Kinetics of steady-state currents and charge movements associated with the rat Na+/glucose cotransporter.

The rat Na+/glucose cotransporter (SGLT1) was expressed in Xenopus oocytes and steady-state and transient currents were measured using a two-electrode voltage clamp. The maximal glucose induced Na+-dependent inward current was ∼300-500 nA. The apparent affinity constants for sugar (α-methyl-D-glucopyranoside; αMDG) (KαMDG0.5) and sodium (KNa0.5) at a membrane potential of −150 mV were 0.2 mM and 4 mM. The KαMDG0.5 increased continuously with depolarizing potentials reaching 40 mM at −30 mV. KαMDG0.5 was steeply voltage dependent, 0.46 mM at −30 mV and 1 mM at −10 mV. From all tested monovalent cations only Li+ could substitute for Na+, but with lower affinity. The relative substrate specificity was D-glucose > αMDG ≈ D-galactose > 3-O-Me-Glc β-naphthyl-D-glucoside uridine. Phlorizin (Pz), the specific blocker of sugar transport, showed an extremely high affinity for the rat cotransporter with an inhibitor constant (KPzi) of 12 nM. SGLT1 charge movements in the absence of sugar were fitted by the Boltzmann equation with an apparent valence of the movable charge of ∼1, a potential for 50% maximal charge transfer (VαMDG) of −43 mV, and a maximal charge (Qmax) of 9 nanocoulombs. The apparent turnover number for the rat SGLT1 was 30 s−1. Model simulations showed that the kinetics of the rat SGLT1 are described by a six-state ordered nonrapid equilibrium model, and comparison of the kinetics of the rat, rabbit and human cotransporters indicate that they differ mainly in their presteady-state kinetic parameters.

Cotransporters are membrane proteins that use the electrochemical potential gradient for ions to accumulate sugars, amino acids and osmolytes into cells. Using the electrochemical potential gradient for Na ϩ , the Na ϩ /glucose cotransporter (SGLT) 1 accumulates glucose across the brush border membrane of the epithelial cells of the intestine and the proximal tubule of the kidney.
Several members of the SGLT family have been cloned, and these include the high affinity glucose cotransporters (SGLT1, K 0.5 ␣MDG ϳ 0.2 mM) from rabbit small intestine (1) and kidney (2), pig (3), and rat kidney (4), human intestine (5), as well as the low affinity glucose cotransporter (pSGLT2, K 0.5 ␣MDG ϳ2 mM) from pig kidney (6). Mapping the genomic arrangement of the human SGLT1 gene, Turk et al. (7) showed that SGLT1 is a single-copy gene, so that the amino acid sequences from various tissues of a given species are identical. Comparison of the amino acid sequences from the rat, human and rabbit clones reveal 86 -87% identity and 93-94% similarity. How does this high degree of homology between the three clones affect their kinetic properties? In this study, we characterized the presteady-state and steady-state kinetics of the rat SGLT1 clone with a view to understand the relationship between structure and function of members of the SGLT1 family.

MATERIALS AND METHODS
The pBluescript II SK plasmid containing the coding sequence for rat SGLT1 (4) was linearized with SalI and transcribed in vitro with T3 RNA polymerase (11). The cRNA was overexpressed in Xenopus oocytes and protein function studied 5-10 days after injection using the twomicroelectrode voltage clamp (8,11). To obtain a current-voltage (I-V) relationship, the membrane voltage was stepped for 100 ms to various test values (V t ) between 50 and Ϫ150 mV in 20-mV decrements and returned to the holding potential (Ϫ50 mV). Averaged currents from three sweeps were low-pass filtered at 500 Hz by an 8-pole Bessel filter and digitized at 100 s/point. In experiments to study the substrate and cation specificity (see Figs. 3 and 4) the currents were continuously monitored on a chart recorder.
Nonlinear regression analyses were performed using the software ENZFITTER (Elsevier-Biosoft, Cambridge, UK), and the fitting routines in Sigmaplot (Jandel Scientific, San Rafael, CA). The Marquardt-Levenberg algorithm was used by both programs.
