCl2 and MnCl2, for 10 min at 308Cinanin vitro protein kinase assay. addition, we found that Src phosphorylated GST – 32P-labeled proteins were resolved by SDS – PAGE and visualized RACK1 (Figure 1a, lane 4) to higher stoichiometry by autoradiography. Approximate molecular masses of the ki- than did the other tyrosine kinases tested. Src also nases (kDa): EGFR, 70; c-Src, 60; c-Abl, 150; PDGFR, 180. phosphorylated GST (Figure 1a, lane 3), but at The smeared bands in lanes 5 and 6 may represents poor migra- significantly lower levels than it phosphorylated tion of the Abl kinase through the gel. RK, RACK1. (b) Phos- phorylation of GST – RACK1 by Src immunoprecipitates. GST – RACK1. To determine whether Src was phos- Proteins were immunoprecipitated with a monoclonal antibody phorylating RACK1 or the GST portion of the fusion specific for Src, MAb 327 (lanes 1 – 6 and 8 – 10) or IgG (lane protein, we proteolytically-cleaved RACK1 from 7) from NIH3T3 cell lysates expressing Y527F c-Src (lanes 1, 2 GST – RACK1 using thrombin, incubated the purified and 8), wild-type c-Src (lanes 3, 4 and 9) or vector alone (lanes 5 – 7 and 10). Immunoprecipitates of Src (lanes 1 – 6), or IgG (lane RACK1 with Src and performed an in vitro kinase 7), were incubated with equivalent amounts of purified GST (odd assay. We observed phosphorylation of the purified lanes) or GST – RACK1 (even lanes) and subjected to in vitro RACK1 by Src (data not shown). Therefore, under the protein kinase assays as described in (a), or subjected to immuno- conditions that we used, Src phosphorylates RACK1 in blot analysis with anti-Src (lanes 8 – 10) vitro, whereas the other non-receptor and receptor tyrosine kinases that we tested, do not. Interestingly, the Abl tyrosine kinase, which is evolutionarily close to tates with GST or GST – RACK1 and performed in Src in terms of catalytic-domain substrate specificity, vitro kinase assays (Figure 1b). We found that did not phosphorylate GST – RACK1 (Figure 1a, lane immunoprecipitates of Y527F Src phosphorylated 6). Previously, we showed that the Abl SH2 domain GST – RACK1 (Figure 1b, lane 2) but not GST does not binds to GST – RACK1 either (Chang et al., (Figure 1b, lane 1). A longer exposure of the 2001). autoradiogram (not shown) revealed that immunopre- To determine whether immunoprecipitates of Src cipitates of wild-type Src also phosphorylated GST – phosphorylate GST – RACK1, and whether the specific RACK1, although at significantly lower levels than did activity of Src affects its ability to phosphorylate those of Y527F Src. As expected, we observed RACK1, we immunoprecipitated similar amounts of autophosphorylation of Y527F and wild-type Src Src protein from NIH3T3 cells overexpressing Y527F (Figure 1b, lanes 1 – 4). The amount of Src present in c-Src (a partially activated and transforming c-Src each immunoprecipitate was similar (Figure 1b, lanes mutant) or wild-type c-Src, incubated immunoprecipi- 8 – 10). Thus, immunoprecipitates of Src phosphorylate

Oncogene RACK1: a novel substrate for the Src kinase BY Chang et al 7621 GST – RACK1, and immunoprecipitates of Src kinases immunoprecipitates (Figure 2, lanes 1, 3 and 5). Thus, that have elevated specific activity phosphorylate RACK1 appeared to be an endogenous substrate for v- GST – RACK1 at higher levels than does normal c-Src. Src. We also tested the following c-Src kinases, each with a different level of specific activity, for their ability to RACK1 is an endogenous substrate for Src phosphorylate and associate with endogenous RACK1: To determine whether RACK1 is an endogenous (1) Y527F c-Src, a partially activated and transforming substrate for Src, we assessed the tyrosine phosphor- c-Src mutant with a specific activity that is about one- ylation of RACK1 in cell lines that expressed Src with half that of v-Src (Cartwright et al., 1987; Kmiecik and different levels of specific activity. We and others have Shalloway, 1987; Piwnica-Worms et al., 1987; Reynolds been unsuccessful in attempts to generate antibodies et al., 1987); (2) Y416F/Y527F c-Src, a partially that immunoprecipitate RACK1 efficiently from cell activated c-Src mutant with a specific activity that is lysates. Thus, to study tyrosine-phosphorylated about one-half that of Y527F c-Src (Kmiecik and RACK1, we immunoprecipitate proteins with a Shalloway, 1987); (3) Y416F c-Src, a c-Src mutant with monoclonal antibody that is specific for phosphotyr- a specific activity that is about one-tenth that of Y527F osine and perform immunoblot analyses with a c-Src and about equal to that of wild-type c-Src monoclonal antibody that is specific for RACK1 (Kmiecik and Shalloway, 1987); and (4) wild-type c- (Chang et al., 1998, 2001). Using this approach, we Src. To assess the tyrosine phosphorylation of RACK1 made RIPA lysates of NIH3T3 cells, NIH3T3 cells that by the c-Src mutants, we immunoprecipitated proteins were expressing v-Src, or CHO cells that were with anti-phosphotyrosine or IgG from RIPA lysates expressing Y527F c-Src and HA – RACK1, immuno- of NIH3T3 cell lines that were stably-expressing one of precipitated proteins with anti-phosphotyrosine or IgG the mutant Src kinases and performed immunoblot and performed immunoblot analysis with anti-RACK1 analysis with anti-RACK1 (Figure 3a). We observed (Figure 2). Y527 c-Src was transiently expressed at the highest level of tyrosine phosphorylation on significantly higher levels in the CHO cells than v-Src RACK1 in cells expressing Y527F c-Src (Figure 3, was stably expressed in the NIH3T3 cells (data not shown). HA – RACK1 contains a nine-amino acid, influenza virus hemagglutinin (HA)-tag; thus, it is approximately 1 kDa larger than endogenous RACK1. We detected tyrosine-phosphorylated endogenous RACK1 (arrowhead) and HA – RACK1 (arrow) in CHO cells that were overexpressingY527F Src and HA – RACK1 (Figure 2, lane 6). We detected a tyrosine-phosphorylated protein in v-Src-transformed NIH3T3 cells (Figure 2, lane 4) that co-migrated with endogenous RACK1 in CHO cells (Figure 2, lane 6) and is, therefore, presumably RACK1. The protein was not detected in anti-phosphotyrosine immunoprecipitates of NIH3T3 cells (Figure 2, lane 2), or in IgG

