Activity, phosphorylation state and subcellular distribution of GLUT4-targeted Akt2 in rat adipose cells

All constructs were generated in the pCIS2 mammalian expression vector (a generous gift from Dr C. Gorman). The wild type (wt), constitutively active(DE, i.e. S474D, T309E), kinase dead (KD: K179A) and triple mutant (A3: K179A,S474A, T309A) forms of human Akt2 cloned into the vector pLNCX (generous gifts from Dr Morris J. Birnbaum) were digested with BglII and BamHI followed by Klenow fill-in reaction and then NcoI digestion, resulting in the inserts of ΔPH-Akt2 with 5′-NcoI and 3′ blunt ends. The HA-GLUT4-NheI(CT)(GLUT4 containing NheI site in the C terminus) in pCIS2 was digested with NheI and HpaI to generate the vector fragment for subcloning. To fuse the NcoI end of ΔPH-Akt2 inserts to the NheI end of HA-GLUT4 vector an oligonucleotide adapter(NheI-NcoI: 5′ CTAGCGGGGCCGGGGCCGGGGC 3′ forward and 5′ CATGGCCCCGGCCCCGGCCCCG 3′ reverse) was used. TheΔPH-Akt2-wt, ΔPH-Akt2-DE, ΔPH-Akt2-KD and ΔPH-Akt2-A3 were then fused in-frame to HA-GLUT4 in pCIS2 using NheI and HpaI (blunt) sites to generate HA-GLUT4-Akt2-wt, HA-GLUT4-Akt2-DE,HA-GLUT4-Akt2-KD and HA-GLUT4-Akt2-A3. HA-GLUT4-Akt2-S474A and HA-GLUT4-Akt2-T309A mutants were generated by site-directed mutagenesis using the QuickChange kit according to manufacturer's instructions (Stratagene, La Jolla, CA), and mutations were confirmed by BigDye sequencing. Construction of the HA-tagged GLUT4 and dominant-negative K44A dynamin1 in pCIS2 has been described previously (Quon et al.,1994; Al-Hasani et al.,1998). The myc-tagged GLUT4 (myc-GLUT4) construct was generated by restriction digestion of GLUT4 via a SauI site contained within the extracellular loop between transmembrane domains one and two. For transfection experiments, the plasmids were purified in mg quantities using a maxiprep kit(Qiagen).

Preparation of isolated rat epididymal adipose cells from male rats (CD strain, Charles River Breeding Laboratories, Inc.) was performed as described previously (Quon et al.,1994). Isolated cells were washed twice with Dulbecco's modified Eagle's medium containing 25 mM glucose, 25 mM Hepes, 4 mM L-glutamine, 200 nM N6-(2-phenylisopropyl)-adenosine, and 75 μg/ml gentamycin, and resuspended to a cytocrit of 40% (5-6×10⁶ cells/ml). 200 μl of the cell suspension were added to 200 μl of Dulbecco's modified Eagle's medium containing 100 μg of carrier DNA (sheared herring sperm DNA; Boehringer Mannheim) and expression plasmids as indicated. The total concentration of plasmid DNA in each cuvette was adjusted to 1 μg/cuvette for HA-GLUT4-Akt2 fusion proteins and 4.5 μg/cuvette for dominant-negative K44A dynamin1. In experiments where myc-GLUT4 and HA-GLUT4-Akt2 fusion proteins were cotransfected the total concentration of plasmid DNA in each cuvette was adjusted to 0.2 μg/cuvette and 0.5 μg/cuvette, respectively. Electroporation was carried out in 0.4-cm gap-width cuvettes (Bio-Rad) using a T810 square wave pulse generator (BTX). After applying three pulses (12 mseconds, 200 V), the cells were washed once in Dulbecco's modified Eagle's medium, pooled in groups of 4-10 cuvettes, and cultured at 37°C, 5%CO₂ in Dulbecco's modified Eagle's medium containing 3.5% bovine serum albumin.

