Portal de Periódicos da CAPES

Mutational Analysis of the Hexose Transporter of Plasmodium falciparum and Development of a Three-dimensional Model

2002; Elsevier BV; Volume: 277; Issue: 34 Linguagem: Inglês

10.1074/jbc.m204337200

1083-351X

Suzanne Kathryn Manning, Charles J. Woodrow, Felipe Zúñiga, P. Iserovich, Jorge Fischbarg, Abraham I. Louw, Sanjeev Krishna,

Vibrio bacteria research studies

Plasmodium falciparuminfection kills more than 1 million children annually. Novel drug targets are urgently being sought as multidrug resistance limits the range of treatment options for this protozoan pathogen. PfHT1, the major hexose transporter of P. falciparum is a promising new target. We report detailed structure-function studies on PfHT1 using site-directed mutagenesis approaches on residues located in helix V (Q169N) and helix VII (302SGL → AGT). Studies with hexose analogues in these mutants have established that hexose recognition and permeation are intimately linked to these helices. A “fructose filter” effect results from the Q169N mutation (abolishing fructose uptake but preserving affinity and transport of glucose, as reported in Woodrow, C. J., Burchmore, R. J. S., and Krishna, S. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 9931–9936). Associated changes in competition for glucose uptake by C-2, C-3, and C-6 glucose analogues compared with native PfHT1 indicate subtle alterations in substrate interaction in this mutant. The Km values for glucose uptake in helix VII mutants are also similar to native PfHT1. Hydrogen bonding to positions C-5 and C-6 in glucose analogues becomes relatively more important in these mutants compared with native PfHT1. To increase understanding of hexose permeation pathways in PfHT1, we have developed the first three-dimensional model for PfHT1. As predicted for GLUT1, the principal mammalian glucose transporter, PfHT1 contains a main and an auxiliary channel. After modeling, the Q169N mutation leads predominantly to local structural changes, including displacement of neighboring helix IV. The 302SGL position in helix VII lies in the same plane as Gln-169 in helix V but is also adjacent to the main hexose permeation pathway, consistent with results from experiments mutating this triplet motif. Furthermore, there are obvious structural and functional differences between GLUT1 and PfHT1 that can now be explored in detail using the approaches presented here. The development of specific inhibitors for PfHT1 will also be aided by these insights. Plasmodium falciparuminfection kills more than 1 million children annually. Novel drug targets are urgently being sought as multidrug resistance limits the range of treatment options for this protozoan pathogen. PfHT1, the major hexose transporter of P. falciparum is a promising new target. We report detailed structure-function studies on PfHT1 using site-directed mutagenesis approaches on residues located in helix V (Q169N) and helix VII (302SGL → AGT). Studies with hexose analogues in these mutants have established that hexose recognition and permeation are intimately linked to these helices. A “fructose filter” effect results from the Q169N mutation (abolishing fructose uptake but preserving affinity and transport of glucose, as reported in Woodrow, C. J., Burchmore, R. J. S., and Krishna, S. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 9931–9936). Associated changes in competition for glucose uptake by C-2, C-3, and C-6 glucose analogues compared with native PfHT1 indicate subtle alterations in substrate interaction in this mutant. The Km values for glucose uptake in helix VII mutants are also similar to native PfHT1. Hydrogen bonding to positions C-5 and C-6 in glucose analogues becomes relatively more important in these mutants compared with native PfHT1. To increase understanding of hexose permeation pathways in PfHT1, we have developed the first three-dimensional model for PfHT1. As predicted for GLUT1, the principal mammalian glucose transporter, PfHT1 contains a main and an auxiliary channel. After modeling, the Q169N mutation leads predominantly to local structural changes, including displacement of neighboring helix IV. The 302SGL position in helix VII lies in the same plane as Gln-169 in helix V but is also adjacent to the main hexose permeation pathway, consistent with results from experiments mutating this triplet motif. Furthermore, there