Serum- and glucocorticoid-inducible kinase SGK phosphorylates and negatively regulates B-Raf.

Phosphorylation can both positively and negatively regulate activity of the Raf kinases. Akt has been shown to phosphorylate and inhibit C-Raf activity. We have recently reported that Akt negatively regulates B-Raf kinase activation by phosphorylating multiple residues within its amino-terminal regulatory domain. Here we investigated the regulation of B-Raf by serum and glucocorticoid-inducible kinase, SGK, which shares close sequence identity with the catalytic domain of Akt but lacks the pleckstrin homology domain. We observed that SGK inhibits B-Raf activity. A comparison of substrate specificity between SGK and Akt indicates that SGK is a potent negative regulator of B-Raf. In contrast to Akt, SGK negatively regulates B-Raf kinase activity by phosphorylating only a single Akt consensus site, Ser(364). Under similar experimental conditions, SGK displays a measurably stronger inhibitory effect on B-Raf kinase activity than Akt, whereas Akt exhibits a more inhibitory effect on the forkhead transcription factor, FKHR. The selective substrate specificity is correlated with an enhanced association between Akt or SGK and their preferred substrates, FKHR and B-Raf, respectively. These results indicate that B-Raf kinase activity is negatively regulated by Akt and SGK, suggesting that the cross-talk between the B-Raf and other signaling pathways can be mediated by both Akt and SGK.

Diverse extracellular stimuli activate the Ras-mitogen-activated protein (MAP) 1 kinase pathway, also known as extracellular signal-regulated kinase (ERK). The Raf protein kinase plays an essential role in transmitting signals from Ras-GTP to activation of MEK (also known as MAP kinase kinase) and MAP kinase. Activation of the Ras-MAP kinase pathway has been implicated in the modulation of a wide variety of cellular responses including cell proliferation, differentiation, cell death, and development (1)(2)(3)(4). The Ras-MAP kinase pathway is therefore subject to tight regulation in response to a combination of extracellular stimuli. Cross-talk between different signaling pathways is likely to be a critical mechanism for the proper regulation of this multifunctional pathway (5,6).
All MAP kinase modules consist of three kinases acting in a sequence and are conserved during evolution. The regulation of the Raf-MEK-ERK cascade has been studied extensively and is one of the best understood MAP kinase cascades (7,8). MEK directly phosphorylates ERK on a conserved TXY motif between the kinase subdomain VII and VIII and activates ERK (9). Similarly, Raf directly phosphorylates MEK on two serine residues between kinase subdomain VII and VIII and activates MEK (10 -12). However, the regulation of Raf is much more complex and not fully understood.
Three isoforms of Raf proteins have been found in the human genome: A-Raf, B-Raf, and C-Raf (also known as Raf-1) (13). Both B-Raf and C-Raf have been implicated in phosphorylating and activating MEK, but the function of A-Raf has been less documented. Structurally, Raf kinases contain three conserved regions: CR1, CR2, and CR3 (14). The CR1 region consists of a Ras binding domain and a cysteine-rich domain, both of which bind Ras and are important for Raf activation by Ras (15)(16)(17)(18). The CR2 region is rich in serine and threonine residues, and multiple phosphorylation sites within this region are responsible for Raf kinase activation (4). The CR3 region is the kinase catalytic domain, and phosphorylation of residues within this region also responsible for Raf activation (4). Deletion of the amino-terminal regulatory domains of Raf results in activation of Raf kinase activity, suggesting that the amino-terminal domain of Raf inhibits its kinase activity (19). In fact, viral oncogenic Raf was originally isolated as an oncogene, which lacks the amino-terminal regulatory domain (20).