Figure 3 Phosphorylation of RACK1 by c-Src in vivo.(a) Tyro- sine phosphorylation of endogenous RACK1 in NIH3T3 cell lines expressing c-Src mutants with various levels of specific activity. Proteins were immunoprecipitated with MAb Py20 (lanes 1 – 4) or IgG (lanes 5 – 7) from RIPA lysates of NIH3T3 cell lines expressing Y527F c-Src (lanes 1 and 5), Y416F/Y527F c-Src (lanes 2 and 6), Y416F c-Src (lanes 3 and 7) or wild-type c-Src (lane 4), and subjected to immunoblot analysis with anti-RACK1. Relative specific activities of c-Src mutants: Y527F, one-half that of v-Src; Y416F/Y527F, one-half that of Y527F; Y416F, one-tenth that of Y527F and about equal to that of wild-type c-Src. wt, wild-type. A nonspecific band running slightly above RACK1 was noted in some IgG immunoprecipitates (lanes 6 and 7). (b) Binding of en- Figure 2 Phosphorylation of RACK1 by v-Src in vivo. Proteins dogenous RACK1 that has been phosphorylated by c-Src mu- were immunoprecipitated with IgG (odd lanes) or a monoclonal tants, to Src’s SH2 domain. Lysates of the cell lines described antibody specific for phosphotyrosine (even lanes) from RIPA ly- in (a) were incubated with a GST fusion protein containing Src’s sates of NIH3T3 cells (lanes 1 and 2), NIH3T3 cells expressing v- SH2 domain (lanes 5 – 8) or GST alone (lanes 9 – 12). Protein Src (lanes 3 and 4) or CHO cells expressing Y527 c-Src and HA – complexes were collected on glutathione-agarose beads and sub- RACK1 (lanes 5 and 6), and subjected to immunoblot analysis jected to immunoblot analysis with anti-RACK1. Lanes 1 – 4: with a monoclonal antibody specific for RACK1. pTyr, anti- Equivalent amounts of cell lysate were loaded directly on the phosphotyrosine MAb Py20 gel prior to transfer and immunoblot analysis with anti-RACK1

Oncogene RACK1: a novel substrate for the Src kinase BY Chang et al 7622 lane 1), intermediate levels of tyrosine phosphorylation on RACK1 in cells expressing Y416F/Y527F c-Src (Figure 3, lane 2), and low level of tyrosine phosphorylation on RACK1 in cells expressing Y416F (Figure 3, lane 3) or wild-type c-Src (Figure 3, lane 4). A nonspeciﬁc band running slightly above RACK1 was noted in some IgG immunoprecipitates (Figure 3, lanes 6 and 7). Thus, kinase-active c-Src phosphorylates endogenous RACK1. To assess the ability of endogenous RACK1 that had been phosphorylated by the c-Src mutants, to bind Src’s SH2 domain, we incubated GST or GST-Src-SH2 (a GST fusion protein containing the SH2 domain of Src) with RIPA lysates of the NIH3T3 cell lines that were stably- expressing one of the mutant c-Src kinases, collected protein complexes on glutathione-agarose beads and performed immunoblot analysis with anti-RACK1 (Figure 3b). We observed the most binding of GST- Src-SH2 to RACK1 from cells expressing Y527F c-Src (Figure 3b, lane 5) or Y416F/Y527F c-Src (Figure 3b, lane 6), and the least binding of GST-Src-SH2 to RACK1 from cells expressing Y416F (Figure 3b, lane 7) or wild-type c-Src (Figure 3b, lane 8). The amount of RACK1 expressed in each cell line was similar (compare Figure 3b, lanes 1 – 4), and lysate RACK1 did not bind to GST alone (Figure 3b, lanes 9 – 12). Thus, endogenous RACK1 that had been phosphorylated by the c-Src mutants binds to Src’s SH2 domain. Figure 4 Phosphorylation of RACK1 by Src and binding of Moreover, the higher the kinase activity of Src, the RACK1 to Src following PMA treatment. CHO cells were transfected with vector alone, HA – RACK1 or HA – RACK1 together more tyrosine-phosphorylation of RACK1 that occurs with wild-type or mutant c-Src: Y527F c-Src (constitutively-ac- (presumably by Src), and the more tyrosine-phosphor- tive), K297M c-Src (kinase-inactive) or dl155 c-Src (contains a ylation of RACK1 that occurs, the more binding there three amino acid deletion in the phosphotyrosine binding pocket is of RACK1 to Src. The higher levels of Src kinase of the SH2 domain, and is defective for binding to RACK1). Cells were treated with PMA for 15 min. (a) Phosphorylation of activity could be due to higher speciﬁc activity of the RACK1 by Src. Proteins were immunoprecipitated from RIPA ly- kinase and/or to higher levels of Src protein expression. sates with anti-phosphotyrosine and subjected to immunoblot The higher levels of tyrosine-phosphorylated RACK1 analysis with anti-RACK1. (b) Binding of RACK1 to Src. Lysate detected could be due to phosphorylation of a greater proteins (lanes 1, 13 and 14) or proteins immunoprecipitated with number of RACK1 molecules and/or to phosphoryla- MAb 327 (lanes 2 – 7) or IgG (lanes 8 – 12) were subjected to immunoblot analysis with anti-RACK1. (c) Src protein levels. Pro- tion of a greater number of tyrosines on individual teins were immunoprecipitated with MAb 327 from cell lysates RACK1 molecules. Together, the results from Figures and subjected to immunoblot analysis with anti-Src 2 and 3 show that RACK1 is an endogenous substrate for Src. Moreover, endogenous RACK1 that has been phosphorylated by Src, binds to Src’s SH2 domain. showed that the dl155 Src mutant has reduced binding to RACK1 (Chang et al., 2001). We observed the Src activity is necessary for both the tyrosine highest levels of tyrosine phosphorylation on RACK1 phosphorylation of RACK1 and the binding of RACK1 to in cells expressing Y527F c-Src (Figure 4a, lane 4), Src’s SH2 domain that occur following PKC activation intermediate levels of tyrosine phosphorylation on To determine whether Src activity is necessary for the RACK1 in cells expressing wild-type c-Src (Figure 4a, tyrosine phosphorylation of RACK1 that occurs lane 5), low levels of tyrosine phosphorylation on following PKC activation, we expressed HA – RACK1 RACK1 in cells expressing dl155 c-Src (Figure 4a, lane or HA – RACK1 together with kinase-active or inactive 3) and undetectable levels of tyrosine phosphorylation Src mutants in CHO cells, treated cells with PMA for on RACK1 in cells expressing kinase-inactive K297M 15 min, immunoprecipitate proteins with anti-phospho- Src (Figure 4a, lane 6), HA – RACK1 alone (Figure 4a, tyrosine and performed immunoblot analysis with anti- lane 2) or vector alone (Figure 4a, lane 1). The amount RACK1 (Figure 4a). We used three Src mutants: (1) of HA – RACK1 expressed in each transfection was Y527F c-Src (constitutively-active) (Cartwright et al., similar (Figure 4b, compare lanes 1, 13 and 14). Thus, 1987); (2) K297M Src (kinase-inactive) (Snyder et al., Src activity is necessary for the tyrosine phosphoryla- 1985); and (3) dl155 Src (contains a three amino acid tion of RACK1, because RACK1 is not deletion in the phosphotyrosine binding pocket of the phosphorylated by kinase-inactive Src. To determine SH2 domain) (Moyers et al., 1993). Previously, we whether Src activity is required for the binding of

