These authors contributed equally to this work
The mTORC1 kinase complex regulates cell growth, proliferation, and survival. Because mis-regulation of DEPTOR, an endogenous mTORC1 inhibitor, is associated with some cancers, we reconstituted mTORC1 with DEPTOR to understand its function. We find that DEPTOR is a unique
The mammalian/mechanistic target of rapamycin complex 1 (mTORC1) is a large (~1 MDa) multiprotein complex consisting of two copies of three subunits: the evolutionary conserved Serine/Threonine protein kinase mTOR (a member of the phosphoinositide-3-kinase-related kinases superfamily of protein kinases, PIKKs), RAPTOR and mLST8 (
In addition to its three core components, mTORC1 associates with various subunits that regulate mTORC1 activity or direct its cellular localization: nutrients such as amino acids promote the association of RAPTOR with heterodimeric Rag GTPases, which, along with the Ragulator complex, recruit mTORC1 to lysosomal surfaces. There, the small GTPase RHEB (Ras homolog enriched in brain) in its GTP-bound state activates mTORC1 via direct interaction with the mTOR catalytic subunit (
Interactions of DEPTOR with the mTOR complexes are less clear. DEPTOR is a 46 kDa protein that consists of three distinct and highly conserved regions (from N- to C-terminus): two tandem DEP domains (a DEPt), an unstructured linker of ~100 residues (residues 228–323, which we have named long-linker) and a C-terminal PDZ domain. The long-linker contains multiple serine phosphorylation sites as well as a consensus βTrCP1-binding site, SSGYFS (referred to as the DEPTOR phosphodegron). Previously, it was suggested that DEPTOR inhibits mTOR function in vivo via an interaction of its PDZ domain with mTOR FAT domain (
Under starvation conditions, DEPTOR binds to mTOR and inhibits its kinase activity, whereas under nutrient replete conditions mTOR-dependent phosphorylation at its degron site marks DEPTOR for degradation (
To characterize the mechanism of mTORC1 inhibition by DEPTOR, we carried out reconstituted inhibition assays with purified recombinant mTORC1, DEPTOR and two major mTORC1 substrates, 4EBP1 (wild-type) and S6K1 (GST-tagged S6K1367-404 polypeptide). Using immunoblotting with antibodies specific for the phosphorylated substrates, we found that DEPTOR inhibited mTORC1 with half maximal inhibitory concentration (IC50) of 14 µM for wild-type 4EBP1, and 51 µM for S6K1367-404 (
(
(
Partial enzyme inhibitors bind to the enzyme and decrease its activity, but still allow substrate turnover, even with the inhibitor bound (
To test the possibility that the residual activity is caused by half-site reactivity in mTORC1 via allosteric communication across the dimer interface, we tested DEPTOR inhibition of a monomeric form of mTOR (the mTORΔN-mLST8 complex, which lacks the N-terminal 1375 residues of mTOR and the RAPTOR subunit). The IC50 for the monomeric mTORΔN-mLST8 was 0.4 µM, and the residual activity 33% (
To determine DEPTOR’s mechanism of inhibition of mTORC1 in more detail, DEPTOR deletion variants as well as a 13 S/T→A mutant with most of the phosphorylation sites in DEPTOR removed (
(
(
DEPTOR is phosphorylated in an mTOR-dependent manner in cells, and phosphorylation sites have been identified in the long-linker region (
We next assayed phosphorylation of full-length DEPTOR and its deletion variants via Phos-tag SDS PAGE. To increase the signal of phosphorylated DEPTOR for reliable detection, we used the hyperactive mTORC1 A1459P. The isolated PDZ as well as tandem DEP domains were not phosphorylated at any detectable rate in an mTOR dependent manner (
We determined the structure of mTORC1 bound to full length DEPTOR by electron cryo-microscopy (cryo-EM). Using cross-linked mTORC1/DEPTOR, we generated a 4.3 Å resolution reconstruction of mTORC1 in a complex with DEPTOR (
(
(
(
Data collection | ||
---|---|---|
Protein details | mTORC1 WT/DEPTOR | mTORC1 A1459P/DEPTOR |
Microscope | Titan Krios (FEI) | Titan Krios (FEI) |
Voltage (kV) | 300 | 300 |
Detector | Gatan K2 Summit | Gatan K3 |
Pixel size (Å) | 1.43 | 1.1 |
Defocus range (μm) | −1.6 to −3.2 | −1.4 to −3.0 |
Movies | 2370 | 4759 |
Frames/movie | 20 | 50 |
Exposure rate (e– / Å2/ s) | 2.5 | 22.4 |
Total dose (e– / Å2) | 40 | 56 |
Number of particles | 491,404 | 97,314 |
Energy filter slit width (eV) | 20 | 20 |
Model composition | ||
Non-hydrogen atoms | 28917 (refined as a monomer) | 28784 (refined as dimer) |
Protein residues | 3640 | 7196 |
Ligands/ ions | - | - |
Density refinement | ||
Resolution (Å) | 4.2 | 4.7 |
Sharpening B-factor (Å) | 283.5 | 145.7 |
Model refinement | ||
Root-mean-square deviation | ||
Bond lengths (Å) | 0.0131 | 0.0089 |
Bond angles (°) | 1.73 | 1.76 |
Molprobity score | 1.47 | 1.4 |
Clashscore, all atoms | 0.95 | 1.3 |
Favored rotamers (%) | 98.0 | 89.4 |
Ramachandran plot (%) | ||
Favored | 92.9 | 92.2 |
Allowed | 6.4 | 7.6 |
Outliers | 0.75 | 0.2 |
The density for the DEPTOR PDZ domain was poorly resolved. As crystallization of the 324–409 construct remained unsuccessful, the structure of the DEPTOR PDZ domain was determined using NMR spectroscopy (
We identified a second density adjacent to the mTOR FRB domain, covering the lipophilic patch formed by the FRB domain residues Y2038, F2038, Y2105, and F2108 (
When the structures of free mTORC1 (PDB 6BCX), RHEB-bound mTORC1 (PDB 6BCU) and mTORC1 bound to DEPTOR are aligned locally on the C-lobe of the kinase domain, it is clear that the ATP-binding sites of the free mTORC1 and DEPTOR-bound mTORC1 are very similar and both are distinct from the RHEB-bound conformation (
Typically, PDZ domains bind their targets via the C-terminal tail of the target protein binding in the PDZ αB/βB binding groove (
(
Normalized differential line-broadening analysis of deuterated DEPTOR PDZ domain binding to mTORC1. Ratios of the 2H13C15N PDZ (residues 324–409) peak intensity in the 15N-1H HSQC lower than one indicate line broadening and therefore the binding surface of mTORC1 interaction with isolated DEPTOR PDZ domain (residues 324–409). Measured peak intensity was normalized to the peak of the C-terminal residue.
