Comprehensive chimeric analysis of amino acid residues critical for high affinity glucose transport by Hxt2 of Saccharomyces cerevisiae.

Chimeras of Hxt2 and Hxt1, high affinity and low affinity glucose transporters, respectively, of Saccharomyces cerevisiae, were previously constructed by random replacement of each of the 12 transmembrane segments (TMs) of Hxt2 with the corresponding region of Hxt1. Characterization of these chimeras revealed that at least TMs 1, 5, 7, and 8 of Hxt2 are required for high affinity transport activity. To determine which amino acid residues in these TMs are important for high affinity glucose transport, we systematically shuffled all of the 20 residues in these regions that differ between Hxt2 and Hxt1. Analysis of 60 independent mutant strains identified as expressing high affinity and high capacity glucose transport activity by selection on glucose-limited agar plates revealed that Leu-201 in TM5 of Hxt2 is most important for such activity and that either Cys-195 or Phe-198 is also required for maximal activity.

Hexose transport across the plasma membrane is a necessary step in the utilization of monosaccharides by living cells. The yeast Saccharomyces cerevisiae is able to take up glucose over a wide range of extracellular concentrations with the use of abundant hexose transporters (Hxt1 to -17, Gal2) (1,2). These transporters belong to the major facilitator superfamily (MFS) 1 and contain 12 transmembrane segments (TMs) (3). Hxt2 is a major high affinity facilitative glucose transporter, whereas Hxt1 is a low affinity facilitative glucose transporter (2,4). The numbers of amino acids in each putative TM and inter-TM loop of Hxt2 are identical to those in the corresponding regions of Hxt1, and the two proteins share ϳ70% amino acid identity in these regions.
We previously investigated which TMs of Hxt2 are important for high affinity glucose transport by the new approach of TM shuffling (4). We randomly replaced, at the DNA level, each of the 12 TMs of Hxt2 with the corresponding segments of Hxt1. Clones encoding transporters with high affinity for glucose were selected by plating transformants on carbon source-limited agar plates. Our results demonstrated that a minimal combination of TMs 1, 5, 7, and 8 of Hxt2 is necessary for high affinity glucose transport; the chimeric transporter C1578, in which all TMs but 1, 5, 7, and 8 of Hxt2 were replaced with the corresponding TMs of Hxt1, thus exhibited high affinity and high capacity glucose transport activity similar to that of Hxt2. Among these four TMs, TM5 seems to be the most important, because all of the chimeras that exhibited high affinity transport possessed this segment of Hxt2 (4).
We have now investigated which of the amino acid residues in TMs 1, 5, 7, and 8 of Hxt2 are important for high affinity, high capacity glucose transport. Our results indicate that Leu-201 in TM5 of Hxt2 is most critical for such activity, with either Cys-195 or Phe-198 also being necessary to support maximal activity.

EXPERIMENTAL PROCEDURES
Construction of Vectors-Construction of the plasmid Hxt2mnx-pVT, which comprises HXT2 under the control of the ADH1 promoter in the multicopy vector pVT102-U (YEp URA3 bla), was previously described (4). In brief, HXT2 was modified by creating MroI, NheI, XhoI, and ClaI sites in the nucleotide sequences corresponding to the NH 2 -terminal end of TM4, the loop between TM6 and TM7, the loop between TM9 and TM10, as well as immediately downstream of the termination codon, respectively. The expression vector C1578-pVT, which encodes the chimeric transporter C1578 (in which all of the TMs of Hxt2, with the exception of TMs 1, 5, 7, and 8, have been replaced with those of Hxt1), was also described previously (4).
