The human tumour suppressor gene SLC5A8 expresses a Na+–monocarboxylate cotransporter

Authors

  • Michael J. Coady,

    1. Groupe d'étude des protéines membranaires, Pavillon Paul-G-Desmarais, 2960 chemin de la Tour, Université de Montréal, Montreal, QC, Canada, H3T 1J4
    Search for more papers by this author
  • Min-Hwang Chang,

    1. Physiology and Biophysics, Case Western Reserve University, Cleveland, OH 44106, USA
    Search for more papers by this author
  • Francois M. Charron,

    1. Groupe d'étude des protéines membranaires, Pavillon Paul-G-Desmarais, 2960 chemin de la Tour, Université de Montréal, Montreal, QC, Canada, H3T 1J4
    Search for more papers by this author
  • Consuelo Plata,

    1. Physiology and Biophysics, Case Western Reserve University, Cleveland, OH 44106, USA
    Search for more papers by this author
  • Bernadette Wallendorff,

    1. Groupe d'étude des protéines membranaires, Pavillon Paul-G-Desmarais, 2960 chemin de la Tour, Université de Montréal, Montreal, QC, Canada, H3T 1J4
    Search for more papers by this author
  • Jerome Frank Sah,

    1. Ireland Cancer Center and Department of Medicine, University Hospitals of Cleveland and Case Western Reserve University, Cleveland, OH 44106, USA
    Search for more papers by this author
  • Sanford D. Markowitz,

    1. Ireland Cancer Center and Department of Medicine, University Hospitals of Cleveland and Case Western Reserve University, Cleveland, OH 44106, USA
    2. Howard Hughes Medical Institute
    Search for more papers by this author
  • Michael F. Romero,

    1. Physiology and Biophysics, Case Western Reserve University, Cleveland, OH 44106, USA
    Search for more papers by this author
  • Jean-Yves Lapointe

    1. Groupe d'étude des protéines membranaires, Pavillon Paul-G-Desmarais, 2960 chemin de la Tour, Université de Montréal, Montreal, QC, Canada, H3T 1J4
    Search for more papers by this author

Corresponding author J.-Y. Lapointe: Physics Department, Université de Montréal, PO Box 6128, succ. ‘centre-ville’, Montreal, QC, Canada, H3C 3J7. Email: jean-yves.lapointe@umontreal.ca

Abstract

The orphan cotransport protein expressed by the SLC5A8 gene has been shown to play a role in controlling the growth of colon cancers, and the silencing of this gene is a common and early event in human colon neoplasia. We expressed this protein in Xenopus laevis oocytes and have found that it transports small monocarboxylic acids. The electrogenic activity of the cotransporter, which we have named SMCT (sodium monocarboxylate transporter), was dependent on external Na+ and was compatible with a 3 : 1 stoichiometry between Na+ and monocarboxylates. A portion of the SMCT-mediated current was also Cl dependent, but Cl was not cotransported. SMCT transports a variety of monocarboxylates (similar to unrelated monocarboxylate transport proteins) and most transported monocarboxylates demonstrated Km values near 100 μm, apart from acetate and d-lactate, for which the protein showed less affinity. SMCT was strongly inhibited by 1 mm probenecid or ibuprofen. In the absence of external substrate, a Na+-independent leak current was also observed to pass through SMCT. SMCT activity was strongly inhibited after prolonged exposure to high external concentrations of monocarboxylates. The transport of monocarboxylates in anionic form was confirmed by the observation of a concomitant alkalinization of the cytosol. SMCT, being expressed in colon and kidney, represents a novel means by which Na+, short-chain fatty acids and other monocarboxylates are transported in these tissues. The significance of a Na+–monocarboxylate transporter to colon cancer presumably stems from the transport of butyrate, which is well known for having anti-proliferative and apoptosis-inducing activity in colon epithelial cells.

The SLC5 gene family encompasses a variety of eukaryotic and prokaryotic cotransporter proteins which convey neutral, positively charged and negatively charged substrates across cell membranes (Wright & Turk, 2004). While most of the well-characterized SLC5 cotransporters have been shown to operate electrogenically with a 2 : 1 stoichiometry between Na+ ions and the cotransported substrate, pre-steady-state and leak currents have also been associated with the activity of these proteins; in addition, glucose-dependent currents exist in the absence of glucose transport for one member of this gene family (Diez-Sampedro et al. 2003). Although the functions of some of the members of this gene family are well understood, there remain a number of orphan cotransporters for which either no transport function is known or for which functional evidence of transport is scant (Wright & Turk, 2004).

The cDNA for one of the orphan cotransporters, from the SLC5A8 gene, was originally isolated from renal tissue and described as expressing a Na+-dependent iodide transporter, based on measurements of iodide uptake into transfected cells (Rodriguez et al. 2002). Although the protein sequence clearly belongs to the SLC5 family of Na+-coupled cotransporters and the protein was shown to be expressed at the apical membrane of thyroid cells, the increase in iodide uptake associated with expression of the transfected protein was extremely small. More recently, we have shown that the SLC5A8 gene is silenced by methylation in most human colon tumours (Li et al. 2003) and that reintroduction of this gene leads to growth suppression.

