Regulation of Na⁺-coupled glucose carrier SGLT1 by AMP-activated protein kinase

Abstract

AMP-activated protein kinase (AMPK), a serine/threonine kinase activated upon energy depletion, stimulates energy production and limits energy utilization. It has previously been shown to enhance cellular glucose uptake through the GLUT family of facilitative glucose transporters. The present study explored the possibility that AMPK may regulate Na⁺-coupled glucose transport through SGLT1 (SLC5A1). To this end, SGLT1 was expressed in Xenopus oocytes with and without AMPK and electrogenic glucose transport determined by dual electrode voltage clamping experiments. In SGLT1-expressing oocytes but not in oocytes injected with water or expressing constitutively active ^γR70QAMPK (α1β1γ1(R70Q)) alone, the addition of glucose to the extracellular bath generated a current (I_g), which was half maximal (K_M) at ≈ 650 μM glucose concentration. Coexpression of ^γR70QAMPK did not affect K_M but significantly enhanced the maximal current (≈ 1.7 fold). Coexpression of wild type AMPK or the kinase dead ^αK45RAMPK mutant (α1(K45R)β1γ1) did not appreciably affect I_g. According to confocal microscopy and Western Blotting, AICAR (1 mM), phenformin (1 mM) and A-769662 (10 μM) enhanced the SGLT1 protein abundance in the cell membrane of Caco2 cells suggesting that AMPK activity may increase membrane translocation of SGLT1. These observations support a role for AMPK in the regulation of Na⁺-coupled glucose transport.

Keywords::

• AMPK
• SGLT1
• glucose transport

Introduction

The AMP-activated protein kinase (AMPK) is activated upon increase in the cytosolic AMP/ATP concentration ratio and thus senses the energy status of the cell [1–3]. Functionally, AMPK has been demonstrated to play roles in the regulation of cellular glucose uptake, glycolysis, fatty acid oxidation and of enzymes required for ATP production [3–6]. Accordingly, AMPK enhances the capacity of the cell to generate ATP [7]. By the same token, AMPK downregulates energy-utilizing mechanisms including protein synthesis, gluconeogenesis and lipogenesis [3,4,7]. Thus through its roles in promoting ATP production and conservation AMPK supports cell survival during energy deprivation [7–9].

Cellular glucose uptake is mediated by two families of glucose transporters; the GLUT family that facilitates the passive transport of glucose across the membrane and the Na⁺ glucose cotransporters which mediate secondary active transport driven by the coupling of glucose uptake with Na⁺. It is widely accepted that acute activation of AMPK leads to increased glucose transport in a variety of cells predominantly through the stimulation of the GLUT family of transport [10–21]. However, during prolonged ischemia low oxygen supply is paralleled by decreased extracellular glucose concentrations. Thus, the nutrient uptake with facilitative glucose carriers alone may fail to provide the cell with sufficient glucose supply in ischemic tissue. Alternatively, in response to chronic hypoxia cellular glucose transport may be achieved by Na⁺-coupled transporters, such as SGLT1 [22]. Transport through SGLT1 is indirectly energy-consuming as transport is facilitated through energy maintained through the active transport of Na⁺ across the basolateral membrane by the energy-consuming Na⁺/K⁺ ATPase. At least in some cells, AMPK inhibits Na⁺, K⁺ ATPase activity, which would partially dissipate the electrochemical Na⁺ gradient [23–25].

The present study thus explored whether the metabolic-sensing kinase AMPK similarly stimulates electrogenic uptake of glucose.

Materials and methods

Constructs

For generation of cRNA, constructs were used encoding wild type human SGLT1 (SLC5A1) [26], wild type AMPKα1-HA, AMPKβ1-flag, AMPKγ1-HA [27], constitutively active ^γR70QAMPKγ1-HA [28] and kinase dead mutant ^αK45RAMPKα1-HA [29]. The cRNA was generated as described previously [30].

