Interaction of α-Lipoic Acid with the Human Na+/Multivitamin Transporter (hSMVT)*

Background: Transport of α-lipoate by hSMVT and its stereospecificity have been elusive. Results: Using hSMVT expressed in oocytes and in Pichia pastoris yielded detailed information about the stereospecificity of hSMVT-mediated lipoate transport and binding. Conclusion: hSMVT can bind two molecules of R-(+)-α-lipoate, its physiological substrate. Significance: In addition to biotin, pantothenate, and iodide, hSMVT mediates the transport of lipoate. The human Na+/multivitamin transporter (hSMVT) has been suggested to transport α-lipoic acid (LA), a potent antioxidant and anti-inflammatory agent used in therapeutic applications, e.g. in the treatment of diabetic neuropathy and Alzheimer disease. However, the molecular basis of the cellular delivery of LA and in particular the stereospecificity of the transport process are not well understood. Here, we expressed recombinant hSMVT in Pichia pastoris and used affinity chromatography to purify the detergent-solubilized protein followed by reconstitution of hSMVT in lipid bilayers. Using a combined approach encompassing radiolabeled LA transport and equilibrium binding studies in conjunction with the stabilized R-(+)- and S-(−)-enantiomers and the R,S-(+/−) racemic mixture of LA or lipoamide, we identified the biologically active form of LA, R-LA, to be the physiological substrate of hSMVT. Interaction of R-LA with hSMVT is strictly dependent on Na+. Under equilibrium conditions, hSMVT can simultaneously bind ∼2 molecules of R-LA in a biphasic binding isotherm with dissociation constants (Kd) of 0.9 and 7.4 μm. Transport of R-LA in the oocyte and reconstituted system is exclusively dependent on Na+ and exhibits an affinity of ∼3 μm. Measuring transport with known amounts of protein in proteoliposomes containing hSMVT in outside-out orientation yielded a catalytic turnover number (kcat) of about 1 s−1, a value that is well in agreement with other Na+-coupled transporters. Our data suggest that hSMVT-mediated transport is highly specific for R-LA at our tested concentration range, a finding with wide ramifications for the use of LA in therapeutic applications.

␣-Lipoic acid (thioctic acid; 5-(1,2-dithiolan-3-yl)pentanoic acid; LA 4 ) is a naturally occurring substance that is involved in the aerobic metabolism of all cells. LA is synthesized in small amounts in mitochondria of mammals and plants as well as in bacteria (1). The biologically active form of LA, the R-enantiomer, is endogenously synthesized de novo from octanoic acid (2), and is covalently bound to specific LA-dependent enzyme complexes (3). Most notably, R-LA serves as a cofactor in the pyruvate dehydrogenase and ␣-ketoglutarate dehydrogenase complexes (4), enzymes that catalyze central redox reactions related to energy production in the cell as well as in oxidative decarboxylations of ␣-keto acids and amino acids (5).
Because of the presence of two juxtaposed thiol groups, free LA (in its oxidized as well as in its reduced form as dihydrolipoic acid) possesses anti-inflammatory and antioxidant properties due to its ability to scavenge reactive oxygen species and reactive nitrogen species, regenerate endogenous antioxidants, and chelate toxic metals (6,7). Synthetic LA, which predominantly exists as a racemic mixture of LA (R,S-LA), has been successfully used as a therapeutic agent for about 60 years in the treatment of e.g. diabetic neuropathy, peripheral artery disease, and various skin and liver diseases (8 -10). Recent studies indicate that LA may interfere with the pathogenesis and/or progression of Alzheimer disease (11), Parkinson disease (12), and multiple sclerosis (13). Both forms of LA, R-LA and R,S-LA, are widely available as dietary antioxidants (14,15), and it is suggested that LA (in free form or as lipoyl protein) from foodstuff is absorbed in the intestine by the Na ϩ /multivitamin transporter (SMVT) (16,17) due to the fact that LA inhibited the uptake of the other two SMVT substrates, biotin and pantothenic acid, in concentration-dependent manner. However, evidence for the transport of lipoic acid or the (stereo)specificity of the interaction of lipoic acid with SMVT is rather limited (18) mostly due to the lack of suitable study systems that allow the detailed kinetic characterization of the target protein without interfering proteins found in cell or membrane preparations.
To shed light on the specifics of LA transport by the human SMVT (hSMVT), we have measured radiolabeled LA uptake in the established Xenopus oocytes system expressing hSMVT and tested the effect of the racemic or enantiomeric forms of LA or lipoamide (5-(1,2-dithiolan-3-yl)pentanamide) on the transport reaction. Furthermore, we have expressed hSMVT in Pichia pastoris cells and purified the recombinant protein from the membrane of the yeast in active form, which allows for the functional characterization of hSMVT in detergent-solubilized form and reconstituted in nanodiscs and proteoliposomes.

