SUMO modifies GβL and mediates mTOR signaling

GβL, the constitutively bound regulatory subunit of mTOR, is modified by SUMO

To understand the role of SUMOylation on mTOR activity, we first tested whether any mTOR components can be SUMO modified. To do so, we transiently overexpressed HA-GβL, myc-Rictor, myc-Raptor, FLAG-PRAS40, and FLAG-mTOR in the presence or the absence of His-SUMO1 in human embryonic kidney 293 (HEK293) cells, followed by nickel–nitrilotriacetic acid (Ni–NTA) denaturing enrichment (38). Since SUMOylated proteins exist at low stoichiometry and SUMO proteases rapidly remove SUMOylation in native lysates (39), this strategy helped to prevent cleavage of SUMO from corresponding target substrates and allowed for further enrichment of potential SUMO substrates. As shown in Figure 1, high molecular weight conjugates of GβL are enriched only in the presence of His-SUMO1 (S∗ GβL, Fig. 1A), starting with a primary band at ∼49 kDa that is consistent with the ∼11 kDa SUMO1 modification size. SUMO1 lacking the essential diglycine conjugation motif (HisSUMO1-AA), which cannot be covalently attached to target lysines (40), failed to enrich high molecular weight species of GβL (Fig. 1A). This result suggests that SUMO modification on GβL is specific and requires the covalent attachment of SUMO1 to GβL. Preliminary data also indicated that GβL is heavily modified by SUMO1 compared with other mTOR components (mTOR, Raptor, Rictor, or PRAS40; Fig. 1, B–D). Furthermore, high molecular weight conjugates of GβL are enriched in the presence of His-SUMO1, His-SUMO2, and His-SUMO3, suggesting all SUMO isoforms can modify GβL (Fig. 1E). Together, these biochemical observations reveal that GβL, a primary core component of mTOR, is strongly modified by SUMO1, SUMO2, and SUMO3.

GβL is SUMO modified at K86, K215, K245, K261, and K305

To identify the potential lysine residue(s) on GβL that are modified by SUMO1, we performed lysine to arginine (K to R) site-directed mutagenesis. GβL contains eight surface-exposed lysine residues distributed across the primary sequence (K86, K158, K213, K215, K245, K261, K305, and K313; Fig. 2, A and B) (41), with three predicted SUMO consensus motifs (K158, K261, and K305). To identify potential lysine modification sites, we generated single, double, quadruple, and full KR mutants of GβL. First, we transfected HEK293 cells with WT HA-GβL or different KR HA-GβL mutant constructs together with His-SUMO1, followed by enrichment of SUMO conjugates using Ni–NTA pulldown. If one or more of these lysines are conjugated by SUMO1, then arginine substitution should prevent conjugation, and Ni–NTA enrichment of the corresponding construct should be impaired compared with WT GβL. Mutagenesis of single lysines showed a varied degree of deficit in GβL SUMOylation (Fig. 2C). The deletion of SUMO consensus sites (K158, K261, and K305) did not abrogate GβL SUMOylation, indicating that either (1) SUMOylation shifted to nonconsensus sites in the mutants or (2) GβL SUMOylation can occur on nonconsensus sites. Double and quadruple lysine mutation sites revealed that the predominant SUMO1 acceptor sites on GβL are K213, K215, K305, and K313 (Fig. 2D), indicated by the reduced accumulation of high molecular weight conjugates when these sites were mutated. Mutation of all lysine sites (HA-GβL full KR) eliminated the SUMO modification of GβL (Fig. 2E), indicating that SUMOylation of GβL is dynamic and fluid in nature. To better elucidate potential SUMO modification sites on GβL, we performed mass spectrometry (MS) on Ni–NTA pulldowns of HEK293 cells transfected with HA-GβL and His-mSUMO3 (compatible for MS detection (42, 43, 44)) and confirmed that GβL is SUMO3 modified at residues K86, K215, K245, K261, and K305 (Fig. 2F and Data S1). Our previous study also identified that GβL is SUMOylated in HEK293 cells stably expressing SUMO3 (43). Furthermore, endogenous GβL was also shown to be SUMO modified at K305 by MS analysis (45). Taken together, these results from site-directed mutagenesis and MS analysis demonstrat that GβL is SUMOylated at multiple lysines, which may in turn regulate mTOR signaling via modulating protein–protein interactions.