Raf activation by Ras is a feature conserved in Caenorhabditis elegans, Drosophila, and mammals. The interaction between Raf and active Ras is common for all Raf family kinases, and although Ras-Raf binding is necessary, it is not sufficient for Raf activation (21)(22)(23)(24). Studies performed with C-Raf have demonstrated that phosphorylation of Ser 338 and Tyr 341 are essential for Raf activation by Ras, growth factor, and phorbol 12-myristate 13-acetate (25)(26)(27). Using phosphospecific antibodies, Mason et al. (27) have shown that Ras induces predominantly Ser 338 phosphorylation, whereas Src mainly stimulates Tyr 341 phosphorylation. Synergistic activation of Raf was observed when both sites were phosphorylated. Recently, we have demonstrated that phosphorylation of Thr 598 and Ser 601 in B-Raf, which correspond to Thr 491 and Ser 494 of C-Raf, are essential for B-Raf activation (28). Phosphorylation of these two sites is stimulated by active Ras. Simultaneous substitutions of Thr 598 and Ser 601 by acidic residues result in a signif-icant elevation of B-Raf activity. Interestingly, Thr 598 and Ser 601 of B-Raf are conserved in all Raf family kinases and are likely to be common phosphorylation sites for Raf activation (51).
Raf kinase activity is also negatively regulated by phosphorylation. PKA phosphorylates Ser 43 of C-Raf and inhibits its activity (29 -32). Recently, Akt has been reported to phosphorylate Ser 259 , which also inhibits C-Raf kinase activity (33). We have shown that B-Raf kinase activity is inhibited by Akt through multiple sites within the CR2 region (34).
Activation of the phosphatidylinositol 3-kinase (PI3K) leads to an increase in production of phosphoinositol phosphate second messages (35). These lipid messages regulate the activity and localization of a number of target proteins, including those containing pleckstrin homology (PH) domains. Akt, a Ser/Thr kinase, also known as protein kinase B, was identified as a viral transforming oncogene and plays an important role in promoting cell survival (36). Akt contains a PH domain and is a major downstream target of PI3K. Akt is phosphorylated and activated by phospholipid-dependent kinase (PDK) (37,38), which is regulated by the phosphoinositol phosphate lipid messages. Several Akt targets have been identified, including GSK3 (glycogen synthase kinase-3) and p70S6 kinase, transcription factor FKHR, proteins associated with apoptosis, and Raf kinase (36). The serum-and glucocorticoid-inducible kinase (SGK), a novel member of the serine/threonine protein kinase gene family, shares significant sequence identity with Akt (39). Several groups have reported that SGK could also be activated by the PI3K pathway (40,41). PDK1 is likely to directly phosphorylate and enhance SGK activity (41,42). However, physiological targets of SGK are largely unknown. Recently, Brunet et al. (43) reported that SGK could phosphorylate and inhibit the FKHRL1 transcription factor. Interestingly, SGK and Akt selectively phosphorylate different residues in FKHRL1, demonstrating that Akt and SGK coordinately regulate the function of FKHRL1 by phosphorylating this transcription factor on distinct sites.
We investigated the role of SGK in B-Raf regulation. Our results demonstrated that SGK is potent in the phosphorylation and inactivation of B-Raf. A preferential association of SGK and B-Raf supports the effect of SGK on inhibition of B-Raf. SGK inhibits B-Raf mainly through phosphorylation of Ser 364 . Under similar experimental conditions, Akt displayed less ability to inhibit B-Raf. In contrast, Akt appears more effective in phosphorylating and inhibiting the FKHR transcription factor than SGK. This report demonstrates that SGK plays an important role in the negative regulation of Raf.
Kinase Assays-Lysates for immunoprecipitation experiments were prepared from transfected HEK293 cells grown on 6-well plates or 10-cm tissue culture dishes in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Briefly, cells were washed twice with cold phosphate-buffered saline and lysed on ice in 1 ml of lysis buffer (25 mM HEPES, pH 7.5, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, 0.5 mM EDTA, 0.025% mercaptoethanol, 1 mM NaF, 200 M Na 3 VO 4 , 200 M phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 10 g/ml aprotinin). The lysates were cleared by centrifugation at 14,000 ϫ g for 10 min at 4°C to remove insoluble debris. Immunoprecipitation was performed by incubation with appropriate antibodies for 1.5 h at 4°C followed by incubation for a further 1 h at 4°C with protein G-Sepharose. Immunoprecipitates were washed three times with lysis buffer and once with phosphate-buffered saline. B-Raf kinase activity was measured by coupled assay using GST-MEK, GST-ERK, and GST-Elk as sequential substrates as described previously (28). ERK activity was determined using GST-Elk as a substrate (28).