Oncogene RACK1: a novel substrate for the Src kinase BY Chang et al 7623 RACK1 to Src’s SH2 domain that occurs following PKC activation, we immunoprecipitated Src from cell lysates and performed immunoblot analysis with anti- RACK1 (Figure 4b). We observed the most binding of RACK1 to Y527F Src (Figure 4b, lane 5), less binding of RACK1 to wild-type Src (Figure 4b, lane 6) and the least binding of RACK1 to dl155 Src (Figure 4b, lane 4) and K297M Src (Figure 4b, lane 7). In fact, the amount of RACK1 present in dl155 and K297M Src immunoprecipitates was not more than could be accounted for by endogenous Src (Figure 4b, lane 3). As expected, RACK1 did not bind to IgG immunoprecipitates (Figure 4b, lanes 8 – 12). The amount of Src present in each immunoprecipitate was similar (Figure 4c, lanes 3 – 6). Thus, Src activity is necessary for both the tyrosine phosphorylation of RACK1 and the binding of RACK1 to Src’s SH2 domain that occur following PKC activation.

Src phosphorylates RACK1 in the carboxy-terminal region of the molecule To begin identifying the Src phosphorylation site(s) on RACK1, we generated a series of N- or C-terminal truncated mutants of RACK1 (see Figure 6). Initially, we co-expressed the N-terminal mutants or full-length RACK1 together with Y527F c-Src and performed immunoblot analysis with anti-phosphotyrosine (Figure 5a). RACK1 mutant 5.1 lacks amino acids Figure 5 Src phosphorylation of RACK1 in the carboxy-terminal region of the molecule. CHO cells were transfected with vec- 1 – 41, 5.2 lacks amino acids 1 – 93, and 5.3 lacks amino tor alone, HA – RACK1, Y527 Src and HA – RACK1, or Y527F acids 1 – 137. As expected, we observed more tyrosine- Src and various N-terminal truncation mutants of RACK1. phosphorylated proteins, and higher levels of tyrosine- RACK1 mutant 5.1 lacks amino acids 1 – 41, 5.2 lacks amino phosphorylation on proteins, in cells that were acids 1 – 93 and 5.3 lacks amino acids 1 – 137. (a and c) Lysate transfected with Y527F c-Src (Figure 5a, lanes 3 – 6) proteins were subjected to immunoblot analysis with anti-phosphotyrosine. (b) The blot shown in (a) was stripped of antibody than in cells that were not transfected with Src (Figure and reprobed with anti-RACK1. (d) A parallel gel to that shown 5a, lanes 1 and 2). In cells expressing Y527F c-Src in (c) was subjected to immunoblot analysis with anti-RACK1 (Figure 5a, lanes 3 – 6), we also observed a tyrosine- phosphorylated protein with a molecular mass that was appropriate for endogenous RACK1. In cells expressing Y527F c-Src and full-length, wild-type HA – proteins that we thought were RACK1 proteins on the RACK1 (Figure 5a, lane 3), we detected a tyrosine- anti-phosphotyrosine blot were, in fact, RACK1 phosphorylated protein that was slightly larger than because they co-migrated with RACK1 on the anti- endogenous RACK1 and had a molecular mass that RACK1 blot. In a repeat experiment, the RACK1 – 5.3 was appropriate for HA – RACK1. In cells expressing mutant was expressed at higher levels (Figure 5d, lane Y527F c-Src and RACK1 – 5.1 (Figure 5a, lane 4), we 4) and was more clearly tyrosine-phosphorylated observed a tyrosine-phosphorylated protein with a (Figure 5c, lane 4). RACK1 – 5.3 lacks WD repeats molecular mass that was appropriate for a RACK1 1 – 3 and contains WD repeats 4 – 7 (see Figure 6). protein lacking 41 amino acids. In cells expressing Thus, Src phosphorylates RACK1 on a tyrosine(s) Y527F c-Src and RACK1 – 5.2 (Figure 5a, lane 5), we located in one of the four carboxy-terminal WD observed a tyrosine-phosphorylated protein with a repeats. molecular mass that was appropriate for a RACK1 protein lacking 93 amino acids. Finally, in cells Src phosphorylates RACK1 in the sixth WD repeat, on expressing Y527F c-Src and RACK1 – 5.3 (Figure 5a, Tyr 228 and/or 246 lane 6), we observed a tyrosine-phosphorylated protein with a molecular mass that was appropriate for a To further identify the tyrosine(s) on RACK1 that is RACK1 protein lacking 137 amino acids. The blot was phosphorylated by Src, we expressed the N- or C- then stripped of antibody and re-probed with anti- terminal truncated mutants of RACK1 as GST fusion RACK1 (Figure 5b). We observed that the mobility of proteins and tested them for phosphorylation by Src each RACK1 mutant was appropriate for its predicted (which had been expressed in baculovirus and puriﬁed molecular mass. Overlaying the autoradiograms from from Sf9 cells) in an in vitro kinase assay. The results Figure 5a,b revealed that the tyrosine-phosphorylated are summarized schematically in Figure 6 (left column).

