Leucine facilitates insulin signaling through a Gαi protein-dependent signaling pathway in hepatocytes

increases through

signaling pathway. This is the first evidence showing that macronutrients like amino acid leucine can facilitate insulin signaling through G proteins directly.
Increased plasma levels of branched-chain amino acids such as leucine have been shown to be associated with insulin resistance in humans and animals (1)(2)(3)(4). Therefore, leucine has been proposed to be a contributor to the development of insulin resistance. This assertion is supported by many studies in animals and cultured cells. For example, infusion of amino acids or branchedchain amino acid leucine can induce insulin resistance (5)(6)(7)(8)(9). Exposure of cultured skeletal muscle cells, adipocytes, or hepatocytes to amino acids or leucine can induce insulin resistance (10)(11)(12). In contrast, some studies have shown that leucine supplementation does not have an effect on the development of insulin resistance or can even improve insulin sensitivity and prevent the diet-induced obesity (13,14). Explanations for these conflicting results are lacking and exact mechanism by which leucine modulates insulin sensitivity has not been established.
It has been shown that infusion of leucine can impair glucose uptake in skeletal muscle and adipose tissue (15) likely through stimulating insulin secretion (16,17) because it is known that excess exposure to insulin can induce insulin resistance (18)(19)(20). Theoretically, leucine is capable of modulating insulin signaling through mTOR and S6K-mediated serine phosphorylations of IRS1/2 (3,21). It is noteworthy that supplementation of leucine can increase insulin sensitivity in obese animals (13,14), and the increased level of branched-chain amino acids by knocking out the rate-limiting enzyme that catabolizes branched-chain amino acids is associated with increased instead of decreased insulin sensitivity (22). The increased insulin sensitivity by branched-chain amino acids can be at least partially explained by decreased food intake and elevated energy expenditure (13,14,22). Similar to leptin, leucine is capable of activating mTOR that is known to be involved in inhibition of appetite (23). Moreover, leucine and isoleucine can actually stimulate glucose uptake in skeletal muscle (24)(25)(26), and leucine can facilitate the insulin-mediated phosphorylation of Akt in a PI3K-indepndent pathway in adipocytes (27). The direct effect of leucine on hepatic insulin signaling and the associated mechanism are currently unknown. In this study, we addressed the direct effect of leucine on insulin signaling and associated mechanisms.
Cell culture: Hepa1c1c7 and C2C12 myocyte cell lines were obtained from The American Type Culture Collection (ATCC) and cultured in DMEM supplemented with 10% fetal calf serum, 2 mM L-glutamine, penicillin (100 U/ml) and streptomycin (100 ug/ml) (Invitrogen, Paisley, UK) at 37°C in a 5% CO 2 humidified incubator at 2x10 5 cells/ml density. Cells were split every two days.
Measurement of glucose production: Glucose production was assayed as described before (37). Briefly, hepatocytes were pre-treated with 1ug/ml pertussis toxin (PT) for 30 min or Gαi1 for 36 h, then treated with 0.6 mM leucine and/or 10 nM insulin for 30 min in the presence or absence of PT, followed by stimulation with cAMP/dexamethasome (10 µM/50 nM) in glucose-free DMEM containing 2 mM sodium lactate for 2.5 hour in the continuous presence or absence of insulin, leucine and PT. The culture media were subsequently collected for measuring glucose and cells were lysed with the lysis buffer for protein determination using the Pierce BCA protein assay kit (Thermo Scientific). Glucose concentrations were determined with an YSI glucose analyzer (YSI, Yellow Springs, OH) and normalized to protein concentrations.
Gene Silencing with siRNA: The siRNA against Rictor or scrambled siRNA was introduced into hepa1c1c7 hepatocytes by reverse transfection with Lipofectamine™ RNAiMAX transfection reagents with the standard approach. In brief, the siRNA transfection mixture was applied to 6 well-plates right before plating hepa1c1c7 hepatocytes in complete DMEM without antibiotics. The siRNAs against Gαi1, Gαi2, Gαi3 or scrambled siRNA was introduced into primary mouse hepatocytes by forward transfection with Lipofectamine™ RNAiMAX transfection reagents with the standard approach. In brief, the siRNA transfection mixture was added to 6 well-plates the day after plating primary hepatocytes in complete Williams' Medium E without antibiotics. After 36 h, cells were treated with leucine or/and insulin as noted. Statistical Analysis: Data are presented as mean ± SE of at least three independent experiments. Data were analyzed with one-way ANOVA, followed by LSD or Dunnett's. The SPSS 15.0 software package was used for all statistical analyses. Differences at values of p<0.05 were considered significant.

