Insulin promotes the cell surface recruitment of the SAT2/ATA2 system A amino acid transporter from an endosomal compartment in skeletal muscle cells.

SAT1-3 comprise members of the recently cloned family of System A transporters that mediate the sodium-coupled uptake of short chain neutral amino acids, and their activity is regulated extensively by stimuli such as insulin, growth factors, and amino acid availability. In skeletal muscle, insulin stimulates System A activity rapidly by a presently ill-defined mechanism. Here we demonstrate that insulin induces an increase in the plasma membrane abundance of SAT2 in a phosphatidylinositol 3-kinase-dependent manner and that this increase is derived from an endosomal compartment that is required for the hormonal activation of System A. Chloroquine, an acidotropic weak base that impairs endosomal recycling of membrane proteins, induced a complete inhibition in the insulin-mediated stimulation of System A, which was associated with a loss in SAT2 recruitment to the plasma membrane. The failure to stimulate System A and recruit SAT2 to the cell surface could not be attributed to a block in insulin signaling, as chloroquine had no effect on the insulin-mediated phosphorylation of protein kinase B or glycogen synthase kinase 3 or upon insulin-stimulated GLUT4 translocation and glucose transport. Our data indicate strongly that insulin increases System A transport in L6 cells by stimulating the exocytosis of SAT2 carriers from a chloroquine-sensitive endosomal compartment.

The plasma membranes of mammalian cells possess multiple transport systems for the cellular exchange of amino acids (1), and of the classical amino acid transport systems described the best studied has been the System A amino acid carrier, which mediates the sodium-coupled uptake of short-chain neutral amino acids (e.g. alanine). System A is expressed in many cell types. Its ability to mediate the uptake of amino acids with N-methyl substitutions has enabled the discrimination of System A from other amino acid transporters using N-methylaminoisobutyric acid (Me-AIB) 1 as a paradigm substrate. Func-tional studies using this substrate have shown that System A activity is highly pH-sensitive and that the carrier is subject to both long-and short-term modulation (1). Numerous functional studies have demonstrated, for example, that System A activity is subject to adaptive up-regulation by amino acid deprivation in multiple cell types and that this response is dependent upon gene expression, since it is inhibited by cycloheximide and actinomycin D, inhibitors of translation and transcription, respectively (recently reviewed by us (2)). Whether such modulation results from the increased expression of System A carriers or that of regulatory molecule(s) capable of effecting changes in transporter activity has remained unclear. However, the recent cloning of three isoforms of the System A transporter (SAT, also known as amino acid transporter A (ATA)), termed SAT1-3/ATA1-3 (3)(4)(5)(6), has meant that the molecular regulation of this transport system can now be investigated in detail. The availability of molecular probes to System A has allowed us to demonstrate recently that the adaptive increase in transport activity triggered by amino acid deprivation stems from a selective up-regulation in the expression of SAT2 protein in the plasma membrane of muscle and fat cells (7). This finding is fully consistent with the observation of others that depriving human fibroblasts of amino acids induces an increase in SAT2 mRNA levels (8).
In addition to adaptive regulation, System A can also be stimulated acutely by hormones such as insulin (9 -11), growth factors (12), and cell stresses (11)(12)(13). These stimuli induce a rapid increase in the V max of System A transport that does not rely upon synthesis of new carriers, but which is thought to involve the acute modulation of carrier function effected by molecules participating in early insulin and growth factor signaling. We have shown previously that the stimulatory effects of insulin and IGF-1 on System A can be blocked by inhibitors of phosphatidylinositol 3-kinase (PI3K) (10 -11) and that expression of a constitutively active form of protein kinase B (PKB), which lies downstream of PI3K in the insulin signaling cascade, mimics the effect of insulin on this amino acid transporter (14). Precisely how the hormonal activation of PI3K and PKB are linked to an increase in functional System A activity is not yet understood, but it is plausible that signaling from these molecules either stimulates the activity of SAT proteins resident in the plasma membrane or that additional SAT carriers are recruited to the cell surface from an intracellular compartment. The concept of carrier recruitment from intracellular compartments in response to insulin is well established for the GLUT4 glucose transporter (15). However, it remains currently unknown whether SAT proteins can be recruited to the plasma membrane in a manner analogous to GLUT4. In an attempt to address this issue, we have investigated the effects of the acidotropic weak base, chloroquine (CQ), which impairs endosomal function and hence protein recycling (16,17), and cytochalasin D (CD), which disrupts the actin-dependent endocytosis of cell surface proteins (18,19) on the insulin-mediated activation of System A and upon the subcellular distribution of SAT2. We demonstrate, for the first time, that disrupting endosomal trafficking results in the loss of insulin-stimulated System A transport and that this correlates with a failure to recruit the SAT2 System A carrier from endosomal membranes.