Recombinant GβL cannot be SUMOylated in vitro

SUMOylation of target substrates requires specific conjugation machinery to occur in vivo, and E3-SUMO ligases facilitate substrate specificity and enhance SUMO transfer from the E2-SUMO-conjugating enzyme, Ubc9. We tested whether GβL could be SUMOylated using purified recombinant E1 and E2 in vitro. As shown before, glutathione-S-transferase (GST)-RanGAP1 is rapidly SUMOylated in the presence of E1 and E2, and this effect, as expected, is enhanced in the presence of the E3 ligase, Rhes (15, 34) (Fig. 3A). In contrast, when purified recombinant GST-GβL or His-GβL were incubated in the presence of E1 and E2 and Rhes, we did not observe any high molecular weight conjugates indicative of SUMOylation of GβL (Fig. 3, B and C). Furthermore, immunoprecipitation (IP) of HA-GβL from HEK293 cells and subsequent in vitro SUMOylation revealed that GβL was not SUMOylated in the presence of E1 and E2 or in the presence of the E3 ligase, Rhes (Fig. 3D). These data suggest that SUMO E1 and E2 are not sufficient to SUMOylate GβL in vitro, indicating that an E3 ligase other than Rhes is required for the SUMOylation of GβL.

SUMO regulates AA-induced activation of mTORC1 signaling

Since we found that the mTOR regulatory subunit GβL is SUMOylated, we then examined if the loss of Sumo1 influences mTORC1 activity. To test this hypothesis, we generated Sumo1 WT (Sumo1^+/+) and Sumo1 KO (Sumo1^−/−) primary mouse embryonic fibroblasts (MEFs) at E13.5 and measured mTORC1 activity (pS6K^T389, pS6^S235/236, and p4EBP1^S65) under conditions of essential AA starvation (−AA) in the presence or absence of leucine stimulation (+Leu). We observed a slight decrease in the phosphorylation of the S6K target, pS6^S235/236, in AA-starved Sumo1^−/− MEFs, compared with Sumo1^+/+ MEFs (Fig. 4, A and B). Upon leucine stimulation, we found a strong reduction in the phosphorylation of direct mTOR targets, pS6K^T389 and p4EBP1^S65, in Sumo1^−/− MEFs compared with Sumo1^+/+ (Fig. 4, A and B). We also observed a noticeable decrease in pS6^S235/236 upon leucine stimulation (Fig. 4, A and B). Furthermore, insulin-induced phosphorylation of the mTORC2 target (pAkt 473) or PI3K target (pAkt 308) was unaltered in Sumo1^−/− MEFs, compared with Sumo1^+/+ (Fig. S1). These results indicate that loss of Sumo1 impairs mTORC1 activity, but not mTORC2 or PI3K mediated Akt phosphorylation.

Because previous studies have shown that SUMO2 and SUMO3 can compensate for SUMO1 loss of function in mice (29, 30), we hypothesized that SUMO2 or SUMO3 might compensate for SUMO1 and/or in addition contribute in regulating mTORC1 activity. To experimentally test this hypothesis, we employed SUMO1/2/3-depleted striatal neuronal cells (STHdh^Q7/^Q7) generated using CRISPR–Cas-9 technology (SUMO1/2/3Δ cells), which displayed ∼40% reduction in unconjugated SUMO1 and SUMO2/3 levels compared with control cells (46) (Fig. 5, A and B). We then compared mTORC1 signaling between control and SUMO1/2/3Δ cells in full media conditions (serum + AAs), AA starvation (Krebs media), and starvation followed by leucine stimulation (+Leu). We found a significant decrease in the phosphorylation of mTOR at Ser²⁴⁴⁸ (a target of PI3K (47)) in SUMO1/2/3Δ cells compared with control cells in all conditions (Fig. 5, A and C). However, the phosphorylation of mTOR at Ser²⁴⁸¹, an autophosphorylation site (48), is not significantly affected in full media or by starvation or leucine stimulation in SUMO1/2/3Δ cells (Fig. 5, A and C). In full media and AA starvation conditions, the phosphorylation of both the mTORC1 target (p4EBP1^T37/46) and S6K target (pS6^S235/236) were significantly attenuated in SUMO1/2/3Δ cells, compared with control cells (Fig. 5, A and D). Upon leucine stimulation, while control cells showed a strong activation of mTORC1 signaling (pS6K^T389, pS6^S235/236, and p4EBP1^T37/46), this was markedly decreased in SUMO1/2/3Δ cells (Fig. 5, A and D). These results indicate that depletion of all three SUMO isoforms robustly decreases AA-induced activation of mTORC1 in striatal neuron cells.