are obvious structural and functional differences between GLUT1 and PfHT1 that can now be explored in detail using the approaches presented here. The development of specific inhibitors for PfHT1 will also be aided by these insights. P. falciparum hexose transporter 1 mammalian facilitative hexose transporter O-methylglucose 5-AHM, 2,5 anhydro-d-mannitol deoxy-d-glucose hexose transporter primers containing Kozak and BglII sites amino acids Plasmodium falciparum infection causes malaria, which afflicts 200–300 million people and kills 1–2 million individuals annually (1Phillips R. Clin. Microbiol. Rev. 2001; 14: 208-226Crossref PubMed Scopus (180) Google Scholar). Multidrug resistant parasites are now established in many parts of the world, limiting significantly the range of chemotherapeutic options. Antimalarial vaccines are still some years away from being useful (2Nosten F. van Vugt M. Curr. Opin. Infect. Dis. 1996; 9: 429-434Crossref Scopus (12) Google Scholar). The recent discovery of the P. falciparum hexose transporter, PfHT11 (3Woodrow C.J. Penny J.I. Krishna S. J. Biol. Chem. 1999; 274: 7272-7277Abstract Full Text Full Text PDF PubMed Scopus (116) Google Scholar), has provided a new avenue for drug target development based on the following facts. First, intra-erythrocytic P. falciparum depends totally on glucose from its host for energy (4Kirk K. Horner H.A. Kirk J. Mol. Biochem. Parasitol. 1996; 82: 195-205Crossref PubMed Scopus (79) Google Scholar). Parasites increase glucose consumption of erythrocytes by up to 100-fold and cannot tolerate decreased extracellular glucose concentrations without suffering immediate falls in intraparasitic pH (4Kirk K. Horner H.A. Kirk J. Mol. Biochem. Parasitol. 1996; 82: 195-205Crossref PubMed Scopus (79) Google Scholar). The uptake of glucose by parasites is likely to be solely via PfHT1, because it is encoded by a single copy gene with no close paralogues (5Woodrow C.J. Burchmore R.J.S. Krishna S. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9931-9936Crossref PubMed Scopus (75) Google Scholar). Furthermore, mRNA for PfHT1 is expressed during the ring stage of parasite development in anticipation of large increases in demands for glucose by trophozoites. PfHT1 has a relatively high affinity for hexoses with a 10-fold higher affinity for d-glucose compared with the major mammalian erythrocyte and cerebral blood vessel glucose transporter, GLUT1 (3Woodrow C.J. Penny J.I. Krishna S. J. Biol. Chem. 1999; 274: 7272-7277Abstract Full Text Full Text PDF PubMed Scopus (116) Google Scholar, 6Krishna S. Woodrow C.J. Cardew G. Transport and Trafficking in the Malaria-infected Erythrocyte. 226. John Wiley & Sons, Ltd, London1999: 126-144Google Scholar). PfHT1 also transports fructose in contrast to mammalian facilitative hexose transporters (GLUTs), which generally transport either glucose or fructose. No structural information is available for PfHT1; therefore, indirect approaches are needed to identify key amino acid residues responsible for PfHT1-hexose interactions. Mutagenesis approaches have been extensively applied to GLUT1 and include glycosylation-scanning (7Hresko R.C. Kruse M. Strube M. Mueckler M. J. Biol. Chem. 1994; 269: 20482-20488Abstract Full Text PDF PubMed Google Scholar), cysteine-scanning (8Hruz P.E. Mueckler M.M. J. Biol. Chem. 1999; 274: 36176-36180Abstract Full Text Full Text PDF PubMed Scopus (54) Google Scholar), alanine-scanning (9Seatter M.J. Kane S. Porter L.M. Arbuckle M.I. Melvin D.R. Gould G.W. Biochemistry. 1997; 36: 6401-6407Crossref PubMed Scopus (13) Google Scholar), and site-directed mutagenesis (10Mueckler M. Weng W. Kruse M. J. Biol. Chem. 1994; 269: 20533-20538Abstract Full Text PDF PubMed Google Scholar), as well as construction of chimeras (11Seatter M.J., De La Rue S. Porter L.M. Gould G.W. Biochemistry. 1998; 37: 1322-1326Crossref PubMed Scopus (94) Google Scholar). Results have provided insight into secondary and tertiary structures of GLUT1 and their role in determining substrate selectivity (12Hruz P.W. Mueckler M.M. Mol. Membr. Biol. 2001; 18: 183-193Crossref PubMed Scopus (139) Google Scholar). A key insight into the hexose permeation pathway of mammalian GLUTs has developed from the observation that GLUTs 1, 3, and 4 transport glucose only and all contain a conserved QLS motif in helix VII, whereas GLUTs 2 and 5 are capable of fructose transport and do not have this conserved motif (11Seatter M.J., De La Rue S. Porter L.M. Gould G.W. Biochemistry. 