Experimental Procedures
Instrumentation for Chemical Syntheses-A Jasco Analytical Instruments Model P2000 digital polarimeter was used. For normal phase HPLC, an Agilent 1100 HPLC system with diode array detection was used for chiral chromatography in combination with a Daicel Chiralpak AD-H column (Table 1) and validated reference standards by a method developed at GeroNova Research, Inc. Melting points were performed on a Melt-Temp melting point apparatus and are uncorrected.

Syntheses of R-(ϩ)-␣-Lipoic Acid; S-(Ϫ)-␣-Lipoic Acid; and R-(ϩ)-, S-(Ϫ)-, R,S-(ϩ/Ϫ)-Lipoamide-R,S-(ϩ/Ϫ)-␣-Lipoic
acid (R,S-LA) (Tohope Pharmaceuticals, Jiangsu, China) was recrystallized twice from cyclohexane/ethyl acetate. 206.33 g (1 mol) of R,S-LA was dissolved at 30 -35°C in 2 liters of toluene. 0.54 mol of R-(ϩ)-␣-methylbenzylamine (RAMBA; obtained from BASF Corp. and Ͼ99% enantiomerically pure) in 0.1 liter of toluene was added dropwise from a dropping funnel in a steady stream over the course of 5 min with vigorous overhead stirring, causing a slight exothermic reaction (T ϭ 40 -45°C) without exceeding 45°C. The reaction mixture was stirred for 30 min at 40 -45°C and gradiently cooled to room temperature over the course of 2 h (cooling rate of ϳ0.17°C/min). The precipitate was filtered and washed twice with 35 ml of previously chilled toluene. The moist salt pair was recrystallized three to five times from 400 ml of ethyl acetate and then dried in a vacuum at Ͻ45°C, yielding 95 g of the R-LA-RAMBA diastereomeric salt pair (R-LA-RAMBA ϭ 95 g, melting point ϭ 110 -115°C, [␣] 20 D ϭ ϩ74.8°(c ϭ 1; ethanol)) ( Fig. 1A). The above process was repeated with fresh R,S-LA solutions and S-(Ϫ)-␣-methylbenzylamine (SAMBA), yielding 92.5 g of the S-LA-SAMBA salt pair (S-LA-SAMBA ϭ 92.5 g, melting point ϭ 108 -114°C, [␣] 20 D ϭ Ϫ73.4°(c ϭ 1; ethanol)). The salt pair was suspended in 3 liters of water at 25°C and 1 liter of cyclohexane with vigorous overhead stirring. The mixture was slowly adjusted to pH 3 by 1 N citric acid and to pH 1 with 5% sulfuric acid with ice cooling. The mixture was then allowed to warm to room temperature. The phases were separated, and the aqueous phase was re-extracted once with 300 ml of cyclohexane. The combined cyclohexane extracts were dried, filtered, and cooled to 5-10°C. The cyclohexane was seeded with R-LA or S-LA crystals post-stirred at this temperature for 5 h for crystallization. The precipitate was filtered off, rewashed with 30 ml of cyclohexane, and dried at 25°C in a vacuum. The filtered solutions were combined and concentrated with rotary evaporation for further processing. The toluene mother liquor was washed with dilute HCl and H 2 O and analyzed for the enantiomer composition, and appropriately adjusted amounts of RAMBA or SAMBA were added to isolate the crude and opposite enantiomer as in the initial resolution step. In each case, for the test compounds utilized in this study, only the initially isolated salt pairs were used after three to five recrystallizations. One mole of R,S-LA yielded 48.0 g of R-LA (23.6% based on R,S-LA, melting point ϭ 44 -46°C (sinter), [␣] 20 D ϭ ϩ121°(c ϭ 1; ethanol)) and 45.4 g of S-LA (22% based on R,S-LA, melting point ϭ 43-45°C (sinter), [␣] 20 D ϭ Ϫ118°(c ϭ 1; ethanol)) ( Fig. 1A).
The three amides (R,S-(ϩ/Ϫ)-lipoamide (R,S-LAM), R-(ϩ)lipoamide (R-LAM), and S-(Ϫ)-lipoamide (S-LAM)) were made by a modification of the procedure originally patented by E. Merck, Darmstadt, Germany for preparation of R,S-lipoylpyridoxamide (19) (Fig. 1B)  pressure with the bath temperature not exceeding 45°C. The evaporation flask was cooled, and the crude LAMs were redissolved in ethyl acetate. The ethyl acetate solutions were washed once with H 2 O, extracted twice with dilute (5%) NaHCO 3 solution to remove any unreacted LA, dried with Na 2 SO 4 , and filtered into a dry evaporating flask. The ethyl acetate solution was reduced to ϳ25% of its initial volume and allowed to cool to room temperature. Crude LAM was filtered and rinsed with chilled, fresh ethyl acetate with the solvent collected into a dry evaporator flask. The process was repeated to reduce the volume by ϳ25% and cooled. The crude LAMs from the two filtrations were combined and recrystallized once from ethanol and once from acetone, yielding sparkling yellow flakes. The crystals were filtered and dried under vacuum (Fig. 1B).

