Functional expression of tagged human Na⁺—glucose cotransporter in Xenopus laevis oocytes

Abstract

1. High-affinity, secondary active transport of glucose in the intestine and kidney is mediated by an integral membrane protein named SGLT1 (sodium glucose cotransporter). Though basic properties of the transporter are now defined, many questions regarding the structure- function relationship of the protein, its biosynthesis and targeting remain unanswered. In order to better address these questions, we produced a functional hSGLT1 protein (from human) containing a reporter tag.

2. Six constructs, made from three tags (myc, haemaglutinin and poly-His) inserted at both the C- and N-terminal positions, were thus tested using the Xenopus oocyte expression system via electrophysiology and immunohistochemistry. Of these, only the hSGLT1 construct with the myc tag inserted at the N-terminal position proved to be of interest, all other constructs showing no or little transport activity. A systematic comparison of transport properties was therefore performed between the myc-tagged and the untagged hSGLT1 proteins.

3. On the basis of both steady-state (affinities for substrate (glucose) and inhibitor (phlorizin) as well as expression levels) and presteady-state parameters (transient currents) we conclude that the two proteins are functionally indistinguishable, at least under these criteria. Immunological detection confirmed the appropriate targeting of the tagged protein to the plasma membrane of the oocyte with the epitope located at the extracellular side.

4. The myc-tagged hSGLT1 was also successfully expressed in polarized MDCK cells. α-Methylglucose uptake studies on transfected cells showed an exclusively apical uptake pathway, thus indicating that the expressed protein was correctly targeted to the apical domain of the cell.

5. These comparative studies demonstrate that the myc epitope inserted at the N-terminus of hSGLT1 produces a fully functional protein while other epitopes of similar size inserted at either end of the protein inactivated the final protein.

The Na⁺-dependent pathway for glucose transport is a basic feature of intestinal and renal physiology. This transport system plays a major role in the homeostasis of glucose by mediating intestinal absorption as well as renal reabsorption of glucose. The gene encoding the protein responsible for this function, named SGLT1, was first cloned by Hediger and co-authors in 1987 from a rabbit intestinal cDNA library, using Xenopus laevis oocytes as the expression system (Hediger et al. 1987). The genes encoding SGLT1 proteins from different species were then cloned from intestine as well as from kidney cDNA libraries. With time, the Na⁺ solute cotransport system has expanded to a family of related proteins (Turk & Wright, 1997) comprising cotransporters for different solutes (sugars, amino acids, ions and other substrates yet to be determined) found in living forms ranging from bacteria to man. The identification of SGLT1 remains a milestone in the field of Na⁺-dependent cotransporters, and today is considered a prototype as well as a reference for other cotransport systems.

Our understanding of SGLT1 has increased dramatically in recent years through approaches such as electrophysiology on Xenopus oocytes, which enables the characterization of steady-state as well as presteady-state events associated with SGLT1 function. SGLT1 is now believed to present 14 membrane-spanning domains and one glycosylation site, and to have both N- and C-termini facing the extracellular side of the membrane (Turk et al. 1996). Recent data indicate that the site for glucose binding is located in the C-terminal half of the protein, between transmembrane domains 10 and 14, while sodium binds to the N-terminal half (Panayotova-Heiermann et al. 1997). Our understanding of SGLT1 has also benefited from a rare human disease, glucose-galactose malabsorption syndrome (GGM), in which affected patients present mutated SGLT1 along with partial or total loss of SGLT1 function. Currently, over 30 different mutations are known (Martin et al. 1996). Of these, 19 are missense mutations thus demonstrating the susceptibility of the protein to minor modifications. Mistargeting of the protein resulting in its intracellular trapping has been suggested as a consequence of some of these mutations (Lostao et al. 1995; Martin et al. 1996).

A powerful tool for investigation for both cell physiology and structural studies is provided by tagging the native protein with an immunoreactive reporter sequence, which helps to identify the protein in Western blot analysis or immunohistochemistry. The use of tags such as myc or haemaglutinin (HA) epitopes was shown to be valuable for studying both cell targeting (Robertson et al. 1995; DeWitt et al. 1996) and the structural biochemistry of proteins, including transporters, in various expression systems (Lemas et al. 1994; DeWitt et al. 1996; Yang & Verkman, 1997). In 1996, Turner and co-workers successfully expressed SGLT1 tagged with the VSVG (vesicular stomatitis virus G protein) epitope in Caco-2 (human colon) cells (Turner et al. 1996). Although functional expression was achieved, the VSVG-SGLT1 was mistargeted to the basolateral membrane. It was concluded that the VSVG signal for basolateral targeting overwhelmed that for apical targeting which is already present in the sequence of SGLT1 (Kong et al. 1993). Recently Vayro and co-workers (Vayro et al. 1998) tagged the rabbit SGLT1 with the myc sequence in the N-terminal position. When expressed in COS (African green monkey kidney) cells, myc-SGLT1 demonstrated an affinity for α-methylglucose (α-MG) similar to that found for the wild-type protein, as well as a similar selectivity amongst different sugars. Unfortunately, the topic of the apical targeting of the protein could not be addressed in COS cells since these cells do not produce typical apical membrane polarization with brush border membranes as seen in cell types naturally expressing SGLT1.

The present paper reports our studies on the functional expression of tagged human SGLT1 (hSGLT1) with the specific goal of comparing many functional characteristics between the tagged and the wild-type proteins.