RESULTS
Leucine facilitates the insulin-induced phosphorylation of Akt in a dose-and timedependent manner. To investigate the direct effect of amino acid leucine on insulin signaling, hepa1c1c17 hepatocytes were treated with increasing amount of leucine in the presence or absence of insulin for 30 min, followed by evaluating phosphorylation of Akt at residue 473. As shown in Fig. 1A, leucine alone did not stimulate phosphorylation of Akt at residue 473 while insulin did. Addition of leucine enhanced the insulin-mediated phosphorylation of Akt at residue 473 in a dose-dependent manner and 0.3 mM of leucine showed an obvious effect while 0.6 mM of leucine induced a maximal level of effect. (Note: The concentration of leucine in the culture media used in this study was 157 µM; and the plasma level of leucine in normal subjects ranges between 132.5-176.6 µM but between 150.2-234.5 µM in obese subjects) (4, 38). To determine the optimized time point of leucine effect, hepa1c1c7 cells were treated with 0.6 mM of leucine with or without insulin for different amount of time (10-60 min), followed by evaluating phosphorylation of Akt. As shown in Fig. 1B, leucine facilitated the insulin-mediated phosphorylation of Akt at residue 473 at 30 min clearly, but not at 10 min, 20 min, or 60 min. Similarly, leucine facilitated the insulin-induced phosphorylation of Akt at residue 308 (Fig. 1C). The facilitating effect of leucine in insulin-mediated Akt phosphorylation was also observed in differentiated C2C12 myocytes (Fig. 2D). Together, these results show that leucine can facilitate the insulin-mediated phosphorylation of Akt in a dose-and timedependent manner.

Leucine facilitates the insulin-induced phosphorylation of ERK1/2 in a dose-and time-dependent manner.
To investigate the direct effect of amino acid leucine on the other main branch of insulin signaling, hepa1c1c17 hepatocytes were treated with increasing amount of leucine in the presence or absence of insulin for 10 min, followed by evaluating phosphorylation of ERK1/2. As shown in Fig. 2A, leucine alone did not stimulate phosphorylation of ERK1/2 while insulin did clearly. Addition of 0.3 mM leucine appeared to enhance the insulin-mediated phosphorylation of ERK1/2 but did not reach a statistical significance until 0.6 mM. To determine the optimized time point of leucine effect, hepa1c1c7 cells were treated with 0.6 mM of leucine with or without insulin for different amount of time (10-60 min), followed by evaluating phosphorylation of ERK1/2. As shown in Fig. 2B, leucine facilitated the insulin-mediated phosphorylation of ERK1/2 at 10 min clearly and faded gradually. Together, these results show that leucine can facilitate the insulin-mediated phosphorylation of ERK1/2 in a dose-and timedependent manner.
The leucine-facilitated insulin-induced phosphorylation of Akt 473 is mTORCindependent. It has been shown that phosphorylation of Akt at residue 473 requires mTORC2 (39). So we went to determine whether mTORC2 activation was required for the leucinefacilitated and insulin-induced phosphorylation of Akt at residue 473 by knocking down the core component of mTORC2, Rictor, with specific siRNA. As shown in Fig. 3A, leucine alone, insulin alone, or leucine plus insulin all stimulated phosphorylation of mTOR. The siRNA knocked down Rictor efficiently, but had no effect on the leucine-and insulin-induced phosphorylation of Akt at residue 473 (Fig. 3B). Similarly, knockdown of the core component of mTORC1, Raptor, had no effect on leucine-and insulininduced phosphorylation of Akt (Fig. 3C).
Together, these results demonstrate that leucine facilitates the insulin-mediated phosphorylation of by guest on March 22, 2020 http://www.jbc.org/ Downloaded from Akt at residue 473 through a pathway independent of mTORC complexes.
Leucine facilitates insulin signaling through a Gαi1-dependent pathway. Since phosphorylation of Akt at residue 473 induced by leucine and insulin was not mediated by mTORC as shown in Fig. 3, we examined the signaling pathways upstream of Akt by using inhibition of c-Src, PI3K, or Gαi proteins. As shown in Fig.  4A, phosphorylation of Akt at residue 473 induced by insulin alone was prevented by PI3K inhibitor LY294002 (LY) but not by c-Src inhibitor PP2 or Gαi protein inhibitor pertussis toxin (PT). However, phosphorylation of Akt at residue 473 facilitated by leucine in the presence of insulin was prevented by either PI3K inhibitor LY or pertussis toxin (Fig. 4B). Leucine or insulin alone stimulated phosphorylation of ERK1/2 and leucine plus insulin together further enhanced phosphorylation of ERK1/2 (Fig. 4B). Inhibitors against c-Src (PP2), PI3K (LY), or Gαi (PT) all appeared to decrease phosphorylation of ERK1/2 induced by leucine plus insulin (Fig. 4B). To determine the isoform of Gαi proteins involved in the leucine effect, we used siRNA to knock down each of the 3 known Gαi proteins. As shown in Fig. 4C, the siRNA against each isoform decreased the level of the target protein significantly, but only the knockdown of Gαi1 prevented the facilitating effect of leucine in phosphorylation of Akt. Together, these results show that leucine facilitates insulin signaling through Gαi1 protein.

