Phlorizin recognition in a C-terminal fragment of SGLT1 studied by tryptophan scanning and affinity labeling.

SGLT1 as a sodium/glucose cotransporter is strongly inhibited by phlorizin, a phloretin 2'-glucoside that has strong interactions with the C-terminal loop 13. We have examined phlorizin recognition by the protein by site-directed single Trp scanning mutagenesis experiments. Six mutants (Q581W, E591W, R601W, D611W, E621W, and L630W) of truncated loop 13 (amino acids 564-638) were expressed in Escherichia coli and purified to homogeneity. Changes in Trp quenching and positions of the emission maxima were determined after addition of phlorizin. D611W displayed the largest quenching of 80%, followed by R601W (67%). It also exhibited the maximum red shift in Trp fluorescence ( approximately 14 nm), indicating an exposure of this region to a more hydrophilic environment. Titration experiments performed for each mutant showed a similar affinity for all mutants, except for D611W, which exhibited a significantly lower affinity (Kd approximately 54 microm). Also the maximum change in the collisional quenching constant by acrylamide was noted for D611W (KSV = 11 m-1 in the absence of phlorizin and 55 m-1 in its presence). Similar results were obtained with phloretin. CD measurements and computer modeling revealed that D611W is positioned in a random coil situated between two alpha-helical segments. By combining gel electrophoresis, enzymatic fragmentation, and matrix-assisted laser desorption ionization mass spectrometry, we also analyzed truncated loop 13 photolabeled with 3-azidophlorizin. The attachment site of the ortho-position of aromatic ring B of phlorizin was localized to Arg-602. Taken together, these data indicate that phlorizin binding elicits changes in conformation leading to a less ordered state of loop 13. Modeling suggests an interaction of the 4- and 6-OH groups of aromatic ring A of phlorizin with the region between amino acids 606 and 611 and an interaction of ring B at or around amino acid 602. Phloretin seems to interact with the same region of the protein.

In mammals, transepithelial transport of D-glucose is mediated by SGLT1 (sodium/glucose cotransporter-1), which can be found in the brush-border membranes of the small intestine and kidney. The transporter facilitates the effective uptake of glucose into cells driven by the electrochemical potential difference of sodium. Coupling and translocation are supposed to be accompanied by conformational changes in the protein. Such changes could be induced, for example, by Na ϩ , which increases the affinity of the cotransporter for sugar (1,2). Extensive mutagenesis studies revealed that the N-terminal half of the protein contains the Na ϩ -binding sites, whereas glucose binds and permeates through the C-terminal half of the cotransporter (3,4). Amino acids 162-173 apparently constitute part of an external Na ϩ pore in the SGLT1 protein, whereas sugar binding is controlled by Gln-457 and Thr-460 (5,6).
The cotransport system is inhibited by glucosides with either aromatic or aliphatic aglucon residues. Phlorizin, a ␤-glucoside of the aromatic compound phloretin, is the most potent competitive inhibitor, with an apparent K i of 1 M (7). It is proposed that the phlorizin binding to SGLT1 is a two-step process: rapid formation of an initial collision complex, followed by a slow isomerization process that occludes phlorizin within its receptor site (8). Phlorizin is thereby supposed to bind to both the sugar-binding site and the aglucon-binding site, the latter with a hydrophobic/aromatic surface (9,10).
Site-directed mutagenesis studies suggest that a hydrophobic region located in the C-terminal loop 13 (amino acids 604 -610) is critically involved in the binding of phlorizin (11). The binding of phlorizin to loop 13 could be confirmed in solution by monitoring phlorizin-dependent fluorescence quenching of the endogenous Trp-561. It has been demonstrated further that the phloretin (but not the glucose) moiety of phlorizin interacts with loop 13 (12).