In this study, we show that the SLC5A8 gene expresses a Na+–monocarboxylate cotransporter with a number of unique features and that this cotransporter exhibits a substrate specificity similar to that of the Na+-independent monocarboxylate transporters (Halestrap & Meredith, 2004). The transport of short-chain fatty acids (SCFA), such as butyrate, from the colonic lumen may be responsible for the anti-proliferative effects of expression of this protein.

Methods

Xenopus oocyte isolation

Oocytes were removed from gravid Xenopus laevis frogs (Connecticut Valley Biological Supply Co., Southampton, MA, USA, and Xenopus Express, Plant City, FL, USA) under anaesthesia (0.13% tricaine (3-aminobenzoic acid ethyl ester) applied to water). The individually dissected oocytes were placed into a Ca2+-free buffered saline solution (200 mosmol (kg H2O)−1) and defolliculated by collagenase digestion. The oocytes were maintained at 18°C in Barth's solution (88 mm NaCl, 1 mm KCl, 0.82 mm MgSO4, 0.41 mm CaCl2, 0.33 mm Ca(NO3)2, 10 mm Hepes, pH 7.6) supplemented with 100 U ml−1 penicillin, 0.1 mg ml−1 streptomycin and 0.2 mg ml−1 kanamycin but without added sodium pyruvate, or in sterile, filtered ND96 medium (96 mm NaCl, 2 mm KCl, 1 mm MgCl2, 1.8 mm CaCl2, 5 mm Hepes-Tris, pH 7.6).

All experiments were performed in accordance with the regulations of either the Comité de déontologie de l'expérimentation sur les animaux of the Université de Montréal or the Institutional Animal Care and Use Committee of Case Western Reserve University.

SLC5A8 mRNA preparation

Human renal cortex was generously provided by Dr Alain Bonnardeaux from Maisonneuve-Rosemont Hospital in agreement with the hospital's ethics committee. mRNA was isolated using TRIzol reagent (Invitrogen, Burlington, ON, Canada) and mRNA was purified with oligo-dT cellulose chromatography. cDNA was synthesized using a Stratagene cDNA synthesis kit (Stratagene, San Diego, CA, USA). The coding region of the SLC5A8 gene was obtained by performing reverse transcription-polymerase chain reaction using Pfu polymerase (Stratagene) on the human renal cDNA using the primers GATCTATGGACACGCCACGGGGCATCG and CTAGTTCACAAACGAGTCCCATTGCTCTTGC. The 1.8 kb product was digested with Exonuclease III to generate cohesive ends (Kaluz et al. 1992) and was ligated into the vector pT7TS (kindly provided by Dr Paul Krieg, University of Texas), which had been digested with Bgl II and Spe I. The identity of the cDNA insert was confirmed by dideoxy sequencing. The recombinant vector was cleaved with Sal I and capped mRNA was transcribed in vitro using the T7 mMessage mMachine kit (Ambion, Woodward, TX, USA). Oocytes were injected with 46 nl aliquots containing the mRNA at 0.1 μg μl−1 at 1–2 days following surgical removal. Other oocytes were injected with water for use as negative controls. The oocytes were used in experiments at 3–7 days following injection.

Two-microelectrode recordings

Currents across oocyte plasma membranes were measured with a standard two-microelectrode voltage clamp technique as previously described (Coady et al. 2002) with minor modifications. In brief, a commercial amplifier and a data acquisition system were used to send voltage pulses to the oocyte and simultaneously record membrane current and voltage signals. Unless described otherwise, the oocyte was superfused (∼1.5 ml min−1) with ND96 medium. After microelectrode impalement, a membrane potential stabilization period of 1–10 min was observed. The membrane potential was then clamped at −50 mV, following which a voltage range from +50 mV to −150 mV was covered in 20 mV steps. The oocyte membrane potential was stepped to each level for 150 ms with 400 ms intervals at −50 mV before clamping to the next potential, and traces were analysed by averaging the signal in a window of 50 ms positioned after the decay of capacitive transient currents. The measurements are generally taken in the absence and in the presence of a particular substrate, and the substrate-specific current is determined by subtraction of one current from the other. When the concentration of Na+ was diminished for measurement of currents or for radiolabel uptake experiments (see below), it was isotonically replaced with choline; similarly, chloride replacement was done isotonically with cyclamate. Membrane potentials measured in the presence of diminished Cl concentrations are corrected for the change in liquid junction potential at the bath electrode, which was measured using a 1 m KCl flowing electrode. All of the experiments were performed at room temperature. The measurements used in this article employed the shortest possible exposure to substrates (20–60 s), and oocytes were discarded after exposure to high concentrations of substrate.

