Akt Protein Kinase Inhibits Rac1-GTP Binding through Phosphorylation at Serine 71 of Rac1*

A putative Akt kinase phosphorylation site (64ydRIRplSYp73) was found in Rac1/CDC42 and Rho family proteins (RhoA, RhoB, RhoC, and RhoG). Phosphorylation of Rac1 by Akt kinase was assayed with recombinant Rac1 protein and the fluorescein-labeled Rac1 peptide. It was shown that the Rac1 peptide and the recombinant protein were phosphorylated by the activated recombinant Akt kinase and the lysate of SK-MEL28 cells, a human melanoma cell line. The phosphorylation of Rac1 inhibited its GTP-binding activity without any significant change in GTPase activity. Both the GTP-binding and GTPase activities of Rac1 S71A protein (with the serine residue to be phosphorylated replaced with alanine) were abolished regardless of the treatment of Akt kinase. Akt kinase activity and Rac1 peptide phosphorylation were down-regulated by the treatment of SK-MEL28 cells with wortmannin or LY294002 (a phosphoinositide 3-kinase inhibitor), but JNK/SAPK kinase activity was up-regulated. Thus, the results suggest that Akt kinase of the phosphoinositide 3-kinase signal transduction pathway phosphorylates serine 71 of Rac1 as one of its authentic substrates and modulates the Rac1 signal transduction pathway through phosphorylation.

Rac1 and CDC42 belong to a subgroup of the Rho GTPase family that binds to and hydrolyzes GTP (1,2). Several studies have shown that GTP-bound Rac1 is active and plays an important role in the control of cell shape, adhesion, movement, endocytosis, secretion, and growth while regulating various aspects of actin cytoskeleton organization (3,4). GTP binding to Rac1 is controlled by Ͼ15 GTP/GDP exchange factors (GEF) 1 and ϳ10 GTPase-activating proteins, depending on the external or internal signals (1,5,6). The role of Rac1 in the signal transduction pathway was also intensively characterized (1)(2)(3)(4)(5)(6). It was reported that activated Rac1 binds to and activates PAK65 protein kinase (7,8). Activated PAK65 can stimulate MEKK1, which in turn phosphorylates and activates SEK/JNK kinase (9,10). The active SEK/JNK kinase phosphorylates JNK/SAPK, which in turn binds to and phosphorylates the N-terminal region of c-Jun (11). Rac1 also activates MEKK3, which in turn phosphorylates MKK3, which mediates the p38 signal transduction pathway (1,12). Thus, these results revealed that Rac1 plays a pivotal role in the signal transduction pathway that regulates multiple cellular functions.
The relationship between the Rac1 and PI3K signal pathways was still unclear. It was suggested that the activation of CDC42 or Rac1 disrupts the normal polarization of mammary epithelial cells in a collagen matrix and promotes motility and invasion with activating PI3K (13). Other data revealed that PI3K and Rac1 form a complex and that Akt kinase and Rac1 are on separate pathways downstream of PI3K (14). It was also reported that p85 (the regulatory subunit of PI3K) binds to GTP-bound Rac1/CDC42 and regulates Rac1/CDC42 downstream function (15,16). Thus, these studies indicated that PI3K is a downstream effector of Rac1. However, most studies suggested that PI3K functions upstream of Rac1 (2,6,8). In the T-cell signal pathway, it was reported that PI3K regulates Rac1 function (17). Other data also revealed that lipid products of PI3K interact with Rac1 to stimulate GDP dissociation (18) and that the direct PI3K activation is sufficient to disrupt epithelial polarization and to induce cell motility and invasion, indicating that PI3K inhibition disrupts actin structures (19). Thus, although it is unclear whether PI3K is upstream of the Rac1 signal transduction pathway or not, it seems that the PI3K signal pathway cross-talks with Rac1.