The strong sensitivity of Trp fluorescence intensity to protein microenvironment is routinely exploited to follow a variety of protein conformational changes, e.g. ligand/substrate binding, folding/unfolding, etc. (13). We used this signal to characterize the structure-function relationship of loop 13, with the refinement to place Trp residues at different positions of the molecule as reporter groups. Wild-type truncated loop 13 and six mutants were expressed in Escherichia coli and purified, and the activity and signal produced by each single Trp peptide were monitored. Changes in Trp fluorescence and the different accessibilities of each Trp mutant to acrylamide in the presence and absence of phlorizin were thereby used to probe the sequence-specific interactions. The results show that phlorizin binding affects the various mutants differently. The largest changes were found for D611W. A qualitatively similar effect on D611W was observed with phloretin.
To identify unequivocally the interaction sites of phlorizin during its binding to SGLT1, we also performed photoaffinity labeling of truncated loop 13 with a photoreactive 3-azido analog of phlorizin. Taken together, the data suggest that an insertion of phlorizin between two ␣-helical segments of loop 13 occurs together with strong interactions between the OH groups of aromatic ring A of phlorizin and the area between amino acids 601 and 611. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.  Taq DNA polymerase, T4 DNA ligase, restriction enzymes, pCR2.1-TOPO cloning vector, dNTPs, and sequencing grade  modified trypsin were obtained from Promega (Mannheim, Germany), New England Biolabs Inc. (Schwalbach, Germany), or Invitrogen (Karlsruhe, Germany). Oligonucleotides were synthesized by MWG Biotech (Ebersberg, Germany). The pGEX-4T-1 vector and BL21(DE3) cells were from Amersham Biosciences, Freiburg, Germany and Novagen (Madison, WI), respectively. Lysozyme and protease inhibitor mixture were purchased from Sigma (Taufkirchen, Germany) and Roche Applied Science, Mannheim, Germany, respectively. Thrombin was purchased from Sigma. The following chemical reagents were obtained from Gerbu Biotechnik (Gailberg, Germany), Sigma, or Roth (Karlsruhe, Germany) in the highest purity available: EDTA, Tris base, glycine, NaCl, L-Trp, phlorizin, isopropyl-␤-D-thiogalactopyranoside, 2,5-dihydroxybenzoic acid, and ␣-cyano-4-hydroxycinnamic acid.
Molecular Biology-Mutagenesis experiments were performed using a Chameleon double-stranded, site-directed mutagenesis Kit (Stratagene) according to the manufacturer's instructions. The following oligonucleotides with mutated nucleotides (underlined) were used for mutagenesis: Q581W, 5Ј-GGA GAG GAA GAC ATT TGG GAA GCT CCA GAG GAG-3Ј; E591W, 5Ј-GAG GCC ACT GAC ACA TGG GTT CCT AAG AAG AAG-3Ј; R601W, 5Ј-AAG AAA GGA TTC TTC TGG CGG GCC TAT GAC CTG-3Ј; D611W, 5Ј-CTG TTT TGT GGG CTG TGG CAG GAT AAG GGA CCC-3Ј; E621W, 5Ј-CCC AAG ATG ACC AAG TGG GAG GAG GCT GCC ATG-3Ј; and L630W, 5Ј-GCC ATG AAG CTG AAG TGG ACA GAC ACC TCC GAG-3Ј. The ApaI site (underlined) was silently introduced in a selection primer as an aid in screening with the oligonucleotide primer 5Ј-GCA AAT CGC GCT GTT AGC GGG ACC ATT AAG TTC TGT CTC GGC-3Ј. The PCR product was purified using the normal primers flanking the mutation site to produce an insert with BamHI and EcoRI restriction sites. The resulting 222-bp fragment was ligated into the pCR2.1-TOPO cloning vector and sequenced prior to mutagenesis. The confirmed cDNA was ligated into the pGEX-4T-1 expression vector. The mutant cDNAs were sequenced to verify the desired mutations at various points. The recombinant plasmids were then transformed into BL21(DE3) cells for expression of the mutant truncated C terminus of SGLT1.