Voltage clamp in Xenopus oocytes

Xenopus oocytes were prepared as previously described [31]. A total of 4.6 ng of each cRNA encoding AMPKα1-HA + AMPKβ1-flag + AMPKγ1-HA (^WTAMPK), AMPKα1-HA + AMPKβ1-flag +^γR70QAMPKγ1-HA (^γR70QAMPK) or ^αK45RAMPKα1- HA + AMPKβ1-flag +AMPKγ1-HA (^αK45RAMPK) were injected with or without 5 ng cRNA encoding SLC5A1(SGLT1) on the day of preparation of the Xenopus oocytes. All experiments were performed at room temperature 3 days after injection. Two-electrode voltage-clamp recordings were performed at a holding potential of −70 mV. The data were filtered at 10 Hz and recorded with a Digidata A/D-D/A converter and Clampex V.9 software for data acquisition and analysis (Axon Instruments). The control superfusate (ND96) contained 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl₂, 1 mM MgCl₂ and 5 mM HEPES, pH 7.4. Glucose was added to the solutions at a concentration of 10 mM unless otherwise stated. Where indicated, 1 mM ouabain (Sigma, Schnelldorf, Germany) was added to the control superfusate. The pump current generated by readdition of extracellular K⁺ was determined by subsequent exposure of the oocytes for 5 h to K⁺-free perfusate, for a further 5 min to Ba²⁺ (5 mM) containing K⁺-free perfusate and then to Ba²⁺ containing K⁺-containing (5 mM) perfusate. The flow rate of the superfusion was approx. 20 ml/min, and a complete exchange of the bath solution was reached within about 10 s.

Immunofluorescence

Caco2 cells were cultured in DMEM medium containing 4.5 g/l glucose, 20% FBS, 1% L-glutamine, 1% non-essential amino acids and 1% penicillin/streptomycin. The cells were grown on 12-mm glass coverslips (neoLab Migge Laborbedarf-Vertriebs GmbH, Heidelberg, Germany) in 12-well plates (5 × 10⁴ cells/well/coverslip). Two days after plating, the cells reached 80–90% confluence. Prior to immunhistochemistry, the cells were incubated in the absence or presence of 1 mM AMPK activator 5-aminoimidazole-4-carboxamide-1-beta-D-ribofuranoside (AICAR) (Calbiochem, Bad Soden, Germany), of 10 μM A-769662 (Tocris Biosciences) or of 1 mM phenformin hydrochloride (Sigma) for 6 h. The cells were then washed with PBS and fixed for 20 min in 4% paraformaldehyde with PBS/0.1% Triton. The cells were washed again and then blocked in 5% bovine serum albumin with PBS/0.1% Triton for 1 h at room temperature. The cells were incubated overnight at 4°C with goat anti-SGLT1 (1:300, Santa Cruz). After incubation, the cells were rinsed three times with PBS/0.1% Triton and incubated with the secondary FITC anti-goat antibody (1:500, Invitrogen) for 1 h at room temperature. After washing with PBS/0.1% Triton cells were incubated with rabbit anti-alpha tubulin (1:50, Cell Signalling Technology) antibody for 2 h at room temperature. The cells were again rinsed with PBS/0.1% Triton and incubated with Cy5 anti-rabbit (1:400, Jackson ImmunoResearch Laboratories) for 1 h at room temperature followed by washing with PBS/0.1% Triton. The nuclei were stained with DRAQ-5 dye (1:1000, Biostatus, Leicestershire, UK) for 5 min at 37°C. The slides were mounted with Prolong antifade reagent (Invitrogen). The images were taken on a Zeiss LSM 5 EXCITER Confocal Laser Scanning Microscope, equipped with a 405–633 nm laser (Carl Zeiss MicroImaging GmbH, Germany) using a water immersion Plan-Neofluar 63/1.3 NA DIC. Three independent experiments were performed.