1998; 37: 1322-1326Crossref PubMed Scopus (94) Google Scholar) (see Fig. 1). Chimeric GLUT2/GLUT3 constructs confirm that fructose selectivity depends in part on the QLS motif in helix VII (13Arbuckle M.I. Kane S. Porter L.M. Seatter M.J. Gould G.W. Biochemistry. 1996; 35: 16519-16527Crossref PubMed Scopus (59) Google Scholar). In GLUT1, helix VII is also important in exofacial binding because mutations reduce affinity for exofacial site-specific ligands 2-N-4-(1-azi-2,2,2-trifluoroethyl) benzoyl-1,3-bis(d-mannos-4-yloxy)-2-propylamine (ATB-BMPA) and 4,6-O-ethylidene-α-d-glucose (13Arbuckle M.I. Kane S. Porter L.M. Seatter M.J. Gould G.W. Biochemistry. 1996; 35: 16519-16527Crossref PubMed Scopus (59) Google Scholar, 14Hashiramoto M. Kadowaki T. Clark A. Muraoka A. Momomura K. Sakura H. Tobe K. Akanuma Y. Yazaki Y. Holman G. Kasuga M. J. Biol. Chem. 1992; 267: 17502-17507Abstract Full Text PDF PubMed Google Scholar). In contrast to these observations, in PfHT1 helix V is clearly critical for fructose transport, since this is selectively ablated by a single amino acid (Q169N) mutation in this helix (5Woodrow C.J. Burchmore R.J.S. Krishna S. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9931-9936Crossref PubMed Scopus (75) Google Scholar). Affinity for glucose as well as uptake is preserved in this mutant, indicating that different mechanisms might determine hexose selectivity in the mammalian and parasite transporters. In another pathogenic protozoan, Trypanosoma brucei, the hexose transporter THT1 can also transport both glucose and fructose with an affinity for 2-deoxy-d-glucose that is comparable to that of PfHT1 (Km ∼0.5 mm) whereas its affinity for fructose is higher (Ki ∼2.6 mm) than that of PfHT1 (Ki ∼11 mm) (15Tetaud E. Bringaud F. Chabas S. Barrett M.P. Baltz T. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 8278-8282Crossref PubMed Scopus (43) Google Scholar). Preliminary studies substituting the PfHT1 helix VII 302SGL by the GLUT1 equivalent (QLS) did not result in measurable glucose uptake (data not shown). We have therefore changed this PfHT1 helix VII302SGL motif to that found in the corresponding region of helix VII in THT1 (302SGL → AGT, see Fig. 1) Additional single mutations (S302A and L304T) within the helix VII SGL motif in PfHT1 are included to establish more precisely the contribution of individual amino acids to substrate-protein interaction. A detailed kinetic analysis of the previously described helix V mutant Q169N is also presented to differentiate the roles of these two helices with respect to fructose transport and substrate-transporter interactions. Recently, a three-dimensional model of GLUT1 has been published describing a helical arrangement that is consistent with experimental observations as well as theoretical structural considerations (16Zuniga F.A. Shi G. Haller J.F. Rubashkin A. Flynn D.R. Iserovich P. Fischbarg J. J. Biol. Chem. 2001; 276: 44970-44975Abstract Full Text Full Text PDF PubMed Scopus (50) Google Scholar). This model predicts two transport channels, both of them partially limited by the QLS motif in helix VII. We have now combined our site-directed mutagenesis studies with similar molecular modeling of PfHT1 to increase our understanding of the hexose permeation pathway of this important parasite transporter. PfHT1 helix VII mutants were created using a nested PCR strategy. Complementary PCR primers (Table I) were designed to introduce a double point-mutated construct of amino acids Ser302 and Leu304 to produce 302SGL → AGT and single point-mutated S302A and L304T constructs with wild type PfHT1 DNA as template using Pfu polymerase (Stratagene, La Jolla, CA; Table I for primer sequences). Restriction sites for confirmation of mutations were included directly over the mutated site. HTKB primers (hexose transporter primers containingKozak and BglII site) used for the amplification of the full-length product includeBglII restriction sites at their 5′-ends, and the forward primer (HTKB1) includes a strong eukaryotic Kozak consensus in front of the start codon (CACCATG). The resulting PCR products were ligated into the pSP vector containing 5′- and 3′-untranslatedXenopus β-globin sequences as previously described (17Gould G. Lienhard G.E. Biochemistry. 