Expression of hSMVT in Xenopus laevis
Oocytes and Functional Studies-The mRNA transcript for hSMVT was produced in vitro with the Ambion MegaScript kit using the T3 RNA polymerase and an RNA cap analog following NotI linearization of pMV1 (18). X. laevis oocytes were injected with 50 ng of mRNA transcript for hSMVT and subsequently incubated in Barth's medium containing gentamicin (5 mg/ml) at 18°C for 3-6 days.
Oocytes were removed from the assay buffer after the indicated period of time and subsequently washed with ice-cold assay buffer without radiotracer followed by scintillation counting after solubilization of individual oocytes with 100 l of 5% SDS. All uptake experiments were performed at least in duplicate with oocytes from different donor frogs. Data are expressed as mean Ϯ S.E. of Ն7 oocytes, kinetic constants were obtained by non-linear regression fitting of the data with Prism 5 (GraphPad), and errors indicate the S.E. of the fit.
Production of hSMVT in P. pastoris-For immunological detection and affinity chromatography, we engineered a recombinant hSMVT construct that carries a C-terminal FLAG epitope and a hexahistidine (His) tag by using standard PCR protocols with synthetic mutagenic oligonucleotides using pMV1 as a template. All modifications in the resulting plasmid pMV1-FH were verified by DNA sequencing, and activity measurements in oocytes revealed virtually indistinguishable activity between native and recombinant hSMVT. The cDNA of recombinant hSMVT was introduced into methanol-inducible pPICHOLI vector (MoBiTec, Inc.) and introduced into P. pastoris KM71H cells. Transformed cells were cultured by consecutive growth at 29°C, and the expression of hSMVT was induced by replenishing the inducer methanol (at 0.5% (v/v)) every 12 h at 20°C for 1.5 days in growth medium with 50% reduced yeast extract and Bacto TM tryptone (BD Biosciences).
Purification of Recombinant hSMVT-P. pastoris cells expressing recombinant hSMVT were disrupted by a 5-fold passage through an EmulsiFlex-C3 homogenizer (Avestin) at To increase the enantiomeric excess, the method was repeated using the opposite resolving agent and seeded with preformed crystals of the undesired salt pair. The hydrocarbon solutions were extracted with aqueous acid to remove excess chiral amine, washed with H 2 O, dried, evaporated, and recrystallized, yielding essentially enantiomerically pure (Ͼ98%) crystals of R-LA or S-LA. B, chemical scheme to prepare R,S-LAM, R-LAM, and S-LAM. The enantiomerically pure acids were used to make the amides via the mixed anhydride method (19). LA was dissolved in THF and incubated with equimolar triethylamine and ethyl chloroformate. The intermediate mixed anhydride was incubated with 28% aqueous ammonia.
Immunological Detection of hSMVT-Immunological detection of recombinant hSMVT in oocytes was performed with the mouse monoclonal anti-FLAG M2 IgG (Sigma) or as control with a rabbit polyclonal antibody (SMVT-11A; Alpha Diagnostic, San Antonio, TX; data not shown) after solubilization of the oocytes as described (18). Immunoreactive bands were visualized with suitable horseradish peroxidase-conjugated secondary antibodies in conjunction with the enhanced chemiluminescence (ECL) method (SuperSignal West Pico kit, Thermo Scientific). The monoclonal anti-FLAG M2 antibody was used to detect recombinant hSMVT in membrane vesicles of P. pastoris or in detergent-solubilized form during purification.