Oncogene RACK1: a novel substrate for the Src kinase BY Chang et al 7624 A We observed that all RACK1 mutants that retained the sixth WD repeat were phosphorylated by Src in vitro, whereas those that had lost this repeat were not. Next, we tested the truncated RACK1 mutants for phosphorylation by Src in vivo. To do so, we co- expressed Src and the HA – RACK1 mutants transiently in CHO cells, treated cells with PMA for 15 min, immunoprecipitated proteins with anti-phosphotyrosine and performed immunoblot analysis with anti- RACK1. HA – RACK1 protein levels were measured by immunoblot analysis of cell lysates with anti- RACK1. Because the amount of mutant HA – RACK1 expressed in cells varied with each transfection, data were quantified by scanning densitometry of bands, expressed as a ratio of tyrosine-phosphorylated mutant HA – RACK1 to levels of mutant HA – RACK1 protein expressed in cells, and compared with the ratio for wild-type RACK1. We found that RACK1 mutants that contained the sixth WD repeat and were phosphorylated by Src in vitro, were also phosphorylated by Src in vivo (Figure 6, right column). In B contrast, RACK1 mutants that lacked the sixth WD repeat and were not phosphorylated by Src in vitro, were not phosphorylated by Src in vivo. Thus, Src phosphorylates RACK1 on a tyrosine(s) located in the sixth WD repeat. We also subjected lysates of cells expressing C- terminal truncation mutants of RACK1 to immunoblot analysis with anti-phosphotyrosine (Figure 6b, lanes 1 – 6). In lysates containing wild-type RACK1 (1 – 317), we detected a tyrosine-phosphorylated 36 kDa protein (Figure 6b, lane 6), which is the predicted size for a full-length RACK1 protein. In lysates containing a RACK1 mutant that lacks WD repeat seven (1 – 258), we detected a tyrosine-phosphorylated 28 kDa protein (Figure 6b, lane 5), which is the predicted size for a RACK1 protein lacking 59 amino acids. However, in lysates containing a RACK1 mutant that lacks WD Figure 6 Src phosphorylation of RACK1 in the sixth WD re- repeats six and seven (1 – 225), we did not detect a peat. A series of N- and C-terminal truncated RACK1 mutants tyrosine-phosphorylated 25 kDa protein (Figure 6b, were generated. (a) Left column: Src phosphorylation of truncated RACK1 mutants in vitro. Truncated RACK1 mutants or lane 4), which is the predicted size for a RACK1 full-length, wild-type RACK1 (1 – 317) were expressed as GST fu- protein lacking 92 amino acids. Nor did we detect a sion proteins and analysed for phosphorylation by purified Src in tyrosine-phosphorylated 20 kDa protein in lysates an in vitro protein kinase assay. Right column: Src phosphoryla- containing a mutant that lacks repeats five – seven tion of truncated RACK1 mutants in vivo. CHO cells were co- (1 – 183; Figure 6b, lane 3), a tyrosine-phosphorylated transfected with Src and wild-type or mutant RACK1. Cells were treated with PMA for 15 min. Proteins were immunoprecipitated 14 kDa protein in lysates containing a mutant that with anti-phosphotyrosine and subjected to immunoblot analysis lacks repeats four – seven (1 – 125; Figure 6b, lane 2) or with anti-RACK1. Data on levels of Src, mutant RACK1 and a tyrosine-phosphorylated 11 kDa protein in lysates tyrosine-phosphorylated mutant RACK were quantified by scan- containing a mutants that lacks repeats three – seven ning densitometry of bands. Because the amount of Src and mutant RACK1 protein expressed in cells varied with each (1 – 97; Figure 6b, lane 1). Together, our results show tranfection, the phosphotyrosine signals were normalized to the that Src phosphorylates RACK1 on a tyrosine(s) amount of HA – RACK1 and Src expressed. Results were then located in the sixth WD repeat. compared to those for wild-type HA – RACK1. Results were simi- To identify the specific tyrosine(s) in RACK1 that is lar for duplicate transfections for each RACK1 mutant, and for phosphorylated by Src in vivo, we performed site- two independent experiments performed. (+): RACK1 was tyrosine phosphorylated; ( – ): RACK1 was not tyrosine phosphory- directed mutagenesis, substituting phenylalanine for lated; (NT): not tested. (b) Src phosphorylation of RACK1 tyrosine at individual and multiple sites in RACK1. mutants in vivo. CHO cells were co-transfected with Src and The RACK1 mutants were then transiently co- wild-type or mutant RACK1, and treated with PMA as described expressed with Src in CHO cells and proteins were above. Cell lysates were subjected to immunoblot analysis with anti-phosphotyrosine. Predicted molecular masses (kDa) of immunoprecipitated from RIPA lysates with anti- RACK1 proteins: 1 – 317 (full-length), 35; 1 – 258, 28; 1 – 225, phosphotyrosine and subjected to immunoblot analysis 25; 1 – 183, 20; 1 – 125, 14; 1 – 97, 11 with anti-RACK1 (Figure 7a). We found that the

Oncogene RACK1: a novel substrate for the Src kinase BY Chang et al 7625 A mold and yeast). We found that all six tyrosines in RACK1 are highly conserved from one species to another (data not shown). Moreover, the sequence flanking Tyr 246 is highly conserved from yeast to human (Figure 7b). The sequence flanking Tyr 228 is less well conserved (Figure 7b), and the sequence flanking other tyrosines in RACK1 is not conserved between species (data not shown). Thus, the Src substrate and binding sites in RACK1, and the sequences that flank them, have been conserved during eukaryotic evolution.

Discussion

This study shows that: (1) RACK1 is a Src substrate; (2) Src activity is necessary for both the tyrosine phosphorylation of RACK1 and the binding of RACK1 to Src’s SH2 domain that occur following PKC activation; and (3) Src phosphorylates RACK1 in B the sixth WD repeat, on highly-conserved Tyr 228 and/ or Tyr 246. Evidence that RACK1 is a Src substrate is that Src and not the Abl, EGFR or PDGFR tyrosine kinases phosphorylated RACK1 in vitro (Figure 1), and that kinase-active forms of Src (v-Src and Y527F c-Src) phosphorylated endogenous RACK1 in vivo (Figures 2 and 3). Evidence that Src activity is necessary for the tyrosine phosphorylation of RACK1 that occurs following PKC activation is that kinase- active forms of Src phosphorylated RACK1 following PMA stimulation, whereas kinase-inactive K297M Src did not (Figure 4a). Evidence that Src activity is Figure 7 Src phosphorylation of RACK1 on highly-conserved Tyr 228 and/or Tyr 246. (a) Src phosphorylation of RACK1 on required for the binding of RACK1 to Src’s SH2 Tyr 228 and/or 246. Site-directed mutagenesis of RACK1 was domain that occurs following PKC activation is that performed, substituting phenylalanine for tyrosine at individual kinase-active forms of Src bound to RACK1 following and multiple sites. RACK1 mutants were co-expressed with Src PMA stimulation, whereas kinase-inactive K297M Src, in CHO cells. Cells were treated with sodium vanadate for 30 min and with PMA for 15 min prior to lysis. Proteins were im- and a Src mutant containing a small deletion in the munoprecipitated with anti-phosphotyrosine and subjected to im- phosphotyrosine-binding pocket of the SH2 domain, munoblot analysis with anti-RACK1. Data were analysed as did not (Figure 4b). Evidence that Src phosphorylates described in the legend to Figure 6. Results were similar for du- RACK1 in the sixth WD repeat is that all truncated plicate plates of cells co-transfected with Src and each RACK1 RACK1 mutants tested that contained the sixth WD mutant, and for two independent experiments performed. (b) Conservation of Tyr 228 and Tyr 246 during eukaryotic evolu- repeat were phosphorylated by Src (both in vivo and in tion. RACK1 sequences from eight eukaryotic organisms were vitro), whereas all mutants that lacked the sixth WD aligned. Comparative sequence for the sixth WD repeat are repeat were not (Figures 5 and 6). Finally, evidence shown that Src phosphorylates RACK1 on Tyr 228 and/or Tyr 246 is that all RACK1 mutants tested that RACK1 mutants that contained phenylalanine at contained tyrosine at either one or both of those positions 228 and 246 (Y140-302F, Y140-246F and positions were phosphorylated by Src, whereas all Y228-302F), were not phosphorylated on tyrosine by RACK1 mutants that contained phenylalanine at both Src. In contrast, the RACK1 mutants that contained position 228 and 246 were not (Figure 7a). tyrosine at either one (Y246F or Y228F; Figure 6b, The findings reported here extend our previous lanes 7 and 8 respectively) or both (Y52-194, 302F) of findings that PKC activation induces the intracellular these positions, were phosphorylated on tyrosine by movement and co-localization of RACK1 and Src, and Src. Thus, Src phosphorylates RACK1 on Tyr 228 the tyrosine phosphorylation of RACK1 (Chang et al., and/or 246. Previously, we showed that Tyr 228 and/or 2001). They provide new information about one kinase 246, when phosphorylated, interact with Src’s SH2 involved in the tyrosine phosphorylation of RACK1. domain (Chang et al., 2001). Other findings, demonstrating that RACK1 interacts To determine whether the Src substrate and binding with the receptor tyrosine phosphatase PTPm, and that sites in RACK1 have been conserved during eukaryotic this interaction is disrupted by the presence of evolution, we compared RACK1 sequences from eight constitutively-active Src (Mourton et al., 2001), species (human, rat, chicken, zebrafish, fly, tobacco, support our findings that RACK1 is tyrosine phos-

