点击显示 收起
1 From the Department of Cellular and Molecular Physiology, The Pennsylvania State University College of Medicine, Hershey, PA
2 Presented at a memorial symposium held in honor of Vernon R Young in Cambridge, MA, 12 November 2004. 3 Supported by grant DK15658 from the National Institutes of Health. 4 Address reprint requests to LS Jefferson, The Department of Cellular and Molecular Physiology, The Pennsylvania State University College of Medicine, 500 University Drive, Hershey, PA 17033-0850. E-mail: jjefferson{at}psu.edu.
ABSTRACT
Amino acids act to regulate multiple processes related to gene expression, including modulation of the function of the proteins that mediate messenger RNA (mRNA) translation. By modulating the function of translation initiation and elongation factors, amino acids regulate the translation of mRNA on a global scale and also act to cause preferential changes in the translation of mRNAs encoding particular proteins or families of proteins. However, amino acids do not directly regulate the function of translation initiation and elongation factors, but instead modulate signaling through pathways traditionally considered to be solely involved in mediating the action of hormones. The best-characterized example of amino acid–induced regulation of a signal transduction pathway is one involving a protein kinase referred to as the mammalian target of rapamycin (mTOR), through which the branched-chain amino acids, particularly leucine, act to modulate the function of proteins engaged in both global mRNA translation and the selection of specific mRNAs for translation. Less understood at this point in time is evidence suggesting that amino acids also act to regulate mRNA translation through mTOR-independent mechanisms. The goal of the present review is to briefly summarize studies, primarily those performed in the laboratories of the authors, that focus on the role of the branched-chain amino acids in the regulation of mRNA translation in skeletal muscle.
Key Words: Branched-chain amino acids • leucine • mTOR • mammalian target of rapamycin • insulin • mRNA translation
INTRODUCTION
A growing number of examples illustrate that the expression of many and perhaps all genes is regulated at multiple steps that include transcription, posttranscriptional processing, nuclear export, stability, and translation of mature messenger RNA (mRNA) molecules into protein. The translation step itself is regulated by a diverse array of mechanisms that act not only on initiation [ie, the recruitment of the initiator methionyl–transfer RNA (met-tRNAi) and mRNA to the 40S ribosomal subunit followed by the joining of the 60S ribosomal subunit to complete assembly of a translationally competent 80S ribosome], but also during elongation (ie, movement of the ribosome along the mRNA as amino acids are incorporated into the growing peptide chain) and termination (ie, the release of the full-length protein and the 40S and 60S ribosomal subunits from the mRNA). The mechanisms that act on initiation are regulated by numerous agents, including hormones, macronutrients, energy state, and others. Early work from our laboratory showed a key role for the branched-chain amino acids, particularly leucine, in the regulation of initiation.
Vernon Young, in yet another example of his truly remarkable vision, was quick to recognize the importance of this regulatory mechanism in gene expression and whole-body protein metabolism and as a result provided much guidance and encouragement to the author (LSJ) over the years, culminating in the request to have the work presented at this symposium. Our intent in presenting the work is to first give a brief overview of the effector mechanisms involved in the initiation of mRNA translation and the signaling pathways that regulate them. We will then briefly review recent work conducted in our laboratories or through collaboration with other laboratories on the role of the branched-chain amino acids in the regulation of mRNA translation in skeletal muscle. The focus here is on skeletal muscle because of the overall contribution of this tissue to whole-body protein and amino acid metabolism, a topic of considerable interest to Vernon Young during his illustrious career.
OVERVIEW OF EFFECTOR MECHANISMS INVOLVED IN THE INITIATION OF mRNA TRANSLATION
The translation of mRNA into protein occurs through a series of events that are classically grouped into 3 stages: initiation, elongation, and termination (reviewed in reference 1). Most of the regulatory mechanisms identified thus far mediate control of the initiation of mRNA translation, and therefore this stage and the signaling pathways that regulate it will be the primary focus of the presentation that follows. The reactions that mediate the initiation of mRNA translation can be summarized as follows: met-tRNAi binds to the 40S ribosomal subunit forming the 43S preinitiation complex, the complex binds to mRNA and localizes to the AUG start codon, and the initiation factors are released from the 40S ribosomal complex, which allows the 60S ribosomal subunit to join to form the 80S ribosomal complex that is competent to proceed to the elongation stage of translation. The binding of met-tRNAi to the 40S ribosomal subunit is mediated by a binary complex consisting of eukaryotic initiation factor 2 (eIF2) and GTP. In a later step, the GTP bound to eIF2 is hydrolyzed to GDP, and the eIF2 · GDP complex is released from the 40S subunit, leaving met-tRNAi behind. Exchange of GDP bound to eIF2 for GTP is mediated by the guanine nucleotide exchange factor eIF2B and allows eIF2 to participate in another round of initiation. The -subunit of eIF2B is the catalytic subunit that mediates the guanine nucleotide exchange reaction, whereas the other 4 subunits modulate the activity of the protein.
