Amino Acid Residues in Transmembrane Segment IX of the Na+/I– Symporter Play a Role in Its Na+ Dependence and Are Critical for Transport Activity*

The Na+/I– symporter (NIS) is a key plasma membrane glycoprotein that mediates Na+-dependent active I– transport in the thyroid, lactating breast, and other tissues. The OH group of the side chain at position 354 in transmembrane segment (TMS) IX of NIS has been demonstrated to be essential for NIS function, as revealed by the study of the congenital I– transport defect-causing T354P NIS mutation. TMS IX has the most β-OH group-containing amino acids (Ser and Thr) of any TMS in NIS. We have thoroughly characterized the functional significance of all Ser and Thr in TMS IX in NIS, as well as of other residues in TMS IX that are highly conserved in other transporters of the SLC5A protein family. Here we show that five β-OH group-containing residues (Thr-351, Ser-353, Thr-354, Ser-356, and Thr-357) and Asn-360, all of which putatively face the same side of the helix in TMS IX, plus Asp-369, located in the membrane/cytosol interface, play key roles in NIS function and seem to be involved in Na+ binding/translocation.

spread and highly successful use of radioiodide in the diagnosis and treatment of major thyroid diseases, including thyroid cancer and its metastases (4).
NIS is a member of the SLC5A transporter family (5), which includes the high and low affinity Na ϩ /glucose co-transporters (sodium/glucose co-transporters 1 and 2), the Na ϩ /myoinositol transporter (SMIT), and the Na ϩ /monocarboxylate transporter (SMCT), among others. All these proteins couple the inward transport of Na ϩ , which occurs in favor of its Na ϩ /K ϩ -ATPase-generated electrochemical gradient (6), to the simultaneous inward translocation of the corresponding solute against its gradient (5,6). NIS activity is electrogenic with a 2Na ϩ to 1I Ϫ stoichiometry. The current NIS secondary structure model (Fig. 1A) depicts NIS as a protein with 13 TMS, the amino terminus facing the extracellular milieu and the carboxyl terminus facing the cytosol; the location of both termini has been confirmed experimentally (7). The 13-TMS model has been generally regarded as a typical pattern for all members of the SCL5A family, with one extra TMS in some members (8). Although several common structural and functional patterns (such as Na ϩ dependence and conformational changes in the transporter after binding of the substrates) and high homology (identity ranges from 17.5% between NIS and SMIT to 55.5% between NIS and SMCT) exist among these proteins, little is known about the regions involved in the Na ϩ or solute translocation pathways.
Twelve cases of congenital I Ϫ transport defect (ITD), a condition that leads to hypothyroidism when untreated, have been reported as a direct cause of NIS mutations (2,3). We have characterized the ITD-causing T354P NIS mutation at the molecular level (9). Although inactive when transiently transfected into COS-7 cells, the T354P NIS protein is normally expressed, post-translationally processed, and correctly targeted to the plasma membrane. We showed the ␤-carbon OH at residue 354 to be essential for NIS function (9). Thr-354 is located in NIS TMS IX, which has the most ␤-OH group-containing amino acids (Ser and Thr) of any NIS TMS (Fig. 1). We have investigated the role of these and other residues within TMS IX. We found that some of them are involved in Na ϩ coupling and/or translocation. Our data suggest that the equivalent amino acids may have a similar function in other members of the SLC5A family of transporters.

EXPERIMENTAL PROCEDURES
Site-directed Mutagenesis-Site-directed mutagenesis was carried out as described previously (9). The oligonucleotides used for mutagenesis contained mutations for changing the codon to a selected amino acid codon. Final PCR products were subcloned into WT rNIS cDNA (pSVSport-rNIS) and sequenced to verify the substitutions.
Transient Transfections-COS-7 cells were transfected using the Lipofectamine reagent (Invitrogen) enhanced with Plus reagent (Invitrogen) according to the manufacturer's instructions. Flow cytometry was performed to determine transfection efficiency. After 2 days, transfected cells were assayed for I Ϫ uptake, cell surface biotinylation, immunoblot analysis, and immunofluorescence.
Iodide Transport-Cells transiently transfected with WT or mutant NIS cDNAs were assayed as described (10 PO 4 , and 5.55 mM glucose in 10 mM Hepes, pH 7.5 (HBSS). Cells were incubated in HBSS containing 20 M KI supplemented with 1 Ci of carrier-free Na 125 I to yield a specific activity of 100 Ci/mmol. For steadystate experiments, incubations proceeded for 1 h at 37°C in a humidified atmosphere and were terminated by aspirating the radioactive medium and washing twice with 1 ml of ice-cold HBSS.