Data presented in Figs. 3, 4, and 5A were carried out on the same oocyte. Similar results were obtained by repeating the experiments 2-4 times on oocytes from different donors. Fig. 1A (left panel) shows the current records from a rat SGLT1 cRNA-injected oocyte bathed in the 100 mM NaCl buffer in absence of sugar. The membrane potential was held at Ϫ50 mV and then stepped for 100 ms to test potentials (V t ) of 30, Ϫ10, Ϫ50, Ϫ90, and Ϫ150 mV. The current relaxation during both the ON and OFF current responses consisted of the capacitive transient followed by a slower decay to the steady state. The slow decay is the presteady-state current of SGLT1 and has been observed in the human and rabbit intestinal Na ϩ /glucose cotransporters (8 -11). Addition of ␣MDG (400 M) to the bath solution generated an inward Na ϩ -current, and abolished the presteady-state current (Fig.  1A, right panel). The sugar-induced steady-state current and the presteady-state currents were not observed in noninjected oocytes.

Steady-state Kinetics-
The steady-state current-voltage (I-V) relationship of the sugar-induced current is the difference in steady-state current in the absence, and in the presence of ␣MDG. Fig. 1B shows a family of sigmoidal I-V curves obtained as [␣MDG] o increased from 31 M to 20 mM. At each test potential (V t ), increasing [␣MDG] o increased the sugar-induced current until saturation was reached at 5 mM. For each ␣MDG concentration, as the test potential was made more negative, the current increased for V t between 0 and Ϫ100 mV, and then became independent of membrane voltage.
The voltage dependence of the apparent K 0.5 for sugar (K 0.5 ␣MDG ) is shown in Fig. 1C. Between Ϫ150 and Ϫ50 mV, K 0.5 ␣MDG was relatively insensitive to membrane voltage. However, between Ϫ30 and Ϫ10 mV, K 0.5 ␣MDG increased steeply with depolarizing potentials, from 0.46 Ϯ 0.03 at Ϫ30 mV to 1.0 Ϯ 0.2 mM at Ϫ10 mV. The calculated maximal current I max ␣MDG at Ϫ150 mV for this oocyte was Ϫ265 nA.
To determine the Na ϩ -dependence of the sugar-evoked currents, the steady-state inward currents were measured as [Na ϩ ] o was varied from 0 to 100 mM while [␣MDG] o was maintained at 5 mM. Fig. 2A shows the Na ϩ -dependent sugarevoked current at V t ϭ Ϫ30, Ϫ50, and Ϫ70 mV. At each V t the current was described by the Hill equation (see legend to Fig.  2B). There was a steep voltage dependence of the apparent affinity for sodium (K 0.5 Na ). K 0.5 Na increased from 4 Ϯ 0.6 mM at Ϫ150 mV to 40 Ϯ 2 mM at Ϫ30 mV. The Hill coefficient (1.8 Ϯ 0.3) was independent of voltage for V t between Ϫ150 to Ϫ50 mV, and the I max Na was Ϫ318 nA at Ϫ150 mV. Substrate and Cation Specificity- Fig. 3A shows the currents induced by various substrates with the oocyte membrane potential held at Ϫ50 mV. Since the experiment was performed on the same oocyte, the magnitude of the current induced by the different substrates indicates the relative affinity of the cotransporter for the substrates. The current induced by Dglucose, ␣MDG, and D-galactose were the largest (Ϸ200 nA). The L-isomer of glucose is a poor substrate (5 nA). 3-O-Methyl-D-glucopyranose acts as a substrate with moderate activity (Ϫ160 nA). The Na ϩ /myo-inositol cotransporter shares high amino acid sequence homology (46% identity) to SGLT1 (12), and myo-inositol has been shown to be a substrate of rabbit SGLT1 (13). Fig. 3A shows that myo-inositol is not transported by the rat SGLT1. The Na ϩ /glucose cotransporter also shows high amino acid sequence homology to the Na ϩ /nucleoside cotransporter (14). Competition experiments using radioactive tracers showed that uridine inhibited the uptake of ␣MDG, ). C, voltage dependence of the apparent affinity for ␣MDG (K 0.5 ␣MDG ). At each membrane potential (V t ) the ␣MDG-induced inward currents (I) were fitted to the equation: where I max ␣MDG is the apparent maximal current at saturating ␣MDG concentrations and K 0.5 ␣MDG is the sugar concentration at 50% I max ␣MDG . K 0.5 ␣MDG was 0.2 mM at Ϫ150 mV, increased to 0.46 mM at Ϫ30 mV and 1 mM at Ϫ10 mV. I max ␣MDG was Ϫ265 nA at Ϫ150 mV. Similar results were obtained in two other experiments. . I max Na is the maximal current at saturating Na ϩ concentrations, K 0.5 Na is the value of [Na] o at 50% I max Na , and n is the apparent coupling coefficient for Na ϩ . K 0.5 Na was more voltage-dependent at depolarizing membrane potentials. The values for K 0.5 Na were 4 Ϯ 0.6 mM at Ϫ150 mV and 40 Ϯ 2 mM at Ϫ30 mV, but n remained independent on voltage. As an example, at V t ϭ Ϫ50 mV all three parameters were: I max Na ϭ Ϫ221 Ϯ 2 nA; K 0.5 Na ϭ 26 Ϯ 2 mM; n ϭ 1.8 Ϯ 0.3. The errors are errors of the fit. Almost identical findings were obtained from two additional oocytes.