The binding of mRNA to the 43S preinitiation complex is mediated by a heterotrimeric complex referred to as eIF4F. The 3 subunits of eIF4F are as follows: eIF4A, an RNA helicase that unwinds secondary structure in the 5- noncoding region of the mRNA; eIF4E, the mRNA cap-binding protein; and the scaffolding protein eIF4G that acts as a molecular bridge between the eIF4E · mRNA complex and the 40S ribosomal subunit. eIF4G also binds to the poly(A)-binding protein (PABP), an interaction that is thought to circularize the mRNA, bringing the 3- end of the mRNA near the 5- end and allowing for more efficient reinitiation of translation.
The association of eIF4E with eIF4G, and therefore binding of mRNA to the 40S ribosomal subunit, is regulated in part by the association of eIF4E with several eIF4E binding proteins, of which eIF4E binding protein (4E-BP)1 is best characterized. Each of the 4E-BPs binds to the same site on eIF4E, as does eIF4G, thereby preventing assembly of the active eIF4F complex. Phosphorylation of the 4E-BPs on multiple residues results in dissociation of the eIF4E · 4E-BP complex, freeing eIF4E to bind to eIF4G. Assembly of the eIF4F complex may also be regulated through phosphorylation of eIF4E and eIF4G, although the mechanisms involved are poorly defined.
SIGNALING PATHWAYS THAT REGULATE EFFECTOR MECHANISMS FOR INITIATION OF mRNA TRANSLATION
mTOR-dependent signaling
With regard to the regulation of mRNA translation, the signaling pathway that has generated the most interest is that involving the protein kinase termed the mammalian target of rapamycin (mTOR) (for recent reviews, see references 2-4). mTOR directly phosphorylates the 4E-BPs as well as the ribosomal protein (rp) S6 kinase S6K1 (Figure 1). Phosphorylation of the 4E-BPs results in release of eIF4E from the inactive 4E-BP · eIF4Ecomplex, which allows eIF4E to bind to eIF4G and form the active eIF4F complex. Phosphorylation of S6K1 by mTOR is one of the final events in its activation. S6K1 phosphorylates several proteins, including rpS6, eIF4B, S6K1 Aly/REF-like target (SKAR), and eukaryotic elongation factor 2 kinase, and therefore affects both the initiation and elongation stages of mRNA translation.
FIGURE 1.. Regulation of signaling through the mammalian target of rapamycin (mTOR) in skeletal muscle. Signaling through mTOR is enhanced by hormones [eg, insulin and insulin-like growth factor I (IGF-1)] and nutrients (in particular leucine) through repression of the GTPase activator activity of the tuberin · hamartin complex toward Rheb (Ras homolog enriched in brain) as described in more detail in the text. mTOR exists in 2 complexes: one containing raptor and GßL (G protein ß-subunit-like protein) and the other containing rictor and GßL. The rictor · mTOR · GßL complex signals to the actin cytoskeleton, whereas the raptor · mTOR · GßL complex signals to several biomarkers of mRNA translation, including 4E-BP1, S6K1, and eukaryotic elongation factor 2 (eEF2) kinase. Lines ending in arrowheads denote activation, whereas lines ending in perpendicular lines denote inhibition. Dashed lines represent incompletely defined steps in a pathway. AMPK, AMP-activated protein kinase; ERK1/2, extracellular signal-regulated kinases 1 and 2; 4E-BP1, eIF4E binding protein 1; PKA, protein kinase A; PKB, protein kinase B; PI 3-kinase; phosphoinositol 3-kinase; rpS6, ribosomal protein S6; S6K1, rpS6 kinase.
