14-3-3β Is a p90 Ribosomal S6 Kinase (RSK) Isoform 1-binding Protein That Negatively Regulates RSK Kinase Activity*

p90 ribosomal S6 kinase 1 (RSK1) is a serine/threonine kinase that is activated by extracellular signal-related kinases 1/2 and phosphoinositide-dependent protein kinase 1 upon mitogen stimulation. Under basal conditions, RSK1 is located in the cytosol and upon stimulation, RSK1 translocates to the plasma membrane where it is fully activated. The ability of RSK1 to bind the adapter protein 14-3-3β was investigated because RSK1 contains several putative 14-3-3-binding motifs. We demonstrate that RSK1 specifically and directly binds 14-3-3β. This interaction was dependent on phosphorylation of serine 154 within the motif RLSKEV of RSK1. Binding of RSK1 to 14-3-3β was maximal under basal conditions and decreased significantly upon mitogen stimulation. After 5 min of serum stimulation, a portion of 14-3-3β and RSK1 translocated to the membrane fraction, and immunofluorescence studies demonstrated colocalization of RSK1 and 14-3-3β at the plasma membrane in vivo. Incubation of recombinant RSK1 with 14-3-3β decreased RSK1 kinase activity by ∼50%. Mutation of RSK1 serine 154 increased both basal and serum-stimulated RSK activity. In addition, the epidermal growth factor response of RSK1S154A was enhanced compared with wild type RSK. The amount of RSK1S154A was significantly increased in the membrane fraction under basal conditions. Increased phosphorylation of two sites essential for RSK1 kinase activity (Ser380 and Ser363) in RSK1S154A compared with RSK1 wild type, demonstrated that 14-3-3 interferes with RSK1 phosphorylation. These data suggest that 14-3-3β binding negatively regulates RSK1 activity to maintain signal specificity and that association/dissociation of the 14-3-3β-RSK1 complex is likely to be important for mitogen-mediated RSK1 activation.

The p90 ribosomal S6 kinase (RSK) 1 family members have a role in mitogen-activated cell growth and proliferation, differ-entiation, and cell survival. RSK is a serine/threonine kinase that is a substrate of ERK1/2 and lies downstream of the Raf/MEK/ERK protein cascade (1). Recent studies have provided insight into mechanisms responsible for activation of RSK1. RSK1 has two distinct domains, both of which are catalytically functional. Phosphorylation of RSK1 by ERK on the carboxyl-terminal catalytic loop activates the RSK1 carboxylterminal kinase domain. This induces a conformational change abrogating the negative effect of this domain, allowing activation of the amino-terminal kinase domain by PDK1 (2). Cellular localization is thought to be important in the regulation of RSK activity. Upon mitogen stimulation RSK undergoes a rapid and transient localization to the plasma membrane, which requires ERK docking and phosphorylation. At the plasma membrane PDK1 inputs as well as ERK-and PDK1independent events lead to full activation of RSK (3). RSK can then phosphorylate its substrates, which include the Na ϩ /H ϩ exchanger NHE1 isoform at the plasma membrane, nuclear transcription factors, and transcriptional coactivator proteins (1, 4).