To address how the PI3K and Rac1 signal transduction pathways are connected to each other, we investigated the relationship between Rac1 and Akt kinase instead of PI3K since Akt kinase has been characterized as the primary signal transducer of the PI3K signal pathway (20,21). Akt kinase in the PI3K signal pathway has been reported to deliver a survival signal while protecting cells from the apoptotic cell death induced by growth factor withdrawal (21). Moreover, the specific amino acid sequence (xxRxRxx(S/T)xx, with the hydrophobic amino acid underlined) that can be phosphorylated by Akt kinase was characterized from all known Akt kinase substrate proteins (20 -25). Thus, the investigation of whether a protein contains the sequence xxRxRxx(S/T)xx may help to determine whether the protein is a possible Akt kinase substrate. From Rac1, CDC42, RhoA, RhoB, RhoC, and RhoG proteins, the specific sequence ( 64 ydRIRplSYp 73 ) that could be phosphorylated by Akt kinase was identified (26 -31). The putative Akt kinase phosphorylation site (serine 71) is located between the effector protein-binding domain and the GTP-binding domain (32). Thus, this finding led us to determine whether Akt kinase phosphorylates Rac1 as one of its substate proteins and whether Rac1 phosphorylation by Akt kinase is one of the signal cross-talks between the Rac1 and PI3K signal transduction pathways. To demonstrate this, we performed an Akt kinase assay with recombinant Rac1 protein and the fluorescein-conjugated Rac1 peptide ( 64 ydRIRplSYp 73 ) and observed the phosphorylation of both recombinant Rac1 protein and the fluorescein-conjugated Rac1 peptide. To characterize how the phosphorylation by Akt kinase modulates Rac1 function, we compared Rac1 GTPase activity and its GTP binding with and without Akt kinase treatment. Thus, we observed that Rac1-GTP binding was significantly inhibited by Akt kinase phosphorylation, without a change in GTPase activity. With the Rac1 S71A mutant (with serine replaced with alanine, the mutant was not phosphorylated by Akt kinase), we observed that both its GTPase activity and GTP binding were abolished. In addition, we investigated whether Rac1 phosphorylation and JNK/SAPK activity are also modulated by Akt kinase activity by treatment with wortmannin or LY294002 (an Akt kinase inhibitor) in the SK-MEL28 cell line (20 -23). Furthermore, we observed the down-regulation of both Rac1 phosphorylation and Akt kinase activity and the activation of JNK/ SAPK by treatment with wortmannin or LY294002. Therefore, our observations strongly suggest that Akt kinase phosphorylates Rac1 protein as its target protein, resulting in the inhibition of Rac1-GTP binding.

EXPERIMENTAL PROCEDURES
Cell Culture-SK-MEL28 cells (a human melanoma cell line) were purchased from American Type Culture Collection (Manassas, VA). Media and supplements were obtained from Life Technologies, Inc. The cell line was maintained in Dulbecco's modified essential medium containing 10% heat-inactivated (30 min at 56°C) fetal bovine serum, 100 units/ml potassium penicillin, 100 g/ml streptomycin, 2 mM glutamine, and 20 mM sodium bicarbonate. The cells were incubated at 5% CO 2 and 95% humidity in a 37°C chamber. The growth medium was changed every 3 days.
Akt Protein Kinase Assay-The Akt kinase assay was performed following the protocol provided by Promega (Madison, WI) with the PepTag nonradioactive protein kinase C assay system, except for the substrate peptides (24). For Akt kinase substrates, the fluoresceinconjugated (the amino terminus of the peptide was conjugated with fluorescein isothiocyanate) IRS-1 ( 30 RKRSRKESYS 39 ) and Rac1 ( 64 ydRIRplSYp 73 ) oligopeptides were purchased from Peptron Co. (Daejun, Korea) (25-31). 5 g of fluorescein-conjugated oligopeptide was incubated with 10 l of differentially treated cell lysates or the activated Akt kinase in 20 l of protein kinase reaction mixture (20 mM HEPES, pH 7.2, 10 mM MgCl 2 , 10 mM MnCl 2 , 1 mM dithiothreitol, 0.2 mM EGTA, 20 M ATP, 1 g of phosphatidylserine, and protein kinase activator) at 30°C for 30 min. The reactions were stopped by heating at 95°C for 10 min. The phosphorylated peptide was separated on 0.8% agarose gel at 100 V for 15 min. The phosphorylated products, which gained one more negative charge, migrated to the anode. After the gel was photographed on a transilluminator, the phosphorylated peptide band was sliced out. The absorbance at 488 nm of the phosphorylated substrate was measured with the spectrophotometer following the protocol provided by Promega.