METHODS

DNA constructs and RNA synthesis

The human Na⁺-glucose transporter hSGLT1 cDNA (Bissonnette et al. 1996b; Chen et al. 1997) was inserted into the pBS vector (Stratagene, San Diego, CA, USA) along with the poly-A tail from the pSP64 poly-A vector (Promega, Madison, WI, USA). The final construct was linearized with EcoRI and RNA was synthesized using T3 polymerase according to the manufacturer's recommendations (Ambion, Austin, TX, USA). Tagged hSGLT1 was constructed by subcloning hSGLT1 into the pJ3M vector for addition of the myc tag (MEQKLISEEDL; Sells & Chernoff, 1995), into pJ3H for addition of the HA tag (MYYPYDVPDYA; Sells & Chernoff, 1995) or into pJ3XH for addition of the poly-His tag (MHHHHHHHHHH). For construction of the pJ3XH vector, two oligonucleotides (5′-TCGACGCCGCCATGGCCGGATCACATCACCATCACCATCACCATCACCATCACT-3′ and 5′-CTAGAGTGATGGTGATGGTGATGGTGATGGTGATGTGATCCGGCCATGGCGGCG-3′) were synthesized (AlphaDNA, Montréal, QC, Canada), phosphorylated and hybridized together after heat denaturation. The double-stranded product was ligated into Sal I-Xba I-cleaved pJ3M, producing a construct (pJ3XH) containing a 5′ 10-His epitope tag. Insertion of tags also resulted in the addition of two amino acids (Ser and Arg) between the amino tags and the initial Met of hSGLT1.

Oocyte preparation, injection and maintenance

Xenopus laevis frogs (Xenopus One, Ann Arbor, MI, USA) were anaesthetized with 2-aminobenzoic acid ethyl ester and ovary nodes were surgically removed. Oocytes were individually separated, as previously described (Chen et al. 1997), except that the defolliculation procedure was performed in Ca²⁺-free Barth's solution. After surgery, the frogs were allowed to recover and eventually reused 3 or 4 times. A 4 month recovery period was respected between operations. When required, the frogs were killed by prolonged incubation in anaesthetic solution. Manipulations (anaesthesia and surgery) and caring of the animals were performed in accordance with the Canadian guidelines and ethics committee from the Université de Montréal. After a 24 h recovery period, oocytes were injected with 5 or 50 ng of cRNA (as specified in the figure legends) at a concentration of 1 μg μl⁻¹ in water using a Drummond microinjector (Broomall, PA, USA) and further incubated in Barth's solution (mm: 88 NaCl, 3 KCl, 0.82 MgSO₄, 0.41 CaCl, 0.33 Ca(NO₃)₂ and 5 Hepes, pH 7.6) supplemented with penicillin (100 u ml⁻¹), streptomycin (0.1 mg ml⁻¹), kanamycin (0.1 mg ml⁻¹), sodium pyruvate (2.5 mm) and 5 % horse serum. Control oocytes were injected with an equivalent volume of water. Under these conditions, oocytes were viable for up to 2 weeks but were studied at 5-7 days after injection.

Electrophysiological studies

Oocyte currents were measured with a two-microelectrode voltage-clamp technique as described by Chen et al. (1997). Current and voltage electrodes were filled with 1 M KCl and resistances were between 4 and 14 MΩ. The bath reference and current electrodes were Ag-AgCl pellets. The oocyte bath was continuously perfused with Barth's solution without any supplement. Only oocytes with resting potentials more negative than -30 mV were considered. Unless otherwise specified in the figure legends, the pulse protocol designed consisted of 11 successive 250 ms pulses separated by 500 ms periods at the resting potential (-50 mV) covering a voltage range of +50 to -175 mV scaled in 25 mV increments. Current and voltage data were analysed by averaging values in a 20 ms window starting at 210 ms after initiation of the pulse. hSGLT1 activity was assayed using 5 mm D-glucose as the substrate. For determination of kinetic parameters, glucose was added to Barth's solution in increasing concentrations up to 5 mm (0.1, 0.2, 0.5, 2 and 5 mm). No compensation for osmolarity was made since the variations were minimal and without any significant effects. For determination of phlorizin (Pz) affinity, the inhibitor was presented in concentrations up to 2 μm (0, 0.05, 0.1, 0.2, 1 and 2 μm) in Barth's solution containing a constant 0.5 mm glucose concentration.

Presteady-state currents were estimated using a similar protocol; 50 ms pulses were applied at 25 mV increments between +50 and -175 mV. The electrodes used displayed resistances between 1 and 2 MΩ and current values determined were averaged from three subsequent runs on the same oocyte. Non-hSGLT1-induced charge movements were determined by testing in the presence of 0.2 mm Pz in Barth's solution. Specific hSGLT1 presteady-state currents were determined by point-to-point subtraction of current in the presence of Pz from that in the absence of inhibitor (Chen et al. 1996).

Pz binding

Pz binding studies were performed with control (non-injected), hSGLT1-injected and myc-hSGLT1-injected oocytes (50 ng cRNA) at 5-7 days after injection. Oocytes were pre-incubated in groups (8-12 per vial) at room temperature for 15 min in non-supplemented Barth's solution prior to assay. The pre-incubation medium was replaced with 0.5 ml of medium containing tracer amounts of [³H]Pz (New England Nuclear, Boston, MA, USA), creating a 0.036 μm Pz concentration in the medium (2 μCi ml⁻¹, specific activity 55 Ci mmol⁻¹), and oocytes were further incubated for 15 min. Binding was stopped by removing the incubation medium and rinsing 4 times with 2 ml of ice-cold Barth's solution. Oocytes were then transferred individually into scintillation vials and dissolved in 0.2 ml 10 % SDS for 4 h, to which was then added 5 ml of Beta-Blend scintillation cocktail (ICN Pharmaceuticals, Montréal, QC, Canada). Tritium content was determined using a Beckman LS6000SC scintillation counter. Non-specific Pz binding was evaluated by tracer displacement using a saturating concentration of inhibitor (50 μm).