Mutagenesis-Site-directed mutants were prepared with a PCRbased approach. Because a single mutagenic primer was not sufficiently long to cover all the mutagenic sites in each TM, we performed PCR in two steps with a GeneAmp PCR system 2400 (Applied Biosystems). In the first step, performed with ExTaq polymerase (Takara, Otsu, Japan), two DNA fragments that were designed to possess 7-10 overlapping nucleotides were prepared for mutants of each TM. These fragments encoded all the possible combinations of amino acid residues at the sites that differ between Hxt1 and Hxt2. Mutagenic primers were constructed by replacing each target codon with degenerate sequences. For example, for TM5 mutants, we used TKC for site 198, KTA for site 201, and TWT for site 215 (see Fig. 1); for site 195, we used two primers, one for expressing the Hxt2-derived amino acid, Cys (TGT), and the other for expressing the Hxt1-derived amino acid, Leu (TTG), because degenerate sequences at this site would result in the incorporation of unnatural amino acids. In the second PCR step, performed with native Pfu polymerase (Stratagene), the two fragments produced by the first PCR were joined together without a template, as described previously (4). Each final product was used to replace the corresponding region of Hxt2 in C1578-pVT with the use of two restriction enzymes: EcoRI and MroI for TM1 mutants, MroI and NheI for TM5 mutants, and NheI and XhoI for TM7 or TM8 mutants. After amplification in Escherichia coli, plasmids were introduced into S. cerevisiae strain KY73 (MAT␣ hxt1⌬::HIS3::⌬hxt4 hxt5::LEU2 hxt2⌬::HIS3 hxt3⌬::LEU2::⌬hxt6 hxt7⌬::HIS3 gal2⌬::DR ura3-52 MAL2 SUC2 MEL) (5).
Plate Selection-Transformants that possessed high affinity, high capacity glucose transport activity were selected after incubation of yeast cells for 3 or 4 days at 30°C on glucose-limited (glucose, 1 mg/ml) agar plates containing a synthetic medium supplemented with adenine and amino acids but not with uracil (S0.1D plates) (4). KY73 cells are not able to grow on S2D plates (glucose, 20 mg/ml), in which glucose is the only carbon source. In parallel, the number of transformants was counted on S2Mal plates (maltose, 20 mg/ml). Modified portions of all clones selected in this study were verified by DNA sequencing with an automated sequencer (model 310, Applied Biosystems).
Transport Assay-Cells harboring plasmids were grown to log phase (optical density at 650 nm, 0.3-0.4) at 30°C in S2Mal synthetic liquid medium. Glucose transport by the cells was measured at 30°C for 5 s as described (6,7). Transport activities measured at a D-[ 14 C]glucose concentration of 0.1 mM were expressed as picomoles of glucose per 1 ϫ 10 7 cells per 5 s and were corrected for the background activity determined either in the presence of 0.5 mM HgCl 2 or with 0.1 mM L-[ 14 C]glucose. In some experiments, transport activity was calculated as a percentage of that obtained with cells expressing C1578.
Construction of a Three-dimensional Model of Hxt2-The crystal structure of LacY (pdb code, 1PV7) was used as the basis for construction of a structural model of Hxt2. Hydrophilic residues in the putative transmembrane helices of Hxt2 were aligned with those of LacY, and the structural model of Hxt2 was generated with the Biopolymer module of Insight II (version 2000; Accelrys, San Diego, CA). The initial model structure was energy-minimized with Discover 3 (version 2000; Accelrys) by fixing the C␣ atoms of the model until the final root mean square deviation became Ͻ0.1 kcal mol Ϫ1 Å Ϫ1 . The optimized complex structure was selected from 100 energy-minimized structures sampled by the molecular dynamics calculations, which were performed with a temperature of 300 K, a cutoff distance of 10 Å, a distance-dependent dielectric constant, and a time step of 1 fs for 100 ps by sampling the conformation every 1 ps with CVFF (Consistent Valence Forcefield) parameters in Discover 3.
