Amino Acid Sensing by mTORC1: Intracellular Transporters Mark the Spot

Abstract

Cell metabolism and growth are matched to nutrient availability via the amino-acid-regulated mechanistic target of rapamycin complex 1 (mTORC1). Transporters have emerged as important amino acid sensors controlling mTOR recruitment and activation at the surface of multiple intracellular compartments. Classically, this has involved late endosomes and lysosomes, but now, in a recent twist, also the Golgi apparatus. Here we propose a model in which specific amino acids in assorted compartments activate different mTORC1 complexes, which may have distinct drug sensitivities and functions. We will discuss the implications of this for mTORC1 function in health and disease.

Figure 1

Amino-Acid-Dependent mTORC1 Signaling and Its Subcellular Control (A) Microenvironmental inputs, including specific amino acids (labeled 1, 2, and 3), are integrated by mTORC1 to control metabolic and cellular pathways that drive cell and organismal growth. CAD, carbamoyl-phosphate synthase 2 (Robitaille et al., 2013); TFEB, transcription factor EB; ULK1, Unc-51-like autophagy activating kinase 1. (B) The multi-hub model suggests that functional specificity could be achieved if distinct inputs regulate more than one mTORC1 signaling hub (labeled A and B), possibly in specific subcellular regions. Subunits of the mTORC1 complex are listed below. Note that mTOR-containing mTORC2 includes different components, such as Rictor, and regulates upstream Akt signaling (reviewed in Masui et al., 2014).

Figure 2

Model of Amino-Acid-Dependent mTORC1 Regulation from LELs Schematic depicts GTP/GDP loading of Rag and Rheb G proteins and protein-protein interactions taking place in the presence of amino acids and growth factors, leading to full activation of mTORC1 (green). Inhibitory interactions that happen in the absence of these positive regulators are shown by dotted crossbars. Multiple sensors are indicated in red text: the transporters (red wavy lines) SLC38A9 (SNAT9) and PAT1 (SLC36A1), and the cytosolic sensors leucyl-tRNA synthetase (LRS), folliculin and its binding partner (FLCN-FNIP), Sestrin 2, and CASTOR1 (pink ovals). They respond to specific amino acid inputs to recruit and activate mTORC1 on the surface of LELs. Green arrow indicates the interaction between PAT1 (SLC36A1) and an mTORC1 supercomplex, which is less stable than for SLC38A9. TSC localization on the LEL surface reduces amino-acid- and/or growth-factor-dependent mTORC1 signaling. Topologically equivalent extracellular space and intracellular compartment lumens are indicated in pale blue in this and subsequent figures. Subunits of mTORC1 regulatory components are listed below. Schematic adapted from model presented in Chantranupong et al. (2015).

Figure 3

Growth Regulation by PATs (A) Overexpressing one (PAT ↑) or two (PAT ↑↑) copies of a PAT amino acid sensor in vivo has different effects on growth of developing structures in the fly. Within the postmitotic cells of the compound eye, ommatidia (unit eyes) and overall eye size progressively enlarge. In the wing, overexpression of one copy of the PAT gene also enhances growth, primarily through increased cell proliferation, but two copies reduce it. Similarly, using cultured HEK293 cells, modest human PAT1 overexpression in a stable cell line (PAT ↑) increases mTORC1 signaling and cell proliferation; however, signaling and proliferation decrease with transient (high level) expression, probably through a dominant-negative mechanism. (B) PAT overexpression in the fly eye leads to a mild, but measurable, increase in growth (b versus a), in contrast to loss of PTEN, a PI3K antagonist, which has a pronounced growth-stimulatory effect, disturbing the hexagonal array of ommatidia (d). Increased PI3K signaling significantly enhances PAT-induced growth (e versus b), promoting intracellular localization of a GFP-tagged PAT: GFP-PAT is marked at the cell surface (white arrow) and inside (yellow arrow) larval fat body cells (f versus c; images from Ögmundsdóttir et al., 2012). Scale bar, 100 μm (a, b, d, and e) and 20 μm (c and f).

Figure 4

Amino Acid Transport and Sensing (A) Amino acid transporters were originally classified by their ability to translocate specific groups of amino acids across the lipid bilayer (left). However, some can activate amino-acid-dependent signaling, either in the presence or absence of transport. These so-called “transceptors” may be the precursors to modern-day receptors (right). Black arrows represent amino acid transport and green arrows signal transmission. (B) Amino acid (AA) transport has been studied in vitro using Xenopus oocytes, facilitated by their large size (∼1,000 μm diameter) and little background transport activity; in reconstituted proteoliposomes (up to 500 μm diameter); or in human cells (∼20 μm diameter), which may contain multiple endogenous transporters. Note that in proteoliposomes, external medium may be topologically equivalent to cytosolic side of lipid bilayer, unlike the other two models.

Figure 5

Intracellular Amino Acid Transporters and Multi-Hub mTORC1 Regulation Schematic model in which amino acid transporters (red wavy lines) act as integral components of multiple intracellular sensing supercomplexes (pink spots). These transporters sense the amino acid content of different subcellular compartments, acting in conjunction with cytosolic amino acid sensors to control the activity of specific mTORC1 hubs (green). They appear to respond to the specific amino acids indicated. Other transporters act as conduits (blue rectangles) bringing amino acids into specific compartments: for the plasma membrane, CD98 coupled with SLC1A5; and for LELs, the LAPTM4b-associated CD98 heterodimer. Equivalent conduit-like transporters for the Golgi are not yet identified. Unlike the SLC38A9- and PAT1-regulated LEL-localized sensing complexes, mTORC1 hubs controlled by ARF1 and Rab1A/PAT4 appear to be Rag independent and are therefore not under the control of the cytosolic sensors shown in Figure 2. ARF1 is typically found on Golgi membranes but is reported to control a lysosomal sensing complex. Reactive oxygen species (ROS)-activated, peroxisome-bound TSC blocks activation of mTORC1, but how this message is relayed remains unclear.