Determination of Pz K_i values

The K_i values for Pz were determined by measuring the inhibition of the current induced by 5 mm glucose in the presence of increasing amounts of the inhibitor (protocol for electrophysiological determination was as described above). The data were analysed according to the equation for competitive inhibition and the glucose K_m values, which are required for this analysis, were determined on the same oocytes.

MDCK transfection and transport

MDCK cells strain II were routinely grown in 75 mm plastic flasks using high-glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 7.5 % cool-calf II serum (Sigma) at 37°C in a 95 % air-5 % CO₂ atmosphere. Culture medium was changed every other day. For the transfection procedure, cells were seeded onto coverslips in 6-well plates (Falcon) at a density of 0.35 × 10⁶ cells per well and transfected the next day (about 40 % of surface covered). Transfection was performed using the calcium phosphate precipitation technique with both pJ3M-hSGLT1 vector (4 μg per well) for expression of myc-hSGLT1 and pcDNA3.1/neo vector (1 μg per well) (Invitrogen, Carlsbad, CA, USA) for selection of transfectants. Two days after transfection, cells were presented with medium containing a low serum concentration (0.5 %) along with neomycin (G-418, 0.5 mg ml⁻¹) for selection of transfectants. Four days after transfection, cells were restored to normal serum conditions (7.5 %) with neomycin to maintain selection. Transfected MDCK cells were then maintained and propagated (without subcloning) in order to generate material for transport assays.

α-Methylglucose (α-MG) uptake on monolayers was performed according to Bissonnette et al. (1996a). MDCK cells were seeded at 0.3 × 10⁶ cells onto filters (0.4 μm pore size; Falcon) and assayed for α-MG transport on days 8-9 (5 days post-confluency) in order to obtain a fully polarized monolayer. The integrity of the cell monolayers was assessed by evaluation of transepithelial resistance. Uptake was performed at room temperature for 30 min using 0.5 μCi ml⁻¹[U-¹⁴C]α-MG (New England Nuclear) in a modified Krebs solution (see Bissonnette et al. 1996a). Filters were presented with 1 ml solution at the apical side and 2 ml at the basolateral side; radiolabelled substrate was supplied to either the apical or the basolateral side. Corrections for the non-specific fraction of uptake were made by subtraction of values for similar uptakes in the presence of 0.5 mm Pz. Data are expressed as counts per minute per filter.

Immunofluorescence studies

Oocytes

Immunofluorescence detection on oocytes was performed using both intact oocytes and slices of fixed oocytes. Oocytes were injected with 5 ng of either myc-tagged or untagged hSGLT1 and processed for immunofluorescence after 5 days of expression. In all cases, only oocytes that tested positive for Na⁺-glucose activity with glucose-specific currents (I_Glc) greater than 1.2 μA at -50 mV were selected. For immunofluorescence on slices, oocytes were rinsed 3 times with Barth's solution and fixed for 15 min in ice-cold methanol solution containing 1 % formaldehyde. Fixed oocytes were rinsed again 3 times and incubated overnight in a 30 % sucrose Barth's solution. Oocytes were embedded in Tissue-tek embedding medium (Sakura Finetek, USA), frozen, sliced (10 μm thickness) on a cryostat and mounted on slides. The samples were either immediately processed or stored at -20°C for no more than 2 weeks. Immunolabelling was performed as follows. Slices were rinsed in three consecutive PBS baths for 5 min periods containing, in sequence, low-high-low amounts of NaCl (low, 8 g l⁻¹; high, 27 g l⁻¹) and blocked for 30 min at room temperature against non-specific binding of antibody with 2 % BSA solution in PBS. Incubation with anti-myc antibody (from clone 9E10, Biomol Research, Plymouth Meeting, PA, USA) at a dilution of 1:500 in PBS was performed in a wet chamber at room temperature for 1 h, followed by three rinses with PBS in the sequence described above. Incubation with secondary antibody (anti-mouse IgG coupled to fluorescein isothiocyanate (FITC), 1:250; Jackson ImmunoResearch Laboratories Inc., West Grove, PA, USA) was performed as for the primary antibody and slices were rinsed accordingly. An anti-quenching agent (propyl gallate in a 1:1 solution of glycerol/100 mm Tris pH 8.0) was then added prior to observation. The technique for detection of myc-hSGLT1 on intact oocytes is based on that of Firsov et al. (1996). Intact oocytes (non-fixed) were rinsed twice in Barth's solution supplemented with 5 % horse serum and incubated in microfuge tubes for 15 min on ice. The medium was then replaced with a similar solution containing anti-myc antibody (1:500) and incubated on ice for 1 h. Oocytes were then rinsed 5 times with Barth's solution plus serum and incubated with secondary antibody (same as above) for 1 h on ice. Oocytes were again rinsed 5 times and mounted on slides onto which spacers had been placed in order to avoid breaking the oocytes when covered with a coverslip. This system provided oocytes that were slightly compressed, and were found to be adequate for microscopic observation.

