Fluorescent indicators for Akt/protein kinase B and dynamics of Akt activity visualized in living cells.

Akt/protein kinase B (PKB) is a serine/threonine kinase that regulates a variety of cellular responses. To provide information on the spatial and temporal dynamics of Akt/PKB activity, we have developed genetically encoded fluorescent indicators for Akt/PKB. The indicators contain two green fluorescent protein mutants, an Akt/PKB substrate domain, flexible linker sequence, and phosphorylation recognition domain. A phosphorylation of the substrate domain in the indicators caused change in the emission ratio based on fluorescent resonance energy transfer between the two green fluorescent protein mutants. To let the fluorescent indicators behave as endothelial nitric-oxide synthase and Bad, which are endogenous Akt/PKB substrates, they were fused with the Golgi target domain and mitochondria target domain, respectively. The indicators thus colocalized with the endogenous substrates conferred their susceptibilities to phosphorylation by Akt/PKB. We showed that the Golgi-localized indicator responded to the stimulation with 17beta-estradiol (E2) and insulin in endothelial cells. In addition, E2 elicited the phosphorylation of the mitochondria-localized indicator in the endothelial cells, but no phosphorylation was observed by E2 or by insulin of the diffusible indicator that has no targeting domain. The difference in the results with the three indicators suggests that the activated Akt/PKB is localized to subcellular compartments, including the Golgi apparatus and/or mitochondria, rather than diffusing in the cytosol, thereby efficiently phosphorylating its substrate proteins. E2 triggered the phosphorylation of the mitochondria-localized indicator, whereas insulin did not induce this phosphorylation, which suggests that the localization of the activated Akt/PKB to the mitochondria is directed differently between insulin and E2 via distinct mechanisms.

The Akt/protein kinase B was identified as a serine/threonine protein kinase with high homology with the protein kinases A and C. At the same time, this kinase was identified as the cellular homologue of the viral oncoprotein v-Akt. Mammals have three closely related Akt isoforms, encoding Akt1, Akt2, and Akt3. All the three isoforms contain an N-terminal Pleckstrin homology domain, followed by a kinase domain and a C-terminal regulatory tail. Akt is an important regulator of various cellular processes including glucose metabolism, cell survival, and angiogenesis (1)(2)(3). Akt becomes activated by a wide variety of stimuli, including growth factors, cytokines, steroid hormones, and cellular stresses. Recent studies on the mechanism of Akt activation indicate that the phosphoinositide 3-kinase and its product phosphoinositide-3,4,5-triphosphate promote translocation of Akt to the plasma membrane and the phosphorylation at the two sites, Thr-308 and Ser-473, in Akt protein. The phosphorylation of the both sites is blocked by pretreatment of the cells with a phosphoinositide 3-kinase inhibitor, wortmannin. The activated Akt then phosphorylates substrates including glycogen synthase kinase-3, Bad, eNOS, 1 caspase-9, and forkhead transcription factors (4 -8). To achieve signaling specificity when the Akt signaling pathway is activated by stimuli that elicit different cellular responses, the exact location of the substrates in living cells may dictate which pathways are activated, but the precise mechanism remains poorly understood because of the lack of sufficient methods to study it.
Recently, we and other groups have reported genetically encoded fluorescent indicators for detecting protein phosphorylation in a single living cell (9 -12). The approaches for detecting protein phosphorylation based on the fluorescent indicators have provided new information on the spatial and temporal regulation of activities of protein kinases and phosphatases.