To determine the amount of 125 I accumulated in the cells, 500 l of cold ethanol was added for 20 min at 4°C, and radioactivity was quantified in a ␥-counter. DNA was determined by the diphenylamine method after trichloroacetic acid precipitation. Uptake was expressed as picomoles of I Ϫ per g of DNA.
Results are the average of at least five different experiments performed in triplicate or sextuplicate. Data were analyzed with GraphPad Prism (Intuitive Software for Science, San Diego). I Ϫ uptake values are the means Ϯ S.E. Statistical significance was determined by t test analysis (twotailed), and differences were considered significant at p Ͻ 0.05.
For I Ϫ -dependent kinetic analysis, cells were incubated with the indicated concentrations of I Ϫ (2.5-300 M) and 140 mM NaCl for 4 min. Initial rate data were analyzed by a nonlinear regression using the following equation for I Ϫ -dependent The terms 0.06 ϫ [I Ϫ ] ϩ 0.60 correspond to background adjusted by least squares of the data obtained with nontransfected cells. For Na ϩdependent kinetic analysis, cells were incubated with the indicated concentrations of Na ϩ (0 -280 mM) and 20 M I Ϫ for 4 min. Osmolarity was kept constant with choline. Initial rate data were analyzed by a nonlinear regression using the following equation for Na ϩ - The terms 0.001 ϫ [Na ϩ ] ϩ 0.87 correspond to background adjusted by least squares of the data obtained with nontransfected cells. Results were analyzed by nonlinear regression in GraphPad Prism software. K m and V max values are the average of at least three experiments and are expressed as mean Ϯ S.E.
Immunoblot-SDS-PAGE and electroblotting to nitrocellulose were performed as described previously (7,11). All samples were diluted 1:2 with loading buffer and heated at 37°C for 30 min prior to electrophoresis. Immunoblots were carried out with the corresponding amount of 2 nM anti-Ct-rNIS Ab and a 1:3000 dilution of a horseradish peroxidase-linked sheep antirabbit IgG (Chemicon International). Both incubations were performed for 1 h. Polypeptides were visualized by enhanced chemiluminescence Western blot detection system (Amersham Biosciences). Nitrocellulose membranes were stripped and reprobed with anti-tubulin (total protein blots) or anti-Na ϩ /K ϩ -ATPase ␣-subunit (biotinylated protein blots) Ab (not shown).
Immunofluorescence and Confocal Microscopy-Cells were grown on glass coverslips, washed two times with phosphatebuffered saline (PBS), and fixed in 2% paraformaldehyde in PBS for 20 min, quenched with 50 mM NH 4 Cl for 10 min, and washed three times with PBS and then permeabilized with 0.2% (v/v) Triton X-100 in PBS containing 0.5% (w/v) bovine serum albumin (Sigma), 0.1 mM CaCl 2 , and 0.1 mM MgCl 2 (PBSACMT) for 5 min. Coverslips were incubated with 4 nM anti-Ct-rNIS Ab PBSACMT for 1 h, washed twice with PBS, and incubated with fluorescein-conjugated goat anti-rabbit (Vector Laboratories) PBSACMT for 1 h. Coverslips were mounted and examined using a Bio-Rad Radiance 2000 laser- scanning confocal microscope, using excitation wavelengths of 488 and 568 nm.
Biotinylation-Cell surface proteins were labeled with the membrane-impermeable biotinylation reagent Sulfo-NHS-SSbiotin (Pierce) as described (12). Transiently transfected cells were rinsed twice with PBS containing 1 mM MgCl 2 and 0.1 mM CaCl 2 , pH 7.4 (PBSCM), at 4°C. Cells were incubated with 1 mg/ml Sulfo-NHS-SS-biotin (Pierce) in a biotinylation medium containing 2 mM CaCl 2 , 150 mM NaCl, and 20 mM Hepes, pH 8.5, for 30 min at 4°C with gentle shaking. The reagent was quenched by washing twice with 100 mM glycine in PBSCM for 10 min. Cells were lysed with a buffer containing 150 mM NaCl, 5 mM EDTA, 1% (v/v) Triton X-100, 1% (v/v) SDS, and protease inhibitors in 50 mM Tris-HCl, pH 7.5 (lysis buffer), for 15 min at 4°C. SDS was diluted 10 times in lysis buffer without SDS. Cell surface proteins were isolated from the cell extract with streptavidin-agarose beads (Pierce) incubated overnight with rotation at 4°C. Beads were rinsed three times with lysis buffer without SDS, twice with high salt buffer containing 5 mM EDTA, 0.1% (v/v) Triton X-100, and 500 mM NaCl in Tris-HCl, pH 7.5, and once with 50 mM Tris-HCl, pH 7.5. The beads were eluted in SDS-PAGE sample buffer containing 200 mM dithiothreitol, heated for 5 min at 75°C, chilled for 10 min on ice, and loaded onto the gel.