MDCK cells

For immunofluorescence assays, transfected MDCK cells (2 passages after transfection and maintained in G-418-supplemented medium) and non-transfected controls were seeded at low density (0.35 × 10⁶ cells per well) and maintained for 4 days prior to assay. Cells on coverslips were treated as described in the previous section for oocyte slices. All steps of this procedure, except for fixation, were performed at room temperature. Coverslips were rinsed twice with PBS and cells were fixed for 20 min at -20°C with a 1 % formaldehyde solution in methanol. Fixed cells were then rinsed twice with PBS, incubated for 5 min with 50 mm NH₄Cl in PBS and rinsed twice again with PBS. The cells were then immediately processed for immunofluorescence staining. Non-specific site blocking was done using a 2 % BSA solution in PBS for 20 min. Incubation with the primary antibody (anti-myc, 1:500 dilution in blocking solution) was performed in a wet chamber for 1 h. Coverslips were then rinsed 5 times with PBS and blocked again with 2 % BSA in PBS. Incubation with secondary antibody (anti-mouse IgG coupled to FITC, 1:250 in blocking solution) was performed as for anti-myc labelling. Coverslips were then rinsed 5 times and mounted on slides with the anti-quenching solution mentioned above.

Plasma membrane preparation of oocytes and Western blot detection

Preparation of an enriched plasma membrane fraction was performed according to Geering et al. (1989), yielding a fraction of about 20 μl from 40 oocytes sampled. The samples were immediately processed for SDS-PAGE (3 μl per lane) and Western blot detection. Proteins were transferred from the gel onto Hybond nitrocellulose and immediately stained with Ponceau Red to validate the quality of the transfer procedure. The nitrocellulose membrane was then blocked for non-specific binding using 5 % non-fat milk in TBS (blocking solution) for 1 h at room temperature. Incubation with the primary antibody (anti-myc, 1:500 in blocking solution) was performed overnight at 4°C. The membrane was then washed 4 times for 15 min each in TBS with 0.1 % Tween and blocked again for 1 h at room temperature. Secondary antibody incubation (HRP-conjugated anti-mouse IgG, 1:1000 in blocking solution) was also performed at room temperature for 1 h after which the membrane was washed again (4 × 15 min). Detection of HRP activity was performed using an enhanced chemiluminescence kit (Phototope-HRP, New England Biolabs, MA, USA) and Biomax MR Kodak film.

Data analysis

Data for electrophysiological studies were collected and analysed using commercial software (RC Electronics, Santa Barbara, CA, USA). Curve fitting of current values for determination of both K_m (glucose) and K_i (Pz) was performed using Fig. P software (Biosoft, Ferguson, MO, USA). The area under the curve of presteady-state currents was measured to obtain Q, the displaced charge. Q was fitted with a Boltzmann equation:

(1)

where Q_hyp and Q_max are Q values at the limit of the large hyperpolarizing voltage and the maximal amplitude of the displaced charge, respectively; z is the valency of the equivalent charge that would cross the entire membrane electrical field; F and R are the Faraday and gas constants and T is the absolute temperature; and V_0.5 is the voltage for which Q=Q_hyp+Q_max/2.

All experiments were performed at least 3 times on oocytes from different donors. The data presented are typical results and were shown to be reproducible within different donors.

RESULTS

Activity of hSGLT1 constructs

A total of six constructs were made out of hSGLT1 using three different reporter tags (myc, HA and poly-His) inserted at either the N- or C-terminal ends of the protein. These constructs had a 12 amino acid sequence (10 for the tags and 2 for the link to hSGLT1) in addition to the original 662 residues found in hSGLT1. Transport activity of the constructs was compared with that found with the wild-type hSGLT1 using electrophysiological methods. No activity higher than our detection limit (2-3 nA) could be detected for any constructs with C-terminal tags. Efforts were thus re-centred around constructs with tags in the N-terminal position. When expressed in Xenopus oocytes, hSGLT1 can induce currents well over 1 μA in amplitude (at -175 mV) once maximal expression levels are obtained, a condition achieved after 5 days of incubation with an injection of 5 ng cRNA (Fig. 1, •). The myc-hSGLT1 construct, when expressed under the same conditions, generated I-V curves with a similar profile and amplitude (Fig. 1, ^). On the other hand, HA-hSGLT1 barely gave 2 % of the signal found with wild-type hSGLT1 (50 nA at -175 mV) while the poly-His construct gave no significant signal at all (inset to Fig. 1). In view of these results, the HA-tagged protein construct was investigated further but the poly-His construct was removed from further consideration. The myc-hSGLT1 construct, the only fully active form of tagged hSGLT1, was on the other hand fully investigated along with the wild-type protein.

Figure 1. I-V plots of glucose-induced currents for wild-type and N-terminally tagged hSGLT1.

Oocytes were injected with 5 ng of either non-tagged (•) or tagged (^, myc; ▵, poly-His; □, HA) hSGLT1 cRNA. Oocytes were incubated for 5 days prior to electrophysiological assays. Glucose-induced currents were determined as described in Methods using a 5 mm glucose Barth's solution. Determination of specific currents was made by subtraction of currents measured in the absence of substrate from those in the presence of substrate. Inset, I-V plots for poly-His- and HA-tagged hSGLT1 shown on an expanded scale. Data are means ±s.e.m. of 5 oocytes from a single donor.