1989; 28: 9447-9452Crossref PubMed Scopus (55) Google Scholar). All products were verified by sequence analysis.Table ISequences of primers used in nested PCR as described in methodsPrimer namePrimer sequence (5′ → 3′)Restriction sitesfAGTGGATGTTTGCTAGCTGGTACACAACAATTTACAGGNheIrAGTCCTGTAAATTGTTGTGTACCAGCTAGCAAACATCCfAGLGGATGTTTGCTAGCTGGTCTGCAGCAATTTACAGNheI, PstIrAGLCTGTAAATTGCTGCAGACCAGCTAGCAAACATCCfSGTGGATGTTTGCTATCTGGTACCCAACAATTTACAGKpnIrSGTCTGTAAATTGTTGGGTACCAGATAGCAAACATCCfHTKBACGTACAGATCTCACCATGACGAAAAGTTCGAAAGBglIIrHTKBACGTACAGATCTTCATACAACCGACTTGGTC Open table in a new tab Xenopus oocytes were assayed as previously described (3Woodrow C.J. Penny J.I. Krishna S. J. Biol. Chem. 1999; 274: 7272-7277Abstract Full Text Full Text PDF PubMed Scopus (116) Google Scholar). cRNAs were transcribed (MEGAscriptTM SP6, Ambion, Austin, TX) from XbaI-linearized plasmids and 10 ng of each RNA in 30–40 nl of RNase-free water was injected into the oocytes with RNase-free water-injected oocytes as controls. Uptake assays were performed at room temperature for 20 min, 24–96 h after injection on groups of 6–8 oocytes in Barth's medium containing permeant d-[U-14C]glucose (310 mCi·mmol−1), 2-deoxy-d-[U-14C]glucose (2-DOG; 58 mCi·mmol−1), 3-O-[14C]methyl-d-glucose (3-OMG) (50 mCi·mmol−1), ord-[U-14C]fructose (289 mCi·mmol−1) from Amersham Biosciences). All uptakes were linear for the times used in these assays. Uptakes of substrate after correction for uptakes in water-injected controls were assayed in experiments with incubation times and experimental conditions as previously described (3Woodrow C.J. Penny J.I. Krishna S. J. Biol. Chem. 1999; 274: 7272-7277Abstract Full Text Full Text PDF PubMed Scopus (116) Google Scholar). A Michaelis-Menten model (PRISM Ver. 2, Graphpad, San Diego) was used for all estimations of kinetic parameters by non-linear regression analysis. All studies with competitors were carried out 24–48 h after injection for 20 min in Barth's medium containing radiolabeled d-glucose (2.69 μm, 323 mCi·mmol−1d-[U-14C]glucose and 35 μmunlabeled d-glucose) and varying amounts of competitor (all from Sigma). Results were analyzed with PRISM using a one-site competition model. To determine the degree of correspondence between the helical assignments for the GLUT1 structure and PfHT1, we analyzed amino acid sequence similarities between the GLUT family (isoforms 1–5), three trypanosomatid transporters and PfHT1. We first performed an alignment between all sequences (ClustalW) (18Thompson J.D. Higgins D.G. Gibson T.J. Nucleic Acids Res. 1994; 22: 4673-4680Crossref PubMed Scopus (56003) Google Scholar) and then used the program TMpred (19Pasquier C. Promponas V.J. Palaios G.A. Hamodrakas J.S. Hamodrakas S.J. Protein Eng. 1999; 12: 381-385Crossref PubMed Scopus (145) Google Scholar) to predict in each sequence the location of the transmembrane regions. The putative PfHT1 structure was modeled by homology, using the three-dimensional model for GLUT1 (16Zuniga F.A. Shi G. Haller J.F. Rubashkin A. Flynn D.R. Iserovich P. Fischbarg J. J. Biol. Chem. 2001; 276: 44970-44975Abstract Full Text Full Text PDF PubMed Scopus (50) Google Scholar) (PDB code 1JA5) as a template. The GLUT1 model and its validation have been described in detail. We used the SwissModel module from the program Swiss-PDB Viewer (vs. 3.5) to align the GLUT1 and the PfHT1 (Swiss Protein Data Base O97467) sequences. We then obtained an improved alignment by including gaps and matching in both sequences the corresponding conserved and semi-conserved amino acids using the program ClustalW (18Thompson J.D. Higgins D.G. Gibson T.J. Nucleic Acids Res. 1994; 22: 4673-4680Crossref PubMed Scopus (56003) Google Scholar). Subsequently, these gaps were manually introduced into the alignment in the Swiss-PDB environment. This resulted in a tentative structure for PfHT1, which presented five gap regions (between PfHT1 residues 19–32, 190–201, 207–208, 278–279 and 326) all located in loops. We utilized a Silicon Graphics Octane work station with InsightII software (Accelrys Inc.). Using the Biopolymer module, we loaded the tentative structure for PfHT1, added the missing amino acids in the gaps, and performed energy minimization using the Discovery module (CVFF force field; 200 iterations with the steepest descent algorithm, and 500 more with the conjugate gradient algorithm). Subsequently, we verified that the residues in the