EXPERIMENTAL PROCEDURES
Materials-Culture media (␣-minimal essential media, ␣-MEM), fetal calf serum, and antimycotic/antibiotic solution were obtained from Invitrogen. Wortmannin, insulin, chloroquine, and cytochalasin D were purchased from Sigma-Aldrich. Complete protease inhibitor tablets were obtained from Roche Molecular Biochemicals. Phospho-specific antibodies to PKB and GSK3 were obtained from New England Biolabs (Hertfordshire, UK). Horseradish peroxidase conjugated anti-rabbit IgG were obtained from Scottish Antibody Production Unit (Carluke, Lanarkshire, UK). Hybond nitrocellulose membrane was obtained from Amersham Biosciences, and reagents for ECL were purchased from Pierce & Warriner (Chester, UK).
Cell Culture-Monolayers of L6 muscle cells were cultured to the stage of myotubes as described previously (14 -20) in ␣-MEM containing 2% (v/v) fetal calf serum and antimycotic/antibiotic solution (100 units/ml penicillin, 100 g/ml streptomycin, and 250 ng/ml amphotericin B) at 37°C in an atmosphere of 5% CO 2 , 95% air. Upon formation of myotubes, cells were deprived of serum by incubating muscle cells in serum-free ␣-MEM for 4 h followed by a 1 h incubation in amino acid-free HEPES-buffered saline (HBS, 20 mM HEPES-Na (pH 7.4), 140 mM NaCl, 2.5 mM MgSO 4 , 5 mM KCl, 1 mM CaCl 2 ). Additions (e.g. insulin, CQ, and cytochalasin D) to the cells were made at times and at concentrations indicated in the figure legends.
Subcellular Fractionation of L6 Myotubes-Subcellular membranes from L6 myotubes were isolated as described previously (21)(22)(23)(24). Briefly, 2 h prior to fractionation, cells were washed, and growth medium (containing 5 mM glucose) was supplemented with chloroquine (100 M). In some dishes, insulin (100 nM) was added 30 min prior to cell harvesting; control dishes received vehicle alone. At the end of the incubation period, cells from five (15 cm) dishes were scraped off the plates with a rubber policeman, pooled, and gently pelleted. The cell pellet was resuspended in ice-cold buffer (250 mM sucrose, 20 mM Hepes, 5 mM NaN 3 , 2 mM EGTA, pH 7.4 plus 1 protease inhibitor tablet/50 ml) and homogenized. The cellular homogenate was subjected to a series of differential centrifugation steps to isolate crude cell membranes that were subsequently fractionated on a discontinuous sucrose gradient (32,40, and 50% sucrose by mass) at 210,000 ϫ g for 2.5 h. Membranes from on top of the 32% sucrose cushion and those at the 32/40% and 40/50% sucrose interfaces were recovered and their protein content determined using the Bradford assay with bovine serum albumin as standard (25).
