Regulation of mTOR by phosphatidic acid

Abstract

mTORC1, the mammalian target of rapamycin complex 1, is a key regulator of cellular physiology. The lipid metabolite phosphatidic acid (PA) binds to and activates mTORC1 in response to nutrients and growth factors. We review structural findings and propose a model for PA activation of mTORC1. PA binds to a highly conserved sequence in the α4 helix of the FK506 binding protein 12 (FKBP12)/rapamycin-binding (FRB) domain of mTOR. It is proposed that PA binding to two adjacent positively charged amino acids breaks and shortens the C-terminal region of helix α4. This has profound consequences for both substrate binding and the catalytic activity of mTORC1.

Keywords: FRB; Rheb; mTOR; phosphatidic acid; phospholipase D; structure.

Figure 1.

Summary of structural studies used in this review. (A) Twenty-five year timeline of mTORC1 structural data achievement. The first crystal structure of an mTOR domain – the FRB domain, in a ternary complex with rapamycin and FKBP12 was achieved in 1996 at a resolution of 2.7 Å. Approximately 25 years later, the structure of whole mTORC1, as well as several mTORC1 subunits in complex with substrates and regulators have been solved at resolutions that range from 1.75 to 3.4 Å. (B) Study description: reference, species, methodology, and resolution for crystal structures. NMR: nuclear magnetic resonance; CDS: circular dichroism spectroscopy; Cryo-EM: cryo-electron microscopy; SPR: surface plasmon resonance.

Figure 2.

Schematic representation of mTORC1 with data provided from structural studies included in this review. (A) mTORC1 contains the subunits mTOR, raptor and mLST8. Raptor and mLST8 are positioned near the kinase domain of mTOR. The mTOR domains N and M HEAT are the farthest from the kinase domain. The FAT domain of mTOR is positioned between the HEAT and the kinase domain of mTOR. The FRB domain is positioned on top, and restricts access, to the catalytic cleft (hollow area) of the kinase domain of mTOR. Substrate phosphorylation by mTORC1 requires 1) substrate binding to raptor through the TOS motif, 2) substrate binding to the FRB recruitment docking site which positions the nearby substrate target loop correctly into the catalytic cleft, and 3) catalytic cleft phosphorylation of substrate target loop. (B) Detailed representation of the kinase domain of mTOR covered by the FRB domain. The catalytic cleft of the kinase domain (hollow area) is enclosed by the FRB α-helices α1 and α4. The crossing of α1-α4 generates a hydrophobic pocket that starts at residue F2108 of α4. α4 ends at residue P2112 [16]. The first two residues outside the pocket are positively charged and highly conserved – R2109 and R2110 [21]. Spatial organization of the side chains of the two adjacent positively charged amino acids R2109 and R2110 can interfere with α4 stability at the C terminal region that follows the hydrophobic pocket.

Figure 3.

Interactions mediated by the recruitment docking site of FRB positioned above the kinase domain of mTORC1. (A) Interactions with rapamycin, S6K1 and PRAS40. Rapamycin binds to the hydrophobic pocket at the crossing of α1 and α4. Rapamycin inhibits substrate recruitment and induces allosteric changes in the kinase domain. A short α-helix of S6K1 makes few contacts with α4, mainly insertion of L396 at the beginning of the hydrophobic pocket. A long α-helix of PRAS40 makes many contacts with α4, including C-terminal areas immediately outside the pocket. (B) Interactions between PA and FRB. The negatively charged PA headgroup makes ionic interactions with two highly conserved positively charged residues, R2109 and R2110, positioned on α4 immediately outside of the hydrophobic pocket. The hydrophobic acyl chains of PA are inserted into the hydrophobic pocket. The interaction of doubly deprotonated PA (−2) with two adjacent positively charged side chains induces side chain spatial rearrangement that destabilizes the C terminal region of α4. PA breaks α4 immediately after the hydrophobic pocket generating a shorter α4 with consequences for substrate binding and tertiary structure conformational changes that affect the kinase domain. (C) PA binding to FRB allows S6K1, but not PRAS40 or rapamycin binding. PA can accommodate S6K1 insertion of L396 into the beginning of the hydrophobic pocket. However, PA disrupts the C-terminal α4 region that makes several interactions with PRAS40. In addition, occupancy of the hydrophobic pocket by the acyl chains of PA prevents many interactions of the pocket with PRAS40 and rapamycin.

Figure 4.

Schematic representation of Rheb activation of mTOR kinase. (A) The catalytic spine of the kinase domain of mTOR is not well aligned, or broken, in the absence of Rheb [10]. (B) Upon Rheb cooperative interaction with distal N and M HEAT domains, a ripple of conformational changes in mTOR ends up aligning critical residues of the catalytic cleft optimally for phosphorylation [10]. This leads to high catalytic turn-over (k_cat) [10]. (C) PA binding to the FRB domain induces conformational changes that similarly to Rheb, align the catalytic spine of the mTOR kinase. PA production at the lysosome is induced by growth-factor activated Rheb. Upon lysosomal PA generation and binding to mTOR at the FRB, Rheb is no longer required for maximal activity of mTORC1 [5].

Figure 5.

Proposed model of PA activation of mTORC1. PA binding to the FRB domain breaks and shortens the C-terminus of the α4 helix, with important consequences for substrate binding and catalytic activity of mTORC1. PA binding reduces the K_m for S6K1 binding but increases the K_m for PRAS40 and rapamycin. PA binding also induces tertiary structure conformational changes that lead to alignment of the catalytic spine of the kinase domain of mTOR, resulting in increased catalytic turn-over (k_cat). The two effects combined lead high mTOR catalytic efficiency (k_cat /K_m) that no longer requires Rheb. Upon PA binding, a fully activated mTORC1 can leave the lysosome and phosphorylate target substrates elsewhere in the cell. Our model is based on structural evidence available in literature, reviewed here.