The activity of mTOR toward downstream targets such as 4E-BP1 and S6K1 is controlled in part through its interaction with the regulatory associated protein of mTOR (raptor) and G protein ß-subunit-like protein (GßL, or mLST8) (3). Evidence linking raptor with amino acid signaling through mTOR is provided by studies wherein raptor expression is down-regulated by using small inhibitory RNAs (siRNAs) (5, 6). In such studies, leucine-induced phosphorylation of S6K1 is greatly repressed to an extent similar to that observed in cells in which mTOR expression is repressed. Like raptor, GßL has been shown to co-immunoprecipitate with mTOR (6). GßL is a positive regulator of mTOR because coexpression of GßL and mTOR results in greatly increased kinase activity of mTOR toward 4E-BP1 and S6K1 compared with expression of mTOR alone. Moreover, reducing GßL expression using siRNA represses leucine- and serum-induced phosphorylation of S6K1, which suggests that GßL is involved in hormone and amino acid signaling though mTOR.
In yeast, TOR exists in 2 complexes, one with KOG1 (raptor) and LST8 (GßL) and the other with LST8, AVO1, AVO2, and AVO3 (7). The second complex is insensitive to rapamycin and associates with the cytoskeleton. Two recent studies identified a protein referred to as the rapamycin-insensitive companion of mTOR (rictor) as the mammalian ortholog of AVO3 (8, 9). When associated with rictor, mTOR binds to GßL but not raptor. Furthermore, as in yeast, the GßL · mTOR · rictor complex associates with the cytoskeleton and does not phosphorylate 4E-BP1 or S6K1. Both the regulation and the function of this complex are presently poorly understood.
The most proximal upstream protein that has been identified in the mTOR signaling pathway is the Ras homolog enriched in brain (Rheb). Rheb is a small GTPase that enhances phosphorylation of S6K1, rpS6, and 4E-BP1 in an mTOR-dependent fashion when overexpressed (10-14). Moreover, in cells overexpressing Rheb, S6K1 phosphorylation is maintained during starvation for amino acids, which suggests that Rheb is involved in transducing signals from amino acids through mTOR. In this regard, recent studies report that Rheb binds directly to mTOR in an amino acid–dependent manner (15-17). Whether amino acids modulate the binding of GTP to Rheb has yet to be firmly established, with one study reporting that amino acids promote the binding of GTP to Rheb (17) and a second study reporting that they do not (16).
Rheb activity is controlled in part by a GTPase-activating protein referred to as tuberin or TSC2. Tuberin and its binding partner harmartin (also known as TSC1) were originally identified as the products of 2 genes that are causative in the autosomal dominant syndrome tuberous sclerosis (reviewed in references 18-22). Mutations in either gene are associated with the widespread development of benign growths in multiple organs and tissues, which suggests that the normal role of these proteins is to restrict cell size and proliferation. This suggestion has been confirmed in studies in which the Drosophila orthologs of hamartin and tuberin, dTsc1 and dTsc2, respectively, were shown to function in a complex that acts downstream of protein kinase B (PKB, also known as Akt) but upstream of Drosophila TOR (dTOR) to restrict cell growth and proliferation (23-25).
Tuberin has been shown to be directly phosphorylated by PKB on multiple serine and threonine residues, and phosphorylation by PKB represses the inhibitory action of the hamartin-tuberin complex on signaling through mTOR to 4E-BP1 and S6K1 (26-29). Likewise, phosphorylation of tuberin on either Ser1210 (30) or Ser1798 (31) by the MAP kinase regulated proteins, MK2 or p90rsk, respectively, inhibits tuberin and leads to activation of mTOR. Phosphorylation of tuberin through combined activation of the PKB and p90rsk signaling pathways results in a greater inhibition of tuberin activity than either one alone. In contrast, phosphorylation by the AMP-activated protein kinase on distinct residues activates tuberin and results in enhanced repression of signaling through mTOR (32).
Studies in both Drosophila (33) and mammalian cells (34) have implicated tuberin and hamartin in amino acid signaling through mTOR. In Drosophila, down-regulated expression of either dTsc2 or dTsc1 causes cells to become resistant to amino acid deprivation (33). Thus, S6K1 phosphorylation is largely maintained during amino acid starvation in cells with reduced expression of either dTsc2 or dTsc1. Similarly, in mammalian cells lacking either tuberin or hamartin, S6K1 phosphorylation is resistant to amino acid deprivation, although prolonged deprivation reduces S6K1 phosphorylation in tuberin-deficient cells (17). Moreover, in mammalian cells in culture, co-overexpression of tuberin and hamartin prevents amino acid–dependent activation of S6K1 (34). Together, these studies strongly suggest that tuberin and hamartin are required for amino acid–induced signaling through mTOR.