Next, we investigated whether SUMOylation affects the activation of mTORC2 (pAkt^S473) or PI3K (pAkt^T308) signaling. In full media conditions, the levels of pAkt^S473 or pAkt^T308 are comparable between control and SUMO1/2/3Δ cells (Fig. 5, A and E). Upon starvation, consistent with previous reports (49, 50, 51), we found that the pAkt^S473 and pAkt^T308 levels are increased in control cells, but not in SUMO1/2/3Δ cells (Fig. 5, A and E). Leucine addition had no further impact on pAkt^S473 and pAkt^T308 levels in control cells or SUMO1/2/3Δ cells (Fig. 5, A and E). These observations indicate that SUMOylation regulates the starvation-induced upregulation of Akt activity mediated by mTORC2 and PI3K. In contrast, we did not observe any changes in the levels of pERK^T202/Y204 in SUMO1/2/3Δ cells compared with control cells (Fig. 5, A and F). As expected, pERK^T202/Y204 is downregulated upon starvation (52), which was similar and unaffected upon Leu addition in both SUMO1/2/3Δ and control cells (Fig. 5, A and F). Moreover, insulin-induced mTORC2 (pAkt⁴⁷³) and PI3K (pAkt³⁰⁸) was similar between control and SUMO1/2/3Δ cells (Fig. S2). These results indicate that SUMO selectively regulates the specific starvation- and AA-mediated phosphorylation status of mTOR signaling without significantly interfering with ERK signaling. Finally, we tested whether diminished mTORC1 signaling in SUMO1/2/3Δ cells affects cell viability or proliferation. We did not observe floating cells indicative of cell death, but we found a decreased cell number in SUMO1/2/3Δ cells compared with control cells (Fig. 5G). As mTOR signaling is implicated in cell growth and proliferation (53), we conclude that diminished proliferation of SUMO1/2/3Δ cells may be due to reduced mTOR signaling.

SUMO regulates the mTORC1 and mTORC2 protein complex formation

We then wondered how SUMO might mechanistically regulate mTOR signaling. One possibility is that SUMO regulates complex formation of mTORC1 and mTORC2. To address this, we carried out IP experiments against mTOR and assessed its interaction with Raptor (mTORC1), Rictor (mTORC2), and GβL (a constitutively bound component of both mTORC1 and mTORC2) in control and SUMO1/2/3Δ cells. As expected, mTOR immunoglobulin G (IgG), but not control IgG, readily immunoprecipitated mTOR and coprecipitated Raptor, Rictor, and GβL (Fig. 6, A and B). However, the abundance of mTOR–Raptor and mTOR–Rictor interactions were diminished in SUMO1/2/3Δ cells compared with control cells, whereas the interaction with mTOR–GβL was unaffected (Fig. 6, A and B). Note, the ∼30% decrease in the interaction is highly significant (Fig. 6, A and B) as SUMO1/2/3Δ cells show only ∼40% depletion of SUMO, compared with control cells (Fig. 5B), indicating that SUMO selectively and strongly regulates the interaction of mTOR with Raptor and Rictor.