Densitometric Analysis-Films were scanned, and the optical density of the bands was measured using the image analysis software ImageQuant (GE Healthcare). Only signals in the linear exposure range of the films were used. The amount of expressed protein (arbitrary units) was compared with WT NIS expression.

RESULTS
Five Amino Acid Residues (Thr-351, Ser-353, Thr-354, Ser-356, and Thr-357) in TMS IX Are Critical for NIS Function on Account of Their ␤-OH Groups-In a previous report on the detailed analysis of the ITDcausing T354P NIS mutant, we showed that the presence of a ␤-carbon OH group at residue 354, located in TMS IX, is essential for NIS function (9), suggesting that TMS IX is, in turn, a potentially significant region for NIS activity. As we noted that TMS IX has the most ␤-OH groupcontaining amino acids (Ser and Thr) of any TMS in NIS ( Fig.  1), we undertook the characterization of the specific roles played by these amino acids (in addition to Thr-354) in NIS Immunoblot and the corresponding surface biotinylation analysis of Ala, Pro, and Cys (C and D) and Ser and Thr (E and F) are shown. Total protein of lysed cells (5 g/lane) was isolated from NT cells or COS-7 cells transfected with NIS cDNA, electrophoresed, electrotransferred, and immunoblotted with 2 nM anti-rNIS Ab. Cell-surface biotinylated proteins were labeled using 1 mg/ml Sulfo-NHS-SS-biotin. Immunoblot analysis of surface-biotinylated polypeptides precipitated with streptavidin-agarose beads was performed with 2 nM anti-rNIS Ab. Nitrocellulose membranes were stripped and reprobed with anti-tubulin (total protein blots) or anti-Na ϩ /K ϩ -ATPase ␣-subunit Abs (biotinylated protein blots) (not shown) to determine the levels of NIS protein expression with respect to WT NIS. G, immunofluorescence analysis of Ala, Pro, Cys, Ser, and Thr substitutions. Permeabilized cells were incubated with 2 nM anti-rNIS Ab, followed by fluorescein-conjugated anti-rabbit Ab.
function. We individually substituted each Thr and Ser in TMS IX with Ala, which lacks an OH group, and carried out transport assays in COS-7 cells expressing the resulting mutant NIS proteins under steady-state conditions at 20 M I Ϫ (a concentration close to the K m for I Ϫ ) and a physiological concentration of Na ϩ (140 mM). Substitutions of Thr-351 ( Fig with Cys caused a decrease in I Ϫ transport similar to that caused by replacing these residues with Ala. However, when we individually substituted Ser-349 ( Fig. 2A, lane 3), Ser-358 (lane 9), or Thr-366 (lane 10) with Ala, we observed that NIS function either decreased only modestly or actually increased (98, 120, and 78% of WT NIS activity, respectively). Similar results were obtained at supersaturating external I Ϫ concentrations (160 M) (data not shown). These results suggest that Ser-349 and Thr-366, which according to our model are located close to the extracellular milieu and the cytosol, respectively (Fig. 1A), do not play a functional role comparable with that of Thr-351, Ser-353, Thr-354, Ser-356, and Thr-357, all of which together form a pocket embedded in the membrane (Fig. 1B). The only exception to this trend was Ser-358, a putative membrane-embedded residue whose substitution with Ala did not lead to loss of NIS activity ( Fig. 2A, lane 9).
To maintain the presence of OH groups at the ␤-carbons, we substituted each Thr with a Ser and each Ser with a Thr in TMS IX, thus generating the mutants T351S (Fig. 2B, , and T357S (lane 5), and we observed that NIS activity ranged from 40 to 65% that of WT NIS ( Fig. 2A, lane 2). This was much higher than that obtained with NIS proteins with Ala/Pro/Cys substitutions at the same positions, which was just 10% or less of WT NIS activity ( Fig. 2A). Although these data suggest that the presence of OH groups at these positions is indeed necessary for NIS activity (compare Fig. 2, A and B), they also demonstrate that such presence alone is not sufficient for full activity.