Protein Expression and Purification-Wild-type and mutant proteins tagged at the N terminus with glutathione S-transferase were expressed using 1 mM isopropyl-␤-D-thiogalactopyranoside induction for 3.5 h at 37°C in a 2-liter culture. Cells were pelleted by centrifugation at 8000 ϫ g for 10 min at 4°C and resuspended in 20 ml of PBS 1 lysis buffer (140 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , and 1.8 mM KH 2 PO 4 adjusted to pH 7.3) supplemented with 5 mM dithiothreitol, 0.1% Triton X-100, 1 mg/ml lysozyme, and protease inhibitor mixture (one tablet). Cells were effectively lysed using the SONOPULS ultrasonic homogenization system under mild conditions. The lysate was centrifuged at 17,000 ϫ g for 45 min at 4°C. The supernatant containing glutathione S-transferase-tagged proteins was applied to a glutathione-Sepharose column (Amersham Biosciences) after passage through a 0.2-m filter. Soluble fractions of wild-type truncated loop 13 and all mutants (amino acids 564 -638) after removing the protease inhibitor were generated by treating the glutathione-Sepharose-immobilized fusion protein, derived from pGEX-4T-1, with 2 units/ml thrombin in PBS for 16 h at 4°C. The purity of the wild-type truncated and mutant proteins was assessed by SDS-PAGE using a precast 4 -12% Bis-Tris gel in Tris/glycine running buffer and staining with Coomassie Blue. All protein concentrations were determined by the Lowry method using bovine serum albumin as a standard.
Steady-state Trp Fluorescence-The fluorescence experiments were performed on a PerkinElmer Life Sciences LS 50B luminescence spectrometer fitted with a 450-watt xenon arc lamp at room temperature. A 0.3-cm excitation and emission path length quartz cell was used for all fluorescence measurements. The excitation wavelength was set at 295 nm for selective excitation of Trp fluorescence. A 290-nm cutoff filter was used to minimize the contribution of scattering signals. Emission spectra were collected from 300 to 400 nm, averaging six scans. The bandwidths for both excitation and emission monochromators were 5 nm. The emission spectra were corrected for background and dilution effects.
Quenching of Intrinsic Protein Fluorescence-Acrylamide-dependent Trp fluorescence quenching experiments were performed in PBS (pH 7.3) for the wild-type and mutant proteins (5 M) in the absence and presence of 50 M phlorizin or phloretin. Acrylamide was added from aliquots of an 8 M stock solution to each protein solution (25°C) at up to 500 mM. The proteins were incubated with the substrate at room temperature for 2 min prior to collecting the fluorescence spectra. The accessibility of Trp were monitored by analyzing the quenching data using a Stern-Volmer equation: , where F 0 is the fluorescence of the protein in the absence of quencher, and F is the observed fluorescence at the concentration of the quencher ([Q]). K SV is the collisional quenching constant, which was determined from the slope of best fit values of the Stern-Volmer plot at a given concentration of the quencher. The difference between inner filter effects of substrate (50 M) from the titration of L-Trp (5 M) performed at an excitation of 295 nm and an emission at 355 nm was found to be much less and hence was not considered in our measurements.
Ligand Binding Assay-The ligand-induced fluorescence quenching as a function of phlorizin or phloretin concentration was monitored as described (12). Increasing amounts of phlorizin or phloretin (1-650 M) were added to all Trp mutants (5 M) in PBS. Briefly, as increasing amounts of phlorizin or phloretin were added, the fluorescence intensity decreased. Apparent binding constants were calculated from titration data by determining the percentage of quenching at saturating phlorizin or phloretin concentrations using a single-site binding equation in the nonlinear regression analysis program Prism (GraphPAD, San Diego, CA). 1 The abbreviations used are: PBS, phosphate-buffered saline; Bis- a The apparent equilibrium dissociation constants (K d ) were determined from the nonlinear regression analysis of the percentage of fluorescence quenching as a function of phlorizin concentration using a computer-based analysis program (Prism). Values are the means Ϯ S.D. of two or three independent experiments.

FIG. 2. Titration of single Trp mutants with phlorizin.
All experiments were conducted as described under "Experimental Procedures." Samples include Q581W (black), E591W (red), R601W (green), D611W (dark blue), E621W (sky blue), and L630W (purple). The percentage of quenching at saturating phlorizin concentrations was calculated from the data, and the values were fitted to a single-site binding equation using Prism to calculate the equilibrium dissociation constant. The K d values for all mutants are summarized in Table I Table  II. The means Ϯ S.D. of two or three independent experiments are given. No difference in quenching was observed for L-Trp (5 M).

