Characterization of Glu126 and Arg144, two residues that are indispensable for substrate binding in the lactose permease of Escherichia coli

Glu126 and Arg144 in the lactose permease are indispensable for substrate binding and probably form a charge-pair [Venkatesan, P., and Kaback, H. R. (1998) Proc. Natl. Acad. Sci. U.S.A. 95, 9802-9807]. Mutants with Glu126-->Ala or Arg144-->Ala do not bind ligand or catalyze lactose accumulation, efflux, exchange, downhill lactose translocation, or lactose-induced H+ influx. In contrast, mutants with conservative mutations (Glu126-->Asp or Arg144-->Lys) exhibit drastically different phenotypes. Arg144-->Lys permease accumulates lactose slowly to low levels, but does not bind ligand or catalyze equilibrium exchange, efflux, or lactose-induced H+ influx. In contrast, Glu126-->Asp permease catalyzes lactose accumulation and lactose-induced H+ influx to wild-type levels, but at significantly lower rates. Surprisingly, however, no significant exchange or efflux activity is observed. Glu126-->Asp permease exhibits about a 6-fold increase in the Km for active transport relative to wild-type permease with a comparable Vmax. Direct binding assays using flow dialysis demonstrate that mutant Glu126-->Asp binds p-nitrophenyl-alpha,D-galactopyranoside. Indirect binding assays utilizing substrate protection against [14C]-N-ethylmaleimide labeling of single-Cys148 permease reveal an apparent Kd of 3-5 mM for lactose and 15-20 microM for beta, D-galactopyranosyl-1-thio-beta,D-galactopyranoside (TDG). The affinity of Glu126-->Asp/Cys148 permease for lactose is markedly decreased (Kd > 80 mM), while TDG affinity is altered to a much lesser extent (Kd ca. 80 microM). The results extend the conclusion that a carboxylate at position 126 and a guanidinium group at position 144 are irreplaceable for substrate binding and support the idea that Arg144 plays a major role in substrate specificity.     

Cysteine scanning mutagenesis of helix V in the lactose permease of Escherichia coli

Using a functional lactose permease mutant devoid of Cys (C-less permease), each amino acid residue in putative transmembrane helix V was replaced individually with Cys (from Met145 to Thr163). Of the 19 mutants, 13 are highly functional (60-125% of C-less permease activity), and 4 exhibit lower but significant lactose accumulation (15-45% of C-less permease). Cys replacement of Gly147 or Trp151 essentially inactivates the permease (< 10% of C-less); however, previous studies [Menezes, M. E., Roepe, P. D., & Kaback, H. R. (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 1638; Jung, K., Jung, H., et al. (1995) Biochemistry 34, 1030] demonstrate that neither of these residues is important for activity. Immunoblots reveal that all of the mutant proteins are present in the membrane in amounts comparable to C-less permease with the exception of Trp151-->Cys and single Cys154 permeases which are present in reduced amounts. Finally, only three of the single-Cys mutants are inactivated significantly by N-ethylmaleimide (Met145-->Cys, native Cys148, and Gly159-->Cys), and the positions of the three mutants fall on the same face of helix V.     

Ligand-induced movement of helix X in the lactose permease from Escherichia coli: a fluorescence quenching study