Expression and Plasma Membrane Targeting of Mutant NIS Proteins Resulting from All TMS IX Substitutions Are Similar to
Those of WT NIS-We subjected lysates from cells expressing either WT or TMS IX mutant NIS to immunoblot analysis with an anti-carboxyl terminus NIS Ab (Fig. 2C, Ala/Pro/Cys substitutions, and Fig. 2E, Ser/Thr substitutions). Polypeptides corresponding to precursor and mature NIS species were detected for all mutant NIS proteins to an extent similar to that of WT NIS. Expression of WT and mutant NIS proteins was compared after normalization using either tubulin for total lysed blots or Na ϩ /K ϩ -ATPase ␣-subunit for surface biotinylation blots (data not shown). We also observed virtually equal plasma membrane localization for mutant NIS proteins as compared with WT NIS by both cell surface biotinylation with a membrane-impermeable reagent (Fig. 2, D and F) and confocal immunofluorescence microscopy (Fig. 2G). These results demonstrate that the decrease in activity observed in the described NIS mutants was not because of either lower expression or a plasma membrane targeting defect.

Functional Role of Residues in the NIS TMS IX
AUGUST 31, 2007 • VOLUME 282 • NUMBER 35

JOURNAL OF BIOLOGICAL CHEMISTRY 25293
Substitutions in TMS IX Do Not Alter the NIS K m Values for I Ϫ -Given the above results, we sought to determine by kinetic analysis whether the partial impairment of function in the mutant NIS proteins was attributable to increased K m values for I Ϫ . We analyzed the kinetic properties of I Ϫ uptake in COS-7 cells expressing all the described TMS IX NIS mutants as compared with WT NIS (Fig. 3, A-C). Initial rates were assessed by measuring I Ϫ accumulation at 4-min time points over a range of I Ϫ concentrations (2.5-200 M), with a constant 140 mM Na ϩ concentration. A kinetic analysis of T354A, T354G, T354P, T354Y, S356P, and T357P NIS was not possible because I Ϫ uptake mediated by these mutants was almost undetectable. No major differences in K m for I Ϫ were observed in cells expressing NIS mutants that displayed significant transport activity with respect to WT NIS (23.7 Ϯ 2.1 M) (Fig. 3A). In contrast, the maximal rate of I Ϫ uptake (V max -I Ϫ ) was different in the various TMS IX NIS mutants (Fig. 3, A-C) and that variation was similar to that observed in I Ϫ uptake at steady state (Fig. 2, A and B). As the decreases in activity did not result from either lower expression or impaired targeting to the plasma membrane, the lower V max values observed most probably reflect a lower turnover rate.
An Increase in the Na ϩ Concentration Rescues I Ϫ Transport in Ala Mutant NIS Proteins That Were Inactive at the Physiological Concentration of Na ϩ -To investigate the possible involvement of residues Ser-349, Thr-351, Ser-353, Thr-354, Ser-356, Thr-357, Ser-358, and Thr-366 in NIS Na ϩ dependence, we carried out I Ϫ transport assays under steady-state conditions at 20 M I Ϫ and 280 mM Na ϩ , compared the results with those obtained at 140 mM Na ϩ (Fig. 2, A and B), and expressed the difference as a ratio. Virtually no change in activity was observed at 280 mM Na ϩ with respect to 140 mM Na ϩ in cells expressing WT NIS (Fig. 4A, lane 1) or in substitutions that displayed I Ϫ uptake activity comparable with that of WT NIS, namely S349A (Fig.  4A, lane 2), S358A (Fig. 4A, lane 8), and T366A NIS (Fig. 4A, lane 9). Interestingly, with the exception of T354A, T354P, and T354Y, all mutant NIS proteins that displayed very low or no I Ϫ uptake at 140 mM Na ϩ exhibited a significant increase in I Ϫ transport at 280 mM Na ϩ , ranging from 1.4-fold in T357P to 2.1-fold in T357A NIS. Furthermore, in two mutants (S356P and T357P NIS) that did not accumulate I Ϫ at 140 mM Na ϩ ( Fig.  2A), there was significant I Ϫ transport at 280 mM Na ϩ . That the impairment of NIS function caused by the absence of OH groups at positions 351, 353, 356, and 357 was partly overcome by a higher Na ϩ concentration suggests that these residues are involved in the Na ϩ dependence of NIS.