Comparison of wild-type hSGLT1 with myc-hSGLT1

In order to better characterize the properties of myc-hSGLT1, we performed a systematic comparison of a host of key functional parameters associated with hSGLT1 activity between the wild-type and the myc-tagged forms of the protein. Figure 2A illustrates I-V curves measured at increasing glucose concentrations (0.1-5 mm). Data collected from these curves were used to derive the affinity of the cotransporters for glucose. In Fig. 2B, current values generated at -50 and -175 mV are plotted against glucose concentration and fitted to the Michaelis-Menten equation:

(2)

where I_Glc represents the current induced by the addition of glucose, and I_max, K_m and [S] are the maximal current, the glucose affinity constant and glucose concentration, respectively. The K_m values thus estimated were found to be 50 % higher at -50 mV (0.76 ± 0.03 and 0.74 ± 0.04 mm for hSGLT1 and myc-hSGLT1, respectively; means ±s.e.m.) than those found at -175 mV (0.56 ± 0.04 and 0.56 ± 0.02 mm for hSGLT1 and myc-hSGLT1, respectively). When comparing glucose affinities between the wild-type and the myc-hSGLT1 proteins (Fig. 2C), it is clear that (1) the K_m values for glucose are basically identical for the two forms of cotransporter and that (2) the two forms of hSGLT1 demonstrate the same behaviour of glucose affinity in response to membrane potential, where K_m values are essentially voltage insensitive at potentials between -75 and -175 mV but exhibit definite increases at more positive membrane potentials. The small difference observed at 0 mV is not a significant finding but rather reflects the difficulties of fitting eqn (2) to the small current amplitudes observed at this potential. For comparison, glucose affinity constants were also determined for the HA-tagged hSGLT1 (Fig. 2C). When compared with wild-type hSGLT1, the affinity constant for glucose was shown to be significantly lower at very negative potentials (-125 to -175 mV). Furthermore, the K_m value for HA-hSGLT1 was observed to be more voltage sensitive in the range -150 to -50 mV.

Figure 2. Determination of the affinity (K_m) of hSGLT1 for glucose.

Oocytes injected with 5 ng cRNA encoding hSGLT1, myc-hSGLT1 or HA-hSGLT1 were tested for glucose-induced currents on day 5. A, glucose-specific currents from hSGLT1-injected oocytes, evaluated at increasing substrate concentration (0.1, 0.2, 0.5, 2 and 5 mm). Current in the absence of substrate was subtracted from total current in order to isolate the glucose-specific portion. B, plots of currents determined at -175, -50 and 0 mV against glucose concentration for oocytes injected with hSGLT1 (•) or myc-hSGLT1 (^) cRNA. The values were fitted according to the Michaelis-Menten equation (eqn (2)). C, affinities (K_m values) for glucose determined at 25 mV increments for membrane potentials between 0 and -175 mV. Values are for oocytes injected with hSGLT1 (•), myc-hSGLT1 (^) or HA-hSGLT1 (□). Data are means ±s.e.m. of 5 oocytes from a single donor.

Evaluation of affinities for Pz was performed using a similar strategy; I-V curves were generated (Fig. 3A) in the presence of increasing amounts of Pz (0-2 μm) while maintaining a constant concentration of glucose (0.5 mm). The current values thus determined were plotted against inhibitor concentration and fitted to the equation for competitive inhibition (Segel, 1975) (Fig. 3B):

(3)

where [I] represents the concentration and K_i the affinity constant of the inhibitor. The K_m value for glucose, required for the determination of the K_i value for Pz, was determined on the same oocytes. The derived K_i values were evaluated at different membrane potentials (Fig. 3C) and were found not to vary in a systematic manner across the range used (-25 to -175 mV). Given the fact that the estimation of K_i depends on the accuracy of the K_m determination, the K_i values for Pz for myc-hSGLT1 (0.13 ± 0.01 μm) were found to be similar to those for the wild-type protein (0.21 ± 0.01 μm, see Table 1). Pz activity was also evaluated through the binding of radiolabelled inhibitor to intact oocytes. As shown in Fig. 4, specific binding of [³H]Pz (assessed by displacement by excess amounts of unlabelled inhibitor) was readily achievable. Furthermore, the specific binding recorded closely followed the levels of cotransporter protein found in the oocyte, as evaluated through glucose-induced currents.

Figure 3. Determination of the affinity (K_i) of hSGLT1 for the inhibitor Pz.

The experimental procedure for determination of the affinity for Pz was similar to that for determination of glucose affinity presented in Fig. 2. Oocytes were injected with 5 ng cRNA encoding hSGLT1 or myc-hSGLT1 and tested for Pz inhibition of glucose-induced currents after 5 days of incubation. A, I-V plots of glucose-induced currents (0.5 mm) in hSGLT1-injected oocytes in the presence of increasing amounts of Pz (0, 0.05, 0.1, 0.2, 1 and 2 μm). Non-specific current, as determined in the absence of substrate, was subtracted from total current. Data are means ±s.e.m. of 5 oocytes from a single donor. B, inhibition curves determined at two membrane potentials (-50 and -175 mV) for both hSGLT1-injected (•) and myc-hSGLT1-injected (^) oocytes. Data were plotted against Pz concentration and fitted with the equation for competitive inhibition (eqn (3)). The K_m values for glucose required in the equation were determined for the same oocytes at each potential. C, K_i values for Pz determined at 25 mV steps for membrane potentials between -175 and -25 mV for oocytes injected with hSGLT1 (•) and myc-hSGLT1 (^).

Table 1.