Other Assays-A crude membrane fraction was prepared from cells as described (8), and immunoblot analysis of this fraction was performed with rabbit polyclonal antibodies specific for the COOH-terminal region of Hxt2 (6) and with 125 I-labeled protein A (IM144, Amersham Biosciences). The intensity of bands corresponding to immune complexes was measured with imaging plates (BAS1800II, Fuji Film) (8), within the range proportional to the amount of protein. Cell number was determined with a particle counter (Z2, Coulter). Protein concentration was measured with bicinchoninic acid (Pierce). TMs 1,5,7, and 8 of Hxt2 Necessary for High Affinity, High Capacity Glucose Transport-In a previous comprehensive study, we showed that a minimal combination of TMs 1, 5, 7, and 8 of Hxt2 is necessary for high affinity, high capacity glucose transport (4). To determine which amino acid residues in these TMs are responsible for such activity, we shuffled all of the 20 residues in these regions that differ between Hxt2 and Hxt1 (Fig. 1). TM1 contains six such residues, and we generated 5,700 transformants expressing C1578-based proteins corresponding to all possible combinations (2 6 ϭ 64) of the Hxt1 and Hxt2 residues at these six sites. Selection on S0.1D plates yielded 120 colonies, of which 63 were subjected to DNA sequencing. Twenty-eight independent transformants manifested high affinity, high capacity glu-cose transport activity (Table I). TMs 5 and 7 possess only four and three residues, respectively, that differ between Hxt2 and Hxt 1; all sixteen possible combinations of these residues for TM5 mutants and eight combinations for TM7 mutants were generated and subjected to plate assays. TM8 contains seven residues that differ between Hxt2 and Hxt1; random saturation mutagenesis of these residues yielded 22,000 transformants and 111 colonies were picked up on S0.1D plates. Subsequent DNA sequencing of 54 clones identified 24 independent mutants that exhibited high affinity, high capacity glucose transport activity. Four residues of Hxt2, Leu-61, Leu-201, Leu-357, and Phe-366, were present in Ͼ80% of the transporters selected by plate assay (Table I), with all high affinity, high capacity transporters possessing Leu-201 of TM5. As a first step in the characterization of amino acid residues responsible for high affinity, high capacity glucose transport, we focused on residues in TM5.

Amino Acid Residues in
Characterization of All 16 Shuffled Mutants of TM5-The amino acid sequence of TM5 of Hxt2 differs at only four positions (sites 195, 198, 201, and 215) from that of the corresponding domain of Hxt1 (Fig. 1). We therefore constructed C1578based proteins corresponding to all 16 possible combinations of the Hxt1 and Hxt2 residues at these four sites. CFLY and lcvf denote the chimeric transporters that contain the Hxt2-derived or Hxt1-derived amino acids at all four of these sites, respectively. The positive clones selected by growth of transformants on carbon source-limited (S0.1D) plates encoded the chimeric transporters CFLY, CFLf, lFLY, lFLf, CcLY, and CcLf (Table  II), all of which contain Leu-201. In addition to Leu-201, either Cys-195 or Phe-198 was required for growth on S0.1D plates.
We prepared a crude membrane fraction from cells expressing each of these 16 chimeric transporters and examined the extent of transporter expression by immunoblot analysis with antibodies to the COOH-terminal region of Hxt2 (Fig. 2). All 16 chimeras yielded a predominant immunoreactive band at a position (47 kDa) corresponding to that of wild-type Hxt2. Quantitative analysis of the bands corresponding to the 47-kDa proteins revealed an expression range of 77-148% (n ϭ 3) relative to the intensity of the band yielded by C1578.
We measured the glucose transport activities of all 16 chimeras with 0.1 mM D-glucose as substrate (Fig. 3). All eight chimeric transporters containing Leu-201 possessed substantial glucose transport activity (Ͼ50% of that of C1578), whereas the activities of the other eight transporters containing Val-201 did not exceed 30% of that of C1578, when transport activities were normalized on the basis of transporter expression as determined by quantitative immunoblot analysis. Glucose transport activities normalized by cell number did not differ significantly from those normalized by the amount of immunoreactive transporter protein.
Kinetic parameters were determined for all 16 chimeras of TM5 of C1578 (Table II). Whereas all the chimeras possessed K m values in the range of high affinity transport, the V max values varied substantially. The values for V max /K m correlated well with growth on agar plates. The 16 chimeras could be divided into three groups characterized by V max /K m values of Ͼ250, 150 -250, or Ͻ150. In the first group, all chimeras contained Leu-201 and either Cys-195 or Phe-198 and grew on both S2D and S0.1D plates. In the third group, all chimeras contained Val-201 and grew on neither S2D nor S0.1D plates. The two chimeras in the second group, which contained Leu-201 and Hxt1-derived amino acids at sites 195 and 198, grew only on S2D.