Because our cryo-EM structure showed a bipartite binding that included the FRB domain, and because the PDZ domain alone is insufficient for mTORC1 inhibition, we attempted to define the region of DEPTOR interacting with mTOR’s FRB. A small stretch of extra cryo-EM density in mTOR's FRB domain suggested an interaction of DEPTOR with this mTOR domain (
(
(
(
To map the regions in DEPTOR interacting with the FRB domain, we carried out NMR experiments with the isolated DEPTOR long-linker and HDX-MS experiments with the full-length DEPTOR. The chemical shift perturbation of the DEPTOR long-linker (residues 228–323) bound to FRB vs. DEPTOR long-linker alone in 1H-15N BEST-TROSY NMR experiments revealed four patches, each of five residues in the long-linker that were altered by FRB binding, residues 228–232, 244–249, 261–264, and 280–285 (
All FRB interacting areas in the DEPTOR linker are within 45 residues of the PDZ domain, which could cover a distance of about 160 Å, if in an extended conformation. As most of the linker showed no indication of secondary structure elements, this length should be sufficient for a bipartite mTORC1 binding interaction involving the DEPTOR PDZ and the long-linker binding either a single mTOR subunit or across the dimer interface (
Secondary chemical shift analysis of the isolated DEPTOR long-linker suggested that the long-linker has no residual secondary structure (
Our structural analysis showed that DEPTOR PDZ binds close to a region on the mTOR FAT domain that undergoes a major conformational change induced by the RHEB-GTP binding (
(
(
(
In order to reconcile our kinetic and binding results suggesting that a cancer-associated, activated mTOR mutant shows increased DEPTOR association, with previous results showing a
(
Our structural and functional analysis of reconstituted mTORC1 inhibition by DEPTOR revealed a unique bipartite binding mechanism, involving an interaction of the DEPTOR PDZ with the FAT domain of mTOR and the DEPTOR long-linker interaction with the FRB. We further identified the long-linker-PDZ as a minimal inhibitory unit. However, our kinetic analysis suggests that the tandem DEP domain region also has a role in inhibition, and work by Wälchli et al. concurrently published with ours suggests a mechanism for this (
Remarkably, DEPTOR only partially inhibits mTORC1 phosphorylation of its two major substrates 4EBP1 and S6K1, independent of the mTOR activation state, while PRAS40 fully inhibits mTOR activity under the identical assay conditions. Partial inhibition of mTORC1 by DEPTOR in cells has been noted (reviewed in
There is evidence that multiple weak interactions may be preferred in biological systems to impart greater specificity (
A major challenge in inhibiting the PI3K/mTOR pathway is the presence of numerous feedback loops that prevent therapeutic success and give rise to resistance against treatment (reviewed in
The discrepancy between the previously reported observation in cells of reduced DEPTOR binding to activated cancer-associated mutants (
We found that phosphorylated DEPTOR no longer inhibits mTORC1. These findings suggest that DEPTOR regulation is performed by mTORC1 itself (
Reagent type | Designation | Source or reference | Identifiers | Additional information |
---|---|---|---|---|
Antibody | Anti-P-4EBP1 (T37/46) (Rabbit, polyclonal) | Cell Signalling | Cat#9459L; RRID: | WB (1:1000) |
Antibody | Anti-4EBP1 (Rabbit, polyclonal) | Cell Signalling | Cat#9452S; RRID: | WB (1:1000) |
Antibody | Anti-P-p70 S6 Kinase (T389) (Rabbit, monoclonal) | Cell Signalling | Cat#9205L; RRID: | WB (1:1000) |
Antibody | Anti-Rabbit IgG, HRP-linked Antibody | Cell Signalling | Cat#7074; RRID: | WB (1:5000) |
Strain, strain background ( | LOBSTR cells | KeraFast | Cat# EC1001 | Chemically competent cells |
Cell line ( | Expi293F cells | Thermo Fisher | Cat#A14527; RRID: | |
Recombinant DNA reagent | pOPL24 (plasmid) | 2xStrepII-1xFlag-Human_Raptor, in pCAG | ||
Recombinant DNA reagent | pOPL25 (plasmid) | Human_mTOR nontagged, in pCAG | ||
Recombinant DNA reagent | pOPL26 (plasmid) | 3xFlag-Human_LST8, in pCAG | ||
Recombinant DNA reagent | pOPL95 (plasmid) | This paper | Human_mLST8 nontagged, in pCAG | |
Recombinant DNA reagent | pOPL119 (plasmid) | This paper | Human_Raptor nontagged, in pCAG | |
Recombinant DNA reagent | pOPL121 (plasmid) | This paper | 2xStrepII(tev)-Human_mTOR, in pCAG | |
Recombinant DNA reagent | pOPL146 (plasmid) | This paper | 2xStrepII(tev)-Human_mTOR_A1459P, in pCAG | |
Recombinant DNA reagent | pOPL151 (plasmid) | This paper | 2xStrepII(tev)-Human_mTOR_1376–2549, in pCAG | |
Recombinant DNA reagent | pOPL107 (plasmid) | This paper | GST(tev)-Human_mTOR_FRB_2015–2114, in pOPTG | |
Recombinant DNA reagent | pOPL40 (plasmid) | This paper | GST(tev)-Human_Rheb1, in pOPTG | |
Recombinant DNA reagent | pMA9 | HisLIP(tev)-Human_PRAS40,in pOPTL | ||
Recombinant DNA reagent | pAB87 (plasmid) | This paper | GST(PreScission)Human_DEPTOR_1–409 (S204, N389 natural variant) (Full length WT), in pGEX-6P1 | |
Recombinant DNA reagent | pOPL111 (plasmid) | This paper | GST(tev)-Human_DEPTOR_228_409 (Linker-PDZ), in pOPTG | |
Recombinant DNA reagent | pOPL127 (plasmid) | This paper | HisLIP(tev)-Human_DEPTOR_324_409 (PDZ), in pOPTL | |
Recombinant DNA reagent | pOPL138 (plasmid) | This paper | GST(tev)-Human_DEPTOR_1–409 (S204, N389 natural variant) (Full length WT), in pOPTG | |
Recombinant DNA reagent | pOPL139 (plasmid) | This paper | GST(tev)-Human_DEPTOR_1-409-13S/T-A mutant (T241A,S244A,T259A,S260A,S263A,S265A,S282A,S283A,S287A, S293A,S297A,S298A,S299A) (Full length 13A), in pOPTG (PCR-ed from Addgene clone 21702) | |
Recombinant DNA reagent | pOPL140 (plasmid) | This paper | GST(tev)-Human_DEPTOR_1–323 (DEP-DEP-linker), in pOPTG | |
Recombinant DNA reagent | pOPL141 (plasmid) | This paper | GST(tev)-Human_DEPTOR_1-323-13S/T-A mutant (DEP-DEP-linker 13A), in pOPTG | |
Recombinant DNA reagent | pOPL143 (plasmid) | This paper | GST(tev)-Human_DEPTOR_228-409-13S/T-A mutant (linker PDZ 13A), in pOPTG | |
Recombinant DNA reagent | pOPL157 (plasmid) | This paper | GST(tev)-Human_DEPTOR_1–220 (DEP-DEP), in pOPTG | |
Recombinant DNA reagent | pOPL159 (plasmid) | This paper | His(tev)-Human_DEPTOR_228–323 (Linker), in pOPTH(tev) | |
Recombinant DNA reagent | pOPL166 | This paper | GST(tev)-Human_DEPTOR_305–409 (extended PDZ), in pOPTG | |
Recombinant DNA reagent | pOPL167 (plasmid) | This paper | GST(tev)-Human DEPTOR_315–409 (extended PDZ), in pOPTG | |
Recombinant DNA reagent | pOPL163 (plasmid) | This paper | GST(PreScission)-Human_S6K1αII,367–404, in pGEX-6P1 | |
Recombinant DNA reagent | pOP826 (plasmid) | This paper | in GST(tev)-Human 4EBP1, in pOPTG | |
Recombinant DNA reagent | pOP854 (plasmid) | This paper | GST(tev)-Human 4EBP1-TOS-less-(114-FEMDI-118/AAAAA), in pOPTG | |
Chemical compound, drug | NEBuilder HiFi assembly Master Mix | New England Biolabs | Cat#E2621S | |
Chemical compound, drug | cOmplete EDTA-free protease inhibitor tablets | Roche | Cat#11873580001 | |
Chemical compound, drug | Universal Nuclease | Pierce | Cat#88702 | |
Chemical compound, drug | Desthiobiotin | IBA | Cat#2-1000-005 | |
Chemical compound, drug | Glutathione Sepharose 4B resin | GE Healthcare | Cat#17-0756-05 | |