Rat adipose cells were harvested 20-24 hours post-transfection and washed in Krebs-Ringer bicarbonate Hepes buffer, pH 7.4, 200 nM adenosine (KRBH buffer) containing 5% bovine serum albumin. The cells from individual cuvettes were transferred into 1.5 ml microcentrifuge tubes. After stimulation with 67 nM (1×10⁴ microunits/ml) insulin for 30 min at 37°C,subcellular trafficking of GLUT4 was stopped by the addition of 2 mM KCN. All of the following steps were performed at room temperature. A monoclonal anti-HA antibody (HA.11, Berkeley Antibody Co.) or anti-myc antibody (Santa Cruz) was added at a dilution of 1:1000 or 1:10, respectively, and the cells were incubated for 1 hour. Excess antibody was removed by washing the cells three times with KRBH, 5% bovine serum albumin. Then 0.1 μCi of ¹²⁵I sheep anti-mouse antibody (Amersham Pharmacia Biotech) was added to each reaction, and the cells were incubated for 1 hour. Finally, the cells were spun through dinonylphtalate oil to remove the unbound antibody,and the cell surface-associated radioactivity was counted in aγ-counter. The resulting counts were normalized to the lipid weight of the samples (Weber et al.,1998). Unless stated otherwise, the values obtained for pCIS-transfected cells were subtracted from all other values to correct for nonspecific antibody binding. Antibody binding assays were performed in duplicate or triplicate.

Transfected cells were harvested after 20 hours of incubation. Crude total membrane fractions were prepared essentially as described previously(Chen et al., 1997). Homogenization and subcellular fractionation of adipose cells were carried out according to the method of Simpson et al.(Simpson et al., 1983). Briefly, cells were washed twice with TES buffer (25 mM Tris-HCl, 250 mM sucrose, 2 mM EDTA, pH 7.4) containing 0.12 mM 4-(2-aminoethyl)-benzenesulphonyl fluoride, 10 g/ml aprotinin, and 10 g/ml leupeptin at 18°C, and homogenized with a Potter-Elvehjem teflon pestle. Subcellular membrane fractions, plasma membranes (PM), high-density microsomes(HDM), and low-density microsomes (LDM), were obtained by differential centrifugation.

Immunoblot analysis of HA-GLUT4-Akt2 fusion proteins was performed with a monoclonal anti-HA antibody (1:1000 dilution) and a polyclonal anti-serum against Akt-phospho-Ser⁴⁷³ or Thr³⁰⁸ (1:750 dilution). The crude total membrane or fractions were separated by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose filters, and the filters were then incubated with 0.2 μCi/ml ¹²⁵I-labeled protein A. HA-GLUT4-Akt2 and Ser⁴⁷⁴- or Thr³⁰⁹-phosphorylated HA-GLUT4-Akt2 fusion proteins were detected using anti-HA antibody and anti-Akt-phospho-Ser⁴⁷⁴ or Thr³⁰⁸-specific antibodies,respectively, following the manufacturer's protocols. The signals of HA-GLUT4 and HA-GLUT4-Akt2 fusion proteins were quantified using Image Gauge V3.12,Science lab 98 (Fuji Photo Film Co., Ltd).

To examine the subcellular localization of Akt2 targeted to the GLUT4-trafficking pathway, HA-GLUT4-Akt2 fusion proteins were transfected into rat adipose cells and their cell-surface expression was detected by anti-HA antibody binding. Total expression was assessed by immunoblotting for the HA-epitope and used as a normalizing factor for cell-surface expression. As shown in Fig. 1, in the basal state (absence of insulin) cell-surface HA-GLUT4 is increased when Akt2-wt (or Akt2-DE, data not shown) is attached compared to that of HA-GLUT4 alone, and similar to that of HA-GLUT4 in the insulin-stimulated state. Insulin produces no further effect. The effect of fusing Akt2-wt to HA-GLUT4 is not blocked by wortmannin (data not shown). However, basal cell-surface HA-GLUT4 is slightly decreased with Akt2-KD (or Akt2-A3, data not shown) attached compared to HA-GLUT4 alone, although it still exhibits a normal insulin-stimulated translocation. Subcellular fractionation reveals these same relationships(data not shown).