Five single-Trp mutants were constructed by replacing Val315, Leu318, Val326, Leu329, or Val331 with Trp in transmembrane helix X of a functional lactose permease mutant devoid of Trp residues (Trp-less permease). Taking into account expression levels, each single-Trp permease except for Val331-->Trp exhibits significant activity. The intrinsic fluorescence emission of each single-Trp mutant does not change significantly after addition of beta-d-galactopyranosyl 1-thio-beta-d-galactopyranoside (TDG), indicating that ligand induces little change in the microenvironment of the Trp residues. However, fluorescence quenching studies with the brominated detergent 7,8-dibromododecyl beta,d-maltoside (BrDM) demonstrate that a Trp residue in place of Val315, Val326, or Val331 becomes less accessible to BrDM in the presence of TDG, while a Trp residue in place of Leu318 or Leu329 becomes more accessible. Acrylamide quenching studies with Leu318-->Trp and Val331-->Trp permeases or 2-(4-maleimidoanilino)naphthalene-6-sulfonic acid (MIANS)-labeled Thr320-->Cys and Glu325-->Cys permeases indicate that positions 318 and 325 also become more accessible to a hydrophobic environment in the presence of TDG, while positions 320 and 331 become less accessible. The findings are consistent with a recently proposed mechanism for energy coupling in lactose permease [Kaback, H. R. (1997) Proc. Natl. Acad. Sci. U.S.A. 94, 5539-5543] in which substrate binding causes a conformational change resulting in movement of Glu325 to a nonpolar environment with a dramatic increase in pKa.     

The substrate-binding site in the lactose permease of Escherichia coli

Site-directed N-ethylmaleimide labeling was studied with Glu-126 and/or Arg-144 mutants in lactose permease containing a single, native Cys residue at position 148 in the substrate-binding site. Replacement of either Glu-126 or Arg-144 with Ala markedly decreases Cys-148 reactivity, whereas interchanging the residues, double-Ala replacement, or replacement of Arg-144 with Lys or His does not alter reactivity, indicating that Glu-126 and Arg-144 are charge-paired. Importantly, although alkylation of Cys-148 is blocked by ligand in wild-type permease, no protection whatsoever is observed with any of the Glu-126 or Arg-144 mutants. Site-directed fluorescence with 2-(4-maleimidoanilino)-naphthalene-6-sulfonic acid (MIANS) in mutant Val-331 --> Cys was also studied. In marked contrast to Val-331 --> Cys permease, ligand does not alter MIANS reactivity in mutant Glu-126 --> Ala/Val-331 --> Cys, Arg-144 --> Ala/Val-331 --> Cys, or Arg-144 --> Lys/Val-331 --> Cys and does not cause either quenching or a shift in the emission maximum of the MIANS-labeled mutants. However, mutation Glu-126 --> Ala or Arg-144 --> Ala and, to a lesser extent, Arg-144 --> Lys cause a red-shift in the emission spectrum and render the fluorophore more accessible to I-. The results demonstrate that Glu-126 and Arg-144 are irreplaceable for substrate binding and suggest a model for the substrate-binding site in the permease. In addition, the findings are consistent with the notion that alterations in the substrate translocation pathway at the interface between helices IV and V are transmitted conformationally to the H+ translocation pathway at the interface between helices IX and X.     

Binding of monoclonal antibody 4B1 to homologs of the lactose permease of Escherichia coli

The conformationally sensitive epitope for monoclonal antibody (mAb) 4B1, which uncouples lactose from H+ translocation in the lactose permease of Escherichia coli, is localized in the periplasmic loop between helices VII and VIII (loop VII/VIII) on one face of a short helical segment (Sun J, et al., 1996, Biochemistry 35;990-998). Comparison of sequences in the region corresponding to loop VII/VIII in members of Cluster 5 of the Major Facilitator Superfamily (MFS), which includes five homologous oligosaccharide/H+ symporters, reveals interesting variations. 4B1 binds to the Citrobacter freundii lactose permease or E. coli raffinose permease with resultant inhibition of transport activity. Because E. coli raffinose permease contains a Pro residue at position 254 rather than Gly, it is unlikely that the mAb recognizes the peptide backbone at this position. Consistently, E. coli lactose permease with Pro in place of Gly254 also binds 4B1. In contrast, 4B1 binding is not observed with either Klebsiella pneumoniae lactose permease or E. coli sucrose permease. When the epitope is transferred from E. coli lactose permease (residues 245-259) to the sucrose permease, the modified protein binds 4B1, but the mAb has no significant effect on sucrose transport. The studies provide further evidence that the 4B1 epitope is restricted to loop VII/VIII, and that 4B1 binding induces a highly specific conformational change that uncouples substrate and H+ translocation.     