Biotinylation and Western blot analysis

10⁹ cells were biotinylated with EZ-link Sulfo-NHS-Biotin (Pierce protein research products, Thermo Scientific) for 30 min at room temperature. Then, the cells were washed three times. The cells were lysed with cell lysis buffer (20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM Na₂EDTA, 1 mM EGTA, 1% Triton, 2.5 mM sodium pyrophosphate, 1 mM beta-glycerophosphate, 1 mM Na₃VO₄, 1 μg/ml leupeptin and 1 mM PMSF, added immediately prior to use) followed by separation of the biotinylated proteins with NeutrAvidin Agarose Resin (Thermo Scientific) at 4°C. The NetrAvidin Agarose Resin was washed three times with ice-cold PBS. The NeutrAvidin agarose-bound protein was separated by addition of 1:5 Laemmli buffer. The samples were boiled for 10 min. The biotinylated protein was subjected to 10% SDS-PAGE gel electrophoresis. Proteins were transferred to a nitrocellulose membrane (Millipore Corp.), and the membranes were then blocked for 2 h at room temperature with 5% non- fat dried milk in Tris-buffered saline (NFDM/TBS) containing 0.1% Tween 20. Incubation with the primary anti-SGLT1 antibody (1:1000, rabbit polyclonal antibody, Millipore) was carried out at 4°C overnight. Specific protein bands were visualized after subsequent incubation with a 1:5000 dilution of anti-rabbit IgG (Cell Signalling) conjugated to horseradish peroxidase and a Super Signal Chemiluminescence detection procedure (GE Healthcare, UK). Afterwards, the same membranes were stripped and reprobed with α-actin antibody (1:1000, rabbit polyclonal antibody, Cell Signalling). The band intensity was determined by Quantity one software (Biorad gel doc system, Chemidoc XRS). Levels of SGLT1 were expressed as the ratio of signal intensity for the target protein relative to that of α-actin.

Statistical analysis

Data are provided as means ± SEM, n represents the number of oocytes investigated. All experiments were repeated with at least two batches of oocytes; in all repetitions qualitatively similar data were obtained. However, the absolute values of the currents may vary appreciably depending on the oocyte batch used (compare Figure 2B and 2D). Data were tested for significance using ANOVA or t-test, as appropriate. Results with p < 0.05 were considered statistically significant.

Results

Constitutively active AMPK stimulates SGLT1-mediated electrogenic glucose transport

Electrogenic glucose transport was minimal in noninjected or water-injected Xenopus oocytes (Figure 1). In Xenopus oocytes expressing SGLT1 (SLC5A1), however, glucose (10 mM) induced an inward current (I_g) reflecting electrogenic entry of Na⁺ and glucose. I_g was significantly enhanced by additional expression of the constitutively active mutant AMP-activated protein kinase ^γR70QAMPK (AMPKα1 + AMPKβ1 + ^γR70QAMPKγ1). Expression of ^γR70QAMPK alone did not lead to appreciable I_g. Thus, ^γR70QAMPK stimulated SGLT1 activity. Wild type AMPK (AMPKα1+AMPKβ1+AMPKγ1) failed to stimulate SGLT1.

Figure 1.  Coexpression of ^γR70QAMPK stimulated electrogenic glucose transport in SGLT1-expressing Xenopus oocytes. (A) Representative original tracings showing glucose (10 mM)-induced current (I_g) in Xenopus oocytes injected with water (a), expressing SGLT1 without (c) or with (d) additional coexpression of wild type AMPK (AMPKα1 + AMPKβ1 + AMPKγ1) or with constitutively-active ^γR70QAMPK (e), or expressing ^γR70QAMPK alone (b). (B) Arithmetic means ± SEM (n = 6–33) of glucose (10 mM)-induced current (I_g) in Xenopus oocytes injected with water (1st bar), expressing SGLT1 without (3rd bar) or with (4th bar) additional coexpression of wild type AMPK (AMPKα1+AMPKβ1+AMPKγ1) or with constitutively-active ^γR70QAMPK (5th bar; AMPKα1+AMPKβ1+^γR70QAMPKγ1), or expressing ^γR70QAMPK alone (2nd bar). ***p < 0.001 indicates statistically significant difference from the absence of SGLT1. ###p < 0.001 indicates statistically significant difference from the absence of ^γR70QAMPK.