mTOR-independent (rapamycin-insensitive) signaling
eIF4G is phosphorylated on multiple residues that cluster in 2 groups: one in the amino-terminal portion of the molecule and the other in a putative hinge region located in the carboxy-terminal one-third of the protein (35). Phosphorylation of 3 residues (Ser1108, Ser1148, and Ser1192) in the latter portion is enhanced by serum stimulation. Interestingly, in addition to enhancing phosphorylation of certain sites on eIF4G, serum also promotes dephosphorylation of other, as yet unidentified, residues. Rapamycin prevents both the serum-induced phosphorylation of Ser1108, Ser1148, and Ser1192 as well as the decrease in phosphorylation of the unidentified sites, which suggests that mTOR is involved in both events. However, deletion of the amino terminus of the protein engenders rapamycin resistance to the phosphorylation events that occur in the carboxy-terminal one-third of the protein, which indicates that mTOR does not directly phosphorylate these sites. Instead, phosphorylation of residues in the amino-terminal portion of the molecule through an mTOR-dependent process is thought to promote a conformational change that alters the accessibility of residues in the carboxy-terminal portion to phosphatases or kinases other than mTOR.
Furthermore, phosphorylation of rpS6 can occur through both rapamycin-sensitive and rapamycin-insensitive mechanisms. S6K1 and S6K2 promote rpS6 phosphorylation on multiple residues, including 2 clusters, Ser235/236 and Ser240/244, in an mTOR-dependent process (36). p90rsk also phosphorylates rpS6, although only on Ser235/236 (37). In addition, p90rsk phosphorylates the S6K1 substrate elongation factor 2 kinase (38) and in rat liver its activation is associated with increased phosphorylation of eIF4B (39). However, p90rsk is not regulated through the mTOR signaling pathway, but instead is regulated by the MAP kinase pathway through ERK1/2 (40). In addition to regulating p90rsk, ERK1/2 promotes phosphorylation of eIF4E through activation of the mitogen-activated protein kinase-interacting kinases 1 and 2 (MNK1/2) (41). MNK1/2 are also activated by the p38 MAP kinase pathway, which results in enhanced phosphorylation of eIF4E (42). Early studies suggested that phosphorylation of eIF4E on Ser209 enhances the affinity of the protein for the m7GTP cap structure, which results in increased mRNA translation (43). More recent studies show that under a variety of conditions that increase protein synthesis, phosphorylation of eIF4E decreases (44, 45). Moreover, a more recent study reported that phosphorylation of eIF4E results in a decrease in its affinity for the mRNA cap structure (46).
Regulation of the met-tRNAi binding step in mRNA translation also occurs through an mTOR-independent process. In this regard, phosphorylation of the -subunit eIF2 on Ser51 by any of the 4 known eIF2 kinases [mammalian general control nonderepressing 2 (mGCN2), heme-regulated inhibitor (HRI), PKR-like endoplasmic reticulum-associated eIF2 kinase (PERK), and protein kinase R (PKR)] converts eIF2 from a substrate for eIF2B into a competitive inhibitor (47). The activity of eIF2B is also regulated by phosphorylation of the catalytic -subunit of the protein. eIF2B is phosphorylated by at least 4 different protein kinases, including glycogen synthase kinase (GSK) 3, dual-specificity tyrosine-phosphorylated protein kinase (DYRK), casein kinase I (CK-I), and CK-II (48). The consequence of phosphorylation by DYRK, CK-I, or CK-II is incompletely defined. Phosphorylation by GSK-3 is thought to inhibit the guanine nucleotide exchange of the protein, although phosphorylation by GSK-3 alone may be insufficient to cause inhibition.
EFFECTOR MECHANISMS THAT MEDIATE DISCRIMINATION IN THE SELECTION OF mRNAS FOR TRANSLATION
Although most mRNAs are translated through an m7GTP cap-dependent mechanism involving the eIF4F complex, other mRNAs may have a greater or lesser requirement for the eIF4F complex to be efficiently translated. For example, mRNAs possessing long, highly structured 5- noncoding regions (eg, those encoding c-myc, ornithine decarboxylase, and cyclin D1) exhibit a greater requirement for the eIF4F complex than do mRNAs with shorter, less structured 5- noncoding regions (49). Thus, under conditions wherein eIF4F availability is limiting, translation of mRNAs with highly structured 5- noncoding regions is repressed to a greater extent than in the general population of mRNAs. In part, this may reflect a requirement of such mRNAs for the RNA helicase activity of eIF4A as well as the helicase-stimulating activity of eIF4B to unwind secondary structure in the 5- noncoding region and allow binding of the 40S ribosomal subunit to the message as well as its migration to the AUG start codon. Although little is known about the mechanism or mechanisms through which eIF4A and eIF4B are regulated, eIF4B is phosphorylated by S6K1 (50), providing a possible link between hormone and nutrient signaling through mTOR and eIF4A/eIF4B function.