To address whether the Thr³⁰⁹ and Ser⁴⁷⁴ residues are phosphorylated when Akt2 is targeted to the GLUT4-trafficking pathway, total cell membrane proteins were isolated from adipose cells transfected with HA-GLUT4-Akt2 fusion proteins and immunoblotted with anti-Akt1-phospho-Ser⁴⁷³ and -Thr³⁰⁸. As shown in Fig. 2,phospho-Ser⁴⁷⁴ signals for both the fusion proteins (∼95 kDa)and endogenous Akt2 (∼62 kDa) are detected. HA-GLUT4-Akt2-wt is phosphorylated on Ser⁴⁷⁴ in the basal state, with no further increase in response to insulin, and this Ser⁴⁷⁴ phosphorylation is not inhibited by wortmannin (data not shown). In HA-GLUT4-Akt2-KD, little or no Ser⁴⁷⁴ phosphorylation is detected in the basal state. Insulin treatment, however, leads to a significant increase in Ser⁴⁷⁴phosphorylation coincidentally with the marked increase in endogenous Akt2 phosphorylation and translocation of the KD fusion protein itself. Interestingly, phosphorylation at Thr³⁰⁹ is not detectable in either HA-GLUT4-Akt2-wt or HA-GLUT4-Akt2-KD, in either the basal or insulin-stimulated state, although significant levels of Thr³⁰⁹phosphorylation of endogenous Akt2 are observed in the insulin-stimulated state (Fig. 2). Subcellular fractionation experiments further demonstrate that the proportions of Ser⁴⁷⁴-phosphorylated-HA-GLUT4-Akt2-wt and -HA-GLUT4-Akt2-KD relative to their respective expression levels are significantly higher in the plasma membranes than the other fractions (data not shown).

Whether Ser⁴⁷³ phosphorylation of Akt occurs through an autophosphorylation mechanism remains to be determined. The data shown above suggest that Akt2 kinase activity might be required for Ser⁴⁷⁴phosphorylation. However, further evidence is needed to determine whether Ser⁴⁷⁴ phosphorylation of HA-GLUT4-Akt2-KD in the insulin-stimulated state is mediated by endogenous activated Akt2 itself rather than another kinase located in the plasma membrane. Thus,HA-GLUT4-Akt2-KD was forced to the cell surface by co-expression of a dominant negative dynamin (Al-Hasani et al.,1998), and cell-surface expression and Ser⁴⁷⁴phosphorylation was examined. As illustrated in Fig. 3A, coexpression of mutant dynamin results in a significant increase in cell surface HA-GLUT4-Akt2-KD in the basal state, which is similar to the insulin-stimulated level. Insulin has no further effect. However, immunoblotting of both the plasma membrane fraction (Fig. 3B) and total cell membranes (data not shown) with anti-Akt1-phospho-Ser⁴⁷³demonstrates that forcing HA-GLUT4-Akt2-KD to the cell surface by a dominant negative mutant dynamin does not itself enhance Ser⁴⁷⁴phosphorylation to the level of GLUT4-Akt2-wt in the basal state;Ser⁴⁷⁴ phosphorylation is still observed after insulin treatment where endogenous Akt2 is also activated.

In order to confirm that Ser⁴⁷⁴ phosphorylation is important for targeting HA-GLUT4-Akt2 to the cell surface and that Thr³⁰⁹phosphorylation might not be necessary for Ser⁴⁷⁴ phosphorylation when Akt2 is targeted to the GLUT4-trafficking pathway, cell-surface expression of the HA-GLUT4-Akt2-S474A and -T309A mutants was examined. The results, shown in Fig. 4,reveal that attachment of Akt2-S474A does not significantly increase the basal cell-surface level of HA-GLUT4 compared to Akt2-wt (t-test, P>0.05), while attachment of Akt2-T309A has an effect not significantly different than that of Akt2-wt (t-test, P>0.05) (Fig. 4A). Moreover, significant levels of Ser⁴⁷⁴ phosphorylation are still observed in HA-GLUT4-Akt2-T309A in the basal state(Fig. 4B).