Engineering a metal binding site within a polytopic membrane protein, the lactose permease of Escherichia coli

Site-directed excimer fluorescence indicates that Glu269 (helix VIII) and His322 (helix X) in the lactose permease of Escherichia coli lie in close proximity [Jung, K., Jung, H., Wu, J., Privé, G.G., & Kaback, H.R. (1993) Biochemistry 32, 12273]. In this study, Glu269 was replaced with His in wild-type permease, leading to the presence of bis-His residues between helices VIII and X. Wild-type and Glu269-->His permease containing a biotin acceptor domain were purified by monomeric avidin affinity chromatography, and binding of Mn2+ was studied by electron paramagnetic resonance (EPR) spectroscopy. The amplitude of the Mn2+ EPR spectrum is reduced by the Glu269-->His mutant, while no change is observed in the presence of wild-type permease. The Glu269-->His mutant contains a single binding site for Mn2+ with a KD of about 43 microM, and Mn2+ binding is pH dependent with no binding at pH 5.0, stoichiometric binding at pH 7.5, and a midpoint at about pH 6.3. The results confirm the conclusion that helices VIII and X are closely opposed in the tertiary structure of lac permease and provide a novel approach for studying helix proximity, as well as solvent accessibility, in polytopic membrane proteins.     

Fluorescence of native single-Trp mutants in the lactose permease from Escherichia coli: structural properties and evidence for a substrate-induced conformational change

Six single-Trp mutants were engineered by individually reintroducing each of the native Trp residues into a functional lactose permease mutant devoid of Trp (Trp-less permease; Menezes ME, Roepe PD, Kaback HR, 1990, Proc Natl Acad Sci USA 87:1638-1642), and fluorescent properties were studied with respect to solvent accessibility, as well as alterations produced by ligand binding. The emission of Trp 33, Trp 78, Trp 171, and Trp 233 is strongly quenched by both acrylamide and iodide, whereas Trp 151 and Trp 10 display a decrease in fluorescence in the presence of acrylamide only and no quenching by iodide. Of the six single-Trp mutants, only Trp 33 exhibits a significant change in fluorescence (ca. 30% enhancement) in the presence of the substrate analog beta,D-galactopyranosyl 1-thio-beta,D-galactopyranoside (TDG). This effect was further characterized by site-directed fluorescent studies with purified single-Cys W33-->C permease labeled with 2-(4'-maleimidylanilino)-naphthalene-6-sulfonic acid (MIANS). Titration of the change in the fluorescence spectrum reveals a 30% enhancement accompanied with a 5-nm blue shift in the emission maximum, and single exponential behavior with an apparent KD of 71 microM. The effect of substrate binding on the rate of MIANS labeling of single-Cys 33 permease was measured in addition to iodide and acrylamide quenching of the MIANS-labeled protein. Complete blockade of labeling is observed in the presence of TDG, as well as a 30% decrease in accessibility to iodide with no change in acrylamide quenching. Overall, the findings are consistent with the proposal (Wu J, Frillingos S, Kaback HR, 1995a, Biochemistry 34:8257-8263) that ligand binding induces a conformational change at the C-terminus of helix I such that Pro 28 and Pro 31, which are on one face, become more accessible to solvent, whereas Trp 33, which is on the opposite face, becomes less accessible to the aqueous phase. The findings regarding accessibility to collisional quenchers are also consistent with the predicted topology of the six native Trp residues in the permease.     