Another example of an effector mechanism that mediates discrimination in the selection of mRNAs for translation is the presence of an uninterrupted stretch of pyrimidine residues (referred to as a TOP motif) at the 5- end of the mRNA (51). Such mRNAs encode proteins that are involved in mRNA translation, including the ribosomal proteins, eukaryotic elongation factors 1A and 2, PABP, and eIF4G. Thus, enhanced translation of TOP mRNAs is one mechanism for increasing ribosome biogenesis and the long-term capacity to synthesize protein. Until recently, it was thought that recruitment of TOP mRNAs into polysomes was the result of activation of S6K1 and subsequent phosphorylation of rpS6. However, studies suggest that activation of S6K1 may not be the only mechanism for enhancing translation of TOP mRNAs (52-54). For example, TOP mRNA translation is regulated normally in cells lacking both S6K1 and S6K2 (37). Moreover, in cells in which the gene encoding the endogenous S6 protein has been mutated so that the 5 known phosphorylation sites are changed to alanine, TOP mRNA translation is regulated normally (55). However, because translation of TOP mRNAs is repressed by the mTOR inhibitor rapamycin, mTOR-dependent, S6K1/S6K2-independent mechanisms must exist for regulating TOP mRNA translation.
Discrimination in the selection of mRNAs for translation can also be mediated through effector mechanisms involving eIF2 and eIF2B. Inhibition of eIF2B represses the translation of most mRNAs, and knockdown of the catalytic -subunit by use of siRNA essentially halts cell growth and triggers apoptosis (56). Similarly, exogenous expression of a dominant-negative eIF2B inhibits isoproterenol-induced hypertrophy in cardiomyocytes (57). Paradoxically, inhibition of eIF2B activity also enhances the translation of a few mRNAs that have open reading frames in the 5- noncoding region (uORF), such as the mRNAs encoding the transcription factors ATF4 (58) and CD36 (59) and the CAT-1 amino acid transporter (60, 61). However, although enhanced translation of mRNAs with uORF sequences has been shown in yeast and in cell lines, a similar phenomenon has not been shown in an intact tissue. Thus, incomplete inhibition of eIF2B caused by partial phosphorylation of eIF2 is another mechanism for altering the pattern of protein expression at the level of mRNA translation.
Finally, another structure that when present in the 5- noncoding region may allow preferential translation when eIF2 is phosphorylated is an internal ribosome entry site (IRES). An IRES allows the ribosome to bind to an internal site in the 5- untranslated region and bypass the normal route of association with the mRNA, ie, binding to the 5- cap structure (reviewed in references 62 and 63). The best characterized IRES-containing mRNA that is regulated by eIF2 phosphorylation is that encoding the cationic amino acid transporter CAT-1 (60). Like many IRES-containing mRNAs, the 5- untranslated region of the CAT-1 mRNA has both an IRES and uORFs, and both elements are required for optimal regulation of its translation. Thus, translation of an uORF adjacent to the IRES promotes a rearrangement of the IRES structure, resulting in its activation. However, this mechanism alone is unlikely to account completely for the enhanced translation of the CAT-1 mRNA because enhanced CAT-1 synthesis is delayed several hours after induction of eIF2 phosphorylation, which suggests that synthesis of another protein might be required for translation of the CAT-1 mRNA.
Proteins that bind to IRES elements and modulate their function are referred to as IRES-transacting factors (ITAFs). Although poorly characterized, it has been suggested that ITAFs function as RNA chaperones that, on binding to the IRES, promote refolding of the domain into the correct structure for 40S ribosome binding. Examples of ITAFs include the polypyrimidine tract binding protein (PTB) and upstream of N-ras (unr) that activate the Apaf-1 IRES (64).