SUMOylation-defective mutant of GβL fails to activate mTOR signaling

Because we found that the mTOR and GβL interaction was unaffected in SUMO-depleted cells (Fig. 6A), we hypothesized that SUMOylation of GβL may participate in the regulation of mTOR activity. To test this, we generated GβL-depleted (GβLΔ) striatal neuronal cells using CRISPR–Cas9 tools. We obtained up to 70% loss of GβL (Fig. 7, A and B). Consistent with previous reports (54, 55), both mTORC1 (pS6K^T389 and pS6^S235/236) and mTORC2 signaling (pAkt^S473) were diminished in GβLΔ cells in all three conditions: full media, AA starved (Krebs), or AA starved and leucine-stimulated conditions compared to control cells (Fig. 7, A and B). We also found that pmTOR^S2448, but not pERK^T202/Y204, was diminished in GβLΔ compared with control cells (Fig. 7, A and B). Thus, we successfully generated GβLΔ cells defective in mTOR activity. To test the role of GβL SUMOylation in regulating mTOR signaling, we transfected HA-GβL WT and HA-GβL full KR mutant (HA-GβL-KR) into GβLΔ cells, followed by AA starvation (–AA, Krebs) and stimulation with 3 mM leucine (+Leu). We investigated mTOR signaling using confocal microscopy/immunofluorescence by measuring intensity levels of pS6^S235/236 (mTORC1) or pAkt Ser⁴⁷³ (mTORC2) in individual GβLΔ cells expressing HA-GβL WT or HA-GβL-KR. In GβLΔ cells expressing HA-GβL WT, we observed enhanced staining of pS6^S235/236 following leucine stimulation (Fig. 7, C and D), whereas HA-GβL-KR had negligible pS6^S235/236 signal (Fig. 7, C and D). Similarly, GβLΔ cells transfected with HA-GβL WT also showed enhanced pAkt^S473 compared with HA-GβL-KR-transfected cells (Fig. 7, E and F). We found no significant changes in mTOR staining between HA-GβL WT or HA-GβL-KR-expressing cells (Fig. 7, G and H). These results suggest that SUMOylation of GβL has an essential role in controlling mTORC1 and mTORC2 signaling.

Reagents and antibodies

All reagents were purchased from Sigma unless indicated otherwise. Antibodies against GβL (3274), mTOR (2972, 2983), pmTOR S2448 (5536), pmTOR S2481 (2974), pS6K T389 (9234), pS6 S235/236 (4858), p4EBP1 T37/46 (2855) p4EBP1 S65 (9451), pAkt S473 (4060), p44/42 ERK1/2 (9101), S6K (9202), S6 (2217), 4EBP1 (9644), Akt (4691), ERK1/2 (4695), Raptor (2280), Rictor (9476), SUMO-1 (4930), and SUMO-2/3 (4971) were obtained from Cell Signaling Technology. 6×-His tag antibody (MA1-21315) was from ThermoFisher Scientific. Antibodies for actin (sc-47778), Myc (sc-40), and GST (sc-138 horseradish perxidase) were obtained from Santa Cruz Biotechnology. FLAG antibody (F7425) was obtained from Sigma–Aldrich. HA-tagged monoclonal antibody (901513) was from BioLegend (previously Covance; catalog no.: MMS-101R). HA-tag polyclonal antibody (631207) was from Clontech.

Cell lines, growth conditions, transfections, and cell proliferation assay

HEK293 cells and MEFs were grown in Dulbecco’s modified Eagle’s medium (DMEM) (11965-092; ThermoFisher Scientific) supplemented with 10% fetal bovine serum (FBS), 1% penicillin/streptomycin, and 5 mM glutamine. For transfections, cells were seeded in 6-well plates or 10 cm plates. After 24 h, transfections were performed with corresponding DNA constructs and PolyFect (Qiagen) using the manufacturer’s instructions. Cells were harvested for experiments 40 h after being transfected. Mouse STHdh^Q7/Q7 striatal neuronal cells were obtained from the Coriell Institute and cultured in DMEM high glucose (10566-016; ThermoFisher Scientific), 10% FBS, 5% CO₂, at 33 °C as described in our previous works (34, 36, 46, 65, 66, 67, 68). GβL (sc-425272) and SUMO1/2/3 deletions were carried out in striatal cells using CRISPR–Cas-9 tools obtained from Santa Cruz, as described before (46, 66). Cell proliferation was tested by cell counting kit-8 assay (no. K1018; ApexBio) as per the manufacturer’s instructions. Briefly, control and SUMO1/2/3Δ striatal neuronal cells were seeded in 96-well plate at a concentration 0f 5000 cells per well. Next day, 10 μl cell counting kit-8 solution was added to each well of the plate and incubated at 33 °C for 2 h, and thereafter, the absorbance was measured at 450 nm. The pH of the cell culture media for MEFs, or striatal neuronal cells, remained within the range of 7.4 to 7.6.