Activation of the Recombinant GST-Akt Kinase Protein-Akt kinaseagarose beads and alkaline protein phosphatase were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY) and activated following the protocol described in our previous report (24). SK-MEL28 cells (10 7 ) grown under serum-free conditions for 24 h were lysed with radioimmune precipitation assay lysate buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM sodium orthovanadate, 100 g/ml phenylmethylsulfonyl fluoride, and 1 g of aprotinin). The cell lysate was precleaned with GST-agarose beads. Akt kinase-agarose beads (20 g) were incubated with 50 l of precleaned cell lysate and protein kinase assay buffer (20 mM HEPES, pH 7.2, 10 mM MgCl 2 , 10 mM MnCl 2 , 1 mM DTT, 0.2 mM EGTA, 20 M ATP, 5 g phosphatidylserine, protein kinase activator) in a final volume of 100 l for 1 h at 30°C. The beads were precipitated and washed three times with excess cell lysis buffer. The final pellet was used for the Akt kinase assay.
Expression and Purification of Recombinant Proteins-GST-wildtype Rac1 and GST-Rac1 S71A fusion proteins were expressed in Escherichia coli and purified on glutathione-Sepharose beads (Amersham Pharmacia Biotech) according to the manufacturer's instructions. The recombinant GTPases were released from the beads by cleavage with human thrombin, and thrombin was removed by adding 10 ml of paminobenzamidine-agarose beads for 30 min at 4°C. Purified proteins were dialyzed against 15 mM Tris, pH 7.5, 150 mM NaCl, 5 mM MgCl 2 , and 0.1 mM DTT and concentrated by ultrafiltration with a Centricon-10 (Amicon, Inc.). Active protein concentrations were determined by a filter binding assay using [ 3 H]GTP as described (32). We performed Akt kinase phosphorylation and GTPase assays with these recombinant Rac1 proteins.
GTP Binding Assay-The assays monitoring the binding of GTP were performed as described previously (26,33,34). The phosphorylated or unphosphorylated protein (35 pmol) was incubated with [␥-32 P]GTP (4 Ci, 20 Ci/mmol; NEN Life Science Products) in 100 l of binding buffer containing 50 mM HEPES, pH 7.6, 0.2 mg/ml bovine serum albumin, and 0.5 mM EDTA at 15°C for the indicated times. The reaction was stopped by the addition of 1 ml of ice-cold wash buffer (50 mM HEPES, pH 7.6, 150 mM NaCl, and 10 mM MgCl 2 ). The samples were then applied to a nitrocellulose membrane that had been rinsed with 3 ml of wash buffer. The filters were immediately washed twice with 3 ml of wash buffer and soaked in 0.2 ml of 0.1 M KOH for 30 min, and the radioactivity was counted by liquid scintillation spectrometry.
GTPase Assay-Rac1 GTPase was assayed as described (33)(34)(35). The phosphorylated or unphosphorylated protein (350 pmol) was loaded with 400 pmol of [␥-32 P]GTP in 100 l of binding buffer at 15°C for 15 min. The reaction was initiated by adding MgSO 4 to 10 mM in 200 l of GTPase buffer (50 mM HEPES, pH 7.6, 1 mM DTT, and 5 mM EDTA). Reaction mixtures were then incubated at 15°C for the indicated times. Aliquots (50 l) were removed at the indicated times and mixed with 750 l of 5% (w/v) Norit A in 50 mM NaH 2 PO 4 . The mixture was centrifuged at 2000 rpm for 5 min, and 400 l of supernatant containing 32 P i was counted by liquid scintillation spectrometry. Site-directed Mutagenesis of Rac1 S71A-The Rac1 point mutation with serine 71 replaced with alanine was introduced with the Chameleon double-stranded, site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. The Rac1 S71A mutation was confirmed by DNA sequencing.
Rac1 and Rac1 S71A Expression Vector Transfection and Purification-Hexahistidine-tagged Rac1, Rac1 S71A, and dominant-positive Akt kinase were each cloned in the pUSEamp(ϩ) mammalian expression vector (Upstate Biotechnology, Inc.). They were transfected or cotransfected in SK-MEL28 cells following the Lipofectin transfection method (Life Technologies, Inc.). Transfected cells (2 ϫ 10 7 ) were lysed in radioimmune precipitation assay lysate buffer, and Ni 2ϩ -agarose beads (20 g) were incubated with 500 l of precleaned cell lysate. The beads were precipitated and washed three times with excess cell lysis buffer. The final pellet was used for the GTPase and GTP binding assays as described above.