Reconstitution of hSMVT-containing Nanodiscs-MSP1E3D1 and lipids were prepared as described (21) with small modifications. Briefly, MSP1E3D1 was expressed in Escherichia coli BL21(DE3) containing the MSP1E3D1-expressing plasmid (Addgene). The protein was purified using the PrepEase Histidine-tagged High Yield Purification Resin (Affymetrix) and dialyzed for 16 h in the presence of tobacco etch virus protease to remove the hexa-His tag from the purified protein. The sample was cleared from non-cleaved MSP1E3D1 and the hexa-His tag peptide by an additional Ni 2ϩ -based chromatography step. E. coli polar lipids, 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1Ј-racemic-glycerol) (POPG) (both from Avanti), and cholesterol (Sigma) were dissolved at the indicated ratios in chloroform and dried under a stream of nitrogen gas. The lipid film was resuspended at a concentration of 20 mM in 10 mM Tris, pH 7.4, 100 mM NaCl, 0.5 mM EDTA, and 0.01% NaN 3 . hSMVT was assembled into nanodiscs at a 0.1:1:500 molar ratio of hSMVT:MSP1E3D1: lipids. Detergent was removed by serial addition of Biobeads SM2 (Bio-Rad). The Biobeads were removed by filtration, and the hSMVT-assembled nanodiscs were resuspended in 100 mM Tris/Mes, pH 7.4 and 1 mM tris(2-carboxyethyl)phosphine after collection by ultracentrifugation at 365,000 ϫ g for 1 h. Samples of the preparation (nanodisc-containing pellet and superna-tant) were collected after the ultracentrifugation step and analyzed by silver staining of SDS-polyacrylamide gels. The protein concentration of all samples was measured with the Amido Black protein assay (22). The concentration of hSMVT in nanodiscs was quantified from silver-stained SDS-polyacrylamide gels with densitometry using NIH ImageJ and known amounts of purified hSMVT as reference.
Scintillation Proximity Assay (SPA)-based Binding Studies-50 ng of hSMVT in purified form or reconstituted in nanodiscs was bound to 250 g of Cu 2ϩ -coated yttrium silicate SPA beads in a 100-l assay volume. R-[ 3 H]LA binding experiments were performed in 50 mM Tris/Mes, pH 7.5, 150 mM NaCl (or 150 mM Tris/Mes), 20% glycerol, 1 mM tris(2-carboxyethyl)phosphine, and 0.1% (w/v) DDM. Equilibrium binding was performed in the dark for 16 h at 4°C with vigorous shaking on a vibrating platform, and the counts/minute (cpm) were determined in the SPA mode of a Wallac 1450 MicroBeta TM plate PMT counter. 800 mM imidazole (which prevents the interaction of the His-tagged protein with the Cu 2ϩ -coated SPA beads) was added to the samples to determine the non-proximity background signal. All experiments were performed at least in duplicate with replicas of Ն3, and data are expressed as mean Ϯ S.E. Data fits of kinetic analyses were performed using non-linear regression algorithms in Prism 5 (GraphPad), and errors represent the S.E. of the fit.

R-[ 3 H]LA Uptake in
Proteoliposomes-Purified recombinant hSMVT was reconstituted at a 1:100 (w/w) ratio in preformed, Triton X-100 (0.12%, w/w)-destabilized liposomes that were prepared from E. coli polar lipid extract in the absence or presence of 10% (w/w) cholesterol in 100 mM potassium phosphate, pH 7.5 and 2 mM ␤-mercaptoethanol. Prior to the uptake measurements, frozen proteoliposomes were thawed at 23°C and extruded through a 400-nm filter. The uptake reaction was initiated by the 20-fold dilution of the proteoliposome suspension (5 mg of lipid/ml) into assay buffer composed of 10 mM HEPES, pH 7.4, 100 mM NaCl (or 100 mM choline chloride), 1 mM CaCl 2 , and 1 mM MgCl 2 plus the indicated concentrations of R-[ 3 H]LA (1 Ci/mmol) at 23°C. Reactions were stopped by the addition of ice-cold 100 mM potassium phosphate, pH 6.0 and 100 mM LiCl and filtered through Millipore 0.22-m nitrocellulose filters. The radioactivity retained on the filters was determined with scintillation counting. Known amounts of radioactivity were used to convert cpm to mol.
The orientation of hSMVT in hSMVT-containing proteoliposomes was determined based on the accessibility of the hexa-His tag at the hSMVT C terminus to Ni 2ϩ affinity resin. 0.01 g of PrepEase Histidine-tagged High Yield Purification Resin was mixed with 1.5 g of hSMVT-containing intact proteoliposomes and incubated for 1 h at 4°C. The resin/protein mixture was loaded onto a disposable column, and the flow-through was collected. The resin was then washed with binding buffer (100 mM KP i , pH 7.5), and bound sample was eluted from the column with 400 mM imidazole-containing binding buffer. All fractions were collected and solubilized with 1% DDM, and normalized volumes were subjected to SDS-PAGE followed by silver staining of the gel.