To examine the role of Leu-201 in Hxt2, we constructed Hxt2(CFvY), in which Leu-201 of Hxt2 was replaced with Val. No significant difference between CFvY and Hxt2(CFvY) was The depiction of each TM is based on a previously described model (15). Filled circles indicate amino acid residues common to both Hxt2 and Hxt1, with the numbers within circles indicating residue positions for Hxt2. found in plate assays and kinetic parameters of glucose transport (Table II). DISCUSSION With the use of newly developed TM shuffling, we previously found that at least four TMs of Hxt2 (TMs 1, 5, 7, and 8) are required for its high affinity glucose transport activity (4). In the present study, we also adopted a comprehensive approach (systematic residue shuffling) to evaluate the contributions of each of the 20 amino acid residues in TMs 1, 5, 7, and 8 of the high affinity glucose transporter Hxt2 that differ from those of the low affinity glucose transporter Hxt1. All possible combinations of residues in each TM (2 6 for TM1, 2 4 for TM5, 2 3 for TM7, and 2 7 for TM8) were generated by saturation mutagenesis (TMs 1 and 8) or specific construction (TMs 5 and 7). Of these 20 residues, Leu-201 of TM5 appeared to be the most important for high affinity, high capacity glucose transport activity, because all such transporters generated in the present study possessed this residue. The remaining 19 Hxt2 residues in TMs 1, 5, 7, and 8 were not absolutely required for high  TMs 1,5,7, and 8 of chimeric transporters exhibiting high affinity, high capacity glucose transport activity Amino acid residues at each site of TMs 1, 5, 7, and 8 that differs between Hxt2 and Hxt1 were shuffled. Transformants that exhibited high affinity, high capacity glucose transport were selected on S0.1D plates and subjected to DNA sequencing. Independent clones (28 for TM1, five for TM5, three for TM7, and 24 for TM8) were tabulated, with Hxt2-derived residues in uppercase within shaded boxes and Hxt1-derived residues in lowercase within open boxes. The percentage of Hxt2-derived amino acids at each residue position is shown in the bottom row. affinity, high capacity glucose transport, with the relative preference for an Hxt2 residue differing at each site. The amino acid residues of several TMs in Hxt2 thus appear to cooperate to achieve high affinity, high capacity transport activity.
We focused on TM5 to identify the residues in this region important for such activity. By replacing each of the four residues of Hxt2 that differ from those of Hxt1 with the corresponding Hxt1 residue, we found that Leu-201 was important for high affinity glucose transport. In addition to Leu at this site, all chimeric transporters selected by growth of transformants on S0.1D plates possessed the Hxt2 residue at either site 195 (Cys) or site 198 (Phe). In contrast, site 215 does not seem to be important for transport activity. The three amino acid residues Cys-195, Phe-198, and Leu-201 are clustered on one face of TM5 of Hxt2 in a helix-packing model (Figs. 1 and 4), suggesting that they may contribute cooperatively to transporter function. Without a comprehensive approach, such as the one adopted in the present study, it would not be possible to detect cooperative effects of more than two residues.
We further examined whether Leu-201 was replaceable by other amino acids by changing Leu-201 in C1578 to each of the other 19 amino acids. Selection on agar plates revealed that none of the cells expressing proteins containing the other 19 residues at site 201 grew on S0.1D plate (data not shown), FIG. 2. Expression of the 16 chimeras corresponding to all possible combinations of Hxt1-or Hxt2-derived amino acids in TM5 of C1578. KY73 cells harboring plasmids encoding each chimera described in Fig. 3 were cultured to log phase at 30°C in S2Mal synthetic medium, after which a crude membrane fraction was prepared. A portion of each fraction (10 g of protein) was subjected to immunoblot analysis with antibodies to Hxt2. FIG. 3. Glucose transport activities of the 16 chimeras corresponding to all possible combinations of Hxt1-or Hxt2-derived amino acid residues in TM5 of C1578. Each residue at sites 195, 198, 201, and 215 of C1578 was replaced with the corresponding amino acid of Hxt1, yielding 16 chimeras (left panel); Hxt2-or Hxt1-derived amino acids are indicated in uppercase and lowercase, respectively. KY73 cells expressing these various proteins were grown to log phase at 30°C in S2Mal synthetic medium, after which glucose transport activity was measured for 5 s at 30°C with 0.1 mM D-glucose as substrate (right panel); transport activities were normalized by cell number (solid bars) or by the mean level of protein expression as determined by quantitative immunoblot analysis (open bars) and are means Ϯ S.E. of values from three or more experiments.