RESULTS
Phosphorylation of Rac1 by Akt Kinase-It has been shown that Akt kinase recognizes a specific peptide sequence as its substrate protein (20 -25, 36 -38). In our previous report, we also showed that human telomerase is activated by Akt kinase phosphorylation (24). Inspecting the amino acid sequences of all known Akt kinase substrates (H2B, BAD, 6-phosphofructo-2-kinase, IRS-1, glycogen synthase kinase-3, caspase-9, and human telomerase reverse transcriptase), we found that the serine residue is found within stretches of homologous amino acids (Table I). The arginine residues at positions Ϫ5 and Ϫ3 and the hydrophobic amino acid at position ϩ2 are conserved relative to the serine/threonine residues that are probably phosphorylated in these proteins. Therefore, the peptide xxRxRxx(S/T)xx (the hydrophobic amino acid is underlined) seems to be conserved in Akt kinase substrates (20 -25, 36 -38). We inspected possible Akt kinase substrate protein sequences with Akt kinase substrate sequence specificity (xxRxRxx(S/ T)xx). Interestingly, we identified the putative Akt kinase phosphorylation site ( 64 ydRIRplSYp 73 ) in Rac1 and all known Rho family proteins (Table I) (26 -31). This finding led us to determine that Akt kinase phosphorylates Rac1, which has been intensively studied as an important regulatory protein of actin organization, motility, invasion, and signal transduction (1)(2)(3)(4)(5)(6).
To demonstrate whether Akt kinase phosphorylates Rac1, we performed an Akt kinase assay with the activated Akt kinase and Rac1 protein in vitro. As shown in Fig. 1A, recombinant Rac1 protein was phosphorylated by Akt kinase. Recombinant Rac1 S71A protein was produced by replacing Ser-71 (to be phosphorylated) with alanine, and the Rac1 S71A mutant was not phosphorylated by Akt kinase. As shown in Fig. 1A and by other data (37,38), Akt kinase was also phosphorylated by itself. Thus, it seems that Akt1 and Akt2 are self-phosphorylated at the intrinsic phosphorylation consensus sequence ( 65 teRpRpnTFi 74 ) in their amino acid sequences (39 -41). However, the function of Akt kinase self-phosphorylation remains to be determined. To confirm Akt kinase protein phosphorylation, the reaction mixture was treated with alkaline phosphatase, and the phosphorylated protein bands disappeared (Fig. 1A).
To further determine that Ser-71 ( 64 ydRIRplSYp 73 ) in Rac1 is phosphorylated by Akt kinase, the fluorescein-labeled Rac1 peptide ( 64 ydRIRplSYp 73 ) was incubated with Akt kinase, which was expressed in E. coli and activated with human melanoma cell lysate. The amount of phosphorylated Rac1 peptide increased depending on the activated Akt kinase concentration (Fig. 1B) and the reaction time (Fig. 1C). For the control, the recombinant GST-Akt kinase protein, which has no kinase activity, was also included, but it gave no detectable amount of phosphorylated product (Fig. 1, B and C). Taken together, these observations strongly demonstrate that Akt kinase phosphorylates Ser-71 of Rac1 as one of its substrate proteins.
Akt Kinase Phosphorylation Inhibits Rac1-GTP Binding, but Not Rac1 GTPase Activity, in Vitro and in Vivo-GTP-bound Rac1 is active and controls the cell shape, adhesion, movement, endocytosis, secretion, and downstream JNK/SAPK activity (1)(2)(3)(4)(5)(6)(7)(8). Thus, we investigated whether Rac1 GTPase activity or its GTP binding was changed with or without Akt kinase pretreatment to determine that the phosphorylation by Akt kinase affects Rac1 function. In the experiments with recombinant Rac1 or Rac1 S71A expressed in E. coli, we observed that Rac1 GTPase activity was unchanged regardless of Akt kinase pretreatment (Fig. 2). To confirm this experiment, we used recombinant Rac1 S71A. The GTPase activity of the Rac1 S71A mutant was abolished regardless of Akt kinase pretreatment (Fig. 2). It seems that the lack of GTPase activity in the Rac1 S71A mutant is similar to that of the Rac1/CDC42 Q61L mutant (15). These results suggest that the phosphorylation by Akt kinase does not inhibit Rac1 GTPase activity in vitro. We also measured Rac1/CDC42 protein phosphorylation by Western blotting (data not shown).