Although eIF2 phosphorylation is one mechanism for enhancing the translation of mRNAs containing IRES, it is not unique. For example, during apoptosis or infection by certain types of viruses, eIF4G is cleaved (65). The normal function of eIF4G is to assemble the translation initiation factors eIF4A and eIF4E and PABP into a complex that mediates the binding of mRNA to the 40S ribosomal subunit. Cleavage of eIF4G separates the binding domain for PABP and the mRNA cap binding protein, eIF4E, from the domains that bind eIF4A and allow ribosome attachment (referred to as the middle fragment of eIF4G or M-FAG). M-FAG generated in etoposide-treated cells promotes the preferential translation of certain, but not all, IRES-containing mRNAs including Apaf-1 and death-associated protein 5 (66). Moreover, several IRES-containing mRNAs are preferentially translated under conditions that promote dephosphorylation or decreased function of eIF4E, for example, when eIF4E is associated with one of the eIF4E-binding proteins such as 4E-BP1. Thus, IRES function can be regulated through multiple mechanisms.
UNIQUE ROLE OF LEUCINE AS A NUTRIENT REGULATOR OF TRANSLATION INITIATION IN SKELETAL MUSCLE
The branched-chain amino acids leucine, isoleucine, and valine are the most abundant of the essential amino acids. In addition to being indispensable for life, the branched-chain amino acids act as nutrient regulators of protein synthesis, protein degradation, and insulin biosynthesis and secretion. Of the 3 branched-chain amino acids, leucine is the most effective with regard to the regulation of these processes.
Studies conducted in our laboratory have established an experimental model system in which the acute effects of leucine on protein synthesis in skeletal muscle can be investigated (67, 68). The model consists of male rats weighing 200 g that are either freely fed (Rodent Chow; Harlan-Teklad, Madison, WI) or deprived of food for 18 h; food-deprived rats are then administered saline, carbohydrate, leucine, isoleucine, valine, or a combination of carbohydrate plus leucine by oral gavage. The amount of leucine administered, ie, 270 mg or 1.35 g/kg, is equivalent to the amount of leucine consumed by rats of the same age and strain during 24 h of free access to food. The same amounts of isoleucine or valine are administered. The carbohydrate administered, ie, 2.63 g, consists of a 50:50 mixture of glucose and sucrose and provides 15% of the daily energy intake by rats of the same age and strain. Initial studies with this model showed that protein synthesis in skeletal muscle (gastrocnemius/plantaris) is reduced in food-deprived rats to 65% of the rate observed in freely fed controls (67). Administration of leucine to the food-deprived rats stimulates protein synthesis to 165% of the untreated control value within 60 min, whereas administration of carbohydrate alone has no effect. Leucine plus carbohydrate has the same effect as leucine alone, and administration of either isoleucine or valine alone has no effect. Thus, leucine is unique among the branched-chain amino acids in regards to its effectiveness as a nutrient regulator of protein synthesis in skeletal muscle. Although leucine administration produces a slight, transient rise in the serum insulin concentration at 30 min, carbohydrate causes a much greater increase that is maintained at 60 min. Thus, the leucine effect cannot be explained by the rise in insulin.
Whether the transient increase in insulin contributes to the leucine effect on protein synthesis was addressed in a separate set of studies in which somatostatin was administered intravenously by primed, constant infusion beginning 60 min before the administration of leucine (69). These studies showed that somatostatin maintains insulin concentrations at the fasting basal level throughout the 60-min time course. It also attenuates the effect of leucine on protein synthesis. Thus, although physiologic increases in serum insulin do not independently stimulate protein synthesis in skeletal muscle of food-deprived rats, a transient increase in the hormone appears to be permissive for the leucine-induced stimulation of protein synthesis.
Additional studies on the relation between insulin and leucine-induced stimulation of protein synthesis in skeletal muscle have been carried out in rats with experimentally induced diabetes (70). These studies showed that protein synthesis in skeletal muscle of diabetic rats is reduced to 35% of the rate observed in food-deprived, nondiabetic controls. Administration of leucine to the diabetic rats stimulates protein synthesis by 50%; however, the rate remains well below that of the untreated, nondiabetic control. The stimulatory response to leucine is enhanced in diabetic rats treated acutely with insulin. However, the recovery of protein synthesis is incomplete, with the rate being equivalent to food-deprived control values but substantially less than values observed in control rats administered leucine. Overall, these studies show that a portion of the protein synthetic response to leucine occurs through an insulin-independent mechanism, because the rate in diabetic rats administered leucine is greater than that of the diabetic controls.