Plasmids and site-directed mutagenesis

pRK5-HA-GβL was obtained from Addgene (1865). His-SUMO1, His-SUMO2, His-SUMO3, and His-SUMO-AA were obtained from Michael Matunis (34). Site-directed mutagenesis was performed on pRK5-HA-GβL using the Quik-Change II Site Directed mutagenesis kit (200523; Agilent Technologies) as per the manufacturer’s instructions. Successful mutagenesis was confirmed by Sanger sequencing (Genewiz).

Ni–NTA denaturing pull down

Ni–NTA pull down of His-SUMO conjugates was performed as previously described (69). Briefly, cells were rinsed in PBS, scraped from 10 cm dishes, and centrifuged at 750g for 5 min. Cell pellets were then directly lysed in pull-down buffer (6 M guanidine hydrochloride, 10 mM Tris, 100 mM sodium phosphate, 40 mM imidazole, 5 mM β-mercaptoethanol, pH 8.0) and sonicated. Lysates were then clarified by centrifugation at 3000g for 15 min. All subsequent wash steps were performed with 10 resin volumes of buffer followed by centrifugation at 800g for 2 min. Ni–NTA Agarose beads (30210; Qiagen) were pre-equilibrated by washing three times with pull-down buffer. After equilibration, beads were resuspended in pull-down buffer as a 50% slurry of beads to buffer. After quantification of cell lysates, 1 mg of lysate was added to 40 μl of Ni–NTA bead slurry to a total volume of 1 ml in Eppendorf tubes. Beads were then incubated overnight at 4 °C mixing end over end. The following day, beads were centrifuged and underwent washing once in pull-down buffer, once in pH 8.0 urea buffer (8 M urea, 10 mM Tris, 100 mM sodium phosphate, 0.1% Triton X-100, 20 mM imidazole, 5 mM β-mercaptoethanol, pH 8.0), and three additional times in pH 6.3 urea buffer (8 M urea, 10 mM Tris, 100 mM sodium phosphate, 0.1% Triton X-100, 20 mM imidazole, 5 mM β-mercaptoethanol, pH 6.3). Elution was performed using 20 μl of elution buffer (pH 8.0 urea uffer containing 200 mM imidazole, 4× NuPAGE lithium dodecyl sulfate (LDS) loading dye, 720 mM β-mercaptoethanol). Samples were then heated at 95 °C for 5 min and directly used for Western blotting. Inputs were loaded as 1% of the total cell lysate.

MS

Identification of SUMO modification lysine sites was carried out as described before (42, 43, 44). Briefly, HEK293 cells were transfected with mSUMO3 and HA-GβL, followed by denaturating Ni–NTA pulldown as described previously. The Ni–NTA resin was extensively washed with 50 mM ammonium bicarbonate to remove traces of Triton, and the proteins were digested with trypsin directly on the Ni–NTA solid support for 16 h at 37 °C. The mSUMO3-modified peptides were immunoprecipitated with a custom anti-NQTGG antibody that recognizes the tryptic remnant created on the SUMO-modified lysine side chain, as described before (42, 43, 44). Samples were analyzed on the Q-Exactive HF instrument (ThermoFisher Scientific), and raw files were processed using MaxQuant and Perseus, as described previously (42, 43, 44).

In vitro SUMOylation

SUMOylation assays were performed as described previously (34) in 20 μl using 1× reaction buffer (20 mM Hepes, 2 mM magnesium acetate, 110 mM KCl, pH 7.4), 1 μg of E1 (Aos1/Uba2), 500 ng of Ubc9, 2 μg of SUMO-1/2, 5 mM ATP, 0.2 mM DTT, and 200 ng of Rhes at 32 °C for 30 min unless noted otherwise. GST-RanGAP1 was used as positive control for SUMO E3 ligase activity of Rhes, as described before (15, 34). To stop reactions, 4× NuPAGE LDS sample buffer was added, and samples were heated at 95 °C for 5 min, followed by separation using SDS-PAGE and immunoblotting.