14-3-3 proteins are a family of 30-kDa acidic proteins that interact with a wide variety of cellular proteins including protein kinases, receptor proteins, enzymes, structural and cytoskeletal proteins, and small G-proteins (5)(6)(7)(8). The predominant 14-3-3-binding motif involves a phosphoserine, but interaction can involve other varied motifs and may not require phosphorylation (9 -11). Because 14-3-3 proteins exist as dimers, they can simultaneously bind multiple phosphorylation sites on the same protein (12,13) or function as scaffolds to form protein-protein interactions (5)(6)(7)(8)14). Other roles of 14-3-3 proteins include regulation of interacting proteins activity, subcellular localization, and/or stability. For example, 14-3-3 proteins are known to act as a scaffold for several proteins in the mitogen-activated protein kinase cascades, including MEKK1, 2, and 3 (15), and to regulate the kinase activity of Raf (16). Work in our laboratory has shown that NHE1 is activated by mitogens through phosphorylation of serine 703 by RSK (4). Subsequently, 14-3-3␤ binds to this site and prevents dephosphorylation of serine 703 (17). In the present study we investigated whether RSK1 was also a binding partner for 14-3-3␤ because it contains a number of putative 14-3-3 interaction motifs. We found that 14-3-3␤ specifically interacts with RSK1 in a phosphorylation-dependent manner, and the site of interaction was identified as serine 154. RSK-14-3-3␤ binding is maximal under quiescent conditions and upon mitogen stimulation; both RSK and 14-3-3␤ translocate to the plasma membrane where they presumably dissociate because binding to 14-3-3␤ is decreased. Furthermore, 14-3-3␤ binding is inhibitory to RSK activity, and RSK1S154A, which does not interact with 14-3-3, has increased activity both basally and upon mitogen stimulation. This study suggests an important role for 14-3-3␤ in the negative regulation of RSK kinase activity.
In Vitro Translation of Full-length RSK1-Full-length rat p90RSK1 under control of the T7 promoter in pcDNA3.1/Myc-His was transcribed and translated in vitro using the TNT T7-coupled reticulocyte lysate system (Promega). The nascent protein was labeled using Transcend biotin-lysyl-tRNA (Promega). Briefly, 40 l of TNT Quick Master Mix, 1 l of methionine (1 mM), 1 g of template DNA, and 1 l of biotin-lysyl-tRNA were mixed in a final volume of 50 l, incubated for 90 min at 30°C, and immediately used for the binding assays.
Preparation of Cell Lysates-The cell monolayers were rinsed with ice-cold phosphate-buffered saline (150 mM NaCl, 20 mM Na 2 PO 4 , pH 7.4) and then scraped in 1 ml of phosphate-buffered saline. After a brief centrifugation, the cells were solubilized in 1 ml of cell lysis buffer (10 mM HEPES, pH 7.4, 50 mM sodium pyrophosphate, 50 mM NaF, 50 mM NaCl, 5 mM EDTA, 5 mM EGTA, 1 mM Na 3 VO 4 , 0.5% Triton X-100 plus 1:1000 protease inhibitor mixture (Sigma)). The cells were sonicated for 20 s (in the case of pull-down studies) or needle-homogenized (in the case of coimmunoprecipitations), agitated on a rotating rocker at 4°C for 30 min, and centrifuged at 12,000 ϫ g for 30 min to remove insoluble cellular debris. For some pull-down experiments, phosphatase treatment of cell lysates was performed by lysing cells in CIAP buffer (20 mM Tris, pH 8, 150 mM NaCl, 1 mM MgCl 2 , 1 mM dithiothreitol, 0.5% Triton X-100, plus 1:1000 protease inhibitor mixture). The lysates were treated with or without 200 units of calf intestinal alkaline phosphatase (Promega) at 37°C for 1 h. Where indicated the inhibitors Na 3 VO 4 (1 mM) and NaF (50 mM) were added.
Immunoprecipitations and Pull-downs-For coimmunoprecipitation studies, HEK293 cells were cotransfected with FLAG-tagged RSK1 and Xpress-tagged 14-3-3␤. Immunoprecipitations were carried out by preclearing cell lysates with protein A/G-agarose (Santa Cruz) for 1 h followed by incubation with anti-Xpress antibody for 3 h and with protein A/G-agarose for a further 1 h or M2 anti-FLAG-agarose beads for 4 h. Pull-downs were performed by incubating 10 g of GST-14-3-3␤ or GST-14-3-3␤ K49Q (produced as described previously (17)) bound to glutathione-Sepharose overnight at 4°C.