Photoaffinity Labeling of Truncated Loop 13 with 3-Azidophlorizin-
Photoaffinity labeling was performed with the photolabile phlorizin analog 3-azidophlorizin (3-AP) synthesized as reported previously (20,21). The photolabeling experiments were carried out in 200 l of PBS (pH 7.3) containing 200 g of truncated loop 13 protein and 1 mM 3-AP. The mixture was preincubated at room temperature in the dark for 5 min. After incubation, photolysis was carried out in a Rayonet RPR-100 photochemical reactor (Southern New England Ultraviolet Co., Branford, CT) fitted with 16, 2800 Å lamps at 22°C for 10 min. As a control, the photoprobe was first exposed to UV light and then added to the protein. No labeling was observed in the absence of UV light. After photoaffinity labeling, the protein was precipitated with chloroform/ methanol (2:1, v/v). A small part of the protein pellet was solubilized with 0.1% trifluoroacetic acid in 50% acetonitrile for MALDI-TOF mass spectrometry analysis, and the rest of the protein pellet was dissolved in sample buffer for SDS-PAGE. Proteins were separated by SDS-PAGE using NuPAGE 4 -12% BisTris gels (Invitrogen). Gels were stained with Coomassie Brilliant Blue G-250 (22). Bands of interest were excised from the gel with a clean razorblade, sliced into 1-mm 3 cubes, incubated overnight at room temperature with 200 l of 50 mM NH 4 HCO 3 (pH 8.5) and 50% acetonitrile, dehydrated with 200 l of acetonitrile, and dried in a centrifugal evaporator. Gel pieces were rehydrated with a small volume of digestion solution (50 mM NH 4 HCO 3 , 5 mM CaCl 2 , and 12.5 ng/l trypsin) at 4°C for 30 min. Trypsin was used at an enzyme/substrate ratio of ϳ1:100 by weight. Digestion was allowed to proceed overnight at 37°C. Peptides were extracted from the gel pieces with 100 l of 20 mM ammonium bicarbonate, followed by two extractions of 1:1 water/acetonitrile plus 1% trifluoroacetic acid and one extraction with 100 l of acetonitrile. All extracts were combined in a fresh tube, flash-frozen, and dried in a centrifugal evaporator. Dried extracts were stored at Ϫ20°C until analyzed. Mass spectra were acquired in the positive ion linear mode on a Voyager DE-PRO MALDI apparatus (PE-Biosystems, Shelton, CA). After mixing the solubilized protein with 2,5-dihydroxybenzoic acid matrix (saturated 2,5-dihydroxybenzoic acid solution in 0.1% trifluoroacetic acid in 50% acetonitrile) for intact truncated loop 13 protein and for acquisition of a mass spectrometric peptide map of the trypsin-digested protein, 0.5-l aliquots of the generated cleavage products were dispensed onto the sample support, followed by 0.5 l of ␣-cyano-4-hydroxycinnamic acid matrix solution (solution of ␣-cyano-4-hydroxycinnamic acid in 0.1% trifluoroacetic acid in 50% acetonitrile). Samples were deposited on a MALDI plate and dried at room temperature prior to collecting the spectra.
Circular Dichroism Analysis-The far-UV CD spectrum was recorded between 190 and 250 nm (350-l sample volume) on a Jasco J-715 spectropolarimeter equipped with a temperature-controlled incubator at 20°C using 1-mm optical path length quartz cells and 10 M wild-type loop 13. The step size was 0.5 nm with a 1.0-nm bandwidth at a scan speed of 50 nm/min. An average of 10 scans was obtained for blank and protein spectra, and the data were corrected for buffer contributions. Measurement was performed under nitrogen flow. The secondary structure percentages were calculated using the K2d computer modeling program. The results are expressed as mean residue ellipticity in units of degrees/cm 2 /dmol. The spectrum was recorded in 10 mM phosphate buffer at pH 6.8.