Herein, we describe genetically encoded fluorescent indicators for Akt/PKB kinase activity, named Aktus (a fluorescent indicator for Akt phosphorylation that can be custom-made), eNOS-Aktus, and Bad-Aktus. eNOS-Aktus and Bad-Aktus are indicators for phosphorylation of eNOS and Bad, respectively. In addition, Aktus is a diffusible cytosolic indicator for phosphorylation by Akt. The common unit of the indicators is Aktus and is based on our general approach for visualizing protein phosphorylation in living cells (9), which contains two GFP mutants, an Akt substrate domain, a flexible linker sequence, and a phosphorylation recognition domain. We demonstrate that a FRET change between the two GFP mutants in the Aktus is induced by the phosphorylation of its Akt substrate domain. Almost all Akt substrates are known to localize to subcellular regions. eNOS is localized to the Golgi apparatus and cholesterol-rich microdomain of plasma membrane, caveolae in cells (13,14), whereas Bad is present in mitochondria outer membrane (15,16). By fusing the Aktus with the respectively subcellular localization domains for eNOS and Bad, the * This work was supported by the Core Research for Evolutional Science and Technology of Japan Science and Technology and grants from the Ministry of Education, Science and Culture of Japan (to Y. U.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed. Tel.: 81-3-5841-4351; Fax: 81-3-5841-8349; E-mail: umezawa@chem.s.u-tokyo.ac.jp. eNOS-Aktus and Bad-Aktus were prepared, respectively, and were compared with the diffusible indicator, Aktus. The colocalized indicators were found to exhibit the extent of phosphorylation of eNOS and Bad, respectively, which was identical to those of the endogenous substrates, whereas Aktus induced no phosphorylation. These results suggest that the activated Akt is localized. The indicators thereby allowed detecting protein phosphorylation by Akt with high spatial and temporal resolution in single living cells.
Plasmid Construction-To construct the genes of the three indicators, fragment cDNAs of two GFP mutants (enhanced CFP and enhanced YFP), a linker and a substrate domain, a phosphoserine recognition domain, an eNOS targeting domain, or a mitochondrial targeting domain fused to the N-terminal of the CFP were generated by PCR and cloned into pBlueScript. All PCR fragments were sequenced with an ABI310 genetic analyzer (Applied Biosystems). The amino acid sequence containing the flexible linker and substrate domain is GGSS-GGSSRGRSRSAP, which is a serine phosphorylation domain recognized by Akt derived from Bad (17). The phosphoserine recognition domain is bovine 14-3-3 protein (amino acid residues 82-235). The amino acid sequence of a targeting domain of eNOS is human eNOS (amino acids 1-35) and a spacer of 10 glycines (18). The amino acid sequence of a mitochondrial targeting domain is Tom20 (amino acids 1-33) (19). All constructs were subcloned into pcDNA3.1 behind a Kozak sequence using HindIII and XhoI sites.
Cell Culture and Transfection-CHO-IR-Akt cells overexpressing both human insulin receptor (IR) and Akt1 were cultured in Ham's F-12 medium supplemented with 10% fetal calf serum, 1% penicillin/streptomycin, 1 mM sodium pyruvate, and 0.1 mM nonessential amino acids at 37°C in 5% CO 2 . The CHO-IR-Akt cells were plated 8 h after transfection onto glass-bottomed dishes or plastic culture dishes for fluorescent imaging of living cells or Western blotting analysis, respectively. Calf pulmonary artery endothelial (CPAE) cells were cultured in Eagle's minimal essential medium supplemented with 20% fetal calf serum at 37°C in 5% CO 2 . Cells were transfected with LipofectAMINE 2000 reagent. The CPAE cells were plated 8 h after transfection onto glass-bottomed dishes and glass coverslips for fluorescence imaging and immunofluorescence, respectively.
Imaging of Cells-CHO-IR-Akt cells were starved in 0.2% bovine serum albumin for 4 -6 h and washed twice with Hanks' balanced salt solution. After transfection, CPAE cells were maintained in the culture medium for 12-24 h, infected with adenoviruses containing Akt1, estrogen receptor (ER), or both for 1 h, and incubated for 12 h in the culture medium. After starvation in a steroid-free medium (phenol red-free Eagle's minimal essential minimum with 2% charcoal-treated fetal calf serum) for 12 h, the cells were washed twice with Hanks' balanced salt solution. The cells were imaged at room temperature on a Carl Zeiss Axiovert 135 microscope with a cooled charge-coupled device camera MicroMaX (Roper Scientific Inc., Tucson, AZ), controlled by MetaFluor (Universal Imaging, West Chester, PA). The exposure time at 440 Ϯ 10 nm excitation was 100 ms. The fluorescent images were obtained through filters at 480 Ϯ 15 nm and 535 Ϯ 12.5 nm with a 40ϫ oil immersion objective (Carl Zeiss, Jena, Germany). The cells were stimulated with 100 nM insulin and 1 M E2.