Purification of His-GβL and GST-GβL

pCMV-His-GβL was cloned from pRK5-HA-GβL into pGEX-6P2 and was transformed into BL21 (DE3) cells (New England Biolabs) and purified using Ni–NTA Agarose beads (30210; Qiagen) according to the manufacturer’s specifications. Briefly, following IPTG induction at 16 °C overnight, BL21 lysate was resuspended in lysis buffer (300 mM NaCl, 50 mM sodium phosphate [pH 8.0], 3% glycerol, 1% Triton X-100, 15 mM imidazole, and 1× protease inhibitor [Roche, Sigma]) and sonicated. Lysate was clarified by spinning at 30,000g for 30 min. Ni–NTA beads were pre-equilibrated by washing three times in 300 mM NaCl and 50 mM sodium phosphate (pH 8.0). Beads were incubated with lysate mixing end over end at 4 °C overnight. The following day, the beads were washed three times in an equilibration buffer, followed by elution using 300 mM NaCl, 50 mM sodium phosphate (pH 8.0), and 1 M imidazole.

pGEX6p2-GST-GβL was cloned from pRK5-HA-GβL into pGEX6p2-GST and was transformed into BL21 (DE3) cells and purified using Glutathione Sepharose beads (45000139; Fisher) as described in our previous studies (34, 35, 68, 70).

Preparation of MEFs

Sumo1 KO mice (Sumo1^−/−) were obtained from Jorma Palvimo (29). MEFs were prepared on E13.5 as described previously (71). Genotyping was performed by PCR with primer pairs for the WT allele (SUMO1 forward: 5′-CTC AAA CAA CAG ACC TGA TTG C-3′; SUMO1 reverse: 5′-CAC TAT GGA TAA GAC CTG TGA ATT-3′) and for the KO allele (Neo1 forward: 5′-CCA CCA AAG AAC GGA GCC GGT T-3′; SUMO-1 reverse: 5′-CAC TAT GGA TAA GAC CTG TGA ATT-3′). The WT of amplicon generated a fragment of 475 bp, whereas the KO amplicon generated a fragment of 550 bp. WT mice (C57BL/6) were obtained from Jackson Laboratory and maintained in our animal facility according to the Institutional Animal Care and Use Committee at The Scripps Research Institute. Mice were euthanized by cervical dislocation, and striatal tissues were dissected and rapidly frozen in liquid nitrogen.

AA treatments

MEFs were seeded in 6-well plates at least 24 h prior. For essential AA starvation, DMEM were replaced with DMEM/F12 lacking l-leucine, l-lysine, l-methionine, and FBS (F12; D9785, Sigma). After 1 h of starvation, cells were lysed or stimulated with 3 mM l-leucine for 15 min (+Leu). AA treatment in Krebs buffer was carried out as described in our previous work (36). Briefly, striatal cells were placed in Krebs buffer medium (20 mM Hepes [pH 7.4], glucose [4.5 g/l], 118 mM NaCl, 4.6 mM KCl, 1 mM MgCl₂, 12 mM NaHCO₃, 0.5 mM CaCl₂, and 0.2% [w/v] bovine serum albumin) devoid of serum and AAs for 1 h to induce full starvation conditions. For the stimulation conditions, after starvation, cells were stimulated for 15 min with 3 mM l-leucine.

Western blotting

MEFs were rinsed briefly in PBS and directly lysed in lysis buffer (40 mM Tris [pH 7.5], 120 mM NaCl, 1 mM EDTA, 0.3% CHAPS, 1× protease inhibitor cocktail [Roche, Sigma], and 1× phosphatase inhibitor [PhosSTOP; Roche, Sigma]). Lysates were passed several times through a 26-gauge needle and clarified by centrifugation at 11,000g for 20 min at 4 °C. Striatal neuronal cells were washed in PBS and lysed in lysis buffer (50 mM Tris–HCl [pH 7.4], 150 mM NaCl, 1% CHAPS, 1× protease inhibitor cocktail, and 1× phosphatase inhibitor [PhosSTOP]), sonicated for 3 × 5 s at 20% amplitude, and cleared by centrifugation for 10 min at 11,000g at 4 °C. Protein concentration was determined with a bicinchoninic acid protein assay reagent (Pierce). Equal amounts of protein (20–50 μg) were loaded and separated by electrophoresis in 4% to 12% Bis–Tris Gel (ThermoFisher Scientific), transferred to polyvinylidene difluoride membranes, and probed with the indicated primary antibodies. Horseradish perxidase–conjugated secondary antibodies (Jackson ImmunoResearch, Inc) were probed to detect bound primary IgG with a chemiluminescence imager (Alpha Innotech) using enhanced chemiluminescence from WesternBright Quantum (Advansta). The band intensities were quantified with ImageJ software (National Institutes of Health). Phosphorylated proteins were then normalized against the total protein levels (normalized to actin).