For the Akt/SGK kinase assays, HA-SGK or HA-Akt immunoprecipitates were incubated for 30 min at 30°C with 40 l of reaction mixture containing 5 mM synthetic peptide (KKRNRRLSVA) as a substrate, 18 mM HEPES, 10 mM MgCl 2 , 1 mM dithiothreitol, 50 mM ATP, 1 M PKA inhibitor peptide (Calbiochem) and 10 Ci of [␥-32 P]ATP. 20 l of 40% trichloroacetic acid was added to the supernatant and incubated for 5 min at room temperature. 40 l of the latter fraction was spotted onto P81 Whatman paper. The paper was washed extensively with 80 mM phosphoric acid and once with ethanol and then counted.
Co-immunoprecipitation Assays-For examination of interaction between B-Raf and SGK or Akt, GST-vector or GST-B-Raf was cotransfected with HA-SGK or HA-Akt in HEK293 cells grown in 10-cm plates. 48 h after transfection, cells were washed twice with ice-cold phosphatebuffered saline and lysed in 10 mM HEPES, pH 7.5, 50 mM NaCl, 1% Triton X-100, 2 mM EDTA, 0.1% mercaptoethanol, 50 mM NaF, 1 mM Na 3 VO 4 , 200 M phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml aprotinin. Glutathione-agarose beads were added to the cell lysates to purify GST-B-Raf. GST-B-Raf was then eluted in 10 mM glutathione in 50 mM Tris, pH 8.0 The glutathione-eluted samples were subjected to SDS-polyacrylamide gel electrophoresis and analyzed by Western blot with anti-GST and anti-HA antibodies. Similar experiments were performed for the association of FKHR and SGK or Akt with cells cotransfected with Flag-FKHR and HA-SGK or HA-Akt, except Flag-FKHR was immunoprecipitated with anti-Flag M2 antibody. To immunoprecipitate endogenous SGK, from two 10-cm plates HEK293 cells were immunoprecipitated using antiserum specific for SGK (43). The immunoprecipitated samples were probed with anti-B-Raf antibody.

RESULTS
Inhibition of B-Raf by SGK-We have shown previously that B-Raf can be phosphorylated and inhibited by Akt (34). To test whether SGK plays a role in B-Raf regulation, we examined B-Raf kinase activity in the presence or absence of SGK. HA-B-Raf was transfected in HEK293 cells with SGK-S422D, which is a constitutively active SGK mutant containing the activation phosphorylation site Ser 422 substituted by an aspartic residue (42). HA-SGK-S422D was used because wild-type SGK has low activity in the absence of serum stimulation. We observed that HA-SGK-S422D decreased the basal kinase activity of B-Raf, which was assayed by an in vitro coupled kinase assay ( Fig. 1, lanes 2 and 3). The effect of SGK on Ras-induced B-Raf activity was also examined. The RasV12-induced B-Raf activity was also inhibited by HA-SGK-S422D (64% inhibition, Fig. 1, lanes 4 and 5). However, under similar conditions, HA-Akt-myr, a constitutively active Akt containing a myristylation signal, also inhibited B-Raf (50% inhibition, Fig. 1, lane  7). As a negative control, the kinase-inactive mutant HA-SGK-K127M was tested and showed no effect on RasV12-induced B-Raf activation (Fig. 1, lane 6). These results demonstrated that kinase activity of SGK is essential for it's ability to inhibit B-Raf activity.
SGK kinase activity is known to be stimulated by serum (40,41). Similarly, B-Raf kinase activity is stimulated by serum. We tested whether SGK can inhibit the serum-induced B-Raf activation. Wild-type SGK, but not the kinase-inactive SGK, effectively blocked B-Raf activation (Fig. 1, lanes 8 -11). Under identical conditions, Akt-myr is less effective. Our data indicate that SGK is a negative regulator of Raf.