To determine whether Akt2 targeted to the GLUT4-trafficking pathway is active in signaling, HA-GLUT4-Akt fusion proteins were co-transfected with myc-GLUT4 and the cell-surface expression of myc-GLUT4 was detected by an anti-myc antibody binding assay. In the absence of insulin, cotransfection of HA-GLUT4-Akt2-wt increases cell-surface myc-GLUT4 almost up to the level observed in the insulin-stimulated state, and insulin does not further enhance this effect. However, this effect is reduced by mutation of Ser⁴⁷⁴to Ala and not observed at all with HA-GLUT4-Akt2-KD(Fig. 5).

In this study, we used HA-GLUT4-Akt2 fusion proteins to directly study the function and subcellular distribution of Akt2 targeted to the GLUT4-trafficking pathway by assessing the phosphorylation state and cell-surface expression of the HA-GLUT4-Akt2 fusion proteins, and determining the effects of the fusion proteins on the translocation of co-transfected myc-GLUT4. Our data show that Akt2 targeted to the GLUT4-trafficking pathway is constitutively phosphorylated on Ser⁴⁷⁴ but not Thr³⁰⁹, resides primarily on the cell surface, and activates myc-GLUT4 translocation even in the absence of insulin. Both phosphorylation state and subcellular localization are critically dependent on the presence of an active kinase domain, suggesting an autocatalytic mechanism; signaling activity is also critically dependent on the presence of an active kinase domain.

Compared to normal HA-GLUT4, the HA-GLUT4-Akt2-wt fusion protein is found predominantly on the cell-surface and in the plasma membrane fraction even in the basal state, and insulin has no further effect. In contrast,HA-GLUT4-Akt2-KD behaves like normal HA-GLUT4, including translocating to the cell surface in response to insulin. We expected that Akt2 fused to GLUT4 would be found in the GLUT4 storage compartment and were surprised that it localized primarily to the cell surface. Because expressed HA-GLUT4 targets to the insulin-responsive compartment and the HA-GLUT4-Akt2-KD construct does so as well, and because Akt2 without the PH domain contains no targeting information, it appears that the two moieties must function together. Furthermore, attaching an Akt2 moiety to HA-GLUT4 does not appear to disrupt normal processing since the HA-GLUT4-Akt2-KD exhibits all of the properties of HA-GLUT4 and GLUT4. These considerations suggest that the trafficking of the GLUT4 moiety provides the mechanism through which the fusion proteins physically associate with the cell surface and that the presence of a functional Akt2 moiety is required for this association.

In order to confirm this interpretation, we examined the phosphorylation state of the Akt2 moiety. Surprisingly, our results show that Thr³⁰⁹ phosphorylation is not detected in HA-GLUT4-Akt2-wt regardless of its targeting to the cell surface in the absence or presence of insulin. In addition, ablation of Thr³⁰⁹ phosphorylation by replacing Thr³⁰⁹ with Ala does not reverse the cell-surface appearance of the HA-GLUT4-Akt2 fusion protein seen with HA-GLUT4-Akt2-wt in the basal state. Finally, translocation of GLUT4-targeted Akt2-wt is not blocked by the PI3-kinase inhibitor wortmannin (data not shown) which is known to inhibit PDK1-activated Thr³⁰⁹ phosphorylation. In contrast,Ser⁴⁷⁴ phosphorylation of HA-GLUT4-Akt2-wt is observed in the basal state at a level similar to that observed in the insulin-stimulated state. This level of Ser⁴⁷⁴ phosphorylation is not reduced by replacing Thr³⁰⁹ with Ala nor inhibiting PI3-kinase with wortamannin (data not shown). However, mutation of Ser⁴⁷⁴ to Ala in the Akt2 moiety of the fusion protein significantly reduces the association of HA-GLUT4-Akt2 with the cell surface in the basal state.