helical segments still fell in allowed Ramachandran regions, and we turned to refining the loop regions using the Swiss-PDB Viewer. We searched for homologous loops with best Ramachandran conformations, avoiding bad contacts and undesirable torsion angles. For each new loop we performed a few steps of energy minimization (GROMOS 43b1 force field; 20 iterations with the steepest descent algorithm). Finally, we performed an energy minimization of the whole protein (in vacuum) using the Discovery module as above, so as to optimize close contacts, bond distances, and angles. Fig. 1 confirms that there are no major differences between helical assignments for different mammalian and parasite transporters, suggesting a good deal of structural conservation between them. For helices V, VII, and VIII, known to be important for the transport of substrates in human GLUTs, the assigned locations are mostly the same, and all conserved amino acids fall inside transmembrane regions as well. This is true also for helices II, III, IV, IX, X, XI, and XII. Only for helices I and VI are there differences in the relative positions of transmembrane regions between GLUTs and the trypanosomatid transporters of T. brucei, T. vivax, and T. cruzi. In contrast to the trypanosomatid sequences, helices for PfHT1 are located in well conserved regions that also align well with those of GLUTs. This reinforced our impression that helical structures of PfHT1 were in positions that corresponded to those in GLUT1 well enough to pursue further development of a putative PfHT1 structure. The PfHT1 structure obtained is depicted in Fig.2 (PDB 1LVI). The quality of this structure was determined using PROCHECK (version 3.3.2) (20Laskowki R.A. MacArthur M.W. Moss D.S. Thornton J.M. J. Appl. Crystallogr. 1993; 26: 283-291Crossref Google Scholar). The distribution thus found for the φ/ψ angles was as follows: core regions, 76.4% (349 aa); allowed regions, 22.5% (103 aa); generously allowed regions, 1.1% (5 amino acids); and disallowed regions, 0%. In contrast to GLUT1, the N-terminal region of PfHT1 is longer, placing it relatively much closer to the C terminus. Interestingly, whereas most loop lengths are similar between GLUT1 and PfHT1, the extracellular loop between helices V and VI is 14 residues longer than in GLUT1 and has a preponderance of polar residues. Fig. 3 places residues that have been mutated in this and previous studies within the hypothetical structure of PfHT1. As in GLUT1, both a main and auxiliary channel are predicted and shown in the plane of the SGL motif that has been mutated in helix VII in PfHT1. The main channel is predicted to act as a conduit for sugars, with the auxiliary channel perhaps permitting also passage of non-sugar substrates. Likewise, QLS in helix VII of GLUT1 is also predicted to contact both channels (Ref. 16Zuniga F.A. Shi G. Haller J.F. Rubashkin A. Flynn D.R. Iserovich P. Fischbarg J. J. Biol. Chem. 2001; 276: 44970-44975Abstract Full Text Full Text PDF PubMed Scopus (50) Google Scholar and see below). Gln169 is predicted to lie in the same horizontal plane and in the vicinity of the SGL motif, but as shown here, points away from the main channel. After dynamic analysis (Fig.4), the Asn mutant side chain (one methylene shorter than the original Gln) appears displaced inwards, as if packed more compactly. However, this new arrangement appears to result in a displacement of helix IV away from the recognition site and the main channel. By contrast, the positions of helices VII and VIII do not change much. The Q169N mutant can therefore be envisaged to lead to alterations in a critical hydrogen bond that perturbs the binding of fructose and not of glucose, as we have previously observed experimentally (5Woodrow C.J. Burchmore R.J.S. Krishna S. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9931-9936Crossref PubMed Scopus (75) Google Scholar). Reconfiguration of the triplet motif (SGL) in helix VII of PfHT1 to an equivalent motif (AGL) found in THT1, a T. bruceihexose transporter (21Barrett M.P. Tetaud E. Seyfang A. Bringaud F. Baltz T. Mol. Biochem. Protozool. 