Western Blotting Analysis-CHO-IR-Akt cells and CPAE cells were starved in 0.2% bovine serum albumin for 4 -6 h and were then stimulated with 100 nM insulin for 15 min at 25°C. After starvation in the steroid-free medium for 12 h, the CPAE cells infected with adenoviruses containing Akt and ER stimulated with 1 M E2 for 20 min at 25°C. The cells were lysed with an ice-cold lysis buffer (50 mM Tris-HCl, pH 7.4, 100 mM NaCl, 1 mM EDTA, 0.1% Triton X-100, 10 mM NaF, 2 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml pepstatin, 10 g/ml leupeptin, and 10 g/ml aprotinin). After immunoprecipitation with the antibody for 1 h at 4°C, the immunocomplex was precipitated using Protein G-Sepharose 4FF beads (Amersham Life Sciences). The sample was separated by SDS-polyacrylamide gel electrophoresis and electrophoretically transferred onto a polyvinylidene fluoride membrane. The obtained signal was quantified using an image analyzer (LAS-1000plus; Fujifilm). Immunofluorescence Microscopy-CPAE cells expressing eNOS-Aktus were fixed with 2% paraformaldehyde and were permeabilized with PBS containing 0.2% Triton X-100 for 10 min. After 1 h of incubation with polyclonal rabbit anti-eNOS (C-20) antibody, the cells were washed with phosphate-buffered saline containing 0.2% fish skin gelatin and incubated with anti-rabbit IgG antibody labeled with Cy5 for 30 min. The coverslips were mounted onto a slide and observed under a confocal laser-scanning microscope (LSM 510; Carl Zeiss).

RESULTS AND DISCUSSION
Design of Aktus, eNOS-Aktus, and Bad-Aktus-The common unit, Aktus, of the present indicators is a five-part tandem fusion protein containing enhanced CFP, a substrate domain, flexible linker sequence, phosphorylation recognition domain, and enhanced YFP. The substrates of Akt including Bad and eNOS contain a common Akt substrate sequence RxRxxS/Ty, where x is any amino acid and y is a bulky hydrophobic residue (20,21). The phosphorylated serine contained in this sequence binds to 14-3-3 protein (22)(23)(24). We chose the RGRSRSAP sequence derived from Bad as the common substrate domain and the 14-3-3 protein as the common phosphorylation recognition domain for Aktus, eNOS-Aktus, and Bad-Aktus. The constructs of the indicators prepared are shown in Fig. 1. The 14-3-3 protein exists naturally as a dimer (25,26), which is mediated by the amino-terminal sequence (27). The dimerization domain is deleted to avoid intermolecular FRET between the indicators.
Protein kinases, phosphatases, and their substrates are often discretely localized in cells, which is essential for efficiency, specificity, and duration of phosphorylation. For example, acylation reactions via myristoylation or palmitoylation of proteins such as Src and eNOS are required for correct cellular trafficking and for coupling intracellular signaling to extracellular stimuli. eNOS is unique among NOS family members, which is localized to specific intracellular domains, including the Golgi and caveolae in cells (13,14). Liu et al. demonstrated that the first 35 amino acids of eNOS fused with GFP were sufficient to target the GFP to the Golgi apparatus and caveolae (18). A phosphorylation of eNOS at Ser-1179 by Akt is a critical control step for NO production in endothelial cells (28), and the localization of eNOS is necessary for its activation by Akt (13,29). By fusing the Aktus with eNOS targeting domain, eNOS-Aktus is formed (Fig. 1), which is capable of detecting the phosphorylation of eNOS.
