Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 Oct;32(10):1980-1988.
doi: 10.1038/s41594-025-01635-0. Epub 2025 Jul 25.

Structural basis for mTORC1 regulation by the CASTOR1-GATOR2 complex

Rachel M Jansen  1   2 , Clément Maghe  1   2 , Karla Tapia  1 , Selina Wu  1 , Serim Yang  1   2 , Xuefeng Ren  1   2 , Roberto Zoncu  3   4 , James H Hurley  5   6   7
Affiliations

Structural basis for mTORC1 regulation by the CASTOR1-GATOR2 complex

Rachel M Jansen et al. Nat Struct Mol Biol. 2025 Oct.

Abstract

Mechanistic target of rapamycin complex 1 (mTORC1) is a nutrient-responsive master regulator of metabolism. Amino acids control the recruitment and activation of mTORC1 at the lysosome through the nucleotide loading state of the heterodimeric Rag GTPases. Under low nutrients, including arginine, the GTPase-activating protein complex GATOR1 promotes GTP hydrolysis on RagA/B, inactivating mTORC1. GATOR1 is regulated by the cage-like GATOR2 complex and cytosolic amino acid sensors. To understand how the arginine sensor CASTOR1 binds to GATOR2 to disinhibit GATOR1 under low cytosolic arginine, we determined the cryo-electron microscopy structure of human GATOR2 bound to CASTOR1 in the absence of arginine. Two MIOS WD40 domain β-propellers of the GATOR2 cage engage with both subunits of a single CASTOR1 homodimer. Each propeller binds to a negatively charged MIOS-binding interface on CASTOR1 that is distal to the arginine pocket. The structure shows how arginine-triggered loop ordering in CASTOR1 blocks the MIOS-binding interface, switches off its binding to GATOR2 and, thus, communicates to downstream mTORC1 activation.

PubMed Disclaimer

Conflict of interest statement

Competing interests: J.H.H. is a cofounder and shareholder of Casma Therapeutics, receives research funding from Hoffmann-La Roche and has consulted for Corsalex. R.Z. is a cofounder and shareholder of Frontier Medicines and a science advisory board member for Nine Square Therapeutics and receives research funding from Genentech. The other authors declare no competing interests.