RESULTS
Protein Expression and Purification-Proteins were expressed in E. coli and purified to homogeneity with a yield of ϳ2.5 mg/liter culture. Plasmid pGEX-4T-1 containing wildtype truncated loop 13 generates a Trp-free protein; thus, the introduced Trp is the only reporter group for each of the mutant proteins.
Mutants Q581W and E591W: Fluorescence and Effect of Phlorizin-The corrected spectra of Trp fluorescence ( ex ϭ 295 nm) of Q581W and E591W are shown in Fig. 1 (A and B,  respectively). The emission maxima of both Trp mutants are in the range of 347-350 nm, which is typical for a hydrophilic environment. 100 M phlorizin was used to monitor the percentage of quenching regarding the various positions of each Trp mutant. Quenching of fluorescence reaching ϳ54% with a Fluorescence parameters were determined from the data shown in Fig. 3. slight red shift (ϳ3 nm) was observed in the spectrum of Q581W, whereas the fluorescence of E591W was quenched by 58%, accompanied by a more noticeable red shift (ϳ5-6 nm).
Mutants R601W, D611W, E621W, and L630W: Fluorescence and Effect of Phlorizin-The corrected Trp fluorescence spectra of R601W, D611W, E621W, and L630W are shown in Fig. 1  (C-F, respectively). The fluorescence maxima for these proteins are in the range of 340 -343 nm, indicating a slightly less polar environment compared with Q581W and E591W. The fluorescence of R601W was quenched upon addition of phlorizin by 67% with a 8-nm red-shifted spectrum, D611W by 80% with a maximum red shift of 14 nm, and E621W by 65% with a red shift of 10 -11 nm. Fluorescence quenching of L630W was only ϳ18%; however, the same red shift as in E621W was observed. The differences between the fluorescence quenching spectra of the various mutants suggest that different segments of loop 13 react differently to the binding of phlorizin to the loop. Such differences could, however, also be due to differences in the phlorizin affinity of the mutants. We therefore determined the apparent binding affinities. Fig. 2. About 90% of the fluorescence of each mutant was quenched at the highest concentration of phlorizin. The estimated equilibrium dissociation constants for all mutants are presented in Table I. For calculation of binding constants, the peak points with shifting toward longer wavelength were collected upon each addition of phlorizin. The mutant proteins (e.g. Q581W, E591W, R601W, E621W, and L630W) showed values very similar to those of wild-type loop 13 (12), indicating that the replacement of amino acids at these positions with Trp residues does change the ability of the protein to bind phlorizin. However, the D611W mutant has a lower affinity, which has to be taken into account when conformational changes are considered (see below).

Effect of the Mutations on Phlorizin Binding-Titration of all single Trp mutants with different concentrations of phlorizin is shown in
Effect of Phlorizin on Trp Accessibility Monitored by Collisional Quenching-The collisional quencher acrylamide was used to detect changes in the "availability" of Trp to the surrounding solvent. The Stern-Volmer quenching plots of each mutant in which F 0 /F is plotted against the acrylamide concentration in the presence and absence of phlorizin were linear and are shown in Fig. 3. The Stern-Volmer constants for all mutants are compiled in Table II. Upon phlorizin binding, the Stern-Volmer constant for Q581W did not change significantly. At 50 M phlorizin, a non-significant protection effect was noted for mutant Q581W. For the E591W and R601W mutant proteins, the quenching constants started to increase, and an  ϳ4 -5-fold increase by phlorizin was found for D611W. E621W showed less quenching, and almost no quenching was observed for L630W. These data suggest that the major conformational changes are initiated by phlorizin at and/or in the vicinity of amino acid 611.
