Insulin increases the association of Akt-2 with Glut4-containing vesicles.

Expression of a constitutively active, membrane-associated Akt-1 (PKB alpha) construct in 3T3L1 adipocytes was shown to induce glucose uptake in the absence of insulin by stimulating Glut4 translocation to the plasma membrane (Kohn, A. D., Summers, S. A., Birnbaum, M. J., and Roth, R. A. (1996) J. Biol. Chem. 271, 31372-31378). However, in rat fat cell the vast majority of Akt-1 is cytosolic and shows no re-distribution to the plasma membrane in response to insulin. On the other hand, little work has been done with other Akt family members such as Akt-2 (PKB beta) or Akt-3 (PKB gamma). In this report, an analysis of the subcellular distribution of Akt-2 in rat adipocytes shows that Akt-2 is present in significant amounts in various membrane compartments, as well as in the cytosol, and the former include the light microsomes where Glut4 is present in the basal state. The distribution of Akt-2 in resting adipocytes was found to substantially overlap with that of Glut4 when light microsomes were subfractionated by a sucrose velocity gradient indicating possible co-localization. We confirmed co-localization of Akt-2 and Glut4 in the basal state by immunopurification of Glut4 vesicles, which exhibited a 5.5-fold increase in Akt-2 in response to insulin relative to the amount of Glut4. These results are consistent with the possibility that Akt-2 may be involved in Glut4 vesicle translocation.

Insulin promotes glucose uptake in adipose tissue and skeletal muscle by increasing the translocation of the Glut4 glucose transporter from an intracellular pool to the plasma membrane (1-3), but the complete signaling pathway responsible for Glut4 translocation remains to be defined. Considerable evidence supports the notion that this event is dependent on insulin-stimulated phosphatidylinositol 3-kinase activity (PI 3-kinase) 1 (4 -8). While the biochemical mechanism by which PI 3-kinase participates in glucose transporter trafficking remains uncertain, there is evidence showing that the intracellular Glut4-containing vesicles undergo an insulin-dependent increase in PI 3-kinase activity (9), suggesting that this might be an important site of lipid kinase action.
The serine/threonine kinase Akt is involved in different cellular processes such as cell growth, cell differentiation, apoptosis, and regulation of glycogen metabolism by insulin (10,(22)(23)(24). In this regard, expression of a constitutively active membrane-bound form of Akt-1 in 3T3L1 adipocytes resulted in stimulated glucose transport and Glut4 translocation to the plasma membrane (25,26). Most of the studies on Akt/PKB have focused on the ␣ isoform, and thus far little is known about the other isoforms. Here, we report the subcellular distribution and regulation of Akt-2 by insulin in rat adipocytes.

EXPERIMENTAL PROCEDURES
Materials-Collagenase and bovine serum albumin were purchased from Boehringer Mannheim. The bicinchoninic acid (BCA) protein determination kit was from Pierce. Aprotinin, leupeptin, and pepstatin A were obtained from American Bioanalytical (Natick, MA). Electrophoresis chemicals were purchased from National Diagnostics (Atlanta, GA).
Adipocyte Incubation with Wortmannin-Isolated adipose cells were incubated with 100 nM wortmannin (Sigma) for 25 min prior to insulin stimulation.
Immunoadsorption of Glut4 Vesicles-Purified anti-Glut4 monoclonal antibody 1F8 (29) as well as nonspecific IgG were coupled to acrylic beads (Reacti-gel GF 2000, Pierce) at a concentration of 0.5 mg of antibody/ml of resin according to the manufacturer's instructions. Before use, the beads were saturated with 1% BSA in PBS for at least 30 min at room temperature and then washed in PBS. Light micro-somes (100 g) were incubated with the beads (20 l) overnight at 4°C. The beads were spun down and washed three times with PBS. The adsorbed material was eluted with 1% Triton X-100 in PBS for 1 h at 4°C and then washed two times with PBS. Finally, SDS-containing Laemmli buffer without ␤-mercaptoethanol (30) was added for 30 min at room temperature. Eluates were analyzed by immunoblotting (see below).
