TBC1D7 is a third subunit of the TSC1-TSC2 complex upstream of mTORC1

Abstract

The tuberous sclerosis complex (TSC) tumor suppressors form the TSC1-TSC2 complex, which limits cell growth in response to poor growth conditions. Through its GTPase-activating protein (GAP) activity toward Rheb, this complex inhibits the mechanistic target of rapamycin (mTOR) complex 1 (mTORC1), a key promoter of cell growth. Here, we identify and biochemically characterize TBC1D7 as a stably associated and ubiquitous third core subunit of the TSC1-TSC2 complex. We demonstrate that the TSC1-TSC2-TBC1D7 (TSC-TBC) complex is the functional complex that senses specific cellular growth conditions and possesses Rheb-GAP activity. Sequencing analyses of samples from TSC patients suggest that TBC1D7 is unlikely to represent TSC3. TBC1D7 knockdown decreases the association of TSC1 and TSC2 leading to decreased Rheb-GAP activity, without effects on the localization of TSC2 to the lysosome. Like the other TSC-TBC components, TBC1D7 knockdown results in increased mTORC1 signaling, delayed induction of autophagy, and enhanced cell growth under poor growth conditions.

Figure 1. Endogenous TBC1D7 binds tightly and ubiquitously to the TSC1-TSC2 heterodimer

(A) Relative abundance of proteins identified in TSC1-TSC2 purifications. Co-expressed Flag-TSC1 and -TSC2 were immunoprecipitated from lysates of HEK-293 cells, and tryptic peptides from 3xFlag-peptide eluates were identified using LC-MS/MS. Proteins with three or more peptides identified in both of two independent purifications, but absent in anti-Flag immunoprecipitates from cells expressing empty vector, are listed. The spectral abundance of a protein is calculated from the number of peptides identified (spectral count) and the amino acid length of the protein. See supporting data in Figure S1A. (B) Reciprocal co-immunoprecipitations of endogenous TBC1D7, TSC1, and TSC2. Endogenous proteins were immunoprecipitated from HeLa cell lysates using the indicated antibodies (listed in Experimental Procedures). See supporting data in Figure S1E. (C) Co-immunoprecipitation of TBC1D7 with TSC1 and TSC2 from mouse tissues. TSC1 and TSC2 were immunoprecipitated from lysates of the indicated mouse tissues. WAT (white adipose tissue). (D) One of two cellular pools of TBC1D7 co-fractionates with TSC1 and TSC2 in a density gradient. HeLa cell lysate was ultracentrifuged over a sucrose gradient and ten sequential fractions of increasing density were collected. Immunoblot band intensities were quantified using ImageJ software and graphed as a percentage of the summed band intensities in all fractions for each protein. (E) TBC1D7 and TSC1 co-immunoprecipitate with TSC2 from high-density fractions. TSC2 was immunoprecipitated from the indicated combined fractions from (D). (F) Stoichiometry of TBC1D7-bound TSC1 and TSC2. Endogenous TBC1D7 (or control IgG) was immunoprecipitated from a HeLa cell lysate and tryptic peptides were detected using LC-MS/MS. The ratio of TSC1 and TSC2 in the TBC1D7 immunoprecipitate (the control lacked these proteins) was estimated using two established quantitative methods, spectral abundance and average total ion current (TIC). Both methods yielded the same ratio as indicated in the schematic summary. (G) TBC1D7 is tightly bound to the TSC1-TSC2 complex. HeLa cells were lysed with buffer lacking NaCl and SDS. A single TSC1 immunoprecipitate from this lysate was split and washed with lysis buffer containing the indicated concentrations of NaCl or SDS.