Uptake of radiolabelled lactate

Groups of six oocytes were placed into 0.5 ml of ND96 media (or an equivalent saline solution in which Na+ or Cl were replaced isotonically) containing various concentrations of lactate along with 130 nCi of l-[U-14C]lactic acid (Na salt) (Perkin-Elmer, Woodbridge, ON, Canada). Oocytes used for measurement of radiolabel uptake under Na+-free conditions were first rinsed twice in a Na+-free ND96 solution; oocytes used for uptake in 2 mm Cl solution were similarly rinsed with 2 mm Cl ND96 solution. Uptakes were performed for 2 min and oocytes were immediately rinsed 5 times with 4 ml of ice-cold ND96. Individual oocytes were then lysed overnight in 0.5 ml SDS prior to addition of Betablend scintillation cocktail (ICN, Costa Mesa, CA, USA) and measurement of radioactivity.

Intracellular pH measurements

Intracellular pH and membrane potential were simultaneously recorded using microelectrodes as reported previously (Romero et al. 1997). pH electrodes were calibrated using pH 6.0 and 8.0 standards (traceable National Bureau Standards, Fisher Scientific, Pittsburgh, PA, USA) followed by point calibration in ND96 (pH 7.50) as previously described (Romero et al. 1997, 1998). The pH electrodes had slopes of at least −54 mV (pH unit)−1.

Statistical methods

Different groups of data were compared using Student's paired or unpaired t tests or an analysis of variance followed by Tukey's HSD test, as appropriate.

Results

Identification of transported substrates

SLC5A8 is most closely related to the Na+–vitamin cotransporter (SMVT) (Prasad et al. 1998) and the Na+–I cotransporter (NIS) (Dai et al. 1996). As the substrates for SMVT and NIS are anionic, we reasoned that SLC5A8 probably cotransports an anionic substrate with Na+. Consequently, oocytes injected with SLC5A8 mRNA were originally superfused with a variety of different anionic species at 1 mm. Exposure of the oocytes to I or Br did not produce current, nor did superfusion of dicarboxylates or tricarboxylates (see Fig. 1A). However, large currents were observed with superfusion of most monocarboxylate anions, even at 100 μm. Only a few of the anions tested are shown in Fig. 1A; among the anions which did not produce current were sulphate, nitrate, formate, phosphate, para-aminohippuric acid, chlorate and glutamate. Superfusion with 1 mm monocarboxylates induced no measurable currents in control oocytes. It is also apparent that the propionate-induced currents were Na+ dependent. We have consequently named this protein the sodium monocarboxylate transporter (SMCT). The cDNA sequence of this clone matches that previously reported (Li et al. 2003), but for a single, silent nucleotide difference (T1617A in GenBank Accession no. AF536216).

Figure 1.

Transporter-mediated currents
A, this current tracing, from an oocyte expressing the SLC5A8 mRNA and voltage clamped at −50 mV, demonstrates inward currents associated with superfusion of 0.1 mm monocarboxylates but not with superfusion of iodide, bromide, dicarboxylates or tricarboxylates (all at 1 mm). It is also apparent that removal of external Na+ eliminates the inward current. The brief, sharp spikes seen are artefacts caused by solution changes. B, oocytes were superfused with 100 μm of each substrate immediately after superfusion with 100 μm propionate. Currents are expressed as a proportion of the current observed with 100 μm propionate. Currents associated with superfusion by formate, succinate and citrate were not significantly different from zero. Values represent mean ±s.d., n= 6–12, from 3 donor animals.

SMCT-expressing oocytes were superfused with 100 μm of different carboxylates immediately following superfusion with 100 μm propionate, and substrate-induced current was expressed as a percentage of the preceding propionate-induced current in order to circumvent changes in transporter activity (see Fig. 1B). Most monocarboxylates evinced currents similar in magnitude to propionate, including butyrate. However, no current was observed with the 1-carbon SCFA (formic acid), and the 2-carbon SCFA (acetic acid) induced markedly smaller currents than did longer SCFAs. Substituted monocarboxylates (lactate and pyruvate, thiol derivatives of propionic acid) also induced large currents through SMCT. The stereospecificity of substrate binding is evident in that d-lactate induced significantly less current than did its l-isomer (P < 0.001, ANOVA and Tukey's HSD). Nicotinic acid also induced a large current, demonstrating that SMCT is capable of accepting a wide variety of monocarboxylates.