Inhibition of B-Raf by SGK Requires the Presence of Ser 364 in B-Raf-Biochemical characterizations demonstrate that SGK recognize a consensus phosphorylation site, RXRXX(S/T), similar to that of Akt (40,41). B-Raf contains three putative Akt/SGK phosphorylation sites. Our previous studies indicated that Ser 364 and Ser 428 are phosphorylation sites by Akt (34). We examined whether SGK also uses similar sites to inhibit B-Raf activity. HA-B-Raf-A, which has the Ser 364 substituted by an alanine residue, is resistant to inhibition by HA-SGK-422D ( Fig. 2A, lanes 7 and 8), indicating that Ser 364 in B-Raf is a target site for SGK. In contrast, HA-B-Raf-AA, which has both Ser 428 and Thr 439 substituted by alanine residues, still can be inhibited by HA-SGK-S422D ( Fig. 2A, lanes 9 and 10). These results suggest that Ser 428 and Thr 439 in B-Raf are not required for its inhibition by SGK. As predicted, HA-B-Raf-AAA, which has all three Akt consensus sites replaced by alanine residues, is also not inhibited by HA-SGK-S422D (Fig.  1, lanes 11 and 12). In the same experiments, the basal activity of wild-type B-Raf is effectively inhibited by HA-SGK-S422D ( Fig. 2A, lanes 1 and 2). Similarly, activation of wild-type B-Raf by serum is also inhibited by SGK ( Fig. 2A, lanes 3 and 4). These results demonstrate that Ser 364 in B-Raf is essential for inhibition by SGK.
We wanted to test whether B-Raf-induced ERK activation is inhibited by SGK in transfected HEK293 cells. Myc-ERK is co-transfected with HA-B-Raf in the presence or absence of SGK. Kinase activity of immunoprecipitated ERK was determined using Elk-1 as a substrate. Wild-type B-Raf co-transfection can induce a low level of ERK activation, which is inhibited by HA-SGK-S422D (Fig. 2B, lanes 1-3). We have previously identified important activation phosphorylation sites in B-Raf (28). Substitution of these sites by acidic residues created a constitutively active HA-B-Raf-ED. ERK activation induced by B-Raf-ED was also suppressed by SGK, whereas kinase-inactive HA-SGK-K127M had no inhibition at all (Fig. 2B, lanes  4 -6). As a comparison, Akt-myr displayed a moderate inhibition of ERK activation by HA-B-Raf-ED mutant (Fig. 2B, lane  7). To further confirm the requirement of Ser 364 in B-Raf inhibition by SGK, mutants of HA-tagged B-Raf-A, B-Raf-AA, and B-Raf-AAA were examined. These alanine mutations were constructed under the wild-type B-Raf background and resulted in an elevation of basal kinase activity (Fig. 2B). SGK inhibited HA-B-Raf-AA but not HA-B-Raf-A or HA-B-Raf-AAA-induced Myc-ERK activation (Fig. 2B, lanes 8 -13). These results are completely consistent with the Raf activity assay in Fig. 2A and demonstrate that Ser 364 in B-Raf is critical for inhibition by SGK.
SGK Stimulates Phosphorylation of Ser 364 in B-Raf-Ser 364 is highly conserved in Raf family kinases, including the C. elegans lin-45, and corresponds to Ser 259 in C-Raf. Ser 259 has been demonstrated to be a negative phosphorylation site. A phosphospecific antibody recognizing Ser 259 of C-Raf is available. We found that this phosphospecific antibody can also recognize B-Raf. To further confirm that Ser 364 in B-Raf is the target phosphorylation site of SGK, immunoblot of B-Raf was performed with the Ser 259 phosphospecific antibody. Co-transfection of HA-SGK-S422D significantly increased the phosphorylation of Ser 364 in B-Raf, whereas the kinase-inactive SGK had no effect at all (Fig. 2C, lanes 2-4). The recognition of the phosphospecific antibody is specific because substitution of Ser 364 by an alanine residue completely eliminated the recognition by this antibody (Fig. 2C, lanes 6 -8). Our observations confirm that Ser 364 in B-Raf is a phosphorylation site by SGK.