Leucine facilitation of insulin-induced phosphorylation of Akt requires β-arrestin 2.
To further investigate the mechanism by which leucine facilitates insulin signaling, the involvement of a Gαi protein scaffold protein, βarrestin 2, was examined. It has previously been shown that β-arrestin 2 as a scaffold of Akt and Src plays an important role in insulin-induced Akt activation (40), but it is unclear whether β-arrestin is involved in the leucine facilitation of insulin signaling. As shown in Fig. 5, leucine facilitated the insulin-mediated phosphorylation of Akt. The effects of leucine and insulin on phosphorylation of Akt were all gone when the β-arrestin 2 gene was knocked down with siRNA. The control siRNA had no effect. Thus, these results show that leucine facilitation of insulin signaling in phosphorylation of Akt involves G protein scaffold protein β-arrestin.
Leucine facilitates insulin inhibition of glucose production through a Gαi1 proteindependent signaling pathway. To further define the mechanism by which leucine facilitated insulin signaling, the role for Gαi protein in leucine facilitation of insulin-mediated inhibition of glucose production was examined in cultured primary mouse hepatocytes. As shown in Fig. 6A, glucose production was stimulated by cAMP and dexamethasome, but the stimulation was significantly inhibited by insulin alone, negligibly reduced by leucine alone, and completely prevented by the combination of insulin and leucine. The combined inhibitory effect of insulin and leucine was also partly blocked by using either pertussis toxin (Fig. 6A) or Gαi1 gene silencing but not by siRNA against Gαi2 or 3 (Fig.  6B). Similarly, the leucine-facilitated inhibition of gluconeogenic gene (G-6-pase and PEPCK) expression was also partially reversed by inhibition of Gαi with PT or siRNA against Gαi1 (Fig. 7). Together, these results further support that leucine facilitates insulin signaling via a Gαi1 protein-dependent signaling pathway.