Immunoprecipitates and pull-downs were then washed four times with 1 ml of cell lysis buffer before the addition of Laemmli sample buffer. After heating at 95°C for 3 min, the proteins were resolved on SDS-PAGE and transferred to nitrocellulose membranes for Western analysis. Immunoblotting was performed with anti-Xpress antibody (Invitrogen), M2 anti-FLAG antibody (Sigma), anti-RSK1 antibody (Santa Cruz), anti-14-3-3␤ antibody (Santa Cruz), or anti-NHE1 antibody (Chemicon), anti-phospho-RSK1 (Ser 363 ) (Upstate Biotechnology), and anti-phospho-RSK (Ser 380 ) (which was a gift from Dr. C. Chrestensen, University of Virginia). Immunoreactive bands were detected with horseradish peroxidase-conjugated secondary antibodies (Amersham Biosciences) and enhanced chemiluminescence.
Preparation of Membrane Fractions-The cell monolayers were rinsed and scraped in 1 ml of ice-cold phosphate-buffered saline. After a brief centrifugation the cells were resuspended in 750 l of membrane fractionation buffer (25 mM Tris-HCl, pH 7.4, 5 mM EGTA, 5 mM EDTA, 100 mM NaF, 5 mM dithiothreitol, and 1:1000 protease inhibitor mixture). The lysates were prepared by needle homogenization and nuclei, and unbroken cells were pelleted at 850 ϫ g for 10 min. Equal amounts of protein in postnuclear supernatants were centrifuged at 100,000 ϫ g for 1 h to obtain soluble fractions and total membrane pellets. The soluble fraction was removed, and the membrane pellets were solubilized in 250 l of buffer containing 1% Triton X-100, sonicated, and rocked for 1 h at 4°C and analyzed by SDS-PAGE as described above.
RSK1 Kinase Assays-Twenty milliunits of recombinant RSK1 (Upstate Biotechnology Inc.) was incubated in 30 l of kinase buffer (30 mM HEPES, pH 7.5, 100 mM NaCl, 10 mM MgCl 2 , 10 mM MnCl 2 , and 0.2 mg/ml bovine serum albumin) overnight with 10 g of GST-14-3-3␤ or GST-14-3-3␤K49A. To initiate the kinase reaction, 10 g of S6 peptide, 7.5 M ATP, and 5 Ci of [␥-32 P]ATP (Amersham Biosciences) were added, and the reaction was incubated for 10 min at 30°C. The reaction was terminated by spotting 25 l of the reaction onto P81 phosphocellulose filter paper. The filters were washed five times in 0.75% phosphoric acid and one time in acetone, and radioactive incorporation was determined by Cerenkov counting.
CREB Reporter Gene Assays-HEK293 cells were plated at 2 ϫ 10 5 cells/well in 24-well plates 24 h prior to transfection. The cells were transfected with different expression plasmids together with the activator plasmid for CREB (pFA2-CREB) and the reporter gene for luciferase (pFR-Luc) (Stratagene). A plasmid expressing the enzyme Renilla luciferase was used as an internal control (pRL-TK; Promega). After 24 h, the cells were serum-starved for 24 h and then stimulated as indicated. Firefly and Renilla luciferase activities were measured using a Dual Luciferase Reporter System (Promega) with a Wallac 1420 luminometer. The data are expressed as firefly luciferase normalized by Renilla luciferase activity.
Immunocytochemistry and Confocal Microscopy-Cos7 cells seeded on glass coverslips were transfected with pCMV-FLAG-RSK1 using LipofectAMINE 2000. After 24 h the cells were serum-starved for a further 12 h and treated with 20% FBS. The cells were fixed in 3% paraformaldehyde, permeabilized with 0.2% Triton X-100, and blocked in phosphate-buffered saline with 5% FBS, 0.2% bovine serum albumin. The cells were stained with anti-FLAG antibody to visualize RSK1 and anti-14-3-3␤ antibody and then incubated in secondary antibodies (Vec-
The Interaction between RSK and 14-3-3␤ Requires Phosphorylation-Many interactions with 14-3-3 are dependent on binding partner phosphorylation. To determine whether this is also the case for the RSK1-14-3-3␤ interaction, PS127 cell lysates were incubated with alkaline phosphatase for 1 h at 37°C before GST-14-3-3␤ pull-down. As shown in Fig showing potential 14-3-3␤-binding motifs underlined. B, HEK293 cell lysates transfected with RSK1 containing serine to alanine mutants of potential 14-3-3␤binding motifs were pulled down with GST-14-3-3␤ beads. Proteins bound to beads were subjected to SDS-PAGE and immunoblotted with anti-RSK1 antibody. C, HEK293 cells were transfected with Xpress-14-3-3␤, and FLAG-RSK1 or FLAG-RSK1S154A. Cells transfected with empty vector were used as a control. The cell lysates were immunoprecipitated (IP) with anti-Xpress antibody and immunoblotted for FLAG-RSK1 (upper panel) or Xpress-14-3-3␤ (middle panel). The lower panel shows equal expression of the RSK constructs.