Characterization of the Transport of LA by hSMVT in
Oocytes-In the presence of Na ϩ , lipoate elicits electrical currents in hSMVT-expressing oocytes, and uptake of radiolabeled pantothenic acid and biotin was shown to be inhibited by lipoate with a half-maximum effective concentration between 1 and 5 M lipoate (16). Direct evidence for the hSMVT-mediated translocation of lipoic acid across the plasma membrane was recently demonstrated with R,S-[ 3 H]LA, the racemic mixture of the S-and R-enantiomers, in hSMVT-expressing oocytes (18). However, to obtain reproducible data with a suitable signal-to-noise ratio, a Ͼ10 M concentration of racemic R,S-[ 3 H]LA was chosen for these uptake studies ( Fig. 2A). In contrast, performing the assay with 1.66 M R-[ 3 H]LA yielded a signal-to-noise ratio of ϳ1:10, a concentration that corresponds approximately to the half-maximum effective concentration determined with indirect measurements (Fig. 2B). The time course of 1.66 M R-[ 3 H]LA uptake in the presence of 100 mM NaCl exhibited a linear rate of transport for the initial 15 min with a steady state of accumulation of about 4 pmol per oocyte after 60 min (Fig. 2C). Assuming the diameter of an oocyte to be 1 mm, its internal volume is 0.523 l; at equilibrium conditions with 1.66 M R-LA inside and outside of the oocyte, there would be 0.87 pmol accumulated in the oocyte. These results demonstrate that hSMVT uses the Na ϩ electrochemical gradient to drive the accumulation of R-LA against its concentration gradient, a key feature of Na ϩ -coupled symport reactions (23). Uptake of R-[ 3 H]LA transport in hSMVT-expressing oocytes was saturable with a K m of 4.0 Ϯ 0.5 M and a maximum velocity of transport (V max ) of 15.5 Ϯ 0.6 pmol ϫ oocyte Ϫ1 ϫ min Ϫ1 (Fig. 2D). Removing Cl Ϫ from the assay buffer had no significant effect, confirming the Cl Ϫ independence of hSMVT-mediated substrate transport (Fig. 2E). Furthermore, the equimolar replacement of NaCl with LiCl or lowering the pH of the uptake assay from pH 7.4 to pH 5.5 (generation of a proton gradient across the oocyte membrane) resulted in R-[ 3 H]LA accumulation in hSMVT-expressing oocytes that was indistinguishable from that observed in control oocytes (Fig. 2E). To test whether the low signal-to-noise ratio of R,S-LA uptake could be attributed to the specific effect of S-LA on R-LA uptake or whether the effective concentration of the transported enantiomer was reduced in the racemic mixture, we performed uptake of 1.  (Fig. 2F).
Functional Expression of hSMVT in P. pastoris-To facilitate the immunological detection of hSMVT in suitable expression systems and its subsequent purification, we introduced a FLAG epitope and a hexa-His tag at the C terminus of hSMVT (hereafter referred to as recombinant hSMVT). As determined with R-[ 3 H]LA uptake studies, recombinant and unmodified hSMVT had virtually indistinguishable transport activity in oocytes (Fig. 3A) and can be immunologically detected with monoclonal anti-FLAG M2 IgG (Fig. 3B). Building on the availability of a fully functional recombinant hSMVT, we tested expression of hSMVT in P. pastoris, a eukaryotic host system that was previously successfully used for the expression of other mammalian (and human) recombinant membrane proteins (24). As judged by Western blot analysis using the anti-FLAG M2 IgG, inducing the expression of recombinant hSMVT from the methanol-inducible alcohol dehydrogenase promotor readily yielded full-length recombinant hSMVT in the plasma membrane of P. pastoris with a similar glycosylation pattern as that observed in oocytes (18). To confirm that the immunoreactive bands at ϳ65 and ϳ58 kDa corresponded to the differently glycosylated versions of hSMVT, we subjected the sample to peptidyl N-glycosidase F treatment. Complete peptidyl N-glycosidase F digestion of the detectable bands resulted in a single band, corresponding to non-glycosylated recombinant hSMVT, at ϳ55 kDa (Fig. 3C). In addition to the differently glycosylated forms of hSMVT at ϳ55, ϳ58, and ϳ65 kDa, we occasionally detected a faint immunoreactive band at ϳ40 kDa that may be a degradation product of recombinant hSMVT containing the C-terminal FLAG epitope. Recombinant hSMVT was purified by anti-FLAG immune affinity chromatography (25) after extracting the protein with DDM (Fig. 3D).
Interaction of LA with Purified Recombinant hSMVT-To test the activity of purified hSMVT in detergent-solubilized form, we assessed binding of R-[ 3 H]LA with the scintillation proximity assay (Fig. 4). In this assay, purified recombinant hSMVT was immobilized on copper-coated scintillation beads (using the hexa-His tag in recombinant hSMVT) that emit light only when a radioactive ligand stably binds to the immobilized target protein (26). 0.83 M R-[ 3 H]LA yielded a robust specific SPA signal when 50 ng (about 0.73 pmol) of purified hSMVT was used per assay (Fig. 4A). Performing the assay in the absence of NaCl yielded a signal that was indistinguishable from the non-proximity (background) signal obtained in the presence of imidazole that competes with the hexa-His-tagged protein for binding to the copper-coated SPA beads, indicative of the strict Na ϩ dependence of the interaction of LA with hSMVT and extending previous direct and indirect studies on the Na ϩ requirement for hSMVT activity (Fig. 4A).