Subcellular Fractionation of Rat Skeletal Muscle-Rat skeletal muscle was excised from the hind limbs of male Sprague-Dawley rats (250 -300 g) and fractionated based on a procedure established by Klip and co-workers (26). The procedure involves skeletal muscle homogenization and a series of differential centrifugation steps that allow isolation of crude muscle membranes. These were subsequently fractionated on a discontinuous sucrose density gradient (25,30, and 35%) as described previously (26,27). This procedure resulted in the separation of three distinct membrane bands; one band was located above the 25% sucrose cushion representing membranes enriched with plasma membranes (denoted as F25), and a second band separated on top of the 30% sucrose layer (F30) contained membranes largely of endosomal origin (27). A third band separated out on top of the 35% sucrose layer (F35) which consisted of membranes largely of intracellular origin that contain the insulin "recruitable" pool of GLUT4 glucose transporters (28). Protein content was determined using the Bradford method (25).
Measurement of System A Amino Acid Transport-System A activity was assayed by measuring the uptake of Me-AIB as described previously (20). Briefly, L6 myotubes were incubated with 10 M [ 14 C]Me-AIB (1 Ci/ml) for 10 min. Nonspecific tracer binding was determined using either [ 3 H]mannitol as an extracellular marker or by determining cell-associated radioactivity in the presence of an excess saturating dose of unlabeled Me-AIB (10 mM). In some experiments glucose transport was measured simultaneously in the same population of muscle cells used to assay System A activity by a dual label approach (12). Uptake of 10 M 2-deoxy-D-[ 3 H]glucose (2DG; 1 Ci/ml, 26.2 Ci/mmol) was measured for 10 min. Carrier-mediated glucose transport was determined by quantitating cell-associated radioactivity in the presence of 10 M cytochalasin B, an inhibitor of facilitative glucose transporters (14). Uptake of both Me-AIB and 2DG was determined by aspirating the radioactive medium, followed by three successive washes in ice-cold isotonic saline solution (0.9% NaCl, w/v). Cells were lysed in 0.05 M NaOH and the associated radioactivity determined by liquid scintillation counting. Total cell protein was determined by the method of Bradford (25).
Analysis of Cell Surface 125 I-Transferrin Binding-The transferrin receptor is known to recycle constitutively between the cell surface and an endosomal compartment. To assess the effects of CQ and CD on endosomal recycling, cell surface transferrin (Tfn) binding in L6 cells was determined essentially as described previously (29). Briefly, human Tfn (Sigma) was iodinated (2.6 ϫ 10 6 cpm/g) in phosphate-buffered saline (PBS) containing 1% bovine serum albumin (w/v), pH 7.4, with carrier free Na 125 I (PerkinElmer Life Sciences) for 10 min. Iodinated Tfn was isolated by passage through a Dowex-1 ion exchange column.
Muscle cells were incubated with CQ and CD for times and at concentrations indicated in the figure legend (Fig. 2) prior to incubation for 30 min in MHB buffer (serum-free ␣-MEM medium, 20 mM HEPES (pH 7.4), 2 mg/ml bovine serum albumin) at 37°C to deplete endogenous transferrin in the culture media. Cells were subsequently incubated in MHB buffer containing 3 g/ml 125 I-Tfn (5 ϫ 10 4 cpm/g) for 60 min at 4°C followed by three successive washes in ice-cold PBS and lysis in 1% Triton X-100 prior to quantitating total cell-associated radioactivity using a Beckman LS6000IC counter. To quantitate the internalized radioactivity, the above protocol was repeated on a duplicate set of cells, but this set was washed three times with PBS/25 mM glacial acetic acid (pH 4.2) following the initial three washes with ice-cold PBS to strip surface-associated 125 I-Tfn. The difference between the two samples gives an indication of surface Tfn binding. Nonspecific surface binding was corrected for by determining radioactivity associated with cells incubated at 4°C with MHB containing 3 g/ml 125 I-Tfn with a 100-fold excess of unlabeled Tfn.
Statistical Analysis-Statistical analysis for multiple comparisons was performed using one-way analysis of variance (ANOVA) followed by the Newman-Keuls test. Data analysis was performed using Graph-Pad Prism software and considered statistically significant at p values Ͻ0.05.