Since SGLT1-mediated glucose transport depends on Na⁺/K⁺ ATPase, further experiments addressed a possible participation of the Na⁺/K⁺ ATPase in the AMPK-dependent stimulation of SGLT1 activity. To this end, SGLT1-dependent currents were determined in the presence and absence of the Na⁺/K⁺ ATPase inhibitor oubain (1 mM). The current was 23.6 ± 1.5 nA (n = 12) in oocytes injected with SGLT1 cRNA in the absence and 26.2 ± 1.9 nA (n = 11) in the presence of 1 mM oubain. Expression of SGLT1 together with ^γR70QAMPK increased the glucose-dependent current to 36.2 ± 1.9 nA (n = 12) in the absence and to 38.3 ± 1.7 nA (n = 12) in the presence of 1 mM oubain. Additional experiments were performed elucidating the effect of AMPK expression on the pump current. The current was 2.5 ± 0.3 nA (n = 5) in AMPK-expressing oocytes and 2.6 ± 0.15 nA (n = 5) in water-injected oocytes. Thus, in oocytes AMPK did not appreciably modify the Na⁺/K⁺ ATPase activity and the Na⁺/K⁺ ATPase was not required for the AMPK-dependent stimulation of SGLT1-mediated currents.

The dead mutant ^K45RAMPK did not stimulate SGLT1 and blunted the stimulating effect of AMPK

As shown in Figure 2A & B, electrogenic glucose transport was not significantly modified by the coexpression of inactive ^αK45RAMPK (^αK45RAMPKα1 + AMPKβ1 + AMPKγ1). Coexpression of SGLT1 with ^αK45RAMPK in addition to constitutively active ^γR70QAMPKα1 (AMPKα1 + AMPK α1 + AMPKβ1 + ^γR70QAMPKγ1) yielded glucose-induced currents, which were significantly smaller than the currents observed with ^γR70QAMPK alone. Thus, ^αK45RAMPK exerted a dominant-negative action on AMPK (Figure 2C & D).

Figure 2.  The catalytically inactive mutant ^K45RAMPK did not stimulate SGLT1. (A) Representative original tracings showing glucose (10 mM)-induced current (I_g) in Xenopus oocytes injected with water alone (a) or expressing SGLT1 without (b) and with (c) inactive ^K45RAMPK. (B) Arithmetic means ± SEM (n = 10–15) of glucose (10 mM)-induced current (I_g) in Xenopus oocytes injected with water alone (1st bar) or expressing SGLT1 without (2nd bar) and with (3rd bar) inactive ^αK45RAMPK (^αK45RAMPKα1+AMPKβ1+AMPKγ1). ***p < 0.001 indicates statistically significant difference from the absence of SGLT1. (C) Representative original tracings showing glucose (10 mM)-induced current (I_g) in Xenopus oocytes injected with water alone (a) or expressing SGLT1 without (b) and with (c) inactive ^αK45RAMPK or with constitutively-active ^γR70QAMPK (d) or with both, ^αK45RAMPK and ^γR70QAMPK (e). (D) Arithmetic means ± SEM (n = 9–15) of glucose (10 mM)-induced current (I_g) in Xenopus oocytes injected with water alone (1st bar) or expressing SGLT1 without (2nd bar) and with (3rd bar) inactive ^αK45RAMPK (^αK45RAMPKα1 + AMPKβ1 + AMPKγ1) or with constitutively-active ^γR70QAMPK (4th bar; AMPKα1 + AMPKβ1 + ^γR70QAMPKγ1) or with both, ^αK45RAMPK and ^γR70QAMPK (5th bar). ***p < 0.001 indicates statistically significant difference from the absence of SGLT1. #p < 0.05 indicates statistically significant difference from the absence of ^γR70QAMPK.