Bad protein is localized to mitochondrial outer membrane (15,16). To localize the Aktus to the mitochondrial outer membrane, Aktus was fused with the mitochondrial targeting domain (amino acids 1-33) derived from Tom20, which binds to the mitochondria outer membrane through the N-terminal ␣-helical hydrophobic transmembrane domain (19). Bad-Aktus is thereby formed (Fig. 1).
Characterization of Aktus-Akt is known to be activated by insulin through a phosphoinositide 3-kinase-dependent pathway (30). To examine whether Aktus was phosphorylated by Akt, we performed immunoblotting analysis using the antiphospho-Bad (Ser-136) antibody. When CHO-IR-Akt cells were stimulated with 100 nM insulin for 15 min, the phosphorylation of Aktus was induced ( Fig. 2A). Pretreatment of the cells with 10 M wortmannin completely blocked the insulin-induced increase in the phosphorylation of Aktus ( Fig. 2A). In the CHO-IR-Akt cells, Akt was phosphorylated at Ser-473 and Thr-308 upon stimulation with insulin, and this phosphorylation of Akt was completely blocked by pretreatment with 10 M wortmannin (Fig. 2B). CHO-IR-Akt cells expressing AktusS136A were prepared in which serine was replaced with alanine at the phosphorylation site within the substrate domain. When this cell was stimulated with insulin, no phosphorylation was detected ( Fig. 2A). In CHO-IR cells, in which Akt was not overexpressed, the insulin-dependent phosphorylation of Aktus was not detected (data not shown). These results indicate that the phosphorylation of Aktus is mediated by the phosphoinositide 3-kinase/Akt pathway. We next examined whether FRET in the Aktus was changed as a result of the phosphorylation of Aktus. Fig. 2C shows images of Aktus expressed in CHO-IR-Akt cells, before and at different times after stimulation with insulin, in which the 480 nm/535 nm emission ratio is represented by pseudocolor images that represent the efficiency of FRET between the GFP mutants within Aktus. When Aktus was transfected into CHO-IR-Akt cells, it was uniformly distributed in the cytosolic compartment but excluded from the nucleus despite the absence of a nuclear-export signal (Fig. 2C), as was expected for its high molecular weight protein. Fig. 2D shows a time course of the change in the cytosolic emission ratio. Upon stimulation with 100 nM insulin, a blue shift of the pseudocolor was observed because of a decrease in the emission ratio (Fig. 2C). The decrease in emission ratio was detectable within several seconds and reached a plateau within 10 min. The insulin-induced change in the cytosolic emission ratio was blocked when the cells were pretreated with 10 M wortmannin (Fig. 2, D and E). No significant change was observed in the cytosolic emission ratio in CHO-IR-Akt cells expressing AktusS136A (Fig. 2, D and E). These results demonstrate that FRET from CFP to YFP within Aktus is increased upon phosphorylation of Aktus in the cytosolic compartment, thereby allowing visualization of protein phosphorylation by Akt in single living cells.
Fluorescent Indicator for eNOS Phosphorylation-One of the major physiological roles of endothelial cells is to mediate the vasodilatory response to various agonists by NO production (31). Endothelial cells treated with insulin or estrogen have shown increased eNOS activity by Akt (4, 6, 32, 33). eNOS is located to specific intracellular domain, including the Golgi apparatus and caveolae (13,14,18). Both co-translational Nterminal myristoylation and post-translational palmitoylation (cysteines 15 and 26) of eNOS are required for the proper subcellular localization of eNOS (18). Previous work has shown that G2A eNOS is not myristoylated and does not localize to the Golgi apparatus or caveolae but is distributed throughout the cytosol (18,34). Furthermore, upon stimulation with vascular endothelial growth factor, G2A eNOS was not phosphorylated at Ser-1179 by Akt (13) and did not produce nitric oxide (29). These results indicate that the localization of eNOS is critical for its phosphorylation and activation. To obtain spatial and temporal information on the dynamics of the phosphorylation of eNOS by Akt, we prepared eNOS-Aktus by fusing Aktus with the eNOS targeting domain. This eNOS-Aktus was in fact colocalized with endogenous eNOS in CPAE cells; the cells were transfected with eNOS-Aktus and immunolabeled with anti-eNOS antibody. Confocal fluorescence images illustrated that eNOS-Aktus was found to be colocalized with endogenous eNOS in CPAE cells (Fig. 3A). eNOS-Aktus and endogenous eNOS were found to localize in the Golgi apparatus (Fig. 3A,  arrowhead), but eNOS-Aktus and endogenous eNOS in the caveolae were not found, which indicates that the majority of eNOS-Aktus and endogenous eNOS is localized in the Golgi apparatus in the CPAE cells.