Current–voltage relationship and kinetic characterization

The nature of the currents induced by 1 mm propionate were examined by voltage-clamping the plasma membrane at different membrane potentials (see Fig. 2A). Large, transient, pre-steady-state currents were observed which were markedly diminished in the presence of substrate, as has been reported for related transporters (Parent et al. 1992; Coady et al. 2002). Following the decay of these transient currents, propionate-dependent steady-state currents were measurable at different membrane potentials. The current–voltage relationship of these steady-state currents displayed a sensitivity to potential similar to that seen with other members of the SLC5 family (see Fig. 2B) (Coady et al. 2002). The monocarboxylate-induced current was voltage dependent throughout the voltage range studied. Current–voltage relationships for the propionate-induced currents were measured at a variety of Na+ and propionate concentrations (see Fig. 3). Michaelis-Menten parameters were accurately fitted to the data obtained across most of the voltage range. As seen in Fig. 3A, the Km for propionate was fairly constant at negative membrane potentials and was 128 ± 4 μm at −50 mV (mean ±s.e.m., n= 9), but increased dramatically at positive membrane potentials. Butyrate and lactate produced similar results, with Km values of 72 ± 8 μm(n= 8) and 159 ± 24 μm(n= 8), respectively, at −50 mV. The SMCT Imax was −474 ± 66 nA (n= 8) at −50 mV and it increased as the membrane potential became more negative (see Fig. 3B). The Km for Na+ was measured in the presence of 0.5 mm propionate and was 44 ± 7 mm at −50 mV (n= 8). As shown in Fig. 3C, it rose progressively as the membrane potential became less negative. Fitting the Hill equation to these currents produced an n value of 1.7 ± 0.3 at −50 mV, suggesting significant cooperativity and the involvement of more than one Na+ ion per cotransport turnover. As shown in the inset to Fig. 3C, the Hill number did not vary considerably as the membrane potential was changed.

Figure 2.

Current–voltage analysis of propionate-induced currents
A, typical currents measured across the plasma membrane of an oocyte expressing SMCT during current–voltage analysis. Currents on the left are in the ND96 saline solution lacking propionate; currents on the right are in the added presence of 1 mm propionate. B, current–voltage representation of the currents measured in the presence and absence of 1 mm propionate as well as of the difference between these two sets of measurements, representing propionate-sensitive current.

Figure 3.

Steady-state kinetics for propionate and Na+
A, the currents induced by addition of various concentrations of propionate were measured in the presence of 96 mm Na+ while the cell was clamped at a series of different membrane potentials. The Michaelis-Menten equation was fitted to the current measurements, from which the derived Km values for propionate are shown. Values shown represent mean ±s.e.m.; n= 9 oocytes, from 3 donor animals. B, the Imax values derived by Michaelis-Menten analysis of the currents from A are shown at different membrane potentials. The data has been standardized to the current observed at −50 mV, to control for variation in SMCT expression between oocytes. C, the external Na+ concentration was varied in the presence of 500 μm propionate and cells were clamped to different membrane potentials; the Michaelis-Menten equation was fitted to the substrate-specific currents and the derived Km values for Na+ are shown. Values shown represent mean ±s.e.m.; n= 8 oocytes, from 3 donor animals. The inset shows the fitting of the Hill equation to the data at different membrane potentials. D, the Imax values derived by Michaelis-Menten analysis of the currents from C are shown at different membrane potentials. The data have been standardized to the current observed at −50 mV, to control for variation in SMCT expression between oocytes.

Inhibitors

We next examined a number of compounds to determine whether any could inhibit SMCT. Compounds that block other SLC5 transporters (e.g. phlorizin, phloretin) or anion transport systems (e.g. probenecid, SITS), as well as substrate analogues (e.g. phenyllactate, ibuprofen), were tested. As Fig. 4 illustrates, ibuprofen efficiently inhibited SMCT currents as did probenecid and valproate, although less potently. Other possible inhibitors showed little or no effect on propionate-induced current, including phloretin and α-cyano-hydroxy-cinnamate, which are potent inhibitors of other monocarboxylate transporters (Poole & Halestrap, 1993). None of the inhibitors tested were able to induce significant currents through SMCT except for 1 mm phenyllactate, which induced 20 nA. We further examined inhibition by both ibuprofen (a propionate derivative) and probenecid (an organic anion exchange inhibitor) by measuring currents induced by various concentrations of propionate in the presence of either 100 μm ibuprofen or 500 μm probenecid. Both molecules appeared to inhibit propionate competitively, ibuprofen with a Ki of 20 ± 3 μm and probenecid with a Ki of 195 ± 33 μm (mean ±s.e.m., n= 7 in both cases, data not shown).

Figure 4.

Inhibitors of SMCT current
The current induced by 100 μm propionate superfusion was measured in the presence and absence of possible inhibitors. All inhibitors were superfused at 1 mm with the exceptions of phloretin and phlorizin (0.1 mm each) and SITS (0.5 mm). 4-CHC is cyano-4-hydroxy-cinnamic acid; SITS is 4-acetamido-4′-isothiocyanostilbene-2,2′-disulphonic acid. Results are expressed as a proportion of the current observed in the presence of 100 μm propionate immediately prior to the addition of inhibitor. No significant inhibition was observed with phlorizin or SITS. Values represent mean ±s.d., n= 5–6, from 3 donor animals.