Phloretin-induced Changes in Trp Fluorescence and Accessibility to Acrylamide-To check the effect of phloretin, the aglucon of phlorizin, on Trp quenching and accessibility, the same experiments were also performed for D611W in the presence of phloretin. The corrected spectra of Trp fluorescence ( ex ϭ 295 nm) of D611W in the presence and absence of 100 M phloretin are shown in Fig. 4A. The fluorescence of D611W was quenched upon addition of phloretin by 69% with a 9 -10-nm red-shifted spectrum. The linear Stern-Volmer quenching plot of D611W in the presence and absence of phloretin is shown in Fig. 4B. An increase in the Stern-Volmer quenching constant for D611W from 11 to 31 M Ϫ1 was observed in the presence of 50 M phloretin.

Detection of Truncated Loop 13 Photocross-linked to 3-AP by MALDI Mass Spectrometry-
The reaction product of photolabeling of truncated loop 13 with 3-AP was analyzed by MALDI mass spectrometry. Fig. 5 (upper panel) shows the expected mass peak for loop 13 at m/z 8825.60, and Fig. 5 (lower panel) demonstrates photoaffinity labeling of the loop with 3-AP. Peak I corresponds to truncated loop 13 (m/z 8825.60), and peak II corresponds to the photolabeled truncated loop 13 protein (m/z 9275.10). The mass difference between peaks I and peak II (Fig. 5, lower panel) corresponds to the mass of photolyzed 3-azidophlorizin (m/z 449.5).
MALDI Mass Spectrometry of Loop 13 Photolabeled with 3-AP-As a next step in the identification of the ligand contact points, photolabeled truncated loop 13 was digested in gel with trypsin. After extraction of the peptides from the gel, the resulting mixture was analyzed in MALDI-TOF mode with unlabeled truncated loop 13 as a reference (Table III). The only additional peak observed in the labeled probe was at m/z 624.52 (Fig. 6, lower panel). This can be explained as an adduct peak to m/z 175.11 (corresponding to Arg-602), with a difference of 449.41. The latter is equal to the mass of photolyzed azidophlorizin. All other peaks appeared also in the control as shown in Fig. 6 (upper panel).
Structural Determination-The CD spectrum of wild-type truncated loop 13 at pH 6.8 is shown in Fig. 7. The CD spectrum shows characteristic minima at 222 and 208 nm, from which the secondary structure was estimated to consist of 37% ␣-helical residues and 63% random coil. The secondary structure agrees well with calculated values based on the amino acid sequence (14). DISCUSSION Elucidation of the structure-function relation of membrane proteins has been hindered by difficulties in expressing the complete protein. We therefore chose to express a fragment supposed to function in the receptor region for the inhibitor phlorizin. Six different single Trp mutants of truncated loop 13 were overexpressed using site-directed mutagenesis to investigate the binding of phlorizin in detail. The use of a truncated loop 13 (amino acids 564 -638) compared with the previously used complete loop 13 was favored to introduce reporter molecules at different positions of the loop, to increase the solubility, and to avoid complications by intraprotein disulfide bridge formation between cysteines 560 and 608 (12). proteins were separated by SDS-PAGE; and protein band of interest was excised, destained, and digested in gel with trypsin. The peptides were extracted from the gel pieces with acetonitrile. Subsequently, 0.5-l aliquots of the generated cleavage products were dispensed onto the sample support, followed by 0.5 l of ␣-cyano-4-hydroxycinnamic acid matrix solution; and the MALDI mass spectra were recorded. Upper panel, truncated loop 13 before photoaffinity labeling; lower panel, truncated loop 13 after photoaffinity labeling with 3-AP. The peak indicated by the asterisk clearly indicates the mass shift due to modification of Arg-602 with 3-AP. Intens., intensity. Truncation of the molecule did not, however, interfere with the ability of the protein to interact with phlorizin. The phlorizin dissociation constants in our studies are very similar to those determined for the formation of the initial collision complex by Oulianova and Berteloot (9) in intact membrane vesicles and also agree with the observed binding constant for complete loop 13 (12).
To facilitate discussion of the secondary structure of trun-cated loop 13 and its conformational changes upon phlorizin binding, the predicted conformation based on the amino acid sequence is shown in Fig. 8A. In agreement with this prediction, the maxima of Trp fluorescence of the various mutants differ. For example, mutants 581 and 591 show a spectrum that has a maximum close to 350 nm, suggesting a random coil orientation of the neighboring segments. For the other mutants e. g. 601, 611, 621, and 630, a more hydrophobic environment of the Trp can be assumed (blue shift in the maxima), suggesting a location in a more ordered, probably ␣-helical structure. The presence of these ␣-helices in solution is confirmed by the CD spectrum, which quite nicely reflects the relative abundance of the various conformations.