Chemical compound, drug | Ni-NTA agarose resin | Qiagen | Cat#30230 | |
Chemical compound, drug | Lysozyme | Sigma | Cat#L6876 | |
Chemical compound, drug | TCEP | Soltec Ventures | Cat#M115 | |
Chemical compound, drug | Bovine Serum Albumin | Thermo Fisher | Cat#BP1605 | |
Chemical compound, drug | PEI (Polyethyleneimine‘MAX’, MW 40,000) | Polysciences | Cat#24765 | |
Chemical compound, drug | SuperSignal West Pico PLUS chemiluminescent substrate | Thermo Fisher | Cat#34577 | |
Chemical compound, drug | LDS sample buffer | NuPAGE | Cat#NP0008 | |
Chemical compound, drug | Triton X-100 | Sigma | Cat#T8787 | |
Chemical compound, drug | Tween 20 | NBS Biologicals | Cat#17767-B | |
Chemical compound, drug | ATP | Jena Bioscience | Cat#NU-1010 | |
Chemical compound, drug | GTPγS | Jena Bioscience | Cat#NU-412–20 | |
Chemical compound, drug | Glutaraldehyde solution, 25% in water | Sigma | Cat#G5882 | |
Chemical compound, drug | D2O | Acros Organics | Cat#351430075 | |
Chemical compound, drug | InstantBlue Protein Stain | Expedeon | Cat#1SB1L | |
Chemical compound, drug | Expi293 Expression Medium | Thermo Fisher | Cat#A1435102 | |
Chemical compound, drug | Yeast Nitrogen Base Without Amino Acids and Ammonium Sulfate | Sigma | Cat#Y1251 | |
Chemical compound, drug | Ammonium15N chloride | Sigma | Cat#299251 | |
Chemical compound, drug | D-Glucose (U-13C699%) | Cambridge Isotopes Laboratores Inc | Cat#CLM-1396–10 | |
Chemical compound, drug | 99.9% 2H atom D2O | Cortecnet | Cat#CD5251P1000 | |
Software, algorithm | ASTRA software package for analysis of SEC-MALS data | Wyatt | ||
Software, algorithm | ProteinLynx Global Server | Waters | 720001408EN; | |
Software, algorithm | DynamX | Waters | 720005145EN | |
Software, algorithm | ImageJ | |||
Software, algorithm | GraphPad Prism | GraphPad Software | ||
Software, algorithm | Topspin | Bruker | ||
Software, algorithm | NMRFAM-Sparky | |||
Software, algorithm | MARS | |||
Software, algorithm | NMRPipe | |||
Software, algorithm | MDD compressed sensing | |||
Software, algorithm | POMONA | |||
Software, algorithm | Rosetta and NMR restraints | |||
Software, algorithm | Random coil shifts | |||
Software, algorithm | RELION three software package | |||
Software, algorithm | Gctf | RELION three software package | ||
Software, algorithm | MotionCor2 | RELION three software package | ||
Software, algorithm | ResMap | RELION three software package | ||
Software, algorithm | Chimera | |||
Software, algorithm | Coot | |||
Software, algorithm | ChimeraX | |||
Software, algorithm | REFMAC5 | |||
Other | 0.45 μM Syringe Filter | Millipore Sigma | Cat#SE2M230I04 | |
Other | 5 μM Minisart Syringe Filter | Sartorius | Cat#17594-Q | |
Other | 5 mL StrepTrap HP column | GE Healthcare | Cat#28-9075-47 | |
Other | 5 mL HiTrap Q HP column | GE Healthcare | Cat#17-1153-01 | |
Other | 5 mL HiTrap Heparin HP column | GE Healthcare | Cat#17-0407-01 | |
Other | 5 mL HisTrap FF column | GE Healthcare | Cat#17-5255-01 | |
Other | HiLoad 16/60 Superdex 200 column | GE Healthcare | Cat#17-1069-01 | |
Other | HiLoad 16/60 Superdex 75 column | GE Healthcare | Cat#17-1068-01 | |
Other | Superose 6 increase10/300 GL | GE Healthcare | Cat#29-0915-96 | |
Other | 4–12% BisTris NuPAGE Protein Gel | Thermo Fisher | Cat#NP0323 | |
Other | SuperSep Phos-tag 7.5% precast gels | FUJIFILM Wako Chemicals | Cat#192–17381 | |
Other | iBlot 0.2 µM pore size nitrocellulose Transfer Stacks | Thermo Fisher | Cat#IB301002 | |
Other | 15 mL Amicon Ultra-153K Centrifugal Filters | Millipore Sigma | Cat#UFC900324 | |
Other | 15 mL Amicon Ultra-1530K Centrifugal Filters | Millipore Sigma | Cat#UFC903024 | |
Other | 15 mL Amicon Ultra-15100K Centrifugal Filters | Millipore Sigma | Cat#UFC910024 | |
Other | 4 mL Amicon Ultra-4 100K Centrifugal Filters | Millipore Sigma | Cat#UFC810024 | |
Other | Enzymate Pepsin Column | Waters | Cat#186007233 | |
Other | Acquity UPLC BEH C18 VanGuard Pre-column | Waters | Cat#186003975 | |
Other | Acquity UPLC BEH C18 column | Waters | Cat#186002346 |
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Roger L. Williams (
The cryo-EM map and the model are deposited with the EMDB (wild-type mTORC1/DEPTOR complex is entry EMD-13099 and A1459P-mTORC1/DEPTOR complex is entry EMD-13097) and PDB (wild-type mTORC1/DEPTOR complex is entry 7OX0 and A1459P-mTORC1/DEPTOR complex entry 7OWG). Backbone assignments of DEPTOR PDZ, mTOR-FRB and DEPTOR linker have been submitted to the BMRB, Biological Magnetic Resonance Bank with the accession numbers 50324, 50325 and 50326, respectively.
For DEPTOR (WT), DEPTOR 13S/T-A, DEPDEP, DEPDEP linker and linker-PDZ, a cell pellet of a 12 L
After cell lysis as described for DEPTOR (WT), the His-lipoyl-tagged DEPTOR PDZ was loaded onto a 5 mL NiNTA column and washed with 75 mL lysis buffer containing 10 mM imidazole prior to eluting the protein with 25 mL lysis buffer spiked with 300 mM imidazole. TEV cleavage was performed overnight as described above. The buffer salt concentration was diluted to 30 mM NaCl and the solution was flown through a 5 mL HiTrap Q HP (GE Healthcare) and a NiNTA column. The flowthrough was collected and concentrated before gel filtration as described for DEPTOR (WT). mTORC1 complexes were expressed by transient transfection of Expi293F cells grown in Expi293 media. A total of 1.1 mg DNA/L cells was co-transfected into cells at a density of 2.5 x106 cells/mL using PEI (Polyethyleneimine ‘MAX’, MW 40,000, Polysciences, 24765, total 3 mg PEI/L cells). After 48 hr, cells were harvested by centrifugation and cell pellets were frozen in liquid N2. Cell pellet from 2 L culture was resuspended in 200 mL of the lysis buffer (50 mM Tris-HCl, pH 8, 500 mM NaCl, 10% glycerol, 1 mM TCEP, 1 mM EDTA, 1 mM EGTA), supplemented with six Complete EDTA-free protease-inhibitor tablets (Roche), 5 µL Universal nuclease (Pierce, 250 U/µL), and 400 µL of a 100 mM PEFA solution, using a Dounce homogenizer (Kontes, 100 mL, Pestle B, small clearance) on ice and sonication (2x 15 s ON at 40% amplitude). The cell lysate was spun at 15,000 rpm for 35 min in a Ti45 rotor, then filtered through Minisart 5 µm filter. Two tandem Strep-Trap HP columns (GE Healthcare 28-9075-48) were equilibrated with lysis buffer, then the filtered lysate was loaded onto the column at a flow of 2.5 mL/min. Extensive washes with lysis buffer (>200 mL), with 50 mL lysis buffer supplemented with 200 mM Li2SO4, and 50 mL of the TEV cleavage buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 10% glycerol, 1 mM TCEP) were performed prior to loading 0.1 mg/mL TEV protease onto the column. The cleavage reaction was performed on the column overnight. The protein was then eluted and the salt concentration was adjusted to 50 mM NaCl. The protein was then loaded onto a 5 mL HiTrap Q column (GE Healthcare) equilibrated with 50 mM HEPES pH 7.5, 50 mM NaCl, and was eluted via a salt gradient. mTORC1-containing fractions were concentrated using an Amicon Ultra-4 100 kDa concentrator, spinning in 3 min intervals at 1000 rcf. Gel filtration was performed using a Superose 6 Increase (10/300) column in gel filtration buffer (50 mM HEPES pH 7.5, 200 mM NaCl, 1 mM TCEP). mTOR complex fractions were pooled and again concentrated using an Amicon Ultra-4 100 kDa concentrator before freezing the protein.