Leak current

Ibuprofen was used to determine whether a leak current passes through SMCT, as is seen for related cotransporters (Umbach et al. 1990; Eskandari et al. 1997; Coady et al. 2002). The ibuprofen-sensitive current observed in the absence of external monocarboxylates was much smaller than that associated with cotransport of external monocarboxylates into the cell, generally yielding currents which were lower than 50 nA (and often zero) at a holding potential of −50 mV, but the magnitude of the current increased markedly at positive membrane potentials (see Fig. 5A). This leak current was not seen in water-injected oocytes. Surprisingly, the current does not appear to be mediated, even partially, by Na+ itself (see Fig. 5B) as the reversal potential for the leak currents did not vary with external [Na+]. Although the currents observed at negative transmembrane potentials were small and relatively constant, the magnitude of the ibuprofen-sensitive leak currents at positive transmembrane potentials did appear to increase with the level of external [Na+]. Examination of the pre-steady-state currents associated with different external [Na+] (in the presence and absence of ibuprofen) showed that the apparent Na+ effect was due to a diminished inhibitory capacity of ibuprofen as external [Na+] was decreased. Sodium therefore does not seem to be required at all for the SMCT leak current. When external [Cl] was altered (see Fig. 5C), the ibuprofen-sensitive leak displayed very little change in inward currents but had a significantly Cl-dependent outward current, without changes in the reversal potential. However, ibuprofen inhibition of the pre-steady-state currents was not affected by external [Cl], indicating that Cl stimulates the leak current at positive transmembrane potentials without itself being transported. Consequently, the ionic species responsible for the leak current remains unidentified. The leak current was larger (at all membrane potentials, data not shown) after the oocyte was superfused with propionate or other SMCT substrates, suggesting that the intracellular SMCT substrates were involved in the leak current.

Figure 5.

Leak current analysis
Leak currents were measured in the absence of propionate. Under these conditions, the leak current is represented by the current inhibited by 1 mm ibuprofen. A, comparison of the currents obtained in the absence of ibuprofen or propionate (ND96), in the presence of 1 mm ibuprofen and in the presence of 100 μm propionate. A typical experiment is shown. B, current–voltage relationships for leak currents (ibuprofen-sensitive in the absence of substrate) from the experiment shown in A, measured in the presence of varying levels of external Na+. Currents were measured with concentrations of ibuprofen in addition to those shown in A. C, current–voltage relationships for leak currents from a typical experiment performed in the presence of varying levels of external Cl.

SMCT stoichiometry

Estimates of cotransporter stoichiometry can be obtained by measuring changes in reversal potential as the external concentration of a substrate is varied, provided that bidirectional transport occurs and that an inhibitor is available (Chen et al. 1995). Oocytes were exposed to various external concentrations of propionate and Na+ while cotransporter currents were measured at different membrane potentials, followed by measurements under the same conditions but in the presence of 1 mm ibuprofen. Ibuprofen-sensitive currents were then used to interpolate the reversal potentials (where the ibuprofen-sensitive current = 0). Decreasing the external [Na+] caused a steep shift in the reversal potential towards more negative values (Fig. 6A). As shown in Fig. 6B, this decrease was less pronounced at lower [Na+] levels, as was seen for the Na+–glucose cotransporter (Chen et al. 1995), presumably due to an increased proportion of leak current in the ibuprofen-sensitive current at low external [Na+]. The rate of decrease in reversal potential at high [Na+] averaged 81.2 ± 14 mV decade−1 (mean ±s.d., n= 8), corresponding to 1.4 ± 0.4 Na+ ions transported per net charge transferred across the membrane. As Fig. 6C shows, increasing external [propionate] increased the magnitude of the ibuprofen-sensitive currents and a reversal potential was observed which increased by 32.8 ± 4.6 mV decade−1 (mean ±s.d., n= 6) as the external [propionate] was elevated (see Fig. 6D), suggesting that 0.56 ± 0.08 propionate molecules crossed the membrane per net unit charge transferred (Chen et al. 1995).

Figure 6.

Effects of varying external Na+ and propionate on current magnitude and reversal potential
A, a typical experiment where external [Na+] was varied between 0 and 96 mm; the tracings shown represent the currents inhibited by 1 mm ibuprofen in the presence of 0.5 mm propionate. B, reversal potentials observed for different external Na+ concentrations (n= 8, data represent mean ±s.e.m.). C, a typical experiment where external propionate was varied between 20 μm and 1 mm. The tracings represent the currents inhibited by 1 mm ibuprofen in the presence of 80 mm Na+. D, reversal potentials observed for different external propionate concentrations (n= 6, data represent mean ±s.e.m.).

Altering external [Cl] caused unexpected changes in the current–voltage relationship of SMCT (see Fig. 7). In the presence of 2 mm external [Cl], there was a substantial current inhibited by ibuprofen, composed of the Na+–propionate cotransport and the (much smaller) leak current. Addition of Cl to the bathing solution produced an additional, superimposed current whose magnitude varied in a roughly linear manner with external [Cl]. Influx of external Cl at negative potentials should diminish the inward current and so the increased current observed could not be due simply to Cl influx. Also, currents seen at different [Cl] had the same reversal potential, confirming that no Cl influx is associated with SMCT transport activity. As the large, Cl-dependent current was not seen in the absence of propionate, both Cl and propionate (and presumably Na+) must be present for this additional current to pass through SMCT. This suggests that Cl has a direct, stimulatory effect on Na+–propionate cotransport.