To investigate mechanisms involved in the stimulatory effect of leucine on protein synthesis in skeletal muscle, studies have been designed to assess modulation of 2 key regulatory steps in the initiation of mRNA translation, a process identified in earlier studies to be responsive to both branched-chain amino acids and insulin. These studies show that leucine has no effect on the met-tRNAi binding step in translation initiation, as assessed by the phosphorylation status of eIF2 on Ser51 of its -subunit and by the guanine nucleotide exchange activity of eIF2B (67). In contrast, leucine has a stimulatory effect on assembly of the eIF4F complex, a key component in the mRNA binding step in translation initiation, as assessed by the phosphorylation status of the eIF4E binding protein 4E-BP1 and by the association of eIF4E with 4E-BP1 and eIF4G. Leucine also has a stimulatory effect on the phosphorylation status of eIF4G as well as S6K1 and its downstream substrate S6 (71).
Because phosphorylation of 4E-BP1, eIF4G, and S6K1 is mediated in part by mTOR, the results suggest that leucine stimulates a signaling pathway involving this serine/threonine protein kinase. This suggestion was confirmed in a separate set of studies in which food-deprived rats were injected intravenously with rapamycin, a specific inhibitor of mTOR, before leucine administration (68). These studies showed that rapamycin blocks entirely the effects of leucine on the downstream targets of mTOR signaling and attenuates the stimulatory effect of the amino acid on protein synthesis. Interestingly, these studies also show that isoleucine, but not valine, mimics to some extent the effect of leucine on mTOR signaling, which is an unexpected result because isoleucine has no effect on protein synthesis.
In the studies described in regard to the interaction between the stimulatory effects of leucine and insulin on protein synthesis in skeletal muscle, somatostatin attenuated the leucine-induced increases in 4E-BP1 and S6K1 phosphorylation and completely blocked the increase in S6 phosphorylation but had no effect on the association of eIF4E and eIF4G (69). Moreover, in the studies in which rats with experimentally induced diabetes were used, leucine administered alone had no effect on mTOR signaling to 4E-BP1 or S6K1 but nonetheless stimulated protein synthesis (70). Leucine administered in combination with insulin infusion sufficient to restore the serum concentration of the hormone to a level equivalent to that present in a food-deprived control rat, or approximately twice that level, rescues or even enhances mTOR signaling to 4E-BP1 and S6K1.
Overall, these results suggest that leucine stimulates protein synthesis in skeletal muscle through both insulin-dependent and insulin-independent mechanisms. The insulin-dependent mechanism is associated with signaling through mTOR to 4E-BP1 and S6K1, whereas the insulin-independent effect is mediated by an unknown mechanism that may involve phosphorylation of eIF4G or its association with eIF4E. Support for the latter suggestion was provided by studies performed in a different experimental model system, ie, a perfused hind-limb preparation from postabsorptive rats (71). In this model, raising the leucine concentration in the perfusate from 1X to 10X in the presence of 1X levels of all other amino acids results in a 60–70% increase in protein synthesis that is associated with no change in mTOR signaling to 4E-BP1 and S6K1. Instead, the higher leucine concentration causes increased phosphorylation of eIF4G on Ser1108 and increased association of eIF4G with eIF4E. Thus, the perfused muscle preparation may provide a good experimental model system in which to investigate the insulin-independent effect of leucine on protein synthesis.
The studies described above examined the responses of translation initiation and protein synthesis in skeletal muscle to the administration of rather large amounts of leucine. To assess the responses to lesser amounts of the amino acid, food-deprived rats were administered by oral gavage leucine in amounts ranging from 0.068 to 1.35 g/kg body wt, the latter amount being equivalent to that administered in the studies described above (72). The results show that stimulation of protein synthesis by leucine 30 min after its administration reaches a maximal value of 135% of the untreated control rate with 0.135 g/kg of the amino acid, ie, 10% of the amount used in the previously described studies. The responses in protein synthesis parallel those of eIF4G phosphorylation and eIF4G association with eIF4E, whereas signaling through mTOR to 4E-BP1 and S6K1 continue to increase in proportion to the increasing amounts of leucine administered and its serum concentrations. An increase in the serum insulin concentration is observed only at the highest amounts of leucine administered, ie, 50% and 100% of that used in the previous studies. Signaling through mTOR appears to be enhanced by the increase in insulin but this is not associated with a further increase in protein synthesis. In summary, these studies again show a closer correlation of protein synthesis to eIF4G phosphorylation and its association with eIF4E than with signaling through mTOR to 4E-BP1 and S6K1.