14-3-3␤ in the absence but not the presence of phosphatase inhibitors. Hence dephosphorylation of RSK1 abolished its ability to bind 14-3-3␤. In support of this conclusion, treatment of intact cells with the broad spectrum serine/threonine protein kinase inhibitor staurosporine also decreased the binding of RSK1 to GST-14-3-3␤ in vitro (Fig. 2B).
Serum Stimulation Reduces the RSK1-14-3-3␤ Interaction-Because serum stimulates RSK1 activity, which then activates NHE1 through phosphorylation of serine 703 (increasing the NHE1-14-3-3 interaction) (4, 17), we investigated the effect of serum on the RSK1-14-3-3␤ interaction. PS127 fibroblasts were stimulated with 20% serum, and RSK association with GST-14-3-3␤ was determined in a pull-down assay. Binding of RSK1 to 14-3-3␤ was decreased by 34.6 Ϯ 7.5% in cells stimulated for 5 min, 36.2 Ϯ 4.1% at 10 min, and 28.2 Ϯ 6.5% at 20 min (p Ͻ 0.05; Fig. 4, A and B). This effect was also observed in HEK293 cells and Cos7 cells overexpressing exogenous RSK1 (data not shown). Preincubating PS127 cells for 30 min with the MEK1 inhibitor PD98059 (30 M), which prevents agoniststimulated ERK1/2 and RSK activation, reversed this inhibition (Fig. 4, A and B). As a control, NHE1 interaction with GST-14-3-3␤ was verified. Binding of GST-14-3-3␤ to NHE1 was stimulated by serum, and this stimulation was significantly inhibited by PD98059 treatment (Fig. 4C, especially at 10 and 20 min). Subcellular Colocalization of RSK1 and 14-3-3-14-3-3 proteins function as scaffolding proteins and have been shown to affect the subcellular localization of interacting proteins (reviewed in Refs. 7 and 14). For example, 14-3-3 can sequester proteins in the cytosol and/or target them to the plasma membrane. RSK1 has previously been shown to be located in the cytosol under basal conditions and to translocate to the plasma membrane after stimulation where it becomes fully activated (3). Therefore, we determined the subcellular distribution of 14-3-3␤ and RSK1 before and after serum stimulation using subcellular fractionation in PS127 fibroblasts. In the absence of serum, RSK1 and 14-3-3␤ were predominantly localized in the cytosol. After stimulation with 20% fetal calf serum for 5 min, there was a significant translocation of both 14-3-3␤ and RSK1 to the membrane fraction (Fig. 5A) In contrast, the distribution of actin was unchanged after fetal calf serum treatment (Fig.  5A). This translocation also occurred with endogenous RSK and 14-3-3␤ in NIH3T3 cells and in HEK293 cells expressing exogenous RSK (data not shown).