Oncogene RACK1: a novel substrate for the Src kinase BY Chang et al 7626 phorylated and that Src is one tyrosine kinase that stream targets and/or by competing with the PDGFR phosphorylates RACK1. Moreover, the observation for binding to Src’s SH2 domain, thereby dissociating that, as cell density increases, cytoplasmic RACK1 Src from the PDGFR and consequently, the mechan- translocates to the plasma membrane and to cell – cell ism by which Src is activated. PTPm and possibly contacts where it associates with PTPm (Mourton et al., other tyrosine phosphatases could, in turn, depho- 2001), resembles our observation that, upon PKC sphorylate RACK1, thereby dissociating RACK1 activation, RACK1 translocates to the plasma from Src and other signaling molecules. membrane where it co-localizes with Src (Chang et Our previous findings that serum or PDGF stimula- al., 2001). Together, the results indicate that, following tion or PKC activation enhance both the tyrosine various stimuli, RACK1 moves to the plasma phosphorylation of RACK1 and the binding of membrane and associates with tyrosine kinases and phosphotyrosines 228 and/or 247 in RACK1 to Src’s phosphatases. Our findings that Src activity is SH2 domain (Chang et al., 2001), suggested that the necessary for both the tyrosine phosphorylation of two are linked; that signals (like PDGFR or PKC RACK1 and the association of RACK1 and Src that activation) that bring Src and RACK1 into close occur following PKC activation, resembles other proximity with each other result in one of two findings that PTPm is necessary for the recruitment of possibilities: (1) Src binds (via its SH2 domain) to RACK1 to both the plasma membrane and cell – cell phosphotyrosines 228 and/or 247 in RACK1 that have contacts that occurs as cell density increases (Mourton been phosphorylated by a tyrosine kinase other than et al., 2001). Together, the findings indicate that Src, and then Src phosphorylates RACK1 on tyrosines tyrosine kinases and phosphatases play an important other than 228 and/or 247; or (2) Src phosphorylates role in recruiting RACK1 to the plasma membrane and RACK1 on Tyr 228 and/or Tyr 247 and then binds to other intracellular sites, and that tyrosine phosphoryla- those same phosphotyrosines. Our new finding, that tion and dephosphorylation of RACK1 at these sites is Src phosphorylates RACK1 on Tyr 228 and/or 247, an important mechanism of protein – protein interac- indicates that the latter of the two possibilities is true, tion and signal transduction. that Src first phosphorylates RACK1 on Tyr 228 and/ RACK1 is a member of an ancient family of or Tyr 247 and then binds to those same phosphotyr- regulatory proteins made up of highly-conserved osines. Historically, there is precedent for cytosolic repeat units ending in Trp-Asp (WD) (Neer et al., tyrosine kinases preferentially binding (via their SH2 1994; Neer and Smith, 1996). Although WD-repeat domains) to sites on substrates that they phosphorylate proteins are known to be serine/threonine phosphory- (Mayer et al., 1995; Pellicena et al., 1998; Zhou et al., lated (Chen et al., 1995), to the best of our 1995). knowledge, RACK1 is the first family member shown A diverse phage library expressing tyrosine-contain- to be tyrosine phosphorylated. We think that tyrosine ing peptides was used to determine preferred substrate phosphorylation of RACK1 and other family sequences for Src-related tyrosine kinases (Schmitz et members may be important ‘switches’ that link these al., 1996). One advantage of this approach was that all molecules to other signaling molecules and relay 20 amino acids were included in the library, whereas signals across many pathways. Because each WD- previously described chemical libraries were lacking repeat protein contains multiple WD domains and tryptophan, tyrosine, serine, threonine and cysteine. multiple tyrosines, the ‘switches’ may be many, and Using the phage display technique, Src was found to the signals may be diverse and amplified. prefer sites containing tryptophan or glycine in Our in vitro results show that the PDGFR tyrosine position +1 and leucine or isoleucine in position – 1 kinase does not directly phosphorylate RACK1 relative to the tyrosine. In another approach, using a (Figure 1). However, serum or PDGF treatment of similar, complete amino acid, combinatorial library cells enhances both the tyrosine phosphorylation of screening, similar results were found: tryptophan was RACK1 and the binding of RACK1 to Src (Chang et important in position +1 and isoleucine in position – 1 al., 2001). Moreover, Src, PLC-g1 and rasGAP are all to the tyrosine (Lou et al., 1996). Our search of the known to associate via their SH2 domains with both amino acid sequence flanking Tyr 228 and 246 of the PDGFR (following PDGF stimulation) and with RACK1 from eight species (ranging from yeast to RACK1 (Chang et al., 1998, 2001). Together, the human) revealed that, in all eight species, Tyr 246 results indicate that, while RACK1 is not directly contains a tyrptophan at position +1 and Tyr 228 phosphorylated by the PDGFR, it is involved in contains a leucine at position – 1 (Figure 7b). Thus, the PDGF-stimulated mitogenic signaling pathways down- sequence of RACK1 surrounding both Tyr 228 and stream of the receptor. One possible sequence of 246 contain preferred substrate sequences for Src events is that activation of the PDGFR leads to tyrosine kinases. activation of Src which then phosphorylates RACK1. Here, we find that Src can phosphorylate either Once RACK1 is tyrosine phosphorylated, it binds to Y228 or Y246, or possibly both tyrosines on RACK1. the SH2 domains of Src, PLC-g1, rasGAP and other We were not able to determine which site was a better signaling molecules and, in turn, regulates their substrate site for Src because both the Y228F and the activity. RACK1 could inhibit Src kinase activity by Y246F RACK1 mutants were phosphorylated by Src inducing a conformational change in Src when it (Figure 7a). Moreover, both tyrosines are flanked by binds, by blocking Src from phosphorylating down- preferred substrate sequence (Figure 7b, and as