It was confirmed by the cotransfection experiment that Rac1 phosphorylation by Akt kinase inhibits Rac1-GTP binding in vivo. When Rac1 was cotransfected with dominant-positive Akt kinase in SK-MEL28 cells, the GTP-binding activity of Rac1 was reduced (Fig. 4B), but its GTPase activity was not likely to be affected (Fig. 4A). The GTPase (Fig. 4A) and GTP-binding (Fig. 4B) activities of Rac1 S71A were abolished and not affected by the cotransfection of Akt kinase, consistent with in vitro assay results (Figs. 2 and 3). Taken together, the results suggest that Rac1 phosphorylation by Akt kinase inhibits its GTP binding without affecting its GTPase activity in vitro and in vivo.
Drugs Affecting Akt Kinase Activity Also Regulate Rac1 Phosphorylation and JNK/SAPK Activity-To further demonstrate that Akt kinase activity regulates Rac1 phosphorylation in vivo, wortmannin and LY294002 (an Akt kinase inhibitor) were used to modulate Akt kinase activity. Because it has been reported that wortmannin and LY294002 are PI3K/Akt kinase pathway inhibitors that result in specific Akt kinase inactivation (36 -38), we assayed Rac1/CDC42 phosphorylation in the SK-MEL38 cell line with wortmannin (0, 50, and 100 nM for a 2-h treatment) or LY294002 (0, 5, and 10 M). To monitor any change in Akt kinase activity upon wortmannin or LY294002 treatment, the fluorescein-conjugated IRS-1 peptide was used to monitor the activity of Akt kinase as a substrate (25). As shown in Fig. 5A (upper panel), wortmannin and LY294002 inhibited Akt kinase in a dose-dependent manner.
To determine whether Rac1 phosphorylation is also modulated along with a change in Akt kinase activity, we assayed Rac1/CDC42 phosphorylation in the human melanoma cell line with wortmannin (0, 50, and 100 nM for a 2-h treatment) or LY294002 (0, 5, and 10 M). Instead of the fluorescein-conjugated IRS-1 peptide, the fluorescein-labeled Rac1 peptide ( 64 ydRIRplSYp 73 ) was used as an Akt kinase substrate. As Akt kinase activity was inhibited by wortmannin or LY294002, Rac1 phosphorylation was also inhibited in a dose-dependent manner (Fig. 5A, lower panel). Thus, the results suggest that Akt kinase activity and Rac1/CDC42 phosphorylation are positively correlated in the human melanoma cell line.
It is known that the GTP-bound form of Rac1 is active and regulates its downstream biological function. Recently, several reports revealed the inhibition of JNK/SAPK kinase by wortmannin, indicating that PI3K also regulates JNK/SAPK kinase (43,44). It was thus necessary to determine how the phosphorylation of Rac1/CDC42 by Akt kinase would affect its downstream JNK/SAPK kinase. The JNK/SAPK activity was measured with the nonradioactive JNK/SAPK assay kit after modulating Akt kinase with wortmannin or LY294002 (20 -23). Rac1/CDC42 phosphorylation and its GTPase activity changes were also measured in the same samples. As shown in Fig. 5B, JNK/SAPK was activated by wortmannin and LY294002 in a dose-dependent manner. These results suggest that JNK/ SAPK activity is negatively correlated with Akt kinase activity and Rac1 phosphorylation. Therefore, the inhibition of Akt kinase activity and Rac1 phosphorylation by wortmannin and LY294002 seems to enhance Rac1/CDC42-GTP binding, which in turn up-regulates JNK/SAPK activity. Together, our data indicate that Akt kinase phosphorylates serine 71 of Rac1 and inhibits Rac1-GTP binding, resulting in the activation of JNK/SAPK. DISCUSSION The small GTP-binding protein Rac1/CDC42 plays a pivotal role in the regulation of diverse physiological events, including reorganization of the actin cytoskeleton, cell cycle progression, and transformation (1)(2)(3)(4)(5)(6)(7)(8). However, the relationship between the PI3K and Rac1 signal pathways seems to be ambiguous. Several reports have revealed that the GTP-binding protein Rac1 mediates some of the biological effects of PI3K and have led to the suggestion that Rac1 may be a common mediator of a variety of responses mediated by PI3K (13)(14)(15)(16). In contrast, most reports revealed that Rac1 and RhoA are downstream targets of PI3K (2,3,(17)(18)(19). Thus, the ambiguous relationship between the PI3K and Rac1 signal pathways still remains to be demonstrated.