To test the interaction of stereoisomers of LA and its derivative, we performed binding of 1.  (Fig. 4B).
We then measured equilibrium binding of increasing concentrations of R-[ 3 H]LA to 50 ng of purified hSMVT in the SPA (Fig. 4C). Reminiscent of substrate binding by the related solute:sodium symporter members vSGLT and PutP (27), binding of increasing concentrations of R-[ 3 H]LA by purified hSMVT revealed a biphasic saturation binding curve under equilibrium conditions. This curve was best fit with two separate one-site models (Fig. 4C), yielding a B max of 1.84 Ϯ 0.09 molecules of LA per molecule of hSMVT. Fitting the data of the first phase (data points between 0 and 5 M) preferred a single site model with Hill slope when compared with a single site model without Hill slope (p Ͻ 0.0001) and yielded a K d of R-LA binding to hSMVT of 0.9 Ϯ 0.03 M with a Hill slope of 1.99 Ϯ 0.11. Data points for the second phase (data points between 5 and 25 M) were fitted with a single site model, yielding a K d of 7.38 Ϯ 1.44 M.
These data indicate that the solubilization and purification of hSMVT with DDM did not adversely affect its activity. We then tested the effect of lipids on the activity of the protein reconstituted into nanodiscs of defined lipid composition. Purified hSMVT was incorporated into nanodiscs composed of polar E. coli lipids supplemented with 0, 6, or 10% (w/w) cholesterol or those composed of POPC/POPG (at a 3:2 (w/w) ratio) supplemented with 6% cholesterol. Incorporation of hSMVT into nanodiscs was highest in those composed of E. coli polar lipids in the presence of 6% cholesterol Ͼ E. coli polar lipids with 10% cholesterol Ն POPC/POPG with 6% cholesterol Ͼ E. coli polar lipids (Fig. 4D); however, the lipid composition had virtually no effect on the binding activity at a saturating concentration of R-[ 3 H]LA by hSMVT when the samples were normalized with regard to the amount of incorporated hSMVT (Fig. 4E).  6). B and C, immunological detection of the recombinant hSMVT (hS) in membranes of P. pastoris with monoclonal anti-FLAG M2 IgG after subjecting samples to 11% SDS-PAGE. B, membranes from P. pastoris cells harboring a control plasmid devoid of the recombinant hSMVT gene served as control (co). C, recombinant hSMVT in P. pastoris membrane vesicles (Ϫ) was subjected to peptidyl N-glycosidase F treatment (indicated by "ϩ"). Protein standards are shown between B and C. D, purification of hSMVT. Membranes were solubilized with 1.5% DDM (Sol), and hSMVT was bound to the ␣-FLAG affinity resin, washed (Wa), and eluted by competition with the FLAG peptide (El). 2 g of purified hSMVT (El) and samples with comparable fractional protein amount were subjected to 14% SDS-PAGE followed by silver staining.
hSMVT-mediated Uptake in Proteoliposomes-Guided by the results of the binding experiments performed in nanodiscs, we reconstituted affinity chromatography-purified hSMVT (see Fig. 3D) in liposomes premade with polar E. coli lipids that contained 10% (w/w) cholesterol. Previous studies (28) revealed the location of the hSMVT C terminus in the cytoplasm. Assessing the accessibility of the His tag (located at the C terminus of hSMVT) in hSMVT-containing proteoliposomes with Ni 2ϩ chelate affinity-based immobilization revealed that the transporter is unidirectionally oriented in outside-out orientation in the proteoliposome membrane (Fig. 5A). Uptake of 0.83 M R-[ 3 H]LA in hSMVT-containing proteoliposomes was strictly dependent on the presence of Na ϩ as the time course of uptake in the absence of NaCl was virtually indistinguishable from that observed in control liposomes lacking hSMVT (Fig.  5B). In the presence of 100 mM NaCl, the time course of R-[ 3 H]LA uptake peaked at about 10 min and showed a decline to equilibrium conditions at Ն3 h. Using the initial linear portion of uptake for the determination of the kinetics of R-LA transport, we measured 1-min intervals of R-[ 3 H]LA transport with varying concentrations ranging from 0.2 to 53 M. hSMVT-mediated transport was concentration-dependent with a concentration constant (K m ) at 50% of the V max of 3.9 Ϯ 0.6 M and a V max of 983.8 Ϯ 43.4 nmol ϫ mg of hSMVT Ϫ1 ϫ min Ϫ1 (Fig. 5C). Transforming this number revealed a k cat of transport of 1.13 Ϯ 0.05 s Ϫ1 . Given the orientation of hSMVT in the proteoliposome membrane (Fig. 5A), this result indicates that the determined k cat reflects the activity of hSMVT under physiological conditions, i.e. mediating the intracellular delivery of lipoic acid from the external medium.