Figure 2. TBC1D7 binds to and is stabilized by TSC1 within the TSC1-TSC2 complex

(A) TSC1-dependent association of TBC1D7 with TSC2 in MEFs. Immunoprecipitations with TBC1D7, TSC1, TSC2, or control IgG (C) antibodies were performed using lysates of littermate-derived wild-type (+/+) and either Tsc1−/− (left) or Tsc2−/− (right) MEFs. (B) TSC1-dependent association of TBC1D7 with TSC2 in HeLa cells. Immunoprecipitations as in (A) were performed using lysates of HeLa cells with siRNA-mediated knockdown of TSC1 or TSC2, compared to control (siC). (C) High-density fractionation of TBC1D7 is dependent on TSC1 and TSC2. Lysates of HeLa cells with siRNA-mediated knockdown of TBC1D7, TSC1, and TSC2 were ultracentrifuged over a sucrose gradient and ten sequential fractions of increasing density were collected. TBC1D7 immunoblot band intensities were quantified using ImageJ software and graphed as a percentage of the summed band intensities in all fractions for each protein. See supporting data in Figure S2A. (D) Loss of TSC1 decreases the stability of both TBC1D7 and TSC2. HeLa cells with siRNA-mediated knockdown of TSC1 were incubated with 100 μM cyclohexamide (chx) for the indicated duration. Immunoblot band intensities were quantified as in (C) and graphed as a percentage of the untreated (0 h) samples. (E) The “free” pool of TBC1D7 is unstable. HeLa cells were incubated with cyclohexamide or water (vehicle) for 3 h and fractionated as in (C). Fractions 3 and 7 from both treatments are also compared on the same immunoblot (right).

Figure 3. Knockdown of TBC1D7 leads to decreased association of TSC1 and TSC2 without effects on TSC2 localization

(A) Knockdown of TBC1D7 leads to a decrease in the co-immunoprecipitation of TSC1 and TSC2. Immunoprecipitations with TSC1, TSC2, or control IgG (C) antibodies were performed using lysates of HeLa cells with siRNA-mediated knockdown of TBC1D7, compared to control (siC). (B,C) Knockdown of TBC1D7 alters the fractionation patterns of both TSC1 and TSC2. Lysates from HeLa cells with siRNA-mediated knockdown of TBC1D7, TSC1, or TSC2 were ultracentrifuged over a sucrose gradient and ten sequential fractions of increasing density were collected. Immunoblot band intensities for TSC1 (B) and TSC2 (C) were quantified using ImageJ software and graphed as a percentage of the summed band intensities in all fractions. See supporting data in Figure S3A–C. (D) Localization of endogenous TSC2. HeLa cells with siRNA-mediated knockdown of TSC2 compared to a control (siC) were serum starved for 16 h and immunofluorescently labeled with a TSC2 antibody. An enlarged view of a single cell is shown below. (E) Knockdowns of TBC1D7 and TSC1 do not have gross effects on the localization of TSC2, including its co-localization with LAMP2 at the lysosome. HeLa cells with the indicated siRNA-mediated knockdowns were prepared as in (D), but co-immunolabeled for TSC2 (red) and LAMP2 (green). In merged images, yellow and orange regions indicate co-localization, and inset boxes show planes perpendicular to the main image that were constructed from z-stacks (arrows indicate cell shown in inset). Corresponding immunoblots and an enlarged view of a single cell are shown to the right. See supporting data in Figure S3D,E.