To demonstrate colocalization of RSK1 and 14-3-3␤ in intact cells, Cos7 cells were transfected with FLAG-RSK1, and RSK1 and 14-3-3␤ were visualized by immunofluoresence staining with anti-FLAG antibody and anti-14-3-3␤, antibody respec- FIG. 4. Serum stimulation inhibits the interaction between RSK1 and 14-3-3␤. PS127 cells were stimulated with 20% FBS for 0, 5, 10, and 20 min with and without pretreatment for 1 h with 30 M PD98059, and the cell lysates were pulled down with GST-14-3-3␤ beads. A, proteins bound to beads were subjected to SDS-PAGE and immunoblotted with anti-RSK1 antibody. Relative binding of RSK1 to 14-3-3␤ is shown as mean Ϯ S.E. of five experiments (bottom panel). B, densitometric analysis of relative binding of RSK1 to 14-3-3 interaction after serum stimulation. The values are normalized to the 0 time point, which was set at 100%. The results are expressed as the means Ϯ S.E. *, p Ͻ 0.05 compared with control (n ϭ 5). C, proteins bound to GST-14-3-3␤ beads were subjected to SDS-PAGE and immunoblotted with anti-NHE1 antibody.
FIG. 5. After serum stimulation RSK1 and 14-3-3␤ translocate to the membrane fraction. Postnuclear supernatants of PS127 cells were separated into cytosol and membrane fractions. Total, cytosol, and membrane fractions were subjected to SDS-PAGE and immunoblotted as shown. Note that four times more of the membrane fraction (v/v) was loaded as compared with cytosol and total fraction. tively (Fig. 6). In unstimulated Cos7 cells, RSK1 and 14-3-3␤ appeared diffusely distributed in the cytoplasm with significant colocalization (Fig. 6A). In cells stimulated with 20% fetal calf serum for 5 min, both 14-3-3␤ and RSK1 translocated to the periphery of the cells (Fig. 6B). Overlaying the two images demonstrated significant colocalization, which was greatest at the plasma membrane (Fig. 6B).
14-3-3␤ Inhibits the Kinase Activity of RSK1-It is possible that RSK1 binding to 14-3-3 could either increase or decrease kinase activity based on previous reports (reviewed in Ref. 14). Because overexpressing a dominant negative 14-3-3 construct in cells would inhibit the Raf-MEK1-ERK pathway (20) and therefore decrease RSK kinase activity, it was not satisfactory to perform this experiment. Therefore, we determined the effect of incubating active recombinant RSK1 with 14-3-3␤ in vitro. Recombinant RSK1 was preincubated with GST alone, GST-14-3-3␤ K49Q, or GST-14-3-3␤ for 4 h, and then a kinase assay was performed using S6 peptide as the substrate. The activity of RSK1 in the presence of GST-14-3-3␤ was significantly reduced by ϳ50% as compared with GST alone or GST-14-3-3␤ K49Q (Fig. 7A). This finding suggests that binding of RSK1 to 14-3-3 may suppress RSK1 activity in vivo.
To assess the functional significance of the RSK-14-3-3 binding in vivo, we compared the kinase activity of wild type RSK1 to the non-14-3-3␤-binding RSK1S154A mutant using a luciferase reporter assay system to measure the activity of CREB. CREB is activated by cAMP-dependent protein kinase and by all three members of the RSK family (RSK1-3) in cells stimulated by activators of the Ras/MEK/ERK1/2 cascade (21)(22)(23). As shown in Fig. 7B, CREB activity in cells expressing RSK1 wild type (RSK1WT) was increased after serum stimulation. However, in cells overexpressing RSKS154A, there was an 80% increase in CREB basal activity and a 114% increase in serumstimulated activity relative to RSK1WT. This suggests that 14-3-3 binding to RSKS154 inhibits RSK activity both basally and after activation with serum. We next investigated the effect of the RSK1S154A mutation on EGF-stimulated CREB activity. The cells were transfected with RSK1WT or RSK1S154A and stimulated for 6 h with different concentrations of EGF, after which CREB activity was measured. In cells transfected with RSK1S154A, the EGF dose-response curve was significantly shifted to the left compared with cells expressing RSK1WT (EC 50 of 1 ng/ml versus 6 ng/ml; Fig. 7C), demonstrating that 14-3-3␤ binding inhibits EGF-stimulated RSK1 activity and thus CREB activation.