Discussion
R-LA is an essential cofactor in many biological enzymatic reactions. In particular, R-LA serves as cofactor in central redox reactions related to energy production in the cell (4,7). Furthermore, free LA is an effective biological antioxidant due to the high energy content of the disulfide group in LA. The disulfide in LA is in a dithiolane, a five-membered ring, with a dihedral angle of 35°. This constellation results in an energetically unfavorable interaction of the orbital of the lone p z electron pairs of the sulfur atoms that in turn exerts a high energy content of the disulfide group of LA with oxidizing species (29). Because of this potent hydroxyl scavenger and chelating potential, LA has gained a salutatory role as a food supplement and in the treatment of several clinical conditions such as Amanita mushroom poisoning (30), diabetic neuropathy (31), Alzheimer disease (11), Parkinson disease (12), and multiple sclerosis (13). However, although only R-LA has been identified as the biologically active form of LA (32), R,S-LA has been administered in therapeutic applications as well as in food supplements simply due to more cost-efficient achiral production processes that were introduced in the 1950s (33). Whereas S-LA (as free form or in R,S-LA) is considered non-toxic, there is evidence that S-LA can adversely affect the activity of R-LA (34). To test whether this effect can be attributed to the cellular delivery of LA by the Na ϩ /multivitamin transporter, the transport system suggested to be central in the transport of LA (16), we characterized the interaction of LA with hSMVT using transport and binding studies.
hSMVT is a member of the SLC5 gene family (SLC5A6) (35) of the solute:sodium symporter family (36). hSMVT has been identified in essentially every tissue of the body (16) where it mediates the Na ϩ -dependent transport of the two vitamins biotin and pantothenic acid and iodide (I Ϫ ) (16,18). In analogy to Na ϩ -coupled substrate uptake of related SLC5 members SGLT1 (the Na ϩ /glucose transporter 1; SLC5A1) and the Na ϩ / iodide symporter (SLC5A5) (for recent reviews, see Refs. 37 and 38)), hSMVT has been suggested to couple the flux of its substrates to the flux of Na ϩ . By using radiolabeled LA, we recently showed that in addition to biotin, pantothenic acid (16), and iodide (18) hSMVT transports racemic LA (18). However, in contrast to the favorable signal-to-noise ratios observed for the transport of biotin, pantothenic acid, and I Ϫ at concentrations close to their reported (apparent) affinities, uptake of racemic R,S-LA exhibited a rather low signal-to-noise ratio ( Fig. 2A). Using radiolabeled R-LA, the biologically active form of LA, the signal-to-noise ratio of the uptake measurements performed with hSMVT-expressing oocytes increased dramatically, thus suggesting that R-LA is indeed the physiological substrate of hSMVT. Competition experiments involving the S and R,S isoforms of LA confirm this notion. Testing the effect of biotin and pantothenic acid, the other two organic substrates of hSMVT, on R-LA uptake revealed different levels of inhibition. This may indicate that biotin is competing with R-LA for the same site(s). Conversely, the relatively low inhibitory effect observed for pantothenic acid may indicate that different sites are used for R-LA and pantothenic acid transport. This complex interaction of hSMVT with its substrates requires a thorough analysis that is the subject of our ongoing investigations.
To assess the stereospecificity of the interaction of hSMVT with LA, hSMVT was successfully expressed in P. pastoris and purified in functional form. This allowed for the direct characterization of the LA-specific binding kinetics in purified, detergent-solubilized form as well as reconstituted in nanodiscs (21). Performing the reconstitution in this artificial membrane system of defined lipid composition showed that the addition of cholesterol increased the incorporation efficiency of hSMVT in the membrane lipid bilayer; however, the lipid composition had no significant effect on the activity of hSMVT per se. Guided by these results, we measured the LA uptake kinetics of hSMVTreconstituted proteoliposomes containing 10% cholesterol. Transport of R-LA by hSMVT was Na ϩ -dependent and yielded a K m of 3.9 Ϯ 0.6 M, a value virtually identical to that obtained in hSMVT-expressing oocytes (4.0 Ϯ 0.5 M). The catalytic turnover number of hSMVT-mediated transport (about 1 molecule of R-LA per s) is well in agreement with the turnover numbers described for other Na ϩ -dependent solute transporters (26, 39 -43). Furthermore, accessibility measurements of the His tag, which served as a reporter for the location of the C terminus (located in the cytoplasm in cellular systems), revealed that hSMVT is inserted in physiologically "correct" orientation; i.e. its cytoplasmic face is located in the lumen of the proteoliposomes. Thus, the k cat determined in our measurements can be considered as the catalytic turnover number of the uptake reaction that mediates the translocation of LA from the extracellular milieu into the cytoplasm. Taken together, these results strongly suggest that the expression of hSMVT in P. pastoris, its subsequent detergent-based extraction from the P. pastoris membrane, and purification maintain hSMVT in fully functional form. Note that the glycosylation pattern in oocytes is similar to that observed in the yeast.