To examine the response of eNOS-Aktus to insulin, CPAE cells expressed with eNOS-Aktus were observed under a conventional fluorescence microscope, and the CFP/YFP emission ratio of the eNOS-Aktus in the Golgi apparatus was measured. Insulin stimulation triggered a rapid change in the emission ratio in CPAE cells infected with adenovirus-expressing Akt (Fig. 3B). Omission of the infection of Akt (Fig. 3B) abolished this FRET response to insulin, indicating that the response of eNOS-Aktus to insulin is caused by the Akt-dependent phosphorylation of eNOS-Aktus. To examine the response of the diffusible indicator Aktus to insulin, the cytosolic region of the CPAE cells transfected with Aktus was observed. Aktus exhibited no response to insulin in the CPAE cells infected with adenovirus containing Akt (Fig. 4). The different results between eNOS-Aktus and Aktus indicate that insulin-stimulated phosphorylation by Akt occurs in the Golgi apparatus but not in the cytosol. The increase in the efficiency of phosphorylation by Akt in the Golgi apparatus indicates that the activated Akt is also localized to the Golgi apparatus.

Fluorescent Indicators for Akt/Protein Kinase B
viruses expressing ER and Akt were transfected with eNOS-Aktus. E2 stimulation caused a significant decrease in the emission ratio of eNOS-Aktus in 3-8 min in the Golgi apparatus of the CPAE cells (Fig. 3B). When CPAE cells were expressed with either ER or Akt alone and stimulated with E2, respectively, no significant response of eNOS-Aktus was obtained ( Fig. 3B), indicating that the response of eNOS-Aktus to E2 was caused by both ER-and Akt-dependent phosphorylation of eNOS-Aktus. On the other hand, Aktus under otherwise identical conditions exhibited no response to E2 (Fig. 4A). The different responses between eNOS-Aktus and Aktus suggest that Akt activated by E2 stimulation is localized to the Golgi apparatus.
To confirm that the responses of eNOS-Aktus to stimulation of insulin and of E2, respectively, reflect that eNOS in the Golgi apparatus is phosphorylated, immunoblotting analyses with anti-phospho-eNOS (Ser-1179) antibody were performed using the CPAE cells transfected with eNOS. The CPAE cells infected with Akt exhibited insulin-dependent phosphorylation of eNOS (Fig. 3C, top). However, treatment with insulin of CPAE cells that were not infected with Akt did not induce the phosphorylation of eNOS (Fig. 3C, top). Moreover, the CPAE cells infected with both ER and Akt induced E2-dependent phosphorylation of eNOS, but the phosphorylation of eNOS upon stimulation with E2 was not detected in the CPAE cells infected with either ER or Akt alone (Fig. 3D, top). The results are consistent with the responses of eNOS-Aktus, indicating that the response of eNOS-Aktus represent the phosphorylation of eNOS. In addition, to confirm that the response of the cytosolic indicator Aktus reflects the phosphorylation of GSK3␤, which is a substrate for Akt in the cytosol, immunoblotting analysis with anti-phospho-GSK3␤ antibody was performed using the CPAE cells transfected with GSK3␤. Consistent with the response of Aktus, no phosphorylation of GSK3␤ was detected in the CPAE cells infected with Akt. The phosphorylation of GSK3␤ was detected in CHO-IR-Akt cells (Fig. 4B, top), where the response of Aktus was observed (Fig. 2D). The results indicate that the response of Aktus reflects the phosphorylation of the cytosolic protein GSK3␤.