Functional parameters for wild-type hSGLT1, myc-hSGLT1 and HA-hSGLT1

	hSGLT1	myc-hSGLT1	HA-hSGLT1
Steady-state parameters
Km glucose (mm)	0.56 ± 0.04	0.56 ± 0.02	0.22 ± 0.08
Imax (μA)	1.52 ± 0.01	1.43 ± 0.03	0.091 ± 0.007
K1 Pz (μm)	0.21 ± 0.01	0.13 ± 0.01
Substrate specificity	Glc>α-MG> 3-OMG>>	Glc>α-MG>3-OMG>>
	2-DG>Man	2-DG>Man
Presteady-state parameters
Qmax (nC)	25.8 ± 1.0	26.8 ± 0.5
z value	0.77 ± 0.03	0.76 ± 0.06
Qmax/Imax(nC μA−1)	9.9	10.7

Steady-state parameters were determined at −175 mV (see Methods). Values are means ±s.e.m. of 5 oocytes from a single donor. Glc, glucose; α-MG, α-methylglucose; 3-OMG, 3-Omethylglucose; 2-DG, 2-deoxyglucose; Man, mannitol.

Figure 4. Binding of Pz to oocytes.

Binding of [³H]Pz was performed on control (non-injected), hSGLT1- or myc-hSGLT1-injected (50 ng cRNA) oocytes. Oocytes were incubated in groups (10 oocytes per vial) at room temperature for 15 min in Barth's solution containing the tracer. Oocytes were then rinsed 4 times with ice-cold Barth's solution and counted individually (see Methods). Data are means ±s.e.m. of counts per minute (CPM) from 8-10 oocytes from the same donor. □, binding values for specific binding as determined by subtraction of the non-specific fraction from total binding (see Methods). , corresponding values of currents elicited by addition of 5 mm glucose at -175 mV.

Determination of presteady-state activities enables evaluation of some functional parameters for transporters such as the number of active transporters in the membrane and, more importantly, the net charge transferred as the transporter changes its conformation (Parent et al. 1992; Chen, 1996). In Fig. 5A, a typical transient current profile is shown for myc-hSGLT1. The currents presented are the Pz-sensitive current in the absence of glucose for imposed membrane potentials ranging between -175 and +50 mV. In Fig. 5B, a plot of the charge moved against membrane potential describes typical Boltzmann curves for both wild-type and myc-tagged hSGLT1. It is readily observable that the two proteins exhibit similar charge movement, as can be seen by the Q and z values obtained (Table 1).

Figure 5. Evaluation of transient currents.

A, transient current profiles determined at membrane potentials varying between -175 and +50 mV. Presteady-state currents found in normal Barth's solution were corrected for non-specific activity by subtraction of currents determined in the presence of 0.2 mm Pz. B, comparison of transient currents between oocytes expressing hSGLT1 (•) and myc-hSGLT1 (^). Oocytes were injected with 5 ng cRNA and tested on day 5. Pulse protocol and calculations of net charge transfer are described in Methods. Values are means ±s.e.m. for 3 oocytes from the same donor.

The final element of comparison consisted of the determination of sugar substrate specificity. Figure 6 shows the specific currents elicited by a variety of sugars that are known substrates for different Na⁺-dependent and Na⁺-independent sugar transporters. No discrepancy was found between the two types of hSGLT1, either in the sequence of sugar specificity or in the level of current induced.

Figure 6. Evaluation of substrate specificity.

Both hSGLT1 (□) and myc-hSGLT1 () were tested for currents induced by various sugars (Man, mannitol; 2-DG, 2-deoxyglucose; 3-OMG, 3-O-methylglucose; α-MG, α-methylglucose; and GLC, glucose. Oocytes injected with 5 ng cRNA were tested on day 5 with 5 mm substrate. Data are presented as the percentage of current found with glucose and were corrected for non-specific current by subtraction of current in the absence of sugar. Values are means ±s.e.m. of 5 oocytes from a single donor.

Immunodetection of myc-hSGLT1 in oocytes

The myc-hSGLT1 construct expressed in oocytes can be detected using either Western blot detection on purified membrane preparations or in situ immunolabelling techniques such as immunofluorescence. Purified preparations of oocyte plasma membrane (Geering et al. 1989) were used for Western blot detection with both types of hSGLT1 (Fig. 7). Using the anti-myc antibody, a unique band spanning from 72 to 80 kDa was revealed, corresponding to myc-tagged hSGLT1. On the other hand, no band was found when wild-type hSGLT1 was expressed.

Figure 7. Western blot analysis of myc-hSGLT1 with anti-myc antibody.

Plasma membrane fractions were purified from hSGLT1-injected (A) or myc-hSGLT1-injected (B) oocytes (50 ng cRNA) after 5 days of incubation. All oocytes had tested positive for glucose-induced currents of at least 1.5 μA. Samples loaded (3 μl per lane) were equivalent to 15 oocyte extracts and were shown to contain an equivalent amount of protein by Ponceau Red staining. Molecular masses indicated on the left (in kDa) were obtained using prestained molecular mass standards from New England Biolabs.

Figure 8 shows in situ detection of myc-hSGLT1 on semi-thin sections (10 μm) of oocytes injected with either wild-type (Fig. 8A) or myc-tagged (Fig. 8B) hSGLT1. When using the 9E10 antibody against the myc epitope, immunofluorescence could only be detected with tagged hSGLT1. This signal was found to be largely localized to the plasma membrane domain, though some intracellular labelling could also be observed. As expected, no specific fluorescence could be detected in hSGLT1-injected oocytes. Immunofluorescence detection has also proved to be successful in vivo using non-fixed oocytes. With this approach, where non-fixed, non-permeabilized oocytes were tested, immunofluorescence could be visualized on the entire surface of myc-hSGLT1-injected oocytes (Fig. 8D) while no signal could be detected on wild-type hSGLT1-injected oocytes (Fig. 8C). This clearly indicates that the epitope tag is accessible from the extracellular side of the membrane, in agreement with the membrane topology recently proposed (Turk et al. 1996).