RESULTS AND DISCUSSION
To investigate the effects of CQ on System A activity in L6 myotubes we first performed dose and time response studies with this acidotropic agent. Muscle cells were preincubated with 100 M CQ for up to 3 h, after which period basal and insulin-stimulated Me-AIB uptake was assayed. In the absence of any pretreatment with CQ, insulin increased Me-AIB uptake by ϳ65%, but this stimulation was progressively lost upon pretreatment of cells with CQ (Fig. 1A). Under these conditions CQ also induced a modest reduction (ϳ25%) in basal System A transport activity (Fig. 1A). Insulin-stimulated System A activity was virtually undetectable in cells pre-exposed to 100 M CQ and was maximally suppressed following preincubation with CQ for 2 h (Fig. 1, A-C). These findings indicate that System A transporters may recycle constitutively at the cell surface and that insulin stimulates System A by enhancing its exocytosis from a CQ-sensitive endosomal compartment. If this supposition is correct, then System A activity in the plasma membrane will depend on the relative rates with which the carrier protein is exocytosed and endocytosed.
Studies in yeast and mammalian cells indicate strongly that In some experiments, muscle cells were washed following pretreatment with CQ and allowed subsequently to recover. Cells were exposed subsequently to insulin (100 nM, 30 min) at times indicated during the post-wash period and insulin-stimulated Me-AIB uptake assayed. Results represent the mean Ϯ S.E. for 3-8 experiments, each performed in triplicate. The asterisks represent a statistically significant change (p Ͻ 0.05) from the appropriate untreated control.
actin filaments and actin-based motor proteins play a key role in receptor-mediated endocytosis (18 -30). Thus, to examine further the importance of exo-and endocytic events in the regulation of System A activity, we investigated the effects of treating muscle cells with cytochalasin D, a reagent that induces depolymerization of actin filaments, in addition to CQ. To assess how each drug affects the recycling of membrane proteins we monitored surface binding of 125 I-transferrin as an index for plasma membrane transferrin receptor (TfR) abundance, a paradigm marker protein that recycles constitutively at the cell surface. Fig. 2 shows that surface Tfn binding was increased significantly (by 73%) following incubation of cells with CD, whereas it fell by ϳ40% in cells that had been exposed to CQ (Fig. 2). These findings are consistent with the view that CD and CQ inhibit endocytosis and exocytosis of TfR, respectively. Previous work using rat adipocytes has shown that the CQ-induced disruption in IGF-II receptor recycling is reversible, suggesting that endosomal function can be restored upon withdrawal of CQ from the incubation medium (31). To test for whether this was also the case in L6 myotubes, muscle cells were washed free of CQ, and Tfn binding was assayed during the post-wash recovery period. Withdrawal of CQ from the incubation medium for periods of up to 2 h led to a restoration in Tfn binding to a level that was not significantly different from cells that had not been exposed to CQ (Fig. 2).
Having established that CD inhibits TfR internalization, we assessed the effects of this drug on System A transport activity. Preincubation of L6 myotubes with 2 M CD alone induced a near 30% stimulation in Me-AIB uptake (Fig. 3). This increase most likely arises through a modest up-regulation in cell surface System A carrier number that is associated with inhibition of their internalization, whereas carrier exocytosis from the endosomal compartment remains unaffected. This proposition is supported by the finding that the CD-induced increase in System A activity did not take place when muscle cells were treated with CQ just prior to incubation with CD (Fig. 3), under which conditions both endo-and exocytosis are inhibited. Treatment of cells with insulin following incubation with CD results in a further increase in System A activity over and above that elicited by CD but to a level no greater than that observed in the presence of insulin alone (Fig. 3) This finding provides further support for the idea that exocytosis of carriers from an endosomal pool is the mechanism by which insulin stimulates System A transport.
PI3K has been implicated strongly in the control of both basal and insulin-stimulated exocytosis of recycling proteins such as TfR (32,33). We hypothesized that if the stimulation of System A elicited by both CD and insulin relies upon carrier recruitment from an endosomal compartment, then it ought to be blocked by the PI3K inhibitor, wortmannin. Consistent with this idea, wortmannin prevented stimulation of Me-AIB uptake in response to CD and insulin (Fig. 3), and use of the structurally unrelated PI3K inhibitor, LY 294002, yielded similar results (data not shown).