We have shown that the Akt activated by both E2 stimulation and insulin stimulation is localized to the Golgi apparatus, whereas E2 or insulin stimulation activates Akt in a different time-dependent manner. The different time dependence between stimulation with E2 and insulin is explained by the difference in the time dependence of phosphoinositide-3,4,5triphosphate production, which precedes the Akt activation, as reported previously (6). It has been discovered that E2-induced activation of Akt occurs through a mechanism independent of insulin-induced activation of Akt (6).
Fluorescent Indicator for Bad Phosphorylation-Apoptosis is fundamental in the regulation of development and control of tissue homoeostasis under conditions of cellular stress. Akt is known to exert antiapoptotic effects through several downstream targets. Among molecules central to the regulation of apoptosis in eukaryotes are members of the Bcl-2 family of proteins, including Bad, Bcl-2, and Bcl-X L . Bad is known to exert its apoptosis-promoting effects by hetrodimerizing with Bcl-2 or Bcl-X L (36). Unphosphorylated Bad is capable of forming heterodimers with Bcl-2 or Bcl-X L localized to the mitochondrial outer membrane (15,16). Datta et al. (17) showed that phosphorylation on Ser-136 of Bad by stimulation with platelet-derived growth factor occurs via the phosphoinositide 3-kinase/Akt pathway in vitro and in vivo. Phosphorylated Bad is complexed with 14-3-3 protein and no longer interacts with Bcl-2 and Bcl-X L , allowing the inhibition of apoptosis (24).
To observe the phosphorylation of Bad in endothelial cells, Bad-Aktus, which is colocalized with endogenous Bad at the mitochondria outer membrane, is prepared by fusing Aktus with the mitochondrial targeting domain. To confirm that the Bad-Aktus in CPAE cells was localized in the mitochondria, CPAE cells transfected with Bad-Aktus were stained with TMRE, a mitochondrial marker. Confocal fluorescence images illustrated that the localization of Bad-Aktus found was the same as that of the mitochondria stained with TMRE (Fig. 5A). This merged image indicates that Bad-Aktus was localized in the mitochondria.
Estrogen is known to inhibit apoptosis of endothelial cells (37)(38)(39), but its mechanism of signal transduction pathways activated by estrogen is incompletely characterized. Although it was reported recently that E2 activates Akt in endothelial cells (6,32,35), it is not known whether stimulation with E2 induces Bad phosphorylation by Akt in endothelial cells. To examine whether Bad is phosphorylated by Akt upon stimulation with E2, CPAE cells infected with adenoviruses containing Akt and ER were transfected with Bad-Aktus. E2 was found to trigger a significant decrease in the CFP/YFP emission ratio of Bad-Aktus in 3-10 min in the mitochondria in CPAE cells (Fig.  5B). When CPAE cells were expressed with either ER or Akt alone, and stimulated with E2, respectively, no significant response with Bad-Aktus was observed (data not shown). This result clearly indicates that E2 stimulation induces the phosphorylation of endogenous Bad via an Akt-and ER-dependent pathway in the endothelial cells. The present result with Bad-Aktus suggests that E2-mediated inhibition of endothelial cell apoptosis may require the phosphorylation of Bad by Akt. Aktus exhibited no response to E2 in the cytosolic region of the CPAE infected adenovirus expressing ER and Akt (Fig. 4). The different results between Bad-Aktus and Aktus described above indicate that the E2-induced phosphorylation of Bad by Akt occurs in the mitochondria, which further suggests that the E2-activated Akt is also localized to the mitochondria.