The T351S, S353T, T354S, S356T, and T357S mutant NIS proteins (which, because of the presence of OH groups at these positions, exhibited 40 -65% of WT NIS activity at 140 mM Na ϩ (Fig. 2B)) also displayed increased I Ϫ transport activity when assayed at 280 mM Na ϩ (Fig. 4B). The increased activity at 280 mM Na ϩ for most mutants was significantly more pronounced in the Ala/Cys than in the Thr/Ser mutants (compare Fig. 4, A  and B). Thus, the increase was 1.8-fold in S353A versus 1.4-fold experiments; in each experiment, activity was analyzed in triplicate. Statistical significance was evaluated by t test analysis (two-tailed); differences were considered significant at p Ͻ 0.05 (*), 0.05-0.001 (**), and Ͻ0.001 (***). C, to assess Na ϩ dependence of I Ϫ uptake, cells were incubated for 4 min with the indicated concentrations of Na ϩ ; isotonicity was maintained constant with choline chloride. Na ϩ dependence data were analyzed using the equation v ϭ (V max ϫ [Na ϩ ] 2 )/ (K m 2 ϩ [Na ϩ ] 2 ) ϩ 0.001 ϫ [Na ϩ ] ϩ 0.87. The term 0.001 ϫ [Na ϩ ] ϩ 0.87 corresponds to the background adjusted by linear regression analysis obtained with NT cells. V max -Na ϩ and K m -Na ϩ values are indicated in the table. Values are the determination Ϯ S.E. of 3-16 experiments. Graphs correspond to a representative experiment. * indicates the extremely low levels of transport activity by these NIS mutants precluded accurate determination of kinetic parameters. in S353T; 1.7-fold in S356A versus no increase in S356T; and 2.1-fold in T357A versus 1.6-fold in T357S. The exception again was Thr-354. The T351S, T354S, and T357S, but not S353T and S356T NIS mutants, recovered full activity when assayed at 280 mM Na ϩ , which may reflect the specific properties of these amino acids (see "Discussion"). Clearly, substitution of Thr-351, Ser-353, Ser-356, and Thr-357 with residues that are devoid of OH groups impairs the Na ϩ dependence of NIS to a greater extent than substitutions that maintain OH groups.
The K m for Na ϩ Is Altered in TMS IX NIS Mutant Proteins-We carried out a kinetic analysis of Na ϩdependent I Ϫ transport in COS-7 cells expressing TMS IX NIS substitutions as compared with WT NIS. Initial rates were assessed by measuring I Ϫ accumulation at 4-min time points over a range of Na ϩ concentrations of 0 -300 mM, with a constant 20 M I Ϫ concentration, and keeping an isotonic medium with choline chloride (Fig. 4C). Most tellingly, significant variations in the K m(Na ϩ ) values with respect to WT NIS (64.7 Ϯ 2.8 mM) were observed in cells expressing the various TMS IX mutant NIS proteins. More pronounced increases in K m(Na ϩ ) values were measured in cases where lower steady-state I Ϫ transport was observed, reaching values that were even higher than the physiological extracellular Na ϩ concentration (137-142 mM) (13) (Fig. 4C). Variations in the maximal rate of Na ϩ -dependent I Ϫ uptake (V max -Na ϩ ) in the TMS IX NIS mutant proteins were quite similar to those observed in the steady-state analysis using 20 M I Ϫ and 280 mM Na ϩ (Fig. 4, A and B). These results suggest that Ser and Thr in TMS IX may be involved in Na ϩ -dependent transport.