TABLE II
Growth on glucose-limited plates and kinetic parameters of glucose transport for the 16 chimeras of TM5 KY73 cells expressing the 16 chimeras of TM5 of C1578, Hxt2, Hxt2(CFvY), or Hxt1 were subjected to plate assays with S2D and S0.1D medium. Cell growth or no growth after incubation for 3-4 days at 30°C is indicated by (ϩ) or (Ϫ), respectively. The cells were grown to log phase at 30°C in S2Mal synthetic liquid medium, after which glucose transport activity was measured for 5 s at 30°C. The K m and V max values (means Ϯ S.E., n Ն 3) were determined with 1 to 100 mM D-glucose as substrate.
where is transport rate at substrate concentration S, K m is the Michaelis constant, and V max is the maximal velocity. The V max /K m value is a good measure of transport efficiency at low substrate concentrations (S Ͻ Ͻ K m ). A similar term, k cat /K M , has been used as a measure of catalytic efficiency (16). For the selection condition of growth on S0.1D plates, which contain a glucose concentration of 0.1% (5.5 mM), the glucose concentration surrounding the colonies on the plates would be expected to be substantially lower than the K m .
indicating that Leu at this position is essential and irreplaceable for high affinity glucose transport with a capacity to sustain cell growth under glucose-limited conditions. The reason for this requirement for Leu at position 201 is not immediately clear, but the size of its aliphatic side chain may be important to maintain the protein conformation necessary for high affinity, high capacity transport activity. We previously showed that one specific amino acid residue of Hxt2, Phe-431 in TM10, is absolutely required for recognition of the difference between glucose and galactose, with no other residue being able to perform this task (9). In addition, Tyr-440 in TM10 is necessary for high capacity glucose transport (9). On the other hand, the recognition and transport of glucose with high affinity appear to require a precise coordination of several amino acid residues in the various TMs that form the permeation pathway, because four TMs of Hxt2 are required for high affinity transport. The critical role of Leu-201 in C1578 was also observed in Hxt2, because Hxt2(CFvY) showed almost the same characteristics as CFvY of C1578. In this respect, it is of considerable interest that Leu in this position is conserved in all the yeast MFS sugar transporters except for Hxt1 (1).
The crystal structures of the bacterial MFS transporters LacY (10) and GlpT (11), both recently determined with a resolution of Ͻ4 Å, reveal that the configurations of the TMs in the two proteins are highly similar. TMs 1, 5, 7, and 8 form the central permeation pathway. The aromatic ring of Trp-151 in TM5 of LacY is implicated in interaction with a substrate analog, galactopyranosylthiogalactopyranoside. The importance of TM5 was also demonstrated in the Tet(C/B) tetracycline transporter; mutation of Gly-452 in this segment was found important for tetracycline resistance (12). A structural model for OxlT, an MFS oxalate transporter of Oxalobacter formigenes, based on electron crystallographic analysis placed TM5 along the substrate pathway (13). TM5 of the mammalian glucose transporter GLUT1 is an amphipathic helix that is thought to form part of the sugar permeation pathway (14).
Although the alignment of bacterial and eukaryotic MFS transporters is not without ambiguity (15), we constructed a working structural model of Hxt2 based on the backbone structure of LacY revealed by crystallography. The positions of Leu-201, Phe-198, and Cys-195 of TM5 as well as of Phe-431 and Tyr-440 of TM10 in this model are shown in Fig. 4. All five residues indicated are situated in the cytoplasmic half of the inner core of the protein, and their arrangement suggests that they all likely interact directly with the substrate. Together with the results of previous studies, our present observations indicate that TM5 of yeast MFS transporters is important for the recognition and passage of substrate. Our present results demonstrate that Leu-201 in TM5 plays a central role in the high affinity, high capacity glucose transport activity of Hxt2.