Oncogene RACK1: a novel substrate for the Src kinase BY Chang et al 7627 discussed above). Previously, we found by peptide Antibodies competition assays that Src’s SH2 domain can bind to either phosphotyrosine 228 or 246 on RACK1 (Chang Src monoclonal antibody (MAb) 327 (Lipsich et al., 1983) et al., 2001). However, our mutational analyses was used for immunoprecipitation and immunoblot analyses. revealed that phosphorylation of Y246 is essential for RACK1 MAb (Transduction Laboratories, Lexington, KY, USA) was used for immunoblot analyses. Anti-phosphotyr- RACK1 binding to Src, whereas phosphorylation of osine MAb PY20 (Transduction Laboratories; Glenny et al., Y228 is not (Chang et al., 2001). Thus, Y246 is 1988) were used for immunoprecipitation and immunoblot probably the primary, and Y228 is a secondary Src analyses. substrate and binding site on RACK1. Perhaps either Y246 or Y228 alone can serve as a Src substrate and a Protein extractions, immunoprecipitations and in vitro protein Src SH2 binding site, but normally only Y246 does so kinase assays in cells. In summary, we have shown that RACK1 is a Src For co-immunoprecipitation of Src and RACK1, cells were substrate. Moreover, Src activity is necessary for both lysed in NP-40 buffer (0.5% NP-40, 20 mM Tris (pH 8.0), the tyrosine phosphorylation of RACK1 and the 100 mM NaCl, 1 mM EDTA, 100 M sodium vanadate, 50 mM sodium fluoride, 50 mM leupeptin, 1% aprotinin and 1 mM binding of RACK1 to Src’s SH2 domain that occur dithiothreitol (DTT)) (Chang et al., 1998, 2001). For other following PKC activation. Src phosphorylates RACK1 experiments, cells were lysed in modified RIPA buffer (0.1% on Tyr 228 and/or 246, highly-conserved tyrosines SDS, 1% NP-40, 1% sodium deoxycholate, 150 mM NaCl, located in the sixth WD repeat that interact with Src’s 10 mM sodium phosphate (pH 7.0), 100 mM sodium vanadate, SH2 domain. We believe that RACK1 is an important 50 mM sodium fluoride, 50 mM leupeptin, 1% aprotinin, Src substrate that signals downstream of growth factor 2mM EDTA and 1 mM dithiothreitol (DTT) (Cartwright et receptor tyrosine kinases and is involved in the al., 1985, 1986, 1987, 1990; Park and Cartwright, 1995). regulation of Src function and cell growth. Lysates were centrifuged at 14 000 g for 1 h at 48C. Protein concentrations were measured by the BCA protein assay (Pierce, Rockford, IL, USA) and, unless otherwise stated, samples were standardized to equal amounts of total cellular Materials and methods protein. Lysates were incubated for 3 h at 48C with excess antibody (1 mg of MAb 327, MAb PY20 or IgG) and protein Cell Culture complexes were collected with the addition of 30 ml of protein NIH3T3 cell lines NIH(pcsrc), NIH(pcsrc527/foc/EP)B1, A/G Sepharose beads (Pharmacia, Biotech, Piscataway, NJ, NIH(pcsrc416/pSV2neo/MC)A, and NIH(pc416/527/foc)A, USA). which overexpress wild-type, Y527F, Y416F or Y527F/ Purified tyrosine kinases or Src immunoprecipitates were Y416F chicken c-Src respectively (gifts from David Shallo- tested for ability to phosphorylate purified GST – RACK1 way, Cornell University, Ithaca, NY, USA), and NIH3T3/c- (wild-type or mutant) or GST in in vitro protein kinase assays Src, NIH3T3/Y527F and NIH3T3/neo (Cartwright et al., (Chang et al., 1998, 2001). The purified protein-tyrosine 1987) were cultured in Dulbecco’s modified Eagle medium kinases that we tested included: Src (expressed from (DMEM) (Mediatech, Herndon, VA, USA) supplemented recombinant baculovirus and purified from Sf9 cells) with 10% calf serum (Sigma, St. Louis, MO, USA), and (Upstate Biotechnology, Lake Placid, NY, USA), c-Abl (gift maintained in G418 (200 mg/ml) (Gibco – BRL, Life Tech- from Jean Wang, University of California, San Diego, CA, nologies, Gaithersburg, MD, USA). NIH3T3 cells and USA), Epidermal Growth Factor Receptor (EGFR) (Prome- NIH(pmvsrc/foc/EP)A1 (gift from David Shalloway) were ga, Madison, WI, USA) and Platelet-Derived Growth Factor cultured in DMEM supplemented with 10% calf serum. Receptor (PDGFR) (Upstate Biotechnology). An equivalent CHO cells (American Type Culture Collection, Rockville, specific activity (5 U) of purified Src, Abl, EGFR or PDGFR MD, USA) were cultured in Ham’s F-12 medium (Media- kinase, or Src immunoprecipitates, were incubated with 1 mg tech) supplemented with 10% fetal bovine serum (FBS) of GST – RACK1 or GST for 10 min at 308Cin30mlof (Sigma). kinase buffer containing 50 mM piperazine-N-N’-bis (2- ethanesulphonic acid) (pH 7.0), 10 mM manganese chloride, 10 mM magnesium chloride, 10 mM DTT, 10 mM ATP and Plasmids 25 Ci [g-32P]ATP (4000 Ci/mmol; ICN, Costa Mesa, CA, Plasmids encoding wild-type (pM5H) or mutant dl155-157 USA) (Cartwright et al., 1985, 1986, 1987, 1990; Park and (pM155) chicken c-src were gifts from Sarah Parsons Cartwright, 1995). Proteins were resolved by SDS – PAGE. (University of Virginia, Charlottesville, VA, USA; Moyers Gels were stained with Coomassie brillant blue G-250 to et al., 1993). pGEM K297M src was a gift from Tony Hunter confirm that equivalent amounts of GST – RACK1 or GST (The Salk Institute for Biological Studies, La Jolla, CA, were present in each lane. Radiolabeled proteins were USA; Snyder et al., 1985). Src inserts from these plasmids or detected with Kodak XAR or Fuji FX film and an Y527F c-src (Cartwright et al., 1987) were subcloned into intensifying screen at 7708C. pcDNA3 (Introgen, La Jolla, CA, USA) to create pcDNA3 c-src, dl155 c-src, K297M c-src and Y527F c-src, and Immunoblot analysis transiently expressed in CHO cells (Chang et al., 1998, 2001). The influenza virus hemagglutinin (HA) tag was Src, Py20 or IgG immunoprecipitates were resolved on 10% inserted into the BglII site of pcDNA3 – RACK1 to create SDS-polyacrylamide gels (acrylamide-bisacrylamide, 29 : 0.8). pcDNA3-HA – RACK1 as previously described (Chang et al., Proteins were transferred to polyvinylidene difluoride 1998, 2001). pGEXsrc-SH2 and pGEX – RACK1 were membranes (Immobilon-PTM; Millipore, Bedford, MA, constructed and used to generate GST-fusion proteins as USA) in transfer buffer (25 mM Tris-HCl (pH 7.4), 192 mM described (Chang et al., 1998, 2001). glycine and 15% methanol) using a Trans-Blot apparatus