suggesting that the nucleoside may be a substrate (4). We tested whether uridine and formycin B, substrates of the nucleoside transporter, are transported by rat SGLT1. Fig. 3B shows that both compounds are transported by rat SGLT1, but with very low affinity since the current generated by 50 mM uridine or 10 mM formycin B was only Ϫ7 and Ϫ11 nA, respectively. The low affinity for uridine transport by rat SGLT1 is also indicated by the observation that uridine caused a weak inhibition of the current induced by 400 M ␣MDG (Fig. 3B, Consistent with reports about the rat intestinal absorption of glucose-conjugated compounds (15) we observed ␤-naphthyl ␤-D-glucopyranoside induced Na ϩ inward currents. Addition of 20 mM of this compound induced about 18 -20% (Ϫ50 nA at Ϫ150 mV) of the recorded currents induced by D-glucose (Fig. 3A).
The ability of the monovalent cations Li ϩ , K ϩ , Rb ϩ , and Cs ϩ to substitute for Na ϩ was also examined. NaCl in the 100 mM NaCl buffer was replaced isoosmotically by LiCl, KCl, RbCl, and CsCl and the currents induced by addition of 25 mM ␣MDG was measured. In the presence of 100 mM KCl, RbCl or CsCl, no detectable inward currents were generated by 25 mM ␣MDG, indicating that they cannot support sugar transport by rat SGLT1. Li ϩ was found to be able to support sugar transport (Fig. 4). There was an inward current upon substitution of LiCl for choline. Similar to the Na ϩ leak, this current was also blocked by 50 M phlorizin (not shown), and indicates that there is a leak of Li ϩ by rat SGLT1 in the absence of sugar. This Li ϩ leak was Ϸ50% of the Na ϩ leak. The current carried by Li ϩ in 25 mM ␣MDG was about 25% (Ϫ25 versus Ϫ130 nA, V t ϭ Ϫ50 mV) of the current carried by Na ϩ , suggesting lower affinity for sugar in LiCl, as detected for rabbit SGLT1 (16). At V t ϭ Ϫ150 mV, the ␣MDG-induced currents were Ϫ150 nA in 100 mM Li ϩ and Ϫ250 nA in 100 mM Na ϩ .
Phlorizin Sensitivity-Phlorizin is a high affinity competitive inhibitor of Na ϩ -dependent glucose transport in renal and intestinal epithelia (K 0.5 Pz Ϸ 10 M, 17). Fig. 5A shows that addition of 5 M phlorizin into the bath medium inhibited the currents induced by a saturating concentration of ␣MDG (5 mM) by 85%. The inhibition was complete at 10 M phlorizin (data not shown). Fig. 5B shows the voltage dependence of the K 0.5 for the inhibition by phlorizin (K 0.5 Pz ) studied at external ␣MDG concentrations of 5 mM and 0.25 mM. K 0.5 Pz was 0.9 M at 5 mM sugar and decreased to 0.09 M when the ␣MDG concentration was 250 M.