Figure 7.

Effects of varying external Cl on SMCT current magnitude and reversal potential
A, external [Cl] was varied between 2 and 104 mm; the currents shown represent currents inhibited by 1 mm ibuprofen in the presence of 0.5 mm propionate. B, reversal potentials observed for different external Cl concentrations (n= 8, data represent mean ±s.e.m.).

Lactate transport

To confirm that radiolabelled monocarboxylates actually pass through SMCT, we measured the uptake of radiolabelled lactate. At both 1 μm and 5 mm external [lactate], uptake rates were constant for 5 min, after which they diminished. All uptake experiments were therefore performed for 2 min. As Fig. 8 demonstrates, there was a large transport of radiolabelled lactate into SMCT oocytes that was both Na+ dependent and absent from control oocytes. In addition, the transport of 100 μm lactate was inhibited 85 ± 5% by the presence of 1 mm ibuprofen (data not shown). The apparent Km for the uptake of lactate was 200 ± 50 μm (n= 3 experiments with 6 oocytes per datum point), similar to the value measured for the lactate-induced current through SMCT, and the uptake levels were commensurate with the currents described above (100 pmol min−1 is equivalent to 320 nA, given two net charges moving across the membrane per lactate). A significant Na+-independent transport of lactate through SMCT was also observed (see Fig. 8); the Km appears to be close to 1 mm. Lactate uptake was also shown to be Cl dependent; when external Cl was removed, 1 μm lactate transport diminished from 411 ± 45 to 206 ± 22 fmol min−1 (mean ±s.e.m., n= 3 experiments, with 6 oocytes per measurement, P < 0.005). This indicates that the increased current seen through SMCT in the presence of chloride is due to increased Na+–monocarboxylate cotransport. We also measured the uptake of radiolabelled propionate (not shown) but encountered difficulties due to the large, Na+-independent uptake of propionate (passive diffusion and/or partition into the membrane) into oocytes.

Figure 8.

Uptake of radiolabelled lactate
The uptake of radiolabelled lactate was measured at various external concentrations of lactate. A typical experiment is shown. The data were analysed using the Michaelis-Menten equation and produced a Km value of 252 ± 54 μm and a maximal rate of transport of 116 ± 9 pmol oocyte−1 min−1. A small, Na+-independent uptake of lactate can also be observed in SMCT-expressing oocytes. The two sets of data shown from control oocytes overlap considerably. Each point represents mean ±s.e.m. for 6 oocytes; data are representative of three experiments.

Intracellular pH and SMCT

As the transport of SCFAs is known to stimulate acidification of the interior of cells (Chu & Montrose, 1997), we investigated the effect of butyrate transport through SMCT on cellular pH. Addition of butyrate has been used previously to lower intracellular pH (pHi) (Romero et al. 1998; Stewart et al. 2001; Flagg et al. 2002). This intracellular acidification is concentration and pHo dependent (Romero et al. 1998; Flagg et al. 2002). Addition of 1 mm butyrate to a water-injected control (Fig. 9A) demonstrated the passive diffusion of non-ionized butyric acid into the oocyte inducing slow, intracellular acidification (Fig. 9A). This butyrate addition did not alter membrane potential (Vm) in control oocytes. When butyrate was transported through SMCT, however, there was a rapid pHi increase (Fig. 9B). This alkalinization was due to the SMCT-mediated transport of ionized butyrate which then partially reassociated with H+ inside the cell. Since multiple Na+ ions enter with each butyrate, the SMCT oocyte also depolarized (by 55 mV). It can also be seen that the SMCT-mediated alkalinization could be reversed by removing external Na+. The average intracellular acidification of control oocytes due to passive diffusion of 1 mm butyrate occurred at a rate of −19 × 10−5 pH units s−1 while the transport of 1 mm butyrate through SMCT-expressing oocytes caused alkalinization by +52 × 10−5 pH units s−1.

Figure 9.

Intracellular pH measurements in Xenopus oocytes
A typical experiment is shown involving simultaneous measurements of intracellular pH (pHi) and membrane potential (Vm) for a water-injected control oocyte (A) and SMCT-injected oocyte (B). Oocytes were continually superfused as indicated in Methods. Butyrate (1 mm) was added as indicated, followed by Na+ removal (0 Na+, choline replacement) in the continued presence of butyrate.

Temporal changes in SMCT activity

One considerable difficulty in performing the experiments described in this paper resulted from changes in SMCT activity during the course of the experiment. Initial exposure to low concentrations of substrates caused a stable current which displayed the characteristics of a Michaelis-Menten response to substrate (see Fig. 10). With prolonged exposure to higher concentrations (1 mm or more), the currents exhibited a peak in electrogenic activity, followed by a prolonged diminution of current. At times, the initial peak could present as a sharp spike. We are not yet certain of the reason for this change in activity. A change in the baseline current (in saline) after exposure to high concentrations of monocarboxylates was also often encountered, although the change could be in the form of increased or decreased conductance for a given substrate. Throughout this study, care was taken to avoid substrate concentrations larger than 1 mm and to reduce the length of time oocytes were exposed to 0.5 and 1 mm concentrations to an absolute minimum.