Tilting of helix I and ligand-induced changes in the lactose permease determined by site-directed chemical cross-linking in situ

The N-terminal six transmenbrane helices (N6) and the C-terminal six transmembrane helices (C6) of lactose permease, each with a single Cys residue, were co-expressed, and cross-linking was studied. The proximity of paired Cys residues in helices I (positions 11, 14, 15, 18, 25, 28, 29, or 32) and VII (positions 227, 231, 232, 234, 235, 238, 239, 241, 242, 245, or 246) was studied by using homobifunctional thiol-specific chemical linkers of different lengths and chemical properties. The results demonstrate that Cys residues on one face of the periplasmic half of helix I (positions 32, 29, 28, or 25) cross-link to Cys residues on one face of the periplasmic half of helix VII (242 or 245). In contrast, no cross-linking is evident with paired Cys residues in the cytoplasmic halves of helices I (positions 11, 14, 15, or 18) and VII (positions 227, 230, 231, 232, 234, 235, 238, or 239). The results indicate that helices I and VII are in close proximity only at their periplasmic halves. Ligand binding decreases cross-linking efficiency of the Cys pair 28/245 or 25/242 with N, N'-o-phenylenedimaleimide (rigid 6 A) and increases efficiency with N,N'-p-phenylenedimaleimide (rigid 10 A) or 1,6-bismaleimidohexane (flexible 16 A), indicating that the inter-thiol distance is about 6 A in the absence of ligand and that ligand binding increases the distance up to 10 A. The inter-thiol distance for Cys pairs 29/245 or 32/245 is less than 6 A in the absence of ligand, and in the presence of ligand, distance increases to between 6 and 10 A. Taken together, the results indicate that ligand binding induces a translational or scissors-like rigid body movement of helix I and/or VII at the periplasmic interface between the helices.     

Proximity of helices VIII (Ala273) and IX (Met299) in the lactose permease of Escherichia coli

Three double-Cys mutant pairs--Ala273-->Cys/Met299-->Cys, Thr266-->Cys/Ile303-->Cys, and Thr266-->Cys/Ser306-->Cys--were constructed in a functional lac permease construct devoid of Cys residues, and the excimer fluorescence or electron paramagnetic resonance (EPR) was studied with pyrene- or spin-labeled derivatives, respectively. After reconstitution into proteoliposomes, excimer fluorescence is observed with mutant Ala273-->Cys/Met299-->Cys, but not with the single-Cys mutants nor with mutants Thr266-->Cys/Ile303-->Cys or Thr266-->Cys/Ser306-->Cys. Furthermore, spin-spin interaction is also observed with mutant Ala273-->Cys/Met299-->Cys, but only after the permease is reconstituted into proteoliposomes. The results provide independent support for the conclusions that helix VIII is close to helix IX and that the transmembrane helices of the permease are more loosely packed in a detergent micelle as opposed to a phospholipid bilayer.     

Binding of ligand or monoclonal antibody 4B1 induces discrete structural changes in the lactose permease of Escherichia coli

By using Cys-scanning mutagenesis with site-directed sulfhydryl modification in situ [Frillingos, S., & Kaback, H. R. (1996) Biochemistry 35, 3950-3956], conformational changes induced by binding of ligand or monoclonal antibody (mAb) 4B1 in the lactose permease of Escherichia coli were studied. Out of 31 single-Cys replacement mutants in helices I, V, VII, VIII, X, or XI, 4B1 binding alters the reactivity of Val238-->Cys (helix VII), Val331-->Cys (helix X), or single-Cys355 (helix XI) permease with N-ethylmaleimide (NEM) in right-side-out membrane vesicles. In addition, site-directed fluorescence spectroscopy shows that mAb 4B1 binding causes position 331 (helix X) in the permease to experience a more hydrophobic environment. In contrast, ligand binding elicits more widespread changes, as evidenced by enhancement of the NEM reactivity of Ala244-->Cys, Thr248-->Cys (helix VII), Thr265-->Cys (helix VIII), Val315-->Cys (helix X), Gln359-->Cys, or Met362-->Cys (helix XI) permease, none of which are altered by 4B1 binding. Furthermore, no effect of 4B1 is observed on the reactivity of Cys148 (helix V), Val264-->Cys, Gly268-->Cys, or Asn272-->Cys (helix VIII), positions which probably make direct contact with substrate. With respect to the N-terminal half of the permease, 4B1 binding causes a small increase in the reactivity of mutants Pro28-->Cys or Pro31-->Cys (helix I), while ligand binding causes much greater increases in reactivity. The findings indicate that 4B1 binding induces a structural change in the permease that is much less widespread than that induced by ligand binding.     