Several Residues in NIS TMS IX Are Highly Conserved in Other Na ϩdependent Transporters-It is interesting to note that TMS IX is one of the regions of NIS that exhibit the highest overall homology with other members of the Na ϩ -dependent transporter family (Fig. 5). OH-containing Thr and Ser residues (such as the above-described Thr-351, Ser-353, Thr-354, Ser-356, and Thr-357), whose substitutions gave rise to mutant proteins with higher K m values for Na ϩ and which are likely to be involved in Na ϩ dependence, are among the conserved ones (Fig. 5A). This suggests that corresponding amino acids in other proteins in the family may also participate in Na ϩ dependence. Given this notion, we set out to examine whether a high degree of conservation of certain NIS TMS IX residues, independently of whether or not they contain OH FIGURE 5. A, multiple sequence alignment of TMS IX in some members of the SLC5A gene family. The sequence alignment was made using ClustalW with the PAM250 weight table (25). The amino acid numbering (345-367) corresponds to the rNIS sequence. The most conserved (black) and the closely related amino acids (gray) in NIS homologues are highlighted. Only Ser-353, Asn-360, and Glu-369 are conserved throughout. Thr-354, Ser-356, and Thr-357 are Ser or Thr in all homologues, except in SMIT transporters. The gene names and the gi accession numbers are written next to the transporter names. Two species of each transporter were selected for the analysis. A sequence pattern that contains all critical residues for transport in this paper is indicated at the bottom. B, NIS TMS IX and A. aeolicus LeuTAa TMS VIII sequence alignment. Note that the OH-containing NIS residues 353 and 354 align with the OH-containing LeuTAa residues 354 and 355. Also, the NIS amide sidechain residue Asn-360 aligns with the LeuTAa amide side-chain residue Gln-361 and NIS Glu-368 and Asp-369 align with LeuTAa Asp-369 and Glu-370. LeuTAa contains 6 of 10 residues in the sequence pattern that we defined as potentially involved in Na ϩ transport and 5 of 6 of the critical residues for NIS transport within the pattern. C, multiple sequence alignment of TMX VIII in some members of SLC6 family. Black triangles indicate important residues for NIS activity. Residues that coordinate Na ϩ in the LeuTAa crystal structure are indicated in gray. The star shows a residue that forms a salt bridge suggested to be an intracellular gate (20) that controls the access of Na ϩ and substrates in the SLC6 family (21). groups, may be an indicator of the possible functional significance of these residues vis à vis Na ϩ dependence. Asn-360 and Asp-369 are 100% conserved in all members of the Na ϩ -dependent transporter family (Fig. 5A). In contrast, Cys-346 is only present in three other transporters (sodium/glucose cotransporter 1, SMCT, and SMCT2), and Ala-364 only in SMCT2. The rest of the members have a Val at position 346 and a Thr at 364 (Fig. 5A).
We generated C346A, N360A, N360Q, N360E, A364T, A364G, E368A, D369A, and E368A/D369A mutant NIS cDNAs; transfected each one individually into COS-7 cells; and assayed the cells for I Ϫ transport activity. All resulting NIS proteins were expressed and targeted to the plasma membrane similarly to WT NIS (Fig. 6, A and B). However, all substitutions, except C346A and E368A, caused a decrease in I Ϫ transport activity (11-47% in Asn-360 substitutions, 40 -76% in Ala-364 substitutions, 44% in Asp-369, and only 8% in the ED368/369AA double mutant) (Fig.  6C). I Ϫ uptake at 280 mM Na ϩ increased with respect to that at 140 mM Na ϩ only in cells expressing NIS proteins with substitutions of the conserved Asn-360 and Asp-369 but not in those expressing NIS with substitutions of the nonconserved Cys-346, Ala-364, or Glu-368 (Fig.  6D). Furthermore, a kinetic analysis of Na ϩ -dependent I Ϫ transport demonstrated a higher K m value for Na ϩ in NIS molecules with replacements at positions 360 and 368 than in WT NIS, whereas the K m values for Na ϩ in NIS with Cys-346 and Ala-364 substitutions were similar to those of WT NIS (Fig. 6F). Also, none of these replacements led to significantly different K m values for I Ϫ from those of WT NIS (Fig. 6E). These findings suggest that the highly conserved Asn-360 and Asp-369 may play a role in the Na ϩ translocation pathway, whereas the less conserved Cys-346 and Ala-364 do not.

DISCUSSION
The study of NIS mutations that cause congenital ITD has yielded extremely valuable structure/function information on NIS. Perhaps the most intriguing mutation is T354P, the analysis of which demonstrated that the presence of a ␤-OH group at position 354 is required for NIS to be functional (9). Several studies have shown that OH-containing amino acids may participate in ion translocation in other transporters, such as the Na ϩ -dicarboxylate co-transporter, the glucose transporter, the Na ϩ /K ϩ ATPase, the glutamate transporter (GLT-1), and the bacterial Aquifex aeolicus Na ϩ -dependent leucine transporter, whose high resolution structure has recently been determined (14 -18). Significantly, NIS TMS IX, where Thr-354 is located, contains the highest number of amino acids with OH groups at the ␤-carbon of any TMS in the protein, and many of these residues are conserved in all the family's transporters (Fig. 5A). In this study, we have analyzed the functional role of all OH-containing and conserved residues in NIS TMS IX (Fig. 1). Remarkably, all mutant NIS proteins we generated exhibited normal maturation, expression, and targeting to the plasma membrane (Fig.  2, C and G, and Fig. 6, A and B), making it clear that the observed changes in transport properties are a direct consequence of local structural changes caused by the engineered amino acid substitutions.