Human RHEB was cloned into a pOPTG vector encoding an N-terminal GST-tag followed by a TEV cleavage site for proteolytic removal. The fusion protein was overexpressed in
Human PRAS40 was purified as described previously (
All reactions were performed in kinase buffer (KB) consisting of 25 mM HEPES pH 7.4, 75 mM NaCl, 0.9 mM TCEP, 5% glycerol, at 30°C for a duration of 45 min for non-activated mTORC1, 2 to 4 min for activated mTORC1, and at 20°C for 20 min for the unstable DEPTOR PDZ construct using non-activated mTORC1. Reactions were set up by preincubating 100 nM non-activated mTORC1 with either 5 µM 4EBP1, 10 µM TOS-less 4EBP1 or 30 µM GST-S6K1367-404 peptide as substrates and various concentrations of the inhibitors for 10 min on ice. After 30 s temperature equilibration at 30°C, reactions were started by the addition of 75 µM ATP and 10 mM MgCl2. For the activated mTORC1 complexes (mTOR-A1459P mutant or in the presence of RHEB-GTP), 30 µM 4EBP1 or GST-S6K1367-404 peptide and 20 nM mTORC1 were used and the reaction was started with 500 µM ATP and 10 mM MgCl2. For RHEB-activated mTORC1, RHEB was preincubated for 1 hr with a 30-fold molar excess of GTPγS (Jena Bioscience NU-412–20, lot IT008-18). Next, mTORC1, DEPTOR, 4EBP1 and RHEB-GTPγS were mixed in the order of mentioning and preincubated for 10 min on ice. After 30 s temperature equilibration at 30°C, reactions were started by the addition of 500 µM ATP and 10 mM MgCl2. All reactions were quenched with 2x SDS sample buffer and resolved on a 4–12% NuPage Bis-Tris gel. Western blots were performed using a 0.2 µM pore size nitrocellulose membrane (Invitrogen IB301002) and the iBlot dry blotting transfer system (Invitrogen). Blocking was performed using 5% Marvel in TBST buffer (100 mM Tris-HCl, 150 mM NaCl, 0.1% Tween 20). Antibodies were obtained from Cell Signalling (P-4EBP1 (T37/46) Rabbit AB, 9459L, RRID:
For the inhibition of mTORC1 A1459P mutant by DEPTOR measured by Phos-tag SDS PAGE, the reactions containing 30 nM mTORC1, 15 µM 4EBP1, 250 mM ATP, 10 mM MgCl2 and varied concentration of DEPTOR were incubated for 4 min at 30°C and then quenched by the addition of 2X SDS sample buffer. The samples were analyzed by SuperSep Phos-tag (50 µM), 7.5% precast gels (192–17381), with MOPS –EDTA running buffer supplemented with 5 mM sodium bisulphate. Staining was done using InstantBlue and quantification followed using the Bio-Rad ChemiDoc-Touch Imaging System.
To analyze phosphorylation of DEPTOR by mTORC1, assays contained 30 nM mTORC1-A1459P mutant, 250 µM ATP, 10 mM MgCl2 and 20 µM of each DEPTOR variant. Phosphorylation was detected by SuperSep Phos-tag precast gels using InstantBlue stain and Bio-Rad ChemiDoc-Touch Imaging System.
To compare the inhibition by DEPTOR vs. phospho-DEPTOR, reactions were set up using a dilution series of DEPTOR (30–0 µM) and 60 nM mTORC1-A1459P mutant. Reaction mix was split in two, one reaction was started by adding 20 mM MgCl2 and 500 µM ATP (to produce phospho-DEPTOR), the second mix was spiked with an equal volume of KB (DEPTOR as a control). After 2 hr, an aliquot of each reaction was quenched in SDS buffer and analyzed by Phos-tag gel for the completeness of mTORC1 dependent DEPTOR phosphorylation (≥80%). At the same time, a second aliquot of each reaction was added into substrate mix resulting in a final concentration of 30 µM GST-S6K, 500 µM ATP, and 20 mM MgCl2 in KB to test for inhibition of phospho-DEPTOR vs. DEPTOR. The reaction was quenched after 2 min, and the results were analyzed as described above by western blots.
All IC50 were determined by the non-linear regression y=ymin + (ymax-ymin)/(1+([DEPTOR]/IC50)).
Purified wild-type mTORC1 (0.7 µM) and DEPTOR (1.4 µM) were mixed in ~300 µL and incubated for 30 min on ice. Following the addition of 0.02% glutaraldehyde (Sigma G5882) from a 1% glutaraldehyde stock in GraFix buffer A (50 mM HEPES pH 7.5, 0.1 M NaCl, 1 mM TCEP, 10% glycerol), the sample was immediately subjected to a gradient fixation (GraFix) (
Purified mutant mTORC1 A1459P (1.2 µM) and DEPTOR (6.8 µM) were preincubated in the presence of 1 mM MgCl2 and 500 µM AMP-PNP (Jena Biosciences) for 20 min and used for cryo-EM grid preparation.
Holey carbon Quantifoil Au R 1.2/1.3 (300 mesh) grids were glow-discharged using an Edwards Sputter Coater S150B for 60 s at 40 mA. The grids were covered with graphene oxide as previously described (
A 3 μL aliquot of freshly-prepared crosslinked mTORC1/DEPTOR complex at OD280 = 0.11 was added to graphene oxide-covered grids and blotted for 11–13 s at 4°C and then plunge-frozen in liquid ethane using a custom-fabricated manual plunger (MRC Laboratory of Molecular Biology). A total of 2370 micrographs of the human mTORC1/DEPTOR complex were acquired on a FEI Titan Krios electron microscope operated at 300 keV. Zero-energy loss images were recorded on a Gatan K2 Summit direct electron detector operated in super-resolution mode with a Gatan GIF Quantum energy filter (20 eV slit width) using SerialEM (
A 3 µL aliquot of mutant mTORC1 A1459P/DEPTOR complex was added onto UltrAuFoil R 1.2/1.3 Au 300 mesh (Quantifoil Micro Tools GmbH) after the grid was glow-discharged using an Edwards Sputter Coater S150B for 60 s at 40 mA. Plunge freezing was performed using Vitrobot (Thermo Fisher Scientific) with a blotting time of 2 s at 14°C and 95% humidity and a force of −15. A total of 4759 micrographs of the mutant mTORC1 A1459P /DEPTOR complex were acquired on a FEI Titan Krios electron microscope operated at 300 keV. Zero-energy loss images were recorded on a Gatan K3 Summit direct electron detector operated in super-resolution mode with a Gatan GIF Quantum energy filter (20 eV slit width) using EPU (Thermo Fisher Scientific) for automated collection. Images were recorded at a calibrated magnification of 81,000 (pixel size of ~1.1 Å) with a dose rate of ~1.12 electrons/Å2/s. An exposure time of 3.25 s was fractionated into 50 movie frames adding to a total dose of 56 electrons/Å2. For data collection, a defocus-range was set to −1.4 to −3.0 µm.