Oncogene RACK1: a novel substrate for the Src kinase BY Chang et al 7628 (BioRad, Hercules, CA, USA) for 2 h at 60 V (Cartwright et constructed by complete digestion of pcDNA3-RACK1 54 – al., 1985, 1986, 1987, 1990; Park and Cartwright, 1995). 317 with XbaI. C-terminal RACK1 mutants: 1 – 258, 1 – 225, Protein binding sites on the membranes were blocked by 1 – 125, 1 – 97 and 1 – 36 were synthesized by PCR amplifica- incubating membranes overnight in TNT buffer (10 mM Tris- tion using pcDNA3 – RACK1 as the template, HCl (pH 7.5), 100 mM sodium chloride, 0.1% (v/v) Tween 20 CTGGTTCCGCGTGGATCCCCGAAT as the 5’ primer, (Sigma)) containing 3% nonfat, powdered milk (blocking and one of the following 3’ primers: 1 – 258: AGCG- buffer). Membranes were incubated with MAb RACK1 GCCGCTTTCTTGCTTCAGTTCATC; 1 – 225: AGCGG- (0.08 mg/ml), affinity-purified MAb 327 ascites (2 mg/ml) or CCGCTTTTGCCTTCGTTGAGATC; 1 – 125: AGCGGCC- MAb PY20 (0.08 mg/ml) for 1 h, washed in TNT buffer with GCTCAGCTTGCAGTTAGCCAG; 1 – 97: AGCGGCCGC- changes every 5 min for 30 min, and incubated with horse- TTTTGCACACACCCAGGGT; 1 – 36: AGCGGCCGCT- radish peroxidase-conjugated donkey anti-mouse IgM GGTGGTGCCCGTTGTG. (Zymed, San Francisco, CA, USA) for RACK1 blots or Oligonucleotide-directed mutagenesis was used to substi- goat anti-mouse IgG (BioRad) for MAb 327 or PY20 blots tute phenylalanine for tyrosine at residue 52, 140, 194, 228, (Cartwright et al., 1985, 1986, 1987, 1990; Park and 246 and/ or 302 of RACK1, utilizing the Transformer Site- Cartwright, 1995). Proteins were detected by enhanced Directed Mutagenesis Kit according to the manufacturer’s chemiluminescence (ECL) (Amersham, Arlington Heights, protocol (Clontech, Palo Alto, CA, USA), and as previously IL, USA) according to the manufacturer’s protocol. described (Chang et al., 2001). The sequences of all RACK1 mutants were confirmed by automated DNA sequencing (Protein and Nucleic Acid Facility, Stanford University, GST-fusion protein assays Stanford, CA, USA). RACK1 mutants were subcloned into Cultures of Escherichia coli DH5a containing pGEX-Src-SH2 pGEX-3X vector for production of GST fusion proteins or or pGEX RACK1 plasmids were induced with 0.1 mM into pcDNA3 vector for expression in mammalian cell lines. isopropyl-b-D-thiogalactopyranoside (United States Biochemical, Cleveland, OH, USA) for 3 h at 308Cas Transfection assays described (Chang et al., 1998, 2001). Bacteria were harvested, resuspended in Tris-buffered saline (TBS) containing 1% CHO cells were transfected with pcDNA3, pcDNA3-HA – Triton X-100 and 100 mM EDTA and sonicated. After RACK1, or pcDNA3-HA – RACK1 (wild-type or mutant) centrifugation at 12 000 g for 10 min to remove debris, the and pcDNA3src (wild-type or mutant) using lipofectamine supernatant was incubated with glutathione-agarose beads (Gibco – BRL) according to the manufacturer’s protocol and (Sigma) for 2 h at 48C with agitation. Beads were washed as described (Chang et al., 1998, 2001). Briefly, 26105 cells three times with TBS. GST fusion proteins were eluted by the were seeded in six-well plates in Ham’s F-12 medium addition of 100 mM Tris-pH 8.0, 120 mM NaCl and 20 mM containing 10% FBS. Twenty-four hours later, transfections glutathione and dialyzed four times against TBS. were performed using 0.5 to 1 mg of plasmid DNA and 10 ml Purified GST fusion proteins (1 – 5 mg) were incubated with of lipofectamine in serum-free media. Five hours later, cells purified kinases or Src immunoprecipitates and subjected to were placed in fresh media containing 10% FBS. Cells were in vitro protein kinase assays (Cartwright et al., 1985, 1986, lysed 48 h after transfection. Cells were treated with PMA 1987, 1990; Park and Cartwright, 1995). Other GST fusion (10 ng/ml) (Gibco – BRL) for 15 min prior to lysis, as proteins were incubated with cell lysates for 3 h at 48Cas previously described (Chang et al., 2001). described (Chang et al., 1998, 2001). Protein complexes were collected with the addition of 30 ml of glutathione beads, Tyrosine phosphorylation of HA – RACK1 mutants by Src in washed four times in buffer containing 0.5% Nonidet P-40 vivo (NP-40), 20 mM Tris pH-8.0, 100 mM sodium chloride (NaCl) and 1 mM EDTA, and boiled in sodium dodecyl To assess tyrosine phosphorylation of HA – RACK1 mutants sulphate (SDS) sample buffer. Proteins were resolved by by Src in vivo, CHO cells were co-transfected with Src and SDS – PAGE, subjected to immunoblot analysis and detected mutant HA – RACK1 and treated with sodium vanadate by ECL. (100 mM) (Sigma) for 30 min and with PMA (10 ng/ml) for 15 min prior to lysis (Chang et al., 2001). Cells harvested with trypsin for 2 min, collected by centrifugation and lysed Construction and expression of RACK1 mutants in boiling SDS sample buffer. Lysates were divided into three N-terminal truncation mutants of RACK1: 42 – 317 (5.1), equal parts, which were loaded on three separate gels and 94 – 317 (5.2), 138 – 317(5.3), 181 – 317 (5.4) , and 223 – 317 subjected to immunoblot analysis with anti-Src, anti-RACK1 (5.5) were made by PCR amplification using pcDNA3 and anti-phosophotyrosine. Data on levels of Src, mutant RACK1 as the template, a common 3’ primer HA – RACK1 and tyrosine-phosphorylated mutant HA – (CCCGCGGCCGCGGGACCCAAGCTTGGTACC) and RACK1 were quantified by scanning densitometry of bands. one of the following 5’ primers: 5.1: GGATCCGGATC- Because the amounts of Src and mutant HA – RACK1 CATGTGGAAACTGACCAGG; 5.2: GGATCCGGATCC- protein expressed in cells varied with each transfection, the ATGACGGGCACCACCACG; 5.3: GGATCCGGATCCA- phosphotyrosine signals were normalized to the amount of TGGTGTGCAAATACACT; 5.4: GGATCCGGATCCAT- HA – RACK1 and Src expressed. Results were then compared GAACTGCAAGCTGAAG; 5.5: GGATTCCGGATCCAT- to those for wild-type HA – RACK1. GAACGAAGGCAAACAC. The five PCR fragments were subcloned into pcDNA3 to create pcDNA3 – RACK1-5.1, 5.2, 5.3, 5.4 and 5.5. RACK1 Acknowledgments mutant 54 – 317 was constructed by digesting full length We thank David Shalloway for NIH3T3 cell lines pcDNA3 – RACK1 with EcoRI, thus cleaving the N-terminal expressing v-Src or c-Src mutants Y527F, Y527F/Y416F 1 – 54 fragment, and reannealing the cDNA. RACK1 mutant or Y416F. We are grateful to Jean Wang for purified Abl 54 – 292 was made by partial digestion of pcDNA3-RACK1 kinase, Sarah Parsons for pMdl155c-src and Tony Hunter 54 – 317 with BamHI. RACK1 mutant 54 – 183 was for pGEM K297M. We thank Daria Mochly-Rosen, Anson