To determine the inhibitor constant K i for phlorizin inhibition (K i Pz ) we performed a series of Dixon plots. We plotted the reciprocal of the currents (1/I) against the phlorizin concentration. The lines in Fig. 5C were obtained by linear regression on phlorizin inhibition of the steady-state currents (V t ϭ Ϫ150 mV) generated by 1 mM and 0.4 mM ␣MDG. The lines intersect at a phlorizin concentration of Ϫ0.012 M. Thus the inhibitory constant K i Pz is 0.012 M (at Ϫ150 mV). It remained slightly voltage dependent in the range Ϫ150 mV to Ϫ50 mV, increasing to 0.053 Ϯ 0.003 M and 0.030 Ϯ 0.010 M at Ϫ70 mV and Ϫ50 mV. The errors are S.E. from three oocytes. FIG. 4. Cation specificity of the sugar transport. Continuous current record showing the effects of Na ϩ and Li ϩ in ␣MDG on the inward currents mediated by the same rat SGLT1 cRNA-injected oocyte, described in Fig. 1. V h was Ϫ50 mV and the oocyte was first perfused with choline (base line is indicated as a dashed line), followed by a 1-min equilibration in 100 mM solution of the appropriate cation (large arrows). At the time indicated by the small arrows, 25 mM ␣MDG was added. The substrate was continuously washed out until currents returned to the base line. The record for V t ϭ Ϫ50 mV is shown. The sugar-induced currents in the presence of lithium and sodium were Ϫ25 nA and Ϫ130 nA, respectively, and the leak currents were Ϫ8 and Ϫ20 nA. Presteady-state Charge Movements-In absence of sugar the Na ϩ /glucose cotransporter exhibits a presteady-state current after step changes in membrane voltage. The presteady-state current records from rat SGLT1 in Fig. 6A were obtained when the membrane voltage was stepped from the holding (Ϫ100 mV) to test voltages 50, Ϫ10, Ϫ50, and Ϫ150 mV. The pre-steady-state currents were completely blocked by 10 M phlorizin (data not shown) and were not observed in noninjected oocytes.
The dependence of the relaxation time constant of the ON transients () on test voltage V t is presented in Fig. 6B. decreased monotonically from 13.5 Ϯ 2 ms at Ϫ50 mV to 2.6 Ϯ 0.1 ms at 50 mV. In the OFF response, was independent of the test voltage V t and was 53 Ϯ 2 ms over the voltage range Ϫ50 to 50 mV. Error bars are S.E. from three oocytes. of the oocyte capacitive current ( 1 ) was independent of the membrane potential (0.6 Ϫ 0.8 ms). Fig. 6C shows the dependence of the total charge (Q, integral of the current transients) on membrane voltage. The curve was  6. Characterization of the presteady-state currents in the absence of sugar. A, presteady-state current records. The presteadystate current records were obtained by subtracting the capacitive (I 1 e Ϫt/1 ) and the steady-state currents (I ss ) from the total current as described in B, and in Loo et al. (8). V h was Ϫ100 mV. The traces at V t ϭ 50, Ϫ10, Ϫ50, and Ϫ150 mV are presented beginning 1 ms after applying the pulse. The inset shows the pulse protocol. B, kinetics of the presteady-state current relaxation. The time constants of relaxation for the ON and OFF current transients () for each tested membrane potential (V t ) were obtained by fitting the measured current (I) to the equation: I ϭ I 1 e Ϫt/1 ϩ I 2 e Ϫt/2 ϩ I ss , where I 1 is the oocyte capacitive current with time constant 1 , I 2 is the rat SGLT1 transient current with time constant 2 , before decaying to the steady-state currents (I ss ). V h was Ϫ100 mV. C, charge-voltage relationship of the current transients. The integrals of the ON and OFF transient currents (Q) due to rat SGLT1 were plotted as a function of the applied test voltage V t . The smooth curve was obtained by fitting the average (f) of these charges to the Boltzmann equation: is the maximal charge transfer, Q dep and Q hyp are the charge movements at the depolarizing and hyperpolarizing limits, F is the Faraday's constant, R is the gas constant, T is the absolute temperature, V 0.5 is the potential for 50% Q max , and z is the apparent valence of the movable charge. Shown are the data from a single oocyte (V h ϭ Ϫ50 mV) with parameters: z ϭ 0.85; V 0.5 ϭ Ϫ46 mV, and Q max ϭ 11 nanocoulombs. drawn according to the Boltzmann relation (see legend to Fig.  6C) to estimate the parameters Q max (maximal charge transferred), z (apparent valence of the movable charge), and V 0.5 (voltage for 50% Q max ). The maximal charge transferred is: Q max ϭ Q dep Ϫ Q hyp , where Q dep and Q hyp are the charges transferred at the tested voltage limits. Since Q max depends on the level of expression, to compare oocytes with differing levels of expression, the data were normalized between 0 and 1 using the relation: (Q Ϫ Q hyp )/Q max . Mean values from five different oocytes were V 0.5 ϭ Ϫ43 Ϯ 3 mV, z ϭ 1.0 Ϯ 0.15, and Q max ϭ 9.0 Ϯ 2.5 nanocoulombs. DISCUSSION The archetypical member of the Na ϩ -dependent family of transport proteins is SGLT. SGLTs have been cloned from rabbit, rat, human, and pig. This family also includes the transporters for myo-inositol and nucleosides (18). In this study, we characterized the kinetics of the Na ϩ /glucose cotransporter cloned from rat kidney. Our goal is to understand the structure-function relations of Na ϩ -dependent glucose transport by comparing and contrasting the kinetics of highly homologous proteins of this gene family.