To gain more insight into the relationship between two particular signal pathways, we intended to determine whether Akt kinase phosphorylates Rho family GTPases (including RhoA, RhoB, RhoC, RhoG, and Rac1/CDC42) directly, instead of interacting with the other kinases. Upon inspecting Rac1/CDC42 family protein amino acid sequences, we identified the specific sequence ( 64 ydRIRplSYp 73 ) that is possibly phosphorylated by Akt kinase in all known Rac1/CDC42 family proteins (Table I) (26 -30). Thus, we performed the Akt kinase assay with recombinant Rac1 protein and the Rac1 peptide ( 64 ydRIRplSYp 73 ) and demonstrated that these are phosphorylated by Akt kinase (Fig. 1). To characterize the major effect of phosphorylation on Rac1 by Akt kinase, we demonstrated that Rac1-GTP binding was inhibited by Akt kinase phosphorylation (Fig. 3) without a change in GTPase activity (Fig. 2). Moreover, we confirmed these observations with the Rac1 S71A mutant, which is not phosphorylated by Akt kinase. The mutant GTPase or GTP binding was abolished regardless of Akt kinase phosphorylation (Figs. 2 and 3). Therefore, our data demonstrate that Akt kinase mediates/modulates Rac1-GTP binding through phosphorylation of Rac1 serine 71. We assume that the inhibition of Rac1-GTP binding by Akt kinase phosphorylation is partially due to the reduction of its affinity for GTP, resulting from the electrostatic charge increase in Rac1 after Akt kinase phospho-rylation. However, it still remains unclear why Akt kinase phosphorylation inhibited Rac1-GTP binding (Fig. 3) without a change in GTPase activity (Fig. 2).
Because active Rac1 (the GTP-bound form) controls its downstream JNK/SAPK kinase activity, we performed the JNK/ SAPK kinase activity assay to determine whether the change in Akt kinase activity with wortmannin or LY294002 also modulates JNK/SAPK kinase activity. Consistent with other observations (44 -47), we observed that wortmannin or LY294002 activates JNK/SAPK kinase activity (Fig. 5B). Thus, our results demonstrate that Akt kinase modulates Rac1-GTP binding through phosphorylation and, in turn, controls JNK/ SAPK kinase activity, indicating that these two signal pathways cross-talk through phosphorylation. Moreover, our data support that Akt kinase is upstream of Rac1. However, we cannot rule out the other possibility that Akt kinase is downstream of Rac1 and that the phosphorylation of Akt kinase is one of the feedback regulation mechanisms.
Akt kinase is involved in signaling for a multitude of important cellular events. The activation or inactivation of Akt kinase in the cell is one of the critical regulatory points to deliver either a survival or an apoptotic signal. Thus, the upstream signal transduction pathway through which Akt kinase is regulated in the cell is an intensively interesting research area (20 -25, 36 -38). To understand how Akt kinase function in the PI3K pathway contributes to cell proliferation/survival, the identification of substrate protein that is phosphorylated by Akt kinase and the characterization of how Akt kinase phosphorylation modulates the protein function (either activation or inhibition) seem to be also important.