The first key finding of this study is that the presence of five ␤-carbon-OH-containing residues (Thr-351, Ser-53, Thr-354, Ser-356, and Thr-357) located in NIS TMS IX is critical for the protein to be functional. Activity was no higher than 10% of WT NIS when Ala, Pro, or Cys were individually substituted into these positions ( Fig. 2A). Only when we kept an OH group at the ␤-carbon by substituting Thr with Ser or Ser with Thr at these positions were we able to recover transport activity, which ranged from 40 to 65% of WT NIS (Fig. 2B). However, the presence of these OH groups alone was not enough for full activity. Ala substitutions at the positions of the other three OH-containing residues in NIS TMS IX (Ser-349, Ser-358, and Thr-366) did not significantly alter NIS function and caused only a modest decrease or increase in activity (Fig. 2A). These results suggest that these residues do not play any functional role.
None of the substitutions significantly altered the I Ϫ K m value as compared with WT NIS (15-35 M) (Fig. 3). In some cases (i.e. T351A in Fig. 3A), we obtained a higher K m(I Ϫ ) with respect to WT NIS, although it generally occurred with substitutions that yielded proteins with low V max(I Ϫ ) values. This attribute made it very difficult to obtain accurate values for kinetic parameters for initial rates, a point reflected in the higher S.E. values obtained (Fig. 3). The observed V max variations are in close relation to steady-state I Ϫ uptake (Fig. 2, A and  B). Next, we investigated whether these residues are involved in the Na ϩ dependence of I Ϫ transport. Ala-substituted NIS proteins exhibited a statistically significant increase in I Ϫ transport (1.5-2.2-fold) when the Na ϩ concentration was raised from 140 to 280 mM (Fig. 4A). Two substitutions (S356P and T357P) that were inactive at 140 mM of Na ϩ recovered partial activity when the Na ϩ concentration was increased to 280 mM. With the sole exception of S356T, all replacements that preserved the ␤-OH group either recovered full activity (T351S, T354S, and T357S) or increased I Ϫ transport 1.5-fold (S353T) (Fig. 4, B and  C). That Ser can completely recover WT NIS activity at saturating Na ϩ concentrations when placed at positions where the original amino acid was Thr, but not the other way around, may be because in the context of a helix Thr and Ser have different properties; the side chain of Thr is constrained to a single rotamer, whereas the OH of Ser can adopt any of the three possible rotamers (g ϩ , g Ϫ , and t). When the WT sequence has a Ser, it seems that interactions with surrounding residues restrict the position of the OH to the rotamer allowed for Thr, so substitution by Thr does not change the position of the OH. When the WT sequence has a Thr, it appears that the position of the OH is not determined by interactions with other residues but rather by the ␤-branched side chain. The replacement of Thr with Ser, which can adopt other rotamers, would thus weaken the interaction with Na ϩ .
The Na ϩ dependence of I Ϫ transport in most NIS mutant proteins showed higher K m(Na ϩ ) values (90 -164 mM) than WT NIS (Fig. 4C). These results suggest that the studied NIS TMS IX residues are involved in Na ϩ binding and/or translocation during transport. Unlike the situation with the I Ϫ K m , and given that the physiological concentration of Na ϩ is ϳ140 mM (13), the observed increases in the Na ϩ K m are highly significant, because some of these mutants are essentially inactive at 140 mM Na ϩ (i.e. T354C, S356P, and T357P). This is so not only because of the apparent decrease in Na ϩ affinity but also because the partial or total impairment of the Na ϩ coordination may ultimately impair substrate coupling. Kanner et al. (19) have similarly shown that amino acid substitutions in TMS VIII of the eukaryotic GABA transporter GAT1 have an effect on cation specificity and cause a 2-3-fold increase in the Na ϩ K m and a 2.5-8-fold increase in the GABA K m , even though TMS VIII residues do not interact with GABA.
Finally, we examined whether our results were applicable to other proteins in the SLC5A family. NIS TMS IX alignment with the corresponding TMS in other transporters exhibits a high overall homology (Fig. 5A). All proteins contain a very high number of OH-containing residues, and most of the functionally critical OH-containing residues in NIS are conserved in the other transporters. We analyzed other residues in NIS TMS IX as indicators of their possible functional role in Na ϩ transport. Interestingly, substitutions of the NIS residues Asn-360 and Asp-369, which are 100% conserved in all family members, increased the K m value of Na ϩ , even though transport activity was less than half that in WT NIS under normal conditions. However, replacement of Cys-346 with Ala (a site where other family members have a Val or Gly), Ala-364 for Thr or Gly (where most members have a Thr), and E368A (where other members have Met, Ile, or Leu) (Fig. 5A), yielded K m values for Na ϩ that were no different from those of WT NIS (Fig. 4C). These data indicate that the highly conserved Asn-360 and Asp-369 may also influence Na ϩ dependence, whereas the less conserved Cys-346, Ala-364, and Glu-368 do not. The single mutant D369A retained 50% of WT activity, even though the data indicate that Asp-369 is critical for the Na ϩ pathway. This is most probably because the functional role of a negatively charged residue in that region may be compensated by the neighboring Glu-368 residue. As predicted by this hypothesis, the double mutant E368A/D369A NIS barely transported I Ϫ .