Differential Selectivity of Akt and SGK toward FKHR-The

FIG. 3. SGK exhibits lower activity to phosphorylate FKHR than Akt.
A, in vitro kinase activity of SGK and Akt using peptide substrate. SGK and Akt were transfected into HEK293 cells. The proteins were immunoprecipitated, and kinase activity was assayed using a peptide substrate (see "Experimental Procedures"). Kinase activities were measured by incorporation of 32 P into substrates. Similar amounts of protein were used in kinase assays for SGK and Akt. Vector transfection was included as a background control. B, in vitro phosphorylation of FKHR by SGK and Akt. Immunoprecipitated Akt or SGK was assayed in vitro using purified GST fusion of the amino-terminal or the carboxyl-terminal fragment of FKHR as substrate. Phosphorylation of FKHR fragments is shown in the upper panel. N denotes the amino terminus of wild-type FKHR, and NЈ denotes the same amino-terminal fragment containing T24A mutation. Similarly, C and CЈ denote the wild-type and S256A/S319A mutant of the carboxyl-terminal fragment of FKHR, respectively. Kinase proteins used in phosphorylation assays are shown in the lower panel as indicated. Quantitation of phosphorylation of the carboxyl-terminal fragment of FKHR is indicated underneath the upper panel for lanes 7, 11, 15, and 19. The value for wild-type SGK was arbitrary set as 1. C, inhibition of FKHR transcriptional activity by SGK requires the intact phosphoacceptor sites. A FKHR reporter was co-transfected with various forms of SGK and wild type (lanes 1-4) or the constitutively active FKHR mutant (lanes 5-7). FKHR-AAA has serine to alanine substitutions at all three putative Akt phosphorylation sites and is constitutively active. The relative luciferase activity is measured and normalized with co-transfected ␤-galactosidase internal control. The error bars are derived from three independently duplicated experiments. Expression levels of SGK and Akt are shown in the bottom panel.
apparent high potency of SGK in the inhibition of B-Raf activation could be because of the following two reasons. SGK may have a higher kinase activity than Akt. Alternatively, SGK may have higher substrate selectivity than Akt toward B-Raf. To compare the kinase activity of SGK and Akt, we performed in vitro kinase assays of SGK and Akt using a peptide substrate that has the consensus recognition sequence of both SGK and Akt. In vitro kinase assay demonstrated that wild-type SGK has a low basal activity, whereas SGK-S422D displayed a much higher kinase activity (Fig. 3A). Akt-myr showed kinase activity similar to that of SGK-S422D when the peptide substrate was used (Fig. 3A). The in vitro kinase assays show that SGK-S422D has activity comparable with AKT-myr toward the peptide substrate, suggesting that SGK has a high substrate selectivity toward B-Raf.
The forkhead family of transcription factors have been shown to be physiological targets of Akt. We have previously identified that FKHR can be phosphorylated and inhibited by Akt (44). To further compare the substrate selectivity of Akt and SGK, we examined the phosphorylation of recombinant FKHR fragments by immunoprecipitated Akt and SGK. FKHR fragments were expressed as GST fusion in E. coli and purified. Wild-type SGK displayed little activity toward the amino-terminal fragment of FKHR (Fig. 3B, lanes 5 and 6), which contains one Akt phosphorylation site, Thr 24 . Similarly, SGK-S422D showed limited phosphorylation on the amino-terminal fragment of FKHR (Fig. 3B, lanes 9 and 10). In contrast, Akt-myr phosphorylated the amino-terminal fragment of FKHR, albeit weakly (Fig. 3B, lanes 17-20). The phosphorylation likely occurred on Thr 24 because substitution of this residue by alanine eliminated the phosphorylation by either Akt or SGK (Fig. 3B, lanes 10 and 18).