The subcellular distributions show that a significantly high proportion of expressed HA-GLUT4-Akt2-wt in both the basal and insulin-stimulated states,and HA-GLUT4-Akt2-KD in the insulin-stimulated state are phosphorylated on Ser⁴⁷⁴ in the plasma membrane fraction as compared to the other fractions, suggesting that Ser⁴⁷⁴ phosphorylation occurs at the plasma membrane (Carvalho et al.,2000; Scheid et al.,2002). Because the Akt2 moiety cannot be released from the HA-GLUT4, Ser⁴⁷⁴ phosphorylation appears to stabilize the fusion protein complex on the cell surface.

The present study using rat adipose cells provides evidence for the concept that Ser⁴⁷⁴ phosphorylation is catalyzed by Akt2 itself. For instance, in the basal state, HA-GLUT4-Akt2-wt shows Ser⁴⁷⁴phosphorylation and a cell-surface localization, while HA-GLUT4-Akt2-KD mainly remains in the intracellular sites and does not show Ser⁴⁷⁴-phosphorylation. In response to insulin, Ser⁴⁷⁴phosphorylation of HA-GLUT4-Akt2-KD is significantly increased, coincident with both translocation to the plasma membrane and activation of endogenous Akt2. To rule out the possibility that other kinases at the plasma membrane are responsible for Ser⁴⁷⁴ phosphorylation of HA-GLUT4-Akt2-wt, we co-expressed dominant negative dynamin-K44A to increase basal cell-surface HA-GLUT4-Akt2-KD without activating endogenous Akt2(Al-Hasani et al., 1998; Kao et al., 1998). This manipulation fails to enhance Ser⁴⁷⁴ phosphorylation of HA-GLUT4-Akt2-KD in the basal state despite the marked increase in cell-surface HA-GLUT4-Akt2-KD. Significant levels of Ser⁴⁷⁴phosphorylation of HA-GLUT4-Akt2-KD are observed only when endogenous Akt2 is activated in response to insulin, and because cell-surface GLUT4 do not recycle in the presence of the mutant dynammin, this phosphorylation appears to take place at the plasma membrane.

These data suggest that Akt2 activity itself is required for Ser⁴⁷⁴ phosphorylation and that Ser⁴⁷⁴ is phosphorylated through an autophosphorylation mechanism. Additional support for this concept comes from the findings that Ser⁴⁷⁴ phosphorylation is correlated with Akt kinase activity. For instance, ML-9, an inhibitor of Akt activity,inhibits insulin stimulation of Akt kinase activity, GLUT4 translocation, and Akt phosphorylation on Ser⁴⁷⁴, but not on Thr³⁰⁹ in brown adipocytes (Hernandez et al.,2001). A similar relationship is observed in adipocytes from type II diabetic subjects, i.e., both Ser⁴⁷⁴ phosphorylation and full activation of Akt are impaired, while Thr³⁰⁹ phosphorylation is much less affected (Carvalho et al.,2000).

A key finding in this study is that targeting Akt2 specifically to the GLUT4-trafficking pathway results in an insulin-like effect on Akt2 activation and GLUT4 translocation. The targeting of Akt to membranes has previously been achieved by replacing the PH domain with a membrane-targeting sequence, such as a src myristylation signal or a viral gag sequence(Cross et al., 1984; Ahmed et al., 1993). Expression of Akt constructs with a myristylation site produces a constitutive insulin-like effect similar to that observed here(Kohn et al., 1996; Kohn et al., 1996). In this case, however, the constructs are found to be bound indiscriminately to all membranes and thus do not provide information regarding a specific site of action. In the present study, targeting Akt2 specifically to the GLUT4-trafficking pathway is a direct way to examine a specific site of Akt2 action. Our data show that HA-GLUT4-Akt2-wt stimulates the cell-surface translocation of co-transfected myc-GLUT4 to a level similar to that of myc-GLUT4 itself in the insulin-stimulated state. This effect is not observed with HA-GLUT4-Akt2-KD and is decreased with the HA-GLUT4-Akt2-S/A mutant. The effects of the fusion proteins on myc-GLUT4 translocation are comparable to the behavior of the fusion proteins themselves. These data suggest that targeting Akt2 specifically to the GLUT4 trafficking pathway activates Akt2 and stimulates GLUT4 translocation in the absence of insulin. Moreover, these results together with the phosphorylation and subcellular distribution data above for the fusion proteins indicate that the plasma membrane localization is associated with Akt2 activation. With myr-Akt1, both Thr³⁰⁹ and Ser⁴⁷⁴ are phosphorylated in the absence of insulin; GLUT4-targeted Akt2-wt is only phosphorylated on Ser⁴⁷⁴, but still shows a full effect on GLUT4 translocation. This suggests that Thr³⁰⁹phosphorylation is necessary for targeting and Ser⁴⁷⁴phosphorylation, for Akt2 activity. However, because the Ser/Ala mutant is still partially active, it is possible that the remaining activity and cell-surface localization of HA-GLUT4-Akt2-S/A is attributable to Tyr⁴⁷⁵ phosphorylation since it has recently been reported that Tyr⁴⁷⁵ phosphorylation is associated with Akt2 activation(Conus et al., 2002).