Several major spectral changes are induced by phlorizin binding to truncated loop 13. First, a red shift in the fluorescence maximum indicates a transition between an ordered state (hydrophobic) to a less ordered state (more hydrophilic) of the loop. The most evident shift was observed for mutant D611W. Thus, phlorizin binding appears to induce the opening of the region around amino acid 611 of loop 13, which before seemed to be buried between the two ␣-helices (see model in Fig. 8A).
The shift in the maximum of Trp fluorescence toward a longer wavelength could, however, also be caused by a strong interaction of phlorizin-associated water with the loop. Such a complex between phlorizin and water molecules can be formed by H-bonding between OH groups of phlorizin (e.g. mainly by the 4-and 6-OH groups of ring A and 4-OH of ring B). This suggests the position of Asp-611 in the binding pocket. A critical role of this region of loop 13 in phlorizin binding has also been proposed from studies on the intact carrier using sitedirected mutagenesis (11) and from the biophysical studies using isolated loop 13 in solution (12).
The similar red shift in the fluorescence maximum of L630W is difficult to explain by the above assumption. Perhaps in this part of loop 13, bound water molecules situated around the protein contribute mostly to the signal. In other studies, ligandinduced conformational changes were also deduced from the fluorescence spectra. In this instance, both water around the ligand and water around the protein contributed to the shift; the contribution of each is in most cases difficult to estimate (18).
Furthermore, fluorescence quenching was observed. We would like to correlate the quenching with the proximity of the indole ring of Trp to the phlorizin molecule. The main mechanism responsible for fluorescence quenching is charge/energy transfer between phlorizin and Trp residues. Similar effects were observed for collagen fluorescence in the presence of varying concentrations of hypericin (16,17). The maximum fluorescence quenching at position 611 could then indicate that phlorizin ring A is very close to the free indole ring of tryptophan. For the amino acids located in region 606 -609, also an interaction, although with slightly less affinity, would ensue. The binding site is thereby defined. The minimum quenching observed for mutant L630W suggests a positional arrangement of this part of loop far from the binding region. In the case of mutants Q581W and E591W, significant quenching signals in the presence of phlorizin can be attributed to a conformational change that brings this part of loop 13 somehow close to the binding site.
Mutation D611W resulted in a decrease in phlorizin affinity. This can be explained by the fact that, because of the replacement of aspartate and the removal of the negative side chain, the network of neighboring amino acids is disturbed, which plays an important role in establishing the preformed pocket for phlorizin binding during the initial complex formation.
A surprising result of this study is that the region between residues 591 and 621 is more accessible to acrylamide in the presence of phlorizin than in its absence. In our previous studies on phlorizin binding to loop 13, a protection effect of phlorizin was observed, leading to a decreased accessibility of Trp at position 561 and its resonance partner Tyr at position 604 (12). The difference might be due to the fact that, in the former studies, the protein contained a disulfide bond, which can alter the conformational changes significantly. The significant change observed for mutant L630W is in agreement with its position in the ␣-helical region of the protein and also suggests its location to be far from aromatic ring A of phlorizin compared with the location of Q581W (Fig. 8A). The high accessibility of D611W to acrylamide after phlorizin binding suggests that the interaction with phlorizin occurs in an aqueous environment; hence, no protection effect is observed.