Figure 10.

Temporal modification of SMCT currents
A typical experiment is shown where initial exposure to low concentrations of propionate (or to other substrates with high affinity) induced currents of magnitudes consistent with Michaelis-Menten activation. Exposure to large concentrations of substrate caused a rapid, persistent decrease in current.

Discussion

We have established the function of the protein encoded by the SLC5A8 gene as a Na+–monocarboxylate cotransporter. Many biological monocarboxylates are hydrophobic, which is why they were long considered to enter cells solely through passive diffusion (Al-Awqati, 1999). However, the SLC16 family of proteins (MCT1–4) has been shown to transport monocarboxylates (Halestrap & Meredith, 2004). SMCT represents a unique, Na+-coupled mechanism by which monocarboxylates can enter cells. There is evidence of Na+–monocarboxylate cotransport in various tissues and cells (Nord et al. 1983; Nakhoul & Boron, 1988; Poustis-Delpont et al. 1988; Siebens & Boron, 1989; Mengual et al. 1990; Kenyon et al. 1994; Bachowska-Mac et al. 1997), most particularly in the apical membranes of renal proximal tubules (Barac-Nieto et al. 1982; Ullrich et al. 1982a,b; Wright, 1985).

The precise means by which SCFAs are absorbed from the colonic lumen remains controversial. SCFA absorption is generally considered to be electroneutral and is known to interact with Na+, Cl and bicarbonate absorption (Sellin, 1999). A widely accepted model of SCFA absorption in the colon posits electroneutral exchange of intracellular bicarbonate ions for SCFAs via MCT1, in parallel with electroneutral Na+–H+ and Cl–SCFA exchange (Kunzelmann & Mall, 2002). There is evidence both to support and refute passive diffusion of SCFAs, the electroneutral exchange model and absorption through SMCT.

Despite the lack of sequence identity between SMCT (SLC5A8) and the SLC16 transporter family, there is remarkable similarity in the specificity of monocarboxylates transported by SMCT and MCT1 (a ubiquitously expressed member of the MCT family, known to be expressed in colon apical membranes) (Ritzhaupt et al. 1998). In both cases, formate is a very poor substrate whereas acetate and longer, unbranched aliphatic monocarboxylates are transported, along with several common, branched monocarboxylates such as lactate and pyruvate (Halestrap & Meredith, 2004). SMCT also transports thiol-substituted monocarboxylates and, to a lesser extent, d-lactate. These results are similar to what has been seen for colon apical membrane vesicles, albeit in the absence of Na+ (Schroder et al. 2000). Future experiments comparing SMCT and MCT1 will be required to establish specific substrates for each of the transporters.

Discrimination between the activities of SMCT and MCT1 may already be possible by the use of inhibitors. MCT1 has been shown to be potently inhibited by both phloretin and α-cyano-hydroxy-cinnamate (4-CHC) (Broer et al. 1997, 1999). As can be seen in Fig. 4, neither 4-CHC nor phloretin had a large effect on transport through SMCT; we do not yet know whether ibuprofen, the most potent inhibitor of SMCT, is able to inhibit MCT1. It is likely that a judicious use of specific inhibitors will permit experimental verification of the roles of SMCT and MCT1 in the apical transport of SCFAs in the colon.

It should be noted that one of the two sequence motifs conserved in the SLC16 family may also occur in SMCT. This motif (D/E)-G-(G/S)-(W/F)-(G/A)-W, occurs at the beginning of the first transmembrane region of the SLC16 proteins (Halestrap & Meredith, 2004). The sequence found at the beginning of the first transmembrane region of SMCT (human/mouse) is G-(T/S)-F-(T/V)-V-W, which bears an intriguing similarity to the MCT sequence motif. It is tempting to speculate that this motif plays a role in the binding or transport of the anion.

The stoichiometry of Na+–monocarboxylate cotransport can be inferred from the ratios of 0.56 ± 0.08 propionate per unit charge and 1.4 ± 0.4 Na+ ions per unit charge. These results are in accordance with a stoichiometry of 3 Na+ per monocarboxylate, producing a net transport of 2 positive charges per turnover. This is in contrast to the other members of the SLC5 family, for which a 2 : 1 stoichiometry has been established for most of members for which stoichiometry has been thoroughly investigated, apart from the anomalous SGLT3 protein (Diez-Sampedro et al. 2003). It is possible that the basic mechanism of transport for the SLC5 family is conserved in SMCT by coordinated apposition of the monocarboxylate with one of the Na+ ions, so that the protein regards it as essentially a neutral substrate being cotransported along with two other Na+ ions.