When the carboxyl-terminal fragment of FKHR, which contains two Akt consensus sites, Ser 256 and Ser 319 , was used in the kinase assay, some background phosphorylation was observed in the negative control experiments (lanes 3 and 4). However, SGK-S422D caused a considerable increase in phosphorylation of the carboxyl-terminal fragment of FKHR (Fig.  3B, lanes 11 and 12). The negative control of the kinase-dead SGKK127M mutant displayed no activity above the background (lanes 15 and 16). Interestingly, Akt was much more active toward the carboxyl-terminal fragment of FKHR than SGK-S422D (lanes 11 and 19). Under similar conditions, phosphorylation of the carboxyl-terminal fragment by AKT was 2-fold stronger than by SGK. The phosphorylation of FKHR carboxyl-terminal fragment was eliminated when Ser 256 and Ser 319 were replaced by alanine residues (lanes 12 and 20). These observations demonstrate that the Akt consensus sites were utilized by both AKT and SGK. Furthermore, Akt is a more potent kinase toward FKHR than SGK.
Phosphorylation of FKHR results in a decrease of the transcriptional activity of FKHR. We compared the effect of Akt and SGK on FKHR activity in vivo. Expression of Akt-myr effectively inhibited the FKHR dependent transcription (Fig.  3C, lanes 1 and 4). Similarly, expression of HA-SGK-S422D inhibited FKHR activity, although less effective than Akt, whereas the kinase-inactive mutant SGK did not inhibit FKHR activity (Fig. 3C, lanes 2 and 3). We have shown that elimination of the Akt phosphorylation sites in FKHR dramatically increases the transcription activity of FKHR (44). The mutant FKHR-AAA molecule is not inhibited by Akt (Fig. 3C, lanes 5  and 7). Similarly, the transcription activity of FKHR-AAA is not inhibited by SGK (Fig. 3C, lane 6), supporting the notion that SGK inhibits FKHR via phosphorylation of the Akt con- sensus sites. The above data in combination with results from Fig. 2 clearly demonstrate that Akt preferentially phosphorylates and inhibits FKHR, whereas SGK is more active than Akt toward B-Raf.
Differential Association of Akt and SGKT with FKHR and B-Raf-Direct protein-protein interaction has been found for many protein kinases and their respective substrates. We examined whether SGK and AKT showed a differential interaction with B-Raf and FKHR. GST-B-Raf was co-transfected with HA-SGK or HA-Akt in HEK293 cells. GST-B-Raf was purified by glutathione-agarose resin and subjected to Western blotting with anti-HA for co-purified SGK or Akt. Fig. 4A clearly shows that complex between SGK and B-Raf is more stable under these conditions than the interaction between Akt and B-Raf (Fig. 4A, top panel) although both SGK and Akt were expressed at a similar level (Fig. 4A, bottom panel). The association of SGK and Akt with B-Raf is specific because the negative control of GST co-precipitated neither Akt or SGK.
Similar co-immunoprecipitation was performed with FKHR. Flag-FKHR was co-expressed with HA-Akt or HA-SGK in HEK293 cells. Cell lysates were immunoprecipitated with anti-Flag to precipitate Flag-FKHR and Western blotted with anti-HA antibody for HA-SGK or HA-Akt. The results in Fig. 4B show that Akt preferentially associates with FKHR. To further test the association between SGK and B-Raf, immunoprecipitation of endogenous proteins was performed. SGK-specific antibody precipitated B-Raf from untransfected HEK293 cells (Fig. 4C, lane 2), whereas the control serum did not precipitate SGK or B-Raf. These results confirm that SGK can form a complex with B-Raf in vivo. The preferential association between SGK and B-Raf is consistent with a strong inhibition of B-Raf by SGK. DISCUSSION Recently Akt has been implicated in negative regulation of B-Raf and C-Raf (33,34). Akt is reported to phosphorylate the CR2 region and inhibit Raf activity. SGK belongs to a new family of protein kinases that is closely related to Akt (39). SGK recognizes the RXRXX(S/T) consensus sequence (40,41). The biological function of SGK is not clear, partly because few physiological substrates of SGK have been identified. Our results demonstrate that B-Raf is a potential substrate of SGK. Active SGK mutant effectively inhibits B-Raf activation in response to serum or Ras by phosphorylation of Ser 364 . Mutation of Ser 364 by alanine renders B-Raf resistant to inhibition by SGK, indicating that phosphorylation of Ser 364 is responsible for the inhibition of B-Raf activity. In contrast, mutation of other Akt consensus sites, Ser 428 and Thr 439 , has no effect on the ability of B-Raf to be inhibited by SGK. In contrast to Akt, SGK can phosphorylate serine phosphoacceptor sites that do not have a bulky hydrophobic amino acid residue immediately carboxyl-terminal to the phosphoacceptor residue (40). The sequence surrounding Ser 364 matches preferred SGK phosphorylation sites, including an alanine at position 365 in B-Raf.