The increased CREB activity in cells transfected with RSK1S154A versus those transfected with RSK1WT combined with the finding that 14-3-3␤ and RSK1 interact maximally under basal conditions suggest that 14-3-3␤ inhibits RSK1 by preventing its activation rather than by interfering with RSK1substrate interaction. To verify this hypothesis we used phospho-specific antibodies targeting two phosphorylation sites on RSK1, Ser 380 and Ser 363 . Ser 380 is autophosphorylated basally, and it is further phosphorylated by the carboxyl-terminal RSK1 kinase domain upon docking and activation by ERK (1,2,24,25). Ser 363 phosphorylation is thought to occur subsequently, after translocation of the RSK-ERK complex to the plasma membrane (3). We found that in RSK1S154A constructs, phosphorylation of both Ser 380 and Ser 363 was increased basally (by 27.3 Ϯ 2.6% and 137.9 Ϯ 10.6% respectively, n ϭ 3) and after 5 min of serum stimulation (by 35.9 Ϯ 3.9 and 28.1 Ϯ 2.8%, respectively, n ϭ 3; Fig. 8, A and B), compared with RSK1WT. These data demonstrate that 14-3-3 inhibits the activation of RSK1.
Because full activation of RSK requires plasma membrane translocation (3), the increased activity of RSK1S154A suggested an increase in the amount of this mutant at the plasma membrane compared with RSK1WT. Therefore, we performed subcellular fractionation in HEK293 cells overexpressing RSK1WT or RSK1S154A. The amount of RSK1S154A in the membrane fraction under basal conditions was increased by 28.3 Ϯ 2.9% (p Ͻ 0.05) as compared with RSK1WT, whereas upon serum stimulation the relative amounts of RSK1WT and mutant in the membrane fraction were similar (Fig. 9, A and B).
14-3-3 proteins have previously been shown to bind mainly through the motifs RSXpSXP and RXXXpSXP, where pS denotes phosphoserine and X is any amino acid; however, there are strong preferences for particular amino acids over others in the X positions (12,13). To locate binding sites in RSK1 we made eight serine to alanine mutations at close matches to these consensus motifs (RLS 154 KEV, RLGS 307 GP, RDS 363 PGI, RGFS 380 FV, RDPS 457 EE, REAS 513 FV, RIS 630 S 631 GK, and RKLPS 732 TT). Only mutation of serine 154 greatly diminished binding of RSK1 to 14-3-3␤ both in vitro and in vivo, suggesting that phosphorylated serine 154 is the site of 14-3-3␤ interaction with RSK1. Interestingly, this motif lies within a predicted coiled-coil domain of RSK (amino acids 151-173; analyzed using Lupas's method at the EMBnet-CH website). Coiled-coil domains are known to mediate protein-protein interactions (27) and have previously been shown to be involved in the interaction of ␥-aminobutyric acid, type B receptors with 14-3-3 proteins (28).
The activation of RSK is dependent on a combination of both phosphorylation and plasma membrane localization events. ERK phosphorylates RSK1 within the carboxyl-terminal domain, and the phosphoinositide dependent kinase PDK1 phosphorylates RSK1 within the NH 2 -terminal domain (1, 2). Richards et al. (3) demonstrated recently that stimulated RSK1 transiently associates with the plasma membrane before accumulating in the nucleus. They also showed that the activity of a kinase inactive RSK1 mutant lacking the ERK docking site could be restored by addition of a membrane-targeting myristoylation site to the level of myristoylated wild type RSK1. This suggests that ERK has a role in escorting RSK to a plasma membrane-associated complex, although the exact mechanism of localization of RSK remains unclear (3).
14-3-3 proteins have been shown to play an important role in the subcellular distribution of a variety of signaling proteins. They appear to have the ability to both sequester proteins in the cytosol (for example protein kinase C; (29), dictoystelium myosin II heavy chain protein kinase C (30), and Raf1 (16,31)) and to cause translocation to the plasma membrane (for example testicular protein kinase 1) (32). Data in the present study suggest that 14-3-3␤ may play a role in sequestering RSK1 in the cytosol under basal conditions. RSK1-14-3-3␤ binding is maximal when cells are quiescent, and significant colocalization can be seen using immunofluoresence.