While this work was in progress, the Tarvare laboratory published a report of a similar project using morphological rather than biochemical assays, and 3T3-L1 adipocytes rather than primary rat adipose cells(Ducluzeau et al., 2002). In this study, the similar fusion proteins of HA-GLUT4-Akt1-wt and -KD both showed a predominantly cytosolic localization, and the distribution of HA-GLUT4-Akt1-KD was not changed in response to insulin(Ducluzeau et al., 2002). These discrepancies are most likely related to a combination of the different approaches used for the construction of the fusion proteins, the different cell types used, and the quantitative differences in the assay techniques. In the Tavare study, Akt1 was fused to the N terminus of GLUT4 while we fused Akt2 to the C terminus of GLUT4. As we observed previously, fusion of GFP to the N terminus, but not the C terminus, of GLUT4 interferes with the targeting of GLUT4 to the cell surface (Dawson et al., 2001). This observation suggests that the discrepancy possibly lies in the different constructs used in the two studies although the behavior of Akt fusion proteins might not be exactly the same as that of GFP fusion proteins. In addition, our observations demonstrate that Akt2 kinase activity is required for Ser⁴⁷⁴ phosphorylation and association of Akt2 with the plasma membrane. For example, fusion of Akt2-KD to HA-GLUT4 does not increase basal cell-surface levels of HA-GLUT4 as the fusion of Akt2-wt does, nor is HA-GLUT4-Akt2-KD phosphorylated on Ser⁴⁷⁴ in the absence of insulin. Finally, we find that GLUT4-targeted Akt2 stimulates GLUT4 translocation without the addition of insulin.

In summary, we have fused the kinase domain of Akt2-wt and various Akt2 mutants to the C terminus of HA-GLUT4 in order to investigate the activity,phosphorylation state and subcellular localization of Akt2 targeted to the GLUT4-trafficking pathway. We observe that HA-GLUT4-Akt2-wt, but not HA-GLUT4-Akt2-KD spontaneously associates with the plasma membrane in a manner similar to the response to insulin and becomes phosphorylated on Ser⁴⁷⁴, but not Thr³⁰⁹. HA-GLUT4-Akt2-KD translocates normally with insulin and becomes phosphorylated on Ser ⁴⁷⁴ in response to insulin, but is still not phosphorylated on Thr³⁰⁹. HA-GLUT4-Akt2-wt, but not HA-GLUT4-Akt2-KD stimulates the translocation of cotransfected myc-GLUT4. These data demonstrate that targeting Akt2 to the GLUT4-trafficking pathway induces Akt2 activation and GLUT4 translocation. Ser⁴⁷⁴ phosphorylation is an autocatalytic reaction requiring an active kinase, and kinase activity is associated with a plasma membrane localization. Fusion of Akt2 to the C terminus of GLUT4 appears to substitute for Thr³⁰⁹ phosphorylation in activating the autocatalytic process.

We thank Dena R. Yver, Mary Jane Zarnowski, and Steven R. Richards for expert technical assistance.