Regulation of the creatine transporter by AMP-activated protein kinase in kidney epithelial cells

Abstract

The metabolic sensor AMP-activated protein kinase (AMPK) regulates several transport proteins, potentially coupling transport activity to cellular stress and energy levels. The creatine transporter (CRT; SLC6A8) mediates creatine uptake into several cell types, including kidney epithelial cells, where it has been proposed that CRT is important for reclamation of filtered creatine, a process critical for total body creatine homeostasis. Creatine and phosphocreatine provide an intracellular, high-energy phosphate-buffering system essential for maintaining ATP supply in tissues with high energy demands. To test our hypothesis that CRT is regulated by AMPK in the kidney, we examined CRT and AMPK distribution in the kidney and the regulation of CRT by AMPK in cells. By immunofluorescence staining, we detected CRT at the apical pole in a polarized mouse S3 proximal tubule cell line and in native rat kidney proximal tubules, a distribution overlapping with AMPK. Two-electrode voltage-clamp (TEV) measurements of Na(+)-dependent creatine uptake into CRT-expressing Xenopus laevis oocytes demonstrated that AMPK inhibited CRT via a reduction in its Michaelis-Menten V(max) parameter. [(14)C]creatine uptake and apical surface biotinylation measurements in polarized S3 cells demonstrated parallel reductions in creatine influx and CRT apical membrane expression after AMPK activation with the AMP-mimetic compound 5-aminoimidazole-4-carboxamide-1-beta-D-ribofuranoside. In oocyte TEV experiments, rapamycin and the AMPK activator 5-aminoimidazole-4-carboxamide-1-beta-D-ribofuranosyl 5'-monophosphate (ZMP) inhibited CRT currents, but there was no additive inhibition of CRT by ZMP, suggesting that AMPK may inhibit CRT indirectly via the mammalian target of rapamycin pathway. We conclude that AMPK inhibits apical membrane CRT expression in kidney proximal tubule cells, which could be important in reducing cellular energy expenditure and unnecessary creatine reabsorption under conditions of local and whole body metabolic stress.

Fig. 1.

Creatine transporter (CRT) and AMP-activated protein kinase (AMPK) localize at the apical pole in rat kidney proximal tubules. A–C: confocal microscopic images of control rat kidney immunolabeled using an anti-CRT antibody alone (A, green) or an anti-CRT antibody amplified by tyramide signal amplification (B) and then using an anti-AMPK-α antibody (C, red). D: regions of colocalization of CRT and AMPK-α at the apical pole of proximal tubule cells (yellow). Scale bars, 100 μm. E–G: little nonspecific staining in tissues incubated in the absence of primary antibody.

Fig. 2.

AMPK activation inhibits CRT-dependent currents in Xenopus laevis oocytes. A: representative traces of CRT-dependent inward (downward) currents at different concentrations of buffer creatine in the presence of potassium 5-aminoimidazole-4-carboxamide-1-β-d-ribofuranosyl 5′-monophosphate (ZMP) or potassium gluconate (KG). B: CRT-dependent currents (means ± SE) were significantly inhibited (30–40%) by the AMPK activator ZMP compared with KG control after 6 h of treatment in the presence of 30 and 300 μM creatine. *P < 0.05 relative to KG (2-tailed, unpaired t-tests; n = 24 oocytes, N = 3 batches per condition). C: relative CRT-dependent currents measured at various creatine concentrations 6 h after injection of ZMP or KG and then fitted by nonlinear least-squares regression to the Michaelis-Menten equation. V_max for each experiment was normalized by the mean V_max for KG (n = 24 oocytes, N = 3 batches per condition). The AMPK activator ZMP caused a 38 ± 2% reduction in V_max but had no effect on K_m.

Fig. 3.

AMPK activation inhibits apical pole localization of CRT in mouse S3 proximal tubule cells, as determined by immunofluorescence confocal microscopy. A: time course of 5-aminoimidazole-4-carboxamide-1-β-d-ribofuranoside (AICAR)-mediated AMPK activation, as measured by immunoblotting of phosphorylated (Thr¹⁷²) AMPK-α (pT172) in S3 cell lysates. AICAR treatment (1 mM) for 2–16 h induced a significant upregulation of phosphorylated (Thr¹⁷²) AMPK-α compared with untreated cells at time 0 under conditions of equivalent protein loading (β-actin control). B: AICAR-induced cytoplasmic redistribution of CRT in polarized S3 cells grown on Transwell filters, as shown by x-z reconstructions of confocal microscope image stacks of S3 cell monolayers immunolabeled using anti-CRT antibody (green) alone or anti-CRT antibody and an anti-ZO-1 antibody (red) grown under control conditions (Con) or in the presence of 1 mM AICAR for 2 h. C: Cy3-coupled phalloidin staining of S3 cell monolayers treated with vehicle (Con) or 1 mM AICAR for 2 h. Images are representative of ≥3 different filter sets and treatments. Scale bar, 10 μm.

Fig. 4.