Binding of R-LA (and R,S-LA; not shown) is strictly dependent on Na ϩ . Assessing saturation binding of R-LA with the purified, detergent-solubilized protein yielded an R-LA-to-hSMVT stoichiometry of ϳ2 with a biphasic binding isotherm. This result is in line with our recent findings (27) according to which the related solute:sodium symporter members vSGLT and PutP possess two substrate binding sites that are important for substrate translocation by proteins with a LeuT fold (see below). These studies extend our earlier studies on LeuT that reveal the importance of two substrate binding sites for Na ϩcoupled substrate transport (44) and for the conformational changes that are associated with the translocation process (45).
Taking advantage of the availability of functional hSMVT and the stable chiral forms of LA and LAM (46), we were able to test the interaction of the different racemic and enantiomeric forms of the two compounds with hSMVT in transport and binding assays. First, LAM did not inhibit R-LA binding and transport, indicating that the free carboxyl group of LA is important for hSMVT-mediated transport and that LA as lipoyl protein is not transported by the transporter. Adumbrating that hSMVT may share an overall similar structure with LeuT (47), referred to as the common LeuT fold that was also identified for the related solute:sodium symporter member vSGLT (48 -50), it is feasible to speculate that R-LA, like the substrates in LeuT, interact via ionic interactions with a Na ϩ bound in close proximity. This notion is supported by the fact that, like in LeuT (45), replacing Na ϩ with Li ϩ impaired transport of R-LA. For example, in LeuT, a molecule of leucine is bound in the socalled central or primary substrate binding site located in the core of the protein and formed by a specific set of residues. In addition, its carbonyl oxygen is coordinated by (or coordinates) the Na ϩ in the proximal Na1 site (47). Because our data highlight the significance of the hydroxyl oxygen of R-LA for its interaction with hSMVT, a similar molecular interaction of substrate and Na ϩ as observed for LeuT is conceivable. Second, the fact that S-LA did not inhibit hSMVT-mediated R-LA uptake and binding may indicate that R-LA is the physiological substrate of hSMVT. Consistent with this notion, the racemic mixture (at approximately equimolar ratio of R-LA and S-LA) shows partial inhibition of R-[ 3 H-]LA uptake and binding when compared with the inhibition caused by R-LA.
LA was found to be absorbed into the cells as the free form and later to be oxidized to dihydrolipoic acid or incorporated into enzymatic complexes (51). hSMVT may thus play an important role in the absorption of R-LA in the intestine and other tissues in the body (16). Studies in healthy volunteers showed R-LA in plasma and that its bioavailability was higher than S-LA after administration of racemic LA (52,53), suggesting that hSMVT transports exclusively R-LA across epithelial cells. However, the detailed pharmacokinetics of LA in humans is not fully characterized. Because of its high polymerization tendency, R-LA is relatively unstable and is poorly dissolved in aqueous solutions, reasons that may lead to poor absorption and low bioavailability. Dietary supplements that typically range from 50 to 600 mg are the primary sources of LA in animal models or clinical studies (14). A recent study shows that the plasma R-LA levels in healthy humans after administration of 600 mg of R-LA peak at about 15 min at a concentration of about 15 g/ml, which corresponds to a molar concentration of ϳ80 M R-LA (15), which is about 10% of the total administered R-LA. This result corresponds to previous studies that revealed that ϳ80% of LA was renally excreted (14,54). Although the amount of R-LA absorption in plasma and tissues was higher than that of S-LA, both enantiomers were found in plasma at a similar time after oral administration of the racemic mixture to healthy volunteers (52,53). Thus, the question remains of whether there are other/additional mechanisms involved in the absorption and secretion of LA at such high concentrations found in food supplements. Takaishi et al. (55) determined the transport of 500 M [ 14 C]LA in human intestinal Caco-2 cells. The transport was Na ϩ -independent and not inhibited by either biotin or pantothenic acid, thus suggesting that SMVT was not the main means of LA translocation in those studies. The authors proposed that the monocarboxylic acid transporter was probably involved in LA transport in the intestinal cell line (55). It may be possible that monocarboxylic acid transporter or another, yet unknown fatty acid transporter mediates the uptake of LA at concentrations Ն60 M, the highest concentration used in our assays. However, although there is currently no robust evidence for the involvement of another transport mechanism associated with the intracellular delivery of LA, we cannot rule out that such (additional) mechanisms exist.