AMPK enhanced the maximal current

Kinetics of glucose-induced currents in SGLT1-expressing Xenopus oocytes (Figure 3) yielded a maximal current of 0.098 ± 0.001 μA. The data was fitted to a hyperbola. The glucose concentration needed for halfmaximal current (K_M) was 650 ± 28 μM. The coexpression of constitutively-active ^γR70QAMPK did not significantly modify K_M (581 ± 29 mM), but significantly increased the maximal current (to 0.160 ± 0.002 μA).

Figure 3.  AMPK enhanced the maximal current without appreciably affecting affinity. Arithmetic means ± SEM (n = 16–22) of I_g as a function of glucose concentration in Xenopus oocytes expressing SGLT1 without (open diamonds) and with (closed squares) constitutively-active ^γR70QAMPK (AMPKα1 + AMPKβ1 + ^γR70QAMPKγ1). The difference from the absence of ^γR70QAMPK is significant p < 0.05 at every glucose concentration applied.

SGLT1 protein abundance is enhanced by activation of AMPK

The enhanced SGLT1 activity could result from increased carrier protein abundance in the plasma membrane. To test for that possibility, immunohistochemistry and confocal microscopy were performed. As illustrated in Figure 4A a 6 h treatment of Caco2 cells with the AMPK activator AICAR-5′-Phosphate (1 mM) was followed by a significant increase in the SGLT1 protein abundance. The results were confirmed by biotinylation of the surface proteins and subsequent Western Blotting (Figure 4B).

Figure 4.  Cell surface SGLT1 protein abundance in Caco2 cells was upregulated by the AMPK activator AICAR. A) Confocal microscopy of Caco2 cells incubated for 6 h in medium without (upper panel) or with (lower panel) AMPK activator AICAR (1 mM). The Caco2 cells were subjected to immunofluorescent staining using FITC-conjugated anti-SGLT1 (green), Cy5-conjugated anti-alpha tubulin (red) and DRAQ-5 dye (blue) for nuclear staining. The images are representative for three independent experiments. (B) Original Western blot of SGLT1 protein abundance in Caco2 cells (upper panel) treated as in A. Caco2 surface proteins were biotinylated and subsequently subjected to Western Blotting. Expression of actin served as loading control (middle panel). The lower panel depicts the arithmetic means ± SEM (n = 12) of the SGLT1 density of AICAR (1 mM)-treated Caco2 cells relative to the non-treated Caco2 cells. The SGLT1 density of non-treated Caco2 cells was set to 1. **p < 0.01 indicates statistically significant difference from the absence of AICAR (t-test). (This figure is reproduced in colour in Molecular Membrane Biology online.)

Similarly, activation of AMPK by 1 mM phenformin or 10 μM A-769662 increased the SGLT1 membrane abundance (Figure 5).

Figure 5.  Cell surface SGLT1 protein abundance in Caco2 cells was similarly upregulated by phenformin and A-769662. Confocal microscopy of Caco2 cells incubated for 6 h in medium without (left panel) or with (middle panel) AMPK activator phenformin (1 mM) or with (left panel) AMPK activator A-769662 (10 μM). The Caco2 cells were subjected to immunofluorescent staining using FITC-conjugated anti-SGLT1 (green) and Cy5-conjugated anti-alpha tubulin antibody (red) and DRAQ-5 dye (blue) for nuclear staining. (This figure is reproduced in colour in Molecular Membrane Biology online.)

Discussion

The sodium glucose cotransporter SGLT1 is a high affinity glucose transporter that is expressed predominantly in the brush border of the small intestine and the proximal tubule within the kidney. The uphill reabsorption of glucose is coupled to Na⁺ down its electrochemical potential gradient across the plasma membrane. The regulation of the activity has been demonstrated to be responsive to phosphorylation by kinases including protein kinase A (PKA) and protein kinase C (PKC) [32,33].