All image-processing steps were done using the RELION three software package (
For the wild-type mTORC1/ DEPTOR data set, 491,404 particles were extracted with a particle box size of 400 by 400 pixels. Two rounds of reference-free 2D classification (using a mask with a diameter of 350 Å) resulted in a selection of 390,636 particles. This set of particles was subjected to a 3D classification over 30 iterations in point group C1 using a low-pass filtered (50 Å) ab-initio reference which was created using the SGD algorithm for de novo 3D model generation introduced in Relion3. Selection of reasonably looking classes by visualization in Chimera (RRID:
To correct for beam-induced particle movements, increase the signal-to-noise ratio for all particles, and to apply radiation-damage weighting, the refined particles were further 'polished' using the Bayesian approach implemented in Relion3. Following this step, another 3D autorefinement in C2 using a mask around the mTORC1/DEPTOR complex as well as applying solvent-flattened FSCs yielded a reconstruction of 4.3 Å resolution (FSC = 0.143 criterion). After a CTF- and beamtilt-refinement for the estimation of per-particle defoci and beamtilt values for the complete set of selected particles, a subsequently performed 3D autorefinement resulted in a mTORC1/DEPTOR reconstruction of 4.2 Å resolution (FSC = 0.143 criterion). A similar processing strategy was performed for the mutant mTORC1 A1459P/DEPTOR complex data set, after automated micrograph assessment using MicAssess (
To improve resolution (especially in the DEPTOR region), we expanded the wild-type mTORC1/DEPTOR dataset using the relion_particle_symmetry expand command while applying C2 symmetry and performed focussed classfication with signal subtraction (
To improve the density map of DEPTOR PDZ region, we performed another focused classification with signal subtraction on the monomer particles. A mask was applied to the region of interest on the PDZ (DEPTOR residues 324–409) and surrounding mTOR domains (N-heat and FAT domain, 61–903 and 1474–1644, respectively), particles were 3D classified without image alignment, and the best class was selected for further refinement of the original (unmasked) particles. This resulted in smaller subsets of the original 223,576 particles; in which the PDZ density was better defined. A 3D refinement of the above selected particles resulted in a map at an overall 4.3 Å resolution, based on the gold-standard FSC = 0.143 criterion (
After correction for the detector modulation transfer function (MTF) and B-factor sharpening, the post-processed map was used for inspection in Chimera (
For the A1459P mutant mTORC1 in a complex with DEPTOR, the wild-type monomer mTORC1/DEPTOR was used as an initial model and placed into a monomer density in the A1459P mTORC1 cryo-EM map with C2 symmetry. The model was adjusted manually and refined using ISOLDE. The ISOLDE-refined monomer was further refined in REFMAC5 using the C2 dimer density, with strict C2 symmetry constraints and with the wild-type monomer for external restraints.
The density for the PDZ domain was clearer in the mTORC1-A1459P/DEPTOR structure than it was in the wild-type structure, probably reflecting a greater affinity of this activated construct for the DEPTOR PDZ. To obtain an unbiased verification of the placement of the PDZ domain we masked out all density due to mTORC1, leaving only density in the region between the FAT domain and the N-heat, and this density was about the size that we would expect for the PDZ domain. We used Chimera to do a global search of this density to find the optimal translation and orientation of the PDZ model in the density. We placed the NMR model in 100,000 initial orientations/translations and carried out a local search for each placement. The initial placements were random orientations and random translations, with the translations restricted so that the center of the PDZ model would be within 10 Å of the center of the putative PDZ density. The orientation/translation that yielded the highest correlation coefficient (0.86) corresponded to the placement that we chose when manually fitting the density. Among 100,000 trials carried out in the global search, this placement was located 4873 times. It is likely that this represents the global maximum correlation placement of the NMR model in the cryo-EM density, since the next most common placement had a lower correlation (0.82) and was reached only 874 times. The lowest correlation coefficient observed in the broad search was 0.73 (found one time in 100,000 trials). The results suggest that our model of the PDZ interaction with mTORC1 is unique and optimal.
Multibody refinement analysis was performed following the protocol described previously (
For NMR experiments, isotopically labeled DEPTOR PDZ and linker-PDZ were expressed in
To produce perdeuterated protein, cells were adapted for growth in deuterated M9 minimal media with 1.7 g/L YNB on agar plates containing 10%, 44% and 78% D2O, before switching to 100% perdeuterated media in a 50 mL starter culture. The starter culture was then innoculated into 1 L M9 minimal media prepared in 100% D2O, supplemented with 1.7 g/L YNB, 1 g/L 14NH4Cl and 4 g/L 2H13C-glucose. Cells were grown at 220 rpm in a shaking incubator and expression were induced with 1 mM IPTG at OD600 = 0.8 at 25°C for 18 hr.
The isolated mTOR FRB domain (residues 2015–2114) was purified from a 6 L culture of C41(DE3)RIPL cells transformed with plasmid pOPL107, grown to OD600 = 0.8 and induced with 0.3 mM IPTG at 16°C for 18 hr. Cells were lysed by sonication in a GST-A buffer (50 mM Tris pH 8, 100 mM NaCl, 1 mM TCEP) supplemented with 0.25 mg/mL lysozyme and 2 µL/100 mL of Universal nuclease. Following ultracentrifugation at 35 k rpm in Ti45 rotor for 35 min, the supernatant was purified by affinity chromatography on the Glutathione-Sepharose 4B beads (GE Healthcare, GE17-0756-05) equilibrated in GST-A buffer. The GST-FRB fusion was eluted with 20 mM glutathione in the same buffer. The eluate was diluted with two volumes of dilution buffer (50 mM Tris pH 8, 1 mM TCEP), loaded on a 5 mL HiTrapQ column and eluted with a 0–1M NaCl gradient. The fractions containing GST-FRB were concentrated in a 30K Ultra-15 concentrator and further purified by gel filtration on a Superdex 75 16/60 column equilibrated in 50 mM HEPES pH 8, 100 mM NaCl, 1 mM TCEP.
The DEPTOR linker (residues 228–323) was purified from a 6 L culture of C41(DE3)RIPL cells transformed with plasmid pOPL159, grown to OD600 = 0.8 and induced with 0.3 mM IPTG at 16°C for 18 hr. The His6-tagged protein was purified by affinity chromatography on Ni-NTA agarose beads (Qiagen 30230) and cleaved on beads with TEV protease o/n at 4°C. The cleaved protein was diluted with 1 vol of 50 mM Tris pH8, 1 mM TCEP (to achieve ~ 50 mM NaCl) and loaded on a 5 mL HiTrapQ column equilibrated in 50 mM HEPES pH8, 25 mM NaCl, 1 mM TCEP. The flow-through fraction containing DEPTOR linker was concentrated in a 3K Amicon Ultra-15 concentrator and further purified by gel filtration on Superdex 75 16/60 column equilibrated in 50 mM HEPES pH 8, 100 mM NaCl, 1 mM TCEP.
All NMR data sets were collected at 278K using Bruker Avance II+ 700 MHz or 600MHz Avance III spectrometers with TCI triple resonance cryoprobes unless otherwise stated. All samples were prepared with 5% D2O as a lock solvent, at pH 8 with 50 mM HEPES and 200 mM NaCl.