The data presented in Fig. 2 indicate that endosomal function can be restored slowly when muscle cells are washed and allowed to recover for 2 h following pretreatment with CQ. We postulated that if the hormonal activation of System A was dependent on a functional endosomal compartment, then the ability of insulin to stimulate System A activity should also be recoverable during the post-wash period. Fig. 4 shows that CQ suppressed completely the hormonal activation of System A but that this stimulation was restored progressively in cells that had been washed "free" of the drug and allowed to recover subsequently in media lacking CQ for up to 2 h.
It is plausible that the inhibitory effects exerted by CQ on the hormonal activation of System A might involve direct effects of the drug on membrane transporters or through a blockade in insulin signaling. To examine these possibilities we investigated the effects of CQ on insulin-stimulated glucose transport and upon the phosphorylation of two important insulin signaling intermediates, PKB and glycogen synthase kinase-3 (GSK3) (34). Although CQ inhibited the hormonal stimulation of System A transport, the drug had no apparent effect on insulin-stimulated glucose transport when assayed in the same cell population (Fig. 5A). This observation is in agreement with previous work using adipocytes showing that although CQ disrupts the recycling of IGF-II receptors, it has little impact on insulin-stimulated glucose transport or upon transporter recruitment to the cell surface (31)(32)(33)(34)(35). These findings imply, first, that although a proportion of the insulin-regulated glucose transporter, GLUT4, localizes to recycling endosomes (36), the effects of CQ on GLUT4 recycling are likely to be minimal. Second, because CQ inhibits insulin-stimulated System A transport and influences the recycling dynamics of TfR and the IGF-II receptor (31), this indicates that such proteins are present in a distinct population of endosomal vesicles from which GLUT4 is excluded. To assess the possibility that CQ may have impaired insulin signaling, we monitored the insulin-induced During the last 10 min of this incubation, muscle cells were exposed to insulin (100 nM) and then lysed. Lysates were subjected to SDS-PAGE and immunoblotting with phospho-specific antibodies to PKB-Ser 473 and GSK3 or an antibody to the C-terminal domain of PKB␣. The blots are representative of data from three experiments. phosphorylation of PKB, an insulin signaling intermediate implicated in the regulation of diverse end point responses to the hormone such as glycogen synthesis (37), glucose transport, and System A transport (14). Fig. 5B shows that insulin promoted the phosphorylation of PKB and also that of its downstream physiological target, GSK3. This finding is fully compatible with the observed stimulation in glucose transport elicited by the hormone (Fig. 5A) and indicates strongly that CQ is unlikely to have inhibited activation of System A through suppression of early insulin signaling events.
Three SAT isoforms (SAT1-3, also known as ATA1-3) have recently been cloned (3)(4)(5)(6). Of these, SAT2 appears to be the most abundantly expressed isoform in skeletal muscle, whereas SAT3 expression occurs at a much lower level, and that of SAT1 is not detectable (6 -7). Recent work from our laboratory has shown that SAT2 protein expression is enhanced in muscle and fat cells following a period of amino acid deprivation, suggesting that changes in SAT2 expression are likely to underpin the adaptive increase in System A activity (7). Moreover, the observation that SAT3 is a relatively poor mediator of Me-AIB uptake compared with the other isoforms (38) implies that SAT2 is the most likely isoform to mediate insulin-stimulated Me-AIB uptake in muscle. To test this proposition, we investigated the effects of insulin and chloroquine treatment on the subcellular distribution of SAT2 in L6 myotubes. For these studies a 32, 40, and 50% (w/w) discontinuous sucrose gradient was utilized. The membranes that float on top of the 32% sucrose cushion have been shown previously to be enriched with plasma membrane (PM) markers such as the ␣1 subunit of the Na,K-ATPase (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22). Membranes from the 32/40% and 40/50% sucrose interfaces are depleted in PM markers but are enriched with proteins of intracellular origin. Fig. 6A shows the relative abundance of annexin II, a protein that is a major component of fusogenic endosomal vesicles and thus a representative endosomal marker, in the three membrane fractions. Annexin II was detected in all three membrane fractions, but its abundance was greatest in membranes recovered from the 32/40% sucrose gradient interface (Fig. 6A), indicating that this fraction was enriched with endosomal membranes.