DISCUSSION
Reports about the regulatory effect of leucine on insulin sensitivity are currently conflicting and the associated mechanisms have not been fully established. In this study, we investigated the direct effect of leucine on insulin signaling and associated mechanisms in cultured hepatocytes. We made some important findings.
Leucine facilitates insulin signaling through a Gαi1 protein-dependent signaling pathways. Leucine has been shown by many to impair insulin signaling (10)(11)(12). The presumed mechanism is associated with activations of by guest on March 22, 2020 http://www.jbc.org/ Downloaded from mTOR/S6K and consequent serine phosphorrrylations of IRS1/2 (10,21). Nevertheless, elevated plasma levels of branched-chain amino acids including leucine by knocking out the ratelimiting enzyme of the catabolic pathway of branched-chain amino acids is linked to increased insulin sensitivity (22). Others have shown that branched-chain amino acids including leucine and isoleucine promote glucose uptake in skeletal muscle cells (24)(25)(26). Leucine has also been shown to facilitate insulin-induced phosphorylation of Akt via a PI3K-independent manner in adipocytes (27,39,41). It is noteworthy that activation of mTOR may be able to promote phosphorylation of Akt at residue 473 and consequently facilitate phosphorylation of Akt at residue 308 (39). The full activation of Akt requires phosphorylations of both 308 and 473 residues (39). The mTORpromoted activation of Akt 473 has been shown to be mediated by mTORC2 (39). In this study, we found that the leucine-facilitated insulin-induced phosphorylation of Akt at residue 473 was not affected by knocking down the core components of mTORC1 or 2, Rictor or Raptor, suggesting that the leucine facilitated insulin signaling is not mediated by mTOR. Moreover, our results clearly show that the leucine-facilitated insulin-induced phosphorylation of Akt at residue 473 was c-Src and PI3K-dependent. These results together implicate that the effect of leucine on insulin signaling comes from somewhere upstream of mTOR and Akt. Interestingly, we observed that leucine facilitation of insulin-mediated phosphorylation of Akt was prevented by the Gαi protein inhibitor, pertussis toxin or siRNA against Gαi protein 1 or β-arrestin, but not by siRNA against Gαi2 or 3. It is noteworthy that it has previously been shown that amino acids such as leucine may be able to trigger intracellular signaling through amino acid transporters, small G proteins, or currently unidentified membrane receptor(s) (42-44). However, the involvement of G protein in the leucine-mediated intracellular signaling has not been reported. Furthermore, our results clearly demonstrate that treatment of cultured hepatocytes with leucine further increases insulin suppression of glucose production and expression of key gluconeogenic genes. The leucine-facilitated insulin suppression of glucose production and gluconeogenesis was blunted by inhibition of Gαi protein. Finally, our results show that leucine can also facilitate insulin-induced phosphorylation of ERK1/2. Therefore, we have found a brand new signaling pathway by which leucine facilitates insulin signaling.
In summary, the findings in this study will not only help explain how macronutrients like leucine directly regulate insulin signaling and metabolism but also help explain why results are sometimes conflicting among different studies.    Hepa1c1c7 hepatocytes were treated with leucine, insulin, or insulin plus leucine for 30 min, followed by evaluations of phosphorylated mTOR, total mTOR, and β-actin by immunoblotting and then quantified. (B) Hepa1c1c7 cell were transfected with siRNAs against Rictor or scrambled siRNA for 36 h, and were then treated with leucine or insulin alone or insulin plus leucine for 30 min, followed by evaluations of phosphorylated Akt (at serine 473), total Akt, total Rictor, and β-actin by using immunoblottings with specific antibodies. (C) Hepa1c1c7 cell were transfected with siRNAs against Raptor or scrambled siRNA for 36 h, and were then treated with leucine (0.6 mM) or insulin (10 nM) alone or insulin plus leucine for 30min, followed by evaluations of phosphorylated Akt (at serine 473), total Akt, total Raptor, and β-actin by using immunoblottings. Results were presented as the mean ± SE of 4 independent experiments, each in duplicate. Bars not sharing a letter are statistically different (p <0.05). insulin together for 30 min as noted, followed by evaluations of phosphorylated and total Akt, each Gαi isoform protein, and β-actin using immunoblotting and then quantification. Results were presented as the mean ± SE of 4 independent experiments, each in duplicate. Statistical significance is denoted by the presence of different letters above the bars. Bars not sharing a letter are statistically different (p <0.05).

Figure 5. Leucine facilitation of insulin-mediated phosphorylation of Akt requires β-arrestin 2.
Hepa1c1c7 hepatocytes were first transfected with siRNA against β-arrestin 2 or scrambled siRNA for 36 h, and then treated with leucine (0.6 mM) or insulin (10 nM) alone or leucine plus insulin together for 30 min as noted, followed by evaluations of phosphorylated and total Akt, β-arrestin 2, and β-actin using immunoblotting and then quantification. Results were presented as the mean ± SE of 3 independent experiments, each in duplicate. Statistical significance is denoted by the presence of different letters above the bars. Bars not sharing a letter are statistically different (p <0.05).