Fractionation of Microsomes in a Sucrose Velocity Gradient-Light microsomes were resuspended in HES buffer, pH 7.4, including protease and phosphatase inhibitors, and an equal amount of protein from each treatment was loaded onto the continuous sucrose gradient, ranging from 10 to 30% (w/v). Membranes were centrifuged at 100,000 ϫ g for 1 h at 4°C. Fractions approximately of 200 l in volume were collected from the gradient. Protein content was measured, and the position of Glut4 was determined by Western blotting.
Western Blot (Immunoblot) Analysis-Samples were separated on SDS-8 and 10% polyacrylamide gels and then electrophoretically transferred to nitrocellulose (0.2 m, Schleicher & Schuell) at 4°C from 1 to 4 h. The membranes were incubated in phosphate-buffered saline containing 0.05% Tween 20, pH 7.4 (PBST) containing 5% non-fat dry milk for 1 h at 25°C to block nonspecific binding, then probed with a polyclonal anti-Akt-1 antibody (Upstate Biotechnology Inc., Lake Placid, NY), or an affinity-purified polyclonal anti-Akt-2 antibody (prepared against the peptide CDQTHFPQFSYSASIRE fused to keyhole limpet hemocyanin via the amino-terminal Cys) or an anti-Glut4 antibody (29). To detect the antigen-bound antibody, the blots were treated with secondary antibody conjugated with horseradish peroxidase. Immunoreactivity was detected using the enhanced chemiluminescence (ECL) detection system (Amersham). Quantitative analysis was performed using a scanning densitometer (Molecular Dynamics, Sunnyvale, CA). The value of control (basal) Akt-2 or Glut4 was set to 100% in all cases, and the percent increase or decrease Ϯ the standard error of the mean (S.E.) was calculated. The statistical significance of differences between values was performed by using the Student's paired t test.

RESULTS
We first studied the subcellular distribution of Akt-2 in freshly isolated rat adipocytes and compared it with the previously reported localization of Akt-1. Adipocytes were fractionated into cytosol and different membrane fractions, as described under "Experimental Procedures." Equal amounts of protein from each fraction were separated by SDS-PAGE and immunoblotted with specific antibodies for Akt-1, Akt-2, or Glut4. As shown in Fig. 1, top panel, Akt-2 was found in all subcellular fractions. The highest Akt-2 signal was found in heavy microsomes, followed by cytosol, plasma membrane, light microsomes, and mitochondria/nuclei. However, the figure was normalized for equal protein, and when one takes into account that the protein yield of the cytosol is 5-10-fold higher than the membrane fractions, the cytosol contains about half of the total Akt-2, and, except for the mitochondria/nuclei fraction, there are roughly equal amounts in the other three membrane fractions. As expected, Akt-1 was detected exclusively in the cytosolic fraction, and in the absence of insulin, Glut4 was localized only in light and heavy microsomes (Fig. 1, middle  and bottom panels, respectively).
The stimulation of Akt-1 by insulin has been reported in different cell types (16,22,31,32) and is accompanied by a change in the electrophoretic mobility of the protein (Fig. 1,  middle panel, and Refs. 31 and 32). This shift in response to the hormone indicates multiple phosphorylated states of Akt-1 (17,31,33). Stimulation of adipocytes with insulin also resulted in activation of cytosolic Akt-2, as indicated by the appearance of a slower migrating form. The Akt-2 content of light microsomes isolated from insulin-treated adipocytes was increased to 145 Ϯ 10% over the untreated control (mean Ϯ S.E., n ϭ 3; *, p Ͻ 0.01), but cellular insulin exposure had a nonobservable effect on Akt2 levels in the plasma membrane, heavy microsomes, or mitochondria/nuclei (Fig. 1, top panel). The source of Akt-2 causing the increased light microsome content is presumably the cytosol, but as noted above, the cytosol contains over half of the Akt, and one cannot detect the small decrease in this fraction that would give rise to the clearly detectable increase in the light microsome fraction. Consistent with many previous reports (1, 29, 34 -39), Glut4 demonstrated insulin-responsive translocation from the light microsomes to the plasma membrane (Fig. 1, bottom panel).