Functional consequences of changing proline residues in the phenylalanine-specific permease of Escherichia coli

The PheP protein is a high-affinity phenylalanine-specific permease of the bacterium Escherichia coli. A topological model based on genetic analysis involving the construction of protein fusions with alkaline phosphatase has previously been proposed in which PheP has 12 transmembrane segments with both N and C termini located in the cytoplasm (J. Pi and A. J. Pittard, J. Bacteriol. 178:2650-2655, 1996). Site-directed mutagenesis has been used to investigate the functional importance of each of the 16 proline residues of the PheP protein. Replacement of alanine at only three positions, P54, P341, and P442, resulted in the loss of 50% or more activity. Substitutions at P341 had the most dramatic effects. None of these changes in transport activity were, however, associated with any defect of the mutant protein in inserting into the membrane, as indicated by [35S]methionine labelling and immunoprecipitation using anti-PheP serum. A possible role for each of these three prolines is discussed. Inserting a single alanine residue at different sites within span IX and the loop immediately preceding it also had major effects on transport activity, suggesting an important role for a highly organized structure in this region of the protein.     

Characterization of the gntT gene encoding a high-affinity gluconate permease in Escherichia coli

During the cooking of meats, several highly mutagenic heterocyclic amines (HCAs) are produced. Three HCAs, IQ, MeIQx, and PhIP have been under study for carcinogenicity in cynomolgus monkeys, and to date, IQ has been shown to be a potent hepatocarcinogen. Concomitantly, the metabolic processing of these HCAs has been examined. Metabolism studies show that the potent hepatocarcinogenicity of IQ is associated with the in vivo metabolic activation of IQ via N-hydroxylation and the formation of DNA adducts. In monkeys undergoing carcinogen bioassay with IQ, N-hydroxylation was confirmed by the presence of the N-hydroxy-N-glucuronide conjugate of IQ in urine. The N-hydroxylation of IQ appears to be carried out largely by hepatic CYP3A4 and/or CYP2C9/10, and not by CYP1A2, an isoform not expressed in liver of this species. Notably MeIQx is poorly activated in cynomolgus monkeys and lacks the potency of IQ to induce hepatocellular carcinoma after a 5-year dosing period. The poor activation of MeIQx appears to be due to the lack of constitutive expression of CYP1A2 and an inability of other cytochromes P450, such as CYP3A4 and CYP2C9/10, to N-hydroxylate the quinoxalines. MeIQx is detoxified in monkeys largely by conjugation with glucuronide at the N-1 position. Although the carcinogenicity of PhIP is not yet known, the metabolic data suggest that PhIP will be carcinogenic in this species. PhIP is metabolically activated in vivo in monkeys by N-hydroxylation, as discerned by the presence of the N-hydroxy-N-glucuronide conjugate in urine, bile, and plasma. PhIP also produces DNA adducts that are widely distributed in tissues. The results from these studies support the importance of N-hydroxylation in the carcinogenicity of HCAs in nonhuman primates and by analogy, the importance of this metabolic activation step in the possible carcinogenicity of dietary HCAs in humans.     

Escherichia coli inorganic pyrophosphatase: site-directed mutagenesis of the metal binding sites

Aspartic acids 65, 67, 70, 97 and 102 in the inorganic pyrophosphatase of Escherichia coli, identified as evolutionarily conserved residues of the active site, have been replaced by asparagine. Each mutation was found to decrease the k(app) value by approx. 2-3 orders of magnitude. At the same time, the Km values changed only slightly. Only minor changes take place in the pK values of the residues essential for both substrate binding and catalysis. All mutant variants have practically the same affinity to Mg2+ as the wild-type pyrophosphatase.     