Consistent with our recent work (7), SAT2 was detected in all three membrane fractions but was enriched in membranes recovered from the 32/40% sucrose interface (Fig. 6A). When SAT2 abundance was investigated in the three fractions following isolation of cells treated with insulin, we did not observe any changes in SAT2 content in membranes recovered from the 40/50% sucrose interface (data not shown). However, data from three separate experiments revealed that the hormone induced a greater than 2-fold increase in SAT2 in the PM-enriched fraction, Fig. 6B (this was associated with an increase in the V max of System A transport from 388 Ϯ 10 pmol/min/mg to 584 Ϯ 16 pmol/min/mg (p Ͻ 0.05), with no significant change in K m ). The increase in cell surface SAT2 was associated with a concomitant reduction in SAT2 abundance in membranes recovered from the 32/40% sucrose interface (Fig. 6B). These findings imply that insulin promotes recruitment of SAT2 to the PM from the 32/40% fraction, which is enriched with endosomal membranes. Pretreatment of L6 cells with CQ for 2 h prior to subcellular fractionation reduced the amount of SAT2 recovered in the PM fraction, a finding consistent with the idea that this compound prevents the delivery of proteins to the cell surface from the endosomal compartment without affecting the internalization of PM proteins that are constitutively recycling at the cell surface. The loss in PM SAT2 was not recovered in any of the other membrane fractions isolated from the sucrose gradient, raising the possibility that, once internalized, SAT2 may have been routed for lysosomal degradation. Unlike SAT2, pretreatment of L6 cells with CQ had no detectable effect on FIG. 6. Effects of insulin and CQ on the subcellular distribution of SAT2 and GLUT4 in L6 myotubes and SAT2 distribution in rat skeletal muscle. Confluent L6 myotubes were subjected to subcellular fractionation on a discontinuous sucrose gradient (32,40, and 50% sucrose (w/w)). Membranes were prepared from either untreated muscle cells or following treatment with 100 M CQ for 2 h and/or with 100 nM insulin (30 min). Plasma membranes recovered from the top of the 32% sucrose cushion and those from the 32/40% and 40/50% sucrose interfaces were resolved on SDS-gels and immunoblotted with antibodies to annexin II, the a1 subunit of the Na,K-ATPase, or SAT2. Representative immunoblots from three separate experiments show the relative distribution of annexin II and SAT2 in the three fractions (A) and the effects of insulin and CQ treatment on SAT2 distribution in the PM and the 32/ 40% membrane fraction (B). C, PM were isolated from untreated L6 myotubes or following pretreatment with 100 nM insulin (30 min) either alone or in the presence of CQ (100 M, 2 h) or 100 nM wortmannin (45 min) and immunoblotted with GLUT4 antibodies. D, membrane fractions (20 g protein) isolated from rat skeletal muscle (see "Experimental Procedures") were immunoblotted with antibodies to the ␣1-subunit of the Na,K-ATPase, transferrin receptor (TfR), SAT2, and GLUT4. the insulin-dependent recruitment of the GLUT4 glucose transporter to the PM (Fig. 6C). The increase in GLUT4 abundance elicited by insulin was, however, prevented by the PI3K inhibitor, wortmannin (Fig. 6C). These findings are consistent with the observation that CQ does not impair insulin-stimu-lated glucose transport (Fig. 5A) and with numerous reports showing that insulin-stimulated GLUT4 translocation requires PI3K activity (39).