Upon activation by serum, a portion of both RSK1 and 14-3-3␤ translocate to the membrane fraction, and they partially colocalize in vivo at the plasma membrane. Serum also reduces 14-3-3␤ binding to RSK1 by 30 -40%. These results suggest that 14-3-3␤ translocates with RSK1 to the plasma membrane, and once there the two proteins dissociate. This dissociation may be regulated in two ways. First, phosphatases are likely to play a role in dephosphorylating the RSK1-binding site (serine 154), and second, other 14-3-3␤ binding partners located at the plasma membrane may compete for 14-3-3␤ interaction. It is interesting to note that the Na ϩ /H ϩ exchanger NHE1 has previously been shown to be a RSK substrate and that upon mitogen stimulation, 14-3-3␤ binds at the site of RSK phosphorylation (4,17). Therefore, once phosphorylated by RSK, NHE1 may be a competing binding protein for 14-3-3␤ at the membrane.
In this study, incubation of RSK1 with GST-14-3-3␤ inhib-FIG. 7. 14-3-3␤ inhibits RSK1 kinase activity through interaction at S154. A, recombinant RSK1 was incubated overnight with 10 g of GST alone, GST-14-3-3␤ K49Q, or GST-14-3-3␤ at 4°C. Kinase activity was determined using S6 peptide as the substrate. The data are expressed as relative radioactivity (cpm) incorporated into S6 peptide. The results are the means Ϯ S.E. *, p Ͻ 0.05. B, HEK293 cells were transfected with the activator plasmid for CREB (pFA2-CREB), the reporter gene for luciferase (pFR-Luc), and Renilla plasmid (pRL-TK). The cells were also transfected with either RSK1 wild type or RSK1 S154A mutant. After 24 h of serum starvation, the cells were stimulated with 20% serum for 6 h. CREB activity was measured by assaying for firefly luciferase values. Renilla luciferase activity was measured as an internal control, and the results are presented as relative luciferase activity. The results are the means Ϯ S.E. *, p Ͻ 0.05 (n ϭ 4). C, CREB activity was assayed as described in B at increasing doses of EGF. The results are the means Ϯ S.E. (n ϭ 3-6).
ited RSK activity by 50%. It has been demonstrated that multiple other signaling proteins bind to 14-3-3 and that their activities are inhibited by 14-3-3. For example, 14-3-3 inhibits the activity of phosphatidylinositol 3-kinase (33), protein kinase C isoforms (29) and (34), dictoystelium myosin II heavy chain protein kinase C (30), and testicular protein kinase 1 (32). We have identified Ser 154 as a major 14-3-3-binding site and found that RSK1-14-3-3 interaction is reduced in S154A mutants. Using the activity of CREB, a transcription factor downstream of Raf (1, 22) as a functional readout for RSK activity, we further demonstrated that the S154A mutation increased basal and serum-stimulated RSK activity. In addition when a dose-response curve for EGF-dependent CREB stimulation was performed, the EC 50 for RSK1S154A was significantly lower than that of wild type RSK1. Increased phosphorylation of two sites essential for RSK1 kinase activity (Ser 380 and Ser 363 ) in the RSK1S154A mutant as compared with RSK1WT both basally and after mitogen stimulation suggests a role for 14-3-3␤ in the inhibition of RSK1 kinase activity by interfering with its phosphorylation.