Several aspects of SMCT activity represent anomalies within the SLC5 family. One such intriguing property is the Cl sensitivity that was observed for SMCT-mediated transport of radiolabelled lactate, the cotransport current and the leak current observed at positive membrane potential. SMCT is only the second member of the SLC5 protein family to demonstrate both Na+- and Cl-dependence (Okuda et al. 2000) and the relevance of the Cl dependency to the physiological activity of SMCT is not clear. SMCT is expressed in the colon and kidney, presumably at the apical membranes, based on immunohistochemical data (Rodriguez et al. 2002) and measurement of transport activities (Storelli et al. 1980; Ullrich et al. 1982a; Poustis-Delpont et al. 1988). Consequently, SMCT will be exposed to radically different ionic environments between the colonic lumen and the lumen of renal tubules (Holtug et al. 1996). It is conceivable that Cl (and/or another anion) plays a regulatory role in the activity of this protein or in the efficient coupling between the anionic exchanger and SMCT. The dependence of SMCT current upon external [Cl] is reminiscent of that seen for the Na+/Cl/GABA transporter GAT1 (Loo et al. 2000).

Another peculiarity of SMCT is that Na+ does not appear to be responsible for the leak current, as we had previously speculated (Li et al. 2003). The 3- to 4-fold increase in intracellular [Na+] which we had previously observed probably resulted from the 3–4 day incubation in media containing 2 mm pyruvate (Na+–pyruvate cotransport) rather than from a Na+ leak current as displayed by other SLC5 members. Although we have no data identifying the ionic species comprising the leak current through SMCT, the augmentation of leak current by pre-loading the cell with monocarboxylates suggests that the leak current is actually composed of monocarboxylates. This current may represent both coupled efflux of Na+ and monocarboxylates or an uncoupled flow of monocarboxylates, which SMCT is capable of performing (see Fig. 8). It is not certain whether external Na+ is required for the leak current to occur; external Na+ was required for our experiments in order that ibuprofen might bind and inhibit the leak current.

It is perplexing that the apparent Km values for Na+–monocarboxylate cotransport by SMCT are in the 50–100 μm range. For the colon, total monocarboxylates (acetate, propionate and butyrate) account for ∼60–100 mm of the colonic content. That is, the Km values for these SCFAs is at least 100-fold less than the physiological concentration. We do not yet know how these substances compete for cotransport by SMCT. Does this discrepancy ensure that these SCFAs are absorbed by the colon and/or kidney? Or do the discrepancies serve a protective function in limiting the uptake of these fuel sources which also have some cellular transformation properties. Additional experiments should reveal much about colon and renal physiology. The role of SMCT in the colonic absorption of Na+ will also bear examination, since Na+ absorption is decreased by about 65% when SCFAs are absent (Zaharia et al. 2001).

The temporal inactivation of SMCT is a poorly understood phenomenon. It depends on the level of external substrate and may be affected by the level of SMCT expression (B. Wallendorff, unpublished observation). The simplest explanation is that high internal concentrations of monocarboxylates and/or Na+ cause some form of inhibition of SMCT. Given the very high concentrations of SCFAs in the colonic lumen, presumably accompanied by a high intracellular concentration within colonocytes, there must be some caution in predicting the precise functional activity of colonic SMCT in vivo.

Nonetheless, the biochemical finding that SMCT can efficiently mediate the uptake of butyrate suggests the highly attractive hypothesis that transcriptional silencing of SMCT in the colon probably confers a proliferative advantage to colonic epithelial cells by virtue of increasing their resistance to the differentiating effects of luminal butyrate. Butyrate has a well-characterized ability to induce differentiation and apoptosis of colon epithelial cells, including some colon cancer cell lines (Heerdt et al. 1994). The silencing of SMCT is an early event in colon neoplasia, being demonstrable in aberrant crypt foci that are the earliest, microscopic neoplastic lesion of the colonic epithelium (Li et al. 2003). We hypothesize that such microscopic proliferations could be the functional result of the loss of butyrate uptake via the SMCT pathway. Alternatively, activation of SMCT rather than inactivation, as in colon cancer/neoplasia, could be a cellular transport signal for the differentiation of epithelial tissues. In vivo models, such as knockout mice, will ultimately help determine the relative contribution of SMCT to colonocyte uptake of butyrate and whether butyrate, or some other monocarboxylate substrate, provides the connection between SMCT silencing and colonic neoplasia versus SMCT activation and epithelial differentiation.

Note added in proof

Since this work was submitted for publication, S. Miyauchi, E. Gopal, Y. J. Fei & V. Ganapathy have published an article describing transport by the SMCT protein (J Biol Chem279, 13293–13296 (2004)). Although their findings agree with the monocarboxylate specificity of SMCT reported here, significant differences exist between the two studies regarding Cl sensitivity and proposed transport stoichiometry.

Appendix

Acknowledgements

We wish to acknowledge the financial support of the Canadian Institutes for Health Research (grant no. MOP-10580) and the National Institutes of Health for awards CA 67409 (S.D.M.) and DK-56218 (M.F.R.). S. D. M. is an investigator of the Howard Hughes Medical Institute.

Ancillary