Akt-1 activation by insulin is sensitive to wortmannin, indicating that this protein is downstream of PI 3-kinase. As shown in Fig. 1, top panel, preincubation with wortmannin abolished the insulin-dependent mobility shift of Akt-2 in the cytosolic fraction. The increase in Akt-2 content induced by insulin in light microsomes was also prevented by wortmannin. As expected, the insulin-stimulated Glut4 translocation was abrogated by the PI 3-kinase inhibitor wortmannin (Fig. 1, bottom  panel).
Considering previous reports about the role of Akt/PKB in the translocation of Glut4 and our data on Akt-2 localization in light microsomes, we investigated if Akt-2 is present in vesicles with the same density as those containing Glut4. Light microsomes were fractionated by a sucrose velocity gradient, proteins were separated by SDS-PAGE, and Western blotting was performed with anti-Glut4 or -Akt-2 antibodies. Results presented in Fig. 2 show that Glut4-containing vesicles sediment faster than the majority of total light microsomes. Akt-2 was present in fractions with the highest levels of Glut4. This result indicates that either Akt-2 is present in vesicles with the same density but different from those that contain Glut4 or that both Akt-2 and Glut4 are present in the same vesicles.
To directly test if Akt-2 is associated with Glut4-containing vesicles, we immunoisolated these vesicles from the light microsomes with an anti-Glut4 antibody (1F8 antibody) covalently coupled to acrylic beads. Vesicle proteins were extracted from the immunoadsorbed Glut4-containing vesicles with 1% Triton X-100, followed by elution of Glut4 with SDScontaining Laemmli buffer (29). As shown in Fig. 3, Akt-2 was clearly detected in Glut4-containing vesicles. Insulin induced a wortmannin-sensitive, 360 Ϯ 62% increase (mean Ϯ S.E., n ϭ 4; *, p Ͻ 0.005) in Akt-2 levels associated with Glut4-containing vesicles, as measured by densitometric scanning of the protein bands. As expected, the amount of Glut4 adsorbed on 1F8 beads decreases by 36% after insulin administration due to the translocation of Glut4-containing vesicles from intracellular microsomes to the plasma membrane, and when one takes this into account, the ratio of Akt-2 to Glut4 is increased 5.5-fold by insulin. This result is consistent with the stimulatory effect of

Akt-2 in Glut4
Vesicles 7202 insulin on the Akt-2 content of the light microsomal fraction as shown in Fig. 1, top panel.

DISCUSSION
To date, most studies with Akt have focused on Akt-1, the first characterized member of the family. The other close relatives Akt-2 and Akt-3 have been poorly studied. In this report, we examined the subcellular distribution and regulation of Akt-2 by insulin in isolated rat adipocytes. We found that Akt-2 was present in cytosol and all membrane fractions, which markedly contrasts with the distribution of Akt-1, which is overwhelmingly cytosolic. The differences in the pattern of subcellular distribution between Akt-1 and Akt-2 can derive from structural differences between these proteins. A sequence comparison between Akt-1 and Akt-2 reveals that the last portion of the amino-terminal domain (amino acids 107-147) is the most divergent between Akt molecules (13, 40). Datta et al. (41) showed that Akt forms multimeric complexes through the interaction between AH domains (Akt homology domain, spans amino acids 1-147). This interaction is highly specific, as determined by the failure of the Akt-1 domain to bind Akt-2 in COS-1 cells. The differences in subcellular distribution between these isoforms suggest that Akt-1 and Akt-2 may interact with different cellular targets and/or can be differentially regulated. Also, it has been shown that the expression of Akt-1 is much higher than that of Akt-2 in brain and lung, whereas Akt-2 is highly expressed in testis while Akt-1 is very low in this tissue (15). Distinct expression of the Akt/PKB isoforms may reflect different function of these proteins. Meier et al. (42) demonstrated that Akt-2 is translocated into the nucleus by stimulation of HEK-293 and REF-52 cells with insulin-like growth factor I and fetal calf serum, respectively. However, we were unable to detect any nuclear translocation of Akt-2 after stimulation of fat cells with insulin ( Fig. 1, top panel). The explanation for this difference is unclear, but it might be due to the fact that fat cells are terminally differentiated, whereas HEK-293 and REF-52 cells are proliferating cell lines.