Figures

Fig. 1
Fig. 1. Cryo-EM structure of GATOR2–CASTOR1 complex.
a, Domain organization of subunits within the GATOR2–CASTOR1 structure. b–d, Composite map and reconstructed model for the GATOR2–CASTOR1 complex viewing from the front face (b), side view (c) and back face (d). Focused maps for different portions of the complex were combined to generate a composite map containing the highest-resolution information for each subunit.
Fig. 2
Fig. 2. CASTOR1 triggers a structural rearrangement in GATOR2 complex.
a,b, Comparison of the front and back faces of the GATOR2–CASTOR1 (a) complex and (b) GATOR2apo complex (b). CASTOR1 is removed for visualization in the GATOR2–CASTOR1 complex. Changes in the MIOS subunits are highlighted in boxes below complex. Key junctions connecting the inner cage are indicated.
Fig. 3
Fig. 3. CASTOR1 interacts with MIOS through negatively charged pocket.
a, Overview of GATOR2–CASTOR1 complex. Front-face MIOS subunits (blue) interact with CASTOR1 (yellow). b, Close-up view of CASTOR1 interaction with MIOS β-propellers. c, Blade diagram for a front-face MIOS β-propeller, highlighting CASTOR1-interacting loops. d,e, Close-up views of the CASTOR1–MIOS interaction shown with CASTOR1 surface view and MIOS ribbon view (d) and CASTOR1 surface colored on the basis of electrostatic potential (e). The units of the scale are kcal (mol·e)−1 at 298 K. f, Ribbon view highlighting specific residues in MIOS loops residues 110–114 and 134–140 (blue) interacting with CASTOR1 residues (yellow). g, HEK293T cells transiently expressing the indicated Flag-tagged wild-type (WT) and MIOS-binding interface (MBI)-mutant CASTOR1 constructs or Flag-tagged METAP2 as a control were starved of arginine for 50 min and, where indicated, restimulated for 10 min. Flag immunoprecipitates (IP) were generated and analyzed by immunoblotting for the indicated proteins. h, HEK293T cells transiently expressing CASTOR1–HA and Flag-tagged wild-type MIOS, Flag-tagged MBI-mutant MIOS constructs or Flag-tagged METAP2 as a control. Cells were starved of arginine for 50 min and, where indicated, restimulated for 10 min. Hemagglutinin (HA) immunoprecipitates were generated and analyzed by immunoblotting for the indicated proteins. i, CASTOR1-knockdown HEK293T cells transiently expressing the indicated Flag-tagged wild-type and MBI-mutant CASTOR1 constructs or Flag-tagged METAP2 as a control were starved of arginine for 50 min and, where indicated, restimulated for 10 min. Anti-HA immunoprecipitates were prepared and analyzed by immunoblotting for the indicated proteins and phospho-proteins. All cell-based assays were performed three times with similar results. Source data
Fig. 4
Fig. 4. CASTOR1 interaction with arginine triggers closing of GATOR2-interacting pocket.
a, Diagram of CASTOR1 interaction with MIOS β-propellers and location of arginine pocket and MIOS-binding interface. b, Electrostatic surface cartoon of CASTOR1apo and close-up view of GATOR2-interact pocket. The units of the scale are kcal (mol·e)−1 at 298 K. Key residues in CASTOR1 that form the pocket are indicated. c, Electrostatic surface cartoon of CASTOR1Arg and close-up view of GATOR2-interacting pocket. The units of the scale are kcal (mol·e)−1 at 298 K. Key residues in CASTOR1 that block the pocket are indicated. d, HEK293T cells transiently expressing the indicated Flag-tagged wild-type and MIOS-releasing loop (MRL)-mutant CASTOR1 constructs or Flag-tagged METAP2 as a control were starved of arginine for 50 min and, where indicated, restimulated for 10 min. Flag immunoprecipitates were generated and analyzed by immunoblotting for the indicated proteins. e, CASTOR1-knockdown HEK293T cells transiently expressing the indicated Flag-tagged wild-type and MRL-mutant CASTOR1 constructs or Flag-tagged METAP2 as a control were starved of arginine for 50 min and, where indicated, restimulated for 10 min. Anti-HA immunoprecipitates were prepared and analyzed by immunoblotting for the indicated proteins and phospho-proteins. f, Overlay of CASTOR1apo (yellow) and CASTOR1Arg (cyan). Rotations in ACT2 and ACT4 α-helices are enlarged for visualization. g, Surface view of CASTOR1apo and CASTOR1Arg arginine-binding pocket. CASTOR1apo is modeled with arginine in the binding pocket. h, Ribbon view of arginine-binding pocket in CASTOR1apo and CASTOR1Arg. i, Overall model for arginine-dependent CASTOR1 interaction with GATOR2. All cell-based assays were performed three times with similar results. Source data