1998; 91: 195-205Crossref PubMed Scopus (59) Google Scholar), resulted in approximately half the overall uptake rates for d-glucose when compared with native PfHT1, 48 h after microinjection of equivalent concentrations of cRNA (mean ± S.E. values for uptake were 77.4 ± 5.9 pmol/oocyte/h for PfHT1 and 42.3 ± 2.4 pmol/oocyte/h for the302SGL → AGT construct). The double point mutation in the above construct was segregated to produce two new mutants S302A and L304T in order to examine if a single amino acid within the triple motif construct was responsible for functional changes. Uptake measurements with the same batch of oocytes as before confirmed that both single amino acid mutations were functional at levels comparable to the 302SGL → AGT mutant (at 48 hd-glucose uptake values were 48.3 ± 6.2 and 35.9 ± 3.9 pmol/oocyte/h, respectively) when injected with equivalent amounts of cRNA. These uptake rates were ample enough to allow the more detailed kinetic analyses that follow. Table II summarizesKm values of mutants of PfHT1 for different substrates and compares these to values previously obtained for native PfHT1. The mutation 302SGL → AGT in helix VII does not alter the affinity for d-glucose (mean ± S.E.Km = 1.3 ± 0.3 mm for the mutant compared with a wild type Km = 1.0 ± 0.2 mm). The segregated helix VII mutants S302A and L304T also show similar affinities for d-glucose (Km = 0.47 ± 0.1 mm and 0.68 ± 0.04 mm, respectively).Table IIKinetic analysis with operative analogues on helix V and VII mutantsSubstratePfHT12-aPreviously reported data.Q169N302SGL → AGTS302AL304TAldose analoguesd-glucoseKm = 1.0 ± 0.2Km = 1.2 ± 0.22-aPreviously reported data.Km = 1.3 ± 0.3Km = 0.47 ± 0.1Km = 0.68 ± 0.04(n = 4)C-1 position1-deoxy-d-glucoseKi = 15.312-bMean of two independent experiments.Ki = 13.6 ± 2Ki = 32 ± 6.4Ki = 32.7, 20.5Ki = 19 ± 2.1C-2 position2-deoxy-d-glucoseKm = 1.3Km = 3.4 ± 0.3Km = 5, 2.2Ki = 0.7 ± 0.15Ki = 1.2 ± 0.3C-3 position3-O-methyl-d-glucoseKm = 1.3 ± 0.3Km = 7.3, 5.5Km = 4.5 ± 0.8Ki = 2, 12-bMean of two independent experiments.Ki = 1.2 ± 0.2Ki = 1.0 ± 0.1C-5 position5-thio-d-glucoseKi = 3.22-bMean of two independent experiments.Ki = 0.7, 0.8Ki = 14 ± 2.7Ki = 11.2 ± 1.6Ki = 4, 5.9C-6 position6-deoxy-d-glucoseKi = 2.2 ± 0.9Ki = 9.3, 6.4Ki = 6.8 ± 1Ki = 3.9 ± 1.6Ki = 3.8 ± 0.7Ketose analoguesd-FructoseKm = 11.5 ± 1.6No transportKm = 12.5 ± 2.0Ki = 11.22-bMean of two independent experiments.Ki = 19.7 ± 4.3Ki = 19.9 ± 1.7Ki = 17.1 ± 3C-2 position2,5-anhydro-d-mannitolKi = 1.4 ± 0.3Ki = 282-aPreviously reported data.Ki = 1.7 ± 0.15Ki = 2.2, 1.2Ki = 2.7, 1.02-a Previously reported data.2-b Mean of two independent experiments. Open table in a new tab The affinity of d-fructose for the 302SGL → AGT mutant (Km = 12.5 ± 2, Table II) is similar to its affinity for native PfHT1 (Km = 11.5 ± 1.6). The Ki for fructose for the mutant is higher but this difference is not significant (Ki = 19.7 ± 4.3; p = 0.19). The segregated helix VII mutants S302A and L304T haveKi values for d-fructose comparable to the AGT mutant (Ki = 19.9 ± 1.7 mmand 17.1 ± 3.0 mm). Results for substrate transport for the Q169N helix V mutant have been published previously (fructose uptake is abolished whereas glucose uptake is preserved with an affinity similar to that of native PfHT1, Ref. 5Woodrow C.J. Burchmore R.J.S. Krishna S. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9931-9936Crossref PubMed Scopus (75) Google Scholar). Deoxy-d-glucose (DOG) analogues provide insight into the interactions of six possible hydrogen bonding sites on glucose with PfHT1 (or GLUT1). These potential hydrogen-bonding sites are interrupted by removal of the hydroxyl group (on carbons 1–4 and 6) and by replacing the oxygen on C-5 with a sulfur atom. Previous studies with PfHT1 showed hydrogen bonding to be important at positions C-1 and C-3 with ≥10-fold increase in Ki observed with 1-DOG (Ki ∼15 mm) and 3-DOG (Ki ∼15 mm) when compared withd-glucose (Km ∼1 mm) (5Woodrow C.J. Burchmore R.J.S. Krishna S. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9931-9936Crossref PubMed Scopus (75) Google Scholar). For the helix V Q169N mutation, no change in Ki was observed for 1-DOG (mean Ki = 13.6 mm) compared with the Ki for 1-DOG measured in native PfHT1 (see above). An almost 4-fold decrease inKi for 5-thio-d-glucose was observed for the Q169N mutant (mean Ki = 0.75 mm) compared with wild type (Ki = 3.2 mm). A 4-fold increase was observed in the Ki for 6-DOG when compared with the corresponding Km orKi values for PfHT1 (Table II). DOG studies on the helix VII 302SGL → AGT mutation showed slightly higher Ki values than observed for native PfHT1 for d-glucose positions C-1 (2-fold increase inKi) and C-6 (3-fold increase inKi; p = 0.026). TheKm for 2-DOG was increased ∼3-fold (meanKm = 3.6 mmversus 1.3 mm). Assay of the S302A mutation suggests that the 2-fold increase inKi with 1-DOG seen in the 302SGL → AGT mutant can mostly be attributed to this single point mutation (meanKi = 27 mm, Table II)) and not to the second change (L304T, mean Ki = 19 mm). The Ki values for 2-DOG obtained for S302A and L304T mutants are similar to the Kmvalue for 2-DOG obtained for native PfHT1. The Kis observed for C-5 and C-6 analogues (5-thio-d-glucose and 6-DOG) for the S302A and L304T mutants were higher than for native PfHT1 and additionally synergized to decrease affinity for both analogues in the 302SGL → AGT mutant. The helix VII Q169N mutation reduces the affinity for 3-OMG (mean Km = 6.4 mm) compared with the wild type (mean Km = 1.3 mm), suggesting increased hydrophobic interactions of this mutant with the C3 methyl group. The 302SGL → AGT mutant has a significantly higher Km value (meanKm = 4.5 mm) for 3-OMG compared with wild type PfHT1 (mean Km = 1.3 mm;p = 0.024). The Ki values for S302A and L304T are not significantly different (mean Kivalues = 1.2 and 1, respectively) compared with native PfHT1, suggesting both mutations act together to increase theKi for 3-OMG in the 302SGL → AGT mutant. As PfHT1 transports d-fructose in its furanose form, 2,5-AHM (2,5-anhydro-d-mannitol), a fixed fructofuranose analogue, was used to assess interactions with mutants (5Woodrow C.J. Burchmore R.J.S. Krishna S. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9931-9936Crossref PubMed Scopus (75) Google Scholar). Previous work established that helix V Q169N mutation abolished fructose transport. All three helix VII mutants showed similarKi values to native PfHT1 for 2,5-AHM (Table II). These findings are consistent with kinetic data for fructose itself and suggest that fructose is also transported by helix VII mutants of PfHT1 in the furanose configuration. Without crystallographic data, information relating structure and function in facilitative hexose transporters relies upon indirect approaches such as mutagenesis, the use of analogues, and molecular modeling. In this work we have used all three approaches to examine the importance of a triple amino acid motif located in helix VII and a single amino acid mutation in helix V, in permeation of hexoses. Both helices V and VII have previously been implicated in substrate binding or selectivity of glucose versus fructose in different facilitative hexose transporters (5Woodrow C.J. Burchmore R.J.S. Krishna S. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9931-9936Crossref PubMed Scopus (75) Google Scholar, 11Seatter M.J., De La Rue S. Porter L.M. Gould G.W. Biochemistry. 1998; 37: 1322-1326Crossref PubMed Scopus (94) Google Scholar, 13Arbuckle M.I. Kane S. Porter L.M. Seatter M.J. Gould G.W. Biochemistry. 1996; 35: 16519-16527Crossref PubMed Scopus (59) Google Scholar). We have also developed the first model for PfHT1 and examined its consistency with our experimental findings. In this model for PfHT1, two regions studied by mutagenesis (Q169N in helix V and 302SGL → AGT in helix VII) are both located in the same plane and are in proximity to the main predicted permeation channel in PfHT1 (Figs. 2 and 3). As the Q169N mutation abolishes fructose transport leaving glucose transport unaffected, perhaps the angular rotation of helix V in this model should be modified so that Gln169 faces the main channel. However, this modification is energetically unfavored (data not shown). Furthermore, our modeling of the Q169N mutant (Fig. 4) reinforces our interpretation that this mutation results in local structural changes (perhaps also involving helix IV), which may lead to the observed relative loss of fructose transport. Alteration of specific hydrogen-bond interactions with fructofuranoses in the mutant may be sufficient to explain the loss of fructose binding while glucopyranose transport is retained, in conjunction with relatively minor changes in helix VII in this mutant. Exofacial binding of GLUT1 was examined with 4,6-O-ethylidene-α-d-glucose (22Mueckler M. Makepeace C. J. Biol. Chem. 1997; 272: 30141-30146Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar), but this compound does not compete for glucose transport by PfHT1 and cannot be used to probe exofacial binding in PfHT1 (5Woodrow C.J. Burchmore R.J.S. Krishna S. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9931-9936Crossref PubMed Scopus (75) Google Scholar). Consequently, analogues of glucose were used to explore hexose permeation further. The helix V mutant showed increases in Km orKi values for 2-DOG, 3-OMG, and 6-DOG and a decreased Ki for 5-thio-d-glucose compared with native sequence PfHT1. A more recent analysis of GLUT1 structure 2Iserovich, P., Wang, L., Ma, L., Yang, H., Zuniga, F. A., Pascual, J. M., Kuang, K., De Vivo, D. C., and Fischbarg, J. (2002) J. Biol. Chem. 277,30991–30997. places the main channel in GLUT1 between helices VII, V, VIII, and X extending to the endofacial loop region. This channel is connected to the auxiliary channel by a prolongation that crosses the endofacial loop and forms an H-like structure connected together around helix VII where the QLS motif plays a critical role by being in contact with both channels. The QLS motif is important in determining fructose versusglucose selectivity in mammalian GLUTs. To confirm that substrate selectivity for glucose versus fructose in PfHT1 is dependent primarily upon helix V, we examined three helix VII mutants (302SGL → AGT, S302A, and L304T) using 2,5-AHM, a fixed furanose analogue of d-fructose, as a sensitive measure of interaction between helix VII and fructofuranoses. There was no significant difference compared with the affinity of 2,5-AHM for native PfHT1. The Km values of all three helix VII mutants ford-glucose are similar to the wild type. Previous studies with PfHT1 showed important hydrogen bonding to glucose at positions C-1 and C-3 (5Woodrow C.J. Burchmore R.J.S. Krishna S. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9931-9936Crossref PubMed Scopus (75) Google Scholar). Studies with the helix VII AGT mutant now show that hydrogen bonding with positions C-5 and C-6 are also important because of higher Ki values for C5 and C6 analogues compared with the Km for d-glucose. Analogue studies for S302A and L304T single mutants establish that the two mutations act synergistically to alter interactions with C-5 and C-6. These mutations therefore appear to result in a different positioning (“footprint”) of glucose in the substrate occupancy site that could be due to protein conformational changes induced by the different physical properties of substituent compared with native amino acids. For example, alanine and threonine are shorter by one carbon atom than serine (Ser302) and leucine (Leu304) that they replace respectively. For S302A there is a loss of a hydroxyl group whereas for the L304T mutant there is a gain of a hydroxyl group. These substitutions may affect hydrophilic as well as hydrophobic interactions in this region. There are clearly fundamental differences in the hexose permeation pathways of P. falciparum PfHT1 and the mammalian GLUTs. From our analysis, it is likely that the capability of PfHT1 to transport fructose has evolved via a different mechanism to that employed by GLUTs 2 and 5. Thus interaction with extracellular hexoses may take place first with helix V (the “fructose” filter) and subsequently with helix VII and the main permeation channel, because the Q169N mutant selectively abolishes fructose transport but preserves glucose transport. Interactions with other C-positions in glucopyranoses are altered less importantly in this mutant. In contrast, in the helix VII mutants, interactions with fructofuranoses are preserved, whereas it is possible to map differences in interactions with certain positions of glucopyranoses compared with native PfHT1. There are also significant structural differences between GLUT1 and PfHT1 exposed by modeling studies, for example the longer and more polar extracellular loop between helices V and VI in PfHT1 compared with GLUT1, the relative proximity of the N- and C-terminal regions and the relative positioning of helix V and helix VII. These modeling studies also point to residues that may line the main permeation channel in several helices (including helices VIII and IX). Mutational studies can test these hypotheses. Furthermore, as selective inhibition of PfHT1 compared with GLUT1 is actively under investigation, working models of PfHT1 and GLUT1 will prove invaluable in advancing inhibitors through the process of drug development. We thank Drs. U. Eckstein-Ludwig, T. Joët, and F. Joubert for discussions.