Steady-state Parameters-The estimated apparent affinities for Na ϩ and glucose (K 0.5 Na and K 0.5 ␣MDG ) for the rat SGLT1 show moderate differences compared to those of rabbit and human transporters for these ligands. Fig. 2B shows that in the voltage range more positive than Ϫ50 mV there is a steeper voltage dependence for the binding of the Na ϩ -ions (40 mM at Ϫ30 mV) to the rat transporter. This accounts for higher voltage sensitivity of the apparent K 0.5 for ␣MDG in the same voltage range, shown in Fig. 1A.
Lee et al. (4) found that the K 0.5 ␣MDG was 397 M at Ϫ60 mV and our value of 300 M is in agreement with the value reported. In this study, the steady-state current induced by ␣MDG was three times higher than the study of Lee et al. (4), and we were able to obtain the voltage dependence of the K 0.5 for sugar and sodium. The stoichiometry from the Hill analysis was 2 Na ϩ :1 sugar molecule, and is similar to that of the rabbit and human (9, 10).
Lee et al. (4) also observed a K 0.5 Pz two orders of magnitude less than the value of 10 M for rabbit SGLT1 (9). The K 0.5 Pz based on inhibition of 50 M [ 14 C]␣MDG uptake was 0.17 M. Our estimate of the real inhibitory constant K i Pz was 0.012-0.03 M. It was recently observed (19), that the Na ϩ -dependent glucose transport system in sheep tracheal epithelium also has a high affinity for phlorizin (K i Pz ϳ 0.020 M). The species differences in the affinity to phlorizin observed here are almost certainly due to differences in the amino acid sequence. For the rat SGLT1, the estimated real K i Pz can also be regarded as binding/dissociation constant and used in future determination of the number of phlorizin molecules binding per cotransporter molecule. The rat, rabbit, and human clones all exhibit a phlorizin sensitive Na ϩ leak current, which is about 15-20% of the maximal ␣MDG-induced current.
Presteady-state Parameters-In the rat SGLT1, as in the rabbit (11,20) and human (8) Na ϩ -dependent glucose transport systems, we observed transient charge movements in sugar free solutions which were completed within 50 ms. These currents were abolished in all three clones by either addition of sugar substrates or the competitive inhibitor phlorizin. The estimated Q-V curves and the resulting parameters (z, Q max , and V 0.5 ) give the functional relation between the moved charge and the membrane potential. z is the average number of net elementary charges (q) apparently moved through a distance (␦) across the membrane (field) in each transporter molecule. This means, that either a single elementary charge moves completely through the membrane electric field, or two elementary charges each move 50% of this distance, or any other combination of negative or positive charges move such that ⌺q i ␦ i ϭ 1. The voltage dependence of the charge movement is shown in Fig. 6C, where Q is plotted as a function of the test potential for the rat SGLT1, and is compared for all three cotransporters in Fig. 7A. There is a displacement of about 40 mV to more negative potentials of the Q-V curves for the rat/human transporters in comparison to the rabbit. Table I shows a comparison of the parameters (z, Q max , and V 0.5 ) from fitting the Q-V relations to the Boltzmann equation. The apparent valence of the movable charge is the same for all three transporters (z ϳ 1). V 0.5 , the voltage at 50% Q max , was similar (ϷϪ40 mV) for the rat and the human clones, whereas the rabbit clone was about 40 mV more positive. Fig. 7B presents the time constants () of the presteady-state current relaxation of the three clones in the membrane voltage range Ϫ50 to 50 mV. for the ON currents of human or rat SGLT1 transients decrease progressively as the test voltage was made more positive. In contrast, for the rabbit SGLT1 transients increased, reached a maximum at 10 mV, and decreases with more depolarizing potentials. Compared to the -V curve of the rabbit, the curves of the human/rat transporters are shifted Ϸ50 mV to more negative potentials.