Figure 8. Immunofluorescence detection of myc-hSGLT1 in oocytes with anti-myc antibody.

In situ identification of myc-hSGLT1 on semi-thin sections (A and B) or on untreated, intact oocytes (C and D). Oocytes were injected with hSGLT1 (A and C) or myc-hSGLT1 (B and D) and tested positive for glucose-induced currents (I_Glc > 1.5 μA at -175 mV). Anti-myc antibody was used at 1:500 dilution and FITC-conjugated anti-mouse IgG at 1:250. Images were taken with a Nikon × 40 oil-immersion objective. Scale bar for A-D, 20 μm.

Expression of myc-hSGLT1 in MDCK cells

MDCK cells were transfected with the pJ3M-hSGLT1 vector, which encodes myc-hSGLT1, and tested for both immunofluorescence and α-MG transport. Evidence for myc-hSGLT1 expression in MDCK cells can be obtained in situ through immunofluorescence detection. Following transient transfection, positive transfectants were easily visualized using anti-myc antibody (Fig. 9A) while no signals were detected in controls (Fig. 9B). Since adequate targeting of SGLT1 to the apical membrane is only present after confluency is reached (Suzuki et al. 1996), uptake studies on confluent monolayers were assayed. Uptake of α-MG into confluent cultures of control and transfected MDCK cells grown on permeable supports showed expression of myc-hSGLT1 with apical polarization (Fig. 10). Pz-sensitive α-MG uptake was stimulated 48-fold above that of control MDCK cells when tracer substrate was presented at the apical side. On the other hand, no specific uptake was detected when the substrate was presented at the basolateral side either in control or in myc-hSGLT1-transfected cells.

Figure 9. Immunofluorescence detection of myc-hSGLT1 in MDCK cells.

MDCK cells transfected with myc-hSGLT1 (A) and non-transfected cells (B) were incubated with anti-myc antibody (1:500) and FITC-conjugated anti-mouse IgG at 1:250. Cells were seeded on coverslips at 0.35 × 10⁶ per well in 6-well plates and tested for immunofluorescence on day 4. Images were taken with a Nikon × 40 oil-immersion objective. Scale bar for A and B, 20 μm.

Figure 10. α-MG uptake in transfected MDCK cells.

Polarized MDCK cells transfected with myc-hSGLT1 were tested for [¹⁴C]α-MG uptake activity and compared with non-transfected cells. Cells were grown on filters for 5 days after confluency and tested for both apical (□) and basolateral () uptake. The non-specific uptake, as determined by addition of 0.2 mm Pz in the uptake medium, was subtracted from total uptake measured in the absence of Pz. A 48-fold accumulation of α-MG was observed in myc-hSGLT1-transfected cells when substrate was presented on the apical side. No specific uptake of substrate was demonstrated into control cells or from the basolateral compartment of myc-hSGLT1-expressing cells.

DISCUSSION

In the past decade, our understanding of SGLT1 structure and functionality, including overall topology (Turk et al. 1996) and determination of glucose and Na⁺ binding sites (Panayotova-Heiermann et al. 1997), has greatly improved. The SGLT1 protein has been expressed in a number of different systems including Xenopus oocytes and MDCK, COS, Caco-2 and Sf9 (Spodoptera frugiperda ovary) cells (Smith et al. 1992; Kong et al. 1993; Turner et al. 1996; Suzuki et al. 1996; Vayro et al. 1998). In all cases, the functional characteristics reported were shown to be in accordance with those of the native protein as found in the intestine or the kidney. The basic characteristics of SGLT1, such as electrogenicity, sodium dependency, substrate specificity and Pz sensitivity, were always respected. The present study on tagged SGLT1 was motivated by the need for new tools of investigation but also due to a recent study where VSVG-tagged hSGLT1 (tag positioned at the C-terminus of the protein) was mistargeted (Turner et al. 1996). In this study, a VSVG-SGLT1 protein was expressed in both COS-1 and Caco-2 cells. The tagged SGLT1 expressed in COS cells displayed typical characteristics of the native protein, though a reduction in substrate affinity was found. Surprisingly, once expressed in Caco-2 cells, the protein was mistargeted to the basolateral domain. Although the choice of Caco-2 cells for expression of SGLT1 is questionable, since this cell line already displays a number of Na⁺-dependent α-MG transport pathways (Blais et al. 1987; Bissonnette et al. 1996), the revealed mistargeting is somewhat intriguing. Results from the laboratory of A. Berteloot (Université de Montréal, unpublished data) demonstrate that the endogenous Na⁺-dependent α-MG transport in Caco-2 cells is expressed exclusively at the apical membrane, thus demonstrating the suitability of these cells for properly targeting SGLT1. Turner et al. (1996) concluded that mistargeting of VSVG-SGLT1 was due to the presence of the tag sequence which includes a basolateral targeting signal that would then be stronger than that of the native SGLT1 protein. However, it should be noted that there is an example of a VSVG-tagged protein (NHE3) adequately targeted to the apical membrane in MDCK and opossum kidney (OK) cells (Noël et al. 1996).

hSGLT1 and myc-tagged hSGLT1 expressed in Xenopus oocytes display identical functional characteristics, as summarized in Table 1. Steady-state (affinities for both glucose and Pz) as well as presteady-state (transient current parameters) properties demonstrate that the two proteins are functionally indistinguishable. Hence, the myc tag expressed at the N-terminus does not functionally interfere with hSGLT1, which indicates that (1) the N-terminal section of the protein can be manipulated without affecting its function and (2) this part of the protein is unlikely to comprise part of the actual active site(s) of the transporter. On the other hand, the N-terminal site should not be considered readily accessible for all modifications, as evidenced by the total (poly-His tagged) or almost total (HA tagged) lack of expression of some tagged hSGLT1 constructs. For these two constructs, we do not know whether the impaired function was due to mistargeting of the protein or to actual inactive proteins correctly targeted to the plasma membrane. Some modifications at the N-terminal region, though not generating additional amino acids in the sequence, have been shown to induce mistargeting of the protein, as demonstrated by the GGM impaired mutation D28N in which the protein is believed to be retained in intracellular compartments (Lostao et al. 1995).