Figure 4. The TSC-TBC complex responds to growth cues and has Rheb-GAP activity

(A) TBC1D7 is associated with TSC2 that is phosphorylated by Akt in response to insulin. Serum starved (18 h) HeLa cells were stimulated with insulin (100 nM) for 10 min prior to lysis and immunoprecipitation (IP) with TBC1D7, TSC1, or control IgG (C) antibodies. (B) TBC1D7 is associated with TSC2 that is phosphorylated by AMPK in response to energy stress. MEFs were grown in the presence of fresh serum (10%) with or without phenformin (1 mM) for 2 h prior to lysis and IP as in (A). (C) TBC1D7 is associated with TSC2 that has Rheb-GAP activity. HeLa cells with stable shRNA-mediated knockdown of TSC2 (shTSC2) or a firefly luciferase control (shLUC) were serum starved for 6 h prior to lysis and IP as in (A). Immunopurified complexes were subjected to Rheb-GAP assays using recombinant GST or GST-Rheb pre-loaded with GTP[α-³²P]. Rheb-bound GTP and GDP were separated by thin layer chromatography (TLC) and quantified using a phosphoimager in duplicate experiments. Mean GDP levels were graphed as percent of total GTP plus GDP (%GDP), with error bars for standard error of the mean. (D) Knockdown of TBC1D7 leads to a net decrease in Rheb-GAP activity of TSC2. Rheb-GAP assays were performed, analyzed, and graphed as in (C) except using HeLa cells with stable shRNA-mediated knockdown of TBC1D7 (shTBC7) or a GFP control (shGFP), and twice the amount of cells. See supporting data in Figure S4B.

Figure 5. TBC1D7 negatively regulates mTORC1 signaling under poor growth conditions

(A) Growth factor-independent mTORC1 signaling in cells with stable knockdown of TBC1D7. HeLa cells with stable shRNA-mediated knockdown of TBC1D7 (shTBC7) or a GFP control (shGFP) were serum starved for 16 h and, where indicated, treated with rapamycin (20 nM) for the final 15 min. (B) Growth factor-independent activation of mTORC1 signaling in a variety of cell lines with knockdown of TBC1D7. The indicated cell lines with siRNA-mediated knockdown of human (293E, RPE1, HeLa) or mouse (MEF) TBC1D7 compared to controls (siC) were serum starved for 16 h prior to lysis. (C) Dual knockdown of TBC1D7 with TSC1 or TSC2 does not have an additive effect on mTORC1 signaling. HeLa cells were co-transfected with the indicated siRNAs and treated as in (B). (D) mTORC1 signaling in TBC1D7-deficient cells remains Rheb-dependent. HeLa cells were co-transfected with the indicated siRNAs and treated as in (B). (E) mTORC1 signaling in TBC1D7-deficient cells is resistant to growth factor and glucose-withdrawal but not amino acid withdrawal. MEFs were treated as in (B) but following serum starvation, the indicated samples were either starved of all amino acids or glucose for an additional 3 h. Light and dark exposures of the same phospho-S6K blot are shown. (F) Co-localization of mTOR and LAMP2 is not affected by loss of TBC1D7. HeLa cells were serum starved for 16 h and immunofluorescently labeled with mTOR and LAMP2 antibodies. In merged images of mTOR (red) and LAMP2 (green), co-localization is represented by yellow and orange pixels. Insets in each merged image show planes perpendicular to the primary image that were constructed from z-stacks. Arrows indicate the cells shown in these insets. (G) Autophagy induction is delayed in TBC1D7-deficient cells. HeLa cells with stable shRNA-mediated knockdown of TBC1D7 (shTBC7) or GFP as a control (shGFP) were grown in complete growth media and then starved of all amino acids (-aa) with or without dialyzed FBS (dFBS) for the indicated duration (min). LC3B-II to -I ratios were quantified with ImageJ software and are shown normalized to the 20 min shGFP samples. (H) Knockdown of TBC1D7 increases cell size. HeLa cells were transfected with control (siC) or TBC1D7 (siTBC7) siRNAs and, 24-h post-transfection, were serum starved for 48 h. The distribution of cell diameters (in μm) within the population of cells was measured using a Z2 Coulter counter. p<0.0001 for comparison of siC and siTBC7 size distributions (average of 5 replicates, with 28,400 cells measured in each). See supporting data in Figure S5.

Figure 6. Model of the TSC-TBC Complex

Schematic of the integrated regulation of mTORC1 by the Rag GTPases through the Ragulator, and Rheb through the TSC-TBC complex (or Rhebulator). See text for details.