1H-15N BEST-TROSY (band selective excitation short transients-transverse relaxation optimized spectroscopy) were collected for all samples using an optimized pulse sequence (
The dynamic properties of the DEPTOR PDZ protein were investigated using standard Bruker 15N T1, T2 and 15N{1H}NOE [heteronuclearNOE] experiments. T1 relaxation times were calculated using delays of 10, 20, 40, 80, 160, 320, 640, 1280, and 2000 ms and T2 relaxation times with delays of 16.9, 33.8, 67.6, 101.4, 135.2, 169.0, 202.8, and 253.5 ms, collected with 3 s relaxation delays. Peak intensities and curve fitting were calculated using Sparky. The heteronuclearNOE experiment was collected as a pseudo 3D spectrum, using a 120° proton pulse train with a 5 ms delay for a total of 5 s, with interleaved on and off resonance saturation. The hetNOE values were calculated from peak intensities according to the equation Ion /I off.
To observe the interaction of DEPTOR PDZ with mTORC1, 32 µM of 2H,13C,15N DEPTOR PDZ was added to 3.2 µM of the mTORC1 complex, with the excess PDZ shifting the equilibrium towards a higher percentage of bound state. Here, binding is observed as a residual effect in the unbound pool of PDZ and is manifested as line broadening in the 1H-15N HSQC experiment when compared to free DEPTOR PDZ only. The 15N-1H HSQC experiment was used here instead of the 1H-15N BEST-TROSY to avoid potential solvent exchange bias. Peaks heights were normalized to the signal of the C-terminal residue before the ratio calculated. Peaks that had reduced relative intensity define the interaction surface for the PDZ domain.
Structural differences between the DEPTOR PDZ domain and the DEPTOR linker-PDZ construct were first identified by chemical shift perturbations in 1H-15N BEST-TROSY experiments. In the absence of a complete linker-PDZ assignment each signal of the linker-PDZ spectrum from a 114 μM sample was compared with that of the assigned PDZ domain collected under the same conditions, giving a weighted chemical shift perturbation calculated by
Both the mTOR-FRB domain and the DEPTOR linker construct were assigned in order to identify residues involved in the binding interaction. FRB assignments were obtained at 293K and transferred to the 278K spectra using a temperature titration. The backbone assignment of the 70 µM FRB sample (residues 2015–2114) was completed using 3D HNCO, HN(CA)CO and HNCACB, CBCA(CO)NH experimental pairs - all collected with 1024, 64, and 96 points in the proton, nitrogen and carbon dimensions respectively and with 20–40% non-uniform sampling (NUS). Data were processed using NMRPipe (
Secondary chemical shift analysis to describe conformational preferences for the DEPTOR linker was based on a comparison of assigned backbone carbon α and β shifts with chemical shift values expected for a random coil protein with the same sequence and under the same experimental conditions (pH and temperature), calculated using methods described previously (
Binding of the DEPTOR linker to mTORC1 FRB was observed by 1H-15N BEST-TROSY NMR. FRB residues involved in the binding were identified by the addition of up to 220 µM of unlabeled linker-PDZ to 40 µM of 15N-labeled FRB. Similarly, DEPTOR linker residues involved in binding were identified by the addition of up to 320 µM of unlabeled FRB to a 40 µM 15N-labeled DEPTOR linker sample. Data from both titrations were analyzed using the above equation.
Experiments followed suggested standards by the HDX-MS community (
The HDX binding study of FRB and DEPTOR was performed by preincubating 100 µM FRB with 100 µM DEPTOR in buffer (50 mM HEPES pH 8.0, 100 mM NaCl, 1 mM TCEP) at room temperature for 1 hr. An aliquot of 5 µL was incubated with 45 µL of D2O buffer at room temperature for 3, 30, 300, and 3000 s, the reaction was quenched and treated as described, with the exception of using a 5–36% gradient of acetonitrile in 0.1% v/v formic acid for elution from Acquity UPLC BEH C18 column. Data were collected from 300 to 2000 m/z, and mass analysis was performed as described above. Deuterium incorporation was not corrected for back-exchange and represents relative, rather than absolute changes in deuterium levels. Changes in H/D amide exchange in any peptide may be due to a single amide or a number of amides within that peptide. All time points in this study were prepared at the same time and individual time points were acquired on the mass spectrometer on the same day.
Twin-Strep-tagged wild-type and activated mTORC1 mutant A1459P were purified as described above with the exception that no TEV-protease was added, and the protein was eluted from the Strep-Trap HP columns using 10 mM desthiobiotin in 40 mL elution buffer prior loading onto a 5 mL HiTrap Q column.
SPR was performed using a Biacore T200 using CM5-sensor chips (Cytiva). Both reference control and analyte channels were equilibrated 50 mM HEPES pH 7.5, 100 mM NaCl, 1 mM TCEP. Twin-Strep-tagged mTOR was captured onto a Strep-Tactin XT (IBA Lifesciences) coated surface prepared according to the supplied instructions. SPR runs were performed with analytes injected for 120 s followed by a 300 s dissociation in a 1:2 dilution series with initial concentrations of 20 µM for DEPTOR PDZ (residues 305–409). After reference and buffer signal correction, sensogram data were fitted using GraphPad Prism (RRID:
We acknowledge the MRC - Laboratory of Molecular Biology Electron Microscopy Facility for access and support of electron microscopy sample preparation and data collection. We thank Christos Savva, Rangana Warshamanage, Domagoj Baretić, Xiao-chen Bai, Rafael Fernández-Leiro and Sjors Scheres for expert assistance with cryo-EM data collection and processing. We thank Jake Grimmet and Toby Darling for implementing and maintaining the scientific computing infrastructure at the MRC LMB. We thank Christopher Johnson for training for the differential scanning fluorimetry and maintaining the Biophysics facility at the MRC LMB. We thank Yohei Ohashi for help with the initial mTORC1 purification. We thank Jaslyn Wong for providing purified mTORC1 A1459P and RHEB. MA was supported by FEBS fellowship and EMBO Advanced Fellowship (EMBO ALTF 603–2019). The work was supported by the Medical Research Council (MC_U105184308 to RLW) and Cancer Research UK (grant C14801/A21211 to RLW).
No competing interests declared
Conceptualization, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - original draft, Writing - review and editing
Conceptualization, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - original draft, Writing - review and editing
Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - review and editing
Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - review and editing
Resources, Investigation, Methodology, Writing - review and editing
Formal analysis, Investigation, Methodology, Writing - review and editing
Formal analysis, Investigation, Methodology, Writing - review and editing
Investigation, Methodology, Writing - review and editing
Formal analysis, Investigation, Methodology, Writing - review and editing
Formal analysis, Investigation, Methodology, Writing - review and editing
Formal analysis, Investigation, Methodology, Writing - review and editing
Software, Formal analysis, Investigation, Methodology, Writing - review and editing
Software, Formal analysis, Supervision, Methodology, Writing - review and editing
Formal analysis, Supervision, Investigation, Methodology, Writing - review and editing
Supervision, Investigation, Methodology, Writing - review and editing
Conceptualization, Resources, Formal analysis, Supervision, Funding acquisition, Visualization, Project administration, Writing - review and editing
The cryo-EM map and the model are deposited with the EMDB and PDB, respectively. Backbone assignments of DEPTOR PDZ, mTOR-FRB and DEPTOR linker have been submitted to the BMRB, Biological Magnetic Resonance Bank with the accession numbers 50324, 50325 and 50326, respectively.