Until now, several important regulatory proteins, including BAD, 6-phosphofructo-2-kinase, glycogen synthase kinase-3, IRS-1, caspase-9, and human telomerase reverse transcriptase, have been characterized as the substrates of Akt kinase (Table I). The Akt kinase substrate consensus sequence was proposed from these protein sequences. The serine/threonine residue to be phosphorylated in these proteins is within the consensus amino acid stretches in which the arginine residues at positions Ϫ5 and Ϫ3 are conserved and in which position ϩ2 relative to those of serine/threonine residues is a hydrophobic amino acid (xxRxRxx(S/T)xx). With the Akt kinase substrate consensus sequence (xxRxRxx(S/T)xx), we identified the possible Akt kinase phosphorylation sequences in Dbl ( 619 khRiRedSYi 628 ), Ras-guanine nucleotide release factor ( 738 pvRaRklSLt 747 ), p115 Rho-GEF ( 772 kpRpRpsSTr 781 ), limulus clotting factor C, and CDC24 ( 121 tiReRpsSAi 130 and 181 nmRnRtlSVe 190 ) (34, 35, 48 -52). The function of these regulatory factors was reported to control the GTP/GDP exchange of Rac1 GTPase. The Dbl-like GEF Lbc oncoprotein specifically activates the small GTP-binding protein Rac1 in mammalian fibroblasts to induce transformation and actin stress fiber formation (51). Another Dbl-related molecule, CDC24, stimulates guanine nucleotide exchange of the GTPase CDC42 family to elicit effects on both gene induction and actin-based cytoskeleton change in yeast (Saccharomyces cerevisiae) (52). p115 Rho-GEF , the coupling factor between G protein and Rho GTPase, also stimulates guanine nucleotide exchange of the GTPase family (34,35,51). Thus, the proteins that contain Akt kinase phosphorylation sites are supposed to regulate/modulate Rac1 family protein function. The finding that Akt kinase phosphorylation sequences are present in these proteins indicates that Akt kinase also modulates the Rac1 signal pathway indirectly with the phosphorylation of Rac1 GTPase regulatory proteins, together with direct Rac1 phosphorylation (Figs. 1-3). However, it remains to be demonstrated that these Rac1 GTPase regulatory proteins are phosphorylated by Akt kinase and that the phosphorylation of Rac1  Table I) and Rac1 ( 64 ydRIRplSYp 73 ) peptides (see Table I) were used to monitor Akt kinase activity and Rac1 phosphorylation, respectively. Akt kinase activity (A, upper panel) and Rac1 peptide phosphorylation (lower panel) were down-regulated, whereas JNK/SAPK activity was up-regulated (B) by treatment with wortmannin or LY294002 in a dose-dependent manner. The negative control was a heat-treated (65°C, 10 min) SK-MEL28 cell lysate. regulatory protein by Akt kinase also affects Rac1 GTPase or GTP-binding activity.
Recently, Akt kinase family proteins (Akt1, Akt2, and Akt3) have been characterized (39 -41). Because of their amino acid sequence homology, each Akt kinase-specific function seems to be redundant. It may be that Akt2 and Akt3 are also kinases responsible for Rac1 phosphorylation. However, it remains to be determined that Akt2 and Akt3 also have the same substrate specificity and which Akt kinase has more preference for a certain kind of Rho family proteins.
Since Akt kinase substrate specificity (xxRxRxx(S/T)xx) is conserved in the other proteins as an Akt kinase substrate, the inspection of these consensus amino acid sequences may help to determine whether these proteins are possible Akt kinase substrates (Table I). With the inspection of the Akt kinase consensus sequence, we noticed two sites ( 159 EPRSRHLSVS 168 and 330 DPRGRLRSAD 339 ) in human MEKK3, which mediates the p38/RK signal transduction pathway (53), and one site ( 73 IER-LRTHSIE 82 ) in human JNK-activating kinase (11). Even though the relationship between Akt kinase and MEKK3 or JNK kinase is presently unknown, Akt kinase phosphorylation sites in these proteins may contribute the signal cross-talk between two different signal pathways to protect the cell from apoptosis or to promote cell proliferation. The presence of Akt kinase phosphorylation sites in both human JNK-activating kinase and Rac1 may also provide a clue to explain why the drugs (wortmannin and LY294002) modulating PI3K also affect JNK/SAPK activity (Fig. 5B) (44 -47). We are investigating to determine whether the phosphatidylinositol 3-kinase/Akt pathway antagonizes/modulates the p38 mitogen-activated protein kinase or JNK-activating kinase pathway through Akt kinase phosphorylation.
In summary, upon identifying Akt kinase substrate protein and characterizing the functional role of Akt kinase phosphorylation, we seem to have discovered significant clues to understand how Akt kinase functions to protect cell apoptosis or to promote cell proliferation. The phosphorylation site of Akt kinase seems to be highly conserved as an Akt kinase substrate consensus sequence (xxRxRxx(S/T)xx). Thus, the inspection of protein sequences with this consensus sequence may help to identify Akt kinase substrate proteins. With this information, we have demonstrated that Akt kinase inhibits Rac1-GTP binding, which plays an important role in controlling cell shape and morphology by the phosphorylation of serine 71 of Rac1 as one of the Akt kinase target proteins.