Yamashita et al. (20) recently reported a high atomic resolution structure of a Na ϩ -driven transporter, the A. aeolicus Na ϩ -dependent leucine transporter (LeuTAa), a bacterial homologue of eukaryotic Na ϩ -dependent neurotransmitter transporters. In this high resolution structure (1.65 Å), the substrates (2 Na ϩ ions and 1 Leu) are bound to the transporter. Remarkably, some of the residues (Thr-354 and Ser-355) that are in contact with the second Na ϩ site in this TMS VIII of the protein (20) correspond exactly to the Ser-353 and Thr-354 residues of NIS TMS IX (Fig. 5B). Also, LeuTAa has a Gln residue at position 361 that aligns with NIS Asn-360 (Fig. 5B). Furthermore, LeuTAa has a pair of negatively charged residues after TMS VIII at positions 369 and 370 that align with NIS Glu-368 and Asp-369 (Fig. 5B). These LeuTAa residues are strictly conserved among NSS family members (to which Leu-TAa, the serotonin and the dopamine transporters belong) and have been proposed to be implicated in the salt bridge that forms an intracellular gate that controls the access of Na ϩ and substrates (20). Substitutions of the equivalent residues in dopamine transporters (Asp-436 and Glu-437) impair dopamine transport (21).
Strikingly, there is more similarity between LeuTAa TMS VIII, which contains the residues that coordinate the second Na ϩ , and NIS TMS IX than between LeuTAa TMS VIII and TMS VIII of the eukaryotic neurotransmitter transporters of the SLC6 family, even though LeuTAa is a bacterial homologue of this family (Fig. 5C). Although these could be merely coincidental similarities, they also raise the possibility that the Na ϩ transport and coupling mechanisms may be conserved among several Na ϩ -driven co-transporters of different families. We analyzed a pattern from NIS TMS IX derived from our results, (S/A)X 3 (S/A)X 3 (S/T)(S/T)X(S/T/A)(S/T)X 2 (N/Q)X 2 (S/T/A)-(S/T/A)X 4 (D/E) (Fig. 5A), and we searched it in all protein sequences contained in the data base using the ScanProsite software tool (22,23). Of a total of 312 retrieved sequences, all but 27 proteins belonged to the SLC5A family, of which NIS is a member. Given that SLC5A is a family of Na ϩ -dependent transporters, it is not surprising that a sequence linked to a likely role in Na ϩ binding and/or translocation is present in proteins of this family. Of note, LeuTAa TMS VIII contains 6 of 10 residues of this sequence pattern and 5 of 6 of the critical amino acids.
According to our secondary structure model, amino acids that we found to be important in Na ϩ -dependent I Ϫ transport are present on an extended surface that runs across one face of helix IX. During the transport cycle of NIS, Na ϩ may contact different residues because of movements of Na ϩ and/or rotation of helix IX. The negatively charged Glu-368 and Asp-369 at the end of TMS IX could constitute the Na ϩ gate, similarly to what happens in LeuTAa (20) and dopamine transporters (21).
In summary, we have shown that five OH-containing residues (Thr-351, Ser-353, Thr-354, Ser-356, and Thr-357), Asn-360 in NIS TMS IX, and Asp-369 at the membrane-cytosol interface, are all critical for NIS function and probably for other transporters of the SLC5A family. These amino acids seem to play a critical role in Na ϩ binding and/or the lining of the Na ϩ translocation. Our results also suggest that Na ϩ binding/translocation does not occur solely through the amino-terminal part of the protein, as has been suggested for sodium/glucose cotransporter 1 as a general mechanism for the SLC5A family (24). We propose that the sequence pattern we have identified in TMS IX along with residues from other TMS may define a region involved in Na ϩ binding/translocation in members of the NIS family. The complete mechanism will be elucidated only when atomic resolution structural data from different conformations of these transporters become available.