It is interesting to note that phloretin, the aglucon of phlorizin, induced qualitatively similar changes in D611W as phlorizin, although the effects on fluorescence and acrylamide accessibility were smaller. These results suggest that phloretin might bind to the same site as phlorizin. The larger disturbance of the conformation of loop 13 by phlorizin might be explained by the bulky nature of the glucose residue attached to the two aromatic rings. If phloretin indeed binds to the same site as phlorizin, the noncompetitive nature of the inhibition of D-glucose transport by phloretin might be due to a close coupling between loop 13 and the sugar-translocating positions of SGLT1. Such coupling can occur by the formation of disulfide bonds between the extramembranous loops; direct evidence for the presence of such bonds has been obtained recently. 2 Recently, we also found ligand-induced conformational changes in SGLT1 in intact rabbit renal medullary brush-border membrane vesicles by monitoring the accessibility of extramembranous cysteine residues with polyethylene oxide-maleimide-activated biotin, followed by specific immunodetection of the biotin-coupled protein 2 and by determining binding probability and unbinding force using antibodies against loop 13 attached to an atomic force microscopic cantilever (15).
The conformation of phlorizin in aqueous solutions was previously studied by two-dimensional NMR and pharmacophore analysis by Wielert-Badt et al. (19). Their model indicates that the interactions via hydrogen bonds from the 2-, 3-, 4-, and 6-hydroxyl groups of the pyranoside and the 4-and 6-OH groups of aromatic ring A are essential for phlorizin binding. This information can now be combined with the Trp fluorescence data to estimate the dimensions of the phlorizin-binding and recognition region. We can assume that phlorizin binds to the region very close to position 611 by H-bonding probably with the 4-OH group by acceptor/donor atom O-4 of aromatic ring A. Thus, position 611 is supposed to be very important in phlorizin recognition, as also suggested by the lower affinity of the D611W mutant. The H-bonding of any one hydrophobic amino acid located in region 606 -609 with the 6-OH group by acceptor/donor atom O-6 of aromatic ring A of phlorizin also causes a strong interaction.
The combination of photoaffinity labeling, SDS-PAGE, enzymatic fragmentation, and MALDI mass spectrometry analysis with the resulting allocation of the attachment site of the ortho-position of ring B of phlorizin to Arg-602 defines directly the phlorizin-binding site in loop 13. Thus, an interaction of this region with ring B of phlorizin can be assumed. The binding pocket could thereby accommodate the phlorizin molecule in the predicted conformation.
On the basis of the photoaffinity labeling pattern of truncated loop 13 with 3-AP, we could provide a generalized mechanism by which phlorizin inhibits sugar transport by SGLT1. According to our assumption, a phlorizin-binding region in SGLT1 is located between amino acids 601 and 611; in this region, aromatic rings A and B of phlorizin interact with hydrophobic amino acids, and the glucoside moiety of phlorizin is free to interact with sugar interaction sites of SGLT1 (5,6). Previous mutagenesis studies provided evidence that also Asp-176 might be involved in determining the overall affinity of phlorizin for the transporter (23). Thus, there seems to be close vicinity between loop 13 and Asp-176, again probably caused by disulfide bonds between various parts of the transporter.
According to our modeling assumptions based on secondary structure prediction and CD analysis, the regions (especially those between residues 599 and 607 and residues 620 and 631) around the binding pocket are in ␣-helical conformation. In the absence of phlorizin, the region around amino acid 611 appears to interact with part of the ␣-helical region, in particular with the hydrophobic amino acids between positions 606 and 609. Upon phlorizin binding, this region becomes exposed to an aqueous environment as shown by the direction of the large arrow in Fig. 8A. In addition, the 4-OH group of aromatic ring A comes close to position 611, strongly altering the Trp emission (Fig. 8B). At the same time, aromatic ring B comes conformationally close to position 600. We also propose that phlorizin binding results in reduction of van der Waals forces between parts of the two helices (e.g. interaction between ␣-helical residues 599 -607 and 620 -629), whereas some part of the ␣-helix (e.g. that close to position 630) is accessible to acrylamide irrespective of the presence of phlorizin.
All proteins need a specific three-dimensional structure to perform their particular function. Keeping in mind the difficulties in resolving the structure of mammalian membrane proteins in general and SGLT1 in particular, the use of SGLT1 fragments appears to be advantageous to study their function in more detail using different biophysical methods, although the refinement of the assumed conformation and dimensions of phlorizin-binding sites during SGLT1-inhibitor complex formation by NMR or x-ray crystallography is still required.