The maximal charge Q max depends on the level of expression of SGLT1 in the membrane since Q max ϭ qzC T , where C T is the total number of transporters. The maximal steady-state inward Na ϩ current induced by saturating sugar concentrations (I max ) is proportional to Q max (8,21). I max ϭ kqzC T , where k is the apparent turnover number of the transporter. k for rat SGLT1 was 30 s Ϫ1 and comparable to that of the human and rabbit (Table I).
Kinetic Model-The mechanism of Na ϩ -dependent sugar transport via rat SGLT1 can be explained by a six-state ordered nonrapid equilibrium kinetic model with mirror symmetry sim- FIG. 7. Comparison of the presteady-state current due to rat, rabbit and human SGLT1 transporters. Data for rabbit SGLT1 (f) are based on the estimates shown in Fig. 3, A and B, of Panayotova-Heiermann et al. (11), data for human SGLT1 (å) are taken from Fig.  3B of Loo et al. (8). Data for rat SGLT1 (E) were obtained from a single oocyte as described in Fig. 6. ilar to the proposed models for the rabbit (20) and human (8) transporters. For reviews see Wright et al. (21,22 Computer simulations resulted in a set of rate constants which account quantitatively and qualitatively for the observed presteady-and steady-state kinetics. The results suggest that differences in the kinetics between the rabbit and rat/human cotransporters are due to differences in k 12 and k 16 (Table II). Such changes in the rate constants must be due to differences in structure between the isoforms. Aligning the primary amino acid sequences show that there are different residues at 129 out of 665 positions, and, when conservative substitutions were taken into account (K ϭ R; S ϭ T; D ϭ E; Y ϭ F ϭ W; and I ϭ FIG. 8. Sequence alignment of the human, rat, and rabbit SGLT1 cotransporters. The full sequence is given for the human SGLT1, for the rat and rabbit cotransporters only the nonconserved residues are included. Identical and residues similar to that in human (D ϭ E, R ϭ K, S ϭ T, I ϭ V ϭ L, and Y ϭ F ϭ W) are replaced by dashes (-). The rabbit residues that are significantly different from those in rat and human, i.e. are polar in either rabbit or rat and human, are shown in bold italics. The location of the putative transmembrane domains is indicated by the lower case letters and underlined ( ****** ). The N-and C-terminals in this secondary structure model (7) are placed on the cytoplasmic side of the plasma membrane. V ϭ L ϭ M) this reduces to differences at 76 positions. These 76 are evenly distributed between the N-and C-terminal halves of the protein (Fig. 8), and are mostly confined to hydrophobic loops between the putative transmembrane helices. The cytoplasmic hydrophilic N-terminal and the external loops between helices 5/6 and 11/12 contain 43 of the 76 nonconserved residues. A clue about the residues that may be important in determining kinetic differences comes from consideration of the residues that are identical in pairs of the three transporters. There are 25 residues shared between human and rat, 17 shared between human and rabbit, and 21 between rabbit and rat. Overall, there are 37 residues that are different between the rabbit and the rat and human, and 25 of these are polar (indicated in bold type in Fig. 8). At 10 positions the residues are charged in either the rabbit or in the human and rat, at 7 positions the residues are serines or threonines in either rabbit or human and rat, and at 5 other positions the residues are polar in either rabbit or human and rat. The polar residues are mostly in hydrophilic segments, and half are clustered in external loops between helices 5/6 and 11/12. There are no significant differences between the residues in putative transmembrane domains of the cotransporters (in transmembrane helix 4 the substitution 176D3 N is of no functional significance) (11). This suggest that the polar residues in hydrophilic domains of the protein play an important role in determining differences in kinetics between species, k 12 and k 16 , by determining the three-dimensional protein structure through electrostatic interactions. This could be tested by examining functional properties of clones after either mutating the polar residues or swapping hydrophilic loops between species.
Our conclusion is that the kinetic differences between the human, rat, and rabbit transporters primarily are due to two partial reactions involving binding/dissociation of Na ϩ ions and translocation of the empty carrier. These differences are probably due to polar residues clustered between helices 5/6 and 11/12.