Complex signal pattern of CRT revealed by immunoblotting. HEK-293 (lanes 1–4) or S3 (lanes 5–6) cells were transiently transfected with empty vector (Con) or a plasmid encoding the hemagglutinin (HA)-tagged CRT. One-day-posttransfection cell lysates were separated by SDS-PAGE and analyzed by immunoblotting using the anti-CRT (lanes 1 and 2) or anti-HA (lanes 3–6) antibody. Both antibodies revealed similar immunoblot signal patterns after transfection with HA-CRT in HEK-293 cells, identifying 4 major bands (short arrows on left). However, the signal pattern observed in HA-CRT-transfected S3 cells (lane 6; long arrows on right) differed from that in HA-CRT-transfected HEK-293 cells (lane 4), suggesting that cellular processing of the CRT at the mRNA and/or protein level differs as a function of host cell type.

Fig. 5.

AMPK activation inhibits apical surface expression of CRT in mouse S3 proximal tubule cells, as measured by biotinylation. A: AMPK activity as measured by immunoblotting of phosphorylated (Thr¹⁷²) AMPK-α (pThr-172) in polarized S3 cell lysates treated with vehicle (Con) or 1 mM AICAR for 2 h under conditions of equal protein loading (β-actin). Quantification (means ± SE) of phosphorylated (Thr¹⁷²) AMPK-α signal normalized to β-actin signal is shown relative to control (bottom). *P < 0.05 (n = 4 replicate experiments). B: immunoblot of CRT in 5% of whole cell lysates or biotinylated protein samples from polarized S3 cell monolayers grown under control conditions (Con) or treated with the AMPK activator AICAR. AICAR-mediated AMPK activation inhibited surface CRT expression (biotinylated CRT) by 30–40% relative to vehicle control, as measured by quantification (means ± SE) of CRT band intensity in the biotinylated fraction divided by that in the cell lysate and expressed relative to that in control cells (bottom). *P < 0.05 (n = 4 experiments). Relatively small amount of β-actin present in the biotinylated fraction (<2% of total cellular β-actin) likely represents binding of the cortical actin cytoskeleton to intrinsic membrane proteins (e.g., Na⁺/H⁺ exchanger isoform 3) present at the apical membrane of these proximal tubule-derived S3 cells. A variety of apical membrane transport proteins have been shown to interact directly or indirectly with the actin cytoskeleton (30).

Fig. 6.

AMPK activation inhibits CRT-mediated [¹⁴C]creatine uptake in mouse S3 proximal tubule cells. Polarized mouse proximal tubule S3 segment cells were grown on Transwell filters. A: relative AMPK activity as measured by phosphorylated (Thr¹⁷²) AMPK-α immunoblotting following treatment with AICAR (1 mM, 2 h) vs. vehicle. Values are means ± SE. *P < 0.05 (n = 3). B: relative CRT uptake assays. Krebs-Ringer-HEPES buffer was applied in the presence or absence of AICAR containing 10 μM [¹⁴C]creatine on the apical side for 45 min at 37°C (n = 3 per condition). The competitive CRT substrate β-guanidinopropionic acid (1 mM) was added in parallel CRT uptake flux measurements as a control (not shown). Resulting non-CRT-mediated [¹⁴C]creatine uptake (generally ∼10% of total) was considered background, and this value was subtracted from total uptake values. AICAR induced a ∼20% inhibition in CRT-dependent [¹⁴C]creatine uptake. *P < 0.05 (n = 3).

Fig. 7.

Glutathione S-transferase (GST) pull-down assays detect interaction of CRT with the AMPK-α₁ catalytic subunit. HEK-293 cells were cotransfected with HA-CRT, HA-AMPK-β₁, HA-AMPK-γ₁, and either GST-AMPK-α₁ (AMPK) or GST alone (Con). Cell lysates were used for immunoblot (IB) directly (Input) or for affinity purification using glutathione-agarose beads (GST pull-down). A: immunoblot with HA-tagged antibodies to check for interaction of HA-CRT with GST-AMPK-α₁. B: Western blot analysis of samples in A, with anti-GST antibodies used to verify GST or GST-AMPK-α₁ expression, as well as enrichment by GST pull-down. Images are representative of 3 repeat experiments.

Fig. 8.

AMPK regulation of CRT may involve the mammalian target of rapamycin (mTOR) pathway. Xenopus oocytes were injected with CRT and mTOR cRNAs 5–6 days before 2-electrode voltage-clamp (TEV) measurements and treated with DMSO (vehicle) or 50 nM rapamycin for 3 days before TEV measurements. Oocytes from control and rapamycin-treated groups were then injected with K-ZMP (ZMP) or K-gluconate (KG) 5–8 h before TEV measurements. CRT-dependent current decreased significantly in mTOR-expressing oocytes treated with ZMP compared with control oocytes treated with KG (left). ZMP did not have a significant effect in oocytes expressing mTOR and treated with rapamycin (right). *P < 0.05 and #P < 0.05 vs. mTOR-alone control (n = 17–21 oocytes per condition, N = 3 batches).