Oncogene RACK1: a novel substrate for the Src kinase BY Chang et al 7629 Lowe and Bishr Omary for critical review of the data. We supported by grants from the National Institutes of Health are grateful to Blanca Pineda for assistance with prepara- to CA Cartwright (R01 DK43743) and to BY Chang tion of the manuscript and ﬁgures. This work was (National Research Service Award CA69810).

References

Abram CL and Courtneidge SA. (2000). Exp. Cell. Res., 254, Mochly-Rosen D. (1995). Science, 268, 247 – 251. 1 – 13. Mourton T, Hellberg CB, Burden-Gulley SM, Hinman J, Bjorge JD, Jakymiw A and Fujita DJ. (2000). Oncogene, 19, Rhee A and Brady-Kalnay SM. (2001). J. Biol. Chem., 276, 5620 – 5635. 14896 – 14901. Cartwright CA, Eckhart W, Simon S and Kaplan PL. (1987). Moyers JS, Bouton AB and Parsons SJ. (1993). Mol. Cell. Cell, 49, 83 – 91. Biol., 13, 2391 – 2400. Cartwright CA, Hutchinson M and Eckhart W. (1985). Mol. Neer EJ and Smith TF. (1996). Cell, 84, 175 – 178. Cell. Biol., 5, 2647 – 2652. Neer EJ, Schmidt CJ, Nambudripad R and Smith TF. (1994). Cartwright CA, Kaplan PL, Cooper JA, Hunter T and Nature, 371, 297 – 300. Eckhart W. (1986). Mol. Cell. Biol., 6, 1562 – 1570. Park J and Cartwright CA. (1995). Mol. Cell. Biol., 15, Cartwright CA, Meisler MA and Eckhart W. (1990). Proc. 2374 – 2382. Natl. Acad. Sci. USA, 87, 558 – 562. Pellicena P, Stowell KR and Miller WT. (1998). J. Biol. Chang BY, Chiang M and Cartwright CA. (2001). J. Biol. Chem., 273, 15325 – 15328. Chem., 276, 20346 – 20356. Piwnica-WormsH,SaundersKB,RobertsTM,SmithAE Chang BY, Conroy KB, Machleder EM and Cartwright CA. and Cheng SH. (1987). Cell, 49, 75 – 82. (1998). Mol. Cell. Biol., 18, 3245 – 3256. Reynolds AB, Vila J, Lansing TJ, Potts WM, Weber MJ and Chen RH, Miettinen PJ, Maruoka EM, Choy L and Derynck Parsons JT. (1987). EMBO J., 6, 2359 – 2364. R. (1995). Nature, 377, 548 – 552. Schechtman D and Mochly-Rosen D. (2001). Oncogene, 20, Glenny JR, Zokas L and Kamps MP. (1988). J. Imm. Meth., 6339 – 6347. 109, 277 – 285. Schmitz R, Baumann G and Gram H. (1996). J. Mol. Biol., Kmiecik TE and Shalloway D. (1987). Cell, 49, 65 – 73. 260, 664 – 677. Lipsich LA, Lewis AJ and Brugge JS. (1983). J. Virol., 48, Snyder MA, Bishop JM, McGrath JP and Levinson AD. 352 – 360. (1985). Mol. Cell. Biol., 5, 1772 – 1779. Lou Q, Leftwich ME and Lam KS. (1996). Bioorg. Med. Thomas SM and Brugge JA. (1997). Annu. Rev. Cell Dev. Chem., 4, 677 – 682. Biol., 13, 513 – 609. Martin GS. (2001). Nature Rev. Mol. Cell Biol., 2, 467 – 475. Zhou S, Caraway III KL, Eck MJ, Harrison SC, Feldman Mayer BJ, Hirai H and Sakai R. (1995). Curr. Biol., 5, 296 – RA, Mohammadl M, Schlessinger J, Hubbard SR, Smith 305. DP, Eng C, Lorenzo ML, Ponder BAJ, Mayer BJ and Mochly-Rosen D and Kauvar LM. (1998). Adv. Pharm., 44, Cantley LC. (1995). Nature, 373, 536 – 544. 91 – 145.

Oncogene