Extended Data Fig. 1
Extended Data Fig. 1. Purification for GATOR2 and CASTOR1.
(a) Chromatogram and gel for GATOR2 purification. (b) Chromatogram and gel for CASTOR1 purification. (c) Chromatogram and gel for Sestrin2 purification. (d) Chromatogram and gel for GATOR1 purification.
Extended Data Fig. 2
Extended Data Fig. 2. Data Processing Pipeline for GATOR2-CASTOR1 complex.
(a) Representative micrograph (b) Representative 2D classes (c) Data processing workflow (d) Overall map for GATOR2-CASTOR1 (e) FSC graph (f) Orientation plot.
Extended Data Fig. 3
Extended Data Fig. 3. Local Refinement for GATOR2-CASTOR1.
(a-g) Local refinement for different sections of complex. Including mask (shown in cyan), FSC graph and resulting map.
Extended Data Fig. 4
Extended Data Fig. 4. Local resolution estimation.
(a) Full complex map (b) CASTOR1-MIOS interface and (c) additional local refinement maps for the complex.
Extended Data Fig. 5
Extended Data Fig. 5. Map to model fit.
(a) Map-model FSC (b) CASTOR1 at CASTOR1-MIOS interface, (c) MIOS at CASTOR1-MIOS interface (d) CASTOR1 residues near arginine binding pocket.
Extended Data Fig. 6
Extended Data Fig. 6. GATOR2 Cage Symmetry.
Comparison of cage symmetry for (a) GATOR2 unbound and (b) GATOR2-CASTOR1 complex. For each complex the individual monomers are reflected over the symmetry axis. Regions distal to the alignments region are enlarged for visualization.
Extended Data Fig. 7
Extended Data Fig. 7. GATOR2 WDR59-MIOS CTD-CTD Junctions.
(a) Comparison of the WDR59-MIOS junctions (black dash circle) on GATOR2-CASTOR1 complex and. GATOR2 unbound. (b) Close up view of the changes to the WDR59-MIOS CTD junctions. GATOR2 unbound (grey) overlayed with the GATOR2-CASTOR1 (blue and red).
Extended Data Fig. 8
Extended Data Fig. 8. Data Processing Pipeline for GATOR2-CASTOR1-Sestrin2.
(a) Representative micrograph (b) Representative 2D classes (c) Data processing workflow (d) Overall map for GATOR2-CASTOR1, FSC graph and orientation plot. (e) Data processing for GATOR1 and representative 2D classes of isolated GATOR1 complex particles.
Extended Data Fig. 9
Extended Data Fig. 9. GATOR2-CASTOR1- Sestrin2 interaction.
(a) GATOR2-CASTOR1 structure docked into cryo-EM map of GATOR2-CASTOR1-Sestrin2. (b) Close up of GATOR2-CASTOR1-Sestrin2 cryo-EM density fitted with Sestrin2-WDR24-SEH1L-SEH1L-MIOS AlphaFold model (c) Full Sestrin2-WDR24-SEH1L-SEH1L-MIOS AlphaFold model (ipTM = 0.69). Close up of interface between WDR24 (green) and Sestrin2 (orange) in AlphaFold model in (d) ribbon view and (e) surface view colored by electrostatic potential. pLDDT for Arg 228 is 0.89. (f) HEK-293T cells transiently expressing HA-tagged SESTRIN2 along with the indicated FLAG-tagged WDR24 constructs or FLAG-tagged METAP2 as a control were starved of leucine for 50 minutes. Where indicated, leucine was added to the lysates during immunoprecipitation. FLAG-immunoprecipitates were generated and analyzed by immunoblotting for the indicated proteins. Experiment performed twice with similar results. Source data
Extended Data Fig. 10
Extended Data Fig. 10. qPCR confirmation shCASTOR1.
qPCR against CASTOR1 performed in HEK293T transduced with a shRNA targeting Luciferase (shLUC) or a shRNA targeting CASTOR1. Data were normalized using ACTB and HPRT1 as housekeeping genes and are presented as the mean ± SEM of a biological triplicate. Unpaired two tailed t-test. ***p = 0.0004. Source data
See this image and copyright information in PMC

Update of

References

    1. Ballabio, A. & Bonifacino, J. S. Lysosomes as dynamic regulators of cell and organismal homeostasis. Nat. Rev. Mol. Cell Biol.21, 101–118 (2020). - PubMed
    1. Goul, C., Peruzzo, R. & Zoncu, R. The molecular basis of nutrient sensing and signalling by mTORC1 in metabolism regulation and disease. Nat. Rev. Mol. Cell Biol.24, 857–875 (2023). - PubMed
    1. Condon, K. J. & Sabatini, D. M.Nutrient regulation of mTORC1 at a glance. J. Cell Sci.132, jcs222570 (2019). - PMC - PubMed
    1. Battaglioni, S., Benjamin, D., Walchli, M., Maier, T. & Hall, M. N. mTOR substrate phosphorylation in growth control. Cell185, 1814–1836 (2022). - PubMed
    1. Chiarini, F., Evangelisti, C., McCubrey, J. A. & Martelli, A. M. Current treatment strategies for inhibiting mTOR in cancer. Trends Pharmacol. Sci.36, 124–135 (2015). - PubMed

MeSH terms

Substances

LinkOut - more resources