The following datasets were generated:
Our editorial process produces two outputs: i)
DEPTOR is a regulator of the central mTORC1 and mTORC2 kinase complexes that is both of general interest to biologists and has remained poorly understood despite many years of investigation. Two
Thank you for submitting your article "Bipartite binding and partial inhibition links DEPTOR and mTOR in a mutually antagonistic embrace" for consideration by
The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.
Essential revisions:
1. mTORC1 phosphorylates DEPTOR in the linker region, and the authors show that DEPTOR pre-phosphorylated by activated mTORC1 (phospho-DEPTOR) does not inhibit mTORC1. Hence, the summary model shown in Figure 7B proposes "mTORC1 activity alone is sufficient to reduce DEPTOR inhibition" through direct phosphorylation of DEPTOR's linker. Does this model predict that the 13A mutant constructs of DEPTOR that are not phosphorylated should be more potent inhibitors of mTORC1 – because these constructs will not be subject to phosphorylation-dependent negative feedback? The authors show that the 13A constructs are inhibitors with similar properties to the wild-type sequence. Please comment.
2. Understanding the structural basis of DEPTOR linker binding to the FRB domain of mTORC1 is critical to explaining why the longer linker is necessary for DEPTOR's partial inhibition. The NMR experiments to corroborate the cryoEM density interpretation, however, raise questions. First, regarding the chemical shift potentials (CSPs), please cite the basis for the weighting term. It appears the authors did not divide by two before taking the square root. Does omitting this step increase the apparent CSPs (which are already small)? Related, is there a precedent for using "minimal" CSP maps? If so, please include citations. I am not an expert but it seems dubious to calculate CSPs compared to unassigned spectra. Also, please annotate the CSP figure legends to indicate the "Max" std value for the scale since the CSPs for the PDZ domain appear to be more robust than for the linker.
3. In Figure 5—figure supplement 1, please overlay the red on top of the blue and zoom in on a few data callouts in detail, showing fewer contour levels. As currently depicted, it is challenging to appreciate peak shifts. Also, DEPTOR linker residue 262 is highlighted in the Figure 5B comparison – but is it aliased with F261, F290, and V309? I can't tell from the figure if the peak centers are distinct enough, and I believe aliased peaks should be excluded from the CSP analysis.
4. Figure 5C is confusing. The FRB domain of mTORC1 appears to protect surfaces from solvent exchange throughout full-length DEPTOR (and some regions more strongly than the putative linker binding region).
5. It is concerning that the cryo-EM reconstruction of the A1459P mTORC1 mutant bound to DEPTOR failed to recover any putative linker density bound to the FRB helices (Figure 6C). Integrating all of the above, and in the absence of structure-based point mutants that impair linker binding to the FRB and partial kinase inhibition, the mechanism by which the linker tunes mTORC1 activity remains unclear.
6. In several panels, the phosphor signal is too saturated to appreciate the extent of DEPTOR-dependent inhibition (e.g. 1A, 2B). Shorter exposures should be shown, instead of or along with the current blots.
7. In the section titled "Cryo-EM structure of mTORC1/DEPTOR reveals a bipartite binding mode of DEPTOR to mTOR," the authors indicate specific residues on the PDZ domain of DEPTOR that bind to mTORC1's FAT domain. The placement of this statement suggests that this information is garnered from docking the PDZ NMR structure into the cryoEM map. From the data presented, however, docking the PDZ domain unambiguously into the density of the cryoEM map seems challenging due to the poor resolution of the peripheral densities. Please comment.
8. In figure 2C it would be useful to see the WT DEPTOR plot. The western blot appears to show higher inhibition is achieved with the Linker+PDZ 13A domain as compared to WT DEPTOR or DEPTOR 13A. Do the authors have any models of what would give rise to this heightened effect with the minimal domain?
Essential revisions:
1. mTORC1 phosphorylates DEPTOR in the linker region, and the authors show that DEPTOR pre-phosphorylated by activated mTORC1 (phospho-DEPTOR) does not inhibit mTORC1. Hence, the summary model shown in Figure 7B proposes "mTORC1 activity alone is sufficient to reduce DEPTOR inhibition" through direct phosphorylation of DEPTOR's linker. Does this model predict that the 13A mutant constructs of DEPTOR that are not phosphorylated should be more potent inhibitors of mTORC1 – because these constructs will not be subject to phosphorylation-dependent negative feedback? The authors show that the 13A constructs are inhibitors with similar properties to the wild-type sequence. Please comment.
The reviewer has raised a point that could cause some confusion and we have tried to clarify it. Phosphorylated DEPTOR is not an inhibitor of mTORC1. However, our assays were carried out in steady state conditions. This means that we are measuring initial rates during which there is no appreciable depletion of substrate and the appearance of product is linear with time. Under these conditions, there will be very little appearance of phosphorylated DEPTOR relative to unphosphorylated DEPTOR. It is for this reason that WT and the 13 S/T-A mutant of DEPTOR have similar steady-state properties. Under these conditions, the WT DEPTOR is not significantly phosphorylated. DEPTOR’s characteristics as an mTORC1 substrate are shown in Figure 2—figure supplement 1. This shows that after 15 min only a small fraction of DEPTOR is phosphorylated. Assays of inhibition of DEPTOR constructs by mTORC1-A1459P were performed over a 4 min time interval. In order to observe the loss of inhibition due to phosphorylation of DEPTOR, we had to pre-phosphorylate DEPTOR by incubating it with mTORC1 for a longer time. In order to circumvent any potential confusion, we have expanded our explanation regarding this point in the revised manuscript on page 17.
2. Understanding the structural basis of DEPTOR linker binding to the FRB domain of mTORC1 is critical to explaining why the longer linker is necessary for DEPTOR's partial inhibition. The NMR experiments to corroborate the cryoEM density interpretation, however, raise questions. First, regarding the chemical shift potentials (CSPs), please cite the basis for the weighting term. It appears the authors did not divide by two before taking the square root. Does omitting this step increase the apparent CSPs (which are already small)? Related, is there a precedent for using "minimal" CSP maps? If so, please include citations. I am not an expert but it seems dubious to calculate CSPs compared to unassigned spectra. Also, please annotate the CSP figure legends to indicate the "Max" std value for the scale since the CSPs for the PDZ domain appear to be more robust than for the linker.
The Euclidian weighing term used in the NMR analysis is from the following references that we have added to the manuscript (Amin et al., 2013; Rowe et al., 2009). There are a number of different ways to report weighted chemical shifts in the literature (Shuker et al., 1996), and here we report the Euclidian distances as opposed to the average as suggested by Williamson (Williamson, 2013). There is a precedent in the literature for the minimal shift maps, and we have added these references to the manuscript (Muskett et al., 1998; Williamson et al., 1997). We feel that the minimal shift map is an important tool for describing differences between states where one state is not assignable. The complex spectra of the DEPTOR long linker-PDZ could not be assigned due to the different dynamic qualities of the folded and disordered parts of the construct. As the assignment of the PDZ alone is known, it is reasonable to assume that peaks shifted in relation to the long-linker PDZ protein are shifted as a result of a change in the local environment as a consequence of the presence of the long linker. The minimal map identifies the likely peak locations in the unassigned spectrum in a conservative approach that can only underestimate the maximal chemical shift perturbation. It is true that line broadened peaks are not identified in the minimal map, however, this analysis will highlight line-broadened peaks and as a result residues that are altered in the new sample conditions.
The analysis of CSP values in absolute terms in not straightforward due to the nature of the residues involved, and as a result, comparing the extent of CSP between systems is highly problematic. With this is mind we have not added the “Max” Std values to the CSP legends.
3. In Figure 5—figure supplement 1, please overlay the red on top of the blue and zoom in on a few data callouts in detail, showing fewer contour levels. As currently depicted, it is challenging to appreciate peak shifts. Also, DEPTOR linker residue 262 is highlighted in the Figure 5B comparison – but is it aliased with F261, F290, and V309? I can't tell from the figure if the peak centers are distinct enough, and I believe aliased peaks should be excluded from the CSP analysis.