In contrast to E2 stimulation, Bad-Aktus was not responsive to insulin stimulation in CPAE cells infected with adenoviruses expressing Akt (Fig. 5B). This result indicates that insulin stimulation induces no phosphorylation of endogenous Bad by Akt in endothelial cells. The results with Bad-Aktus suggest that the activated Akt that is induced by E2 is localized to the mitochondria, but the activated Akt that is induced by insulin is not localized to the mitochondria; consequently, no phosphorylation of Bad takes place.
To confirm that the responses of Bad-Aktus reflect the phosphorylation of Bad in the mitochondria, immunoblotting analyses with anti-Bad antibody were performed using the CPAE Fluorescent Indicators for Akt/Protein Kinase B 30950 cells transfected with Bad. E2 stimulation of CPAE cells infected with both adenoviruses expressing ER and Akt caused a modification of Bad such that it migrated slower in SDS-polyacrylamide gels than Bad from unstimulated cells, indicative of E2-mediated phosphorylation of Bad (Fig. 5C). The phosphorylation of Bad upon stimulation with E2 was not detected in the CPAE cells infected with either ER or Akt alone (Fig. 5C). Insulin stimulation of CPAE cells infected with adenovirus expressing Akt was not able to induce the phosphorylation-dependent mobility shift of Bad (Fig. 5D). These results are in agreement with those of Bad-Aktus, indicating that the response of Bad-Aktus in fact represents the phosphorylation of Bad. The phosphorylation-dependent mobility shift of Bad can be regarded as the extent of the phosphorylation of Bad-Aktus. Thus, by using the electrophoretic mobility shift assay of Bad, we examined what level of decrease in the CFP/YFP emission ratio is gained as a result of phosphorylation of the substrate domain in Bad-Aktus. E2 stimulation of CPAE cells infected with both adenoviruses expressing ER and Akt induced an increase in phosphorylated Bad from 19% to 26% of the total Bad proteins (Fig. 5C). These values correspond to the emission ratio of Bad-Aktus before and after E2 stimulation, respectively (Fig. 5B). The level of decrease in the CFP/YFP emission ratio of Bad-Aktus was estimated to be ϳ6.5 ϫ 10 Ϫ3 upon extrapolating the amount of the phosphorylated Bad with the mobility shift assay to 100% of the Bad proteins.
In conclusion, we have developed fluorescent indicators for the serine/threonine kinase Akt. We showed that eNOS-Aktus responded to the stimulation with E2 and insulin at the Golgi apparatus in the CPAE cells. In addition, E2 elicited the phosphorylation of Bad-Aktus at the mitochondria in the CPAE cells. But no phosphorylation of Aktus was observed by E2 or by insulin. The difference in the results with the three indicators suggests that the activated Akt is localized to subcellular compartments, including the Golgi apparatus and/or mitochondria rather than diffusing in the cytosol, thereby efficiently phosphorylating its substrate proteins. E2 triggered the phosphorylation of Bad-Aktus, whereas insulin did not induce the phosphorylation of Bad-Aktus. The results suggest that the localization of the activated Akt is directed differently between stimuli, including insulin and E2, via distinct mechanisms. Akt can phosphorylate substrate proteins, including not only eNOS and Bad but also forkhead transcription factors and IB kinases (4 -8), both of which are localized in the nucleus. By replacing the targeting domains of the present indicators with those of forkhead transcription factors and IB kinases, the genetically encoded indicators for forkhead transcription factors and for IB kinases can also be prepared. Endogenous Akt isoforms are expressed at low levels in CPAE cells. Consequently, the endogenous Akt phosphorylated only a small portion of the expressed indicators. Overexpression of Akt was required to phosphorylate much of the remaining indicators, so that the activation of the Akt as a whole became detectable. Throughout this study, the overexpression was made with Akt1, which was the major isoform. Therefore, the conclusion was drawn basically with Akt1 that the activated Akt was localized in each particular subcellular compartment. Upon replacing Akt1 with Akt2 or Akt3 in the CPAE cells, possible distinctions if any between Akt1, -2, and -3 could be assessed. The present indicators and their applications are thus expected to contribute to the studies of a whole range of dynamics of the activated Akt in living cells.