Interestingly, analysis of SAT2 abundance in subcellular membrane fractions prepared from rat skeletal muscle indicate that, as in L6 cells, the carrier is present in membranes of endosomal origin (F30) based on enrichment of TfR (a marker of recycling endosomes). In contrast, SAT2 immunoreactivity was very low in membranes that harbor the insulin responsive GLUT4 pool (F35) (Fig. 6D). Indeed, in the present studies the carrier was only detected in this fraction upon prolonged exposure of autoradiographic film. Although SAT2 immunoreactivity was greatest in the PM (F25) fraction, it should be noted that the abundance of SAT2 in the F30 fraction is not insignificant, as the total protein recovery of this fraction is ϳ5-fold higher than that of PM enriched fraction (27). These findings thus raise the possibility that SAT2 may also be recruited to the sarcolemma of rat muscle from an endosomal pool as observed in our cell-based studies.
To further strengthen our assertion that SAT2 mediates insulin-stimulated System A transport, we investigated (i) whether the insulin-mediated increase in plasma membrane SAT2 was sensitive to wortmannin, a PI3K inhibitor; and (ii) whether SAT2 translocation could be rescued following CQ wash-out in a time-dependent manner. Fig. 7A shows that the increase in plasma membrane SAT2 elicited by insulin was suppressed by prior treatment of cells with 100 nM wortmannin, a finding consistent with the observation that wortmannin inhibits the hormonal activation of System A transport (Fig. 3). In line with the data presented in Fig. 4, we observed that CQ prevented the insulin-induced increase in cell surface SAT2 but that cells "washed free" of CQ displayed a time-dependent return of SAT2 recruitment (Fig. 7B). Taken together, the findings presented in Figs. 6 and 7 suggest strongly that SAT2 translocation underpins the functional increase in System A transport that is observed in response to insulin in L6 cells.
In summary, we have shown that treatment of skeletal muscle cells with insulin increases System A amino acid transport due to the translocation of SAT2 protein from an intracellular pool (which contains a high abundance of the endosomal marker annexin II) to the plasma membrane. CQ, a substance that disrupts the trafficking of endosomal transferrin receptors to the plasma membrane, prevents both the recruitment of SAT2 from this internal pool and the increase in System A transport following insulin treatment. Furthermore, disrupting actin microfilaments with CD moderately increases both System A activity and transferrin binding, consistent with a role for the actin network in endocytosis from the plasma membrane. Collectively these results imply that insulin stimulates System A activity by increasing the exocytosis of the SAT2 amino acid transporter from an intracellular endosomal location to the plasma membrane. Our present working model for the hormonal regulation of skeletal muscle System A transport is shown in Fig. 8. Future experiments will seek to determine the nature of the endosomal SAT2 vesicles in greater detail and how the endosomal and cytoskeletal systems interact with insulin signaling networks to regulate System A activity. FIG. 7. Insulin-stimulated SAT2 translocation is PI3K-dependent and can be reinstated following recovery of muscle cells from chloroquine. A, plasma membranes were isolated from either untreated L6 myotubes or following treatment with insulin (100 nM, 30 min) alone or with insulin (100 nM, 30 min) and 100 nM wortmannin (45 min) prior to SDS-PAGE and immunoblotting with SAT2 antibodies. B, muscle cells were treated with insulin (100 nM, 30 min) alone or exposed to the hormone during the last 30 min of a 2-h CQ pretreatment period (CQ present at 100 M). Cells were either subjected to subcellular fractionation to isolate PM or washed three times with HBS and allowed to recover for the times indicated prior to a second challenge with insulin alone and subsequent membrane isolation. Isolated PM fractions were resolved on SDS-gels and immunoblotted with antibodies to SAT2 or the ␣1-subunit of the Na,K-ATPase (used as loading control). The data shown in B are from two separate experiments.
FIG. 8. Scheme illustrating our current model for the insulinmediated activation of System A in skeletal muscle. Insulin stimulates the membrane proximal effectors PI3K and PKB, which then promote the translocation of SAT2-containing vesicles from a CQ-sensitive endosomal pool to the PM, facilitating the transport of short chained neutral amino acids (AA) into the cell. CD inhibits actin-dependent endocytosis and thus increases the activity of SAT2 at the plasma membrane. The solid lines represent confirmed steps, and dashed lines represent putative steps.