We have altered the figure as suggested, and included a couple of zoomed in peaks in the figure. The zoom that shows residue M262 highlights that the peak position is unambiguous and so can be included in the analysis.
4. Figure 5C is confusing. The FRB domain of mTORC1 appears to protect surfaces from solvent exchange throughout full-length DEPTOR (and some regions more strongly than the putative linker binding region).
Our HDX-MS results show that in addition to the long-linker, the N-terminal region of DEPTOR (peptide 2-32) shows substantial protection from solvent exchange in the presence of the FRB domain of mTORC1. We did not comment on the changes in this region, since we know that they are not essential for either the DEPTOR-mediated inhibition of mTORC1 or the partial inhibition unique to DEPTOR. However, we can now provide a structural hypothesis for the origin of these changes in protection in the N-terminus of DEPTOR, since while our manuscript was in review, another manuscript describing the structure of the tandem DEP domains at the N-terminus of DEPTOR was published (Weng et al., JMB; 2021, PMID 33865870). In this structure, a short segment (228-235) from the beginning of what we refer to as the DEPTOR long linker (228-323) forms a short helix that interacts with residues in a segment 20-32 at the N-terminus of the first DEP domain. This N-terminal region becomes more protected upon interaction with the mTOR FRB. It is plausible that the FRB interacting with the long linker promotes ordering of the linker and thereby protection of the N-terminal 2-32 peptide, which contacts the long linker. This also might account for the lower IC50 of the full-length DEPTOR compared with the construct lacking the tandem DEP domains (Figure 2C). We have now briefly commented on this in the text on page 14.
5. It is concerning that the cryo-EM reconstruction of the A1459P mTORC1 mutant bound to DEPTOR failed to recover any putative linker density bound to the FRB helices (Figure 6C). Integrating all of the above, and in the absence of structure-based point mutants that impair linker binding to the FRB and partial kinase inhibition, the mechanism by which the linker tunes mTORC1 activity remains unclear.
We have a reasonable explanation for the absence of the DEPTOR linker density at the FRB for the mTORC1-A1459P construct bound to DEPTOR. The mTORC1-A1459P/DEPTOR sample was not cross-linked. When we do not cross-link, the apparent helical density bound to the FRB is not evident. This is not especially surprising given the weak affinity of this FRB site for DEPTOR and other substrates. The original cryo-EM study of the mTORC1 complex by Pavletich and his colleagues (Yang et al., Nature 2017) also showed no density on the FRB in the presence of the substrate 4EBP1, although the kinetic results in the same study show the importance of 4EBP1 interaction with the FRB for maximal rate. The structural demonstration of this interaction with the FRB was shown by fusing the substrate to the FRB and crystallising the fusion FRB construct (Yang et al., Nature 2017). The fusion construct was one way to capture experimentally the interaction. In our work with wild-type DEPTOR, we have captured this interaction by crosslinking. Although this is a weak interaction, it is a classic observation that weak interactions can have profound influences on the overall affinity of protein/protein interactions. This is clear from the thermodynamics of independent interactions, and there is strong evidence that multiple weak interactions may be preferred in biological systems to impart greater specificity. This has been nicely demonstrated by the recent work of Scheepers et al. (PNAS 117, 22690–22697; 2020, PMID 32859760).
It is not clear what the reviewer means by the statement “Integrating all of the above…” Our NMR study characterises the interaction of the DEPTOR linker with the FRB. The HDX-MS is also consistent with this interaction. We did not make point mutations in the DEPTOR linker, because it is clear from the NMR study that there are multiple regions in the linker that can interact with the FRB. We did, however, delete the linker and show that we lose all inhibition. We did not make mutations in the FRB, because it has been shown previously that this site is important for substrate interaction. Mutagenesis would have been far less informative than the extensive work we did to characterise the interactions between DEPTOR and the FRB.
6. In several panels, the phosphor signal is too saturated to appreciate the extent of DEPTOR-dependent inhibition (e.g. 1A, 2B). Shorter exposures should be shown, instead of or along with the current blots.
This was a good suggestion and the figures 1 and 2 have been revised. Both the long and short exposures are now presented, and this makes our conclusions more visually obvious.
7. In the section titled "Cryo-EM structure of mTORC1/DEPTOR reveals a bipartite binding mode of DEPTOR to mTOR," the authors indicate specific residues on the PDZ domain of DEPTOR that bind to mTORC1's FAT domain. The placement of this statement suggests that this information is garnered from docking the PDZ NMR structure into the cryoEM map. From the data presented, however, docking the PDZ domain unambiguously into the density of the cryoEM map seems challenging due to the poor resolution of the peripheral densities. Please comment.
The reviewer is correct. We garnered our initial assessment of residues interacting with the PDZ domain by examining the residues in the model that were closest to the mTOR after fitting the NMR-derived model of the PDZ domain into the cryo-EM density. Subsequent examination of CSP and line-broadening were consistent with this initial estimation. Although the resolution of the cryo-EM density is limited, the fit of the PDZ domain to the density is unique. To show this, we created a new map for the A1459P mutant in which we masked out all density due to mTORC1. This left density only in the region between the FAT domain and the N-heat, and this density was about the size that we would expect for the PDZ domain. We used Chimera to do a global search of this density to find the optimal translation and orientation of the PDZ model in the density. We placed the NMR model in 100000 initial orientations/translations and carried out a local search for each placement. The initial placements were random orientations and random translations, with the translations restricted so that the center of the PDZ model would be within 10 Å of the center of the putative PDZ density. The orientation/translation that yielded the highest correlation coefficient (0.86) corresponded to the placement that we chose when manually fitting the density. Among 100000 trials carried out in the global search, this placement was located 4873 times. The next most common placement had a lower correlation (0.82) and was reached only 874 times. Any random placement with the PDZ domain volume occupying the volume of the density will give a correlation coefficient that is statistically significant (different than zero), but they all have correlation less than the correct solution that is consistent with the NMR measurements. The lowest correlation coefficient observed in the broad search was 0.73 (found one time in 100000 trials). From this unbiased analysis, we believe that our model for the PDZ interaction with mTORC1 is unique and correct. We have briefly described this unbiased analysis in the methods on page 48.
8. In figure 2C it would be useful to see the WT DEPTOR plot. The western blot appears to show higher inhibition is achieved with the Linker+PDZ 13A domain as compared to WT DEPTOR or DEPTOR 13A. Do the authors have any models of what would give rise to this heightened effect with the minimal domain?
The IC50 is greater for the linker-PDZ 13A compared with full-length DEPTOR, suggesting that the truncated construct is a poorer inhibitor than the full-length construct. It is probably a poorer inhibitor because it binds more weakly to mTORC1 (perhaps due to a change in the order of the linker due to loss of its interaction with the tandem DEP domains as referred to in point 4 above). When we include the WT DEPTOR plot in Figure 2C as suggested by the reviewer (see
We do not have a complete structural understanding of how partial inhibition arises. We know only that it requires the linker and that it involves interactions with the FRB. Since the same region on the FRB is also involved in interacting with substrates and the inhibitor PRAS40, we have a rough idea of what it looks like structurally. However, because substrate turnover can still take place with DEPTOR bound (the definition of partial inhibitor), we know that it does not completely overlap with the substrate interaction site. We also know it is distinct from PRAS40, which is a complete inhibitor that prevents substrate binding. Given that we do not have a full understanding of the structural origin of DEPTOR’s partial inhibition, we kept the speculations about it to a minimum. However, in response to the reviewer’s question, we have added some speculation on page 19 in the discussion.