The relative specificity of SGK toward B-Raf is visible when compared with Akt and FKHR. SGK is more potent in B-Raf inhibition, whereas Akt is more effective in FKHR inhibition. Protein-protein interaction data are consistent with the inhibition results and further support the observation that B-Raf is a preferred substrate of SGK. The differential association between SGK and Raf likely contributes to the substrate selectivity of SGK toward B-Raf. Taken together, our data strongly indicate that SGK plays an important role in Raf regulation. Future quantitative analyses of phosphorylation and detailed kinetic studies are required for the further understanding of the relative substrate specificity of AKT and SGK.
Akt and SGK are similarly activated by PDK1 and modu-lated by the PI3K pathway (40,41 (43). In addition, SGK is regulated at both transcriptional and post-translational levels (39,45). The expression of SGK is rapidly induced by a variety of stimuli including serum, steroid hormones, cytokines, and osmotic stress (39, 46 -48). These differences suggest that Akt and SGK may have complementary rather than redundant functions. We propose that both SGK and Akt play roles in the negative regulation of Raf, with SGK being a more potent inhibitor. The fact that SGK is induced by serum and growth factors indicates that SGK may constitute a negative feedback loop to suppress constitutive activation of the Raf-MEK-ERK kinase cascade. Protein kinase A is known to phosphorylate and inhibit Raf activity. In C-Raf, PKA is reported to phosphorylate Ser 43 and inhibit C-Raf activation (29,32). However, recent results dispute the role of Ser 43 in regulation of C-Raf by PKA and indicate that Ser43 is not required for C-Raf inhibition by PKA (49). Furthermore, Ser 43 is not conserved in B-Raf, and yet B-Raf is also negatively regulated by PKA. The effect of cAMP on B-Raf regulation is rather complex, because cAMP can also induce B-Raf activation in some cells such as PC12 (50). Thus, the mechanism of Raf inhibition by cAMP is not completely understood. Interestingly, SGK is activated by intracellular cAMP (48). Therefore, SGK may mediate the inhibitory effects of cAMP on Raf activity. The regulation of SGK by cAMP could provide an important mechanism for cross-talk between signal transductions of the trimeric G-protein-coupled serpentine receptors and the Ras-coupled tyrosine kinase receptors.
The SGK phosphorylation site Ser 364 in B-Raf is conserved in C-Raf, Drosophila Raf, and the C. elegans lin-45 Raf. These conservations suggest that SGK will likely play a role in the negative regulation of all members of the Raf family kinases. We have observed that the corresponding sites in lin-45 Raf indeed play a negative role in lin-45 function in vivo. Substitution of the Akt/SGK recognition site by alanine results in an active lin-45 Raf and induces multi-vulvae phenotypes (51). Interestingly, C. elegans contains a single SGK predicted by the genome sequence. The fact that SGK is regulated by a wide variety of extracellular stimuli provides a point of signal integration. Regulation of SGK at either the transcriptional or post-translational level represents a possible mechanism of cross-talk between other signaling pathways and the Raf-ERK pathway.