mTOR regulates lysosomal ATP-sensitive two-pore Na(+) channels to adapt to metabolic state

Abstract

Survival in the wild requires organismal adaptations to the availability of nutrients. Endosomes and lysosomes are key intracellular organelles that couple nutrition and metabolic status to cellular responses, but how they detect cytosolic ATP levels is not well understood. Here, we identify an endolysosomal ATP-sensitive Na(+) channel (lysoNa(ATP)). The channel is a complex formed by two-pore channels (TPC1 and TPC2), ion channels previously thought to be gated by nicotinic acid adenine dinucleotide phosphate (NAADP), and the mammalian target of rapamycin (mTOR). The channel complex detects nutrient status, becomes constitutively open upon nutrient removal and mTOR translocation off the lysosomal membrane, and controls the lysosome's membrane potential, pH stability, and amino acid homeostasis. Mutant mice lacking lysoNa(ATP) have much reduced exercise endurance after fasting. Thus, TPCs make up an ion channel family that couples the cell's metabolic state to endolysosomal function and are crucial for physical endurance during food restriction.

Figure 1. Endolysosomal ATP-sensitive Channel (lysoNa_ATP)

(A) Whole-endolysosome patch clamp recording. Inward current (negative) denotes positive charge flowing into the cytosol (bath) from the endolysosomal lumen. (B-E) Recordings with ramp protocols (−100 to +100 mV in 1 s, every 10 s, V_h = 0 mV) from mouse peritoneal macrophages. (B) Representative continuous recordings. Recordings at time points 1, 2 and 3 (indicated) were used for the current-voltage (I-V) relationships in (C). (D) Statistics of current amplitudes at -100 mV. (E) Current amplitudes normalized to those obtained without ATP and fitted to the Hill equation (n ≥ 5). (F-K) Representative (F-H) and averaged (I-K) ATP-sensitive currents recorded from cardiac myocytes (F, I), hepatocytes (G, J) and fibroblasts (H, K). Unless otherwise stated, recordings shown were obtained with PI(3,5)P₂ (1 μM) in the bath. Numbers of endolysosomes are in parentheses. Data are presented as mean ± SEM.

Figure 2. TPC1 and TPC2 Form lysoNa_ATP in HEK293T Cells

(A-D) ATP sensitivity of endolysosomal currents was tested in HEK293T cells transfected with GFP (A) or GFP-tagged TRPML1 (B), TPC1 (C) and TPC2 (D). PI(3,5)P₂ (1 μM) was included in the bath during recordings. (A) Mock-transfected HEK293T cells had little lysoNa_ATP current (24.9 ± 8.7 pA without ATP, 16.2 ± 6.2 pA with 1 mM ATP-Mg, at -100 mV). (B) I_TRPML1 is insensitive to ATP, as shown in the representative recordings (left) and the statistics of the averaged current amplitudes (right, at −100 mV). (C, D) ATP-sensitive currents recorded from TPC1- (C) and TPC2-transfected cells (D). The IC₅₀ of ATP on I_TPC2 was 0.55 ± 0.09 mM and 0.92 ± 0.31 mM in presence of 0.1 μM and 1 μM PI(3,5)P₂, respectively (n ≥ 4). (E, F) I_TPC2 is insensitive to ADP (E) and GTP (F). (G) Similar to (D) but 10 mM ATP-Mg was added to the pipette solution. Data are shown as mean ± SEM. See also Figure S1.

Figure 3. mTOR Is Required for lysoNa_ATP's ATP Sensitivity

(A-C) Compared with the control (A, no rapamycin), currents from TPC2-transfected endolysosomes were minimally inhibited by 1 mM ATP in the presence of rapamycin (B, 0.2 μM rapamycin). Statistical data are in (C). (D-F) Similar to (A-C), but recorded from macrophages (1.0 μM rapamycin used in E). (G-J) Infecting HEK293T cells with shRNA lentivirus against mTOR but not with the control shRNA reduced the endogenous mTOR protein level (G) and TPC2 current's ATP sensitivity (H-J). Inset in (J) indicates TPC2 protein levels in control and mTOR shRNA virus-infected cells. (K-L) Knocking down Raptor (L) but not Rictor (K) reduced lysoNa_ATP's ATP sensitivity. Data are represented as mean ± SEM. See also Figure S2.