The present study demonstrates that the metabolic sensing kinase AMP-dependent kinase (AMPK) stimulates the electrogenic glucose transporter SGLT1 (SLC5A1). A catalytically inactive AMPK (αK45R mutation) and the constitutively active AMPK γR70Q mutation were expressed with the SGLT1 transporter in the heterologous oocyte expression system. Expression of constitutively active AMPK resulted in an increase in the maximal current generated by the electrogenic Na⁺ transport coupled to glucose through the SGLT1 transporter, reflecting a 61% increase in membrane transport. The kinetics demonstrated that this was not due to an increased affinity for glucose but rather the result of an increased maximal transport, while Km was not significantly altered by AMPK coexpression. Importantly, the catalytically inactive and the wild type AMPK had no effect on SGLT1 activity suggesting that the change in membrane transport is indeed due to AMPK activity.

Regulation of SGLT1 activity has previously been shown to be mediated via altering the expression of the co-transporter within the plasma membrane. Activation of PKC is able to decrease membrane SGLT1 through both direct and indirect means leading to decreased glucose uptake whereas PKA activates SGLT-dependent currents [32,33]. To determine whether AMPK-mediated regulation of SGLT1 activity was due to altered membrane trafficking, the colorectal adenocarcinoma cell line Caco-2 was stimulated with the AMPK agonists AICAR, phenformin and A-769662, and an increased membrane staining of SGLT1 was observed. These data therefore suggest that alterations in activity are probably due to membrane insertion.

The ability of AMPK to stimulate cellular glucose uptake has been shown before [34]. However, glucose uptake was considered to be due to activation of the GLUT family of facilitative glucose carriers, predominantly mediated by its effects on GLUT1 and GLUT4 [10–21].

During energy depletion, during which AMPK activity is maximal, glucose uptake through those passive carriers appears, at first glance, to be more appropriate than the Na⁺-coupled secondary glucose uptake, which requires ATP-consuming extrusion of Na⁺ by the Na⁺/K⁺ ATPase if cellular K⁺ loss and subsequent depolarization and cell swelling are to be avoided [34]. The events in energy-depleted cells are difficult to assess, as several mechanisms may compromise Na⁺-coupled and/or facilitative glucose transport. ATP depletion may decrease the hexokinase reaction, which otherwise generates a glucose gradient by trapping cellular glucose. In the alveolar epithelium, activation of AMPK may reduce Na⁺, K⁺ ATPase activity [23–25], which would decline the electrochemical gradient for Na⁺-coupled glucose uptake. In any case, decreasing extracellular glucose concentrations, which parallel the lack of oxygen during ischemia, are expected to affect the glucose uptake through the facilitative glucose carriers more severely than Na⁺-coupled glucose uptake. Even at partially dissipated Na⁺ gradients SGLT1 is more prone to accumulate glucose in the cell at minimal extracellular glucose concentrations than facilitative glucose carriers. As the Na⁺, K⁺ ATPase extrudes 3 Na⁺ ions for one ATP, the energy needed for subsequent extrusion of the cotransported Na⁺ is only a fraction of the energy gained, even if glucose is utilized for ATP generation by glycolysis only.

In conclusion, the AMP-dependent kinase AMPK is a powerful regulator of the Na⁺-coupled glucose transporter SGLT1 and thus does not only stimulate passive but secondary-active cellular uptake of glucose.

Acknowledgements

The authors acknowledge the meticulous preparation of the manuscript by Lejla Subasic and Sari Rübe.

Declaration of interest: This study was supported by the Deutsche Forschungsgemeinschaft, GRK 1302, SFB 773, La 315/13-3 and by a IZKF grant of the University of Tübingen to M.F. (No. 1889-0-0). B.E.K. is supported by the Australian Research Council and is an NHMRC Fellow. The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.