These results suggest that 14-3-3 plays a role in maintaining RSK1 signaling fidelity. Under basal conditions it may prevent activation by weak ERK signals. It is possible that RSK1 is held in the cytoplasm in an inactive conformation by 14-3-3␤, thus blocking phosphorylation of RSK1 activation sites. Because activation of RSK requires plasma membrane localization, increased basal activity of the S154A mutant correlates well with increased amounts in the plasma membrane compared with RSK1WT. Therefore, 14-3-3 is not a major determinant of membrane translocation of RSK1, although immunofluorescence and membrane fractionation data suggest that the two proteins move in a complex to the plasma membrane before FIG. 8. 14-3-3␤ inhibits RSK1 kinase activity through inhibition of RSK1 phosphorylation. HEK293 cells were transfected with FLAG-RSK1 or FLAG-RSK1S154A. Cells transfected with empty vector were used as a control. The cells were treated with or without serum for 5 min and RSK1 constructs were immunoprecipitated (IP) with anti-FLAG antibody. A, immunoprecipitates were immunoblotted with p380 RSK antibody. The lower panel shows equal expression of the RSK constructs. B, immunoprecipitates were immunoblotted with p363 RSK antibody. The lower panel shows equal expression of the RSK constructs. The blots are representative of three similar experiments.
FIG. 9. The amount of RSK1S154A in the membrane fraction is increased under basal conditions as compared with RSK1WT. A, postnuclear supernatants of HEK293 cells transfected with RSK1WT or RSK1S154A were separated into cytosol and membrane fractions. The total and membrane fractions were subjected to SDS-PAGE and immunoblotted as shown. B, densitometric analysis of relative amounts of RSK1 and RSK1S154A in the membrane and total fractions basally (Control) and after serum stimulation. The values are relative to RSK1WT at each condition, which was set at 100%. The results are expressed as the means Ϯ S.E. *, p Ͻ 0.05 (n ϭ 5).
dissociating. Rather, 14-3-3 is a modulator of RSK1 kinase activity. At the plasma membrane, 14-3-3 may continue to alter the kinetics of RSK1 activity until complete dissociation of the proteins. Binding to14-3-3␤ appears to mask phosphorylation sites in RSK1 crucial for its activation, and it may also block the access of substrates or ATP to the catalytic site of RSK1 or continue to maintain RSK1 in a partially inactive conformation. In this way, 14-3-3␤ could play an important role in the negative feedback regulation of active RSK1. Our data suggest some similarities between the regulation of RSK1 and Raf by 14-3-3. There are two serine sites on Raf that bind to 14-3-3, serine 259 and serine 621 (12,16,35). The activity of the Raf S259A mutant is increased basally and after EGF stimulation to a greater extent than wild type Raf (16,(35)(36)(37), and the mutant has increased amounts in the plasma membrane, suggesting 14-3-3 binding to Ser 259 antagonizes Raf activity by sequestering it in the cytosol (38,39).
In summary, we demonstrate for the first time that RSK1 is a 14-3-3-binding protein and suggest a functional role of the 14-3-3␤/RSK1 interaction in vivo. Based on the model shown in Fig. 10, we propose that RSK1 and 14-3-3␤ are bound in the cytosol, maintaining RSK1 in an inactive conformation. Upon stimulation with mitogens, both 14-3-3␤ and RSK1 translocate to the plasma membrane. RSK1-14-3-3␤ complexes dissociate at the plasma membrane, allowing full activation of RSK1. Finally, RSK1 phosphorylates substrates such as Ser 703 of NHE1. This study demonstrates added complexity of the cellular regulation of RSK. Further understanding of the mechanisms of 14-3-3 regulation of RSK signaling will give insights into the role of RSK in cell growth and proliferation. Because previous studies by our laboratory have demonstrated increased RSK activity in cells and tissues of hypertensive ani-mals, it will be of interest to study the potential role of 14-3-3 in this altered RSK regulation. In unstimulated cells, RSK1 and 14-3-3␤ bind in cytosol, perhaps maintaining RSK1 in an inactive conformation. On stimulation with mitogens, both 14-3-3␤ and RSK1 translocate to the plasma membrane. RSK 14-3-3␤ complexes then dissociate, allowing full activation of RSK1. RSK1 then phosphorylates plasma membrane substrates such as Ser 703 of NHE1 and translocates to the nucleus to phosphorylate nuclear substrates.