UTILIZATION OF INTACT ANIMAL MODELS TO INVESTIGATE THE REGULATION OF PROTEIN SYNTHESIS IN SKELETAL MUSCLE
The effectiveness of low amounts of leucine in stimulating protein synthesis in skeletal muscle was also shown in a series of studies carried out in collaboration with Teresa Davis at Baylor College of Medicine. In these studies, fasted neonatal pigs were infused intraarterially with leucine to achieve plasma concentrations of the amino acid that were 1.5- to 3-fold basal concentrations (unpublished data, 2005). At the higher infusion rate (ie, 3-fold increase in plasma leucine), protein synthesis was stimulated to 133% of the control rate at 60 min, but returned to the control value at 120 min. The return was probably due to the significant drop in the plasma concentrations of other essential amino acids.
Other studies conducted in collaboration with Dr Davis in which neonatal pigs were used as an experimental model have provided several novel insights into the interaction of hormones and nutrients in the control of protein synthesis in skeletal muscle (73-80). This experimental model is particularly useful because rapid rates of growth and protein deposition, particularly in skeletal muscle, are accompanied by high rates of protein synthesis in neonates. These studies have led to several conclusions, including the following: 1) the developmental decline in the rate of protein synthesis in skeletal muscle from 15%/d in 7-d-old pigs to 4%/d in 26-d-old pigs is accompanied by marked reductions in the activity of eIF2 and the expression or activity of signaling components downstream of phosphoinositol 3-kinase, eg, PKB, mTOR, and S6K1, implying down-regulation of both the met-tRNAi and mRNA binding steps in translation initiation. 2) The decline is accompanied by a blunted response to feeding, ie, a stimulation of the protein synthesis rate from 15% to 24%/d compared with from 4% to 6%/d in 7- compared with 26-d-old pigs, which is due at least in part to the reductions in signaling components and translation initiation factors noted above. 3) Intravenous infusion of insulin or amino acids reproduces the feeding-induced stimulation of protein synthesis, and insulin and amino acids act independently of one another to produce this effect. 4) Feeding, insulin, or amino acids all act in part to stimulate protein synthesis by activating the mTOR signaling pathway to components of translation initiation. Other processes must be involved, however, because the feeding response is only partially inhibited by rapamycin, and under certain conditions, insulin or amino acids stimulate protein synthesis without a change in mTOR-mediated signaling. 5) Compared with adult animals, protein synthesis in skeletal muscle in neonates is much less sensitive to lipopolysaccharide-induced sepsis despite rather marked deactivation of the mTOR signaling pathway to components of translation initiation. 6) Chronic treatment with growth hormone increases the protein synthetic response of skeletal muscle to feeding, which is apparently due to a marked increase in eIF2B activity and thus activation of the met-tRNAi binding step.
In other collaborative work conducted in intact animal models, a study in rats with Jean Grizard in France provided confirmation that both insulin and amino acids are required to reproduce the stimulatory effect of feeding on protein synthesis in skeletal muscle, with the combination of insulin and amino acids acting at least in part to activate the mTOR signaling pathway (81). A study in adult pigs with Robert Wolfe at the University of Texas at Galveston showed that a hemodialysis-induced 40% reduction in plasma amino acid concentrations leads to an inhibition of protein synthesis accompanied by a corresponding fall in eIF2B activity, which suggests a possible role for the met-RNAi binding step in mediating the effect (82).
CONCLUSIONS
The branched-chain amino acid leucine plays a unique role in the regulation of mRNA translation in skeletal muscle. Through activation of both rapamycin-sensitive and rapamycin-insensitive signaling pathways, leucine up-regulates multiple steps in mRNA translation, resulting in acute increases in protein synthesis. The rapamycin-sensitive components of the mTOR signaling pathway are rapidly being identified and characterized. However, the rapamycin-insensitive pathways through which leucine regulates mRNA translation and the downstream effectors of the pathway are still poorly defined. Moreover, the potential role of rapamycin-insensitive mTOR complexes in regulating mRNA translation is relatively unexplored. The protein or proteins that sense leucine availability and the mechanisms through which they act to regulate mRNA translation remain to be identified. Finally, the contribution of leucine-induced changes in the translation of mRNAs encoding specific proteins or families of proteins compared with global changes in mRNA translation warrants further investigation. The work of numerous investigators, many of whom were inspired by the work and ideas of Vernon Young, will help to answer these and other questions related to the regulation of muscle protein synthesis by branched-chain amino acids.
ACKNOWLEDGMENTS
The authors contributed equally to the preparation of the manuscript. Neither author had a personal or financial relation with the sponsoring agency.
REFERENCES