Figure 4. TPCs Form a Complex with mTOR

(A) Immunoblotting (IB) with whole cell lysates or immunoprecipitates (IP) from cells co-transfected with mTOR and GFP or GFP-tagged channels as indicated, showing that both TPC1 and TPC2, but not TRPML1, associate with mTOR. (B-E) Co-transfecting mTOR and TPC2 into HEK293T cells increased mTOR protein levels (B) and TPC2's ATP sensitivity, as shown in representative recordings (C, D) and normalized current sizes at −100 mV (E). (F-H) lysoNa_ATP currents recorded in the presence of rapamycin (1 μM) from cells co-transfected with a rapamycin-resistant mTOR (mTOR^S2035T) (F, H), or a rapamycin-resistant and kinase dead double mutant (mTOR^{S2035T/D2357E}) (G, H). Data are presented as mean ± SEM. See also Figure S3.

Figure 5. LysoNa_ATP Detects Nutrient Deprivation

(A-E) Currents recorded from HEK293T cells transfected with TPC2 alone (A, B), or together with RagB^GTP (C) or RagB^GDP (D), were fed (A, D), or starved (B, C) in medium containing no glucose or amino acids for 60 min before recording. Current sizes normalized to those obtained without ATP (at −100 mV) are summarized in (E). ** indicates statistical significance (p < 0.01). (F-I) lysoNa_ATP currents recorded from TPC2-transfected cells with treatments of glucose removal (F), amino acid removal (G) and amino acid re-fed (10 min) after removal (H). (J) Immunoblotting (IB) with whole cell lysates (lower 3 panels) or immunoprecipitates (IP with anti-GFP; upper 2 panels) from cells co-transfected with mTOR and GFP-tagged TPC2 with no treatment (control, fed), starvation (amino acid deprivation for 3h) or amino acid re-feeding after starvation (10x amino acid stimulation for 10 min). A weak mTOR band in the “starved” lane in the upper panel was visible after longer exposure (not shown). Data are presented as mean ± SEM.

Figure 6. ATP-sensitive Current Is Absent in Endolysosomes of tpc1/tpc2 Knockout Peritoneal Macrophages

Currents were recorded before the addition of PI(3,5)P₂ (basal), after 1 μM PI(3,5)P₂, and after application of ATP, from tpc1/tpc2 dKO peritoneal macrophages. 1 μM PI(3,5)P₂ elicited no inward current (dKO: 32.1 ± 7.4 pA before PI(3,5)P₂, 20.9 ± 3.9 pA after PI(3,5)P₂, at −100 mV, n = 10; WT: 22.8 ± 5.8 pA before PI(3,5)P₂, 142.3 ± 3.5 pA after PI(3,5)P₂, n = 20). See also Figure 1C & D for comparison with WT. (D, E) Rescue of lysoNa_ATP in the knockout macrophages with transfection of human (D, E) or mouse TPC2 (mTPC2) (E). Data are presented as mean ± SEM. See also Figure S4.

Figure 7. lysoNa_ATPs Control Lysosomal Membrane Potentials, Lysosomal pH, and Are Required for Normal Fasting Endurance

(A-C) Membrane potentials of macrophage lysosomes from wild-type (A) and dKO (B) were monitored with current clamp recordings while ATP was added to the bath containing 0 or 10 mM Na⁺ as indicated by the bars above the recordings. (D-F) Lysosomal pH measured with ratio-metric imaging from WT (D, F) and dKO macrophage lysosomes (E, F) before and after starvation. Distribution histograms of pH values (fitted to Gaussian distributions) are in (D, E) and averaged values are in (F). (G) Amino acid analysis. Levels of 15 plasma amino acids (R, K, T, M, F, V, L, I, D, S, Q, G, A, Y, W) of each animal (5 to 6 in each group) were measured after 3 days of fasting and normalized to those at the beginning (6 hr) of fasting. (H, I) Mice were tested before and after fasting for 3 days, and 2 days after re-feeding. (H) Distance traveled at exhaustion. (I) Distance traveled of each mouse (represented by each point) after fasting and 2 days after re-feeding, as normalized to that before fasting. Black and blue circles indicate behavioral tests after fasting and after re-feeding conditions, respectively. Red triangles indicate mean values. Numbers of animals tested are in (H). Several points overlap and are not distinguished. Asterisks indicate statistical significance (*, p<0.05; ***, p< 0.001). NS, not significant. Data are presented as mean ± SEM. See also Figure S5.