Unfortunately, our addition of tags to the C-terminus of hSGLT1 has systematically failed. No activity was found in oocytes injected with cRNA encoding HA-, poly-His- or myc-tagged hSGLT1. These results indicate that the integrity of the C-terminus is probably more vital for expression of active hSGLT1 protein than that of the N-terminus. Still, it should not be concluded that such a task is impossible since, as mentioned earlier, C-terminal tagging of SGLT1 has been successfully achieved by Turner and co-workers (Turner et al. 1996) using a VSVG tag. We conclude that the chance of forming a functional tagged protein depends on both the type of epitope tag and its location, and each combination has to be tested individually for both membrane localization and function.

Functional characterization of myc-hSGLT1

In a recent paper, Vayro and co-workers (Vayro et al. 1998) showed expression of a functional myc-tagged SGLT1 cloned from rabbit in COS cells, which is in accordance with our present results on the functionality of myc-SGLT1 cloned from human and expressed in oocytes. Unfortunately, the choice of COS cells as an expression system could not give information as to the actual apical targeting of such a construct. In addition, very few functional characterizations were performed. Our results, based on transfected MDCK cells, confirm appropriate targeting of myc-hSGLT1 to the apical domain of the cell, as demonstrated by the strict apical pathway for uptake of α-MG (Fig. 10). These results are in agreement with those of Suzuki and co-workers (Suzuki et al. 1996) showing that wild-type SGLT1 transfected into MDCK cells will completely segregate to the apical domain after completion of the tight junction. Immunofluorescence studies on oocytes, as presented in Fig. 8, confirm the adequate plasma membrane targeting of the myc-tagged protein. The fluorescence was shown to be clearly restricted to the cell membrane, although faint signals could sometimes be detected in a location directly underneath the plasma membrane. The positive labelling found in Fig. 8D, where untreated and non-fixed whole oocytes were used, also identified the extracellular position for the N-terminus of the protein, supporting the conclusion of Turk et al. (1996) based on computerized analysis of the secondary structure of the protein and the results of Vayro et al. (1998) using immunogold labelling on COS-7 cells. This result also indicates that the epitope is exposed in an adequate conformation in the functional protein and that it allows good binding of the antibody under physiological conditions for the oocyte. Still, addition of the antibody to a functioning oocyte (electrophysiological assay with 5 mm glucose) did not impede its activity (data not shown), a conclusion also reached by Vayro et al. (1998). This result strengthens the hypothesis that the N-terminus is not directly involved in the protein function.

The present paper reports for the first time a study of Pz binding to oocytes. Specific binding of Pz can be readily achieved in oocytes expressing high levels of hSGLT1 as shown in Fig. 4. The maximal signal can represent up to a 10-fold increase in binding over background, suggesting that adequate specific binding might still be achieved with oocytes expressing less than half-maximal SGLT1 activity. Pz binding correlated with the level of expression of the protein as demonstrated by the good correlation between levels of activity (glucose-induced current) and binding of the inhibitor. This suggests that Pz binding could be used as a means to identify or quantify the level of expressed SGLT1 protein. This approach could be useful for studies of non-functional or mutant SGLT1 in oocytes which could otherwise be impossible to detect. Also, the amplitude of specific Pz binding depicted on oocytes was found to be sufficient to allow accurate affinity studies such as evaluation of K_d constants (data not shown).

SGLT1 will tolerate both minor and major modifications of its structure but results are often unpredictable. While major truncations can lead to expression of a partially functional protein (Panayotova-Heiermann et al. 1997), minute modifications such as point mutations will induce complete loss of activity, as for the 30 different mutations reported for GGM patients (Martin et al. 1996). In an effort to investigate structure-function relationships, our laboratory has generated chimeras between the human SGLT1 and the Na⁺-myo-inositol cotransporter SMIT. Out of these constructs, a chimera named C1 consisting of the last 13 transmembrane domains of hSGLT1 and the N-terminal first transmembrane segment of SMIT is expressed in oocytes as a protein with a close similarity in activity to SGLT1 (Jalal et al. 1996). This protein still demonstrates Na⁺ dependency, α-MG specificity, though with a higher affinity, as well as Pz sensitivity. Even though SMIT and SGLT1 are closely related, the functionality of C1 stresses the relative flexibility of the first transmembrane segment of hSGLT1. The degree to which SGLT1 can accept such modifications of its N-terminal domain without inducing changes to its basic features is not known. The possibility of tagging the protein in a way that will not interfere with its functionality is an attractive means of investigation for structure-function studies. As we demonstrate in this paper, myc-tagged hSGLT1 can adequately replace the wild-type protein without any alteration of function. Creating a protein tagged at the C-terminal position could aid in investigating the biochemistry of SGLT1.