The thermal stability and domain interactions of the mannitol permease of Escherichia coli. A differential scanning calorimetry study

The thermal stability and domain interactions in the mannitol transporter from Escherichia coli, enzyme IImtl, have been studied by differential scanning calorimetry. To this end, the wild type enzyme, IICBAmtl, as well as IICBmtl and IICmtl, were reconstituted into a dimyristoylphosphatidylcholine lipid bilayer. The changes in the gel to liquid crystalline transition of the lipid indicated that the protein was inserted into the membrane, disturbing a total of approximately 40 lipid molecules/protein molecule. The thermal unfolding profile of EIImtl exhibited three separate transitions, two of which were overlapping, that could be assigned to structural domains in the protein. Treatment with trypsin, resulting in the degradation of the water-soluble part of the enzyme while leaving the binding and translocation capability of the enzyme intact, resulted in a decrease of the Tm and enthalpy of unfolding of the membrane-embedded C domain. This effect was much more apparent in the presence of the substrate but only partly so in the presence of the substrate analog perseitol. These results are consistent with a recently proposed model (Meijberg, W., Schuurman-Wolters, G. K., and Robillard, G. T. (1998) J. Biol. Chem. 273, 7949-7946), in which the B domain takes part in the conformational changes during the substrate binding process.     

Preferential mutagenesis of lacZ integrated at unique sites in the Escherichia coli chromosome

To study the variation in spontaneous mutation frequencies in different chromosomal domains, a mini-Mu-kan-lacZ- transposable element was constructed to insert the lacZ-(Trp570 --> Opal) allele into many different loci in the Escherichia coli chromosome. Papillation on MacConkey lactose plates was used to screen for mini-Mu insertion mutants with elevated levels of spontaneous mutagenesis of lacZop --> LacZ+; candidates were then screened for normal mutation frequencies in other genes. Two different insertion mutants with this enhanced mutagenesis phenotype were isolated from 14000 colonies, and named plm-1 (preferential lacZ mutagenesis) and plm-2. The frequency of LacZ- --> LacZ+ mutations in these plm mutants was over 400-fold higher than that in isogenic strains containing mini-Mu-kan-lacZop insertions at other loci. Six Lac+ reversion (or suppression) mutations obtained from each of the two plm mutants were mapped by P1 transduction and all were found to be linked to the Kan(r) gene in the mini-Mu-kan-lacZop, suggesting that a localized mutagenic event is responsible for the preferential mutagenesis. Furthermore, both the LacZ+ --> LacZ- and Kan(r) --> Kan(s) mutant frequencies of these Lac+ revertants were in the range of 10(-3) to 10(-2), indicating that this putative localized mutagenesis is neither allele nor gene specific. To identify the plm loci, the chromosomal regions flanking the mini-Mu insertion sites were cloned and sequenced. A computer-assisted database search of homologous sequences revealed that the plm-1 locus is identical to the mutS gene; the mini-Mu insertion most probably results in the production of a truncated MutS protein. We suggest that the enhanced lacZ mutation frequency in plm-1 may be associated with an active process involving the putative truncated MutS protein. The DNA sequence of the plm-2 locus matched a putative malate oxidoreductase gene located at 55.5 min of the E. coli chromosome.     

Regulation of the raffinose permease of Escherichia coli by the glucose-specific enzyme IIA of the phosphoenolpyruvate:sugar phosphotransferase system

In enteric bacteria, chromosomally encoded permeases specific for lactose, maltose, and melibiose are allosterically regulated by the glucose-specific enzyme IIA of the phosphotransferase system. We here demonstrate that the plasmid-encoded raffinose permease of enteric bacteria is similarly subject to this type of inhibition.