IP

For in vitro SUMOylation assays containing HA-GβL, HEK293 cells were transfected as described previously and lysed directly in lysis buffer (40 mM Tris [pH 7.5], 120 mM NaCl, 1 mM EDTA, 0.3% CHAPS, 1× complete protease inhibitor cocktail, and 1× PhosSTOP). Lysates were passed several times through a 26-gauge needle and clarified by centrifugation at 11,000g for 20 min at 4 °C. IP was performed using Anti-GβL (3274; Cell Signaling Technologies), Protein A/G Plus-Agarose beads (sc-2003; Santa Cruz Biotechnology), and 500 μg of lysate mixing end over end at 4 °C overnight. The following day, beads were washed three times in lysis buffer, followed by a single wash in 1× reaction buffer. Beads were then directly incubated in a reaction buffer containing the aforementioned reaction components for SUMOylation assays.

Striatal cells (2 × 10⁶) were plated in 10 cm dishes, and next day, after leucine stimulation, the cells were washed in cold PBS and lysed in IP buffer (50 mM Tris–HCl [pH 7.4], 150 mM NaCl, 1% CHAPS, 10% glycerol, 1× protease inhibitor cocktail, and 1× PhosSTOP). The lysates were run several times through a 26-gauge needle in IP buffer and incubated on ice for 15 min and centrifuged 11,000g for 15 min. Protein estimation in the lysate supernatant was done using a bicinchoninic acid method, a concentration (1 mg/ml) of protein lysates was precleared with 35 μl of protein A/G Plus-agarose beads for 1 h, supernatant was mixing end over end for 1 h at 4 °C in mTOR IgG (2983, Cell Signaling Technology) or control Rabbit IgG (CST-2729S), and then 60 μl protein A/G beads were added and mixing end over end overnight at 4 °C. After 12 h, the beads were washed five times with IP buffer (without protease/phosphatase inhibitor), and the protein samples were eluted with 30 μl of 2× LDS containing +1.5% β-mercaptoethanol and proceeded for Western blotting as described previously.

Immunostaining

Immunostaining was carried out as described in our previous work (36). Briefly, striatal cells were grown on poly-d-lysine (0.1 mg/ml)–coated glass coverslips, and after 48 h of transfection, the medium was changed to Krebs buffer medium devoid of serum and AAs for 1 h to induce full starvation conditions. For the stimulation conditions, cells were stimulated with 3 mM l-leucine for 15 min. Cells were washed with cold PBS, fixed with 4% paraformaldehyde (20 min), treated with 0.1 M glycine, and permeabilized with 0.1% (v/v) Triton X-100 (5 min). After being incubated with blocking buffer (1% normal donkey serum, 1% [w/v] bovine serum albumin, and 0.1% [v/v] Tween-20 in PBS) for 1 h at room temperature, cells were stained overnight at 4 °C with antibodies against HA (CST, 3724; 1:1000 dilution), pS6^Ser235/236 (CST, 4858S; 1:200 dilution), pAKT^Ser473 (CST, 4060S; 1:500 dilution), and mTOR (CST, 2983S; 1:200 dilution). Alexa Fluor 488 (Invitrogen, A21202; 1:1000 dilution) or Alexa Fluor 568 (Invitrogen, A10037; 1:1000 dilution)–conjugated secondary antibodies were incubated together with the nuclear stain 4′,6-diamidino-2-phenylindole for 1 h at room temperature. Glass coverslips were mounted with Fluoromount-G mounting medium (ThermoFisher Scientific, catalog no.: 0100-01). Images were acquired by using the Zeiss LSM 880 confocal microscope system with 63× objective.

Statistical analysis

Most experiments were performed in triplicate. Images were quantified using ImageJ (FIJI). Data are presented as mean ± SEM. Statistical analysis was performed using Student’s unpaired two-tailed t test or one-way ANOVA followed by Tukey’s multiple comparison test (Prism 7, GraphPad software).