In this study, we demonstrated that insulin activated both Akt-1 and Akt-2 in the cytosolic fraction. Insulin activation induced the shift to a higher molecular weight of both Akt proteins. Previous reports have shown that platelet-derived growth factor stimulated both Akt-1 and Akt-2 isoforms (15). These data are consistent with the proposal that both isoforms are regulated in a similar way but that different subcellular localization may result in distinct cellular responses.
We showed that Akt-2 and Glut4 have a similar distribution in vesicles fractionated by a sucrose velocity gradient and that Akt-2 is present in Glut4-containing vesicles. The localization of Akt-2 in these intracellular membranes may be involved in regulating Glut4 trafficking in adipocytes. These results are consistent with previous reports in which Akt-1 has been implicated in Glut4 translocation as well as glucose uptake (25,26).
The mechanism by which insulin induce Glut4 translocation is not fully understood. Several signaling molecules have been implicated in this process, including PI 3-kinase. Inhibitors of PI 3-kinase have been observed to block the ability of insulin to stimulate glucose transport and Glut4 translocation in rat primary adipocytes and 3T3-L1 differentiated adipocytes (4 -8, 43). Also, the expression of a constitutively active PI 3-kinase induced Glut4 translocation in the absence of insulin (44 -47). The presence of PI 3-kinase (9) and Akt-2 in rat adipocyte Glut4-containing vesicles might be important in the process of Glut4 translocation; however, much additional work will be required to determine a functional role for Akt-2 and PI 3-kinase in Glut4-containing vesicles. Data showing a Glut4 vesicle-associated kinase activity that could be Akt-2 is promising in this regard. 2 FIG. 2. Akt-2 and Glut4 co-localize in vesicles with the same density. Light microsomes (0.5 mg of protein) from insulin-treated (10 nm for 10 min) rat adipocytes were centrifuged in a 10 -30% sucrose gradient (4.6 ml) as described under "Experimental Procedures." Fractions collected from the gradient were subjected to SDS-PAGE and immunoblotted with antibodies to Akt-2 or Glut4 (upper panel). Both proteins were detected with horseradish peroxidase-conjugated secondary antibody and a chemiluminescence detection kit. Lower panel shows distribution of Akt-2 and Glut4. The data shown are representative of three separate experiments.
FIG. 3. Akt-2 is present in Glut4-containing vesicles. Adipose cells were treated or not with 10 nM insulin for 10 min. Wortmannin (100 nM) was added 25 min before stimulation. Cell fractionation was performed as described under "Experimental Procedures." Light microsomes (100 g of protein) were immunoadsorbed with anti-Glut4 antibody (1F8) covalently attached to acrylic beads or with beads linked to nonspecific antibody (IgG). Immunoadsorbed Glut4 vesicles were eluted sequentially with 1% Triton X-100 and SDS-containing Laemmli buffer. Eluates were subjected to SDS-PAGE and immunoblotted with antibodies to Akt-2 and Glut4. The data shown are representative of four separate experiments.